a system dynamics approach to construction safety … · a system dynamics approach to construction...

AA SSYYSSTTEEMM DDYYNNAAMMIICCSS AAPPPPRROOAACCHH TTOO

CCOONNSSTTRRUUCCTTIIOONN SSAAFFEETTYY CCUULLTTUURREE

By

TThhaannwwaaddeeee CChhiinnddaa

B.Eng, M.Eng

A thesis submitted in fulfillment of the requirements for the award of the degree of

DDooccttoorr ooff PPhhiilloossoopphhyy

From

Centre for Infrastructure Engineering and Management

Griffith School of Engineering

Faculty of Science, Environment, Engineering and Technology

GGrriiffffiitthh UUnniivveerrssiittyy

Gold Coast, Australia

October 2007

A System Dynamics Approach to Construction Safety Culture

i

DDEECCLLAARRAATTIIOONN

This work has not previously been submitted for a degree or diploma in any university.

To the best of my knowledge and belief, the thesis contains no material previously

published or written by another person except where due reference is made in the thesis

itself.

________________

Thanwadee Chinda

October 2007

�


ii

AACCKKNNOOWWLLEEDDGGEEMMEENNTT

This research study would not be possible without the encouragement and assistance of

so many individuals and organizations.

First, and foremost, I thank my supervisor, Professor Sherif Mohamed, for the

opportunity, encouragement, guidance, and support he has given me over the years of

this research.

Thanks also to my associate supervisor, Dr. Rodney Stewart, for his assistance. Special

thanks go to Ms. Sandra Paine and Ms. Mary Ping for their help and support throughout

the study. I also thank Professor Yew-Chaye Loo for his vision in establishing the

School of Engineering, Griffith University, on the Gold Coast Campus. I am here

because of his energy and enthusiasm for engineering research.

I thank Mrs. Pullapha Noinonthong for her help in translating the survey questionnaire

into the Thai language. I am grateful to Mrs. Suangsurat Munka and the Department of

Labour Protection and Welfare, Ministry of Labour, Thailand, for helping me in the

process of data collection. Thanks also to the Thai construction organizations that

contributed their knowledge and experience to the questionnaire survey. Invaluable

assistance in data collection was given by Mr. Kumpanath Kaewtungmuang, Dr.

Wanchai Asvapoositkul, and Lt. Chanaton Surarak; thank you all very much.

I also thank my colleagues at the Griffith School of Engineering, especially Mr. Jaeho

Lee, Ms. Le Chen, and Mr. Kriengsak Panuwatwanich for their help in my data

analysis. I am also appreciative of the editorial assistance given by Ms. Carmel Wild

and Dr. Robyn Heales.


iii

I am indebted to P’Kuum for his support, understanding, and thoughtfulness throughout

my Masters and PhD studies. May we remain good friends forever.

For supporting me personally, I thank P’Wut, P’Duang, P’Krieng, P’Ton, P’Sunan and

Bill, Heather, NaToo, NaNueng, NaNuan, and my friends at the Chiangmai Thai

restaurant. Thank you for making my life in Australia wonderful.

Last, but not least, I would like to dedicate this thesis to my beloved family: my Dad,

my Mom, P’Art, P’Aoy, my aunt, and my grandparents. Thank you for loving me,

believing in me, and giving me the strength to complete this large undertaking – a

successful PhD thesis. I love you all.


iv

LLIISSTT OOFF PPUUBBLLIICCAATTIIOONNSS

The following papers were produced to disseminate the concepts and results of the work

undertaken by the author during the course of this PhD study.

Mohamed, S. and Chinda, T., 2005. Organizational safety culture: a system dynamics

approach. Proceedings of the 4th triennial international conference rethinking and

revitalizing construction safety, health, environment and quality, 17-20 May 2005, Port

Elizabeth, South Africa, 282-292.

Chinda, T. and Mohamed, S., 2006. Modelling construction safety culture using system

dynamics. In: D. Fang, R.M. Choudhry, and J.W. Hinze, eds. Proceedings of the CIB

W99 2006 international conference on global unity for safety and health in

construction, 28-30 June 2006, Beijing, China. Beijing: Tsinghua University Press, 165-

172.

Chinda, T. and Mohamed, S., 2007. Causal relationships between enablers of

construction safety culture. In: S.M. Ahmed, S. Azhar, and S. Mohamed, eds.

Proceedings of the fourth international conference on construction in the 21st century,

11-13 July 2007, Gold Coast, Australia. USA: CITC-IV, 438-445.

Chinda, T. and Mohamed, S., 2008. Structural equation model of construction safety

culture. Engineering, construction and architectural management, 15(2), 114-131.

Chinda, T. and Mohamed, S., 2008. System dynamic modeling of construction safety

culture. The 5th international conference on innovation in architecture, engineering, and

construction, 23-25 June 2008, Antalya, Turkey. Paper summitted for publication and

presentation.

�


v

AABBSSTTRRAACCTT

Throughout the world, the construction industry has had a poor safety record, and is

disproportionately more dangerous when compared to other industries. The major cause

of construction accidents is attributed to unsafe behaviours and work practices, which

are viewed as the direct result of having a poor safety culture. The development of a

mature safety culture has been recognized as a vital element in the achievement of high

standards of safety, alongside an effective safety management system. A better

understanding of how to improve safety culture greatly assists an organization to

allocate appropriate safety resources, and thus improve its overall occupational health

and safety performance.

Recently, researches have been undertaken to measure the ‘health’ of construction

safety culture in an attempt to plan for safety culture improvement. Those studies,

however, have focused neither on the interactions among key safety culture elements,

nor on the consequences of safety initiatives being undertaken over time. Importantly,

construction organizations need to be able to measure their current safety culture

maturity level, and identify areas for safety improvement, to enable them to progress

through to higher maturity levels. Such actions are essential, as the implementation of

safety initiatives that do not address prioritized areas for improvement, may add little

value to the organization in its quest to improve its safety culture, and reduce costs in

the long term.

To address these issues, this study developed a model of construction safety culture, and

investigated the interactions and causal relationships between the five enablers (what

the organization should be doing) and Goals (what the organization aims to achieve),

and their consequences over time. The ‘construction safety culture index’, developed

through modelling construction safety culture, was used to measure the level of

construction safety culture maturity in the organization, and identify areas for safety

improvement.


vi

In developing a construction safety culture model, this study adopted an internationally

recognized performance measurement system, the European Foundation for Quality

Management (EFQM) Excellence model, which consists of five key enablers

(Leadership, Policy and Strategy, People, Partnerships and Resources, and Processes),

to achieve a set of predetermined Goals. It was hypothesized that Leadership drives

(influences) the three enablers (Policy and Strategy, People, and Partnerships and

Resources), which, in turn, collectively influence the achievability of predetermined

Goals through the implementation and improvement of suitable Processes.

To confirm the above hypothesis, and to investigate the proposed relationships among

the Enablers and Goals, the statistical techniques of exploratory factor analysis and

structural equation modelling were performed. A questionnaire survey was used for data

collection purposes. The survey was sent to medium-to-large construction-contracting

organizations operating in Thailand. The targeted respondents were selected on the

assumption that they held senior appointments within their respective organizations.

The results revealed that Leadership was the main driver to effective safety culture, and

that this enabler strongly influenced Policy and Strategy, and People, but had a weak

relationship with Partnerships and Resources. Most of Leadership’s influence on

Partnerships and Resources appeared to be mediated through the People enabler. Thus,

it could be postulated that Thai construction managers focus more on human resources

and teamwork than on the provision of safety resources.

The results also showed that People and Policy and Strategy play a key role in

successful safety implementation. The cooperation of people within the organization,

and an effective safety policy and strategy, therefore, influence effective safety

implementation, which, in turn, enhances Goals achieved by the organization.

To capture the interactions among the five enablers and Goals, over a period of time

(e.g. to examine how feedback of Goals affects the implementation of the Leadership

enabler over time), system dynamics modelling was performed. The causal relationships


vii

obtained from structural equation modelling were used to develop a construction safety

culture dynamic model; it was predicted that a higher enablers’ value would result in a

higher Goals value, and ultimately a higher construction safety culture index.

To plan for safety improvement, an organization can facilitate the testing of alternative

strategies, by simulating the developed dynamic model, to improve its construction

safety culture index, and to progress through to higher maturity levels without actually

having to implement them. This approach reduces any costs that may occur from not

implementing the best strategy.

The cyclical style of safety management was also modelled, to reflect real-life

situations, where management tends to withdraw its attention to safety following the

realisation of excellent performance record.

In conclusion, the developed construction safety culture dynamic model provides

insight into the interactions and influences that each enabler has on improving

construction safety culture over time. The construction safety culture index helps an

organization to assess how well its safety implementation is performed, and provides

guidance on how to plan for safety improvements.


viii

TTAABBLLEE OOFF CCOONNTTEENNTTSS

DDEECCLLAARRAATTIIOONN ................................................................................................................................................................................................................................ ii

AACCKKNNOOWWLLEEDDGGEEMMEENNTT ....................................................................................................................................................................................................iiii

LLIISSTT OOFF PPUUBBLLIICCAATTIIOONNSS ............................................................................................................................................................................................ iivv

AABBSSTTRRAACCTT ...................................................................................................................................................................................................................................... vv

TTAABBLLEE OOFF CCOONNTTEENNTTSS ................................................................................................................................................................................................vviiiiii

LLIISSTT OOFF FFIIGGUURREESS ................................................................................................................................................................................................................ xxiivv

LLIISSTT OOFF TTAABBLLEESS ................................................................................................................................................................................................................ xxvviiiiii

AACCRROONNYYMMSS ................................................................................................................................................................................................................................ xxxxii

CCHHAAPPTTEERR 11

IINNTTRROODDUUCCTTIIOONN .......................................................................................................................................................................................................................... 11

1.1 GENERAL OVERVIEW .......................................................................................1

1.2 THE CONSTRUCTION INDUSTRY ...................................................................1

1.3 SAFETY CULTURE ..............................................................................................5

1.4 SAFETY CULTURE IN CONSTRUCTION ORGANIZATIONS .......................8

1.5 MEASURING SAFETY CULTURE ...................................................................10

1.5.1 Wright et al.’s Safety Culture Improvement Matrix (1999) ...................11

1.5.2 Molenaar et al.’s Characteristics of Safety Culture (2002) ....................12

1.5.3 Mohamed’s Balanced Scorecard for Benchmarking Safety Culture

(2003) .....................................................................................................14

1.6 RESEARCH NEED AND RESEARCH AIMS ...................................................15

1.7 THESIS ORGANIZATION .................................................................................16


ix

CCHHAAPPTTEERR 22

RREESSEEAARRCCHH MMEETTHHOODDOOLLOOGGYY ............................................................................................................................................................................ 1199

2.1 GENERAL OVERVIEW .....................................................................................19

2.2 RESEARCH DESIGN AND RESEARCH FRAMEWORK ...............................19

2.2.1 Performance Measurement Systems: The Review .................................22

2.2.2 The Constructs of the EFQM Excellence Model: The Review ..............23

2.2.3 Data Collection: Questionnaire Survey ..................................................24

2.2.4 Data Screening and Preliminary Analyses .............................................28

2.2.5 Exploratory Factor Analysis: The Introduction ......................................28

2.2.6 Structural Equation Modelling: The Introduction ..................................29

2.2.7 System Dynamics Modelling: The Introduction ....................................30

2.3 SUMMARY ..........................................................................................................41

CCHHAAPPTTEERR 33

LLIITTEERRAATTUURREE RREEVVIIEEWW...................................................................................................................................................................................................... 4433


3.2 PERFORMANCE MEASUREMENT SYSTEMS ..............................................43

3.2.1 The Malcolm Baldrige National Quality Award Framework .................44

3.2.2 The Balanced Scorecard Framework ......................................................48

3.2.3 The European Foundation for Quality Management Excellence Model 50

3.3 SELECTION OF A BASIC FRAMEWORK FOR THE CONSTRUCTION

SAFETY CULTURE ............................................................................................52

3.3.1 A Comparison between the MBNQA Framework and the EFQM

Excellence Model ....................................................................................52

3.3.2 A Comparison between the BSC Framework and the EFQM

Excellence Model ...................................................................................54

3.4 THE PROPOSED CONSTRUCTION SAFETY CULTURE MODEL ..............59


x

3.4.1 Leadership ..............................................................................................60

3.4.2 Policy and Strategy .................................................................................62

3.4.3 People .....................................................................................................63

3.4.4 Partnerships and Resources .....................................................................65

3.4.5 Processes .................................................................................................66

3.4.6 Goals .......................................................................................................67

3.5 SAFETY CULTURE MATURITY MODEL ......................................................71

3.5.1 Safety Culture Maturity Levels ..............................................................71

3.5.2 Scoring Each Maturity Level ..................................................................74

3.6 SUMMARY ..........................................................................................................75

CCHHAAPPTTEERR 44

DDAATTAA CCOOLLLLEECCTTIIOONN AANNDD PPRREELLIIMMIINNAARRYY AANNAALLYYSSEESS ............................................................................................ 7777


4.2 QUESTIONNAIRE SURVEY .............................................................................77

4.3 SAMPLE CHARACTERISTICS .........................................................................79

4.4 DATA SCREENING AND PRELIMINARY ANALYSES ................................84

4.4.1 Handling Missing Data ...........................................................................84

4.4.2 Test of Normality ...................................................................................86

4.4.3 Outliers Test ...........................................................................................88

4.4.4 Scale Reliability (Cronbach’s Alpha)......................................................90

CCHHAAPPTTEERR 55

EEXXPPLLOORRAATTOORRYY FFAACCTTOORR AANNAALLYYSSIISS AANNDD SSTTRRUUCCTTUURRAALL EEQQUUAATTIIOONN

MMOODDEELLLLIINNGG .................................................................................................................................................................................................................................. 9933



xi

5.2 THE EXPLORATORY FACTOR ANALYSIS ..................................................93

5.2.1 Assessment of the Suitability of the Data for the Analysis ....................94

5.2.2 Factor Extraction ....................................................................................95

5.2.3 Factor Rotation and Interpretation ..........................................................95

5.2.4 The EFA Results......................................................................................96

5.2.5 Conclusion of the EFA .........................................................................101

5.3 THE STRUCTURAL EQUATION MODELLING ...........................................103

5.3.1 Measurement Model .............................................................................104

5.3.2 Structural Model ...................................................................................111

5.3.3 Conclusion of the SEM ........................................................................115

CCHHAAPPTTEERR 66

SSYYSSTTEEMM DDYYNNAAMMIICCSS MMOODDEELLLLIINNGG OOFF CCOONNSSTTRRUUCCTTIIOONN SSAAFFEETTYY CCUULLTTUURREE --

MMOODDEELL BBUUIILLDDIINNGG .......................................................................................................................................................................................................... 111177

6.1 GENERAL OVERVIEW ...................................................................................117

6.2 SYSTEM DYNAMICS MODELLING .............................................................117

6.3 CAUSAL LOOP DIAGRAMS OF CONSTRUCTION SAFETY CULTURE .119

6.3.1 Causal Loop Diagram ...........................................................................119

6.3.2 A Causal Loop Diagram of the CSC Index ..........................................122

6.4 CONSTRUCTION SAFETY CULTURE DYNAMIC MODEL ......................127

6.4.1 Leadership Dynamic Model .................................................................127

6.4.2 People Dynamic Model ........................................................................130

6.4.3 Partnerships and Resources Dynamic Model .......................................132

6.4.4 Policy and Strategy Dynamic Model.....................................................133

6.4.5 Processes Dynamic Model ....................................................................134

6.4.6 Goals Dynamic Model ..........................................................................135


xii

6.5 DYNAMIC SIMULATION RESULTS .............................................................137

6.5.1 Base Run Results ..................................................................................137

6.5.2 Base Run Results Examination ............................................................143

6.5.3 Model Verification and Validation .......................................................145

CCHHAAPPTTEERR 77

SSYYSSTTEEMM DDYYNNAAMMIICCSS MMOODDEELLLLIINNGG OOFF CCOONNSSTTRRUUCCTTIIOONN SSAAFFEETTYY CCUULLTTUURREE --

MMOODDEELL AAPPPPLLIICCAATTIIOONNSS ........................................................................................................................................................................................ 115511


7.2 POLICY ANALYSIS .........................................................................................151

7.2.1 Base Run ...............................................................................................152

7.2.2 Policy Experiments between Organizations ‘A’ and ‘B’ .....................160

7.3 THE CYCLICAL STYLE OF SAFETY MANAGEMENT ..............................168

7.3.1 The Dynamic Model of the Cyclical Style of Safety Management ......171

7.3.2 The Simulation Results .........................................................................182

7.3.3 Conclusion of the Cyclical Style of Safety Management .....................186

CCHHAAPPTTEERR 88

SSTTUUDDYY FFIINNDDIINNGGSS AANNDD RREECCOOMMMMEENNDDAATTIIOONNSS FFOORR FFUUTTUURREE RREESSEEAARRCCHH ................ 118899


8.2 MAJOR FINDINGS ...........................................................................................189

8.3 CONTRIBUTIONS TO THE EXISTING BODY OF KNOWLEDGE ............194

8.4 IMPLICATIONS FOR THE THAI CONSTRUCTION INDUSTRY ...............195

8.5 LIMITATIONS AND RECOMMENDATIONS FOR FUTURE RESEARCH 197

8.6 CLOSURE ..........................................................................................................199


xiii

AAppppeennddiixx 11 QQuueessttiioonnnnaaiirree SSuurrvveeyy ...................................................................................................................................................... 220000

AAppppeennddiixx 22 RRaaww DDaattaa ............................................................................................................................................................................................ 221111

AAppppeennddiixx 33 SSttaannddaarrddiizzeedd SSccoorreess ((ZZ--SSccoorreess)) ........................................................................................................................ 222200

AAppppeennddiixx 44 MMeeaassuurreemmeenntt MMooddeell RReessuullttss.................................................................................................................................... 222299

AAppppeennddiixx 55 SSttrruuccttuurraall MMooddeell RReessuullttss .............................................................................................................................................. 224411

AAppppeennddiixx 66 SSDD EEqquuaattiioonnss ooff BBaassee RRuunn SSiimmuullaattiioonn .................................................................................................... 225522

AAppppeennddiixx 77 LLiinneeaarr RReeggrreessssiioonn RReessuullttss .......................................................................................................................................... 225577

AAppppeennddiixx 88 SSeennssiittiivviittyy AAnnaallyysseess RReessuullttss wwhheenn tthhee IInniittiiaall VVaalluuee ooff eeaacchh EEnnaabblleerr

iiss CChhaannggeedd .......................................................................................................................................................................................... 226600

AAppppeennddiixx 99 SSeennssiittiivviittyy AAnnaallyysseess RReessuullttss wwhheenn tthhee EExxttrraa EEffffoorrtt GGiivveenn ttoo IImmpprroovvee

eeaacchh EEnnaabblleerr iiss CChhaannggeedd.............................................................................................................................................. 226655

AAppppeennddiixx 1100 SSDD EEqquuaattiioonnss ooff OOrrggaanniizzaattiioonn ‘‘AA’’.................................................................................................................. 227700

AAppppeennddiixx 1111 SSDD EEqquuaattiioonnss ooff OOrrggaanniizzaattiioonn ‘‘BB’’.................................................................................................................. 227755

AAppppeennddiixx 1122 SSDD EEqquuaattiioonnss ooff tthhee CCyycclliiccaall SSttyyllee ooff SSaaffeettyy MMaannaaggeemmeenntt .................................... 228800

RREEFFEERREENNCCEESS ............................................................................................................................................................................................................................ 228855


xiv

LLIISSTT OOFF FFIIGGUURREESS

Figure 1.1 Traditional concept of culture (Clarke, 1999) ...........................................7

Figure 1.2 Safety culture table model (Ho and Zeta, 2004)........................................9

Figure 1.3 Highest levels of safety culture hierarchy (Molenaar et al., 2002)..........12

Figure 1.4 The people branch (Molenaar et al., 2002)..............................................12

Figure 1.5 The process branch (Molenaar et al., 2002) ............................................13

Figure 1.6 The values branch (Molenaar et al., 2002) ..............................................13

Figure 1.7 Safety culture balanced scorecard (Mohamed, 2003) .............................14

Figure 2.1 Research design .......................................................................................20

Figure 2.2 Research activities and expected outputs ................................................21

Figure 2.3 Construction safety literature

(Adapted from Melville and Goddard, 1996)..........................................24

Figure 2.4 Methods of data collection (Adapted from Kumar, 2005) ......................25

Figure 2.5 AMOS window and its drawing tool icons .............................................30

Figure 2.6 STELLA menus of symbols for creating a model (Ithink, 2003)............36

Figure 3.1 Four basic elements of the MBNQA framework (NIST, 1993) ..............47

Figure 3.2 Four key perspectives of the BSC framework

(Lamotte and Carter, 2000) .....................................................................48

Figure 3.3 The EFQM Excellence model (EFQM, 2000).........................................51

Figure 3.4 Mapping the BSC framework onto the EFQM Excellence model

(Sheffield Hallam University, 2003) .......................................................57

Figure 3.5 Links between the safety management system and the EFQM

Excellence model (Adapted from Mbuya and Lema, 2004) ...................58

Figure 3.6 The proposed CSC model........................................................................59


xv

Figure 3.7 Safety culture maturity model (Lardner et al., 2001) ..............................72

Figure 4.1 Years of experience in the Thai construction industry............................79

Figure 4.2 Years of experience in the present organization......................................80

Figure 4.3 Job titles of the respondents ....................................................................80

Figure 4.4 Safety responsibilities..............................................................................81

Figure 4.5 Safety activities engagement ...................................................................81

Figure 4.6 Formal safety policy in the organization .................................................82

Figure 4.7 Safety performance compared to the national average record ................82

Figure 4.8 The most influential enablers in improving safety culture ......................83

Figure 5.1 Baseline model of the CSC....................................................................102

Figure 5.2 The best-fit measurement model ...........................................................109

Figure 5.3 The best-fit structural model..................................................................112

Figure 5.4 The final CSC model .............................................................................113

Figure 6.1 Basic components of a SD model..........................................................118

Figure 6.2 Basic CSC dynamic model ....................................................................119

Figure 6.3 An example of a causal loop diagram ...................................................120

Figure 6.4 A causal loop diagram of the CSC index ..............................................123

Figure 6.5 A causal loop diagram of the five enablers and Goals ..........................126

Figure 6.6 The CSC dynamic model.......................................................................128

Figure 6.7 Leadership dynamic model....................................................................129

Figure 6.8 People dynamic model ..........................................................................131

Figure 6.9 Partnerships and Resources dynamic model .........................................132

Figure 6.10 Policy and Strategy dynamic model ......................................................133

Figure 6.11 Processes dynamic model......................................................................135

Figure 6.12 Goals dynamic model ............................................................................136

Figure 6.13 The CSC index dynamic model.............................................................137


xvi

Figure 6.14 Graphical results of the Enablers score over time.................................138

Figure 6.15 Graphical results of the Goals score over time......................................139

Figure 6.16 Graphical results of the CSC index over time .......................................139

Figure 6.17 Graphical results of the five enablers over time ....................................143

Figure 6.18 Sensitivity results of the ‘used_lds’ value when its initial value is

changed..................................................................................................146

Figure 6.19 Sensitivity results of the CSC index when the initial value of Lds is

changed..................................................................................................147

Figure 6.20 Sensitivity results of the ‘used_lds’ value when the ‘plds’ value is

changed..................................................................................................148

Figure 6.21 Sensitivity results of the CSC index when the ‘plds’ value is changed 148

Figure 7.1 Graphical results of the five enablers of organization ‘A’ over time ....154

Figure 7.2 Graphical results of the Enablers score of organization ‘A’ over time.155

Figure 7.3 Graphical results of the Goals score of organization ‘A’ over time......155

Figure 7.4 Graphical results of the CSC index of organization ‘A’ over time .......156

Figure 7.5 Graphical results of the five enablers of organization ‘B’ over time ....158

Figure 7.6 Graphical results of the Enablers score of organization ‘B’ over time.159

Figure 7.7 Graphical results of the Goals score of organization ‘B’ over time......159

Figure 7.8 Graphical results of the CSC index of organization ‘B’ over time .......160

Figure 7.9 The accident cycle

(Adapted from NPS Risk Management Division, 2006).......................169

Figure 7.10 The normal accident cycle (Adapted from Jones, 2007) .......................170

Figure 7.11 The CSC index cycle as management withdraws attention to safety....172

Figure 7.12 The dynamic model of the cyclical style of safety management...........173

Figure 7.13 Leadership dynamic model....................................................................174

Figure 7.14 People dynamic model ..........................................................................176


xvii

Figure 7.15 Partnerships and Resources dynamic model .........................................177

Figure 7.16 Policy and Strategy dynamic model ......................................................178

Figure 7.17 Processes dynamic model......................................................................179

Figure 7.18 Goals dynamic model ............................................................................180

Figure 7.19 Graphical results of the Enablers score as the effect of the attention

withdrawal .............................................................................................184

Figure 7.20 Graphical results of the Goals score as the effect of the attention

withdrawal .............................................................................................184

Figure 7.21 Graphical results of the CSC index as the effect of the attention

withdrawal .............................................................................................185


xviii

LLIISSTT OOFF TTAABBLLEESS

Table 1.1 Definitions of safety culture (Adapted from Potter, 2003)........................6

Table 1.2 Safety culture components (Adapted from Flannery, 2001) ....................8

Table 1.3 Explanation of the scoring scale (Wright et al., 1999) ............................11

Table 2.1 Flow diagram model conventions (Morecroft, 1988)..............................36

Table 2.2 Tests of model structure (Forrester and Senge, 1980).............................38

Table 2.3 Tests of model behaviour (Forrester and Senge, 1980)...........................39

Table 2.4 Tests of policy implications (Forrester and Senge, 1980) .......................40

Table 3.1 The MBNQA categories and items (Pannirselvam and Ferguson, 2001)46

Table 3.2 An example of a BSC template (Lamotte and Carter, 2000)...................49

Table 3.3 Similarities and differences of the MBNQA framework and the

EFQM Excellence model.........................................................................53

Table 3.4 Comparison between core concepts, the MBNQA framework, and the

EFQM Excellence model (Tummala and Tang, 1995)............................53

Table 3.5 Comparison between the BSC framework and the EFQM Excellence

model (Otley, 1999).................................................................................55

Table 3.6 High-level comparisons of the BSC framework and the EFQM

Excellence model (Lamotte and Carter, 2000)........................................56

Table 3.7 Six model constructs and their 34 attributes............................................70

Table 4.1 Missing values .........................................................................................85

Table 4.2 Skewness and kurtosis of the 34 attributes..............................................87

Table 4.3 The mean, the 5% trimmed mean, and the standard deviation of

the 34attributes ........................................................................................89

Table 4.4 Internal consistency of the five enablers and Goals ................................91

Table 5.1 Bartlett’s test of sphericity and the KMO index......................................94


xix

Table 5.2 Three factors extracted from the remaining 25 items..............................97

Table 5.3 Two factors extracted from nine items of Factor 1 of Table 5.2 .............98

Table 5.4 Two factors extracted from nine items of Factor 2 of Table 5.2 .............99

Table 5.5 Five factors extracted from the EFA .....................................................100

Table 5.6 Internal consistency of five factors extracted from the EFA.................101

Table 5.7 The GOF indices of the baseline and the best-fit measurement models107

Table 5.8 Square multiple correlations and standardized coefficients of

observed variables .................................................................................110

Table 5.9 The GOF indices of the best-fit structural model ..................................113

Table 5.10 The direct and indirect path coefficients between the five enablers

and Goals ...............................................................................................115

Table 6.1 Simulation results of the five enablers and Goals .................................140

Table 6.2 Simulation results of the enablers, Goals, and CSC index....................141

Table 6.3 Experimentation with extra efforts given to improve the five enablers 145

Table 7.1 Simulation results of the five enablers of organization ‘A’ ..................153

Table 7.2 Simulation results of the Enablers, Goals, and CSC index of

organization ‘A’.....................................................................................154

Table 7.3 Simulation results of the five enablers of organization ‘B’...................157


organization ‘B’.....................................................................................158

Table 7.5 Simulation results of the five enablers of organization ‘A’ with

‘plds’ = 0.1.............................................................................................162


organization ‘A’ with ‘plds’ = 0.1 .........................................................162

Table 7.7 Simulation results of the five enablers of organization ‘A’ with

‘plds’ = 0.2.............................................................................................163


xx


organization ‘A’ with ‘plds’ = 0.2 .........................................................164

Table 7.9 Simulation results of organization ‘A’ with ‘plds’, ‘pppl’, and

‘ppro’ = 0.2 ............................................................................................165

Table 7.10 Simulation results of organization ‘A’ with ‘plds’ = 0.3, and ‘pppl’

and ‘ppro’ = 0.1 .....................................................................................166

Table 7.11 Simulation results of organization ‘A’ with ‘plds’ and ‘ppro’ = 0.2,

and ‘pppl’, ‘pprs’, and ‘ppol’ = 0.1 .......................................................166

Table 7.12 The CSC index of organization ‘B’ when more of effort is given to

enhance each enabler .............................................................................168

Table 7.13 Simulation results of the cyclical style of safety management..............183


xxi

AACCRROONNYYMMSS

ACSNI Advisory Committee on the Safety of Nuclear Installation

AMOS Analysis of Moment Structure

ANZSIC Australian and New Zealand Standard Industrial Classification

BSC Balanced Scorecard

CFA Confirmatory Factor Analysis

CFI Comparative Fit Index

Co_lds_pol Correlation Value between Leadership, and Policy and Strategy

Co_lds_ppl Correlation Value between Leadership and People

Co_lds_prs Correlation Value between Leadership, and Partnerships and Resources

Co_pol_pro Correlation Value between Policy and Strategy, and Processes

Co_ppl_pro Correlation Value between People and Processes

Co_ppl_prs Correlation Value between People, and Partnerships and Resources

Co_pro_goals Correlation Value between Processes and Goals

Co_prs_pol Correlation Value between Partnerships and Resources, and Policy and

Strategy

C.R. Critical Ratio

CSC Construction Safety Culture

dCSC_Index Desired Construction Safety Culture Index

DF Degree of Freedom

DF_goals_pro Decision Fraction between Goals and Processes

DF_pol_lds Decision Fraction between Policy and Strategy, and Leadership

DF_pol_prs Decision Fraction between Policy and Strategy, and Partnerships and

Resources

DF_ppl_lds Decision Fraction between People and Leadership


xxii

DF_pro_pol Decision Fraction between Processes, and Policy and Strategy

DF_pro_ppl Decision Fraction between Processes and People

DF_prs_lds Decision Fraction between Partnerships and Resources, and Leadership

DF_prs_ppl Decision Fraction between Partnerships and Resources, and People

dgoals Desired Goals Value

dlds Desired Leadership Value

dpol Desired Policy and Strategy Value

dppl Desired People Value

dpro Desired Processes Value

dprs Desired Partnerships and Resources Value

DYNAMO Dynamic Models

EFA Exploratory Factor Analysis

EFQM European Foundation for Quality Management

ggoals Gap of Goals

glds Gap of Leadership

GOF Goodness of Fit

gpol Gap of Policy and Strategy

gppl Gap of People

gpro Gap of Processes

gprs Gap of Partnerships and Resources

IFI Incremental Fit Index

INIT Initial Value

KMO Kaiser-Meyer-Olkin

Lds Leadership

MBNQA Malcolm Baldrige National Quality Award

NFI Bentler-Bonett Normed Fit Index


xxiii

NIST National Institute of Standards and Technology

plds Percentage of More Effort Provided to Improve Leadership

Pol Policy and Strategy

Ppl People

ppol Percentage of More Effort Provided to Improve Policy and Strategy

pppl Percentage of More Effort Provided to Improve People

ppro Percentage of More Effort Provided to Improve Processes

pprs Percentage of More Effort Provided to Improve Partnerships and

Resources

Pro Processes

Prs Partnerships and Resources

RFI Relative Fit Index

rgoals Goals Rate

rlds Leadership Rate

rldsf Leadership Rate Fraction

RMSEA Root Mean Square Error of Approximation

rpol Policy and Strategy Rate

rppl People Rate

rpro Processes Rate

rprs Partnerships and Resources Rate

SD System Dynamics

S.E. Standard Error

SEM Structural Equation Modelling

SMC, R2 Square Multiple Correlation

SMS Safety Management System

SPICE Standardized Process Improvement for Construction Enterprises


xxiv

SPSS Statistical Package for Social Science

Stat. Statistics Value

STELLA Structural Thinking Experimental Learning Laboratory with Animation

TLI Tucker-Lewis Index

used_goals Goals Value

used_lds Leadership Value

used_pol Policy and Strategy Value

used_ppl People Value

used_pro Processes Value

used_prs Partnerships and Resources Value

�2 Chi Square

zacci Z-Score of the ‘Number of Accidents’ Item

zaccn Z-Score of the ‘Accountability’ Item

zalgn Z-Score of the ‘Safety and Productivity Alignment’ Item

zawrn Z-Score of the ‘Safety Awareness’ Item

zbnmk Z-Score of the ‘Benchmarking System’ Item

zcmmt Z-Score of the ‘Commitment’ Item

zcomm Z-Score of the ‘Communication’ Item

zcoop Z-Score of the ‘Stakeholders’ Cooperation’ Item

zcost Z-Score of the ‘Cost of Accidents’ Item

zcstm Z-Score of the ‘Customers’ Expectations’ Item

zdocu Z-Score of the ‘Safety Documentation’ Item

zfdbk Z-Score of the ‘Feedback’ Item

zfinc Z-Score of the ‘Financial Resources’ Item

zhmnr Z-Score of the ‘Human Resources’ Item

zhskp Z-Score of the ‘Housekeeping’ Item


xxv

zimge Z-Score of the ‘Industrial Image’ Item

zinit Z-Score of the ‘Safety Initiatives’ Item

zintg Z-Score of the ‘Safety Integration in Business Goals’ Item

zinvm Z-Score of the ‘Workers’ Involvement’ Item

zjstf Z-Score of the ‘Job Satisfaction’ Item

zldbx Z-Score of the ‘Leading by Example’ Item

zmrle Z-Score of the ‘Workforce Morale’ Item

znobm Z-Score of the ‘No-Blame Approach’ Item

zprcp Z-Score of the ‘Shared Perceptions’ Item

zprsp Z-Score of the ‘Work Pressure’ Item

zresc Z-Score of the ‘Safety Resources’ Item

zresp Z-Score of the ‘Safety Responsibilities’ Item

zrisk Z-Score of the ‘Risk Assessment’ Item

zrlsp Z-Score of the ‘Workers’ Relationships’ Item

zsppt Z-Score of the ‘Supportive Environment’ Item

zstnd Z-Score of the ‘Safety Standards’ Item

zswbh Z-Score of the ‘Safe Work Behaviour’ Item

ztrng Z-Score of the ‘Training’ Item

zwkld Z-Score of the ‘Workload’ Item


xxvi

[THIS PAGE IS INTENTIONALLY LEFT BLANK]


1

11 IINNTTRROODDUUCCTTIIOONN

1.1 GENERAL OVERVIEW

Chapter 1 outlines the background to this study. It provides descriptions and

characteristics of the construction industry, and definitions of safety culture, as well as

previous research undertaken to measure safety culture. The research need, research

aims, and thesis organization are also presented.

1.2 THE CONSTRUCTION INDUSTRY

The Australian and New Zealand Standard Industrial Classification (ANZSIC) system

defines the construction industry as “units mainly engaged in construction, repair,

alteration, and renovation of buildings and other structures, and those engaged in

providing building or construction trade services and specific installation activities”

(Australian Bureau of Statistics, 1993).

The construction industry comprises many organizations, including property developers,

architects, engineers, quantity surveyors, accountants, lawyers, contractors,

subcontractors, labourers, and specialist tradespersons. It operates on international,

national, regional, and local scales, with participants ranging from large multinational

organizations to single person operations.

According to Jaafari (1996), the construction industry is different from the

manufacturing industry due to: 1) its fragmented structure; 2) its diffused responsibility;

3) its prototype nature; 4) its influences of public, regulatory agencies, and interest

groups; 5) its transient and itinerant labour force, which is not trained to operate under


2

the quality assurance mode of construction; and 6) its virtual lack of research and

development. These factors are explained below:

� Its fragmented structure: The bulk of construction businesses are normally

generated by a large number of small firms that are less inclined to formal methods

of work and management.

� Its diffused responsibility: In construction projects, many individual professionals

and firms share responsibility for the specification, design, and construction of the

projects.

� Its prototype nature: Construction projects typically resemble the ‘prototype’

products in the manufacturing industry, carrying unique design features, site

characteristics, and functions. Thus, the potential for errors to creep in is always

presented due to the once-off nature of the relevant activities and production

processes.

� Its influences of public, regulatory agencies and interest groups: These influences

will ultimately affect the functions and configurations of the projects, which include

construction methods and associated safeguards to the environment, third party

issues, and beneficiaries.

� Its transient and itinerant labour force, which is not trained to operate under the

quality assurance mode of construction: The training of skilled labour is generally

based on learning how to do the work, not being one’s own inspector to produce

zero defects.

� Its virtual lack of research and development (R&D): Typically, R&D work in

construction is confined to that undertaken by the manufacturers of materials and

components incorporated into the projects. There is little R&D work on lines of

projects, such as commercial buildings as a ‘product line’, or managerial processes

in infrastructure works, etc.

A construction project is a unique task, has a predetermined date of delivery, is subject

to one or several performance goals (such as resource usage and quality), and consists

of a number of complex and/or interdependent activities (Packendorff, 1995). The

projects may vary from simple dwellings to complex structures, and normally involve


3

many changes, such as frequent teamwork rotations, exposure to weather conditions,

and high rates of unskilled workers (Rosenfeld et al., 2006). To better understand the

construction industry, it is critical to understand its unique characteristics. Maloney

(2003) listed construction organization characteristics, namely:

� Construction projects in general, with a relatively short and finite duration, cause

project team stress by focusing on what has to be accomplished.

� Time pressure dictates management decisions and project payment, so delays are not

tolerated.

� Cost, work schedule, and productivity determine project profit.

� Supervision provides explicit direction as to what is to be accomplished, while

individuals performing the work may determine how the work is to be done.

� There is extensive goal clarification performance, i.e. planning, and scheduling.

� Construction organizations have significant differences in complexity, diversity, and

size, and often being temporary, multi-organizations, with as many as 20 different

contractors working on a single project site at one time.

� The principal or prime contractor (the contractor with whom the client has a

contract) subcontracts approximately 80% of the work to specialty contractors.

� A single subcontractor lacks control of the working environment.

� The operative employment of contractor organizations on a specific construction site

may be short and transient because workers are added to, and released from, crews

in response to project schedules.

These characteristics make the construction industry one of the most hazardous

industries, resulting in high rates of severe and fatal work-related accidents (Maloney,

2003). In the United Kingdom (UK), for example, the industry accounts for one third of

all work-related fatalities and, on average, five construction workers are killed every

two weeks, while one member of the public is killed every month by construction

activities (HSC, 2003). In the United States of America (USA), although construction

jobs account for just 5% of the total workforce, they account for more than 17% of

annual workplace deaths (Goetsch, 2003). In Australia, the construction fatality rate in


4

2001-02 was five per 100,000 employees, which made it double the all-industry average

(NOSHC, 2005).

Emerging economies and less developed countries are no exception to high fatality

rates. In Thailand, for example, the rate of accidents and fatalities in the construction

industry is reported as highest of all industries (International Labour Organization,

2005). Additionally, construction workers are five times more likely to suffer a

permanent disability than are those in other industries. In India, one of the world’s fast

growing economies, the construction industry accounts for a major share of work-

related accidents (Damodaran, 2006). A similar trend is experienced in the international

construction ‘hotspot’ – the United Arab Emirates – where construction accidents

dominate work-related accident records (The UAE Ministry of Labour and Social

Affairs, 2001).

Construction accidents cause many human tragedies, de-motivate workers, disrupt site

activities, delay project progress, and adversely affect the overall cost, productivity, and

reputation of the construction industry (Mohamed, 1999). According to Kartam (1997),

construction accidents may arise from a variety of causes, which can generally be

classified as: 1) physical incidents posing hazardous situations; and 2) behavioural

incidents caused by unsafe acts. The latter has been identified as the main cause of

construction accidents (Sawacha et al., 1999), and is viewed by many as the direct result

of having a poor safety culture (Smith and Roth, 1991).

Since poor safety culture can lead to risks to human lives, much attention has been paid,

over the past few years, to organizational safety culture, especially to its definitions,

dimensions, and enablers, as well as to the development of tools for assessing and

monitoring its ‘health’, in order to identify areas for safety performance improvement.

The establishment of a good culture of safety can undoubtedly help organizations to

control and reduce their construction costs, and increase the efficiency of their

operations in the long term (Fung et al., 2005). The next section (Section 1.3) presents

an overview of previous research into safety culture.


5

1.3 SAFETY CULTURE

The term safety culture was first introduced by the International Atomic Energy Agency

(IAEA) following their analysis into the nuclear reactor accident at Chernobyl, Ukraine,

in 1986 (Gadd and Collins, 2002). The identification of a poor safety culture, as a

contributing factor to this accident, led to a large number of studies investigating and

attempting to measure safety culture in a variety of high-hazard industries (Little, 2002).

No single definition of what constitutes a safety culture exists. However, the majority of

research studies commonly describe it as including norms, rules, and behaviours that are

presented with respect to safety, as well as characteristics, beliefs, and values that are

exhibited (Potter, 2003). One of the most widely used safety culture definitions is that

developed by the Advisory Committee on the Safety of Nuclear Installation (ACSNI,

1993). This broad-based definition was based on the findings of a study group on

human factors, and had been adopted for use in this study (see below). Other safety

culture definitions are listed in Table 1.1.

Safety culture is the product of individual and group values, attitudes, perceptions, competencies, and patterns of behaviour that determine the commitment to, and the style and proficiency of, an organization’ s health and safety management. Organizations with a positive safety culture are characterized by communications found on mutual trust, shared perceptions of the importance of safety, and confidences in the efficacy of preventive measures (ACSNI, 1993).


6

Table 1.1 Definitions of safety culture (Adapted from Potter, 2003)

Definition Source

“That observable degree of effort by which all organizational members

direct their attentions/actions towards improving safety on a daily basis.”

Cooper, 2000

“Those aspects of the organizational culture which will impact on attitudes

and behaviours related to increasing or decreasing risk.”

Guldenmund, 2000

“The attitudes, beliefs, and perceptions shared by natural groups as

defining norms and values, which determine how they act and react in

relation to risks and risk control systems.”

Hale, 2000

“The involving perceptions and attitudes, as well as the behaviour of

individuals within an organization.”

Harvey et al., 2002

“The ideas and beliefs that all members of the organization share about

risk, accidents, and ill health.”

Cooper, 2002

“An environmental setting where everyone feels responsible for safety,

and pursues it on a daily basis, going beyond ‘the call of duty’ to identify

unsafe conditions and behaviours, and intervene to correct them... people

‘actively care’ on a continuous basis for safety... (which) is not a priority

that can be shifted depending on situational demands, rather safety is a

value linked with all other situational priorities.”

Geller, 2001

Safety culture is made up of a collection of individual cultures and other subcultures

within the environmental constraints and promotions of an organization (see Figure

1.1). Developing a safety culture needs the cooperation of both individuals and groups

in an organization. It should not be viewed as a separate process, but one that forms an

integrative part of the wider organizational culture (Clarke, 1999).


7

Individual

Culture

Vocational Culture

Organizational Culture

National Culture

Safety Culture

Defined or Manifested

Extrinsic Elements

Intrinsic Elements

Values

Sum

Norms

Beliefs

Assumptions

Rituals

Symbols

Behaviours

Group

Figure 1.1 Traditional concept of culture (Clarke, 1999)

Many researchers have investigated the components of safety culture, finding

similarities between them. However, there appears to be a moderate difference of

opinion about their compositions. Table 1.2 compares four studies of safety culture

components. Management commitment, communication, and training are critical

components in the development of a good safety culture. Organizations thus need to

concentrate on these factors to improve their safety culture.


8

Table 1.2 Safety culture components (Adapted from Flannery, 2001)

Components Zohar (1980) ICAO (1992) Lardner et al. (2000) INEEL (2004)

Management

commitment to safety*

� � � �

Two-way

communication

� � �

Job satisfaction � � �

Safety training* � � � �

Housekeeping � � �

Learning organization � �

Workers’ involvement � �

Shared perceptions

about safety

� �

Safety resources � �

Personal accountability � �

Workable and realistic

safety rules

� � �

Note: * Common to all four studies

1.4 SAFETY CULTURE IN CONSTRUCTION ORGANIZATIONS

Blockley (1995) proposed that positive changes in safety in the construction industry

will not be fully effective until safety culture is improved. A better understanding of

safety culture will help construction organizations to strategically allocate safety

resources, and thus improve their overall occupational health and safety performance on

sites.

Recently, many research studies have been undertaken in the area of construction safety

culture (CSC). Kartam et al. (2000), for example, studied issues, procedures, and


9

problems of construction safety in Kuwait, and concluded that safety culture

improvement, especially in areas such as management training and commitment in

safety, was needed to prevent construction injuries and accidents. Little (2002)

identified 14 key elements across four themes (ownership and commitment, systems and

procedures, training and competence, and communication) to improve safety culture in

the UK construction industry. He stated that leadership ownership and commitment is

the bedrock upon which improvements in construction safety can be built, and that

commitment must be visible to the workers through, for example, senior management

health and safety tours.

�

Ho and Zeta (2004) studied safety in the Hong Kong construction industry, and

established four key cultural factors (environment, behaviour, organization, and person)

that affect the CSC. They concluded that safety culture and its significance vary from

one country to another due to cultural differences. Indeed workers may behave

differently due to their background differences (race, nation, religion, and community).

According to Hofstede (1980), the cultural diffences are based on five value

dimensions; power distance, uncertainty avoidance, individualism versus collectivism,

masculinity versus femininity, and long-term versus short-term focus. For this reason, a

‘safety culture table model’ was developed (see Figure 1.2). It consists of four main

CSC factors (environment, behaviour, organization, and person). It was claimed that an

organization’s safety culture will fail if it lacks the support of any of these factors.

Figure 1.2 Safety culture table model (Ho and Zeta, 2004)


10

Fung et al. (2005) compared safety culture divergences among three levels of

construction personnel (top management, supervisors, and frontline workers). They

revealed that top management and supervisors have significant safety culture

divergences from frontline workers, especially in areas of organizational commitment,

communication, and the reporting of accidents and near misses. The researchers

proposed that managers and supervisors launch safety promotional campaign to raise

safety awareness. Further, open communication was recommended as a way to decrease

safety culture divergences among these three groups.

1.5 MEASURING SAFETY CULTURE

Such research studies, discussed above, highlight the importance of safety culture in the

construction industry. It is apparent that the enhancement of safety culture helps

organizations to reduce the number of accidents, improve the industry’s image, and

enhance safety performance (Kartam et al., 2000; Tang et al., 2003; Teo et al., 2005).

Before attempting to improve safety culture, organizations need to measure their current

safety culture implementation, and then plan for safety improvements. Unfortunately, as

Speirs and Johnson (2002) attested, safety culture is difficult to measure because it is

not a product or an outcome, but rather a process. Consequently, it is necessary to

distinguish between the outcomes of safety management and the process by which it is

acquired. This, however, is not an easy task to perform. Indeed Flin and Mearns (1999)

questioned whether the ‘state of safety’ could truly be measured in organizations. They

argued that safety attitude surveys, and other such quantitative methods of gauging

safety, are descriptive rather than normative. Thus, they saw the need for qualitative

research that includes researchers spending time in organizations they are studying, to

get a feel for the culture, and to understand how people interact throughout

organizations.

Despite the above arguments, a brief summary of the main studies attempting to

measure, benchmark, and present the ‘aggregate’ scores as an indicator of the ‘health’

of organizational safety culture, is presented below.


11

1.5.1 Wright et al.’s Safety Culture Improvement Matrix (1999)

Wright et al. (1999) developed a so-called ‘safety culture improvement matrix’ based on

an internationally recognized business model, the European Foundation for Quality

Management (EFQM) Excellence model. This matrix, which is a self-assessment tool,

may be ‘scored’ in two methods. The first method simply entails a judgment of whether

each and every criterion of the matrix’s nine criteria (leadership, policy and strategy,

people management, resources, processes, customer satisfaction, employee satisfaction,

impact on society, and behavioural results) has been satisfied. This result is obtained by

asking top management to what extent the particular element has been satisfied.

Elements that are wholly satisfied can be coloured green, partly satisfied yellow, and

unsatisfied red.

The second method entails scoring individual questions (from zero to 100 points, based

on the judgment of the assessors), and calculates both specific element and overall

scores. Once a weighted score for each element is obtained, it is added together to

compute the total safety culture score. An explanation of the scoring scale is presented

in Table 1.3.

Table 1.3 Explanation of the scoring scale (Wright et al., 1999)

Score Explanation

0 None or minimal anecdotal evidence of activity on this point

25 Some evidence of activity on this point, such as some aspects of a results element are

measured, some examples of good leadership communications, etc.

50 Evidence of soundly based approach that delivers about half of the examples cited, but

overlooks some important points

75 Evidence of refinement and improvement of activity; results show sustained high

standards of performance and good integration of activity into normal operations and

planning

100 Systematic and refined activity has been totally integrated into normal working

patterns (but is still visible), with results showing that the organization is the best in its

class, and evidence that the activity will be sustained over time. Excellent comparisons

with internal targets


12

1.5.2 Molenaar et al.’s Characteristics of Safety Culture (2002)

Molenaar et al. (2002) identified a total of 31 characteristics of a positive safety culture,

based on a number of construction organizations with outstanding safety records. These

characteristics were grouped into three branches: people, process, and values (see

Figure 1.3). There are 13 characteristics in the ‘people’ branch (see Figure 1.4), 11

characteristics in the ‘process’ branch (see Figure 1.5), and seven characteristics in the

‘values’ branch (see Figure 1.6).

Figure 1.3 Highest levels of safety culture hierarchy (Molenaar et al., 2002)

Figure 1.4 The people branch (Molenaar et al., 2002)

Safety Culture

People Process Values

People

Top Management Field Personnel Subcontractor

Importance

Initiate

Communication

Training

Accountability

Importance Empowerment Safety Personnel Pre-construction

Past Performance

Incentive

Attendance

Importance


13

�

Figure 1.5 The process branch (Molenaar et al., 2002)

Figure 1.6 The values branch (Molenaar et al., 2002)

Within each branch, the characteristics were organized into a hierarchical structure, with

quantifiable questions to operationally measure the characteristics. All the questions are

based on previously proven research, and the results serve as a ‘snap-shot’ assessment

of organizational safety culture.

Process

Safety Plan Training and Education Disincentives

Involvement

Change

Feedback Change

Enforcement

Duration

Consistency

Dedicated Time Effectiveness

Assessment and Change Incentives

Regularity Value

Values

Safety Values Behaviour-Based Safety

Importance

Actions

Responsibility

Length of Employment

Identification and Correction

Participation

Hazard Prevention


14

1.5.3 Mohamed’s Balanced Scorecard for Benchmarking Safety Culture (2003)

Mohamed (2003) adopted a Balanced Scorecard tool (as shown in Figure 1.7) to

benchmark organizational safety culture. He argued that this tool has the potential to

provide a medium by which to translate safety plans and processes into a clear set of

goals, which are, in turn, translated into a system of performance measures. The tool

offers the advantage of providing a mix of objective and subjective performance

measures that can effectively communicate a powerful strategic focus on safety to the

entire organization. It is also conducive to organizational learning by providing

feedback on targets of performance measures that have not been achieved. Further, it

has a number of different, but complementary, perspectives that help enable

organizations to pursue incremental safety performance improvements.

Figure 1.7 Safety culture balanced scorecard (Mohamed, 2003)

Customer Perspective

Goals Measures

Operational Perspective

Goals Measures

Learning Perspective

Goals Measures

Management Perspective

Goals Measures

What must we do to ensure efficient implementation of rules and procedures?

What must management excel at to achieve zero-accident culture?

How do our employees/ project partners/clients see us?

How do we continue to learn and improve?


15

1.6 RESEARCH NEED AND RESEARCH AIMS

The variety of tools, briefly presented above, are an indication of how researches are

rapidly progressing towards the development of a reliable and valid instrument to

measure organizational safety culture. A major shortcoming with these tools, though, is

the inability to appropriately capture and present causal links between what the

organization is doing and what it aims to achieve (in this study called the Enablers and

Goals, respectively). Another element of weakness lies in a lack of understanding about

the interactions among different organizational safety culture enablers, as well as the

extent of their individual, or combined, effects on the organization’s ability to achieve

safety performance improvements. For example, a safety management system, which is

a complex interactive set of enablers, may not function according to what had been

originally planned and predicted for a variety of reasons, e.g. the difference in

perception of safety culture that has the potential to determine how successful the

process of system implementation is (Dedobbeleer and Beland, 1991).

There has also been little examination of the extent to which there is a consensus among

workers and managers regarding the contributions of the identified enablers in

determining perceptions of safety culture. Based on the SPICE (Standardized Process

Improvement for Construction Enterprises) process improvement framework, it is easy

to argue that implementing safety initiatives that are not addressing prioritized areas for

improvement may add little value to the organization in its quest to improve its safety

culture (Sarshar et al., 2000). In other words, organizations should realistically assess

their organizational safety culture maturity level, and progress sequentially through

different levels of cultural maturity.

In summary, then, there is a need to examine the interactions and interrelationships

among the CSC enablers, so that construction organizations are able to better

understand the influences of enablers on safety culture. To meet the need, this study

aims to contribute a greater understanding of the CSC enablers to the construction

industry. The aims are:


16

� Focusing on the interactions among the CSC enablers and Goals, as well as their

consequences over time;

� Investigating causal relationships between those enablers and Goals;

� Providing a model to measure the CSC maturity levels; and

� Identifying areas for improvement in order to progress through to higher maturity

levels.

With the above aims in mind, this study set out to develop a CSC dynamic model to

simulate the interactions and causal relationships between the CSC enablers, and to

predict the influence of each enabler on safety goals, over a period of time. In

developing the CSC dynamic model, this study utilized a system dynamics (SD)

modelling technique to analyse and solve the problems, with a focus on policy analysis

and design. The details of the SD modelling are described in Chapter 6.

1.7 THESIS ORGANIZATION

The thesis is organized into eight chapters. The structure of each chapter is as follows:

� Chapter 1 describes the characteristics of the construction industry, the descriptions

of safety culture, and the attempts to measure safety culture. Gaps and shortcomings

of previous research studies are identified, and the research needs, as well as the

research aims, are stated.

� Chapter 2 outlines the adopted research methodology, including the research design,

the research activities and expected outputs, the literature review relating to the

development of the CSC model, the data collection methods, and an introduction of

the exploratory factor analysis (EFA), the structural equation modelling (SEM), and

the system dynamic (SD) modelling.


17

� Chapter 3 reviews the literatures on performance measurement systems that were

potential basic models for the CSC. The selection was based on a number of criteria.

The EFQM Excellence model, the selected measurement system, was examined in

detail, including its key constructs and their associated attributes. The CSC model

was then proposed, based on the EFQM Excellence model. Lastly, the five levels of

CSC maturity were introduced at the end of this chapter.

� Chapter 4 details the questionnaire survey’s development and data collection. The

data interpretations, data screening, and preliminary analyses were performed to

increase confidence in the data collected.

� Chapter 5 presents the exploratory factor analysis (EFA), using the SPSS program,

to confirm the construct validity of the five enablers of the proposed CSC model.

Structural equation modelling was then performed to investigate the causal

relationships between the six constructs (five enablers and Goals) of the CSC

model. The final CSC model is presented at the end of this chapter.

� Chapter 6 sees the final CSC model being used to develop the CSC dynamic model

utilizing the system dynamics (SD) modelling technique. The developed dynamic

model was verified and validated to increase confidence in the model. The CSC

index, developed through the SD simulation, was used to measure the current CSC

maturity level of the organization.

� Chapter 7 presents a number of simulations, with different safety policies, that were

undertaken to examine different scenarios to enhance the CSC index and progress

through to higher CSC maturity levels. The cyclical style of safety management was

also modelled to reflect real-life situations where management withdraws attention

from safety, which then leads to a reduced CSC index.

� Chapter 8, the final chapter, concludes with the major findings, contributions to the

existing body of knowledge, implications for the Thai construction industry, and

limitations and recommendations for future research.

The next chapter (Chapter 2) outlines the research methodology adopted in this study.


18



19

2 RREESSEEAARRCCHH MMEETTHHOODDOOLLOOGGYY


As discussed in Chapter 1, the shortcomings of previous research studies lie in their not

considering the interactions and causal relationships between what the organization is

doing, and what it aims to achieve. Moreover, no index is available to use in assessing

CSC maturity levels. Thus, this study aims to develop a CSC dynamic model to explain

the interactions among the key constructs (five enablers and Goals) of the CSC. The

CSC index, developed through the SD simulation, was a tool used to help measure the

CSC maturity level, and to address areas for safety improvement.

This chapter presents different research steps/activities used to achieve the research

aims stated in Chapter 1.

2.2 RESEARCH DESIGN AND RESEARCH FRAMEWORK

The research design (shown diagrammatically in Figure 2.1) involved a review of the

safety culture literature to identify the gaps of previous research studies. The research

needs, followed by the research aims, were identified to fill the research gaps. The

research aims required the development of the CSC model. The data was then collected

via a questionnaire survey. Following the data analyses, the CSC dynamic model and

the CSC index were constructed to fulfil the research aims of the study.


20

Figure 2.1 Research design

The ultimate goal of this study was to develop a CSC index for measuring the CSC

maturity level of an organization. To achieve this goal, a number of research activities

and expected outputs were planned (see Figure 2.2). The details of each activity are

described below.

Literature Review

Research Aims

Model Development

Data Collection

Gaps of Previous Research Studies

Data Analyses

Analysis Results

Research Need

Fulfil Research Aims


21

Figure 2.2 Research activities and expected outputs

The Proposed CSC Model Literature Review on the Performance Measurement Systems

(See Chapter 3)

Data Collection

The Baseline Model of the CSC

The Final CSC Model

The CSC Dynamic Model

The Verified and Validated Model

Literature Review on the Constructs of the EFQM Excellence Model

(See Chapter 3)

Exploratory Factor Analysis (EFA) (See Chapter 5)

Structural Equation Modelling (SEM) (See Chapter 5)

System Dynamics (SD) Modelling (See Chapter 6)

Verification and Validation (See Chapter 6)

Activities Outputs

Enablers, Goals, and their Associated Attributes of the Proposed CSC Model

Questionnaire Survey (See Chapter 4)

The CSC Index SD Simulation (See Chapter 7)

Data Screening and Preliminary Analyses (See Chapter 4)

Screened Data


22

2.2.1 Performance Measurement Systems: The Review

To develop the CSC model, this study considered three major performance

measurement systems, including the Malcolm Baldrige National Quality Award

(MBNQA) framework, the Balanced Scorecard (BSC) framework, and the European

Foundation for Quality Management (EFQM) Excellence model (see details in Chapter

3). While these measurement systems are widely used in measuring an organization’s

performance (Caravatta, 1997; Wongrassamee et al., 2003), only one was selected,

however, to be used for the CSC model development.

The selection of a basic measurement system to develop the CSC model was based on a

comparison of the advantages and disadvantages of each system. The EFQM Excellence

model was the measurement system of choice, for the following reasons:

� The EFQM Excellence model, compared with the MBNQA framework, includes

areas of financial performance and impact on society, which, in the construction

industry, are considered important in order to improve the CSC (Wright et al.,

1999).

� Compared to the BSC framework, the EFQM Excellence model covers more aspects

of Partnerships and Resources, Customer Results, and Society Results. Indeed,

Wright et al. (1999) postulated that resources, such as human and financial

resources, are important in the planning of safety improvements. Further, Mohamed

(2003) suggested that the Customer is one of the four perspectives against which to

benchmark organizational safety culture in construction.

� The EFQM Excellence model identified strengths, and areas for improvement,

across the organization’s processes. This capability satisfied identified research

aims.

� The EFQM Excellence model focused on continuous improvement, including the

use of constructive feedback to improve performance. This capability suited the SD

modelling technique that was used in developing the CSC dynamic model.


23

To better understand the EFQM Excellence model, its key constructs, as well as the

attributes associated with each construct, are examined in detail in Chapter 3.

2.2.2 The Constructs of the EFQM Excellence Model: The Review

The EFQM Excellence model comprises nine constructs, including five ‘enablers’ and

four ‘results’. The five ‘enablers’ are Leadership, Policy and Strategy, People,

Partnerships and Resources, and Processes; while the four ‘results’ are People Results,

Customer Results, Society Results, and Key Performance Results. However, the focus of

this study was mainly on the improvements of, and interactions among, enablers’

criteria, to achieve better results, so that the four ‘results’ criteria could be combined

together into a single construct called Goals (see details in Chapter 3). As a result, the

CSC model was proposed; it consists of six constructs (five enablers and a single set of

Goals). Each construct comprises a number of attributes to explain the construct. For

example, the Leadership construct consists of four attributes (top management

commitment, effective two-way communication, management accountability, and

management leading by example), which were selected by their frequent citations

within recent construction safety literature (such as textbooks, scientific journals,

conference proceedings, theses and dissertations, and company reports). This approach,

advocated by Melville and Goddard (1996), is described in Figure 2.3.

A total of 34 attributes covering the six constructs of the CSC were chosen following

their frequent citations (see the attributes in Table 3.7 of Chapter 3). Within the

constructs, Leadership consists of four attributes, Policy and Strategy five attributes,

People seven attributes, Partnerships and Resources four attributes, Processes seven

attributes, and Goals seven attributes. These attributes were used in developing the

questionnaire survey for data collection; the details of the questionnaire survey are

described next.


24

Textbooks

Company Reports

Scientific Journals

Conference Proceedings

Theses and Dissertations

Construction Safety Literature

Good starting place for safety culture literature, and good sources

for the analyses’ methodologies

e.g. Byrne (2001), Coakes and Steed (2003), Ithink (2003), and

Tabachnick and Fidell (2007)

Up-to-date information

e.g. Engineering, construction and architectural management; Journal of

construction engineering and management; and Safety sciences

Gatherings of researchers in particular field, where scientific

results are presented

e.g. Proceedings of the CIB W99 2006 international conference on global unity for safety and health in construction, 28-30 June 2006,

Beijing, China

Up-to-date, and often unpublished, information

e.g. Hsu (2002), Huy (2002), and Ali (2006)

Practical value, and can be easily downloaded from the

internet

e.g. HSC (2003), INEEL (2004), International Labour

Organization (2005), and NOSHC (2005)

Figure 2.3 Construction safety literature (Adapted from Melville and Goddard, 1996)

2.2.3 Data Collection: Questionnaire Survey

2.2.3.1 Survey Content

In conducting a research study, three different types of information (facts, behaviours,

and opinions) may be obtained from survey respondents (Dane, 1990). Facts are

anything that can be verified independently; they are phenomenon or characteristics

available to anyone who knows how to observe them. The information collected from

facts is often called ‘demographic characteristics’ (such as age, gender, name of the

organization, years of experience, and job title). Behaviours are actions completed by a

respondent. Like facts, behaviours can be verified, but only if they are witnessed or


25

indirect evidence is obtained. Opinions are expressions of a respondent’s preference,

feeling, or behavioural intention. They can be objectively measured. In this study,

respondents’ opinions or perceptions on the CSC are used in a number of data analyses.

2.2.3.2 Survey Methods

According to Kumar (2005), there are two major approaches (secondary and primary

sources, see Figure 2.4) to gathering information about a situation, person, problem, or

phenomenon. Information from secondary sources is normally available and needs only

be extracted. Primary data, on the other hand, can be collected by observations,

interviews, or questionnaires.

Methods of Data Collection

Secondary Sources

Documents

�Government publications�Earlier research�Census�Personal records�Client histories�Service records

Primary Sources

Observation Interviewing Questionnaire

Participant

Non-participant

Mailed questionnaire

Collective questionnaire

Structured

Unstructured

Figure 2.4 Methods of data collection (Adapted from Kumar, 2005)


26

In this study, three methods of data collection (observation, interview, and

questionnaire) were initially considered. Observation is a purposeful systematic and

selective way of watching and listening to an interaction as it takes place. Usually, there

are two types of observations: participant and non-participant. In participant

observations, a researcher participates in the activities of the group being observed in

the same manner as its members, with or without their knowing that they are being

observed. Non-participant observations occur, on the other hand, when the researcher

does not get involved in the activities of the group. According to Kumar (2005), the

problems with using observation as a method of data collection may include the

possibility of observer bias, and the misinterpretation of observations.

A personal interview, sometimes called a face-to-face interview, has a major advantage

in that it allows the researcher to record not only verbal responses, but also any facial or

bodily expressions. These nonverbal responses may give the researcher greater insight

into the respondents’ true opinions and beliefs. The other advantages include: 1) the

respondents can ask for the questions to be clarified; 2) the researchers can ask follow-

up questions, if they think they will provide more reliable data; 3) supplementary

material, such as audio/video materials, can be used to increase the respondents

understanding of the questions; and 4) the response rates are generally high. This type

of survey, however, is time-consuming, and it can be very costly. It may also generate

interviewer bias, if the interviewer is not well trained. In addition, participants may be

more likely to give socially desirable responses, because it is deemed appropriate by

society (Jackson, 2003).

A written questionnaire is self-administered, and can be sent through the traditional mail

system or by email. It is important that a mail survey be clearly written and self-

explanatory because no one will be available to answer questions regarding the survey,

once it has been mailed out. Questionnaire surveys have several advantages, for

example, they generally have less sampling bias (a tendency for one group to be

overrepresented in a sample) than personal interviews. They also allow the researcher to

collect data on more sensitive information. Participants, who may be unwilling to

discuss personal information with someone face-to-face, may be willing to answer such

questions in a written survey. This method is usually less expensive and can cover a


27

large geographical area. Further, the participants can take as much time as they need to

answer the questions without feeling the pressure of someone waiting for the answers.

However, the major problems concern the low response rate, and the misinterpretation

of the questions (McBurney, 1994).

In summary, the three methods of data collection offer different advantages and

disadvantages. For example, the observation is the best approach to collect information

required for the researcher interested in the behaviours rather than the perceptions of

individuals. While the personal interview offers the opportunity for questionnaire

clarification, it is costly and time-consuming, and may suffer from interviewer bias.

Finally, the questionnaire survey is less expensive and has no answers bias, but it does

have a poor response rate.

Due to the limitation of costs and time, this study used the questionnaire survey as the

data collection method of choice. The drop-off and collect approach was used (together

with the mailing method) to increase the response rate. Further, with this approach, help

was provided during the drop-off and collect times to clarify any misunderstandings.

The intention of this research was to study the construction industry sector in Thailand.

Medium to large construction-contracting organizations, with staff of 100 or more, were

selected for the sampling. Targeted respondents were selected on the assumption that

they held senior appointments, such as executive directors, managing directors, and

senior project managers, within their respective organizations. To avoid any

misinterpretation or misunderstanding, each question was written in English, with the

Thai translation underneath. The Thai translation was reviewed and corrected by a local

Thai professional translator.

The questionnaire survey included open-ended, partially open-ended, and rating-scale

questions. The open-ended questions related to data, such as the name of the

organization, job title, and working experience. With the partially open-ended questions,

the respondents were asked to choose one of the provided answers that best represented

their beliefs, or to select the ‘other’ option, if the answers provided were not


28

appropriate. The rating-scale questions asked the respondents to indicate their degree of

agreement or disagreement with each statement provided (that represented each attribute

of the CSC) using a five-point Likert-type scale (points 1 and 5 represented strongly

disagree and strongly agree, respectively). The Likert scale was chosen as it measures

the magnitude of the opinion, not simply the direction (McBurney, 1994). The

questionnaire survey and data collection are described in detail in Chapter 4.

2.2.4 Data Screening and Preliminary Analyses

To increase confidence in the data collected, via the questionnaire survey, it was

examined for characteristics such as response rate, working experience of the

respondents, and the position of the respondents (see Chapter 4). The Statistical Package

for Social Sciences (SPSS) program version 11.5 was used to ensure data consistency,

and to allow the results to be meaningfully interpreted. Thus a number of data screening

and preliminary analyses, including the handling of missing data, the normality test, the

outliers test, and the reliability test, were performed. The screened data were then

further analysed using more complex analyses, including the exploratory factor analysis

(EFA), the structural equation modelling (SEM), and the system dynamics (SD)

modelling.

2.2.5 Exploratory Factor Analysis: The Introduction

According to Seo et al. (2004), the exploratory factor analysis (EFA) is a precursor to

the structural equation modelling. In this research, the EFA was performed to gather

information about the interrelationships among a set of attributes, and to yield a factor-

based scale of the CSC. In other words, the EFA was used to confirm the validity of the

five enabler-constructs of the proposed CSC model.


29

The method of principal axis factoring, with an eigenvalue over one, together with the

varimax rotation, was chosen for the analysis. It was expected that the baseline model of

the CSC would be achieved from this analysis (see Chapter 5).

2.2.6 Structural Equation Modelling: The Introduction

To provide further evidence of construct validity and to investigate the causal

relationships between the six constructs of the CSC (five enablers and the single set of

Goals), the baseline model of the CSC was performed using structural equation

modelling (SEM). SEM comprises two main tests: the measurement model, and the

structural model. Firstly, it is important to confirm the measurement model before the

structural model can be conducted. The results of the SEM helped clarify the causal

relationships, as well as the degrees of influence, of the six CSC constructs (see details

in Chapter 5). These clarified relationships were then used in the CSC dynamic model

development (see Chapter 6).

While a number of programs can be used to conduct the SEM (such as AMOS, CALIS,

EQS, LISREL, Mplus, Mx Graph, RAMONA, and SEPATH) (Kline, 2005), the AMOS

program version 6.0 was chosen for this analysis. AMOS (Analysis of Moment

Structure) is a Microsoft Windows program made up of two core modules: AMOS

Graphics and AMOS Basic (Kline, 2005). In AMOS Graphics, the program provides a

graphical user interface through which the user can specify the model by drawing it on

the screen. Further, all other aspects of the analysis are controlled through this interface.

Indeed a complete set of tools is available under AMOS Graphics for drawing,

modifying, or aligning graphical elements of model diagrams. Each tool is represented

by an icon, and performs one particular function. AMOS window, together with its

palette displaying icons, is shown in Figure 2.5 below.


30

Figure 2.5 AMOS window and its drawing tool icons

2.2.7 System Dynamics Modelling: The Introduction

The system dynamics (SD) modelling was used to develop the CSC dynamic model. It

was first introduced by Forrester (1961) as a method for modelling and analysing the

behaviour of complex social systems, particularly in an industrial context. It has been

used to examine various social, economic, and environmental systems, where a holistic

view is important, and feedback loops are critical to the understanding of the

interrelationships (Rodrigues and Bowers, 1996). Simonovic (2005) stated that a SD

simulation approach relies on an understanding of complex interrelationships existing

among different elements within a system. This understanding is achieved by

developing a model that can simulate and quantify the behaviour of the system over

time. Such simulations are considered essential in understanding the dynamics of the

system.


31

However, it is difficult to evaluate a SD, as there are no performance criteria for such an

evaluation (Barnes et al., 2005). Nevertheless, some of its strengths and limitations are

stated below. The strengths of SD applications are that:

� It looks at the policies as well as the processes: SD enables policies to be included

in the model as well as processes, so that any problems with the policies can be

addressed. The type of policy may be formal or informal. High overtime levels, for

example, may result from an informal policy, interpreted as the feeling one has to

work long hours to look good. Such a policy can be included into a SD model.

� It provokes serious systems thinking: The idea of the SD is to look at the problem as

a whole, including those influencing factors that affect the behaviour of the system

(such as cause and effect interrelationships between system variables). The outcome

is a more consistent solution.

� It includes high (qualitative, conceptual) level, as well as low (quantitative,

detailed) level, analysis: SD incorporates both qualitative analyses, such as causal

loop diagrams, and quantitative analyses, incorporating rates and levels. These are

useful as they provide a good basis for decision-making.

Some limitations of SD applications are listed below:

� It may be difficult to apply at detailed levels: This difficulty results from the amount

of mathematical analysis required, especially when it lacks the use of a computer

program.

� It has a problem with the delay time factor: At the conceptual level, there is no

reference to the length of the delay between two elements; putting a ‘delay’ into the

diagram only proves that there is a delay that will affect the outcome. At the

quantitative level, it may be difficult to accurately predict the length of the delay,

which may then affect the simulation result. A number of simulations may have to

be run, with varying delay lengths, to obtain a general idea of what the effects might

be.

� It is difficult to set the boundary of the system: To set the boundary of the system, all

factors that significantly affect the problem must be represented. In practice, it is

hard to judge which factors should be included or excluded.


32

� It has a problem with the time horizon: Two issues relating to this limitation are the

length of time to which the model relates (that is, two years or 50 years), and the

inability to compare the effect of events that occur at different points of time. If the

model relates to a long period of time, it is likely that the system structure will

change during that time, thus making the results invalid.

Despite these limitations, the SD methodology provides a good basis from which to

make decisions. It allows for the interrelationships among important variables, all of

which affect the problem, thus providing a better understanding of the problems, and the

ways in which it can be solved.

2.2.7.1 SD Modelling in Construction

In the construction domain, many researchers have reported SD modelling applications.

Love et al. (2000), for example, developed a SD model to capture the interrelationships

among factors that contribute to design errors and reworks in construction projects. The

developed model helped unravel a series of complex problems into more manageable

interrelated components. It enabled the design and project managers to better understand

the process of design documentation and how design errors occur in construction

projects. The outcome was more effective design management, and an improvement in

the profitability and competitiveness of the design firm.

Chritamara and Ogunlana (2002) developed a SD model to better understand the

interacting nature of the problems inherent in the design-and-build procurement of

construction projects in Thailand. The dynamic model was validated and calibrated for a

typical large design-and-build infrastructure project using the time and cost overrun

problems that were experienced. Extensive simulations, with many policies,

individually and in various combinations, were made via SD modelling to identify the

best policy.


33

Tang and Ogunlana (2003a) used SD modelling to gain insights into the interactions

between a country’s construction market and the organization’s financial, technical, and

managerial capabilities. They also employed the SD methodology to provide a careful

and holistic evaluation of the improvement policies to enhance organizational

performance (Tang and Ogunlana, 2003b).

2.2.7.2 The use of SD Modelling in this Study

The studies referred to above verify the proven usefulness of SD modelling in the

construction industry. Additionally, according to Sterman (1992), SD modelling has a

number of advantages that make it suitable for an organization in developing strategies

and improving performance. Such SD advantages are:

� The methodology was developed to deal with dynamics.

� The model is well suited to representing multiple interdependencies.

� It is the modelling method of choice where there are significant feedback processes.

� More than any other modelling technique, it stresses the importance of non-linearity

in model formulation.

� Both hard (objective oriented, formal, and quantitative) and soft (learning oriented,

intuitive, and qualitative) data can be used in the model.

� The model can be used to deal with future behaviour of the system.

SD modelling was used in this study to capture the interactions and causal relationships

between the five enablers and Goals of CSC, over a period of time. The reasons for its

use are:

� The CSC includes many variables that can give rise to changes, such as changes in

safety policy and strategy, in the workforce, and in the resources. A change,

including its effects, may cause another change (referred to as a dynamic change);


34

this may, in turn, affect the whole system. Thus, SD modelling can be used to deal

with these dynamic changes.

� There is a need to investigate the interactions and causal relationships between the

CSC enablers, and the feedback of Goals on the enablers. SD modelling can be used

to capture these feedback processes.

� Most of the data in this study are soft data, such as top management commitment to

safety, provision of safety resources, risk and hazard assessment, etc. SD modelling

permits the use of soft data in the modelling.

� SD modelling can facilitate testing alternative strategies to improve the CSC in

organizations without actually having to implement them. This saves money by

eliminating costs that may occur from not implementing the best safety strategy.

2.2.7.3 SD Software

A number of software programs have been developed for SD modelling. Eberlein’s

(2007) list, with brief details, is presented below:

� DYNAMO: DYNAMO (Dynamic Models) was the first SD simulation language; for

a long time the language and the field were considered synonymous. Originally

developed by Jack Pugh at Massachusetts Institute of Technology (MIT), the

language was made commercially available from Pugh-Roberts in the early 1960s.

DYNAMO today runs on PC compatibles under DOS/Windows. It provides an

equation based development environment for SD models.

� Powersim (www.powersim.com): In the mid 1980s, the Norwegian government

sponsored research aimed at improving the quality of high school education using

SD models. This project resulted in the development of Mosaic, an object oriented

system aimed primarily at the development of simulation-based games for

education. Powersim was later developed as a Windows based environment for the

development of SD models that also facilitates packaging as interactive games or

learning environments.


35

� Vensim (www.vensim.com): Originally developed in the mid 1980s for use in

consulting projects, Vensim was made commercially available in 1992. It is an

integrated environment for the development and analysis of SD models. Vensim

runs on Windows and Macintosh computers.

� STELLA (Structural Thinking Experimental Learning Laboratory with Animation)

(www.iseesystems.com): Originally introduced on the Macintosh in 1984, the

STELLA software provided a graphically oriented front end for the development of

SD models. The stock and flow diagrams, used in the SD literature, are directly

supported with a series of tools supporting model development. Equation writing is

made through dialog boxes accessible from the stock and flow diagrams. STELLA

is available for Macintosh and Windows computers.

In this study, the CSC dynamic model is formulated using the dynamics software

package ‘STELLA’ (Ithink, 2003) because of its significantly better interface

capabilities. The graphical depictions of the STELLA models, and the ability to quickly

adjust a model and run it on the software, have proven to be an excellent discussion

medium for model enhancement (Morecroft, 1988).

The STELLA software provides the modeller with a menu of symbols for creating a

system diagram on a computer screen (see Figure 2.6 and Table 2.1). Symbols are

selected and moved onto the screen and then connected. Modellers are constrained by

the SD connection rules to produce diagrams, which connect symbols in a set sequence.

In addition to symbols, STELLA provides guidelines for equation formulation. These

guidelines can be thought of as rules for converting symbols, text and words into

algebra (Morecroft, 1988).


36

Stock Flow Converter Connector Decision Process Diamond Paint Brush

Dynamite Ghost

Figure 2.6 STELLA menus of symbols for creating a model (Ithink, 2003)

Table 2.1 Flow diagram model conventions (Morecroft, 1988)

Symbol Name Description safety culture index

Stock A stock can be defined as a structural term for anything

that accumulates, for example, savings in the bank.

enablers

Flow If stocks are bathtubs, then flows are pipes that feed and

drain them.

leadership

Converter If stocks are names of structure and flows are verbs, then

converters are adjectives and adverbs.

enablers

Cloud A cloud is an infinite reservoir representing the map

boundary. The capacity of cloud is so great that it makes

no sense to worry about filling or draining it.

enablers

leadership

Connector A connector is used to link the sectors and converters to

other converters

Cloud

Graph Pad Table Pad

Numeric Display

Text Box Hand

Navigation Arrows

Map/ Model Toggle

Sector Frame

Graphics Frame

Button


37

2.2.7.4 Model Verification and Validation

The development of the CSC dynamic model is described in detail in Chapter 6.

According to Forrester and Senge (1980), the developed dynamic model must be tested

with two important processes: model verification and model validation, to establish

confidence in the soundness and usefulness of the dynamic model. McLucas (2005)

stated that model verification and validation are the two model building processes of

establishing confidence in the dynamic model. Unfortunately, these two terms are often

used interchangeably, which leads to confusion. Verification and validation involve two

distinctly different types of activities, but they are inseparable when it comes to SD

modelling.

According to Rakitin (2001), verification is defined as “the process of determining

whether or not the products of a given phase of SD modelling development cycle fulfil

the requirements established during the previous phase.” Verification involves

designing and applying a sufficiently exhaustive set of tests that measure how models

behave. This behaviour is compared with the modes of behaviour specified for complete

models. Indeed, there is a variety of tests that may be conducted in the process of

verifying a model. Such tests include logical, extreme-value, and mass-balance tests

(McLucas, 2005).

� Logical tests are designed to assure the parametric verification, dimensional

integrity, and unit consistency, correct sequence of calculation, and

stochastic/statistical character.

� Extreme-value tests are designed to assure stability under exposure to extreme

conditions and extreme policies.

� Mass-balance tests are designed to assure that physical flows do not violate the

basic requirement for physical flows into a model to either accumulate or flow out.

Mass-balance must be assured during every time-step of every simulation run.

Rodrigues and Williams (1998) postulated that the primary purpose of model validation

is to ensure that the model captures the general dynamics of the system behaviour, and


38

produces results that are as close as possible to their real occurrences. In order to

validate the model, two validation steps must be undertaken carefully: 1) the feedback

structure must be able to capture the general dynamics of system behaviour; and 2) the

calibration parameters for a specific situation must be as close as possible to their real

occurrences.

A model is considered behaviourally validated if simulation results display similar

behavioural patterns when compared with observed behaviour in a real system.

Structural validity is attained by first evaluating every relationship and feedback loop in

the dynamic hypothesis to ensure that it captures the general dynamic behaviour of a

construction organization. Second, the parameters and equations used in the system

dynamic model are investigated to ensure that the parameters match the effect of

corresponding parts in the real system (Tang and Ogunlana, 2003a). According to

Forrester and Senge (1980), the focus of validating activities are on three types of tests

(model structure, model behaviour, and policy implications), as demonstrated in Tables

2.2 to 2.4, respectively. These tests aim to identify the cause-and-effect mechanisms of

the model.

Table 2.2 Tests of model structure (Forrester and Senge, 1980)

No. Test Explanation

1 Structure Verification Compares the model structure directly with the structure of

the perceived system the model represents.

2 Parameter Verification Compares model parameters to knowledge of a perceived

system to determine if parameters correspond conceptually

and numerically to real life.

3 Extreme Condition Tests Improve the model in the normal operating region by

examining the effects of extreme conditions.

4 Boundary Adequacy Assesses whether a model aggregation is appropriate, and if

the model includes all relevant structure.

5 Dimensional Consistency Uses dimensional analysis to validate model’s rate

equations.


39

Table 2.3 Tests of model behaviour (Forrester and Senge, 1980)


1. Behaviour Reproduction (Focuses on reproducing historical behaviour)

a. Symptom-Generation

Tests

Examine whether a model recreates the symptoms of

difficulty that motivate construction of the model.

b. Frequency Generation and

Relative-Phasing Tests

Focus on periodicities of fluctuations and phase

relationships between variables.

c. Multiple-Model-Tests Consider whether a model is able to generate more than

one mode of observed behaviour.

d. Behaviour-Characteristics Focus on any peculiar shapes in a fluctuating time

series.

2. Behaviour Prediction (Focuses on future behaviour)

a. Pattern-Prediction Tests Examine whether a model generates believable patterns

of future behaviour.

b. Event-Prediction Tests Focus on a particular change in circumstances.

3. Behaviour Anomaly Tests Trace a behavioural anomaly to the elements of the

model structure responsible for the behaviour.

4. Family Member Tests Assess whether a model is generic to the class of

system to which the particular member belongs.

5. Surprise Behaviour Test Highlights behaviour in the real system that has not

been previously recognized.

6. Extreme Policy Involves altering policy statement in an extreme way,

and runs the model to determine dynamic

consequences.

7. Boundary Adequacy Considers whether a model includes the structure

necessary to address the issues for which it is designed.

8. Behaviour Sensitivity

Focuses on the sensitivity of model behaviour to

changes in parameter values. It ascertains whether

plausible shifts in model parameters can cause a model

to fail behaviour tests that are previously passed.


40

Table 2.4 Tests of policy implications (Forrester and Senge, 1980)


1. System Improvement Considers whether policies found beneficial after

working with a model, when implemented, also

improve real-system behaviour. In time, this test

becomes the decisive test, but only as repeated real-

life applications of a model lead overwhelmingly to

the conclusion that the model points the way to

improved policy. In the meantime, confidence in

policy implications of models must be achieved

through other tests.

2. Changed-Behaviour Prediction Asks if a model correctly predicts how behaviour of

the system changes if a governing policy is changed.

3. Boundary Adequacy Examines how modifying the model will alter policy

recommendations arrived at by using the model.

4. Policy Sensitivity Reveals the degree of robustness of model behaviour,

and indicates the degree to which policy

recommendations may be influenced by uncertainty in

parameter values. Such testing can help to show the

risk involved in adopting a model for policymaking.

In this study, the ‘logical’ test had been used for model verification to assure parametric

verification, unit consistency, and correct sequence of calculation. The ‘behaviour

sensitivity’ test, on the other hand, was used to validate the model, as it is widely used in

many research studies (such as Miller, 1990; Duynisveld, 1999; Huy, 2002; and Tang

and Ogunlana, 2003a). The results of the model verification and validation are shown in

Chapter 6. The verified and validated dynamic model was simulated to achieve the CSC

index. This index was then used, together with the CSC maturity levels (see details in

Chapter 3), to assess the current CSC maturity level of the organization, and prioritise

areas for safety improvement. A number of policy analyses were then performed, with

SD modelling, to achieve the best policy the organization could use to enhance its CSC

index. Moreover, the cyclical style of safety management was modelled to reflect real-


41

life situations where management withdraws its attention to safety, which leads to the

decreased CSC index (see details in Chapter 7).

2.3 SUMMARY

This chapter addressed relevant methodological issues, and described particular

appropriate methodologies used in this study. The key elements of the research

methodologies were the research design, and the research activities and expected

outcome frameworks. The research design framework provided the big picture for this

study, and encompassed the research aims and what needed to be done to fulfil those

research aims.

The research activities and expected outcomes framework, on the other hand, described

each step of the study to fulfil the research aims. The research started with a literature

review of performance measurement systems, to achieve a basic measurement system

for the CSC model development. The EFQM Excellence model, which was the selected

measurement system, was examined in detail, including its key constructs and their

associated attributes. The CSC model was then proposed, based on the EFQM

Excellence model. A questionnaire survey was developed for the data collection. The

data collected were screened to increase confidence in the data. A number of data

analyses were performed, including the EFA, the SEM, and the SD modelling, to

achieve a CSC dynamic model. The dynamic model was verified and validated to

ensure its usefulness. The verified and validated model was simulated to ultimately

achieve the CSC index, which was expected to fulfil the research aims.

The next chapter (Chapter 3) describes the critical literature review undertaken in

relation to the development of the CSC model.


42



43

33 LLIITTEERRAATTUURREE RREEVVIIEEWW


This chapter contains two parts: the literature review and the key constructs. The

literature review covered three performance measurement systems (the MBNQA

framework, the Balanced Scorecard framework, and the EFQM Excellence model).

These measurement systems have the potential to be used as a basic measurement

system for the CSC model development. The advantages and disadvantages of each

measurement system were compared to select the most suitable measurement system for

the CSC.

The second part presents the key constructs of the selected performance measurement

system, the EFQM Excellence model. The attributes associated with each construct are

stated, and the CSC model is proposed. Lastly, the five levels of CSC maturity are

introduced in the end of this chapter.

3.2 PERFORMANCE MEASUREMENT SYSTEMS

There are a number of performance measurement systems that can be used in measuring

an organization’s performance. Wongrassamee et al. (2003) grouped these performance

measurement systems into two broad categories: systems that emphasize self-

assessment and systems designed to help managers in measuring and improving

business processes. Each category is briefly described below:

� Systems that emphasize self-assessment consist of:

o The Deming Prize (www.deming.org);


44

o The Malcolm Baldrige National Quality Award (MBNQA) framework

(www.quality.nist.gov); and

o The European Foundation for Quality Management (EFQM) Excellence model

(www.efqm.org).

� Systems designed to help managers in measuring and improving business processes

include:

o The Capability Maturity Matrices (www.sei.cmu.edu\cmm\);

o The Performance Pyramid;

o The Effective Progress and Performance Measurement (EP2M); and

o The Balanced Scorecard (BSC) framework.

Among the measurement systems listed above, Wongrassamee et al. (2003) observed

that the EFQM Excellence model and the BSC framework have received wide publicity,

and have recently been adopted by many organizations worldwide. Caravatta (1997),

however, argued that the MBNQA framework is one of the best tools for measuring

performance, claiming that there are approximately one million copies of this award’s

criteria in circulation, most of which are being used as a self-assessment tool.

In response to the above statements, this study considers three performance

measurement systems (the MBNQA framework, the BSC framework, and the EFQM

Excellence model), as the most potentially useful model for developing a CSC model.

The details of each system are described below.

3.2.1 The Malcolm Baldrige National Quality Award Framework

The MBNQA was established by the US Congress in 1987, to promote quality

awareness and thus improve the competitiveness of US companies. Since then, it has

become an important catalyst for improving competitiveness, and increasing the


45

awareness of quality improvement methods (Pannirselvam and Ferguson, 2001). Its

criteria have been adopted or used as a model at local, state, and national, as well as

international levels. According to the National Institute of Standards and Technology

(NIST) (1993), the MBNQA criteria have three important purposes in strengthening US

competitiveness, viz�

� To help raise quality performance practices and expectations;

� To facilitate communication and sharing among and within organizations of all

types, based upon a common understanding of key quality and operational

performance requirements; and

� To serve as a working tool for planning, training, assessment, and other uses.

It was expected that the award criteria would help organizations to achieve two

competitive goals, including the delivery of ever-improving value to customers, which

would result in an improved marketplace and operational performances.

The core concepts were embodied in the MBNQA framework, and, according to

Cortada and Woods (1994), they consist of seven categories, namely:

� Leadership;

� Information and analysis;

� Strategic quality planning;

� Human resources development and management;

� Management of process quality;

� Quality and operational results; and

� Customer focus and satisfaction.

Table 3.1 presents these categories and the examination items used under each category.


46

Table 3.1 The MBNQA categories and items (Pannirselvam and Ferguson, 2001)

No. MBNQA Category Examination Item

1. Leadership Senior executive leadership

Management for quality

Public responsibility

2. Information and analysis Management of data

Benchmarks

Company level data

3. Strategic quality planning Performance planning process

Performance plans

4. Human resources development and management Human resource plans

Employee involvement

Employee training

Employee performance

Employee well being

5. Management of process quality Design quality

Process management

Support services management

Supplier quality

Quality assessment

6. Quality and operational results Quality results

Operational results

Business results

Supplier quality results

7. Customer focus and satisfaction Customer expectation

Customer relationship management

Commitment to customers

Satisfaction determination

Satisfaction results

Satisfaction comparison

The categories are grouped into four basic elements (driver, system, measures of

progress, and goals), as shown in Figure 3.1. Each element is described below:


47

� Driver represents the ‘leadership’ category. It focuses on how top management

emphasizes quality at all levels, and communicates this emphasis throughout the

organization.

� System consists of processes for meeting the company’s quality goals and

performance requirements. Those processes are measured by ‘information and

analysis’, ‘strategic quality planning’, ‘human resources development and

management’, and ‘management of process quality’ categories.

� Measures of progress is measured by the ‘quality and operational results’ category.

� Goal includes the ‘customer focus and satisfaction’ category that focuses on

customer expectation, customer relationship, commitment to customer, satisfaction

determination, and satisfaction results and comparisons.

Figure 3.1 Four basic elements of the MBNQA framework (NIST, 1993)

Customer focus and satisfaction

Goal

Quality and operational results

Measures of Progress

Driver

Leadership

System

Strategic quality planning

Information and analysis

Human resources development and

management

Management of process quality


48

The criteria, along with the advantages and disadvantages of the MBNQA framework,

were compared with those of the EFQM Excellence model to select a basic framework

for a CSC model development. The details are discussed in Section 3.3.1.

3.2.2 The Balanced Scorecard Framework

The BSC framework was first introduced by David Norton and Robert Kaplan in 1990

(Wongrassamee et al., 2003). It is a system of linked objectives, measures, targets, and

initiatives, that collectively describe the vision (strategy) of an organization, and how

that vision (strategy) can be achieved. It is a tool designed to enable the implementation

of an organization’s strategy, by translating it into concrete and operational terms,

which can be measured, communicated, and driven to ensure decision-making and

action. According to Lamotte and Carter (2000), the BSC framework is organized

across four key perspectives (financial, customer, internal, and learning and growth), as

shown in Figure 3.2. The details of each perspective are described below:

Figure 3.2 Four key perspectives of the BSC framework (Lamotte and Carter, 2000)

Financial perspective

Customer perspective

Internal perspective

Learning and growth perspective

Vision


49

� Financial (or shareholder) perspective describes the financial objectives that need

to be achieved to meet the expectations of the shareholder. It is normally in the form

of market presence, economic return, or asset utilisation.

� Customer (or external) perspective focuses on describing key attributes of the

products/services offering that represent value for the customer from the customer’s

point of view.

� Internal perspective describes the processes and activities, which, if executed at the

highest level of performance, will drive success in meeting financial and customer

objectives.

� Learning and growth (or innovation) perspective is often referred to as the enablers.

Its objectives may focus on developing specific skills and competencies, knowledge,

and information and culture. It represents the foundation of the company and its

future capability.

An example of a completed BSC template is shown in Table 3.2. The comparisons

between the BSC framework and the EFQM Excellence model, as well as the selection

of a basic model for the CSC, are described in Section 3.3.2.

Table 3.2 An example of a BSC template (Lamotte and Carter, 2000)

Perspective Objective1 Measure2 Target3 Initiative4

Financial

Organic revenue

growth

Revenue from

existing businesses

1998:$800m Re-packaging of

existing products

Customer Consumers first

choice brand

Consumer

satisfaction index

1998:7/10 In-store

presentations

Internal Cross-sell our

products

Percent revenue

from new products

1998:15% Train staff on

new product

offerings

Learning and

Growth

Communicate

strategy

Percent of people touched

by communication

road show

1998:60% Road show

around all

factories

Note: 1 Objective: Statement of what must be achieved if the strategy is to be successful and the vision realised. 2 Measure: How success in achieving the objectives will be measured and tracked. 3 Target: The level of performance or rate of improvement needed over a specific time-scale. 4 Initiative: Key action programmes required to achieve objectives.


50

3.2.3 The European Foundation for Quality Management Excellence Model

The EFQM, a membership based, not-for-profit organization, was created in 1988 by 14

leading European businesses (EFQM, 2000). It has a key role to play in enhancing the

effectiveness and efficiency of European organizations, by reinforcing the importance

of quality in all aspects of their business activities, stimulating, and assisting the

development of quality improvement as a basis for their achievement of organizational

excellence (EFQM, 2000). The EFQM Excellence model has been acknowledged as an

effective way for organizations to improve the quality of their processes. Further, the

model has been used in business generally, as well as in specific industries, such as

hospitality, education, and construction (Camison, 1996; Wright et al., 1999; Sheffield

Hallam University, 2003). Empirical evidence suggests that the application of holistic

management models, such as the EFQM Excellence model, has a positive effect on

organizational performance (Kristensen and Juhl, 1999).

The model recognizes that there are many approaches to achieving sustainable

excellence in all aspects of performance. It is based on the premise that “excellent

results with respect to performance, customers, people, and society are achieved through

leadership driving policy and strategy, that is delivered through people, partnerships and

resources, and processes” (EFQM, 2000). The model consists of nine criteria, five of

which are ‘enablers’ and four of which are ‘results’ (see Figure 3.3). Enablers include

Leadership, Policy and Strategy, People, Partnerships and Resources, and Processes,

and results include People, Customer, Society, and Key Performance results. Put

simply, enablers cover what an organization is doing, while results cover what an

organization aims to achieve. Each criterion, in the context of safety management, is

defined below:


51

Leadership

People

Policy and

Strategy

Partnerships and

Resources

ProcessesKey

Performance Results

PeopleResults

SocietyResults

CustomerResults

Enablers Results

Innovation and Learning

Figure 3.3 The EFQM Excellence model (EFQM, 2000)

� Leadership (Lds) describes how leaders develop and facilitate the achievement of

the mission and vision of health and safety, develop values required for long-term

success, implement them by appropriate actions and behaviours, and personally

involve themselves in ensuring that the organization’s safety management system is

developed and implemented.

� Policy and strategy (Pol) describes how an organization implements its mission and

vision of safety via clear stakeholder focused strategies, which are supported by

relevant policies, plans, objectives, targets, and processes.

� People (Ppl) describes how an organization manages, develops, and releases the

knowledge and full potential of its people at an individual, team-based, and

organization-wide level, and plans these activities to support its policies and

strategies and the effective operation of its processes.

� Partnerships and resources (Prs) describes how an organization plans and manages

its external partnerships with project participants and other stakeholders and

resources to support its safety policies and strategies and the effective operation of

its safety-related processes.


52

� Processes (Pro) describes how an organization designs, manages, and improves its

processes to support its policies and strategies, and fully satisfy and generate

increasing value for its customers, employees, and other stakeholders.

� People results look at what an organization is achieving in relation to its own

employees.

� Customer results look at what an organization is achieving in relation to its external

customers (e.g. clients and project participants), and other stakeholders.

� Society results look at what an organization is achieving in relation to a local

community and society as appropriate.

� Key performance results look at what an organization is achieving in relation to its

planned performance.

The advantages and disadvantages of the EFQM Excellence model, along with those of

two other performance measurement systems (the MBNQA and the BSC frameworks),

are presented in the following section.

3.3 SELECTION OF A BASIC FRAMEWORK FOR THE

CONSTRUCTION SAFETY CULTURE

3.3.1 A Comparison between the MBNQA Framework and the EFQM

Excellence Model

The MBNQA framework and the EFQM Excellence model were both developed to

improve quality management (Tummala and Tang, 1995). Both models are results-

oriented, and give maximum weight to customer satisfaction. The MBNQA criteria,

however, do not include financial performance whereas it is included in the EFQM

Excellence model, thus, making it less broad-based than the EFQM Excellence model.

Further, the EFQM Excellence model, by including the impact on society as one of the


53

nine criteria, covers more aspects (such as preservation of global resources) in a more

detailed fashion than the MBNQA framework.

The similarities and differences of these two models are indicated in Table 3.3. Further,

a more detailed comparison of these two models, against the core concepts of strategic

quality management, is shown in Table 3.4.

Table 3.3 Similarities and differences of the MBNQA framework and the EFQM

Excellence model

Criteria The MBNQA

Framework

The EFQM

Excellence Model

Being results-oriented � �

Giving maximum weight to customer satisfaction � �

Including financial performance �

Including the impact on society �

Table 3.4 Comparison between the core concepts, the MBNQA framework, and the

EFQM Excellence model (Tummala and Tang, 1995)

The MBNQA Framework Core Concepts The EFQM Excellence Model

Leadership Leadership Leadership

Information and Analysis Strategic Quality Planning Policy and Strategy

Strategic Quality Planning Continuous Improvement People

Human Resources Development

and Management

People Participation and

Partnership

Partnerships and Resources

Management of Process Quality

Design Quality, Speed and

Prevention

Processes

Quality and Operational Results Customer Focus People Results

Customer Focus and Satisfaction Fact-Based Management Customer Results

Society Results

Key Performance Results


54

The dashed lines indicate the inclusion of more aspects in Partnerships and Resources,

Society Results, and Key Performance Results criteria of the EFQM Excellence model,

thus showing that all core concepts of strategic quality management are embedded in

the EFQM Excellence model.

The above comparisons confirm that the EFQM Excellence model is a more fitting

basic CSC model, for the following reasons:

� Its inclusion of financial performance is one of the main issues of concern to the

construction industry. Tam et al. (2004) identified that a financial policy for safety

management must be considered during an organization’s goal setting to

appropriately distribute the budgets to aid safety activities.

� A society issue is included, whereas it is not stated in the MBNQA framework. In

the construction industry, organizations should pay attention to the local community

and the environment to: 1) reduce any risks that may affect society and the

environment (Little, 2002); 2) reduce social costs of accidents borne by society and

its institutions (Tang et al., 2003); and 3) improve the organization’s image and

increase the society’s perception of the organization’s competence (Wright et al.,

1999).

3.3.2 A Comparison between the BSC Framework and the EFQM Excellence

Model

Otley (1999) compared the similarities and differences of the BSC framework and the

EFQM Excellence model using five central areas of management control systems. The

results are summarized in Table 3.5. It is important to note that both models provide

broad and non-prescriptive templates, meaning that management can assign their own

measures to suit their corporate situations and environment.


55

Table 3.5 Comparison between the BSC framework and the EFQM Excellence model

(Otley, 1999)

Item The BSC Framework The EFQM Excellence Model

Objectives Multiple objectives based on strategy

and emphasize four key perspectives

Multiple objectives based on TQM

principles and emphasize nine

criteria

Strategies and

Plans

Assign strategic measures by using

strategy map to connect each measure

to strategy

Not particularly addressed but all

weighted criteria and weighted

sub-criteria can be used as

guidance

Targets Not addressed due to non-prescriptive

template. Managers are required to

assign target performance levels

Management can set their expected

performance levels

Rewards Suggests that individual compensation

system should be linked to strategic

measures

Requires an appropriate reward and

recognition system, but no explicit

guidance given

Feedback Requires double-loop learning which is

more complicated than single-loop

feedback

The model itself provides feedback

information as a default of the

assessment method

Lamotte and Carter (2000) showed a comparison of these two models (see Table 3.6).

They concluded that the BSC framework was designed to communicate and assess

strategic performance, whereas the EFQM Excellence model focuses on encouraging

the adoption of good practice across all management activities of an organization.


56

Table 3.6 High-level comparisons of the BSC framework and the EFQM Excellence

model (Lamotte and Carter, 2000)

Item The BSC Framework The EFQM Excellence Model

Origins � Performance measurement, value

creation

� TQM

Aspiration and

Benefits Sought

� Performance improvement

� To translate a company’s strategy

into focused, operational and

measurable terms

� Enabling strategic performance

� Performance improvement

� To identify strengths and areas for

improvement across an

organization’s processes to

encourage best management

practice

� Enabling best management

practice

Deliverables � A set of logically linked strategic

objectives with lead and lag

indicators/targets across four

perspectives

� A benchmark and relative

assessment of the quality of an

organization’s processes and

results by assessing/scoring against

nine criteria

Development

Approach

� Strategy driven, workshop based,

iterative, hypothesis driven,

management team involvement,

macro view, future looking

� Set of objectives and

measurement are unique to every

organization

� Step change in performance

� Process driven, self-assessment

fact gathering, data collection,

scoring based, detail oriented,

present focused

� Set of criteria and measurement

areas are the same for all

organizations

� Continuous improvement

Success Factors � Management team level

sponsorship and commitment

� On-going process embedded in

governance processes

� Management team level

sponsorship and commitment

� On-going process embedded in

day-to-day management

From an assessment of the above two tables (Tables 3.5 and 3.6), it is evident that the

EFQM Excellence model is more suitable to be used in developing a CSC model,

specifically: 1) it is able to identify strengths and areas for improvement across an

organization’s processes; this capability satisfies the research aims; 2) it focuses on


57

continuous improvement; and 3) it adopts sets of criteria and measurement areas that are

the same for all organizations, thus making it easy for comparisons.

Further, Sheffield Hallam University (2003) mapped the BSC framework onto the

EFQM Excellence model, and concluded that the EFQM Excellence model covers more

aspects of Partnerships and Resources, Customer Results, and Society Results, while

they are not explicitly addressed in the BSC framework (see Figure 3.4). Indeed, Wright

et al. (1999), Mohamed (2003), and Tang et al. (2003) suggested that these aspects are

important in the planning of safety improvements in the construction industry. This,

thus, confirms the suitability of the EFQM Excellence model, as an optimum approach,

to use in developing a CSC model.

Internal Business Process Perspective•Process management approach •Process measures

Customer Perspective•Student and customer surveys•Student and customer indicators•Academic outcomes

Funding Provider Perspective•Financial performance•Academic outcomes•Audit outcomes•Institutional review outcomes

•Vision and values •Defining and working with stakeholders•Setting up a management system including process management

•Strategy development and implementation•Balanced Scorecard as approach

Innovation, Learning and Growth •Staff satisfaction•People indicators•People development and involvement•Use of RADAR•Self-assessment scores and outcomes

Leadership

People

Policy and

Strategy

Partnerships and

Resources

ProcessesKey

Performance Results

PeopleResults

SocietyResults

CustomerResults

Enablers Results


Figure 3.4 Mapping the BSC framework onto the EFQM Excellence model

(Sheffield Hallam University, 2003)


58

To further confirm the suitability of the EFQM Excellence model to be used as a basic

model for a CSC model development, Mbuya and Lema (2004) investigated the

relationships between the EFQM Excellence and the safety management system (SMS),

and found links between them (see Figure 3.5). These relationships illustrate that the

enablers of the EFQM Excellence model match with the essential elements of the SMS,

proving the appropriateness of the EFQM Excellence model for the CSC model

development.

Figure 3.5 Links between the safety management system and the EFQM Excellence

model (Adapted from Mbuya and Lema, 2004)

The next section describes the development of the proposed CSC model, based on the

EFQM Excellence model criteria.

Safety Management System Elements Enablers of the EFQM Excellence Model

Policy and Objectives

Organization

Management Review

Practices and Procedure

Implementation and Compliance

Verification and Assessment Processes


People

Policy and Strategy

Leadership


59

3.4 THE PROPOSED CONSTRUCTION SAFETY CULTURE MODEL

The EFQM Excellence model is suitable to be used as a basic model for the CSC (see

Section 3.3). As discussed earlier, the model consists of five enablers that help to

achieve four results. In this study, however, the focus has been mainly on the

improvements of, and interactions among, the enablers’ criteria to achieve better results.

For this reason the four ‘results’ criteria were combined together into a single construct

(referred to hereinafter as Goals). The proposed CSC model is shown in Figure 3.6.

Enablers Goals


Leadership(Lds)

(100 points)

People (Ppl)

(90 points)

Policy and Strategy

(Pol)(80 points)


(Prs)(90 points)

Processes(Pro)

(140 points)

Goals(500 points)

Figure 3.6 The proposed CSC model

The proposed CSC model assumes that leadership drives people management, policy

and strategy, as well as resources, and that these three enablers collectively influence

the ability to achieve pre-determined Goals through the implementation and

improvement of suitable processes (see Figure 3.6). The six theoretical constructs (five

enablers and the single set of Goals) represent the basic elements of the proposed

model.


60

In addition to the enablers and Goals, the criterion weights are also an important part of

the model. As shown in Figure 3.6, a total of 1,000 points of the proposed model was

evenly split (500/500) between the enablers and Goals. The 500 points allocated to the

enablers were distributed as follows: 100 points to Leadership, 80 points to Policy and

Strategy, 90 points to People, 90 points to Partnerships and Resources, and 140 points

to Processes (EFQM, 2000). Importantly, this allocation of points among enablers,

reflecting their relative contribution to the achievement of Goals, is an area of much

debate. For practical purposes, however, this study adopted the original enablers’

allocation promoted by the EFQM Excellence model (see Figure 3.6). The Goals

construct, on the other hand, contained 500 points, which represented the aggregate

scores of people, customer, society, and key performance results.

The criterion weights of the above five enablers and Goals were later used as an input

into the development of the CSC dynamic model utilizing the SD modelling technique

(see Chapter 6).

Each construct of the proposed CSC model comprised a number of its associated

attributes, which were carefully selected, with reference to the frequency of citations in

recent construction safety literature. These attributes represented the items used to

operationalise each construct (as described in the questionnaire survey in Chapter 4).

The details of the six constructs, along with their associated attributes, are briefly

described below.

3.4.1 Leadership

Leadership and management commitment to safety is recognized as a fundamental

component of an organization’s occupational health and safety (Lingard and Blismas,

2006). In the area of construction, a number of research papers support leadership as the

main enabler in developing a good safety culture (Little, 2002; Mohamed, 2002;

Molenaar et al., 2002; Teo et al., 2005; Lingard and Blismas, 2006).


61

Leadership can be examined using four associated attributes, namely top management

commitment, effective two-way communication, management accountability, and

management leading by example. A brief description of each is presented below:

1. Top management commitment: Teo et al. (2005) identified that safety culture in

construction organizations was dependent upon the safety commitment of

management and workers towards safety promotions and campaigns.

Organizations, where their top management gave high levels of safety support

and commitment, were found to have better safety performance and safety

records (Hinze and Reboud, 1988; Boonrod et al., 1998; Lingard and Blismas,

2006).

2. Effective two-way communication: Lardner et al. (2001) argued that to progress

through to higher safety culture maturity levels, an organization needed more

face-to-face communications, both formal and informal, between management

and frontline staff. Additionally, Little (2002) stated that two-way

communication was one of the key factors in improving safety culture. Teo and

Fang (2006), likewise, found that to enhance safety performance, safety

information should be passed down from top management to frontline workers.

3. Management accountability: According to Olcott (1997), safety culture was the

responsibility of management, and management held the accountability for

creating an atmosphere where each individual employee understood and

accepted his/her role in preventing accidents. As a safety program cannot be

successful on an individual basic, so that the responsibility to accomplish safety

activities must be transferred from top management to individuals at lower

levels of authority (Aksorn and Hadikusumo, 2007). In addition, all levels of site

management should be evaluated in terms of health and safety to ensure

appropriate accountability (Dias and Coble, 1996).

4. Management leading by example: Management needs to be a role model in how

to behave safely. This approach is an important key to enhancing safety culture;

a lack of proper modelling will lead to employees not taking the development of

a positive safety culture seriously (Dunlap, 2004).


62

3.4.2 Policy and Strategy

The Policy and Strategy enabler refers to how an organization implements its mission

and vision of safety via clear stakeholder focused strategies, which are supported by

relevant policies, plans, objectives, targets, and processes. It consists of five attributes:

1) safety awareness and promotion; 2) alignment of productivity and safety targets; 3)

safety standards, laws, regulations; 4) safety initiatives to improve safety standards; and

5) safety as an integral part of business goal settings. The following briefly describes

these attributes:

1. Safety awareness and promotion (such as rewards, recognitions, and

punishments): Molenaar et al. (2002) included incentives and disincentives as

one of the characteristics of safety culture in construction organizations, and

stated that those incentive rewards could include informational (e.g. feedback),

social (e.g. praise/recognition) and tangible (e.g. bonuses/awards) reinforcement

for desired health and safety behaviour (Lingard and Blismas, 2006). Gibb and

Foster (1996) claimed that construction projects that use safety incentive

schemes demonstrated increased safety performance.

2. Alignment of productivity and safety targets: Potter (2003) proposed that to

enhance a culture of safety, safety should have the same weight as productivity

and profitability when economic decisions are made. Indeed safety rules should

be adhered to even under production pressures (particularly those imposed by

budgetary constraints) (Hinze and Reboud, 1988; Glendon and Litherland,

2001).

3. Safety standards, laws, and regulations: A good safety culture needs realistic

and workable safety rules that are practical in all situations (Glendon and

Litherland, 2001; Aksorn and Hadikusumo, 2006).

4. Safety initiatives to improve safety standards: Safety initiatives should be

proactively planned to continually improve safety standards (Boonrod et al.,

1998; Teo et al., 2005).

5. Safety as an integral part of business goal settings: Ahmed et al. (2004) stated

that safety is a company’s core issue, and it must be given a top priority in the

company’s goals setting. Dunlap (2004), likewise, stated that, in order for an


63

organization to be effective in health and safety performance, leadership must

include health and safety in short and long term business goal settings.

3.4.3 People

Niskanen (1994) stated that it is not just management participation and involvement in

safety activities that is important, but also the extent to which management encourages

the involvement of the workforce. Management must be willing to devolve some

decision-making power to the workforce rather than simply play the more passive role

of recipient. In this way, workers are more likely to take ownership and responsibility

for their safety (Williamson et al., 1997).

The attributes associated with this enabler are shared perceptions about safety, safety

empowerment and responsibilities, supportive environment, workers involvement,

relationships among workers, workload, and work pressure. These attributes are

presented below:

1. Shared perceptions about safety: Employees with good perceptions of safety

tend to participate more in safety activities (Dedobbeleer and Beland, 1991;

Fang et al., 2006).

2. Safety empowerment and responsibilities: Hudson (2001) noted that there are

three main safety cultural developments, one of which is to involve workers in

the task of regulatory compliance, and encourage them to take personal

responsibility. In a positive safety culture environment, leaders provide a

reasonable rationale for a desired outcome, and then empower employees to

customize methods for achieving that outcome (Geller, 2000). According to

O’Dea and Flin (2001), empowering employees may be achieved by involving

them in decision-making and developing safety interventions and safety policy.

3. Supportive environment: Good teamwork is identified as a necessary

characteristic of a good safety culture (Olcott, 1997). In a good supportive


64

environment, workers are responsible for their own safety, as well as for their

fellow workers’ safety (Teo and Fang, 2006).

4. Workers’ involvement: Mohamed (2002) stated that the higher the level of

workers’ involvement in safety matters, the more positive the safety climate.

Aksorn and Hadikusumo (2006) noted that successful safety programs largely

depend on employees’ involvement, as workers tend to support the activities that

they themselves help to create. For this reason, workers should be given the

opportunity to provide input into the design and implementation of safety

programs, such as being a member of the safety committee, reporting hazards

and unsafe practices to supervisors, identifying training needs, and investigating

accidents.

5. Relationships among workers: Cooperation between members and the

coordination of safety systems, particularly on multi-occupied sites, are

important if safety is to be improved (Langford et al., 2000). Construction

workers who continually interact with coworkers also rely on them to a greater

extent to provide a safer work environment (Olcott, 1997). Sites, where

workmates often give suggestions to each other on how to work safely, report

less accident rates and fewer workers’ distress (such as anxiety, frustration, and

job dissatisfaction) (Siu et al., 2004).

6. Workload: Glendon and Litherland (2001) observed that, to develop a positive

safety climate, workload should be reasonably balanced. Siu et al. (2004)

claimed that workers’ perceptions of high-role overloads are associated with an

increased tendency to engage in unsafe acts, thus, as fatigue and over-exertion

set in, workers will lose their focus, which results in unsafe acts (Cohen, 2002).

7. Work pressure: According to Siu et al. (2004), work pressures are caused by

distress, unworkable schedule times, and workforce instability (high turnover).

They proposed that psychological distress (such as anxiety, frustration, and job

dissatisfaction) predicts accident rates. Indeed workers who report more anxiety

report more injuries, and take fewer safety precautions. Time schedules for

completing work projects should be workable and realistic to enhance safety

climate (Glendon and Litherland, 2001; Aksorn and Hadikusumo, 2004).

Workforce stability is another important factor contributing to successful safety

programs (Cohen, 1977). Plants with low accidents usually have a workforce

composition that includes employees who are recruited or retained because they


65

work safely; these work environments also have lower turnover and absenteeism

(Lee, 1998).

3.4.4 Partnerships and Resources

The Partnerships and Resources enabler describes how an organization plans and

manages its external partnerships with project participants and other stakeholders, and

organizes its resources to support its safety policies and strategies, as well as the

effective operation of its safety-related processes. Four attributes associated with this

enabler are project participants and stakeholders’ cooperation, adequacy of financial

resources dedicated to safety, availability of necessary safety-related resources, and

human resources management. These attributes are briefly described below:

1. Project participants and stakeholders’ cooperation: According to Wright et al.

(1999), cultural norms cannot be defined in isolation by management, but must

instead involve all key stakeholders (such as regulators, customers, staff, and

contractors) in decision-makings. This process ensures that those norms are

appropriate and meet the expectations of all parties. An effective safety culture

should be conceived of as an appropriate match between the behaviours, values,

and attitudes of members of the organization with the expectations of

stakeholders.

2. Adequacy of financial resources dedicated to safety: To achieve the

organization’s safety cultural goals, financial resources should be allocated to

aid health and safety policies (such as training, recruiting, and acquiring

information) (Wright et al., 1999).

3. Availability of necessary safety-related resources: Aksorn and Hadikusumo

(2006) proposed that a successful safety implementation could not be

accomplished by lack of safety resources. Sufficient safety resources should be

allocated to carry out day-to-day activities to accomplish short and long term

safety goals (Wangniwetkul, 2007). Satisfactory safety facilities, including tools,

equipment, and information, should be provided to the staff so that they can

implement safety activities safely (Sorensen, 2002).


66

4. Human resources management: An organization should endeavour to have

adequate staff to ensure that the job is performed safely (Oklahoma Department

of Labour, 1998).

3.4.5 Processes

This enabler describes how an organization designs, manages, and improves its

processes to support its policies and strategies, and to fully satisfy and generate

increasing value for its customers, employees, and other stakeholders. It consists of

seven attributes (safety training, risk and hazard assessment, feedback on safety

implementation, adopting a no-blame approach, site layout planning and good

housekeeping, site safety documentations, and having an effective benchmarking

system), as described below:

1. Safety training: Training is a major factor influencing safety levels, as it helps

personnel carry out various activities effectively, establishes a positive safety

attitude, and integrates safety with construction and quality goals (Jaselskis et

al., 1996; Tam et al., 2004; Teo et al., 2005). An organization with a good safety

culture always ensures that its staff are safety aware and properly trained, so that

they understand the consequences of unsafe acts (Lardner et al., 2001; INEEL,

2004).

2. Risk and hazard assessment: Risk assessment, including all potential risks (such

as accidents and injuries, regulatory issues, and environmental releases) should

be included in safety-planned activities (McDougall, 2004; Berg, 2006). As

organizational changes can have major effects on safety performance, an

auditable process of risk identification, analysis, and review is required to

manage those changes, and to maintain safety performance (Taylor, 2003).

3. Feedback on safety implementation: To achieve a good safety culture,

management should foster a climate that encourages feedback, so that

organizations learn from their experiences (ICAO, 1992). Tam et al. (2004)


67

revealed that safety score in an organization, where feedback is given, is higher

than that in organizations where no feedback is given.

4. Adopting a no-blame approach: In a blame-free environment, workers feel that

they are fairly treated and are not blamed when they report safety incidents (as

those incidents are regarded as learning opportunities rather than occasions to

criticize or blame individuals) (Taylor, 2003). Vecchio-Sadus and Griffiths

(2004) supported the approach that a blame-free environment helps enhance the

CSC.

5. Site layout planning and good housekeeping: Cohen (1977) stated that better

housekeeping, more orderly plant operations, and adequate environmental

qualities are expected in the organizations with successful safety experiences.

An inadequate site layout plan may lead to the injury to construction personnel

or the public, along with damage to either property or the environment (Suraji et

al., 2001).

6. Site safety documentations (e.g. documented risk plans, site safety plans, site

accident logbooks, and minutes of site safety meetings): Pasman (2000)

identified the main elements of a safety management system as process

knowledge and documentation, the records of design criteria, and the records of

management decisions. Speirs and Johnson (2002) added that a good safety

culture organization would generate a substantial number of high quality

incident reports.

7. Having an effective benchmarking system: Taylor (2003) proposed a

benchmarking system as one of the key features of a strong safety culture. He

claimed that if the organization stops searching for new ideas of safety

improvements by means of benchmarking and seeking out best practice, there is

a danger that its safety culture will slip backwards.

3.4.6 Goals

Goals, with respect to employees, customers, society, and business performance,

represent the ultimate objectives an organization endeavours to achieve. This construct

is examined under seven attributes, namely level of job satisfaction, safe work


68

behaviour, reduced number of accidents and safety related incidents, exceeded

customers’ expectations, improved industry image and safety standards, higher

workforce morale, and reduced total costs associated with accidents. Each attribute is

briefly described below:

1. Level of job satisfaction: A person with a high level of job satisfaction normally

holds positive attitudes towards the job, which thus, in turn, assists in reducing

work injuries (Paul and Maiti, 2007). To accomplish a higher job satisfaction,

improvement of safety-related issues is required (Grote and Kunzler, 2000).

Those safety improvements may include the encouragement of two-way

communication, adequate provision of safety training, and a supportive

environment in safety matters.

2. Safe work behaviour: Mohamed (2002) proposed that a higher level of safety

climate is positively associated with a higher level of self-reported safe work

behaviours. Further, Fang et al. (2006) highlighted that, by encouraging positive

safety behaviour and reducing negative behaviour, the safety climate of an

organization could be improved.

3. Reduced number of accidents and safety-related incidents: An organization with

positive safety culture usually has an acute awareness of the high-risk, error

prone nature of its work, which may lead to the reduction of accidents

(Sorensen, 2002; McDougall, 2004; Ho and Zeta, 2004). Teo et al. (2005) stated

that when safety aspects are well managed, the frequency of accident

occurrences may be reduced.

4. Exceeded customers’ expectations: Customer perspective represents the product

of safety culture (Mohamed, 2003). This perspective may be perceived by level

of customer satisfaction, customer feedback, and customer’s expectations. To

provide continuous improvement and maintain the standards expected by

customers, this customer perspective is acquired directly through opinions

expressed during meetings, and indirectly through other points of contact

(Wright et al., 1999).

5. Improved the industrial image and safety standards: An organization with a

good safety performance has a better organization image (Tang et al., 2003).

This image may be assessed by measurement of the public trust using attitude


69

surveys, or by public reaction to statements made by an organization on its

safety performance (Wright et al., 1999).

6. Higher workforce morale: Employees are more likely to participate in safety

activities when there is a positive organizational culture within the workplace, as

reflected in good work relations and morale, and level of control over work

(Wallace and Neal, 2000). According to Mohamed (2003), workforce morale

can be enhanced by the recognition of individuals with an excellent safety

performance.

7. Reduced total costs associated with accidents: Little (2002) stated that ignorance

of health and safety commitments leads to economic risks for organizations. The

improvement of safety culture helps reduce the social costs of accidents usually

borne by the society (such as cost of property losses, cost of accidents and

injuries, cost of adverse publicity, and cost of environmental releases), and the

total costs of accidents usually borne by the organization (such as cost of lost

production, plant damage, and lost time through accidents) (Pasman, 2000; Tang

et al., 2003).

The above definitions of the six constructs (five enablers and Goals), and their 34

associated attributes, are summarized in Table 3.7. These constructs, as well as their

attributes, are later used, in this study, to develop a so-called construction safety culture

index (CSC index), which serves as an indicator for assessing the CSC maturity level in

the organization. It is important that an organization be able to assess its current

maturity level, as the type of improvement method, needed to support safety culture

development, differs as safety culture matures (Lardner et al., 2001). Consequently, a

safety improvement method may fail if it is not matched to the maturity of the

organization’s existing safety culture. The next section details the development of a

safety culture maturity model and its five levels of culture maturity.


70

Table 3.7 Six model constructs and their 34 attributes

Construct Attribute

Leadership (Lds) 1. Commitment

2. Communication

3. Accountability

4. Leading by example

Policy and Strategy (Pol) 5. Safety awareness

6. Safety and productivity alignment

7. Safety standards

8. Safety initiatives

9. Safety integration in business goals

People (Ppl) 10. Shared perceptions

11. Safety responsibilities

12. Supportive environment

13. Workers’ involvement

14. Workers’ relationships

15. Workload

16. Work pressure

Partnerships and Resources (Prs) 17. Stakeholders’ cooperation

18. Financial resources

19. Safety resources

20. Human resources

Processes (Pro) 21. Training

22. Risk assessment

23. Feedback

24. No-blame approach

25. Housekeeping

26. Safety documentation

27. Benchmarking system

Goals 28. Job satisfaction

29. Safe work behaviour

30. Number of accidents

31. Customers’ expectations

32. Industrial image

33. Workforce morale

34. Cost of accidents


71

3.5 SAFETY CULTURE MATURITY MODEL

3.5.1 Safety Culture Maturity Levels

Lardner et al. (2001) developed a safety culture maturity model (as shown in Figure 3.7)

based on the capability maturity model, to be used as a tool to assist organizations in

establishing their current level of safety culture maturity, and in identifying actions

required to improve their safety culture. The model consists of 10 elements: 1) visible

management commitment; 2) safety communication; 3) productivity versus safety; 4)

learning organization; 5) health and safety resources; 6) participation in safety; 7)

shared perception about safety; 8) trust between management and staff; 9) industrial

relations and job satisfaction; and 10) safety training. It is likely that an organization

will be at different levels in these 10 elements.

The safety culture maturity model consists of five levels of maturity (emerging,

managing, involving, cooperating, and continually improving). Deciding which level is

most appropriate needs to be based on the average level achieved by the organization or

site being evaluated. It is suggested that organizations progress sequentially through the

five levels, by building on the strengths, and removing the weaknesses of the previous

level. It is, therefore, not advisable for an organization to attempt to jump or skip a

level. For example, it is important for organizations to go through the managing level

before the involving level, as it is important that managers develop their commitment to

safety and understand the need to involve frontline employees.


72

Em erging Level 1

Cooperating Level 4

Involving Level 3

M anaging Level 2

Continually improving

Level 5

Impr

ovin

g sa

fety

cul

ture

Incr

easin

g co

nsist

ency

Engage all staff to develop cooperation and commitment to

improving safety

Develop consistency and fight complacency

Realise the importance of frontline staff and develop personal

responsibility

Develop management commitment

Figure 3.7 Safety culture maturity model (Lardner et al., 2001)

A brief description of each maturity level is described below.

� Level 1 – Emerging level: At this first level, safety is defined in terms of technical

and procedural solutions and compliance with regulations. Safety is not seen as a

key business risk, and the safety department is perceived to have primary

responsibility for safety. Many accidents are seen as an unavoidable, and as a part of

the job. Most frontline staff are uninterested in safety, and may only use safety as

the basis for other arguments, such as changes in shift systems.

� Level 2 – Managing level: At this level, safety is seen as a business risk, and

management time and effort is put into accident prevention. Safety is solely defined

in terms of adherence to rules and procedures, and engineering controls. Accidents

are seen as preventable. Managers perceive that the majority of accidents are solely

caused by unsafe behaviours of frontline staff. Safety performance is measured in

terms of lagging indicators (such as lost time injuries), while safety incentives are

based on reducing those lagging indicator rates. In addition, senior managers are

reactive in their involvement in health and safety, i.e. they use punishment when

accident rates increase.


73

� Level 3 – Involving level: At this level, the organization is convinced that the

involvement of frontline staff in health and safety is critical if future improvements

are going to be achieved. Managers recognize that wide ranges of factors cause

accidents, and that the root causes often originate from management decisions. A

significant proportion of frontline staff are willing to work with management to

improve health and safety. Further, majority of staff accept personal responsibility

for their own health and safety. Safety performance is actively monitored, and the

data is used effectively.

� Level 4 – Cooperating level: At this level, the majority of staff in the organization is

convinced that health and safety is important from both a moral and economic point

of view. Managers and frontline staff recognize that wide ranges of factors cause

accidents, and that the root causes are likely to come back to management decisions.

Frontline staff accept personal responsibility for their own, and others, health and

safety. The importance of all employees feeling valued and treated fairly is

recognized. The organization puts a significant effort into proactive measures to

prevent accidents. Additionally, safety performance is actively monitored using all

data available. Non-work accidents are also monitored, and a healthy lifestyle is

promoted.

� Level 5 – Continually improving level: At this final level, the prevention of all

injuries or harm to employees (both at work and at home) is a core company value.

The organization uses a range of indicators to monitor performance, but it is not

performance-driven, as it has confidence in its safety processes. The organization is

constantly striving to improve, and find better ways of improving hazard control

mechanisms. All employees share the belief that health and safety is a critical aspect

of their job, and accept that the prevention of non-work injuries is important. The

company invests considerable effort in promoting health and safety at home.

The safety culture maturity model helps senior management in planning and designing a

safety culture improvement initiative appropriated to their local needs and

circumstances. It provides a systematic process to help senior managers understand key

organizational and behavioural aspects of safety, prioritize areas for safety

improvement, and plan how to make those improvements.


74

3.5.2 Scoring Each Maturity Level

To be able to assess the level of CSC maturity, each maturity level needs a score-range

(zero to 1,000 points). According to the EFQM (1998), the total score of 1,000 points

could be divided into five levels, as follows:

� Uncommitted: The score ranges from 0 to 249 points.

� Drifters: The score ranges from 250 to 499 points.

� Improvers: The score ranges from 500 to 749 points.

� Award winners: The score ranges from 750 to 999 points.

� World-class: This level has a single score of 1,000 points.

Many researchers, however, report the use of the EFQM Excellence model with a

number of different levels and respective score ranges. Dale and Smith (1997), for

example, divided the total of 1,000 points into six levels with the score-ranges, as

presented below. They suggested that the levels were a useful way of characterizing

organizations, and helping them to recognize symptoms and develop plans for the

future.

� Unaware or uncommitted: The score-range is between zero - 99 points.

� Initiators: The score-range is between zero - 299 points (covers the immediate

previous and next levels).

� Drifters: The score-range is between 100 - 299 points.

� Improvers: The score-range is between 300 - 649 points.

� Award winners: The score-range is between 650 - 749 points.

� World-class: The score-range is between 750 - 1,000 points.

Ahmed et al. (2003), on the other hand, allocated the 1,000 points, based on the EFQM

Excellence model and the interviews with senior managers and consultants, into seven

levels, to be used as a quality self-assessment. Those levels are:


75

� Uncommitted: The score is between zero - 149 points.

� Drifter: The score is between 150 - 299 points.

� Tool pusher: The score is between 300 - 499 points.

� Improver: The score is between 500 - 649 points.

� Award winner: The score is between 650 and 849 points.

� World-class: The score is between 850 and 999 points.

� Supertitive: This level has a single score of 1,000 points.

In view of the score-range diversity listed above, the author decided to use five levels of

safety culture maturity, as represented in Figure 3.7, with each level having a score-

range of 200 points (as shown below):

� Emerging level: A score ranges between zero - 200 points.

� Managing level: A score ranges between 201 - 400 points.

� Involving level: A score ranges between 401 - 600 points.

� Cooperating level: A score ranges between 601 - 800 points.

� Continually improving level: A score ranges between 801 - 1,000 points.

These score ranges are later used, together with the CSC index developed through SD

modelling (described in Chapter 6), to identify the CSC maturity level in the

organization.

3.6 SUMMARY

In this chapter, three performance measurement systems, including the MBNQA

framework, the BSC framework, and the EFQM Excellence model were considered as

they were likely to be used in developing the CSC model. The advantages and

disadvantages of each performance measurement system were compared, and the

EFQM Excellence model was selected for the CSC model development. Using the

EFQM Excellence model as the basic framework, a CSC model was proposed. It


76

consisted of six constructs, including five enablers (Leadership, Policy and Strategy,

People, Partnerships and Resources, and Processes), and a single set of Goals, to

represent the basic elements of the proposed model.

The proposed CSC model comprised 34 attributes to operationally define its six

constructs (five enablers and Goals). These attributes were used in developing a

questionnaire survey to elicit respondents’ opinions on the different attributes in the

context of their current safety practices and performance. The details of questionnaire

survey development, data collection, as well as the preliminary analyses, are explained

in the next chapter.


77

44 DDAATTAA CCOOLLLLEECCTTIIOONN AANNDD PPRREELLIIMMIINNAARRYY AANNAALLYYSSEESS


The 34 attributes derived in Chapter 3 were used in developing a questionnaire survey,

which is described in this chapter. Response rates, as well as the sample characteristics

are demonstrated. Then the preliminary analyses, including the handling of missing

data, the normality test, the outliers test, and the reliability test were performed to

increase confidence in the data collected.

4.2 QUESTIONNAIRE SURVEY

As stated in Section 2.2.3, this study uses the questionnaire survey to facilitate the

collection of information from construction organizations. Selvanathan and Selvanathan

(2005) stated that a good questionnaire survey could influence a high response rate. The

longer the questionnaire, the lower both the response rate and the quality of the data

collected. They suggested that, in developing the questionnaire, the researcher should:

� Keep the questionnaire as short as possible. This approach encourages respondents

to complete it.

� Ask short, simple, and clearly worded questions, to enable respondents to answer

quickly, correctly, and without ambiguity.

� Start with simple questions to help respondents get started comfortably.

� Arrange the questions in logical order.

� Avoid using leading questions.

� Include a covering letter, which explains the purpose of the survey.


78

� Ensure that the questionnaire does not violate any ethical issues.

The questionnaire survey in this study contained a total of six pages, which could be

considered appropriate (not too long). It comprised two parts. The first part was devoted

to gathering demographical information about the respondents and their respective

organizations to ensure that the respondents have the appropriate backgrounds, and to

determine fundamental organizational safety issues including safety performance

records compared to their peers, safety policy and initiatives, reporting systems, and

devoted safety resources. This part contained nine questions: four open-ended questions

and five partially open-ended questions. Consequently, it was useful in identifying

discrepancies in the received responses.

The second part of the questionnaire covered 34 statements (representing the 34 CSC

attributes, see Table 3.7), to operationally define the six constructs (five enablers and

Goals) of the proposed CSC model. Each statement was designed to elicit respondents’

opinions on the different attributes in the context of their current safety practices and

performance using a five-point Likert scale, with point 1 representing ‘strongly

disagree’ and point 5 representing ‘strongly agree’. This approach enabled the

evaluation of the organization’s perception of, and commitment towards, each construct

to be carried out (see the questionnaire survey in Appendix 1).

With help from the Department of Labour Protection and Welfare, Ministry of Labour,

Thailand, a list of more than 150 medium to large construction organizations, with more

than 100 staff, was prepared and used as the sampling frame. The targeted respondents

were selected on the assumption that they held senior appointments (such as executive

directors, managing directors, and senior project managers) within their respective

organizations to capture a macro-level perspective of safety culture. The questionnaire

survey was both mailed and handed directly to targeted organizations.


79

4.3 SAMPLE CHARACTERISTICS

Two hundred and twenty questionnaires were distributed, with 118 responses

represented a response rate of 53.6% (as research experts have argued that mail surveys

may not be reliable unless they either achieve a minimum of 50% response, or

demonstrate with some form of verification that the nonrespondents are similar to the

respondents) (Erdos, 1970). Up to three usable feedback questionnaires were chosen

from each organization to avoid bias in the data. From the returned responses, only three

were deemed unusable, due to unanswered items (data incompleteness or response

discrepancy), and were subsequently dropped from the data set. As a result, 115 usable

questionnaires provided data for 101 companies for the analyses (Appendix 2 contains

all raw data pertaining to questionnaire results).

�

As shown in Figures 4.1 and 4.2, 75.7% of the respondents had more than five years

working experience in the local Thai construction industry, and 59.1% had been

working for their present organization for at least five years. This result indicates the

reasonably high work experience rate of the respondents.

24.318.3

57.4

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

100.0

0-5 6-10 >10

Years in the construction industry

Res

pond

ents

(%)

Figure 4.1 Years of experience in the Thai construction industry


80

40.9

19.1

40.0

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

100.0

0-5 6-10 >10

Years in the present organization

Res

pond

ents

(%)

Figure 4.2 Years of experience in the present organization

As shown in Figure 4.3, all respondents held senior positions in their organizations.

Most (83%) were also involved actively in site operations as they worked as project

managers, site managers, and safety managers.

Safety manager37%

Project manager26%

Site manager20%

Director17%

Figure 4.3 Job titles of the respondents


81

Most of the respondents had safety responsibilities in planning and auditing safety on

site (see Figure 4.4). In addition, more than 50% of the respondents engaged in safety

related activities (such as training, and meetings reporting) at least once a month (see

Figure 4.5). These figures indicate the positive involvement of the respondents in a wide

range of safety activities.

Planning37%

Supervising23%

Auditing32%

Others8%

Figure 4.4 Safety responsibilities

Every month51%

Irregular23%

Every year17%

Every six months9%

Figure 4.5 Safety activities engagement


82

Almost all of the respondents (93%) reported that their organizations had a formal

safety policy (see Figure 4.6), thus proving the appropriateness of the sampled

organizations involved in the survey. Also, more than 80% of the respondents believed

that their organization’s safety performance was at least as good as the national average

safety record (see Figure 4.7). These results (Figures 4.6 and 4.7) give confidence in the

suitability of the sampled companies to reflect the correct practices of the Thai

construction industry.

No7%

Yes93%

Figure 4.6 Formal safety policy in the organization

Better29%

Same52%

Worse19%

Figure 4.7 Safety performance compared to the national average record


83

The majority of respondents ranked People as the most influential enabler for

significantly improving safety culture; both in the organization and in the construction

industry (see Figure 4.8). This result was consistent with the research findings of

Pipitsupaphol and Watanabe (2000), who investigated the root causes of labour

accidents in the Thai construction industry. They concluded that the major immediate

causes of accidents relate to unsafe acts of workers (such as not wearing personal

protective equipment). Further, they maintained that improvements in these aspects

could reduce more than one-third of the accidents.

14.3

24.1

33.9

18.8

8.9

22.328.6

32.1

7.1 9.8

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

100.0

Leadership Policy andStrategy

People Partnershipsand Resources

Processes

CSC Enablers

Res

pond

ents

(%)

The Industrial Level The Organizational Level

Figure 4.8 The most influential enablers in improving safety culture

In addition to the People enabler, the respondents ranked Policy and Strategy and

Leadership as the other two important enablers for improving safety culture. Leaders,

therefore, should motivate their team members to achieve safety goals (Northouse,

1997). The survey respondents also considered Partnerships and Resources as a

significant factor in improving safety culture at the industrial level, but not at the

organizational level. This outcome may result from the organization not being able to

provide adequate, and necessary, safety resources (such as personal protective

equipment) to all workers because of employee turnover, and the addition of and

released from the project team in response to the work schedules (workforce instability).


84

In conclusion, organizations that participated in the survey claimed that they had safety

policies and safety implementations in place. The respondents had a high level of

experience in the construction industry, were in senior positions, and were thus able to

drive safety improvements.

4.4 DATA SCREENING AND PRELIMINARY ANALYSES

After the data was collected, a number of data examination techniques, ranging from the

simple process of visual inspection of graphical displays to statistical methods. Thus,

statistical methods of the handling of missing data, the normality test, the outliers test,

and the reliability test needed to be performed to increase confidence in the data. Each

statistical method is described in detail below.

4.4.1 Handling Missing Data

Missing data is one of the most pervasive problems in data analysis. Its seriousness

depends on the pattern of missing data, how much is missing, and why it is missing.

According to Tabachnick and Fidell (2007), the pattern of missing data is more

important than the amount missing. Thus, missing values scattered randomly through a

data matrix pose less serious problems than non-randomly missing values, which are

serious, no matter how few they are, because they affect the generalizability of the

results. There are a number of methods used to handle missing data, such as deleting

cases, using mean substitution, using a missing data correlation matrix, and treating

missing data as data. Tabachnick and Fidell (2007), however, claimed that if only 5% or

less of the data points are missing in a random pattern from a large data set, the

problems are less serious, and that almost any procedures for handling missing data

yield similar results. Table 4.1 illustrates the percentage of the missing values for the 34

items (attributes) within the six CSC constructs.


85

Table 4.1 Missing values

No. Attribute (Item) Usable Case(s) Missing Data (Total Cases) Count Percent

1. Commitment 114 1 0.9

2. Communication 114 1 0.9

3. Accountability 115 - -

4. Leading by example 115 - -

5. Safety awareness 113 2 1.7

6. Safety and productivity alignment 114 1 0.9

7. Safety standards 112 3 2.6

8. Safety initiatives 114 1 0.9

9. Safety integration in business goals 114 1 0.9

10. Shared perceptions 115 - -

11. Safety responsibilities 115 - -

12. Supportive environment 114 1 0.9

13. Workers’ involvement 113 2 1.7

14. Workers’ relationships 115 - -

15. Workload 114 1 0.9

16. Work pressure 115 - -

17. Stakeholders’ cooperation 113 2 1.7

18. Financial resources 113 2 1.7

19. Safety resources 114 1 0.9

20. Human resources 115 - -

21. Training 113 2 1.7

22. Risk assessment 113 2 1.7

23. Feedback 115 - -

24. No-blame approach 113 2 1.7

25. Housekeeping 112 3 2.6

26. Safety documentation 114 1 0.9

27. Benchmarking system 115 - -

28. Job satisfaction 112 3 2.6

29. Safe work behaviour 115 - -

30. Number of accidents 115 - -

31. Customers’ expectations 114 1 0.9

32. Industrial image 111 4 3.5

33. Workforce morale 114 1 0.9

34. Cost of accidents 115 - -


86

None of the 34 items (attributes) had more than 5% of their values missing (with the

highest percentage is of the ‘industrial image’ item, see Table 4.1). Thus, any ‘handling

missing data’ method could be employed. For this study, the ‘mean substitution’ method

was chosen.

4.4.2 Test of Normality

The screening of continuous variables for normality is an important early step in almost

every multivariate analysis (Tabachnick and Fidell, 2007). Although the normality of

the variables is not always required for an analysis, the solution is usually more

appropriate if the variables are all normally distributed. For this reason, the normality of

the variables is assessed by either statistical or graphical methods.

Two important components of normality are skewness and kurtosis (Tabachnick and

Fidell, 2007). Skewness relates to the symmetry of the distribution; a skewed variable is

a variable whose mean is not in the centre of the distribution. Kurtosis, on the other

hand, relates to the peakedness of a distribution; a distribution is either too peaked (with

short, thick tails), or too flat (with long, thin tails). When a distribution is normal, the

values of skewness and kurtosis are zero (Pallant, 2005). If there is a positive skewness,

there is a pileup of cases to the left, and the right tail is too long; with negative

skewness, the result is reversed. Kurtosis values above zero indicate a distribution that

is too peaked, while kurtosis values below zero are reversed. Non-normal kurtosis

produces an underestimate of the variance of a variable.

According to Morgan and Griego (1998), if the division of statistics values (Stat.) of

skewness (or kurtosis) and its standard error (S.E.) are not above 5.5, then that skewness

(or kurtosis) is not significantly different from normal. Curran et al. (1996), however,

recommend that the values of skewness < 2.0 and kurtosis < 7.0 are acceptable. Table

4.2 shows the skewness and kurtosis values of the 34 attributes. The results demonstrate

that all 34 attributes show normal distribution, thus increasing confidence in the data.


87

Table 4.2 Skewness and kurtosis of the 34 attributes

No. Attribute (Item) Skewness Kurtosis Stat. S.E. Stat./S.E. Stat. S.E. Stat./S.E.

1. Commitment -1.16 .226 -5.13 1.81 .449 4.03

2. Communication -0.59 .226 -2.61 0.25 .449 0.56

3. Accountability -0.95 .226 -4.20 1.25 .447 2.80

4. Leading by example -1.47 .226 -6.50 2.15 .447 4.81

5. Safety awareness -0.76 .227 -3.35 0.23 .451 0.51

6. Safety and productivity alignment -0.89 .226 -3.94 0.64 .449 1.43

7. Safety standards -0.69 .228 -3.03 0.46 .453 1.02

8. Safety initiatives -1.00 .226 -4.42 1.42 .449 3.16

9. Safety integration in business goals -0.75 .226 -3.32 0.35 .449 0.78

10. Shared perceptions -0.50 .226 -2.21 0.27 .447 0.60

11. Safety responsibilities -0.19 .226 -0.84 -0.14 .447 -0.31

12. Supportive environment -0.78 .226 -3.45 1.31 .449 2.92

13. Workers’ involvement -0.49 .227 -2.16 -0.76 .451 -1.69

14. Workers’ relationships -0.42 .226 -1.86 0.17 .447 0.38

15. Workload -0.77 .226 -3.41 0.10 .449 0.22

16. Work pressure -0.54 .226 -2.39 -0.12 .447 -0.27

17. Stakeholders’ cooperation -0.76 .227 -3.35 0.46 .451 1.02

18. Financial resources -0.32 .227 -1.41 -0.45 .451 -1.00

19. Safety resources -0.72 .226 -3.19 0.05 .449 0.11

20. Human resources -0.79 .226 -3.50 0.81 .447 1.81

21. Training -0.39 .227 -1.72 -0.33 .451 -0.73

22. Risk assessment -0.51 .227 -2.25 -0.42 .451 -0.93

23. Feedback -0.59 .226 -2.61 0.00 .447 0.00

24. No-blame approach -0.43 .227 -1.89 -0.35 .451 -0.78

25. Housekeeping -0.48 .228 -2.11 -0.02 .453 -0.04

26. Safety documentation -1.16 .226 -5.13 1.34 .449 2.98

27. Benchmarking system -0.69 .226 -3.05 -0.28 .447 -0.63

28. Job satisfaction -1.02 .228 -4.47 1.81 .453 4.00

29. Safe work behaviour -0.88 .226 -3.89 1.06 .447 2.37

30. Number of accidents -0.93 .226 -4.12 1.40 .447 3.13

31. Customers’ expectations -0.66 .226 -2.92 0.51 .449 1.14

32. Industrial image -1.04 .229 -4.54 1.34 .455 2.95

33. Workforce morale -1.01 .226 -4.47 1.73 .449 3.85

34. Cost of accidents -1.28 .226 -5.66 2.03 .447 4.54

Note: Stat. = Statistics values, S.E. = Standard error values


88

4.4.3 Outliers Test

An outlier is a case with such an extreme value on one variable (a univariate outlier), or

such a strange combination of scores on two or more variables (multivariate outlier),

that distorts the statistical results (Tabachnick and Fidell, 2007). There are many ways

to test for outliers, such as the use of the 5% trimmed mean, the use of standardized

scores (z-scores), and the use of boxplots (Pallant, 2005; Tabachnick and Fidell, 2007).

In this study, however, the mean, the 5% trimmed mean, and the z-score tests were used

to detect outliers.

4.4.3.1 5% Trimmed Mean

The 5% trimmed mean is a mean calculated from the cases in which 5% of the top and

the bottom of the cases are removed (Pallant, 2005). According to Pallant (2005), the

big difference (> 0.2) between a mean and its 5% trimmed mean may indicate a problem

with an outlier. Table 4.3 illustrates the means, the 5% trimmed means, and the standard

deviations (S.D.) of the 34 CSC attributes. The results show that the mean differences

(�mean) of all attributes are small, providing support for the absence of outliers.

4.4.3.2 Z-Score

To further detect outliers, a standardized score (z-score) test was performed. According

to Tabachnick and Fidell (2007), the cases with z-scores that exceed 3.29, at p < 0.01,

two-tailed test, are the potential outliers. There were 12 z-scores exceeding 3.29, in

which most were from case number ‘76’ (data file number 76, see Appendix 3). As a

result, case number ‘76’ was deleted from the data file, leading to a total of 114 data

retained for further analyses.


89

Table 4.3 The mean, the 5% trimmed mean, and the standard deviation of the 34

attributes

Attribute (Item) Mean 5% Trimmed Mean ��Mean S.D. Commitment 4.19 4.27 0.08 0.819

Communication 3.66 3.70 0.04 0.948

Accountability 4.01 4.07 0.06 0.832

Leading by example 4.25 4.36 0.11 0.926

Safety awareness 3.61 3.68 0.07 1.073

Safety and productivity alignment 3.80 3.87 0.07 0.997

Safety standards 3.81 3.87 0.06 0.916

Safety initiatives 3.82 3.89 0.07 0.895

Safety integration in business goals 3.78 3.85 0.07 1.011

Shared perceptions 3.92 3.96 0.04 0.751

Safety responsibilities 3.72 3.74 0.02 0.732

Supportive environment 3.99 4.04 0.05 0.781

Workers’ involvement 3.73 3.76 0.03 0.813

Workers’ relationships 3.67 3.70 0.03 0.835

Workload 3.83 3.89 0.06 0.861

Work pressure 3.77 3.80 0.03 0.921

Stakeholders’ cooperation 3.63 3.67 0.04 0.928

Financial resources 3.72 3.75 0.03 0.940

Safety resources 3.88 3.93 0.05 0.942

Human resources 3.94 4.01 0.07 0.891

Training 3.81 3.86 0.05 0.921

Risk assessment 3.47 3.50 0.03 1.036

Feedback 3.80 3.84 0.04 0.910

No-blame approach 3.21 3.24 0.03 1.097

Housekeeping 3.75 3.78 0.03 0.800

Safety documentation 3.93 4.02 0.09 0.984

Benchmarking system 3.30 3.34 0.04 1.036

Job satisfaction 3.65 3.70 0.05 0.813

Safe work behaviour 3.86 3.91 0.05 0.826

Number of accidents 4.11 4.17 0.06 0.803

Customers’ expectations 3.89 3.94 0.05 0.849

Industrial image 3.79 3.85 0.06 0.865

Workforce morale 3.81 3.86 0.05 0.819

Cost of accidents 4.02 4.10 0.08 0.917


90

4.4.4 Scale Reliability (Cronbach’s Alpha)

When selecting scales to include in the study, it was important to find the scales that

were statistically reliable. The scale reliability, the proportion of variance attributable to

the true score of latent variable (Pallant, 2005), can be defined as the extent to which a

measure produces similar results over different occasions of the data collection (Seo et

al., 2004). It was thus essential to examine the reliability of each scale whenever a

measurement was involved.

One of the main issues in scale reliability concerns the scale’s internal consistency

(Cronbach’s alpha, �). In a good solution, Cronbach’s alpha (�) ranges between zero

and one - the larger the value, the more stable the factors. A high value means that the

observed variables account for substantial variance in the factor scores, while a low

value means the factors are poorly defined by the observed variables. Generally, the

value of 0.70 is accepted as the minimum desired value of reliability (Pallant, 2005).

In this study, the 34 attributes within the six CSC constructs were tested for internal

consistency, using the retained 114 data. The results, shown in Table 4.4, had values

ranging from 0.83 to 0.89, all of which were considered acceptable. This thus increases

confidence in the contribution of the 34 attributes to the measurement of their respective

constructs.


91

Table 4.4 Internal consistency of the five enablers and Goals

Construct ��

1. Leadership Enabler

(Attributes: commitment, communication, accountability, and leading by

example)

0.837

2. Policy and Strategy Enabler

(Attributes: safety awareness, safety and productivity alignment, safety

standards, safety initiatives, and safety integration in business goals)

0.881

3. People Enabler

(Attributes: shared perceptions, safety responsibilities, supportive environment,

workers’ involvement, workers’ relationships, workload, and work pressure)

0.858

4. Partnerships and Resources Enabler

(Attributes: stakeholders’ cooperation, financial resources, safety resources, and

human resources)

0.874

5. Processes Enabler

(Attributes: training, risk assessment, feedback, no-blame approach,

housekeeping, safety documentation, and benchmarking system)

0.852

6. Goals

(Attributes: job satisfaction, safe work behaviour, number of accidents,

customers’ expectations, industrial image, workforce morale, and cost of

accidents)

0.893

To further confirm this finding, as well as to gain a better understanding of the factor

structure of the CSC scale as a preliminary step toward the SEM, the EFA was

conducted (see following chapter).


92



93

55 EEXXPPLLOORRAATTOORRYY FFAACCTTOORR AANNAALLYYSSIISS AANNDD

SSTTRRUUCCTTUURRAALL EEQQUUAATTIIOONN MMOODDEELLLLIINNGG


Preliminary analyses were conducted to increase confidence in the data (see Chapter 4).

Following on from that work, an exploratory factor analysis (EFA) was performed to

extract attributes into a number of factors that represent the interrelations among the set

of those attributes (as discussed in the following section). The attributes associated with

the five enablers of the proposed CSC model were analysed with the EFA to confirm

the construct validity of those five enablers. To further confirm the construct validity,

and to examine the causal relationships between the six constructs (five enablers and

Goals), structural equation modelling (SEM) was performed, using the AMOS program.

The final CSC model was achieved, as reported at the end of this chapter

The exploratory factor analysis, along with its details, is presented in the next section.

5.2 THE EXPLORATORY FACTOR ANALYSIS

The exploratory factor analysis (EFA) method is often used in the early stages of data

analysis to gather information about interrelationships among a set of variables.

According to Seo et al. (2004), the EFA is a precursor to the SEM. When conducting an

EFA, three main steps are followed: 1) the assessment of the suitability of the data; 2)

the factor extraction; and 3) the factor rotation (Pallant, 2005). The details of each step

are described below.


94

5.2.1 Assessment of the Suitability of the Data for the Analysis

Two main issues facilitate the determination of whether a particular data set is suitable

for factor analysis. The first issue is the sample size. Tabachnick and Fidell (2007)

noted that it is comforting to have at least 300 cases for factor analysis. However, they

conceded that a smaller sample size (e.g. 150 cases) should be sufficient, if the solutions

have several high loading marker variables (above 0.80). Pallant (2005), on the other

hand, recommended that five cases for each item are adequate in most cases. Coakes

and Steed (2003) took a stance in between, arguing that a sample of 100 cases is

acceptable, with a sample of 200 or more cases being preferable. A total of (usable) 114

cases (with case number ‘76’ deleted) were considered acceptable for the analysis in

this study.

The second issue concerns the strength of the inter-correlations among the items.

Bartlett’s test of sphericity and the Kaiser-Meyer-Olkin (KMO) test are normally

applied to assess the factorability of the data (Pallant, 2005). Bartlett’s test of sphericity

should be significant (p < 0.05) for factor analysis to be considered appropriate, while

the KMO index should range from zero to one, with 0.6 suggested as the minimum

value for a good factor analysis (Tabachnick and Fidell, 2007).

In this study, Bartlett’s test of sphericity was significant (see Table 5.1), with the KMO

index being 0.91, thus indicating that the data was suitable for factor analysis.

Table 5.1 Bartlett’s test of sphericity and the KMO index

Test Recommended

Value

Calculated

Value

Bartlett's test of sphericity (significant) < 0.05 0.00

Kaiser-Meyer-Olkin measure of sampling adequacy (KMO) > 0.60 0.91


95

5.2.2 Factor Extraction

Factor extraction, the second step in conducting the EFA, involves determining the

smallest number of factors that can be used to best represent the interrelations among

the set of variables (Tabachnick and Fidell, 2007). There are a variety of approaches

that can be used to extract (or identify) the number of underlying factors. Some of the

most commonly available extraction techniques are principal components, principal axis

factoring, and general least square (Pallant, 2005).

According to Coakes and Steed (2003), the most frequently used techniques are

principal components and principal axis factoring. The goal of both techniques is to

extract the maximum variance from the data set with each component. The principal

axis factoring is, however, widely used, and conforms to the factor analytic model in

which common variance is analysed, with the unique and error variance removed

(Tabachnick and Fidell, 2007). For this reason, the principal axis factoring method was

chosen for the analysis.

It is also important to determine the number of factors requested for factor analysis.

Thus, according to Pallant (2005), three techniques are generally used to assist in

decisions concerning the number of factors to retain. These techniques are: 1) the

Kaiser’s criterion or the eigenvalue rule; 2) the Catell’s screed test; and 3) the Horn’s

parallel analysis. The eigenvalue is the value that represents the amount of total variance

explained by that factor. In this study, the eigenvalue over one (> 1.0) was used as the

criterion for extracting the factors of the CSC model (Tabachnick and Fidell, 2007).

5.2.3 Factor Rotation and Interpretation

After the extraction, factor rotation is ordinarily used to present the pattern of loadings

in a manner that is easy to interpret. Numerous methods of factor rotation are available,

but the most commonly used is ‘varimax’ (Coakes and Steed, 2003; Pallant, 2005).


96

Varimax is a variance maximizing procedure. The goal of varimax rotation is to

maximize the variance of factor loadings by making high loadings higher, and low

loadings lower, for each factor (Tabachnick and Fidell, 2007). The varimax method was

used for the factor analysis in this study.

In summary, principal axis factoring, with an eigenvalue over one, together with the

varimax rotation method, were used to examine the dimensionality of the 27 attributes

of the CSC’s five enablers, and to achieve better interpretability of the factor loadings.

As the remaining seven attributes of the Goals construct were grouped together as one

factor, they were not included in the analysis. The results of the EFA are shown below.

5.2.4 The EFA Results

The 27 attributes of the CSC’s enablers were analysed for factor extraction. Due to the

sample size (114 data points), a cut-off factor loading of 0.45 was used to screen out the

attributes (or items) that were weak indicators of the constructs (as suggested by Hair et

al. (1998)). The first run thus led to dropping of the ‘no-blame approach’ item as it

failed to make the cut-off. Consequently, performing the factor analysis on the

remaining 26 items highlighted another problematic item, namely ‘shared perceptions’.

Based on an eigenvalue greater than one (> 1.0), the remaining 25 attributes were

extracted into three factors, which accounted for 59.73% of the total variance, as shown

in Table 5.2. Factor 1 was predominantly accounted for by nine items, initially

measuring Processes (Pro) and Partnerships and Resources (Prs); Factor 2 by 10 items,

initially measuring Leadership (Lds) and Policy and Strategy (Pol); and Factor 3 by six

items, initially measuring People (Ppl) and Partnerships and Resources (Prs).


97

Table 5.2 Three factors extracted from the remaining 25 items

Item Factor Extracted

1 2 3

Safety resources .791

Risk assessment .759

Benchmarking system .670

Financial resources .651

Workers’ involvement .641

Safety documentation .587

Training .585

Feedback .536

Safety integration in business goals .535

Accountability .731

Safety and productivity alignment .703

Commitment .697

Communication .676

Safety initiatives .609

Leading by example .575

Safety awareness .550

Safety standards .494

Supportive environment .476

Workers’ relationships .465

Workload .775

Human resources .673

Housekeeping .597

Stakeholders’ cooperation .578

Work pressure .568

Safety responsibilities .481

Extraction Method: Principal Axis Factoring. Rotation Method: Varimax with Kaiser Normalization. Rotation converged in seven iterations.

A closer examination of the identified factors revealed the potential for further analysis

to extract the independent factors in line with those of the proposed CSC model (see

Figure 3.6). For this reason, it was decided to further factor-analyse the three factors,

setting their required extraction limit to two new factors each. The nine items of Factor


98

1 (in Table 5.2) thus gave rise to two new factors, accounting for 62.58% of the total

variance, as shown in Table 5.3.

Table 5.3 Two factors extracted from nine items of Factor 1 of Table 5.2


1 2



Risk assessment .689


Training .524



Feedback .610


Extraction Method: Principal Axis Factoring. Rotation Method: Varimax with Kaiser Normalization. Rotation converged in three iterations.

Similarly, two factors were extracted from the 10 items of Factor 2 (in Table 5.2).

However, as the ‘supportive environment’ item failed to make the cut-off factor loading,

it was removed from the data file. The remaining nine items of Factor 2 were

reanalysed, and produced two new factors, accounting for 59.10% of the total variance

(see Tables 5.4). No new factors were extracted from Factor 3, thus this factor was

considered as a single construct.


99

Table 5.4 Two factors extracted from nine items of Factor 2 of Table 5.2


1 2





Commitment .770

Communication .644

Accountability .624



Extraction Method: Principal Axis Factoring. Rotation Method: Varimax with Kaiser Normalization. Rotation converged in three iterations.

In this study, five factors, within the remaining 24 items, were extracted from the EFA

(as summarized in Table 5.5), with Leadership (five items), Policy and Strategy (four

items), People (six items), Partnerships and Resources (five items), and Processes (four

items).

Interestingly, the above analysis led to nine items1, initially assumed to be associated

with a certain enabler, to strongly correlate with another enabler. To illustrate, the

‘safety and productivity alignment’ item appeared to be loading on the Leadership

enabler not the Policy and Strategy enabler, as was initially hypothesized.

1The nine relocated items were: 1) the ‘safety and productivity alignment’; 2) the ‘workers relationships’; 3) the

‘human resources’; 4) the ‘housekeeping’; 5) the ‘stakeholders cooperation’; 6) the ‘risk assessment’; 7) the ‘workers involvement’; 8) the ‘training’; and 9) the ‘safety integration in business goals’ items.


100

Table 5.5 Five factors extracted from the EFA

Item Factors Extracted Lds Pol Ppl Prs Pro

Commitment .770

Communication .644

Accountability .624

Safety and productivity alignment* .551





Workers’ relationships* .515

Workload .775

Human resources* .673

Housekeeping * .597

Stakeholders’ cooperation* .578

Work pressure .568




Risk assessment* .689

Workers’ involvement* .535

Training * .524


Safety integration in business goals* .614

Feedback .610


Note: * Items relocated to another enabler, Lds = Leadership, Pol = Policy and Strategy, Ppl = People, Prs = Partnerships and Resources, Pro = Processes

Following the re-allocation of the nine items, the Cronbach’s alpha (�) test was re-

applied to ensure the appropriateness of the groupings of the five factors extracted. As

shown in Table 5.6, the alpha coefficients ranged from 0.85 to 0.90, all of which were

considered highly reliable.


101

Table 5.6 Internal consistency of five factors extracted from the EFA

Construct Cronbach’s Alpha (��)

1. Leadership Enabler

(Attributes: Commitment, Communication, Accountability, Safety

and productivity alignment, and Leading by example)

0.856

2. Policy and Strategy Enabler

(Attributes: Safety awareness, Safety standards, Safety initiatives,

and Workers’ relationships)

0.849

3. People Enabler

(Attributes: Workload, Human resources, Housekeeping,

Stakeholders’ cooperation, Work pressure, and Safety

responsibilities)

0.896

4. Partnerships and Resources Enabler

(Attributes: Financial resources, Safety resources, Risk assessment,

Workers’ involvement, and Training)

0.886

5. Processes Enabler

(Attributes: Benchmarking system, Safety integration in business

goals, Feedback, and Safety documentation)

0.873

6. Goals

(Attributes: Job satisfaction, Safe work behaviour, Number of

accidents, Customers’ expectations, Industrial image, Workforce

morale, and Cost of accidents)

0.893

5.2.5 Conclusion of the EFA

The EFA gave rise to a total of 24 attributes grouped to explain five latent factors

(enablers), whereas a total of seven attributes were grouped to explain the sixth latent

factor (Goals), leading to the so-called baseline model of the CSC (shown in Figure

5.1).


102

PolSafety initiativesSafety standards

Safety and productivity alignment

Safety awareness

Safety integration in business goals

Ppl

Safety responsibilities

Workers' involvement

Workers' relationships

Prs

Human resources

Safety resourcesFinancial resources

Stakeholders' cooperation

Pro

Feedback

Training

Housekeeping

Safety documentation

GoalsCustomers' expectationsNumber of accidentsSafe work behaviour

Job satisfaction

Industrial imageWorkforce moraleCost of accidents

LdsLeading by example

AccountabilityCommunication

Commitment

WorkloadWork pressure

Benchmarking system

Risk assessment

Note: Lds = Leadership, Pol = Policy and Strategy, Ppl = People, Prs = Partnerships and Resources, Pro = Processes

Figure 5.1 Baseline model of the CSC


103

The ovals and rectangles, shown in the baseline model, symbolise latent and observed

variables, respectively. The former represent the six constructs of the baseline model,

whereas the latter represent their respective attributes.

The arrows connecting the two sets of variables indicate the direction of the

hypothesized influence. For example, it is hypothesized that Leadership is manifested

by the achievement of its five attributes, namely: ‘commitment’, ‘communication’,

‘accountability’, ‘leading by example’, and ‘safety and productivity alignment’; the

arrows are thus shown originating from Leadership to each one of the five attributes.

To provide further evidence of the construct validity of the CSC model, and to

investigate the causal relationships between the five enablers and Goals of the CSC, the

structural equation modelling was conducted next.

5.3 THE STRUCTURAL EQUATION MODELLING

Theoretically, the SEM comprises two types of models: measurement and structural

models. The former is concerned with how well the observed variables measure the

latent factors, addressing their reliability and validity. The latter is concerned with

modelling the relationships between the latent factors, by describing the amount of

explained and unexplained variance, which is akin to the system of simultaneous

regression models (Wong and Cheung, 2005). The details of the measurement and

structural models are explained below.


104

5.3.1 Measurement Model

Testing the structural model would have been meaningless until it was established as a

good measurement model. In this study, a confirmatory factor analysis (CFA) was

undertaken to establish confidence in the measurement model. The CFA specifies the

posited relations of the observed variables to the underlying constructs. It belongs to the

family of SEM techniques, as it allows for the assessment of fit between the observed

data and a priori conceptualized, theoretically grounded model that specifies the

hypothesized causal relationships between latent factors and their observed indicator

variables (Mueller and Hancock, 2004).

According to Byrne (2001), the measurement model should be assessed by five

methods: 1) the feasibility of parameter estimates; 2) the appropriateness of standard

errors (S.E.); 3) the statistical significance of parameter estimates; 4) model fit as a

whole (using goodness of fit, GOF, indices); and 5) square multiple correlation (SMC,

R2). Each method is described in detail below.

5.3.1.1 Feasibility of Parameter Estimates

Further, Byrne (2001) noted that the parameter estimates of a good measurement model

must exhibit the correct sign and size, and be consistent with the underlying theory. Any

estimates falling outside the admissible range signal a clear indication that either the

model is wrong or the input matrix lacks sufficient information. Examples of parameters

exhibiting unreasonable estimates are: 1) correlation values more than one; 2) negative

variances; and 3) covariance or correlation matrices that are not positive definite.

The analysis results of the measurement model, shown in Appendix 4, revealed that all

the parameter estimates were both reasonable and statistically significant, thus

confirming the construct reliability and validity of the baseline model.


105

5.3.1.2 Appropriateness of Standard Errors (S.E.)

Joreskog and Sorbom (1993) confirmed that the standard errors that are excessively

large or small indicate a poor model fit. For instance, if a standard error approaches

zero, the test statistic for its related parameter cannot be defined. Likewise, standard

errors that are extremely large indicate parameters that cannot be determined.

Because the standard errors are influenced by the units of measurement in observed

and/or latent variables, as well as the magnitude of the parameter estimate itself, no

definitive criterion for small and large has been established (Joreskog and Sorbom,

1993). The results from this study demonstrated that all standard errors appeared to be

in good order, proving the construct validity of the baseline model (see Appendix 3).

5.3.1.3 Statistical Significance of Parameter Estimates

The critical ratio (C.R.) was used to test the statistical significance of the parameter

estimates. The C.R. represents the parameter estimate divided by its standard error. It

operates as a z-statistic in testing that the estimate is statistically different from zero.

Based on a probability level of 0.05, the test statistic needed to be > �1.96 before the

hypothesis (that the estimate equals to zero) could be rejected (Byrne, 2001). The results

showed that all the C.R. values of the latent and observed variables of the baseline

model were more than 1.96, therefore, the hypothesis (that the estimate equals to zero)

could be rejected (see Appendix 3).

5.3.1.4 Model Fit as a Whole

The assessment of the overall model fit is considered as a critical issue in relation to any

SEM. There are a number of indices that may be used in assessing the model fit. One of


106

the most widely used fit indices is a model chi square (�2), which tests the closeness of

the fit between the sample covariance matrix and the fitted covariance matrix (Kline,

2005). However, since the formula for computing �2 is directly related to the sample

size, almost all the models are evaluated as incorrect as the sample size increases. For

this reason, the ratio of �2 to the degrees of freedom (�2/DF) has been commonly used

as an alternative fit index. Normally, the value of �2/DF less than two represents the

model as a good fit (Kline, 2005). Garson (2006), however, proposed that the value of

less than three is acceptable.

Another widely used fit index is the Root Mean Square Error of Approximation

(RMSEA) (Byrne, 2001). It has been recognized as one of the most informative criteria

in covariance structure modelling. It is the best fit index where models are very

parsimonious, as it measures the lack of fit per degree of freedom. According to

Tabachnick and Fidell (2007), the value of RMSEA up to 0.05 indicates a good model

fit, with a value up to 0.10 indicating a reasonable error of approximation.

Apart from the �2/DF and the RMSEA, an additional group of fit indices are commonly

used, including a Bentler-Bonett normed fit index (NFI), a comparative fit index (CFI),

an incremental fit index (IFI), a Tucker-Lewis index (TLI), and a relative fit index

(RFI)). Kline (2005) suggested that these indices’ value should be at least 0.90 to

indicate a model fit. Garson (2006), however, used the cut-off value of 0.80 as the

criterion. The GOF indices of the baseline model revealed a need to modify the model

in order to improve the model fit (see Table 5.7).

According to Kline (2005), there are potentially two possible options involved in the

process of model refinement. The first option is to eliminate the links or ‘paths’ with

very low correlations. This option was found not to be applicable to the baseline model

due to the high correlations of all paths (see Appendix 3). The second option is to

remove the observed variables shown by the computed modification indices as having

multicollinearity (e.g. the observed variables/their error variances that correlated to


107

more than two other observed variables/error variances). Garson’s (2006) observations

are that signs of high multicollinearity are indicated by:

Table 5.7 The GOF indices of the baseline and the best-fit measurement models

GOF Index Recommended Level Baseline

Model

Best-Fit Measurement

Model

�2/DF < 2.00 (Byrne, 2001) 2.10 1.65

RMSEA � 0.10 (Tabachnick and Fidell, 2007) 0.10 0.07

NFI > 0.80 (Ullman, 2001) 0.71 0.81

CFI > 0.90 (Kline, 2005) 0.82 0.91

IFI > 0.80 (Garson, 2006) 0.82 0.91

TLI > 0.90 (Kline, 2005) 0.80 0.90

RFI > 0.80 (Garson, 2006) 0.68 0.78

Note: The baseline model is shown in Figure 5.1, while the best-fit measurement model is shown in Figure 5.2.

� Standardized regression weights: Since all latent variables in a SEM model have

been assigned a metric of one, all standardized regression weights should be within

the range of plus or minus one. When there is a multicollinearity problem, the

standardized regression weights may show the values of greater than one and/or less

than minus one.

� Standard errors of the unstandardized regression weights: As with the standardized

regression weights, the unstandardized values of greater than one and/or less than

minus one may indicate a sign of high multicollinearity.

� Covariances of the parameter estimates: The variables with high multicollinearity

may well be reflected in high covariances of the parameter estimates.

� Variance estimates: Negative error variance estimates may also be another effect of

the multicollinearity

According to the second option, seven observed variables that showed signs of high

multicollinearity were removed (see the modification indices in Appendix 3):


108

� Three of those variables (‘workload’, ‘work pressure’, and ‘housekeeping’) were

from the People enabler.

� One variable, which was ‘risk assessment’, came from the Partnerships and

Resources enabler.

� Three variables (‘job satisfaction’, ‘safe work behaviour’, and ‘customers’

expectations’) were from Goals.

Further modifications appeared not to improve the model fit, thus leading to the best-fit

measurement model with the best GOF indices, as shown in Figure 5.2 and Table 5.7,

respectively. Table 5.7 highlights the significant improvement of the GOF indices’

values compared to those obtained for the baseline model.

5.3.1.5 Square Multiple Correlation (SMC, R2)

SMC (R2) is the extent to which a measurement model is adequately represented by the

observed measures. Each R2 is interpreted as the proportion of variance in the indicator

that is explained by the respective latent variable; this is a similar concept to the

communality estimate in the EFA (Ullman, 2001). Normally, the R2 0.5 is used as an

indicator of a reasonably good convergent validity for the model (Kline, 2005).

A summary of R2, together with the standardized path coefficients, of the best-fit

measurement model is shown in Table 5.8. The results showed that most of the R2 of the

observed variables were greater than 0.50, indicating a reasonably good convergent

validity of the fit-measurement model. Moreover, all path coefficients were positive and

statistically significant at p < 0.05, thus their significance to the model was augmented.


109



Safety awareness


PplSafety responsibilities



Prs

Human resources

Safety resourcesFinancial resources


Pro

Feedback

Training


Goals

Number of accidentsIndustrial image

Workforce moraleCost of accidents

LdsLeading by example

AccountabilityCommunication

Commitment

Benchmarking system

Note: Lds = Leadership, Pol = Policy and Strategy, Ppl = People, Prs = Partnerships and Resources, Pro = Processes

Figure 5.2 The best-fit measurement model


110

Table 5.8 Square multiple correlations and standardized coefficients of observed

variables

Regression Path R2 Standardized Coefficient

Leadership

Commitment 0.53 0.73

Communication 0.61 0.78

Accountability 0.70 0.84

Leading by example 0.38 0.61

Safety and productivity alignment 0.40 0.63

Policy and Strategy

Safety awareness 0.59 0.77

Safety standards 0.62 0.79

Safety initiatives 0.75 0.87

Workers’ relationships 0.35 0.59

People

Safety responsibilities 0.37 0.61

Stakeholders’ cooperation 0.76 0.87

Human resources 0.72 0.85


Financial resources 0.81 0.90

Safety resources 0.76 0.87

Workers’ involvement 0.35 0.59

Training 0.48 0.69

Processes

Feedback 0.59 0.77

Safety documentation 0.66 0.81

Benchmarking 0.45 0.67

Safety integration in business goals 0.59 0.77

Goals

Number of accidents 0.55 0.74

Industrial image 0.42 0.65

Workforce morale 0.46 0.68

Cost of accidents 0.36 0.60

Note: All path coefficients were significant at 0.05 probability level


111

5.3.2 Structural Model

Having established confidence in the measurement model, a structural equation model

was developed and tested to examine the direction of the assumed relationships between

the six latent variables (constructs), as reflected by the arrows connecting them (see

Figure 3.6). A fundamental feature of any SEM is the direction of the arrows denoting

the direction of the assumed relationships between the variables as explained below.

In the CSC model, the arrows were assumed to support the argument that Leadership

drives (influences) three enablers (Policy and Strategy, People, and Partnerships and

Resources), and that these enablers collectively influence the ability to achieve pre-

determined Goals through the implementation and improvement of suitable Processes.

As a starting point, bi-directional arrows were used to represent the relationships among

the three enablers (Policy and Strategy, People, and Partnerships and Resources),

without an explicitly defined causal direction. This is because of the variables’ potential

to affect each other. For example, Policy and Strategy might influence People, and/or

vice versa.

To explore this relationship further, and to improve the overall model fit, a number of

model runs (with different arrow directions connecting the enablers) were carried out.

Any links with very low correlations, or items showing signs of multicollinearity, were

deleted. For each run, the GOF indices were computed and compared. According to

Clissold (2004), the model with the best fit would prove the directional influences.

As a result, the ‘leading by example’ item was deleted because of its high

multicollinearity, leading to four observed variables representing the Leadership

construct. The fitted structural model (shown in Figure 5.3), with the best GOF indices

(listed in Table 5.9), was deemed to be the final CSC model (see Figure 5.4) (the full

results of the structural model are presented in Appendix 5).


112



Safety awareness


Ppl

Safety responsibilities


Prs

Human resources

Safety resources

Financial resources


Pro

Feedback

Training

GoalsIndustrial imageWorkforce moraleCost of accidents

LdsAccountability

CommunicationCommitment


Number of accidents

Benchmarking system


Note: Lds = Leadership, Ppl = People, Prs = Partnerships and Resources, Pol = Policy and Strategy, Pro = Processes

Figure 5.3 The best-fit structural model


113

Table 5.9 The GOF indices of the best-fit structural model

GOF Index Recommended Level Best-Fit

Measurement Model

Best-fit

Structural Model

�2/DF < 2.00 (Byrne, 2001) 1.65 1.68

CFI > 0.90 (Kline, 2005) 0.91 0.91

NFI > 0.80 (Ullman, 2001) 0.81 0.81

TLI > 0.90 (Kline, 2005) 0.90 0.90

IFI > 0.80 (Garson, 2006) 0.91 0.91

RFI > 0.80 (Garson, 2006) 0.78 0.78

RMSEA � 0.10 (Tabachnick and Fidell, 2007) 0.07 0.07

Note: The best-fit measurement model is shown in Figure 5.2, and the best-fit structural model is shown in Figure 5.3

The final CSC model (see Figure 5.4) confirmed that Processes had a significant direct

relationship with Goals (with a path coefficient = 0.90), and that Processes explains (or

influences) 82% of the variance in Goals.

Enablers Goals


Leadership(Lds)

People (Ppl)

Policy and Strategy

(Pol)


(Prs)

Processes(Pro) Goals

0.64

0.16

0.59

0.86

0.36

0.46

0.62

0.90

0.41

0.94

0.77

0.99 0.82

(100 points)

(90 points)

(90 points)

(80 points)

(140 points)(500 points)

Note: All path coefficients were significant at 0.05 probability level. The italised numbers show the variance values (R2) of the factors

Figure 5.4 The final CSC model


114

As previously mentioned, the value of SEM lies in its ability to depict both direct and

indirect effects between the variables. In light of this, the final CSC model (see Figure

5.4) appears to indicate that People strongly influences Partnerships and Resources,

whereas Partnerships and Resources moderately affects Policy and Strategy. Both

People and Policy and Strategy were found to have significant direct relationships with

Processes (with path coefficients of 0.46 and 0.62, respectively) at 0.05 probability

level.

No statistically significant relationship, however, was found between Partnerships and

Resources and Processes, indicating the absence of any direct effect. An indirect effect

existed, though, through Policy and Strategy. Further, Partnerships and Resources was

found to have a positive impact on Policy and Strategy (path coefficient = 0.36), which,

in turn, influenced Processes. It was worth noting that the R2 for Processes was 0.99,

demonstrating that 99% of the variance associated with this particular enabler was

accounted for by its two predictors, i.e. the People and Policy and Strategy enablers.

Leadership showed a significant direct relationship with People (path coefficient =

0.64) and Policy and Strategy (path coefficient = 0.59), but surprisingly bore no

statistically significant relationship with Partnerships and Resources (path coefficient =

0.16). The relatively strong influence People had on Partnerships and Resources (path

coefficient = 0.86) suggests that Leadership indirectly influences Partnerships and

Resources through People.

A summary of the direct and indirect path coefficients, together with the values of R2

between the five enablers and Goals, is shown in Table 5.10. Indeed, most of the R2 of

the latent variables were greater than 0.50 indicating a good convergent validity of the

model.


115

Table 5.10 The direct and indirect path coefficients between the five enablers and Goals

Latent Factor Correlation Coefficient R2

Ppl 0.64*Lds 0.41

Prs (0.16*Lds)+(0.86*Ppl)+(0.55*Lds*Ppl) 0.94

Pol (0.59*Lds)+(0.36*Prs)+(0.31*Ppl*Prs) 0.77

Pro (0.46*Ppl)+(0.62*Pol)+(0.29*Lds*Ppl)+(0.37*Lds*Pol)+

(0.22*Prs*Pol)

0.99

Goals (0.90*Pro)+(0.41*Ppl*Pro)+(0.56*Pol*Pro) 0.82

Note: All path coefficients were significant at 0.05 probability level, Ppl = People, Prs = Partnerships and Resources, Pol = Policy and Strategy, Pro = Processes

5.3.3 Conclusion of the SEM

The final CSC model reveals that Leadership strongly influences People and Policy and

Strategy. Leaders should, therefore, be a role model in promoting healthy and safe work

behaviour, ensure that workers accept their safety responsibilities, and set a realistic

safety policy and communicate this policy throughout organizations. It is clear that

Leadership is the main driver to effective safety culture, and the strong commitment of

leaders is crucial in promoting safety culture.

Leadership also has an influence on Partnerships and Resources; however, it appears to

be a relatively weak direct effect. It appears that most of its influence on this particular

enabler is mediated through the People enabler. This indirect effect corroborates well

with the overall perception of Thai construction managers, where human resources and

teamwork are believed to be more crucial to successful safety implementation than the

provision of safety resources (Aksorn and Hadikusumo, 2007). These relationships are

reflected by a strong correlation between People and Partnerships and Resources, and a

weak correlation between Leadership and Partnerships and Resources.


116

Consequently, safety policy and strategy should reflect the need of safety resources, as

requested by the workers, because of the direct and indirect relationships that exist

between Partnerships and Resources and Policy and Strategy, and between People and

Policy and Strategy, respectively.

In conclusion, it can be stated that People and Policy and Strategy play a key role in

successful safety implementation, as verified by the strong links from these two

enablers to Processes. Partnerships and Resources, on the other hand, shows no

significant direct, but indirect, effect on Process through Policy and Strategy.

Therefore, an effective safety policy and strategy will influence an effective safety

implementation, which, in turn, will enhance Goals achievement in organizations.

The final CSC model, as well as the correlation coefficients between its six constructs

(five enablers and Goals), were used in developing the CSC dynamic model (described

in the next chapter).


117

66 SSYYSSTTEEMM DDYYNNAAMMIICCSS MMOODDEELLLLIINNGG OOFF

CCOONNSSTTRRUUCCTTIIOONN SSAAFFEETTYY CCUULLTTUURREE -- MMOODDEELL BBUUIILLDDIINNGG


This chapter outlines the development of the CSC dynamic model. System dynamics

(SD) modelling technique was used to capture the interactions and causal relationships

between the five enablers and Goals of the CSC, over time. The developed dynamic

model was verified and validated to increase confidence in the model.

6.2 SYSTEM DYNAMICS MODELLING

System dynamics (SD) was used to examine the various social, economic, and

environmental systems, where a holistic view is important, and feedback loops are

critical to understanding the interrelationships (Rodrigues and Bowers, 1996).

According to Simonovic (2005), this understanding of interrelationships is achieved by

developing a model that can simulate and quantify the behaviour of the system, once

again over time. The simulation of the model is considered essential to understand the

dynamics of the system. An overview of SD, its applications in the construction, and the

available softwares were presented in Section 2.2.7.

To develop a SD simulation model, five steps must be carried out: 1) understanding the

system and its boundaries; 2) identifying the key variables; 3) representing the physical

processes or variables through mathematical relationships; 4) mapping the structure of

the model; and 5) simulating the model (Simonovic, 2005). Vennix (1996) has

identified the systematic procedural steps in SD modelling as follows:


118

� Problem identification and model purpose;

� System conceptualization;

� Model formulation and parameter estimation;

� Analysis of model behaviour including sensitivity analysis;

� Model evaluation, including model validity and verification;

� Policy analysis testing; and

� Model use or implementation.

In SD modelling, feedback loops are considered critical in understanding the

interrelationships between key elements of the model. These feedback loop structures,

once identified, are translated to so-called stock-flow diagrams, to enable the

simulations (Ford, 1999). The basic components of a typical SD model are shown in

Figure 6.1.

Stock: State or condition of the sy stem

Flow

Conv erter: Inf ormation about lev el of the sy stem

Figure 6.1 Basic components of a SD model

The basic CSC dynamic model (Figure 6.2) in this study showed that the CSC index

was represented by the sum of the five enablers and Goals values (with a maximum

score of 1,000 points). For simplicity, and keeping the basic 50/50 for the

Enablers/Goals ratio, as indicated by the EFQM Excellence model in Figure 3.6, any

increase in the value of the CSC index was assumed to have contributed from both

Enablers and Goals, evenly.

Source

Action


119

CSC INDEX

Enablers Goals

Leadership

Policy and Strategy

People


Processes

Figure 6.2 Basic CSC dynamic model

The following section sheds more light on the relationships between the five enablers

and Goals of the CSC, as modelled by the SD.

6.3 CAUSAL LOOP DIAGRAMS OF CONSTRUCTION SAFETY

CULTURE

6.3.1 Causal Loop Diagram

To conceptualize a real world system under investigation, the SD focuses on the

structure and behaviour (over time) of the system using multiple feedback loops (closed

chains of cause-and-effect links, in which information about the result of actions is fed

back to generate further action). These feedback loops are presented graphically using a

causal loop diagram (see Figure 6.3).


120

-

+ + +

+

+ -

+

-

Arboleda et al., 2003

Arboleda et al., 2003

Dedobbeleer and Beland, 1991

Pipitsupaphol and Watanabe, 2000

Speirs and Johnson, 2002

Siu et al., 2004 Siu et al., 2004

Siu et al., 2004

Gillen et al., 2002

Figure 6.3 An example of a causal loop diagram

The causal loop diagram is a SD tool used to portray a feedback loop in an easy

understanding diagram. A loop is a closed system, comprising a number of elements

and causal relationships. The arrows (as shown in Figure 6.3) indicate the direction of

influence, and plus/minus (+, -) signs indicate the type of the influence (Khanna et al.,

2004). These plus/minus signs have the following meanings (Forrester, 1985):

� A causal link from one element ‘A’ to another element ‘B’ is positive (that is, +), if

either: a) ‘A’ adds to ‘B’; or b) a change in ‘A’ produces a change in ‘B’ in the same

direction. For example, a better ‘perception of safety’ will enhance more

‘participation in safety activities’; this relationship is represented by a plus (+) sign

(see Figure 6.3).

� A causal link from one element ‘A’ to another element ‘B’ is negative (that is, -), if

either: a) ‘A’ subtracts from ‘B’; or b) a change in ‘A’ produces a change in ‘B’ in

the opposite direction. For example, a lesser ‘accident rate’ leads to higher ‘job

satisfaction’; this is represented by a minus (-) sign (see Figure 6.3).

In addition to the signs on each link, the complete feedback loop also is given a sign. If

a particular element starts the loop by changing its value in one direction (e.g. by

Management Commitment

Safety Resources

Safety Related Activities

Perception of Safety

Participation in Safety Activities

Distress

Accident Rate

Job Satisfaction

- - +

Feedback Loop ‘A’ Feedback Loop ‘B’


121

increasing its value), and closes the loop with the value changed in the same direction

(e.g. closes the loop by increasing the value), then the loop is called a positive loop. A

negative loop is the reverse.

According to Ventana Systems, Inc. (2001), the positive and negative loops can also be

determined by counting the number of minus (-) signs appearing on all the links that

make up the loop. Specifically,

� A feedback loop is called positive (indicated by or sign), if it contains an

even number of negative causal links, as shown in feedback loop ‘A’ in Figure 6.3.

� A feedback loop is called negative (indicated by or sign), if it contains an

odd number of negative causal links, as shown in feedback loop ‘B’ in Figure 6.3.

Thus, the sign of a loop is the algebraic product of the signs of its links. To better

understand the feedback loop, an explanation of a positive feedback loop ‘A’, between

‘perception of safety’, ‘participation in safety activities’, ‘distress’, ‘accident rate’, and

‘job satisfaction’, is described in detail below (see Figure 6.3).

An increased ‘perception of safety’ has the potential to increase the level of

‘participation in safety activities’ (representing a positive influence, ‘+’ sign)

(Dedobbeleer and Beland, 1991). A higher level of ‘participation in safety activities’

will tend to decrease the ‘distress’ (representing a negative influence, ‘-’sign), leading to

a reduced ‘accident rate’ (the decrease in distress reduces the accident rate; this

represents a positive influence, ‘+’ sign) (Siu et al., 2004).

A reduced ‘accident rate’ will lead to a higher ‘job satisfaction’ (representing a negative

influence, ‘-’ sign) (Siu et al., 2004), which, in turn, will enhance the people’s

‘perception of safety’ (the higher job satisfaction, the better perception of safety; this

represents a positive influence, ‘+’ sign) (Gillen et al., 2002). Positively enhancing the

‘perception of safety’ closes the loop. Thus, the feedback loop linking ‘perception of

+ +

- -


122

safety’, ‘participation in safety activities’, ‘distress’, ‘accident rate’ and ‘job

satisfaction’ is a positive loop, as shown by the sign (see Figure 6.3).

A negative feedback loop ‘B’ between ‘management commitment’, ‘safety related

activities’, ‘perception of safety’, ‘participation in safety activities’, and ‘safety

resources’ can be described as follows (see Figure 6.3). An increased ‘management

commitment’ towards safety will tend to increase the number and intensity of ‘safety

related activities’ (representing a positive influence, ‘+’ sign), which, in turn, will

significantly enhance the ‘perception of safety’ (representing a positive influence, ‘+’

sign) (Arboleda et al., 2003). A notable increase in the ‘perception of safety’ has the

potential to increase the level of ‘participation in safety activities’ (representing a

positive influence, ‘+’ sign) (Dedobbeleer and Beland, 1991). Thus, more ‘safety

resources’ are required as a result of more people participating in safety activities

(representing a positive influence, ‘+’ sign) (Pipitsupaphol and Watanabe, 2000). This

will, unfortunately, tend to put more pressure on ‘management commitment’ towards

safety (the requirement of safety resources negatively affects management’s

commitment towards safety; this is represented by a ‘-’ sign) (Speirs and Johnson,

2002). This, thus, closes a negative feedback loop (as shown by the sign) between

‘management commitment’ towards safety, ‘safety related activities’, ‘perception of

safety’, ‘participation in safety activities’, and ‘safety resources’ (see Figure 6.3).

6.3.2 A Causal Loop Diagram of the CSC Index

Figure 6.4 shows a causal loop diagram of the proposed CSC index. The loop consists

of seven elements to explain the relationships between the Enablers, Goals, and the

CSC index. These seven elements are:

+

-


123

Gap of CSC Index

Enablers Score

CSC Index Score

Goals Score

Gap of Goals

Desired CSC Index

Desired Goals

+

+

+

-

+

-

-

-

+

+

+

Figure 6.4 A causal loop diagram of the CSC index

1. Enablers score at point (t) in time (maximum 500 points): This score is equal to

the sum of the Leadership (Lds) score (maximum 100 points), the People (Ppl)

score (maximum 90 points), the Partnerships and Resources (Prs) score

(maximum 90 points), the Policy and Strategy (Pol) score (maximum 80 points),

and the Processes (Pro) score (maximum 140 points) (each enabler score is

assigned based on the EFQM Excellence model, see Figure 3.6).

Enablers score = Lds score + Ppl score + Prs score + Pol score + Pro score

2. Goals score at point (t) in time (maximum 500 points)

3. CSC index score at point (t) in time (maximum 1,000 points): This score is equal

to the sum of the Enablers score and the Goals score.

CSC index score = Enablers score + Goals score at point (t) in time

4. Desired goals score: This is the ultimate score that each and every organization

aspires to achieve. The score is set as 500 points.


124

5. Gap of goals at point (t) in time: It is equal to the difference between the Desired

goals score and Goals score at point (t) in time.

Gap of goals = Desired goals score – Goals score

6. Desired CSC index score: This score contains five values: 200, 400, 600, 800,

and 1,000 points, to match the five CSC maturity levels (see Figure 3.7 and

Section 3.5.2). Deciding which value to be used depends on the CSC index score

at that point of time. For instance, if the CSC index score at point (t) in time

equals 100, meaning that the organization is in the first maturity level, and thus

has a maximum score of 200 points. Then, the Desired CSC Index at this point

of time is set as 200 points (representing the threshold for the immediately

following maturity level).

At the first CSC maturity level, the Desired CSC index score = 200

At the second CSC maturity level, the Desired CSC index score = 400

At the third CSC maturity level, the Desired CSC index score = 600

At the fourth CSC maturity level, the Desired CSC index score = 800

At the fifth CSC maturity level, the Desired CSC index score = 1,000

7. Gap of CSC index at point (t) in time: This index score is equal to the difference

between the Desired CSC index score and CSC index at point (t) in time.

Gap of CSC index = Desired CSC index score – CSC index score

The relationships between these seven elements (see Figure 6.4) are as follows. At any

point of time, the CSC index score represents the sum of the Enablers score and the

Goals score. This score is compared with the Desired CSC index score, resulting in a

Gap of CSC index that reflects the difference between these two values. As the CSC


125

index score increases (as a result of an improvement in the Enablers score and the

Goals score), the Gap of CSC index decreases, forming a negative (-) relationship.

Following the continuous improvement cycle, the resulting decrease of the Gap of CSC

index tends to increase the Enablers score (for example, the perception of a smaller Gap

of CSC index results in a better perception of safety, which, in turn, increases the

participation in safety activities; they are then reflected as a higher Enablers score).

Thus, the relationship between the Gap of CSC index and the Enablers score is negative

(-) because a change in the Gap of CSC index results in a change in the opposite

direction of the Enablers score. The increased Enablers score, undoubtedly, enhances

the CSC index score, representing a positive (+) relationship (the changes of the two

elements are in the same direction). This then closes a positive loop ( ) between the

CSC index score, the Gap of CSC index, and the Enablers score (the loop starts and

ends by increasing the CSC index score).

Continuing with the Enablers score, the increased Enablers score improves the Goals

score (represented by a ‘+’ sign). This may be seen, for example, as proper safety

training leads to lower accidents, which, in turn, increases job satisfaction (Teo et al.,

2005). The higher Goals score, when compared with the Desired Goals score (which is

set as 500 points), results in a smaller Gap of Goals (a negative, ‘-’, relationship is

formed). The perceived smaller goals gap will tend to enhance the implementation of

the five enablers (the Enablers score increases). For example, the lower number and

cost of accidents enhance the perception of, and commitment to, safety (Turner, 1991).

Therefore, the loop between the Enablers score, the Goals score, and the Gap of Goals

is considered a positive ( ) loop, i.e. the loop starts and closes by increasing the

Enablers score.

To further explain the interactions among the five CSC enablers, an example of a more

detailed causal loop diagram, showing the relationships among these five enablers and

Goals, is illustrated below.

+

+


126

Safety Implementation(Pro)

Effective Safety Policy(Pol)

Resource Requirements(Prs)

Staff Participation(Ppl)

Management Commitment(Lds)

Cost of Accidents(Goals)

-

+++

+

++

-

+

+

Figure 6.5 A causal loop diagram of the five enablers and Goals

As shown in Figure 6.5, increased management commitment towards safety (Lds) will

tend to increase staff participation in safety activities (Ppl) (a ‘+’ link) (Teo et al.,

2005), which, in turn, will increase resource requirements (Prs) (a ‘+’ link)

(Pipitsupaphol and Watanabe, 2000). This increase in resource requirements (Prs) is

likely to enhance safety policy formulation (Pol) (a ‘+’ link), which sequentially

improves safety implementation (Pro) (a ‘+’ link). This assumption reflects the

recommendations made by Wright et al. (1999), which implied that resource

requirements are a fundamental element in formulating effective policies to improve

safety process implementation.

Further, improved safety implementation (Pro) will tend to reduce the cost of accidents

(Goals) (a ‘-’ link) (Pannirselvam and Ferguson, 2001). Undoubtedly, this reduced cost

of accidents (Goals) leads to more management commitment towards safety (Lds) (a ‘-’

link) (Turner, 1991). Thus, the feedback loop between Lds, Ppl, Prs, Pol, Pro, and

Goals is a positive loop.


127

In this study, the causal loop diagrams of the CSC were converted into a so-called

‘stock-flow’ diagram, using a SD based software package ‘STELLA’ (Ithink, 2003), to

enable the simulations. The details are described in the next section.

6.4 CONSTRUCTION SAFETY CULTURE DYNAMIC MODEL

The formulated CSC dynamic model (as shown in Figure 6.6) captures the interactions

among the five enablers and Goals, where the CSC index represents the sum of the

Enablers score and the Goals score (with an overall score of 1,000 points, see Section

6.3.2). This dynamic model reflects the assumption that the CSC index can be

‘healthier’, provided that the organization focuses on improving the five enablers and,

accordingly, achieves higher safety goals. The dynamic models of the five enablers and

Goals are discussed in detail in the following sections.

6.4.1 Leadership Dynamic Model

The Leadership dynamic model, as shown in Figure 6.7, provides a simple

representation of the stock (leadership) and flow (rlds = leadership rate) diagram (refer

to the Acronyms list – page xxi). In this model the increase in the ‘rlds’ depends on: 1)

the value of the leadership (used_lds); 2) the leadership rate fraction (rldsf); 3) the gap

of goals (ggoals); 4) the gap of leadership (glds); and 5) the percentage of more effort

provided to improve the leadership score (plds) (in the initial base run of the model, the

organization considers all five enablers as having equal significance in improving the

CSC index, so the ‘plds’ is set as zero), as shown in Equations 6.1 and 6.2 (see full

details of the SD equations in Appendix 6).

A S

yste

m D

ynam

ics

App

roac

h to

Con

stru

ctio

n Sa

fety

Cul

ture

12

8

LE

AD

ER

SH

IP

rld

s

use

d l

ds

dld

s

GO

AL

S

rgo

als

CS

C I

ND

EX

gg

oal

sd

go

als

use

d l

ds

use

d p

ol

use

d p

pl

use

d p

rs

use

d p

ro

rld

sf

PE

OP

LE

rpp

l

use

d g

oal

s

Co l

ds

pp

l

DF

pp

l ld

sg

pp

l

use

d p

pl

gg

oal

s

dp

pl

DF

goal

s pro

Co

pro

go

als

PO

LIC

Y &

ST

RA

TE

GY

gp

ol

gp

ol

DF

pol

lds

rpol

DF

po

l p

rsC

o p

rs p

ol

Co

lds

po

l use

d p

ol

dpo

l

PA

RT

NE

RS

HIP

S &

RE

SO

UR

CE

S

rprs

use

d p

rs

dp

rs

gp

rs

DF

prs

lds

Co

ld

s p

rs

EN

AB

LE

RS

DF

prs

ppl

Co p

pl

prs

gp

rs

PR

OC

ES

SE

S

rproD

F p

ro p

pl

Co

ppl

pro

use

d p

rodpro

gp

ro

DF

pro

pol

Co

po

l pro

gp

ro

gld

s

pld

s

pp

ol

pp

pl

pp

rs

pp

ro

Not

e: C

o_ld

s_po

l, C

o_ld

s_pp

l, C

o_ld

s_pr

s, C

o_po

l_pr

o, C

o_pp

l_pr

o, C

o_pp

l_pr

s, C

o_pr

o_go

als,

and

Co_

prs_

pol =

Cor

rela

tions

bet

wee

n Ld

s an

d Po

l, Ld

s an

d Pp

l, Ld

s an

d Pr

s, P

ol a

nd

Pro,

Ppl

and

Pro

, Pp

l an

d Pr

s, P

ro a

nd G

oals

, an

d Pr

s an

d Po

l, re

spec

tivel

y. D

F_go

als_

pro,

DF_

pol_

lds,

DF_

pol_

prs,

DF_

ppl_

lds,

DF_

pro_

pol,

DF_

pro_

ppl,

DF_

prs_

lds,

and

D

F_pr

s_pp

l = D

ecis

ion

frac

tions

bet

wee

n G

oals

and

Pro

, Pol

and

Lds

, Pol

and

Prs

, Ppl

and

Lds

, Pro

and

Pol

, Pro

and

Ppl

, Prs

and

Lds

, and

Prs

and

Ppl

, res

pect

ivel

y. d

goal

s, d

lds,

dpo

l, dp

pl, d

pro,

dpr

s,=

Des

ired

Goa

ls, d

esir

ed L

ds, d

esir

ed P

ol, d

esir

ed P

pl, d

esir

ed P

ro, a

nd d

esir

ed P

rs, r

espe

ctiv

ely.

ggo

als,

gld

s, g

pol,

gppl

, gpr

o, a

nd g

prs

= G

aps

of G

oals

, Lds

, Pol

, Ppl

, Pr

o, a

nd P

rs, r

espe

ctiv

ely.

pld

s, p

pol,

pppl

, ppr

o, a

nd p

prs

= Pe

rcen

tage

eff

ort p

rovi

ded

to L

ds, P

ol, P

pl, P

ro, a

nd P

rs, r

espe

ctiv

ely.

rgoa

ls, r

lds,

rpol

rppl

, rpr

o, a

nd rp

rs =

Goa

ls, L

ds, P

ol,

Ppl,

Pro,

and

Prs

rate

s, re

spec

tivel

y. u

sed_

goal

s, u

sed_

lds,

use

d_po

l, us

ed_p

pl, u

sed_

pro,

and

use

d_pr

s =

Goa

ls, L

ds, P

ol, P

pl, P

ro, a

nd P

rs v

alue

s, re

spec

tivel

y. rl

dsf =

Lds

rate

frac

tion

Figu

re 6

.6 T

he C

SC d

ynam

ic m

odel


129

Leadership

rlds

used lds

dlds

rldsf

ggoals

gldsplds

Leadership Dynamic Model

Note: See Acronyms list for the abbreviations. The symbols are described in Figure 2.6 and Table 2.1.

Figure 6.7 Leadership dynamic model

Eq. 6.1 Leadership(t) = Leadership(t - dt) + (rlds)*dt

Eq. 6.2 rlds = ((used_lds + ggoals)*rldsf) + (glds*plds)

The percentage of more effort provided to improve the Leadership score (plds) is the

effort (rather than what is normally provided) that the organization dedicates to boost

the value of Leadership to achieve its maximum score, i.e. 100 points, in a shorter

period of time. Further details are described in Section 6.5.2.

The value of the leadership rate fraction (rldsf) is constant, and was recommended by

the respondents (participating in the questionnaire survey) to be 0.08. It is the average

score derived from the ratio of budgets organizations reported to be spending in

implementing safety activities to their annual total budgets.

�

Specifically, then, Figure 6.7 may be explained as follows: when ‘ggoals’ is large (in

other words the score of Goals is low compared to the 500 targeted score), leadership

must try hard to reduce this gap. This is achieved by, for example, leaders committing

more to safety, encouraging more two-way communication, assigning safety

accountability to staff, and aligning productivity and safety targets. As a result, the

‘rlds’ increases. Naturally, the increased ‘rlds’ increases the ‘leadership’ stock, which,

in turn, increases the ‘used_lds’ value. The maximum score of ‘used_lds’ is controlled

= 0 (in base run)


130

by the maximum ‘desired value of leadership’ (dlds), which is equal to 100 points (see

Figure 5.4).

Given that Leadership is assumed to drive People, Partnerships and Resources, and

Policy and Strategy, the newly obtained ‘used_lds’ value is transferred to these three

connected dynamic models. This transferred value is, however, influenced by the

strength of the correlation between Leadership and the three enablers (People,

Partnerships and Resources, and Policy and Strategy) (see Figure 5.4). For example,

the ‘used_lds’ value that is transferred to the People dynamic model will have a value

equal to the value of the ‘used_lds’, multiplied by the correlation coefficient between

Leadership and People, which is equal to 0.64 (0.64*‘use_lds’ value). The ‘used_lds’

values transferred to the Partnerships and Resources, and Policy and Strategy dynamic

models, on the other hand, are equal to (0.16*‘use_lds’ value), and (0.59*‘use_lds’

value), respectively.

6.4.2 People Dynamic Model

The relationship between Leadership and People has been confirmed and cited by many

research studies. Little (2002), for example, stated that leaders play an important role in

changing workers’ behaviour through demonstrating strong commitment and

accountability towards safety. The psychological link between management presence on

site and workers safe behaviour gives rise to a good perception about safety, and hence

getting workers to accept more safety responsibilities (Siu et al., 2004).

The People dynamic model, shown in Figure 6.8, illustrates that the ‘used_lds’ value

(obtained from Section 6.4.1) affects the people rate (rppl), as represented in Equations

6.3 and 6.4. This, in turn, influences the ‘used_ppl’ value.


131

People

rppl

Co lds ppl

DF ppl lds gppl

used ppl

dppl

used lds

pppl

People Dynamic Model


Figure 6.8 People dynamic model

Eq. 6.3 People(t) = People(t - dt) + (rppl)*dt

Eq. 6.4 rppl = (used_lds*DF_ppl_lds) + (gppl*pppl)

‘DF_ppl_lds’, shown in Equation 6.4, depends on the ‘gap of people’ (gppl) value, and

the ‘correlation coefficient between the leadership and people’ (Co_lds_ppl) value. As

shown in Figure 5.4, the ‘Co_lds_ppl’ value equals 0.64, leading to Equations 6.5 and

6.6 of ‘DF_ppl_lds’ as below.

Eq. 6.5 DF_ppl_lds = gppl*Co_lds_ppl/100

Eq. 6.6 DF_ppl_lds = gppl*(0.64/100)

The ‘used_ppl’ score is controlled by the ‘dppl’ value (the desired score of the People

enabler, 90 points). This ‘used_ppl’ score is then transferred to the Partnerships and

Resources and Processes dynamic models (see causal links between Ppl and Prs, and

between Ppl and Pro in Figure 5.4).

= 0 (in base run)


132

6.4.3 Partnerships and Resources Dynamic Model

People has a direct effect on Partnerships and Resources, as supported by Pipitsupaphol

and Watanabe (2000), who investigated the root causes of accidents in the Thai

construction industry. They concluded that workers must be provided with adequate

safety resources to facilitate performing the job safely. The ‘used_ppl’ value, therefore,

flows into the ‘partnerships and resources rate’ (rprs), as shown in Figure 6.9 and

Equations 6.7 and 6.8 below.


rprs

used lds

used prs

dprsgprs

DF prs lds

Co lds prs

used ppl DF prs ppl

Co ppl prs

gprs

pprs

Partnerships and Resources Dynamic Model


Figure 6.9 Partnerships and Resources dynamic model

Eq. 6.7 Partnerships_&_Resources(t) = Partnerships_&_Resources(t - dt) + (rprs)*dt

Eq. 6.8 rprs = (used_lds*DF_prs_lds) + (used_ppl*

DF_prs_ppl) + (gprs*pprs)

‘DF_prs_lds’, shown in Equation 6.9, depends on the ‘gap of partnerships and

resources’ (gprs) value, and the ‘correlation coefficient between the leadership and

partnerships and resources’ (Co_lds_prs) value. ‘DF_prs_ppl’, on the other hand,

depends on the ‘gprs’ value, and the ‘correlation coefficient between the people and

partnerships and resources’ (Co_ppl_prs) value (see Equation 6.10).

= 0 (in base run)


133

Eq. 6.9 DF_prs_lds = gprs*Co_lds_prs/100

Eq. 6.10 DF_prs_ppl = gprs*Co_ppl_prs/100

Leadership also has an effect on Partnerships and Resources, although it is a weak one.

Consequently, the ‘used _lds’ value also flows into the ‘rprs’. The increase in the ‘rprs’

enhances the ‘used_prs’ value, and this, in turn, increases the ‘policy and strategy rate’

(rpol) of the Policy and Strategy dynamic model, as described in the following section.

6.4.4 Policy and Strategy Dynamic Model

Leadership and Partnerships and Resources influence the establishment of safety policy

and strategies in the organization (see Figure 5.4), leading to the flows of the ‘used_lds’

and ‘used_prs’ values into the ‘policy and strategy rate’ (rpol), as shown in Figure 6.10

and Equations 6.11 to 6.14.

used ldsPolicy and Strategy

gpolDF pol lds

used prsgpol

rpol

DF pol prs Co prs pol

Co lds pol

used pol

dpol

ppol

Policy and Strategy Dynamic Model


Figure 6.10 Policy and Strategy dynamic model


134

Eq. 6.11 Policy_&_Strategy(t) = Policy_&_Strategy(t - dt) + (rpol)*dt

Eq.6.12 rpol = (used_lds*DF_pol_lds) + (used_prs*

DF_pol_prs) + (gpol*ppol)

Eq. 6.13 DF_pol_lds = gpol*Co_lds_pol/100

Eq. 6.14 DF_pol_prs = gpol*Co_prs_pol/100

The increased ‘rpol’ increases the ‘used_pol’ value. This ‘used_pol’ value is then

transferred to the Processes dynamic model, as described in the next section.

6.4.5 Processes Dynamic Model

Both People and Policy and Strategy play a key role in the successful safety

implementation; this is consistent, to some extent, with the process management studies

of Eskildsen and Dahlgaard (2000), and Pannirselvam and Ferguson (2001), where

process management was found to be directly related to strategic planning and human

resource management.

The Processes dynamic model, as depicted in Figure 6.11, demonstrates that the

increased ‘used_ppl’ and ‘used_pol’ values tend to increase the ‘processes rate’ (rpro)

(see Equations 6.15 to 6.18).

= 0 (in base run)


135

Processes

rpro

used ppl

DF pro pplCo ppl pro

used pro

dpro

gpro

used polDF pro pol

Co pol pro

gpro

ppro

Processes Dynamic Model


Figure 6.11 Processes dynamic model

Eq. 6.15 Processes(t) = Processes(t - dt) + (rpro)*dt

Eq. 6.16 rpro = (used_ppl*DF_pro_ppl) + (used_pol*DF_pro_pol)

+ (gpro*ppro)

Eq. 6.17 DF_pro_ppl = gpro*Co_ppl_pro/100

Eq. 6.18 DF_pro_pol = gpro*Co_pol_pro/100

The increasing of ‘rpro’ will improve the ‘used_pro’ value, which, ultimately, will

enhance the ‘goals rate’ (rgoals), as described in the next section.

6.4.6 Goals Dynamic Model

In the Goals dynamic model, the Processes enabler appears to be very strongly

correlated to Goals (see Figure 5.4). The increase or decrease of the ‘used_pro’ value

has an effect on the ‘goals rate’ (rgoals), as shown in Figure 6.12 and Equations 6.19

and 6.20 below.

= 0 (in base run)


136

Goals

rgoals

ggoals

dgoals

used proused goals

DF goals proCo pro goals

CSC Index

Goals Dynamic Model


Figure 6.12 Goals dynamic model

Eq. 6.19 Goals(t) = Goals(t - dt) + (rgoals)*dt

Eq. 6.20 rgoals = used_pro*DF_goals_pro

The increase of the ‘used_goals’ value depends on the ‘rgoals’ value, which appears to

increase when the ‘used_pro’ value increases. The increased ‘used_goals’ value reduces

the ‘gap of goals’ (ggoals), which then has an effect on the ‘rlds’ of the Leadership

dynamic model (see Figure 6.7).

The simulations of the CSC dynamic model iterate as cycles, from the Leadership to the

Goals dynamic models. In each cycle, the Enablers score and the CSC index are

calculated, as illustrated in Figure 6.13 and Equations 6.21 and 6.22 below. The cycles

continue until the CSC index reaches a score of 800 or more, indicating that the

organization has achieved the ‘continually improving’ (fifth) maturity level’.


137

CSC Index

used lds

used polused ppl used prs

used pro

used goalsEnablers

CSC Index Dynamic Model


Figure 6.13 The CSC index dynamic model

Eq. 6.21 Enablers = used_lds + used_ppl + used_prs + used_pol + used_pro

Eq. 6.22 CSC index = Enablers + used_goals

The next section (Section 6.5) describes the simulation results of the CSC dynamic

model.

6.5 DYNAMIC SIMULATION RESULTS

6.5.1 Base Run Results

The CSC dynamic model was simulated using ‘STELLA’ software (Ithink, 2003). In

the ‘base run’ simulation, the initial values of the five enablers were set as zero to

manipulate the situation of organizations with no prior safety implementation. The

initial value of Goals, however, was not equal to zero. The linear regression was

performed with the SPSS program, using data from the questionnaire survey between

the Enablers score (as a sum data score of the five enablers) and the Goals score (see


138

full details of linear regression results in Appendix 7). The results reveal that when the

Enablers value (which has a maximum score of 500 points) equals the hypothetical

value of zero, the Goals value (which also has a maximum score of 500 points) is equal

to 68 points (see Equation 6.23 below). This may be explained that some of the four

Goals’ attributes (including, ‘number of accidents’, ‘industrial image’, ‘workforce

morale’, and ‘cost of accidents’, see Figure 5.3) do not totally and exclusively depend

on the implementation of the five enablers. For example, organizations may experience

low ‘number of accidents’ due to low workload or sheer luck.

Eq. 6.23 Goals = 68 + (0.79*Enablers)

The initial values of Enablers (which was set at zero) and Goals (68 points) were

substituted in the SD equations (see Appendix 5). The dynamic model was simulated,

and the results are displayed graphically in Figures 6.14 to 6.16, and numerically in

Tables 6.1 and 6.2. The time units used in the simulation can be varied as SD model

allows the users to define their own units. For simplicity, however, the time unit used in

this study will be referred to as ‘years’, henceforth. Nevertheless, it is worth noting that

this ‘years’ unit does not represent the calendar year.

Page 11.00 5.00 9.00 13.00 17.00

Years

1:

1:

1:

0

250

500

1: Enablers Score

1

1

1

1

Figure 6.14 Graphical results of the Enablers score over time


139

Page 2

1.00 5.00 9.00 13.00 17.00

Years

1:

1:

1:

68

284

500

1: used goals score

1 1

1

1

Figure 6.15 Graphical results of the Goals score over time

Page 3

1.00 5.00 9.00 13.00 17.00

Years

1:

1:

1:

68

534

1000

1: CSC Index Score

1

1

1

1

Figure 6.16 Graphical results of the CSC index over time


140

Table 6.1 Simulation results of the five enablers and Goals

Year Score

Lds Pol Ppl Prs Pro Goals

Initial 0.00 0.00 0.00 0.00 0.00 68.00

1 2.64 0.47 0.56 0.21 0.13 68.00

2 5.50 2.36 2.66 1.68 1.82 68.13

3 8.55 6.14 6.30 5.52 7.31 69.12

4 11.73 12.28 11.40 12.50 18.66 72.08

5 14.84 20.91 17.70 22.68 36.73 77.84

6 23.20 31.92 25.42 35.34 60.02 114.61

7 30.80 44.85 35.61 49.67 85.27 156.79

8 42.10 57.14 46.98 63.51 107.57 244.71

9 52.23 66.79 58.16 74.71 123.35 329.25

10 60.37 73.17 68.07 82.23 132.44 381.64

11 70.79 76.76 75.77 86.46 136.86 461.90

12 78.61 78.59 81.40 88.53 138.78 491.36

13 85.54 79.42 85.07 89.43 139.54 498.07

14 92.69 79.78 87.32 89.79 139.83 499.57

15 100.00 79.92 88.62 89.92 139.94 499.91

16 100.00 79.97 89.31 89.97 139.98 499.98

17 100.00 79.99 89.66 89.99 139.99 500.00

18 100.00 80.00 89.83 90.00 140.00 500.00

19 100.00 80.00 89.91 90.00 140.00 500.00

20 100.00 80.00 89.96 90.00 140.00 500.00

21 100.00 80.00 89.98 90.00 140.00 500.00

22 100.00 80.00 89.99 90.00 140.00 500.00

23 100.00 80.00 89.99 90.00 140.00 500.00

24 100.00 80.00 90.00 90.00 140.00 500.00

Note: Maximum scores of Lds, Pol, Ppl, Prs, and Pro are 100, 80, 90, 90, and 140 points, respectively.


141

Table 6.2 Simulation results of the enablers, Goals, and CSC index

Year Enablers Used_Goals CSC Index Level*

Score %Increasing Score %Increasing Score %Increasing

Initial 0.00 - 68.00 - 68.00 - 1st

1 4.00 0.80 68.00 0.00 71.98 0.40 1st

2 14.01 2.00 68.13 0.03 82.14 1.02 1st

3 33.82 3.96 69.12 0.20 102.94 2.08 1st

4 66.57 6.55 72.08 0.59 138.65 3.57 1st

5 112.86 9.26 77.84 1.15 190.71 5.21 1st

6 175.90 12.61 114.61 7.35 290.52 9.98 2nd

7 246.21 14.06 156.79 8.44 403.00 11.25 3rd

8 317.30 14.22 244.71 17.58 562.01 15.90 3rd

9 375.24 11.59 329.25 16.91 704.49 14.25 4th

10 416.27 8.21 381.64 10.48 797.91 9.34 4th

11 446.64 6.07 461.90 16.05 908.55 11.06 5th

12 465.90 3.85 491.36 5.89 957.26 4.87 5th

13 479.00 2.62 498.07 1.34 977.07 1.98 5th

14 489.40 2.08 499.57 0.30 988.97 1.19 5th

15 498.40 1.80 499.91 0.07 998.31 0.93 5th

16 499.23 0.17 499.98 0.01 999.21 0.09 5th

17 499.63 0.08 500.00 0.00 999.63 0.04 5th

18 499.82 0.04 500.00 0.00 999.82 0.02 5th

19 499.91 0.02 500.00 0.00 999.91 0.01 5th

20 499.96 0.01 500.00 0.00 999.96 0.01 5th

21 499.98 0.00 500.00 0.00 999.98 0.00 5th

22 499.99 0.00 500.00 0.00 999.99 0.00 5th

23 499.99 0.00 500.00 0.00 999.99 0.00 5th

24 500.00 0.00 500.00 0.00 1,000.00 0.00 5th

Note: Bold numbers refer to the time unit, where the organization reaches the fifth level of CSC maturity. (*) Level = CSC maturity level

As shown in Figures 6.14 through to 6.16, at the starting point, the Enablers value was

zero and the Goals value was 68, leading to the CSC index of 68 points. At this stage,

the gap of the Goals (ggoals) value was relatively large (500 – 68 = 432 points). This,

then, boosted the value of Leadership (see Equation 6.2), which, in turn, increased the


142

values of the remaining four enablers, i.e. People, Policy and Strategy, Partnerships

and Resources, and Processes.

As the five enablers’ values increased (identifying an improvement in safety culture’s

implementations), the Goals value, and the CSC index increased. The simulation

continued until the CSC index reached the maximum score of 1,000 points (the

Enablers and Goals values reach their maximum 500 points). Table 6.2 showed that it

took 11 years for the organization, with a non-existent safety policy and safety

implementation process, to progress from the first to the fifth levels of CSC maturity

(the CSC index reached 800 points or more at the end of year 11).

The graphs shown in Figures 6.14 through to 6.16 showed similar S-shaped patterns,

with a slow increase at the beginning of the simulation. It took six years for the

organization to progress from the first to the second levels of culture maturity. This

result demonstrated that for an organization with a non-existent safety culture policy

and implementation process, it was hard to improve the CSC in the early stage of the

safety implementation. This is shown by a slow increase in the rate for Enablers, and

the even slower increased rate for Goals (see Table 6.2). After the organization reached

the second maturity level, however, the Enablers and Goals values increased rapidly, as

depicted by the sharp rises in the curves shown in Figures 6.14 and 6.15, respectively.

This, in turn, enhanced the CSC index, which could be seen as a steep incline of the

graph shown in Figure 6.16. The organization progressed from the second to the fifth

maturity levels over five years (at the end of year 11), showing a significant safety

improvement in the organization.

After year 11, it was difficult for the organization to increase the Enablers value, as

most of the safety implementations were accomplished. Moreover, the extra effort

needed to further improve safety in the organization might be switched to other

important areas. This, in turn, slows the increase rates of the Goals value and the CSC

index (see Figures 6.15 and 6.16, respectively). As shown in Table 6.2, the organization

achieved its CSC index of 1,000 points (representing the perfect safety implementation)

at the end of year 24. It appears to be very challenging to reach a perfect safety


143

implementation; however, an organization can plan its safety implementation to

progress through to the fifth CSC maturity level by using the time frame shown in Table

6.2.

6.5.2 Base Run Results Examination

By observing the increasing rate of the five enablers’ values at the early stage of the

simulation (see Table 6.1 and Figure 6.17), it is clear that Leadership was the weakest

enabler in boosting the CSC index, as it produced the least scores compared with the

other four enablers. To explain, at the end of year six (when the organization reached

the second maturity level), the Leadership’s score was 23.20 out of 100 points,

representing 23.2% of the scores produced. On the other hand, People, Policy and

Strategy, Partnerships and Resources, and Processes produced 28.24, 39.90, 39.27, and

42.87% of their maximum scores, respectively. Therefore, to expedite the Enablers

value, and achieve higher CSC index in the early stages, an organization should focus

more on improving its Leadership.

Page 1

1.00 5.00 9.00 13.00 17.00

Years

1:

1:

1:

2:

2:

2:

3:

3:

3:

4:

4:

4:

5:

5:

5:

0

50

100

0

45

90

0

40

80

0

70

140

1: used lds score 2: used ppl score 3: used prs score 4: used pol score 5: used pro score

1

1

1

1

2

2

2

2

3

3

3

3

4

4

44

5

5

5 5

Figure 6.17 Graphical results of the five enablers over time


144

To confirm whether the organization should concentrate on improving the

implementation of Leadership, a number of model runs were needed. First, the

organization was said to allocate 10% of more effort to focus on Leadership

improvement, i.e. apart from normally implementing this enabler, the organization puts

more time and effort (by 10%) into further enhancing this particular enabler’s

implementation. This means that the organization maintained its improvement of the

five enablers, but more attention was given to Leadership.

Consequently, the ‘plds’ value was set to 0.1, while the ‘pppl’, ‘pprs’, ‘ppol’, and ‘ppro’

were still set as zero. The CSC dynamic model was then simulated, and the results were

recorded. Next, the ‘plds’ was set back to zero, then the ‘pppl’ was set as 0.1 (meaning

that the organization now changed its focus, from improving Leadership’s

implementation, to the People enabler). The model was re-simulated, and the results

were recorded, then the ‘pppl’ was set back to zero.

The simulations were performed for all five enablers; the results (shown in Table 6.3)

demonstrate that, by focusing more on Leadership, the organization reached the second

maturity level in a much shorter time (three years), and achieved the fifth level of

maturity four years earlier (it took seven years, instead of 11 years, to achieve the fifth

maturity level). Therefore, for the organization starting at level one of CSC maturity,

attention should be paid, in the main, to improving the key attributes of Leadership to

successfully progress through to higher maturity levels. The leaders should thus take

safety seriously through being a role model (Dunlap, 2004; Teo et al., 2005), assigning

and communicating safety responsibilities clearly to all staff (Lardner et al., 2001), and

ensuring that the workload was reasonably balanced among workers to avoid unsafe

behaviours (Glendon and Litherland, 2001).


145

Table 6.3 Experimentation with extra efforts given to improve the five enablers

Year Lds Pol Ppl Prs Pro

Index* Level* Index Level Index Level Index Level Index Level Initial 68.0 1st 68.0 1st 68.0 1st 68.0 1st 68.0 1st 1 87.0 1st 82.3 1st 85.5 1st 81.7 1st 86.9 1st

2 131.7 1st 109.0 1st 121.5 1st 104.5 1st 113.3 1st

3 210.9 2nd 150.0 1st 177.8 1st 141.1 1st 147.9 1st

4 357.0 2nd 203.1 2nd 272.4 2nd 193.5 1st 189.9 1st

5 540.4 3rd 315.4 2nd 394.6 2nd 290.7 2nd 282.8 2nd

6 735.8 4th 414.3 3rd 549.4 3rd 401.7 3rd 377.4 2nd

7 897.1 5th 561.6 3rd 700.7 4th 558.6 3rd 498.5 3rd

8 970.5 5th 693.8 4th 794.9 4th 699.8 4th 602.1 4th

9 991.1 5th 781.9 4th 903.3 5th 792.6 4th 740.2 4th

10 997.0 5th 876.1 5th 949.4 5th 902.9 5th 797.0 4th

11 998.9 5th 937.2 5th 967.1 5th 951.8 5th 896.9 5th

Note: Bold number refers to the CSC index reaching the fifth level of CSC maturity. (*) Level = CSC maturity level, (*) Index = CSC index

6.5.3 Model Verification and Validation

The CSC dynamic model was verified using the ‘logical’ test to assure its parameters,

the unit consistency, and the correct sequence of calculation (see Section 2.2.7.4). The

six constructs (the five enablers and the single set of Goals) and their attributes were

tested and confirmed by two statistical analyses: the EFA and the SEM. The ‘years’

time unit was consistently used throughout the simulation, and the sequence of the

calculation was correct, following the directional influences shown in Figure 5.4.

To validate the CSC dynamic model, a behavioural sensitivity analysis was conducted

(see Section 2.2.7.4) to test the robustness of the model, by ensuring that the

uncertainties and the estimating errors did not significantly affect the overall behaviour

of the model. It tested the limits of the model and its ability to adjust itself in response

to the changes. According to Tang and Ogunlana (2003a), a model is considered robust


146

if its behaviour does not change drastically, when a parameter or behavioural

relationship is altered.

The sensitivity analysis of the CSC dynamic model was carried out by changing the

initial value of Leadership (by 25, 50, and 75%), meaning that the initial value was

changed from zero to 25, 50, and 75 points (out of a maximum of 100 points),

respectively. The simulation results, displayed graphically in Figures 6.18 and 6.19,

demonstrate that the change in the initial values of the Leadership only numerically

affects the model behaviour, not the pattern of the model, thus validating the model.

Page 2

1.00 5.00 9.00 13.00 17.00

Years

1:

1:

1:

0

50

100

used lds score: 1 - 2 - 3 - 4 -

1

1

1

1

2

2

2 2

3

3 3 34 4 4 4

Note: The numbers 1, 2, 3, and 4 displayed in the figure represent the initial values of Lds of zero, 25, 50, and 75 points, respectively.

Figure 6.18 Sensitivity results of the ‘used_lds’ value when its initial value is changed


147

Page 8

1.00 5.00 9.00 13.00 17.00

Years

1:

1:

1:

0

500

1000

CSC Index Score: 1 - 2 - 3 - 4 -

1

1

1

1

2

2

2 2

3

33 3

4

4 4 4

Note: The numbers 1, 2, 3, and 4 displayed in the figure represent the initial values of Lds of zero, 25, 50, and 75 points, respectively.

Figure 6.19 Sensitivity results of the CSC index when the initial values of Lds are

changed

Sensitivity analyses (changing the initial values of People, Partnerships and Resources,

Policy and Strategy, and Processes) were also conducted; the results show that the

patterns of the model behaviour were not sensitive to the changes in all parameters (see

Appendix 8 for graphical results of the sensitivity analyses).

Changing the percentage of more effort provided to improve Leadership (plds) from

zero to 10, 20, and 30%, respectively, through another sensitivity test achieved the

simulation results shown in Figures 6.20 and 6.21, thus proving the non-sensitivity of

the model behaviour.


148

Page 2

1.00 5.00 9.00 13.00 17.00

Years

1:

1:

1:

0

50

100

used lds score: 1 - 2 - 3 - 4 -

1

1

1

1

2

2

2 2

3

3 3 3

4

4 4 4

Note: The numbers 1, 2, 3, and 4 displayed in the figure represent ‘plds’ of zero, 0.1, 0.2, and 0.3, respectively.

Figure 6.20 Sensitivity results of the ‘used_lds’ value when the ‘plds’ value is changed

Page 8

1.00 5.00 9.00 13.00 17.00

Years

1:

1:

1:

0

500

1000


1

1

1

1

2

2

2 2

3

3

3 3

4

44 4

Note: The numbers 1, 2, 3, and 4 displayed in the figure represent ‘plds’ of zero, 0.1, 0.2, and 0.3, respectively.

Figure 6.21 Sensitivity results of the CSC index when the ‘plds’ value is changed


149

The results of the sensitivity analyses achieved by changing the values of the ‘pppl’,

‘pprs’, ‘ppol’, and ‘ppro’ from zero to 10, 20, and 30%, respectively, are illustrated in

Appendix 9. The results demonstrate the non-sensitivity of the pattern of the model.

The following chapter describes a number of model applications of the CSC dynamic

model. Several policy experiments were performed to underline the areas requiring

safety improvement and an enhancement of the CSC index. The cyclical style of safety

management that reflects management withdrawing attention from safety is also

modelled in the next chapter.


150



151

77 SSYYSSTTEEMM DDYYNNAAMMIICCSS MMOODDEELLLLIINNGG OOFF

CCOONNSSTTRRUUCCTTIIOONN SSAAFFEETTYY CCUULLTTUURREE -- MMOODDEELL AAPPPPLLIICCAATTIIOONNSS


This chapter presents the verified and validated CSC dynamic model experiment, along

with a set of safety policies examining their impact under different scenarios to

highlight areas for safety improvement, and to enhance the CSC index. The cyclical

style of safety management, modelled to reflect real-life situations, where management

tends to withdraw its attention away from safety following the realisation of excellent

performance record, is also discussed.

Section 7.2 (below) describes a number of policy analyses, as well as their simulation

results.

7.2 POLICY ANALYSIS

The five CSC enablers include those necessary elements that enable the organization to

improve its safety performance. Each enabler comprises a number of attributes, such as

management commitment, workers’ involvement, resource availability, and so on. Any

positive interventions in these attributes are expected to enable the organization to meet

its safety goals. These interventions can be in the form of improving safety training,

providing more safety resources, integrating safety in business goals, etc.

This study grouped all such interventions under enabler-specific improvement efforts,

and referring to these as a set of policy experiments. The formulation of an effective set

of policies through the dynamic model simulation for an improved and sustained


152

organizational performance was the main objective of SD modelling (Tang and

Ogunlana, 2003b). The simulation was, therefore, used as a means for experimentation

with the model, understanding its behaviour, and identifying a framework for policy

intervention (Saeed and Brooke, 1996).

A set of policy experiments, along with comparative simulations, to demonstrate

improved performance, are described below. Two organizations (‘A’ and ‘B’) were

randomly chosen, among the 101 construction-contracting organizations responding to

the questionnaire, to represent two different maturity levels: organization ‘A’ is

currently in the second (managing) level of maturity, and organization ‘B’ is in the third

(involving) maturity level. The SD analysis was undertaken for each of the two

organizations in which the simulation results were called ‘the base run’. A number of

policies were then made for each organization in an attempt to enhance their safety

efforts.

Section 7.2.1 describes the base run for each of these organizations (‘A’ and ‘B’),

including the time period each organization (with different initial Enablers and Goals

values) needs to use to progress through to the fifth CSC maturity level.

7.2.1 Base Run

7.2.1.1 Base Run for Organization ‘A’

The initial values of the five enablers and Goals of organization ‘A’ (referring to the

questionnaire database) were as follows:

� ‘used_lds’ = 20.0 (out of 100 points)

� ‘used_ppl’ = 43.2 (out of 90 points)


153

� ‘used_prs’ = 18.0 (out of 90 points)

� ‘used_pol’ = 19.2 (out of 80 points)

� ‘used_pro’ = 37.3 (out of 140 points)

� ‘used_goals’ = 129.0 (out of 500 points)

These scores were summed and gave rise to the base value for the CSC index of 266.7

points (the CSC index is the sum of the Enabler and Goals scores, see Section 6.3.2).

Organization ‘A’ was, therefore, currently at the second maturity level (see the score-

range in each maturity level in Section 3.5.2). The initial values of the five enablers,

Goals, and CSC index were then substituted in the SD equations (see SD equations of

organization ‘A’ in Appendix 10). In the ‘base run’, the organization considered all five

enablers as having equal significance in improving the CSC index, i.e. the ‘plds’, ‘pppl’,

‘ppol’, ‘pprs’ and ‘ppro’ were all set as zero. The dynamic model was run, leading to

the simulation results displayed numerically in Tables 7.1 and 7.2, and graphically in

Figures 7.1 to 7.4, respectively.

Table 7.1 Simulation results of the five enablers of organization ‘A’


Score Gap* Score Gap Score Gap Score Gap Score Gap

Initial 20.00 80.00 19.20 60.80 43.20 46.80 18.00 72.00 37.30 102.7

1 26.73 73.27 32.33 47.67 49.62 40.38 44.20 45.80 69.32 70.68

2 36.89 63.11 47.48 32.52 56.87 33.13 63.08 26.92 97.43 42.57

3 46.23 53.77 60.53 19.47 64.56 25.44 75.59 14.41 117.83 22.17

4 55.50 44.5 69.59 10.41 71.80 18.2 82.99 7.01 129.83 10.17

5 64.95 35.05 74.87 5.13 77.70 12.3 86.87 3.13 135.76 4.24

6 73.20 26.80 77.68 2.32 82.26 7.74 88.71 1.29 138.34 1.66

7 79.89 20.11 79.01 0.99 85.38 4.62 89.50 0.50 139.38 0.62

8 86.62 13.38 79.60 0.40 87.38 2.62 89.81 0.19 139.77 0.23

9 93.79 6.21 79.85 0.15 88.58 1.42 89.93 0.07 139.92 0.08

Note: Bold numbers refer to the time unit where the organization reaches its fifth level of CSC maturity. (*) Gap = the difference between the maximum and the achieved scores


154

Table 7.2 Simulation results of the Enablers, Goals, and CSC index of organization ‘A’

Year Score CSC Maturity Level

Enablers Goals CSC Index

Initial 137.70 129.00 266.70 2nd

1 222.21 155.75 377.96 2nd

2 301.76 226.78 528.54 3rd

3 364.73 300.88 665.61 4th

4 409.70 372.86 782.56 4th

5 440.16 444.63 884.79 5th

6 460.18 487.31 947.49 5th

7 473.15 497.16 970.31 5th

8 483.18 499.37 982.55 5th

9 492.07 499.86 991.93 5th

Note: Bold numbers refer to the time unit where the organization reaches its fifth level of CSC maturity.

Page 1

1.00 3.25 5.50 7.75 10.00

Years

1:

1:

1:

2:

2:

2:

3:

3:

3:

4:

4:

4:

5:

5:

5:

20

60

100

19

50

80

43

67

90

18

54

90

37

89

140

1: used lds score 2: used pol score 3: used ppl score 4: used prs score 5: used pro score

1

1

1

1

2

2

2

2

3

3

3

3

4

4

4 4

5

5

5 5

Figure 7.1 Graphical results of the five enablers of organization ‘A’ over time


155

Page 2

1.00 3.25 5.50 7.75 10.00

Years

1:

1:

1:

138

319

500

1: Enablers Score

1

1

1

1

Figure 7.2 Graphical results of the Enablers score of organization ‘A’ over time

Page 3

1.00 3.25 5.50 7.75 10.00

Years

1:

1:

1:

129

315

500

1: used goals score

1

1

1

1

Figure 7.3 Graphical results of the Goals score of organization ‘A’ over time


156

Page 4

1.00 3.25 5.50 7.75 10.00

Years

1:

1:

1:

267

633

1000

1: CSC Index Score

1

1

1

1

Figure 7.4 Graphical results of the CSC index of organization ‘A’ over time

The results show that it took five years for organization ‘A’ to progress from the second

to the fifth maturity levels. To demonstrate, it took two years for the organization to

reach the third maturity level. The organization then advanced through to the fourth

level in one year, and reached the fifth level of maturity at the end of year five.

When organization ‘A’ reached the fifth CSC maturity level, the scores of Partnerships

and Resources, Processes, and Policy and Strategy were close to their maximum scores

(the gaps between their maximum and achieved values were small, see Table 7.1), while

the gaps of the Leadership and People values were relatively large. Thus, to plan for

safety improvement and achieve the fifth maturity level in a shorter time frame, the

organization should pay more attention to improving the Leadership and People

enablers. This is further explained in Section 7.2.2.

The next section describes the base run simulation of organization ‘B’.


157

7.2.1.2 Base Run for Organization ‘B’

The initial values of organization ‘B’ were: ‘used_lds’ = 85.0 points, ‘used_ppl’ = 43.2

points, ‘used_prs’ = 40.5 points, ‘used_pol’ = 35.2 points, ‘used_pro’ = 56.0 points and

‘used_goals’ = 200.0 points, leading to an initial CSC index of 459.9 points.

Organization ‘B’ was, therefore, at the third level of CSC maturity (see Section 3.5.2 for

the score-ranges of the five maturity levels). The initial value of Leadership was

relatively high (85 out of 100 points), demonstrating a strong management commitment

to safety.

The initial values were substituted in the SD equations, and the dynamic model was

simulated (see SD equations of organization ‘B’ in Appendix 11). The simulation

results are presented numerically in Tables 7.3 and 7.4, and graphically in Figures 7.5 to

7.8, respectively.

Table 7.3 Simulation results of the five enablers of organization ‘B’



Initial 85.00 15.00 35.20 44.80 43.20 46.80 40.50 49.50 56.00 84.00

1 98.72 1.28 59.56 20.44 64.95 25.05 63.96 26.04 91.96 48.04

2 100.00 0.00 72.05 7.95 77.50 12.50 78.85 11.15 118.47 21.53

3 100.00 0.00 77.08 2.92 83.77 6.23 85.71 4.29 131.50 8.50

4 100.00 0.00 78.95 1.05 86.90 3.10 88.44 1.56 136.83 3.17

5 100.00 0.00 79.63 0.37 88.46 1.54 89.45 0.55 138.85 1.15

6 100.00 0.00 79.87 0.13 89.23 0.77 89.81 0.19 139.59 0.41

7 100.00 0.00 79.95 0.05 89.62 0.38 89.93 0.07 139.85 0.15

8 100.00 0.00 79.98 0.02 89.81 0.19 89.98 0.02 139.95 0.05

9 100.00 0.00 79.99 0.01 89.91 0.09 89.99 0.01 139.98 0.02



158

Table 7.4 Simulation results of the Enablers, Goals, and CSC index of organization ‘B’



Initial 259.90 200.00 459.90 3rd

1 379.14 249.42 628.56 4th

2 446.86 347.65 794.51 4th

3 478.06 449.15 927.21 5th

4 491.12 487.88 979.00 5th

5 496.38 497.25 993.63 5th

6 498.49 499.39 997.88 5th

7 499.35 499.86 999.22 5th

8 499.72 499.97 999.69 5th

9 499.87 499.99 999.87 5th


Page 1

1.00 3.25 5.50 7.75 10.00

Years

1:

1:

1:

2:

2:

2:

3:

3:

3:

4:

4:

4:

5:

5:

5:

85

93

100

35

58

80

43

67

90

41

65

90

56

98

140

1: used lds score 2: used pol score 3: used ppl score 4: used prs score 5: used pro score

1

1 1 1

2

2

2 2

3

3

3 3

4

44 4

5

55 5

Figure 7.5 Graphical results of the five enablers of organization ‘B’ over time


159

Page 2

1.00 3.25 5.50 7.75 10.00

Years

1:

1:

1:

260

380

500

1: Enablers Score

1

1

1 1

Figure 7.6 Graphical results of the Enablers score of organization ‘B’ over time

Page 3

1.00 3.25 5.50 7.75 10.00

Years

1:

1:

1:

200

350

500

used goals score: 1 -

1

1

1 1

Figure 7.7 Graphical results of the Goals score of organization ‘B’ over time


160

Page 4

1.00 3.25 5.50 7.75 10.00

Years

1:

1:

1:

460

730

1000

CSC Index Score: 1 -

1

1

1 1

Figure 7.8 Graphical results of the CSC index of organization ‘B’ over time

It took three years for organization ‘B’ to progress from the third (involving) to the fifth

(continually improving) maturity levels (with a CSC index of 927.2 points at the end of

year three, see Table 7.4). The scores of the five enablers at the end of year three

indicated that organization ‘B’ should place more attention on the Processes and People

enablers, as they produced the largest score-gaps compared with the other three enablers

(see Table 7.3).

The following section describes this in more detail with a comparison between the base

run results of organizations ‘A’ and ‘B’. A number of policy analyses were performed

for each organization to enhance its safety performance, achieve a higher CSC index,

and reach the fifth maturity level in a shorter period of time.

7.2.2 Policy Experiments between Organizations ‘A’ and ‘B’

As stated earlier, SD modelling can assist in testing alternative strategies to improve

organizational safety culture, without having to implement them. Consequently,


161

organizations ‘A’ and ‘B’ may need to experiment with different safety policy scenarios

to enhance their CSC indices, and select the best policy that matches their situation.

The next section presents the policy experimentations undertaken by organization ‘A’ to

reach the fifth maturity level within three years; the same as that achieved by

organization ‘B’.

7.2.2.1 Policy Experiments of Organization ‘A’

Organization ‘A’ took five years to mature (to reach the fifth level of CSC maturity, see

Table 7.2). To enable organization ‘A’ to plan to reach the CSC maturity earlier (such

as within three years), a number of policies experimentations need to be conducted.

The gaps of the Leadership and People values were relatively large, when compared

with those of the other three enablers (see Table 7.1). With this in mind, planning for

safety improvement should be performed in the Leadership and People areas, if

organization ‘A’ expects to achieve a higher CSC index value, and reach the fifth

maturity level in a shorter time period. Simulations, with different policy scenarios,

focusing on improving those two enablers (Leadership and People) may be conducted

to achieve the most effective policy. Examples of the policy experiments are described

below.

In reaction to the large gap in the Leadership score, the organization must allocate more

effort to improving the leadership’s four attributes (‘commitment’, ‘communication’,

‘accountability’ and ‘safety and productivity alignment’). The initial values of the five

enablers, Goals, and CSC index reflected the baseline scores listed in Section 7.2.1.1,

with the ‘plds’ then set to, say, as 0.1 (representing the 10% more effort). The dynamic

model was run, and the simulation results predict that the organization reached the CSC

maturity one year earlier (see Tables 7.5 and 7.6) (achieved the fifth level within four

years; one year faster than the base run results, see Table 7.2).


162

Table 7.5 Simulation results of the five enablers of organization ‘A’ with ‘plds’ = 0.1



Initial 20.00 80.00 19.20 60.80 43.20 46.80 18.00 72.00 37.30 102.70

1 34.40 65.60 33.22 46.78 50.41 39.59 44.51 45.49 69.45 70.55

2 51.33 48.67 50.25 29.75 59.72 30.28 64.16 25.84 98.33 41.67

3 67.81 32.19 64.23 15.77 69.35 20.65 77.08 12.92 119.21 20.79

4 82.58 17.42 72.87 7.13 77.44 12.56 84.31 5.69 131.00 9.00

5 95.53 4.47 77.18 2.82 83.15 6.85 87.75 2.25 136.45 3.55

6 100.00 0.00 78.98 1.02 86.55 3.45 89.17 0.83 138.67 1.33

7 100.00 0.00 79.64 0.36 88.28 1.72 89.71 0.29 139.52 0.48

8 100.00 0.00 79.87 0.13 89.14 0.86 89.90 0.10 139.83 0.17

9 100.00 0.00 79.95 0.05 89.57 0.43 89.96 0.04 139.94 0.06



with ‘plds’ = 0.1



Initial 137.70 129.00 266.70 2nd

1 232.00 155.76 387.76 2nd

2 323.78 227.08 550.87 3rd

3 397.70 319.96 717.65 4th

4 448.19 407.42 855.61 5th

5 480.06 477.79 957.86 5th

6 493.38 494.94 988.32 5th

7 497.14 498.87 996.02 5th

8 498.74 499.75 998.49 5th

9 499.43 499.95 999.38 5th


Leadership still had the least score (in other words, the gap of the Leadership score was

the largest), when compared with the other four enablers (see Table 7.5). This outcome


163

may indicate that more effort was still needed to further improve the Leadership value.

Thus, the organization may, for example, provide 20% more effort (instead of the 10%

given in the last simulation) to improving this particular enabler. The ‘plds’ was then set

at 0.2. The model was simulated; the results are shown in Tables 7.7 and 7.8.

Table 7.7 Simulation results of the five enablers of organization ‘A’ with ‘plds’ = 0.2



Initial 20.00 80.00 19.20 60.80 43.20 46.80 18.00 72.00 37.30 102.70

1 41.50 58.50 34.08 45.92 51.16 38.84 44.80 45.20 69.58 70.42

2 63.29 36.71 52.58 27.42 62.10 27.90 65.09 24.91 99.13 40.87

3 83.52 16.48 66.88 13.12 72.79 17.21 78.23 11.77 120.30 19.70

4 99.71 0.29 74.76 5.24 80.71 9.29 85.16 4.84 131.79 8.21

5 100.00 0.00 78.11 1.89 85.37 4.63 88.19 1.81 136.85 3.15

6 100.00 0.00 79.33 0.67 87.70 2.30 89.35 0.65 138.84 1.16

7 100.00 0.00 79.76 0.24 88.85 1.15 89.77 0.23 139.58 0.42

8 100.00 0.00 79.92 0.08 89.43 0.57 89.92 0.08 139.85 0.15

9 100.00 0.00 79.97 0.03 89.72 0.28 89.97 0.03 139.95 0.05



164


with ‘plds’ = 0.2



Initial 137.70 129.00 266.70 2nd

1 241.12 155.76 396.88 2nd

2 342.19 227.36 569.55 3rd

3 421.73 333.83 755.55 4th

4 472.13 431.83 903.96 5th

5 488.53 483.78 972.31 5th

6 495.21 496.32 991.54 5th

7 497.97 499.18 997.15 5th

8 499.11 499.82 998.93 5th

9 499.60 499.96 999.56 5th


By setting the ‘plds’ = 0.2, the organization reached the fifth CSC maturity level in four

years (see Tables 7.7 and 7.8) (which was the same time frame of when the ‘plds’ =

0.1). However, the scores of the five enablers, as well as the CSC index, at the end of

year four, appeared to be higher than those obtained when the ‘plds’ = 0.1 (see Tables

7.5 and 7.6).

Consequently, the organization needed further experiments with different ‘extra’ efforts

for improving the scores of the five enablers to achieve a higher Goals score and reach

the CSC maturity as planned.

Three examples of policy scenarios that organization ‘A’ may apply to enhance the CSC

index, and achieve the fifth CSC maturity level within three years are shown below.

� Providing 20% more of effort to improve the Leadership, People, and Processes

enablers (‘plds’, ‘pppl’, and ‘ppro’ = 0.2). The CSC index at the end of year three


165

was 802.21 points (the organization reached the fifth CSC maturity level in three

years), as shown in Table 7.9.

� Providing 30, 10, and 10% more of effort to increase the Leadership, People, and

Processes scores, respectively (‘plds’ = 0.3, ‘pppl’ = 0.1, and ‘ppro’ = 0.1). The

organization reached the fifth maturity level, with the CSC index of 804.37 points,

at the end of year three (see Table 7.10).

� Providing 20% more of effort to improving the Leadership and Processes enablers

(‘plds’ and ‘ppro’ = 0.2), and 10% more of effort to increase the People,

Partnership and Resources, and Policy and Strategy scores (‘pppl’, ‘pprs’, and

‘ppol’ = 0.1). The CSC index at the end of year three was 805.59 points (see Table

7.11).

Table 7.9 Simulation results of organization ‘A’ with ‘plds’, ‘pppl’, and ‘ppro’ = 0.2

Year Score Level*

Lds Pol Ppl Prs Pro Enablers Goals CSC Index

Initial 20.00 19.20 43.20 18.00 37.30 137.70 129.00 266.70 2nd

1 41.45 34.13 58.68 46.20 84.80 265.26 158.80 424.06 3rd

2 63.67 53.03 72.14 67.97 115.96 372.78 247.84 620.63 4th

3 83.14 67.34 81.33 80.62 131.49 443.92 358.29 802.21 5th

4 99.51 75.02 86.35 86.45 137.39 484.71 466.35 951.07 5th

5 100.00 78.21 88.58 88.74 139.25 494.78 492.42 987.20 5th

6 100.00 79.36 89.45 89.56 139.79 498.16 498.32 996.48 5th

7 100.00 79.77 89.78 89.85 139.94 499.35 499.63 998.98 5th

8 100.00 79.92 89.92 89.95 139.98 499.77 499.92 999.69 5th

9 100.00 79.97 89.97 89.98 140.00 499.92 499.98 999.90 5th

Note: Bold numbers refer to the time unit where the organization reaches the fifth level of CSC maturity. (*) Level = CSC maturity level


166

Table 7.10 Simulation results of organization ‘A’ with ‘plds’ = 0.3, and ‘pppl’ and

‘ppro’ = 0.1

Year Score Level*


Initial 20.00 19.20 43.20 18.00 37.30 137.70 129.00 266.70 2nd

1 48.03 34.92 55.74 45.79 77.66 262.13 157.33 419.46 3rd

2 73.71 54.86 69.36 67.42 109.13 374.48 242.62 617.10 4th

3 94.12 69.04 79.58 80.34 127.60 450.68 353.70 804.37 5th

4 100.00 75.94 85.34 86.35 135.67 483.30 463.91 947.21 5th

5 100.00 78.54 87.95 88.69 138.58 493.76 491.74 985.50 5th

6 100.00 79.48 89.09 89.54 139.55 497.66 498.16 995.82 5th

7 100.00 79.82 89.60 89.84 139.86 499.11 499.59 998.71 5th

8 100.00 79.93 89.82 89.94 139.96 499.66 499.91 999.57 5th

9 100.00 79.98 89.92 89.98 139.99 499.87 499.98 999.85 5th


Table 7.11 Simulation results of organization ‘A’ with ‘plds’ and ‘ppro’ = 0.2, and

‘pppl’, ‘pprs’, and ‘ppol’ = 0.1

Year Score Level*


Initial 20.00 19.20 43.20 18.00 37.30 137.70 129.00 266.70 2nd

1 41.45 39.21 55.07 50.32 85.19 271.25 158.84 430.09 3rd

2 63.66 59.01 67.69 71.53 116.62 378.52 248.19 626.71 4th

3 83.10 71.40 77.80 82.74 131.88 446.91 358.68 805.59 5th

4 99.44 77.05 84.19 87.50 137.53 485.71 466.57 952.27 5th

5 100.00 79.07 87.43 89.21 139.29 495.01 492.48 987.49 5th

6 100.00 79.71 88.87 89.75 139.80 498.33 498.14 996.47 5th

7 100.00 79.91 89.50 89.92 139.95 499.63 499.28 998.91 5th

8 100.00 79.97 89.78 89.98 139.99 499.92 499.71 999.63 5th

9 100.00 79.99 89.90 89.99 140.00 499.98 499.88 999.87 5th



167

In summary, organization ‘A’ is currently in the early stage of its maturity level, with

Leadership found to be crucial if the organization aspires to progress through to higher

maturity levels. The focus of the organization, therefore, should be on a more effective

implementation process of the Leadership’s attributes, as described below:

� Leaders must take safety more seriously (Lingard and Blismas, 2006).

� Leaders must act as a role model in behaving safely (Dunlap, 2004).

� Leaders should continue to encourage workers to give opinions and/or suggestions

on safety matters (Little, 2002).

� Leaders are expected to educate workers, and ensure that they hold safety

responsibilities, for both themselves, and their workmates (Dias and Coble, 1996).

� Leaders should respond quickly to correct safety problems when they are brought to

their attention (Teo et al., 2005).

7.2.2.2 Policy Experiments of Organization ‘B’

The base run results of organization ‘B’ illustrated that it took three years for the

organization to progress from the second to the fifth levels of CSC maturity. For

organization ‘B’ to achieve maturity earlier (less than three years), a number of

sensitivity analyses, with, say, 10% extra effort being given to improve each enabler

(the ‘plds’, ‘pppl’, ‘pprs’, ‘ppol’, and ‘ppro’) need to be undertaken. The analyses help

to identify which enabler has the potential to increase the CSC index so that the

organization reaches the CSC maturity level earlier.

The sensitivity analysis results, illustrated in Table 7.12, demonstrate that, by giving the

10% more effort to enhance the score of Processes (‘ppro’ = 0.1), organization ‘B’

achieved its maturity one year earlier, i.e. within two years. The results of the other four

enablers, however, showed no advancement in the organization achieving the fifth CSC

maturity level earlier.


168

Table 7.12 The CSC index of organization ‘B’ when more of effort is given to enhance

each enabler

Year CSC Index

Base run Plds = 0.1 Ppol = 0.1 Pppl = 0.1 Pprs = 0.1 Ppro = 0.1

Initial 459.90 459.90 459.90 459.90 459.90 459.90

1 628.56 629.71 631.44 632.15 631.63 635.25

2 794.51 794.69 797.20 798.49 797.13 802.76

3 927.21 927.31 928.79 930.02 928.67 942.50

4 979.00 979.04 979.67 980.58 979.67 983.50

5 993.63 993.64 993.89 994.47 993.91 994.93

Note: Bold numbers refer to the time unit where the organization reaches the fifth level of CSC maturity

In conclusion, then, to improve safety performance and achieve the fifth CSC maturity

level, in a shorter time period, organization ‘B’ should focus on enhancing the

improvement of the Processes enabler, as it facilitates a faster CSC maturity

achievement. The organization may need to:

� Provide adequate safety training, especially for new staff, to ensure that the job is

performed safely (Tam et al., 2004);

� Have a routine risk and hazard assessment (Berg, 2006);

� Keep the site housekeeping at a high level (Zohar, 1980);

� Adopt a no-blame approach, and learn from experience (ICAO, 1992); and

� Have a good safety benchmarking system to compare the organization’s safety with

that of other construction organizations (Taylor, 2003).

7.3 THE CYCLICAL STYLE OF SAFETY MANAGEMENT

The policy experiments, described in Section 7.2, demonstrate the value in identifying

areas for safety improvement to progress through to higher CSC maturity levels, and to


169

achieve the maximum score of the CSC index (1,000 points). In real-life situations,

however, the CSC index may never reach its maximum score. One key reason is top

management’s view of the fifth maturity level as a target, not as means of continual

improvement. Once the fifth maturity level is reached, top management tends to slow

the momentum behind all safety activities. This phenomenon is known as ‘attention

withdrawal’, and is illustrated in the accident cycle (shown in Figures 7.9 and 7.10),

where top management gradually and slowly withdraws its attention to safety when

safety performance reflects the highest level of maturity (NPS Risk Management

Division, 2006).

M anagem ent tak es pro m pt safe ty ac tion s

M anagem ent slow ly w ithdraw s its attention to safety

(M anagem en t satisfaction w ith reach ing the fifth C S C m aturity level)

Safety Perform ance

Acc

iden

t dec

reas

es

Acc

iden

t inc

reas

es

U pper lim it(M anagem en t sa tisfaction w ith safety perform ance)

T im e

L ow er lim it

Figure 7.9 The accident cycle (Adapted from NPS Risk Management Division, 2006)


170

Stage 1

Occurrence of a detrimental event

Stage 2

Identification of shortfalls and recrimination

Stage 3

Spotlight firmly on safety

Stage 4

Strongly supported campaign for major safety improvements and implementation

Stage 5

A growing history of safety success comfort

Stage 6

Attention slowly being focused elsewhere

Stage 7

Safety concerns raised but judged as not sufficiently important (or welcomed) in the overall scheme of things

Figure 7.10 The normal accident cycle (Adapted from Jones, 2007)

When the accident rate is high (reaches the lower limit, see Figure 7.9), it becomes top

management’s priority to reduce the number of accidents (as shown in Stages 1 to 3 of

Figure 7.10). Top management identifies the shortfalls and priorities, and then acts

promptly to enhance safety improvements. Indeed, top management may consider, for

example, providing more safety resources to carry out the job safely, empowering safety

responsibilities to staff, and encouraging more two-way safety communication (as

shown in Stage 4 of Figure 7.10).

As top management pays more attention to reducing the number of accidents, the

accident rate decreases (representing Stage 5 of Figure 7.10). This continues until the

accident rate reaches the hypothetical upper limit of management satisfaction with

safety performance (see Figure 7.9), then top management starts to unintentionally

withdraw its attention to safety (as shown in Stage 6 of Figure 7.10). For example, the

financial resources provided to support the acquisition of safety resources may be cut


171

down (see the negative relationship between safety resources and management

commitment shown in the feedback loop ‘B’ of Figure 6.3). Further, top management

may also shift the focus to increasing production (as shown in Stage 7 of Figure 7.10),

in which, then, puts pressure on the staff. As pressure increases, the distress increases,

leading to the intensifying of the accident rate (see the positive relationship between the

distress and accident rate shown in the feedback loop ‘A’ of Figure 6.3).

Consequently, the accident rate increases as management withdraws its safety attention

until the hypothetical lower limit is reached (see Figure 7.9). Top management then

reacts promptly and swiftly by taking actions to reduce the accident rate again, and thus

the cycle continues.

SD modelling is used to better understand the changes in the Enablers, Goals, and CSC

index scores resulting from the cyclical style of safety management. The details of these

changes are discussed below.

7.3.1 The Dynamic Model of the Cyclical Style of Safety Management

The cyclical style of safety management was modelled with SD modelling. The

assumption made in the modelling process was that top management withdraws its

attention to safety when the CSC index reaches 95% of its maximum score

(representing the upper limit of management satisfaction with safety performance, see

Figure 7.11). The 95% level was selected as it represents a very high confidence in the

organization’s safety management ability, and any accidents that might occur could be

largely traced to random events represented by the 5% error level. At this point, top

management starts gradually shifting its safety attention to other areas for improvement,

believing that an adequate safety management system is in place, and the effective

implementation of this system will continue, regardless of the level of management

support/attention.


172

Management takes prompt safety actions

(i.e. safety performance reflects the fourth CSC maturity level)

Management withdraws its attention to safety

(i.e. high CSC index value)

(Management satisfaction with reaching the fifth CSC maturity level)

Upper limit(Management satisfaction with much improved safety performance)

CSC Index

Time

950

800 Lower limit

The

fift

h C

SC m

atur

ity le

vel

Posi

tive

slop

e Negative slope

“Comfort Zone”

Figure 7.11 The CSC index cycle as management withdraws attention to safety

Eighty percent of the maximum score of the CSC index, on the other hand, was chosen

as the lower limit (see Figure 7.11). At this point, the organization is falling into the

lower maturity level, i.e. from the fifth to the fourth maturity levels. Top management

realises the problem, and starts taking actions to improve the CSC index.

The dynamic model of the cyclical style of safety management is shown in Figure 7.12.

The upper and lower limits (the CSC index of 950 and 800 points, respectively) were

used in the SD equations, as described in the dynamic models of the five enablers and

Goals below.


173

LEADERSHIP

rlds

used lds

dlds

GOALS

rgoals

CSC INDEX

ggoals

dgoals

slope

used pol

used ppl used prs

used pro

rldsf

PEOPLE

rppl

used goals

Co lds ppl

DF ppl ldsgppl

used ppl

ggoals

dppl

used ldsDF goals pro Co pro goals

POLICY & STRATEGY

gpol

gpol

DF pol lds

rpol

DF pol prs

Co prs pol

Co lds pol

used pol

dpol

PARTNERSHIPS & RESOURCES

rprs

used pol

used prsdprs

gprs

DF prs lds

Co lds prs

rlds2

ENABLERS

DF prs ppl

Co ppl prs

gprs

PROCESSES

rpro

DF pro ppl

Co ppl pro

used pro

dpro

gpro

DF pro pol

Co pol pro

rlds2

rlds2 rlds2

gpro

rlds2

rppl2

rlds2

rprs2 rpol2

Co lds pol

rpro2

glds

Co lds ppl

Co ppl pro

Co pol pro

slope

slope

desired CSC INDEX

rgoals2 CSC flow

Co lds ppl

CSC INDEX

CSC INDEX

rlds3

rlds3

rlds3rlds3

rlds3

rlds3dCSC INDEX

dCSC INDEX

Co prs pol

Note: Co_lds_pol, Co_lds_ppl, Co_lds_prs, Co_pol_pro, Co_ppl_pro, Co_ppl_prs, Co_pro_goals, and Co_prs_pol = Correlations between Lds and Pol, Lds and Ppl, Lds and Prs, Pol and Pro, Ppl and Pro, Ppl and Prs, Pro and Goals, and Prs and Pol, respectively. DF_goals_pro, DF_pol_lds, DF_pol_prs, DF_ppl_lds, DF_pro_pol, DF_pro_ppl, DF_prs_lds, and DF_prs_ppl = Decision fractions between Goals and Pro, Pol and Lds, Pol and Prs, Ppl and Lds, Pro and Pol, Pro and Ppl, Prs and Lds, and Prs and Ppl, respectively. dgoals, dlds, dpol, dppl, dpro, dprs,= Desired Goals, desired Lds, desired Pol, desired Ppl, desired Pro, and desired Prs, respectively. ggoals, glds, gpol, gppl, gpro, and gprs = Gaps of Goals, Lds, Pol, Ppl, Pro, and Prs, respectively. plds, ppol, pppl, ppro, and pprs = Percentage effort provided to Lds, Pol, Ppl, Pro, and Prs, respectively. rgoals, rlds, rpol rppl, rpro, and rprs = Goals, Lds, Pol, Ppl, Pro, and Prs rates, respectively. used_goals, used_lds, used_pol, used_ppl, used_pro, and used_prs = Goals, Lds, Pol, Ppl, Pro, and Prs values, respectively. rldsf = Lds rate fraction

Figure 7.12 The dynamic model of the cyclical style of safety management


174

7.3.1.1 Leadership Dynamic Model

The Leadership dynamic model is shown in Figure 7.13. The upper and lower limits of

the CSC index (800 and 950 points, respectively) were adopted in the Lds equations

(see full details of SD equations of the cyclical style of safety management model in

Appendix 12). Equations 7.1 to 7.3 demonstrate that there will be no inflows of the Lds

score (‘rlds’ = 0), if the CSC index exceeds the upper limit of management satisfaction

with safety performance (950 points, see Figure 7.11); or the slope of the CSC index is

negative (slope < 0, see Figure 7.11), which represents the onset of withdrawing

attention to safety.

As the attention to safety gets withdrawn, the Lds score decreases (there are outflows of

the Lds score, ‘rlds2’ or ‘rlds3’ >0), leading to a reduced CSC index. The ‘attention

withdrawal’ continues until the CSC index reaches the lower limit (800 points, see

Figure 7.11), then top management reacts through taking actions to increase the CSC

index again (the Lds score increases, ‘rlds’ > 0, and ‘rlds2’ or ‘rlds3’ = 0).

LEADERSHIP

rlds

used lds

dlds

rldsf

ggoals

rlds2

gldsslope

slope

CSC INDEX

CSC INDEX

rlds3

dCSC INDEX

Leadership Dynamic Model


Figure 7.13 Leadership dynamic model


175

Inflows:

Eq 7.1 rlds = IF (CSC_INDEX > dCSC_INDEX) OR ((800 <

CSC_INDEX < dCSC_INDEX) AND (slope < 0)) THEN

(0) ELSE ((used_lds + ggoals)*rldsf

Outflows:

Eq. 7.2 rlds2 = IF (CSC_INDEX < 800) OR ((800 < CSC_INDEX <

dCSC_INDEX) AND (slope > 0)) THEN (0) ELSE ((glds

+ ggoals)*rldsf

Eq. 7.3 rlds3 = IF (CSC_INDEX > dCSC_INDEX) AND (rlds = 0) AND

(rlds2 = 0) THEN ((glds + ggoals)*rldsf) ELSE (0)

The decreased Lds score negatively affects the score of the People enabler, as these two

enablers have a clear positive relationship (see Figure 5.4). The details of this effect are

described in the following section.

7.3.1.2 People Dynamic Model

The People dynamic model is illustrated in Figure 7.14. As top management withdraws

attention to safety (‘rlds2’ or ‘rlds3’ > 0), there tends to be a decrease in people’s

perception of, and participation in, safety (see the feedback loop ‘B’ of Figure 6.3),

which, in turn, reduces the Ppl score (there is an outflow of the Ppl score, ‘rppl2’ > 0)

(see Equations 7.4 and 7.5). Normally, the amount of score reduction depends on the

correlation strength between the Lds and Ppl enablers, as demonstrated in Equation 7.5.


176

used lds

PEOPLE

rppl

Co lds ppl

DF ppl ldsgppl

used ppl

dppl

rlds2

rppl2

rlds3

People Dynamic Model


Figure 7.14 People dynamic model

Inflows:

Eq. 7.4 rppl = (used_lds*DF_ppl_lds)

Outflows:

Eq. 7.5 rppl2 = IF (rlds2 > 0) OR (rlds3 > 0) THEN (rppl + (rppl*

Co_lds_ppl)) ELSE (0)

As both scores of the Lds and Ppl enablers continue to decrease, they tend to reduce the

Partnerships and Resources score (see Figure 6.5 for the positive relationships between

these three enablers). The details for this model are described below.

7.3.1.3 Partnerships and Resources Dynamic Model

The reduction in the Lds score, affected by the cyclical style of safety management,

decreases the Partnerships and Resources score directly and indirectly, through the Ppl

enabler (see Figure 7.15, and Equations 7.6 and 7.7). For instance, the financial

resources provided to support the acquisition of safety resources may be cut down as


177

management shifts its focus to increasing production (management withdraws attention

from safety to other areas needing improvement).

used lds

PARTNERSHIPS & RESOURCES

rprs

used ppl

used prs

dprs

gprs

DF prs lds

Co lds prs

DF prs ppl

Co ppl prs

Co lds ppl

gprs

rlds2

rprs2

rlds3

Partnerships and Resources Dynamic Model


Figure 7.15 Partnerships and Resources dynamic model

Inflows:

Eq. 7.6 rprs = (used_lds*DF_prs_lds) + (used_ppl*DF_prs_ppl)

Outflows:

Eq. 7.7 rprs2 = IF (rlds2 > 0) OR (rlds3 > 0) THEN (rprs + (rprs*

Co_lds_prs) + (rprs* Co_lds_ppl*Co_ppl_prs)) ELSE (0)

The amount of the Prs score decrease depends on the correlation strength between this

particular enabler, and the Lds and Ppl enablers, as illustrated in Equation 7.7. As the

Lds and Prs enablers have a direct effect on the Policy and Strategy enabler, the

decreased Lds and Prs scores, undoubtedly, reduce the Pol score. The details of the

Policy and Strategy dynamic model are described below.


178

7.3.1.4 Policy and Strategy Dynamic Model

The dynamic model of Policy and Strategy (see Figure 7.16) demonstrates that the

increased Lds score increases the Pol score (the Lds enabler has a positive relationship

with the Pol enabler, see Figure 6.5) (see Equation 7.8). As top management’s attention

to safety being withdrawn (in other words, the Lds score is reduced), the Pol score

decreases. The amount of the Pol score decrease depends on the correlation strength

between this enabler, and the Lds and Prs enablers (as Lds has direct and indirect,

through Prs, effects on Pol) (see Equation 7.9).

used ldsPOLICY & STRATEGY

gpol

gpol

DF pol lds

used prs

rpol

DF pol prs

Co prs pol

Co lds pol

used pol

dpol

rlds2

rpol2

Co lds prs

rlds3

Policy and Strategy Dynamic Model


Figure 7.16 Policy and Strategy dynamic model

Inflows:

Eq. 7.8 rpol = (used_lds*DF_pol_lds) + (used_prs*DF_pol_prs)

Outflows:

Eq. 7.9 rpol2 = IF (rlds2 > 0) OR (rlds3 > 0) THEN (rpol + (rpol*

Co_lds_pol) + (rpol* Co_lds_prs*Co_prs_pol)) ELSE (0)


179

The ‘attention withdrawal’ (the decrease in the Lds score) not only has a direct effect on

the Ppl, Prs, and Pol scores, but also has an indirect influence in the reduction of the

Process score through the Ppl and Pol enablers (see Figure 5.4). This decrease is

described in the Processes dynamic model below.

7.3.1.5 Processes Dynamic Model

The Processes dynamic model (see Figure 7.17) shows that the decreased Lds, Ppl, and

Pol scores, as management withdraws its attention to safety, lower the Pro score (see

Equations 7.10 and 7.11). The score decrease depends on the correlation strength

between the Pro and the Lds, Ppl, and Pol enablers.

used ppl

used polPROCESSES

rpro

DF pro ppl

Co ppl pro

used prodprogpro

DF pro pol

Co pol progpro rlds2

rpro2

Co lds ppl

Co lds pol

rlds3

Processes Dynamic Model


Figure 7.17 Processes dynamic model

Inflows:

Eq. 7.10 rpro = (used_ppl*DF_pro_ppl) + (used_pol*DF_pro_pol)


180

Outflows:

Eq. 7.11 rpro2 = IF (rlds2 > 0) OR (rlds3 > 0) THEN (rpro + (rpro*

Co_lds_ppl*Co_ppl_pro) + (rpro*Co_lds_pol*

Co_pol_pro))ELSE (0)

The decrease of the Pro score, undoubtedly, negatively affects the Goals score (as these

two constructs have a strong positive relationship, see Figure 5.4). The details of the

Goals dynamic model are presented below.

7.3.1.6 Goals Dynamic Model

In the Goals dynamic model (see Figure 7.18), the increased Pro score enhances the

Goals score (as they have a strong positive relationship, see Figure 5.4) (see Equation

7.12); the reduced Pro score leads to the decreased Goals score (as demonstrated in

Equation 7.13).

GOALS

rgoals

ggoals

dgoals

used pro

used goalsDF goals pro Co pro goals

rlds2Co lds ppl

Co ppl pro

rgoals2

Co lds pol

Co pol pro

rlds3

CSC INDEX

Goals Dynamic Model


Figure 7.18 Goals dynamic model


181

Inflows:

Eq. 7.12 rgoals = used_pro*DF_goals_pro

Outflows:

Eq. 7.13 rgoals2 = IF (rlds2 > 0) OR (rlds3 > 0) THEN (rgoals + (rgoals*

Co_lds_pol*Co_pol_pro*Co_pro_goals) + (rgoals*

Co_lds_ppl*Co_ppl_pro* Co_pro_goals)) ELSE (0)

The reduction of the Goals score (Equation 7.13), as affected by the ‘attention

withdrawal’, is influenced by Pro, which is also being effected by Lds through Ppl and

Pol. The smaller Goals score, certainly, leads to a larger ‘gap of goals’ (ggoals), which,

in turn, will affect the Lds enabler (as ‘rlds’ depends on ‘ggoals’, see Equation 7.1), and

the simulation repeats as cycles.

The ‘attention withdrawal’ causes a reduction in the CSC index. This continues until

the index score reaches the lower limit (800 points, see Figure 7.11), then management

takes prompt actions to improve safety implementation. Management achieves this, for

example, by getting workers involved in safety activities, encouraging feedback on

safety, providing adequate safety resources, and reassigning safety responsibilities to

staff. As top management returns its focus on safety improvement, the CSC index

begins to increase. This continues until the index score exceeds the upper limit

corresponding to management’s satisfaction with the safety record (assumed herein as

950 points, see Figure 7.11), then management starts to withdraw its attention away

from safety again, and the cycle continues.

The simulation results, as an effect of the ‘attention withdrawal’ phenomenon, are

described in Section 7.3.2 below.


182

7.3.2 The Simulation Results

The dynamic model of the cyclical style of safety management is simulated to

investigate the changes in the Enablers, Goals, and CSC index scores as reflected by

management attention withdrawal. The lower limit used in the simulation was, as stated

earlier, set at the CSC index = 800 points. According to the base run results (previously

shown in Tables 6.1 and 6.2), this index score was almost achieved (the CSC index =

797.90 points) at the end of year 10. Therefore, the scores of the five enablers and Goals

at this point of time were used as the initial values for the simulation. The specifics are

detailed below:

� ‘used_lds’ = 60.37 (out of 100 points)

� ‘used_ppl’ = 68.07 (out of 90 points)

� ‘used_prs’ = 82.23 (out of 90 points)

� ‘used_pol’ = 73.17 (out of 80 points)

� ‘used_pro’ = 132.44 (out of 140 points)

� ‘used_goals’ = 381.64 (out of 500 points)

The initial values of the five enablers and Goals were adopted in the SD equations, and

the model was simulated. The simulation results illustrate that at the beginning of the

simulation, the CSC index, which reflects the fourth CSC maturity level, increased as

the Enablers and Goals’ scores increased (see Table 7.13, and Figures 7.19 to 7.21),

demonstrating safety improvement. At the end of year three, however, the CSC index

reached its specified upper limit (950 points, see Figure 7.11), and management started

to unintentionally withdraw its attention from safety to other areas requiring

improvement. Thus, there was a slight drop in the five enablers’ score, which led to a

decrease in the Goals and the CSC index scores.


183

Table 7.13 Simulation results of the cyclical style of safety management

Year Score Lds Pol Ppl Prs Pro Enablers Goals CSC Index Initial 60.37 73.17 68.07 82.23 132.44 416.28 381.64 797.92 1 72.05 76.76 75.79 86.46 136.87 447.93 426.25 874.18

2 83.05 78.62 81.56 88.54 138.78 470.54 460.13 930.67

3 89.63 79.21 84.07 89.17 139.33 481.42 472.84 954.25*

4 86.19 78.69 81.61 88.55 138.85 473.89 462.28 936.16

5 81.47 77.84 78.31 87.48 138.05 463.15 447.68 910.83

6 74.98 76.52 74.03 85.75 136.74 448.02 427.59 875.61

7 66.05 74.57 68.78 83.06 134.64 427.10 400.15 827.25

8 60.41 73.31 65.88 81.26 133.18 414.04 382.97 797.01**

9 63.96 75.31 70.31 83.95 135.54 429.08 401.28 830.36

10 75.98 77.86 77.65 87.33 138.20 457.02 446.37 903.39

11 85.94 79.11 82.85 88.93 139.31 476.14 471.15 947.29

12 89.27 79.28 83.68 89.12 139.45 480.79 475.09 955.88*

13 86.01 78.80 81.07 88.44 139.05 473.38 465.39 938.77

14 81.57 78.03 77.55 87.32 138.40 462.87 451.97 914.84

15 75.49 76.82 72.98 85.49 137.32 448.11 433.47 881.58

16 67.16 75.03 67.33 82.68 135.60 427.80 408.10 835.90

17 58.95 74.29 65.12 81.37 134.88 414.61 385.58 800.18**

18 71.62 77.28 73.85 85.96 137.86 446.58 437.68 884.26

19 81.81 78.83 80.36 88.30 139.16 468.46 466.43 934.89

20 90.90 79.53 84.64 89.34 139.68 484.10 482.00 966.10*

21 88.44 79.22 82.37 88.83 139.46 478.31 474.97 953.29

22 85.10 78.70 79.25 87.96 139.08 470.10 465.23 935.33

23 80.57 77.88 75.08 86.53 138.45 458.50 451.74 910.24

24 74.38 76.60 69.67 84.27 137.42 442.35 433.14 875.49

25 65.93 74.72 63.04 80.88 135.81 420.38 407.63 828.01

26 60.60 73.51 59.36 78.69 134.72 406.88 391.58 798.46**

27 64.00 75.43 65.01 81.93 136.49 422.86 407.82 830.67

28 75.58 77.90 74.30 86.26 138.54 452.58 450.09 902.67

29 85.25 79.12 80.88 88.45 139.43 473.13 473.18 946.31

30 94.18 79.66 85.05 89.40 139.79 488.08 485.63 973.71*

31 92.33 79.42 82.87 88.94 139.64 483.19 480.02 963.21

32 89.80 79.02 79.79 88.14 139.38 476.13 472.22 948.36

. . . . . . . . .

. . . . . . . . . 112 100.00 80.00 90.00 90.00 140.00 500.00 500.00 1,000.00 Note: * The CSC index reaches its upper limit. ** The CSC index reaches its lower limit.


184

Page 11.00 7.00 13.00 19.00 25.00

Years

1:

1:

1:

350

425

500

1: Enablers Score

11

11

Figure 7.19 Graphical results of the Enablers score as the effect of

the attention withdrawal

Page 41.00 7.00 13.00 19.00 25.00

Years

1:

1:

1:

350

425

500

1: used goals score

1

1

1

1

Figure 7.20 Graphical results of the Goals score as the effect of

the attention withdrawal


185

Page 21.00 7.00 13.00 19.00 25.00

Years

1:

1:

1:

700

850

1000

1: CSC Index Score

1

1

1

1

Figure 7.21 Graphical results of the CSC index as the effect of the attention withdrawal

The CSC index continued to decrease gradually until it reached the lower limit of 800

points at the end of year eight. Subsequently, top management took prompt actions to

improve safety implementation, in response to the fear that the organization would fall

to a lower maturity level (from the fifth to the fourth maturity levels). As a result, there

was a relatively large increment in the Lds score from 60.41 points, at the end of year

eight, to 79.58 points, at the end of year 10 (an increase of more than 19% of its

maximum score, see Table 7.13).

Leadership action to improve safety implementation, obviously, enhanced the

implementation of the Ppl, Pol, Prs, and Pro enablers (as seen by the increase of these

four enablers’ scores, see Table 7.13), leading to a higher Goals score and, ultimately,

the CSC index. The actions taken to improve the CSC index continues until the index

exceeds the assumed upper limit (the CSC index of 950 points), then the ‘attention

withdrawal’ takes place again (top management shifts attention from safety to other

areas for improvement), and the cycle continues.

Simulation results (see Figure 7.21 and Table 7.13) show that the CSC index score

oscillates between the fourth and the fifth CSC maturity levels. However, it slowly aims


186

towards the maximum score of 1,000 point (the maximum score of the CSC index),

over a very long term.

7.3.3 Conclusion of the Cyclical Style of Safety Management

SD modelling was used to model the cyclical style of safety management to imitate the

situation where management withdraws its attention to safety. The ‘attention

withdrawal’ occurs when the CSC index exceeds the assumed upper limit of 950 points.

Top management, then, started to gradually withdraw its attention to safety (for

example, its focus from safety implementation to product enhancement), believing that

safety management system is in place, and the effective safety implementation will

continue, regardless of management support. These leadership actions negatively

affected the workers’ perception of, and participation in, safety, which, in turn, led to

less job satisfaction and lower workforce morale (as indicated by the lower scores of

Ppl, Prs, Pol, and Pro, as well as Goals).

The decreased Enablers and Goals scores undoubtedly reduce the CSC index. This

continues until the index score reaches the assumed lower limit of 800 points, indicating

that the organization has fallen into the lower (the fourth) maturity level. At that stage,

top management takes prompt safety actions to increase the CSC index by: 1)

encouraging staff to put forward their opinions about how to improve safety, 2)

providing safety training, especially to new staff; and 3) promoting safety campaigns to

enhance safety awareness.

Such positive safety implementation improves the Enablers score, which, in turn,

enhances the Goals and CSC index scores. This approach continues until the CSC index

surpasses the upper limit of management satisfaction with its safety performance, then

the ‘attention withdrawal’ starts again, and the cycle continues.

The results from this study show that the CSC index score oscillates between the fourth

and the fifth CSC maturity levels, as an effect of the cyclical nature of safety


187

management. However, the organization, once again, turns towards the maximum index

score of 1,000 point, over a very long period.


188



189

88 SSTTUUDDYY FFIINNDDIINNGGSS AANNDD RREECCOOMMMMEENNDDAATTIIOONNSS

FFOORR FFUUTTUURREE RREESSEEAARRCCHH


This chapter brings together the major findings of this study. A review and synthesis of

the existing body of knowledge is presented below. The implications of this research

and the findings for the Thai construction industry, and the recommendations for future

research are discussed at the end of the chapter.

8.2 MAJOR FINDINGS

The main objective of this study was to investigate the interactions and causal

relationships among the key factors (enablers and Goals) of the construction safety

culture (CSC). An understanding of their individual, or combined, effects on an

organization’s ability to achieve safety performance improvements was also sought.

With this in mind, the following secondary objectives were identified:

� Review the literature regarding the nature of the CSC and its key components, and

identify the tools (well-established performance measurement systems) available for

measuring safety culture.

� Develop a CSC model based on a widely used performance measurement system,

the EFQM Excellence model, and investigate the interactions and causal

relationships among its key factors (CSC enablers and Goals).

� Obtain Thai industry input, via a questionnaire survey, the data acquired to be used

for the statistical analyses and SD modelling.


190

� Perform the EFA and SEM analyses to confirm the construct validity of the

proposed CSC model.

� Develop a CSC dynamic model utilizing SD software ‘STELLA’ to examine the

interactions and causal relationships among the five enablers and Goals, over a

period of time.

� Verify and validate the developed dynamic model.

� Assess CSC maturity levels using the scores from the developed CSC index.

� Identify areas for safety improvement to achieve a higher CSC index, and progress

through to higher maturity levels.

These objectives were successfully achieved, as presented in Chapters 1 to 7. In Chapter

1, a literature review of the characteristics of the construction industry, safety culture

definitions, and the measuring of safety culture was conducted. Two major

shortcomings seemed apparent, viz�

� No safety culture model included either the causal relationships among the key

factors of the CSC, or any feedback mechanism between these factors.

� No tool existed for appropriately assessing the CSC maturity levels or the evolution

to higher maturity levels, over time.

Consequently, these shortcomings became the research objectives, while a number of

research aims were identified to fulfil those research gaps.

Chapter 2 presented the research methodology adopted for this study, including the

research design and the research activities, along with the expected outputs. The

research activities and expected outcomes led to the specific steps required to fulfil the

research aims.


191

As discussed in Chapter 2, a questionnaire survey was adopted as the most appropriate

method for data collection. The techniques for data screening and the preliminary

analyses were explained. In addition, the EFA, SEM, and SD modelling, as well as their

uses in this study were also introduced.

Three well-known performance measurement systems (the MBNQA framework, the

BSC framework, and the EFQM Excellence model), were critically examined and

compared in Chapter 3. Based on this comparison, the EFQM Excellence model was

identified as the basis for the CSC model development. The proposed CSC model

consisted of six constructs, i.e. five enablers, namely Leadership, Policy and Strategy,

People, Partnerships and Resources, and Processes, and the single set of Goals. Their

associated attributes were identified from the literature review, and were used in

developing the questionnaire survey. The survey was drawn up to define and then

operationalise the six constructs of the CSC model.

In Chapter 4, the questionnaire, which consisted of 34 statements covering 34 attributes

of the CSC, was drawn up. It was rated on a five-point Likert scale. The questionnaire

was sent to over 100 Thai construction-contracting organizations, with 53.6% being

returned. Three surveys were unusable due to data incompleteness, and so they were

dropped from the data file. One hundred and fifteen surveys were used in the analyses.

The preliminary analyses and data screening were performed to increase confidence in

the data. The results showed that less than 5% of the missing values were found in each

attribute (item), and that all the attributes displayed a normal distribution (their

skewness and kurtosis values were in acceptable ranges). Confidence in the data was

therefore increased. Only a single outlier was found (in data number ‘76’). As a result,

this case was removed from the data file, leading to a retained total of 114 data sets for

the analyses. The retained data were used to test the internal consistency of the 34

attributes within the six constructs of the CSC. The results demonstrated high reliability

values (Cronbach’s alpha higher than 0.7), and hence increased confidence in the

attributes to the measurement of their respective constructs.


192

The screened data were analysed further, using principal axis factoring with varimax

rotation (see Chapter 5). A total of 27 attributes of the CSC enablers were identified,

categorised into a number of factors that represented the interrelations among the set of

attributes. Three attributes were removed from the data file, leaving 24 attributes

grouped into five enabler-constructs (labelled Leadership, Policy and Strategy, People,

Partnerships and Resources, and Processes). Among those groups, nine attributes were

relocated from one proposed enabler to another.

After this relocation, the SEM was undertaken to provide further evidence for the

construct validity of the CSC model. The 24 attributes, grouped to explain the five

enablers, and the seven attributes grouped to explain Goals, were examined, using the

confirmatory factor analysis (CFA) technique, to further specify the posited relations of

the attributes to their underlying constructs. A number of GOF indices, such as �2/DF,

RMSEA, CFI, and NFI, were used to assess the model fit. The proposed CSC model

was modified by eliminating the links with low correlations, and removing the attributes

with high multicollinearity. In doing so, seven attributes were removed, leading to a

best-fit measurement model that comprised 20 attributes grouped to represent the six

constructs of the CSC model.

The fitted measurement model was subsequently tested to examine the direction of the

assumed relationships between the six constructs, which were reflected by the arrows

connecting them. It was first assumed that bi-directional relationships existed between

the three enablers (Policy and Strategy, People, and Partnerships and Resources), due

to their potential to affect each other. The model was then analysed with different

directional influences among these three enablers. As a result, one attribute with high

multicollinearity was removed from the Leadership construct. The fitted structural

model confirmed the direction of relationships among the five enablers and Goals of the

CSC.

Leadership appears to directly influence the implementation of the People, Policy and

Strategy, and Partnerships and Resources enablers. However, Leadership has indirect

effect on Partnerships and Resources; through the implementation of the People


193

enabler, the relationship was found to be stronger than the direct one. This result fits

well with the assumption that Thai managers consider teamwork more important, in

improving safety implementation, than the provision of safety resources (Aksorn and

Hadikusumo, 2006).

Partnerships and Resources, on the other hand, was found to indirectly affect Processes

through Policy and Strategy, which, likewise, appears to be indirectly influenced by

People. Both People and Policy and Strategy have significant relationships with

Processes, which appears to have a strong effect on the achievement of Goals. The

fitted structural model was labelled the final CSC model. The directions of the

relationships among its five enablers and Goals, as well as their correlation coefficients,

were then used in developing the CSC dynamic model.

In Chapter 6, the CSC dynamic model was formulated to capture the interactions and

causal relationships among the six constructs (five enablers and Goals) of the CSC

model, over a period of time. Model verification and validation were undertaken to

increase confidence in the developed model. The CSC index, developed through the

dynamic model, represented the sum of the five enablers and Goals’ values at a point in

time, and was used together with the five levels of CSC maturity to indicate the current

CSC maturity level.

Base run results revealed that an organization with ad-hoc safety implementation should

primarily focus on enhancing the Leadership and People enablers to successfully

progress through to higher CSC maturity levels in the future. This finding perfectly

matches with the respondents’ perspective that People and Leadership are among the

most influential enablers in significantly improving the CSC.

As presented in Chapter 7, policy analyses were performed with two organizations

(randomly chosen from the data file), currently in the second (managing) and third

(involving) levels of CSC maturity, respectively. A number of safety policies were


194

tested to identify the most effective policy each organization could apply to enhance its

CSC, and progress through to the fifth (continually improving) CSC maturity level.

A cyclical style of safety management was also modelled to imitate real-life situations

where top management gradually withdrew its attention from safety when the CSC

index exceeds the upper limit of management satisfaction with safety performance. This

‘attention withdrawal’ negatively affects the Enablers and Goals scores, and ultimately

the CSC index. The CSC index decreases as top management withdraws its attention

from safety. This decrease continues until the index score reaches the lower limit,

meaning that the organization falls into the lower CSC maturity level (the fourth CSC

maturity level). as a consequence of this fall, top management takes prompt actions to

improve safety implementation, and thus to increase the Enablers and Goals scores, as

well as the CSC index. Once again, the CSC index rises until it exceeds the upper limit,

then management starts to withdraw its attention from safety again, and the cycle

continues.

While being affected by the cyclical nature of safety management, the organization,

however, slowly progresses towards the maximum CSC index score of 1,000 point, over

a very long period of time.

8.3 CONTRIBUTIONS TO THE EXISTING BODY OF KNOWLEDGE

Despite a large number of research studies focusing on measuring safety culture,

virtually no research has been undertaken: 1) to investigate the interactions and causal

relationships among the key factors of the CSC; and 2) to assess the CSC maturity level

and determine areas for improvement to progress through to higher maturity levels, over

a period of time.

The results from this study have contributed to the existing body of knowledge in the

following ways:


195

� Previously, no study had comprehensively modelled the CSC, and most studies had

only tested the interactions among selected enablers, in isolation. Those studies were

unable to identify the causal mechanisms that researchers/managers need to

understand and so enhance safety culture. The CSC model, developed in this study,

explores the causal relationships among the five enablers and Goals, thus extending

knowledge and understanding of the key factors and their respective, as well as

collective, influences on CSC implementation and output.

� The research facilitated the examination of the relationships among the enablers in a

user-friendly graphical format. To the best of the author’s knowledge, this had not

previously been undertaken in the area of the CSC.

� The developed CSC model provides an integrated framework for understanding how

decisions and behaviours of leadership are linked to safety processes, and how these

translate into desired safety goals (better safety performance).

Such contributions provide a strong foundation for understanding the CSC and its key

factors, as well as the relationships among those key factors, thus adding value to future

research.

8.4 IMPLICATIONS FOR THE THAI CONSTRUCTION INDUSTRY

The development of a CSC dynamic model provides a number of benefits to the Thai

construction industry, as discussed below.

� Although extracting from international literature, most of the attributes (more than

82%), associated with each enabler and Goals, in the final CSC model give a good

representation of safety practices in the Thai construction industry. Such attributes

include training, safety standards, human resources, financial resources,

stakeholders’ cooperation, management commitment, workers involvement, workers

relationships, safety standards, safety resources, number of accidents, cost of

accidents, safety awareness, safety initiatives, safety integration in business goals,


196

accountability, safety responsibilities, communication, and feedback (Boonrod et al.,

1998; Pipitsupaphol and Watanabe, 2000; Embassy of Denmark, Bangkok, 2006;

Aksorn and Hadikusumo, 2007; Wangniwetkul, 2007). The utilization of the study’s

results in realistically improving the CSC in Thailand can be said to be reinforced.

� By the literature review, Leadership was the most influential factor for a successful

safety program implementation in the Thai construction industry (Aksorn and

Hadikusumo, 2007). In an organization with no prior safety implementation,

leaders must be a role model in behaving safely. They must also be opened-mind,

and encourage two-way communication, and thus share the ideas of how to improve

safety.

� People and Policy and Strategy also play a key role in successful safety

implementation in the Thai construction industry. Workers involvement in safety

related activities, clear assigned safety responsibilities, and realistic safety rules help

enhance safety performance in an organization (Boonrod et al., 1998; Embassy of

Denmark, Bangkok, 2006).

� Further, the use of personal protective equipment (PPE) helps reduce site accidents

(Wangniwetkul, 2007). While adequately provided, the workers, sometimes, do not

use their PPE because it is inconvenient to do so (Pipitsupaphol and Watanabe,

2000). Consequently, the leaders are the key factor to solving this problem. They

must be role models, wearing the PPE everytime they enter the site. This action will

raise the workers’ perception and awareness of safety, and they will, then, follow the

leaders’ example (this is consistent with the direct relationships among the

Leadership, People, and Partnerships and Resources enablers found in this study).

� SD modelling was a useful tool in understanding the interactions and causal

relationships among the key factors, and has been similarly used in a number of

research projects focused on the Thai construction industry (Ogunlana et al., 1996;

Chritamara and Ogunlana, 2002; Ogunlana et al., 2002). The developed CSC

dynamic model provides insights into, and guidance and understanding of, the

interactions and casual relationships among the five enablers and Goals of the CSC,

over a period of time. In the Thai construction industry, these interactions and causal

relationships are important, as a change in one enabler may largely affect the change

in another enabler(s). This conclusion is supported by Aksorn and Hadikusumo


197

(2004), who stated that a lack of management support and management pressure

was found to be associated with workers’ unsafe acts, which led to high accident

rates. An organization could improve its safety performance if top management

committed more to improving safety, for instance, promoting realistic and workable

safety rules, providing adequate safety resources, and encouraging safety feedback

(Aksorn and Hadikusumo, 2007).

� The developed CSC index will assist organizations in assessing their current CSC

maturity levels, and identifying areas for safety improvement to enable progress

through to higher maturity levels. Organizations with different maturity levels will

need different safety policy and safety implementation processes, which cannot be

imitated. The use of SD modelling, with the developed CSC index, will help

organizations to plan the most effective safety implementation process to achieve

their safety goals within a planned time frame.

� Making investment decisions to improve one or more enabler is one of the most

essential roles played by management. However, these decisions have different

levels of impact on safety goals. Thus, managers who evaluate different policies

must do so based on informed decisions. The developed CSC dynamic model has

the capability to facilitate to test alternative strategies, through a number of model

simulations, to improve safety culture, by do not actually have to implement them.

Nevertheless, this approach helps to save costs that may occur from not

implementing the best safety strategy.

8.5 LIMITATIONS AND RECOMMENDATIONS FOR FUTURE

RESEARCH

The limitations of this study are presented below:

� The data used were based on input provided only by medium-to-large construction

contracting companies operating in Thailand.


198

� The targeted respondents were in senior positions, such as directors and project

managers, to facilitate the capture of a macro-level perspective for the CSC.

Workers were not included in the sample.

� The numbers of attributes used to operationalise each of the six constructs of the

CSC (Leadership, Policy and Strategy, People, Partnerships and Resources, and

Processes) were extracted from the international literature review, and were not

specifically limited to the Thai practices.

� The subcontractors’ role in improving the CSC was not of explicit interest (it is only

included in the ‘stakeholders’ cooperation’ attribute in the Partnerships and

Resources enabler).

� While the ‘workload’ and ‘work pressure’ attributes were key factors influencing

the construction site safety in Thailand (Boonrod et al., 1998; Pipitsupaphol and

Watanabe, 2000), they were not included in the final CSC model as a result of the

EFA and SEM.

� The final CSC model was developed based on the questionnaire surveys targeting

Thai construction organizations, thus, it might not be a best normative model to

prescribe the way of developing CSC in other countries.

� The study offers maximum flexibility to the users in utilizing their best judgement to

determine the relevant contributions each item makes to the operationalisation of a

certain enabler. In other words, the study allows the users to perform the analysis

using different levels of extra efforts per enabler (as a whole) without examining the

contributions made by each item. However, this could be a problem if the user does

not have enough experience or knowledge in planning for safety improvements.

The recommendations for areas of future research are listed below:

� The study was conducted using data from Thai construction organizations, in which

Thailand is considered a developing country. Thus, a comparative study may be

performed between developed and developing countries to investigate the

differences in CSC’s perspectives.


199

� A comparison study between top management and workers’ perceptions of safety

could capture the macro-level, as well as micro-level perspectives, of CSC.

� Different score-ranges of the five levels of CSC maturity might be altered

(whenever appropriate) to investigate the results over a period of time.

� Case studies could be conducted, over a period of time, to examine the deviations

between the simulation and the real life results to help further refine the model.

8.6 CLOSURE

This study made fundamental contributions to the area of construction safety culture

(CSC). The developed CSC dynamic model provided an insight into the interactions and

influences each enabler has in improving the CSC. The developed CSC index will guide

organizations benchmarking their CSC maturity level, and planning safety

improvements.


200

�

AAppppeennddiixx 11

QQuueessttiioonnnnaaiirree SSuurrvveeyy


201

Centre for Infrastructure Engineering and Management Griffith University

Questionnaire Survey

Safety Culture in Construction

Organizations

Thanwadee Chinda

PMB 50 Gold Coast Mail Centre


Griffith University

Gold Coast Campus

Queensland 9726 Australia


202


Griffith University, Gold Coast Campus

PMB 50 Gold Coast Mail Centre 9726, Queensland, Australia

February 28, 2008

Dear Sir/ Madam;

The following questionnaire has been developed to assess the organizational safety

culture in the context of construction. Please set aside 20 minutes for filling out this

questionnaire. The information will be used for academic purposes only, as integral part

of a PhD study. Individual responses will be kept confidential. Only a consolidated

summary of the result may be published.

The questionnaire comprises a series of statements relating to different aspects of

construction safety culture. Please indicate your level of agreement or disagreement

using a five-point Likert scale (for example 1 = strongly disagree; 5 = strongly agree).

Once you have completed the questionnaire, please return it in the envelope provided as

soon as possible. Should you have any questions, please do not hesitate to contact me at

+61-422062007 (Australia); 01-8802408 (Thailand) or e-mail me at:

[email protected] or write to me: Thanwadee Chinda 153/3 Moo 12 Khaowang-

Numpu Rd. Jadeehuk Muang Ratchaburi 70000 Thailand. Thank you in advance for

your cooperation.

Yours sincerely, Endorsement

Thanwadee Chinda Prof. Sherif Mohamed

(PhD Candidate) (Principal Supervisor)


203

��

�� !"� ��"� ��#��$�� %��"��$�

6 &�'�� 2549

��$��(��$��;

�))�)*��+&��$,��-�!.,� ��/+�/� ��%��0� ��%��'�� (safety culture) /�� 1

�� (�� &2�� 20 ��$��/+�/� �� ))�)*��$,!��$3��,��*� 1

�-�3%/+�� . 4��(��,� ��1 �%5��(��.�!� ��-��)� �"��!�� $

�� /��!� “A system dynamics approach to safety culture” !��,��*� � 6)3��(��$��$1

��7��(�8��7��$3�� %��(��,�

�))�)*��+&��$,�)(��%5� 3 �(��

�(��$ 1 �%5��-�*��3%� $�� )�(�� $�(��-��(

�(��$ 2 %�� )3%��!��,�� 34 !�� $�� )��0� ��%��'��/�� 1

�� (�� &2�� )��6��3�(��6��!��(��/��"(��!�� (��) 1

��(� �)3�(��6�� $�&� ��) 3 ��(� �)3�(�+��6��3�(��6�� ) 5 ��(� �)��6�� $�&�)

�(��$ 3 �%5� ��&%7��6�!��(�� /��!��

��(��")�))�)*��$,��6�� &2��(� ��)��/�(/�9��"��"�%:�$��)��1��1

7��-��*3%��)�))�)*��3��*��"�� *��(��$!��%�� /� �&2�"��"(�$

+61-422062007 (Australia)

+66-81-8802408 (Thailand) ��

E-mail ��$ [email protected]

!!)�&2�(��/� ��/��(��

��1 1 1 1 1 1 1 ��)��1

1

��$1��1 1 1 1 1 1 1 ;<=>?1@AB<C>1D=AEFBG1

H�� . 4�I1 1 1 1 1 1 1 H��$%�. 4�I1

�


204

Part I: Personal Information (�� )

Please fill in the blanks or tick the appropriate box. ( �&2�")�-�*��/�+(��(��1�� +(�

�$�(��"�� )

�

1. The name of your organization (optional) ___________________________________

+�� !��(��1H�� ")��3�(")3��I11

2. Your job title: ______________________________________

"-��(�!��(��11

3. How long have you been working in this organization? __________ year/s

�(��-��/�� $,��1 1111 1 1 1 1 11111111%J1

�

4. How long have you been working in the construction industry? __________ year/s

�(��-��/��&"�� (��11 1111111111 1 11111111111%J1

5. Does your organization have a formal safety policy (such as safety manuals)?

�� !��(��$��)�� $�� )��%��'��(��%5�� 3�(1 1

� Yes � No

��$1 1 1111 1 1 11113�(�$1

6. Does your current role have direct safety responsibilities?

)�)��!��(��/�� %K��&)��$�(�� $��!�� )��)7��+)��%��'��3�(1 1

� Yes � No

11111�$1 1 1111 1 1 11113�(�$1

If yes, is it in:

*��$1�%5��)7��+)��1

� Planning and/or reviewing safety policy� � � Safety auditing

��7�1��L��1�)��)��%��'��11 1111"��)��%��'��1

� Supervising safety in site works � � � � Other (please specify) �

�%5��$%�. 4��%��'��/�1site �� 1 1 1111��M1H�%��)&I1

1

�


205

7. How often do you engage in safety related activities (such as safety training, safety

auditing, safety meeting, etc.)?

�(��$�(�� $��!��$�(��(��/� �� %��'��1 H�+(�1 ��NO N�� %P�)�"��1

��%��'��Q1 ��"��)��%��'��Q1 ��!��(��%��+&��%��'��I1 ��

��$��/�1 1 1

� Once a month � Once every 6 months �

�& ��1 111 11111 1 1111�& 1R1��111111 1 1111 1

� Once a year � Other (please specify) _______________

�& %J1 1 1 1 1 1111��M1H�%��)&I1

8. When was the last time you have engaged in safety related activities (such as safety

training, safety auditing, safety meeting, etc.)?

�(��$�(�� $��!��$�(��(��/� �� %��'��1 H�+(� ��NO N�� %P�)�"��1

��%��'��Q1 ��"��)��%��'��Q1 ��!��(��%��+&��%��'��I1 ��,��&��1

��3�1

� Last month � Last 6 months

��$��1 1 1 1 1111R1��$��1

� Last year � Other (please specify) _______________

%J�$��1 1 1 1 1111��M1H�%��)&I1

9. In your opinion, how would you rank your organization’ s safety performance

compared to the national’ s average record?

/��6�!��(��1 �(��/�� 2S��%�� '��%��'��/�� !��(��1

�%�$�)��$�) �)� 2S��8�$�/�%��(��3�1 1

� Above � Same � Below

�$ �(�� 2S��8�$�1 11111 1111��(� �)� 2S��8�$�1 1 1111"-� �(�� 2S��8�$�1


206

Part II: Safety Culture (�� !�� "��)

This part contains 34 statements relating to construction safety culture. Please complete

this part by circling the score that best reflects the level of your agreement or

disagreement with each statement.

(�(��$,%�� )��!��,��1 TU1 !�� $�� )��0� ��%��'��/�� 1

�� (��1 �&2�� )��6��3�(��6��!��(��/��"(��!��)

The meaning of each score is shown below. (��!��"(��)��%5��$,)

1 2 3 4 5 N/A

Strongly Disagree

3�(��6�� $�&�1

Disagree

3�(��6��1

Neither Agree nor Disagree

3�(�+��6��3�(��6��1

Agree

��6��1

Strongly Agree

��6�� $�&�1

!3�(")1

No. Statement (�� ) Score

(�� )

Leadership (�� !#�$��)�

1. In our organization, management takes safety seriously NV��)��/�� !��/��-��W� $�� )��%��'��(�� 1

1 2 3 4 5 N/A

2.

Management encourages workers to give opinions and/or suggestions on safety matters NV��)��+� +��/�� H��,��I ��6��L��!�1��%��'��

1 2 3 4 5 N/A

3. In our organization, management makes sure that workers hold their responsibilities for their own safety NV��)��/�� !��,-��/��(/��(��)7��+)/��$��1��%��'��!�"�1

1 2 3 4 5 N/A

4.

Management acts quickly to correct safety problems when brought to his/her attention NV��)�� 3!%KW��%��'��(��$%KW�� !.,�1

1 2 3 4 5 N/A

Policy and Strategy (�%��&��'��(��)

5. It is our policy to recognize workers with good safe behaviours ��$��)��/� ��+��+��$�$��%��X"��%��'��$��$��1

1 2 3 4 5 N/A

6. It is our policy to give safety the same priority as production ��$��)��/� ��/��-��W��%��'��(� �)�� 7��"1

1 2 3 4 5 N/A

7. Our organization has a safety policy that gets reviewed and upgraded regularly �� !��$��)��%��'��$3��) ��)��%��)%�&�/��$!.,�1��1

1 2 3 4 5 N/A

8. In our organization, safety initiatives are proactively planned in order to continually improve our safety standards /�� !��%��'��3��) ��7��%5��(��$1��%��+��/� ��0��"�#��%��'��1

1 2 3 4 5 N/A

9. In our organization, safety is an integral part in formulating our business decisions and goals /�� !��%��'��*� ��%5��(��.�!��7��%Y��1 &� ��1

1 2 3 4 5 N/A


207

No. Statement (�� ) Score (�� )

People ()��)

10. Our project staff (including workers) believe that our organization is genuinely concerned about workplace safety �� H��,��I�+��(�� !��-��.�*.��%��'��1/��$�-��

1 2 3 4 5 N/A

11. Our project staff (including workers) fully understand their safety responsibilities �� H��,��I!��$��!��/�/��$��)7��+)��%��'��!�"�1

1 2 3 4 5 N/A

12. In our organization, workers can seek advice on safety matters from their immediate boss, such as project manager, safety manager, supervisor etc. /�� !��*!�-��-��%��'��3�� !�"�1 �+(�1�� 7��)�� Q17��)��%��'��Q1�$%�. 4�1��1��M1

1 2 3 4 5 N/A

13. In our organization, project staff (including workers) are involved, formally and/or informally, in safety related issues /�� !��1 �� H��,��I�$�(��(��1 ��,��(��%5�1�� L��3�(�%5�� 1/��!�� $�� )��%��'��1

1 2 3 4 5 N/A

14. In our organization, workmates often give suggestions to each other on how to work safely /�� !��(�� /��-��-�/� ��-��(��%��'�� (1 ��

1 2 3 4 5 N/A

15. In our organization, workload is reasonably balanced among workers so that they can get the job done safely /�� !��1 %��2��$��&�� )�-��(��13��-��(��%��'��1

1 2 3 4 5 N/A

16. Our organization ensures that workers are not under pressure to avoid unsafe behaviours �� !��,-��/��(/��(��3�(3��-��'��/"�� 1��1��$ ��$��X"� ��$��$��"(��%��'��1

1 2 3 4 5 N/A

Partnerships and Resources (*&��+�$�� )��)

17.

Project participants, such as subcontractors, cooperate with us in following our safety standards 7��$�(��(��/�� +(�7��)��(�1 /��(��/� ��%P�)�"�"��1��"�#��%��'��!�� !��1

1 2 3 4 5 N/A

18. In our organization, financial resources are adequately provided to support the implementation of our safety policy �� !��/�� )��&�� (��$��1 ��-�3%/+�/� ��1�-�� "��)��%��'��!�� 1

1 2 3 4 5 N/A

19. Our organization has sufficient necessary safety resources available so that workers can carry out their jobs safely �� !��$&% �2��%��'��$�-��W�(��1��$��1��/+��1��/��-��3��(��%��'��

1 2 3 4 5 N/A

20. Our organization endeavours to have adequate human resources to get the job done safely �� !��$�� &4��(��$��$��-�/��"(��M�-��6�3��(��1%��'��1

1 2 3 4 5 N/A


208

No. Statement (�� ) Score ,�� -

Processes (��)

21. In our organization, we provide adequate training for those performing new tasks safely ��$��/��(�$3�(��-�� (�1 �� !��/�� )��&� ��NO N�1�� %P�)�"��%��'��(��$��1 ��/�� -��%5�3%1�(��%��'��1

1 2 3 4 5 N/A

22. In our organization, risk and hazard assessment is a part of our routine safety planned activities /�� !��1 ��%��7��$��"��$�� !.,�� 1 ��-��1�%5��(��.�!� �� %��'��%��-��1

1 2 3 4 5 N/A

23. Feedback on safety implementation is encouraged within the organization in order to improve safety performance �� !��)��&�/��$ ��+$,��!��$��L��!�) ��(�!� ��-��1��%��'��1 ��7�/� ��-�3%%��)%�&��0��%�� '��1��%��'��

1 2 3 4 5 N/A

24. Our organization adopts a no-blame approach so that workers always report near misses and accidents they experience or witness �� !��-��)) ��3�("��"$��1 3�( �(��41 ��3�(��4��/+�1 ��,�1��!��.��"6�/��&)�"��"&�$"��%��)��)��6��3�(%��%K�1

1 2 3 4 5 N/A

25. In our organization, site housekeeping is maintained at a high level /�� !��1 �� 4�1 ZC[B1 ��/��$��%��'��(/�1��)��1

1 2 3 4 5 N/A

26. Our organization keeps accidents records to investigate their causes �� !�� 6)��&)�"��"&��-�3%/+�/� ��)��"&!�1&)�"��"&��,�M

1 2 3 4 5 N/A

27. Our organization has a good safety benchmarking system to compare with other construction organizations �� !��$��)) ��%�$�)��$�)%��7�H\B]̂AFE<_C]̀I��1%��'��$�$1 ��-�3%/+�/� ��%�$�)��$�)%��7� �)�� 1 �� (��M�

1 2 3 4 5 N/A

Goals ($�� .�� *��)

28. Workers are generally satisfied with the way we currently manage safety in our organization ��$��.��/�/� ��)��%��'��!�� /�1%K��&)��1

1 2 3 4 5 N/A

29. The way we currently manage safety in our organization promotes safe work behaviour ��)��%��'��!�� /�%K��&)��$�(��+(��/� ��1��)��&�/�� X"� ��/� ��-��(��%��'��1

1 2 3 4 5 N/A

30. The way we currently manage safety in our organization helps us reduce the number of severe accidents and safety related incidents ��)��%��'��!�� /�%K��&)��1 �$�(��+(��/� ��1�-��&)�"��"&�$� ��!.,�/�� 1

1 2 3 4 5 N/A

31. The way we currently manage safety in our organization helps us meet our clients’ expectations ��)��%��'��!�� /�%K��&)��1 �$�(��+(��/� ��1 )��&��"�� !�� 1

1 2 3 4 5 N/A

32. Public perceive our organization with a good safety image �� 2+�� !��(��$'��%��'��$�$1

1 2 3 4 5 N/A

33. The way we currently manage safety in our organization has a positive influence on workers’ morale ��)��%��'��!�� /�%K��&)��1 �$7��)� "(1�$� ��!��11

1 2 3 4 5 N/A


209

No. Statement (�� ) Score (�� )

Goals (Cont.) ($�� .�� *��(��)

34. The way we currently manage safety in our organization leads to reduction in the total costs associated with accidents ��)��%��'��!�� /�%K��&)��1 �-�3%��( ��!�1�(�/+��(��,��$� $��!�� )&)�"��"&1 �+(�1 �(��$��Q1 �(�� 4��)��1��M1

1 2 3 4 5 N/A

Your opinions (�� . �*/��)

1. Considering the five enablers listed above (Leadership, Policy and Strategy, People, Partnerships and Resources and Processes), which enabler do you think the Thai construction industry considers being the most influential in significantly improving safety culture? (Please tick only one box) ��2�1a1��!�� !��)�1H9.�%�� )��1��%5�7��-�Q1��)��&� ��"��Q1�� Q1�&��(��17��$�(��(��1��1�� Q1�� )�� I1�(��(��!�3��$�&(��*�� %��0�� -��.��(��$�� $�&�/� ��%��)%�&��0��0� ��1%��'��1H �&2�� $��1b1+(�I11

� Leadership � Policy and Strategy 1111��%5�7��-�1 1 1 1 1111��)��&� ��"��1� People � Partnerships and Resources

1 1111��11111 1 1 111111111111111�&��(��17��$�(��(��1��1�� 1 � Processes ��)�� 1 2. Considering the five enablers listed above (Leadership, Policy and Strategy, People,

Partnerships and Resources and Processes), which enabler do you think your organization considers being the most influential in significantly improving safety culture? (Please tick only one box) ��2�1a1��!�� !��)�1H9.�%�� )��1��%5�7��-�Q1��)��&� ��"��Q1�� Q1�&��(��17��$�(��(��1��1�� Q1�� )�� I1�(��(��!�3��$��-��.��(��$1 �� $�&�1 /� ��%��)%�&��0��0� ��%��'��1 H �&2�� $��1 b1+(�I1

� Leadership � Policy and Strategy 1111��%5�7��-�1 1 1 1 11111��)��&� ��"��1� People � Partnerships and Resources

1 1111��11111 1 1 1 1 11111�&��(��17��$�(��(��1��1�� 1 � Processes ��)�� 1


210

3. How will you rank the five enablers listed above (Leadership, Policy and Strategy, People, Partnerships and Resources and Processes) according to its importance in enabling a more positive safety culture? �(��$��-��)1 a1 ��!�� !��)�1 H9.�%�� )��1 ��%5�7��-�Q1 ��)��&� ��"��Q1�� Q1�&��(��17��$�(��(��1�� Q1�� )�� I1�(��3�1��/+�� 2S��-��W/�1 ��-�/�� 0� ��%��'��)� �� !.,�1

___ Leadership ___ Policy and Strategy 11111111��%5�7��-�1 1 1 11111111��)��&� ��"��1___ People ___ Partnerships and Resources

1 1111111��11111 1 1 1 11111111�&��(��17��$�(��(��1��1�� 1 ___ Processes 11 1111111 ��)�� 1

4. Do you have other suggestions to improve the safety culture on construction sites? �(��$!��/� ��%��)%�&��0��0� ��%��'��1 /�1 ZC[B1 ��!� ��1 �� (��3�(1

� Yes � No 11111�$1 1 1111 1 1 11113�(�$1

If yes, please write them down in the space provided: *��$1 �&2��!$��!��!��(��(��$,1

_____________________________________________________________________________________ __________________________________________________________________________________________________________________________________________________________________________ __________________________________________________________________________________________________________________________________________________________________________

Thank you very much for your time and effort ��&1��2��2*�� 3��


211

AAppppeennddiixx 22

RRaaww DDaattaa

A S

yste

m D

ynam

ics

App

roac

h to

Con

stru

ctio

n Sa

fety

Cul

ture

21

2

case

cm

mt

Com

m.

accn

ld

bx

awrn

al

gn

stnd

in

it in

tg

prcp

re

sp

sppt

in

vm

rlsp

w

kld

prsr

1

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

2 ��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

3 ��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

4 ��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

5 ��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

6 ��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

7 ��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

8 ��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

9 ��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

10

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

11

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

12

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

13

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

14

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

15

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

16

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

17

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

18

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

19

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

20

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

21

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

22

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

23

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

24

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

25

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

26

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

27

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

28

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

29

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

30

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

Not

e: c

mm

t = th

e ‘c

omm

itmen

t’ i

tem

, com

m =

the

‘com

mun

icat

ion’

item

, acc

n =

the

‘acc

ount

abili

ty’

item

, ldb

x =

the

‘lea

ding

by

exam

ple’

item

, aw

rn =

the

‘saf

ety

awar

enes

s’ it

em,

algn

= th

e ‘s

afet

y an

d pr

oduc

tivity

alig

nmen

t’ it

em, s

tnd

= th

e ‘s

afet

y st

anda

rds’

item

, ini

t = th

e ‘s

afet

y in

itiat

ives

’ ite

m, i

ntg

= th

e ‘s

afet

y in

tegr

atio

n in

bus

ines

s go

als’

item

, prc

p =

the

‘sha

red

perc

eptio

ns’

item

, res

p =

the

‘saf

ety

resp

onsi

bilit

ies’

item

, spp

t = th

e ‘s

uppo

rtiv

e en

viro

nmen

t’ it

em, i

nvm

= th

e ‘w

orke

rs’

invo

lvem

ent’

item

, rls

p =

the

‘wor

kers

’ re

latio

nshi

ps’

item

, wkl

d =

the

‘wor

kloa

d’ it

em, p

rsr =

the

‘wor

k pr

essu

re’ i

tem


213

case cmmt Comm. accn ldbx awrn algn stnd init intg prcp resp sppt invm rlsp wkld prsr 31 ��

32 ��

33 ��

34 ��

35 ��

36 ��

37 ��

38 ��

39 ��

40 ��

41 ��

42 ��

43 ��

44 ��

45 ��

46 ��

47 ��

48 ��

49 ��

50 ��

51 ��

52 ��

53 ��

54 ��

55 ��

56 ��

57 ��

58 ��

59 ��

60 ��

Note: cmmt = the ‘commitment’ item, comm = the ‘communication’ item, accn = the ‘accountability’ item, ldbx = the ‘leading by example’ item, awrn = the ‘safety awareness’ item, algn = the ‘safety and productivity alignment’ item, stnd = the ‘safety standards’ item, init = the ‘safety initiatives’ item, intg = the ‘safety integration in business goals’ item, prcp = the ‘shared perceptions’ item, resp = the ‘safety responsibilities’ item, sppt = the ‘supportive environment’ item, invm = the ‘workers’ involvement’ item, rlsp = the ‘workers’ relationships’ item, wkld = the ‘workload’ item, prsr = the ‘work pressure’ item


214


62 ��

63 ��

64 ��

65 ��

66 ��

67 ��

68 ��

69 ��

70 ��

71 ��

72 ��

73 � � ��

74 ��

75 ��

76 ��

77 ��

78 ��

79 ��

80 ��

81 ��

82 ��

83 ��

84 ��

85 ��

86 ��

87 ��

88 ��

89 ��

90 ��



215


92 ��

93 ��

94 ��

95 ��

96 ��

97 ��

98 ��

99 ��

100 ��

101 ��

102 ��

103 ��

104 ��

105 ��

106 ��

107 ��

108 ��

109 ��

110 ��

111 ��

112 ��

113 ��

114 ��

115 ��



216

case coop finc Resc hmnr trng risk fdbk nobm hskp docu bnmk jstf swbh acci cstm imge mrle cost 1 ��

2 ��

3 ��

4 ��

5 ��

6 ��

7 ��

8 ��

9 ��

10 ��

11 ��

12 ��

13 ��

14 ��

15 ��

16 ��

17 ��

18 ��

19 ��

20 ��

21 ��

22 ��

23 ��

24 ��

25 ��

26 ��

27 ��

28 ��

29 ��

30 ��

Note: coop = the ‘stakeholders’ cooperation’ item, finc = the ‘financial resources’ item, resc = the ‘safety resources’ item, hmnr = the ‘human resources’ item, trng = the ‘training’ item, risk = the ‘risk assessment’ item, fdbk = the ‘feedback’ item, nobm = the ‘no-blame approach’ item, hskp = the ‘housekeeping’ item, docu = the ‘safety documentation’ item, bnmk = the ‘benchmarking system’ item, jstf = the ‘job satisfaction’ item, swbh = the ‘safe work behaviour’ item, acci = the ‘number of accidents’ item, cstm = the ‘customers’ expectations’ item, imge = the ‘industrial image’ item, mrle = the ‘workforce morale’ item, cost = the ‘cost of accidents’ item

A S

yste

m D

ynam

ics

App

roac

h to

Con

stru

ctio

n Sa

fety

Cul

ture

21

7

case

co

op

finc

Res

c hm

nr

trng

ri

sk

fdbk

no

bm

hskp

do

cu

bnm

k js

tf

swbh

ac

ci

cstm

im

ge

mrl

e co

st

31

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

32

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

33

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

34

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

35

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

36

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

37

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

38

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

39

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

40

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

41

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

42

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

43

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

44

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

45

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

46

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

47

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

48

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

49

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

50

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

51

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

52

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

53

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

54

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

55

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

56

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

57

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

58

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

59

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

60

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

Not

e: c

oop

= th

e ‘s

take

hold

ers’

coo

pera

tion’

item

, fin

c =

the

‘fin

anci

al r

esou

rces

’ ite

m, r

esc

= th

e ‘s

afet

y re

sour

ces’

item

, hm

nr =

the

‘hum

an r

esou

rces

’ ite

m, t

rng

= th

e ‘t

rain

ing’

item

, ri

sk =

the

‘ris

k as

sess

men

t’ it

em, f

dbk

= th

e ‘f

eedb

ack’

item

, nob

m =

the

‘no-

blam

e ap

proa

ch’ i

tem

, hsk

p =

the

‘hou

seke

epin

g’ it

em, d

ocu

= th

e ‘s

afet

y do

cum

enta

tion’

item

, bnm

k =

the

‘ben

chm

arki

ng s

yste

m’

item

, jst

f =

the

‘job

sat

isfa

ctio

n’ it

em, s

wbh

= th

e ‘s

afe

wor

k be

havi

our’

item

, acc

i = th

e ‘n

umbe

r of

acc

iden

ts’

item

, cst

m =

the

‘cus

tom

ers’

exp

ecta

tions

’ ite

m,

imge

= th

e ‘i

ndus

tria

l im

age’

item

, mrl

e =

the

‘wor

kfor

ce m

oral

e’ it

em, c

ost =

the

‘cos

t of a

ccid

ents

’ ite

m

A S

yste

m D

ynam

ics

App

roac

h to

Con

stru

ctio

n Sa

fety

Cul

ture

21

8

case

co

op

finc

Res

c hm

nr

trng

ri

sk

fdbk

no

bm

hskp

do

cu

bnm

k js

tf

swbh

ac

ci

cstm

im

ge

mrl

e co

st

61

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

62

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

63

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

64

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

65

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

66

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

67

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

68

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

69

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

70

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

71

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

72

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

73

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

74

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

75

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

76

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

77

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

78

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

79

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

80

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

81

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

82

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

83

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

84

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

85

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

86

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

87

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

88

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

89

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

90

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

Not

e: c

oop

= th

e ‘s

take

hold

ers’

coo

pera

tion’

item

, fin

c =

the

‘fin

anci

al r

esou

rces

’ ite

m, r

esc

= th

e ‘s

afet

y re

sour

ces’

item

, hm

nr =

the

‘hum

an r

esou

rces

’ ite

m, t

rng

= th

e ‘t

rain

ing’

item

, ri

sk =

the

‘ris

k as

sess

men

t’ it

em, f

dbk

= th

e ‘f

eedb

ack’

item

, nob

m =

the

‘no-

blam

e ap

proa

ch’ i

tem

, hsk

p =

the

‘hou

seke

epin

g’ it

em, d

ocu

= th

e ‘s

afet

y do

cum

enta

tion’

item

, bnm

k =

the

‘ben

chm

arki

ng s

yste

m’

item

, jst

f =

the

‘job

sat

isfa

ctio

n’ it

em, s

wbh

= th

e ‘s

afe

wor

k be

havi

our’

item

, acc

i = th

e ‘n

umbe

r of

acc

iden

ts’

item

, cst

m =

the

‘cus

tom

ers’

exp

ecta

tions

’ ite

m,

imge

= th

e ‘i

ndus

tria

l im

age’

item

, mrl

e =

the

‘wor

kfor

ce m

oral

e’ it

em, c

ost =

the

‘cos

t of a

ccid

ents

’ ite

m

A S

yste

m D

ynam

ics

App

roac

h to

Con

stru

ctio

n Sa

fety

Cul

ture

21

9

case

co

op

finc

Res

c hm

nr

trng

ri

sk

fdbk

no

bm

hskp

do

cu

bnm

k js

tf

swbh

ac

ci

cstm

im

ge

mrl

e co

st

91

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

92

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

93

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

94

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

95

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

96

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

97

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

98

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

99

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

100

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

101

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

102

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

103

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

104

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

105

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

106

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

107

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

108

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

109

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

110

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

111

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

112

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

113

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

114

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

115

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

Not

e: c

oop

= th

e ‘s

take

hold

ers’

coo

pera

tion’

item

, fin

c =

the

‘fin

anci

al r

esou

rces

’ ite

m, r

esc

= th

e ‘s

afet

y re

sour

ces’

item

, hm

nr =

the

‘hum

an r

esou

rces

’ ite

m, t

rng

= th

e ‘t

rain

ing’

item

, ri

sk =

the

‘ris

k as

sess

men

t’ it

em, f

dbk

= th

e ‘f

eedb

ack’

item

, nob

m =

the

‘no-

blam

e ap

proa

ch’ i

tem

, hsk

p =

the

‘hou

seke

epin

g’ it

em, d

ocu

= th

e ‘s

afet

y do

cum

enta

tion’

item

, bnm

k =

the

‘ben

chm

arki

ng s

yste

m’

item

, jst

f =

the

‘job

sat

isfa

ctio

n’ it

em, s

wbh

= th

e ‘s

afe

wor

k be

havi

our’

item

, acc

i = th

e ‘n

umbe

r of

acc

iden

ts’

item

, cst

m =

the

‘cus

tom

ers’

exp

ecta

tions

’ ite

m,

imge

= th

e ‘i

ndus

tria

l im

age’

item

, mrl

e =

the

‘wor

kfor

ce m

oral

e’ it

em, c

ost =

the

‘cos

t of a

ccid

ents

’ ite

m


220

AAppppeennddiixx 33

SSttaannddaarrddiizzeedd SSccoorreess ((ZZ--SSccoorreess))

A S

yste

m D

ynam

ics

App

roac

h to

Con

stru

ctio

n Sa

fety

Cul

ture

22

1

zc

mm

t Z

com

m

zacc

n zl

dbx

zaw

rn

zalg

n zs

tnd

zini

t zi

ntg

zprc

p zr

esp

zspp

t zi

nvm

zr

lsp

zwkl

d zp

rsr

1 1.

0 1.

4 1.

2 0.

8 0.

4 0.

2 0.

2 0.

2 0.

2 1.

4 0.

4 0.

0 0.

3 -0

.8

0.2

0.3

2 1.

0 0.

4 1.

2 0.

8 0.

4 0.

2 1.

3 0.

2 0.

2 0.

1 -1

.0

0.0

0.3

-0.8

0.

2 0.

3 3

1.0

1.4

1.2

0.8

1.3

1.2

1.3

1.3

1.2

1.4

0.4

0.0

1.6

0.4

1.4

1.3

4 -0

.2

0.4

1.2

-0.3

0.

4 0.

2 1.

3 0.

2 0.

2 0.

1 0.

4 -1

.3

-0.9

-0

.8

0.2

-0.8

5

1.0

0.4

0.0

0.8

0.4

0.2

0.2

1.3

1.2

0.1

0.4

1.3

0.3

1.6

1.4

0.3

6 -0

.2

0.4

0.0

-0.3

-0

.6

-0.8

0.

2 -0

.9

0.2

0.1

0.4

-1.3

0.

3 0.

4 -1

.0

0.3

7 -0

.2

0.4

0.0

0.8

-0.6

-0

.8

0.2

0.2

-0.8

0.

1 -1

.0

0.0

0.3

0.4

0.2

0.3

8 1.

0 1.

4 1.

2 0.

8 1.

3 1.

2 1.

3 1.

3 1.

2 1.

4 0.

4 0.

0 0.

3 0.

4 0.

2 0.

3 9

1.0

1.4

1.2

0.8

1.3

1.2

1.3

1.3

1.2

1.4

1.8

1.3

1.6

1.6

1.4

1.3

10

1.0

-0.7

1.

2 -0

.3

-0.6

-0

.8

-3.1

-3

.2

-0.8

-1

.2

-2.4

1.

3 -0

.9

-0.8

-2

.1

-1.9

11

-0

.2

0.4

-1.2

-0

.3

0.4

1.2

-0.9

0.

2 1.

2 -1

.2

0.4

0.0

1.6

0.4

-1.0

-0

.8

12

-0.2

0.

4 0.

0 -0

.3

0.4

-0.8

-0

.9

0.2

0.2

-1.2

-1

.0

0.0

0.3

0.4

-1.0

-0

.8

13

-0.2

0.

4 0.

0 0.

8 1.

3 0.

2 0.

2 0.

2 0.

2 0.

1 0.

4 1.

3 0.

3 1.

6 0.

2 1.

3 14

-0

.2

0.4

0.0

-0.3

-0

.6

0.2

1.3

0.2

0.2

0.1

-1.0

-1

.3

-0.9

0.

4 0.

2 1.

3 15

1.

0 0.

4 0.

0 -0

.3

-0.6

0.

2 1.

3 0.

2 0.

2 -1

.2

-1.0

-1

.3

0.3

0.4

0.2

1.3

16

-0.2

0.

4 0.

0 -0

.3

0.4

-0.8

0.

2 0.

2 0.

2 0.

1 -1

.0

-1.3

-0

.9

0.4

0.2

1.3

17

-1.5

-0

.7

0.0

0.8

0.4

0.2

0.2

0.2

0.2

0.1

-1.0

0.

0 0.

3 0.

4 0.

2 -0

.8

18

1.0

0.4

1.2

0.8

-0.6

0.

2 0.

2 0.

2 1.

2 0.

1 0.

4 -1

.3

0.3

0.4

0.2

0.3

19

-0.2

-0

.7

0.0

-0.3

0.

4 0.

2 0.

2 0.

2 0.

2 1.

4 1.

8 0.

0 0.

3 -0

.8

0.2

0.3

20

-0.2

0.

4 1.

2 -0

.3

0.4

0.2

0.2

0.2

-0.8

0.

1 0.

4 0.

0 0.

3 0.

4 1.

4 1.

3 21

-0

.2

-0.7

0.

0 0.

8 -1

.5

-0.8

0.

2 0.

2 0.

2 -1

.2

-1.0

0.

0 0.

3 -0

.8

-1.0

0.

3 22

-0

.2

-0.7

-1

.2

-1.4

-1

.5

-1.8

0.

2 0.

2 -0

.8

-1.2

-1

.0

-1.3

-2

.1

-0.8

-1

.0

-0.8

23

-0

.2

-0.7

-1

.2

0.8

-2.4

-1

.8

-0.9

-2

.0

-1.8

-1

.2

0.4

0.0

-0.9

-2

.0

0.2

-0.8

24

1.

0 1.

4 1.

2 -3

.5

-2.4

-1

.8

-0.9

0.

2 -0

.8

0.1

-1.0

0.

0 0.

3 -0

.8

-1.0

-0

.8

25

1.0

-0.7

0.

0 -2

.4

-0.6

-0

.8

-0.9

-0

.9

0.2

-2.6

-1

.0

-1.3

-2

.1

0.4

-1.0

-0

.8

26

1.0

0.4

0.0

0.8

-0.6

0.

2 0.

2 0.

2 0.

2 1.

4 0.

4 0.

0 0.

3 0.

4 0.

2 -0

.8

27

-1.5

0.

4 0.

0 0.

8 0.

4 0.

2 -0

.9

-0.9

-0

.8

0.1

-1.0

0.

0 -0

.9

-0.8

0.

2 -0

.8

28

1.0

1.4

0.0

0.8

-0.6

1.

2 -0

.9

-0.9

-0

.8

-2.6

-1

.0

0.0

-2.1

-2

.0

0.2

0.3

29

-1.5

-0

.7

0.0

-0.3

0.

4 -0

.8

-0.9

-0

.9

-0.8

0.

1 0.

4 0.

0 -0

.9

0.4

0.2

0.3

30

1.0

1.4

0.0

0.8

1.3

1.2

1.3

0.2

1.2

1.4

1.8

1.3

1.6

0.4

1.4

0.3

Not

e: z

cmm

t = z

-sco

re o

f the

‘com

mitm

ent’

item

, zco

mm

= z

-sco

re o

f the

‘com

mun

icat

ion’

item

, zac

cn =

z-s

core

of t

he ‘a

ccou

ntab

ility

’ ite

m, z

ldbx

= z

-sco

re o

f the

‘lea

ding

by

exam

ple’

ite

m, z

awrn

= z

-sco

re o

f the

‘sa

fety

aw

aren

ess’

item

, zal

gn =

z-s

core

of t

he ‘

safe

ty a

nd p

rodu

ctiv

ity a

lignm

ent’

item

, zst

nd =

z-s

core

of t

he ‘

safe

ty s

tand

ards

’ ite

m, z

init

= z-

scor

e of

the

‘saf

ety

initi

ativ

es’

item

, zin

tg =

z-s

core

of

the

‘saf

ety

inte

grat

ion

in b

usin

ess

goal

s’ it

em, z

prcp

= z

-sco

re o

f th

e ‘s

hare

d pe

rcep

tions

’ ite

m, z

resp

= z

-sco

re o

f th

e ‘s

afet

y re

spon

sibi

litie

s’

item

, zsp

pt =

z-s

core

of

the

‘sup

port

ive

envi

ronm

ent’

item

, zin

vm =

z-s

core

of

the

‘wor

kers

’ in

volv

emen

t’ it

em, z

rlsp

= z

-sco

re o

f th

e ‘w

orke

rs’

rela

tions

hips

’ ite

m, z

wkl

d =

z-sc

ore

of

the

‘wor

kloa

d’ it

em, z

prsr

= z

-sco

re o

f the

‘wor

k pr

essu

re’ i

tem

A S

yste

m D

ynam

ics

App

roac

h to

Con

stru

ctio

n Sa

fety

Cul

ture

22

2

case

zc

mm

t zc

omm

za

ccn

zldb

x za

wrn

za

lgn

zstn

d zi

nit

zint

g zp

rcp

zres

p zs

ppt

zinv

m

zrls

p zw

kld

zprs

r 31

-0

.2

-2.8

-1

.2

-2.4

-2

.4

-0.8

-0

.9

-0.9

-0

.8

0.1

0.4

0.0

-0.9

-2

.0

-1.0

-0

.8

32

-0.2

0.

4 0.

0 0.

8 0.

4 0.

2 0.

2 0.

2 -0

.8

0.1

0.4

-1.3

0.

3 0.

4 0.

2 0.

3 33

-1

.5

-1.8

-1

.2

0.8

-1.5

0.

2 0.

2 0.

2 0.

2 -1

.2

0.4

-1.3

-0

.9

-2.0

-1

.0

0.3

34

-0.2

-0

.7

0.0

0.8

-0.6

-0

.8

-0.9

-0

.9

-0.8

0.

1 0.

4 0.

0 0.

3 0.

4 0.

2 0.

3 35

-0

.2

-0.7

0.

0 -0

.3

0.4

0.2

0.2

0.2

-0.8

0.

1 0.

4 0.

0 0.

3 -0

.8

-2.1

-0

.8

36

1.0

1.4

1.2

0.8

1.3

1.2

1.3

1.3

1.2

1.4

1.8

1.3

1.6

1.6

1.4

1.3

37

-2.7

-1

.8

-1.2

-2

.4

-0.6

-1

.8

-3.1

-2

.0

-0.8

0.

1 -1

.0

-1.3

-0

.9

-0.8

0.

2 -3

.0

38

-0.2

-0

.7

0.0

-0.3

-0

.6

0.2

1.3

0.2

0.2

1.4

0.4

0.0

0.3

-0.8

0.

2 0.

3 39

1.

0 0.

4 1.

2 0.

8 0.

4 1.

2 0.

2 0.

2 1.

2 0.

1 0.

4 0.

0 0.

3 -0

.8

0.2

0.3

40

-0.2

0.

4 0.

0 -0

.3

0.4

0.2

0.2

0.2

-0.8

0.

1 0.

4 0.

0 0.

3 0.

4 0.

2 0.

3 41

-0

.2

-0.7

0.

0 0.

8 0.

4 0.

2 0.

2 0.

2 -0

.8

0.1

0.4

0.0

-0.9

1.

6 0.

2 0.

3 42

-0

.2

0.4

0.0

-0.3

0.

4 0.

2 0.

2 0.

2 0.

2 0.

1 -1

.0

-1.3

0.

3 -0

.8

0.2

0.3

43

1.0

1.4

1.2

0.8

0.4

1.2

1.3

1.3

1.2

1.4

1.8

1.3

1.6

-0.8

0.

2 1.

3 44

-0

.2

0.4

1.2

0.8

1.3

0.2

0.2

1.3

1.2

1.4

1.8

0.0

0.3

1.6

1.4

1.3

45

-0.2

-0

.7

0.0

-0.3

0.

4 -0

.8

0.2

0.2

0.2

0.1

-1.0

0.

0 0.

3 -0

.8

-1.0

0.

3 46

1.

0 1.

4 1.

2 -0

.3

0.4

0.2

0.2

0.2

1.2

1.4

0.4

1.3

0.3

0.4

0.2

-0.8

47

1.

0 -0

.7

0.0

-1.4

0.

4 0.

2 0.

2 0.

2 1.

2 0.

1 -1

.0

0.0

0.3

-0.8

1.

4 -0

.8

48

-0.2

0.

4 0.

0 -0

.3

0.4

0.2

0.2

0.2

0.2

0.1

0.4

0.0

0.3

-0.8

-1

.0

0.3

49

1.0

1.4

1.2

0.8

1.3

1.2

1.3

1.3

1.2

1.4

1.8

1.3

1.6

1.6

1.4

1.3

50

-0.2

-0

.7

0.0

-0.3

-1

.5

0.2

0.2

0.2

-0.8

-1

.2

0.4

0.0

0.3

0.4

0.2

-0.8

51

-0

.2

0.4

0.0

-0.3

0.

4 0.

2 0.

2 0.

2 0.

2 0.

1 0.

4 0.

0 0.

3 0.

4 0.

2 0.

3 52

1.

0 1.

4 0.

0 0.

8 0.

4 0.

2 0.

2 1.

3 0.

2 1.

4 1.

8 0.

0 -2

.1

0.4

1.4

1.3

53

-0.2

0.

4 1.

2 0.

8 0.

4 1.

2 1.

3 0.

2 1.

2 1.

4 1.

8 0.

0 1.

6 1.

6 1.

4 1.

3 54

1.

0 0.

4 0.

0 0.

8 0.

4 0.

2 0.

2 0.

2 0.

2 0.

1 0.

4 0.

0 0.

3 0.

4 0.

2 0.

3 55

1.

0 1.

4 1.

2 0.

8 1.

3 1.

2 1.

3 1.

3 1.

2 1.

4 0.

4 0.

0 1.

6 0.

4 1.

4 1.

3 56

1.

0 0.

4 0.

0 0.

8 1.

3 1.

2 0.

2 0.

2 0.

2 0.

1 0.

4 0.

0 0.

3 0.

4 1.

4 0.

3 57

-0

.2

0.4

0.0

-0.3

0.

4 0.

2 0.

2 0.

2 0.

2 0.

1 0.

4 0.

0 0.

3 0.

4 0.

2 0.

3 58

-0

.2

0.4

0.0

0.8

0.4

1.2

0.2

0.2

0.2

0.1

0.4

0.0

0.3

0.4

0.2

0.3

59

1.0

0.4

0.0

0.8

0.4

0.2

0.2

1.3

0.2

-1.2

-1

.0

1.3

-0.9

0.

4 0.

2 0.

3 60

1.

0 -0

.7

-1.2

-1

.4

-0.6

1.

2 -2

.0

1.3

-2.8

-1

.2

-1.0

-1

.3

0.3

-0.8

-3

.3

-1.9

N

ote:

zcm

mt =

z-s

core

of t

he ‘c

omm

itmen

t’ it

em, z

com

m =

z-s

core

of t

he ‘c

omm

unic

atio

n’ it

em, z

accn

= z

-sco

re o

f the

‘acc

ount

abili

ty’ i

tem

, zld

bx =

z-s

core

of t

he ‘l

eadi

ng b

y ex

ampl

e’

item

, zaw

rn =

z-s

core

of t

he ‘

safe

ty a

war

enes

s’ it

em, z

algn

= z

-sco

re o

f the

‘sa

fety

and

pro

duct

ivity

alig

nmen

t’ it

em, z

stnd

= z

-sco

re o

f the

‘sa

fety

sta

ndar

ds’

item

, zin

it =

z-sc

ore

of th

e ‘s

afet

y in

itiat

ives

’ ite

m, z

intg

= z

-sco

re o

f th

e ‘s

afet

y in

tegr

atio

n in

bus

ines

s go

als’

item

, zpr

cp =

z-s

core

of

the

‘sha

red

perc

eptio

ns’

item

, zre

sp =

z-s

core

of

the

‘saf

ety

resp

onsi

bilit

ies’

ite

m, z

sppt

= z

-sco

re o

f th

e ‘s

uppo

rtiv

e en

viro

nmen

t’ it

em, z

invm

= z

-sco

re o

f th

e ‘w

orke

rs’

invo

lvem

ent’

item

, zrl

sp =

z-s

core

of

the

‘wor

kers

’ re

latio

nshi

ps’

item

, zw

kld

= z-

scor

e of

th

e ‘w

orkl

oad’

item

, zpr

sr =

z-s

core

of t

he ‘w

ork

pres

sure

’ ite

m


223

case zcmmt zcomm zaccn zldbx zawrn zalgn zstnd zinit zintg zprcp zresp zsppt zinvm zrlsp zwkld zprsr 61 -1.5 -0.7 0.0 -0.3 0.4 -0.8 -0.9 -0.9 0.2 0.1 0.4 -1.3 -0.9 0.4 -1.0 -1.9 62 -0.2 -1.8 -1.2 -0.3 -0.6 0.2 -0.9 -0.9 -0.8 0.1 -1.0 -1.3 -0.9 -0.8 -1.0 0.3 63 -1.5 -1.8 -2.4 -1.4 -0.6 -0.8 0.2 -0.9 -1.8 -1.2 -1.0 0.0 -2.1 -0.8 0.2 -0.8 64 1.0 0.4 0.0 -0.3 0.4 0.2 -0.9 -0.9 -0.8 0.1 0.4 0.0 0.3 0.4 0.2 1.3 65 -0.2 0.4 0.0 -0.3 0.4 0.2 0.2 0.2 0.2 0.1 0.4 0.0 0.3 0.4 0.2 0.3 66 1.0 -2.8 1.2 0.8 1.3 -2.8 1.3 1.3 1.2 1.4 1.8 0.0 1.6 0.4 1.4 1.3 67 -0.2 -1.8 0.0 -0.3 0.4 0.2 0.2 0.2 0.2 0.1 0.4 0.0 -0.9 -0.8 0.2 0.3 68 -1.5 0.4 1.2 0.8 -0.6 0.2 1.3 0.2 0.2 0.1 0.4 1.3 1.6 -0.8 0.2 0.3 69 -0.2 0.4 0.0 -0.3 1.3 0.2 0.2 0.2 1.2 0.1 0.4 1.3 0.3 0.4 0.2 0.3 70 -0.2 0.4 0.0 -0.3 . 0.2 0.2 0.2 0.2 0.1 -1.0 0.0 -0.9 0.4 0.2 0.3 71 -1.5 0.4 0.0 0.8 -0.6 0.2 . . 0.2 -1.2 -1.0 0.0 -2.1 -0.8 0.2 -0.8 72 -1.5 -0.7 -1.2 -0.3 -0.6 -0.8 -0.9 0.2 0.2 0.1 0.4 1.3 0.3 -0.8 -1.0 -0.8 73 . . -2.4 0.8 -2.4 -0.8 -2.0 -3.2 -2.8 0.1 -1.0 -2.6 -2.1 0.4 -1.0 0.3 74 -0.2 0.4 0.0 0.8 . 0.2 0.2 0.2 1.2 0.1 0.4 1.3 0.3 0.4 -1.0 0.3 75 -0.2 0.4 0.0 -0.3 0.4 0.2 1.3 1.3 . 0.1 0.4 0.0 . 0.4 1.4 1.3 76 -3.9 -2.8 -3.6 -3.5 -2.4 -2.8 -2.0 -3.2 -2.8 -1.2 -2.4 -1.3 -0.9 -3.2 -1.0 -0.8 77 1.0 0.4 1.2 0.8 0.4 1.2 1.3 1.3 1.2 0.1 0.4 1.3 0.3 0.4 . 0.3 78 -0.2 0.4 0.0 -1.4 1.3 1.2 1.3 1.3 -0.8 -1.2 0.4 1.3 0.3 0.4 0.2 -0.8 79 -0.2 0.4 0.0 -0.3 0.4 0.2 . 0.2 0.2 0.1 0.4 0.0 0.3 -0.8 0.2 1.3 80 -0.2 0.4 0.0 -0.3 0.4 . -0.9 -0.9 -0.8 0.1 0.4 . 0.3 0.4 0.2 0.3 81 1.0 1.4 1.2 0.8 1.3 1.2 1.3 1.3 1.2 0.1 0.4 0.0 1.6 0.4 0.2 1.3 82 1.0 1.4 1.2 0.8 1.3 -2.8 1.3 1.3 1.2 0.1 1.8 1.3 1.6 1.6 0.2 1.3 83 -0.2 0.4 1.2 0.8 1.3 1.2 1.3 0.2 0.2 -1.2 -1.0 1.3 0.3 0.4 0.2 1.3 84 -1.5 -1.8 -2.4 -0.3 -2.4 -1.8 -0.9 -0.9 -1.8 -1.2 -1.0 0.0 0.3 -2.0 -1.0 -1.9 85 1.0 0.4 0.0 -0.3 -0.6 0.2 0.2 0.2 0.2 1.4 0.4 1.3 0.3 0.4 1.4 0.3 86 -0.2 -0.7 -1.2 -1.4 0.4 0.2 0.2 0.2 0.2 0.1 0.4 0.0 0.3 0.4 1.4 0.3 87 1.0 1.4 1.2 -0.3 0.4 1.2 -0.9 1.3 1.2 1.4 1.8 1.3 0.3 0.4 1.4 0.3 88 -0.2 1.4 1.2 -0.3 1.3 1.2 1.3 1.3 1.2 1.4 1.8 1.3 1.6 1.6 1.4 1.3 89 -0.2 0.4 0.0 0.8 0.4 -0.8 0.2 0.2 0.2 0.1 -1.0 0.0 1.6 1.6 1.4 1.3 90 -0.2 0.4 0.0 0.8 1.3 0.2 0.2 0.2 0.2 0.1 0.4 1.3 0.3 1.6 1.4 0.3

Note: zcmmt = z-score of the ‘commitment’ item, zcomm = z-score of the ‘communication’ item, zaccn = z-score of the ‘accountability’ item, zldbx = z-score of the ‘leading by example’ item, zawrn = z-score of the ‘safety awareness’ item, zalgn = z-score of the ‘safety and productivity alignment’ item, zstnd = z-score of the ‘safety standards’ item, zinit = z-score of the ‘safety initiatives’ item, zintg = z-score of the ‘safety integration in business goals’ item, zprcp = z-score of the ‘shared perceptions’ item, zresp = z-score of the ‘safety responsibilities’ item, zsppt = z-score of the ‘supportive environment’ item, zinvm = z-score of the ‘workers’ involvement’ item, zrlsp = z-score of the ‘workers’ relationships’ item, zwkld = z-score of the ‘workload’ item, zprsr = z-score of the ‘work pressure’ item


224

case zcmmt zcomm zaccn zldbx zawrn zalgn zstnd zinit zintg zprcp zresp zsppt zinvm zrlsp zwkld zprsr 91 1.0 1.4 0.0 0.8 1.3 1.2 . 0.2 1.2 0.1 0.4 0.0 . -0.8 0.2 0.3 92 -0.2 -0.7 -1.2 -0.3 -1.5 0.2 -0.9 -0.9 0.2 0.1 0.4 1.3 0.3 0.4 -1.0 -1.9 93 -0.2 -0.7 -1.2 -0.3 -0.6 -0.8 -0.9 -0.9 -0.8 -1.2 -1.0 0.0 0.3 -0.8 0.2 0.3 94 1.0 0.4 1.2 0.8 -0.6 1.2 0.2 0.2 -0.8 0.1 0.4 1.3 -2.1 1.6 0.2 0.3 95 1.0 -0.7 0.0 0.8 -0.6 0.2 -0.9 -0.9 -0.8 0.1 -1.0 0.0 -0.9 0.4 -1.0 -0.8 96 1.0 -0.7 1.2 0.8 -0.6 -0.8 0.2 -0.9 0.2 1.4 -1.0 -1.3 -2.1 -0.8 -2.1 -1.9 97 -0.2 -0.7 0.0 -0.3 -0.6 0.2 -0.9 -0.9 -0.8 0.1 0.4 1.3 -0.9 0.4 0.2 -0.8 98 1.0 -0.7 -1.2 -1.4 -0.6 1.2 -2.0 1.3 -2.8 -1.2 -1.0 -1.3 0.3 -0.8 -3.3 -1.9 99 -0.2 -0.7 -1.2 -0.3 -0.6 -0.8 -0.9 -0.9 -0.8 -1.2 -1.0 0.0 0.3 -0.8 0.2 0.3

100 -0.2 -0.7 0.0 -0.3 0.4 -0.8 0.2 0.2 0.2 0.1 -1.0 0.0 0.3 -0.8 -1.0 0.3 101 -1.5 -0.7 0.0 -0.3 0.4 -0.8 -0.9 -0.9 0.2 0.1 0.4 -1.3 -0.9 0.4 -1.0 -1.9 102 -0.2 -0.7 0.0 0.8 0.4 0.2 0.2 0.2 -0.8 0.1 0.4 0.0 -0.9 1.6 0.2 0.3 103 1.0 -0.7 0.0 -1.4 0.4 0.2 0.2 0.2 1.2 0.1 -1.0 0.0 0.3 -0.8 1.4 -0.8 104 -0.2 0.4 0.0 -0.3 0.4 0.2 0.2 0.2 0.2 0.1 0.4 0.0 0.3 -0.8 -1.0 0.3 105 1.0 0.4 1.2 0.8 -0.6 0.2 0.2 0.2 1.2 0.1 0.4 -1.3 0.3 0.4 0.2 0.3 106 1.0 -0.7 0.0 0.8 -0.6 0.2 -0.9 -0.9 -0.8 0.1 -1.0 0.0 -0.9 0.4 -1.0 -0.8 107 1.0 1.4 0.0 0.8 1.3 1.2 1.3 0.2 1.2 1.4 1.8 1.3 1.6 0.4 1.4 0.3 108 -0.2 0.4 0.0 -0.3 0.4 -0.8 -0.9 0.2 0.2 -1.2 -1.0 0.0 0.3 0.4 -1.0 -0.8 109 1.0 0.4 1.2 0.8 -0.6 1.2 0.2 0.2 -0.8 0.1 0.4 1.3 -2.1 1.6 0.2 0.3 110 -0.2 0.4 0.0 0.8 1.3 0.2 0.2 0.2 0.2 0.1 0.4 1.3 0.3 1.6 0.2 1.3 111 1.0 1.4 1.2 0.8 1.3 1.2 1.3 1.3 1.2 1.4 0.4 0.0 0.3 0.4 0.2 0.3 112 -0.2 -0.7 -1.2 0.8 -2.4 -1.8 -0.9 -2.0 -1.8 -1.2 0.4 0.0 -0.9 -2.0 0.2 -0.8 113 -2.7 -1.8 -2.4 -2.4 -1.5 -2.8 -2.0 -2.0 -1.8 -2.6 -2.4 -3.8 -0.9 -0.8 -1.0 -1.9 114 -2.7 -1.8 -2.4 -2.4 -1.5 -1.8 -2.0 -2.0 -1.8 -2.6 -2.4 -2.6 -0.9 -2.0 -2.1 -1.9 115 -2.7 -1.8 -2.4 -2.4 -1.5 -1.8 -2.0 -2.0 -1.8 -2.6 -2.4 -2.6 -0.9 -2.0 -2.1 -1.9 Note: zcmmt = z-score of the ‘commitment’ item, zcomm = z-score of the ‘communication’ item, zaccn = z-score of the ‘accountability’ item, zldbx = z-score of the ‘leading by example’ item, zawrn = z-score of the ‘safety awareness’ item, zalgn = z-score of the ‘safety and productivity alignment’ item, zstnd = z-score of the ‘safety standards’ item, zinit = z-score of the ‘safety initiatives’ item, zintg = z-score of the ‘safety integration in business goals’ item, zprcp = z-score of the ‘shared perceptions’ item, zresp = z-score of the ‘safety responsibilities’ item, zsppt = z-score of the ‘supportive environment’ item, zinvm = z-score of the ‘workers’ involvement’ item, zrlsp = z-score of the ‘workers’ relationships’ item, zwkld = z-score of the ‘workload’ item, zprsr = z-score of the ‘work pressure’ item


225

case zcoop zfinc zresc zhmnr ztrng zrisk zfdbk znobm zhskp zdocu zbnmk zjstf zswbh zacci zcstm zimge zmrle zcost 1 -0.7 -0.8 0.1 -1.1 -0.9 -0.5 -0.9 -0.2 -0.9 -0.9 -0.3 -0.8 -1.0 -0.1 0.1 0.2 0.2 0.0 2 0.4 0.3 0.1 0.1 0.2 0.5 0.2 0.7 0.3 0.1 0.7 0.4 1.4 1.1 0.1 0.2 0.2 1.1 3 0.4 1.4 1.2 1.2 1.3 1.5 1.3 -0.2 0.3 1.1 1.6 1.7 1.4 1.1 1.3 1.4 1.5 1.1 4 -0.7 -0.8 0.1 0.1 1.3 0.5 0.2 -0.2 0.3 0.1 0.7 -0.8 0.2 -0.1 0.1 0.2 -1.0 0.0 5 0.4 1.4 0.1 1.2 0.2 1.5 0.2 -0.2 0.3 1.1 0.7 0.4 0.2 1.1 0.1 0.2 0.2 1.1 6 -0.7 -0.8 0.1 -1.1 -0.9 0.5 0.2 -0.2 -0.9 1.1 -0.3 -0.8 0.2 1.1 0.1 -0.9 -1.0 0.0 7 0.4 0.3 0.1 -1.1 0.2 -0.5 -0.9 -0.2 -0.9 0.1 -0.3 0.4 0.2 -0.1 0.1 -0.9 0.2 0.0 8 0.4 0.3 0.1 1.2 0.2 -0.5 0.2 -0.2 0.3 0.1 0.7 0.4 0.2 -0.1 0.1 0.2 0.2 0.0 9 1.5 1.4 1.2 1.2 1.3 0.5 0.2 -0.2 1.6 1.1 0.7 0.4 0.2 -0.1 0.1 0.2 0.2 1.1

10 -1.8 -0.8 -2.0 -2.2 0.2 -2.4 -2.0 1.6 -0.9 -3.0 -2.2 -2.0 -2.3 -1.4 -1.0 -3.2 -2.2 -3.3 11 -0.7 -0.8 0.1 0.1 0.2 0.5 -0.9 -0.2 0.3 0.1 0.7 -0.8 -1.0 -0.1 0.1 0.2 -1.0 0.0 12 -0.7 0.3 -0.9 0.1 0.2 0.5 0.2 0.7 -0.9 0.1 -0.3 0.4 0.2 -1.4 -1.0 0.2 0.2 0.0 13 0.4 1.4 0.1 0.1 0.2 0.5 0.2 -0.2 0.3 0.1 0.7 0.4 0.2 1.1 0.1 0.2 1.5 1.1 14 0.4 0.3 0.1 -1.1 -0.9 -0.5 -0.9 -0.2 0.3 -0.9 -0.3 -0.8 0.2 1.1 1.3 0.2 0.2 0.0 15 -0.7 0.3 0.1 0.1 -0.9 -0.5 -0.9 0.7 0.3 -0.9 -0.3 -0.8 0.2 1.1 1.3 0.2 0.2 0.0 16 0.4 0.3 0.1 -1.1 0.2 -0.5 0.2 -0.2 1.6 0.1 -0.3 -0.8 0.2 1.1 1.3 0.2 0.2 0.0 17 0.4 0.3 0.1 0.1 0.2 -0.5 -0.9 -0.2 0.3 0.1 -0.3 -0.8 -1.0 -0.1 -2.2 1.4 -3.4 1.1 18 0.4 1.4 1.2 0.1 -0.9 0.5 1.3 -0.2 0.3 1.1 0.7 0.4 1.4 -0.1 0.1 0.2 0.2 0.0 19 0.4 0.3 0.1 0.1 0.2 1.5 0.2 -0.2 1.6 1.1 0.7 0.4 1.4 1.1 1.3 0.2 0.2 0.0 20 -0.7 1.4 0.1 1.2 0.2 0.5 0.2 0.7 1.6 1.1 1.6 0.4 0.2 1.1 1.3 1.4 0.2 0.0 21 0.4 -1.8 0.1 0.1 -0.9 0.5 -0.9 -1.1 0.3 0.1 -0.3 0.4 0.2 -0.1 -1.0 -0.9 0.2 0.0 22 -0.7 -1.8 -2.0 -1.1 -0.9 -0.5 -0.9 -0.2 0.3 0.1 -1.3 0.4 0.2 -0.1 -1.0 -0.9 -1.0 0.0 23 -1.8 -0.8 -0.9 0.1 -0.9 -1.4 -2.0 -0.2 -0.9 -0.9 -1.3 -0.8 -1.0 -0.1 -1.0 -2.1 0.2 0.0 24 -0.7 0.3 -0.9 -1.1 -0.9 0.5 -0.9 0.7 -0.9 0.1 -0.3 -0.8 0.2 -1.4 0.1 -0.9 0.2 -1.1 25 -2.8 -1.8 -0.9 0.1 -2.0 -0.5 0.2 -0.2 0.3 0.1 -0.3 -0.8 -1.0 -0.1 0.1 -2.1 -2.2 0.0 26 0.4 0.3 1.2 -1.1 1.3 0.5 0.2 0.7 0.3 0.1 0.7 -0.8 -1.0 1.1 0.1 0.2 0.2 0.0 27 -0.7 -1.8 -0.9 0.1 1.3 0.5 0.2 -0.2 -0.9 -0.9 0.7 0.4 -1.0 -1.4 -1.0 0.2 0.2 0.0 28 -1.8 -1.8 -2.0 -1.1 -2.0 -1.4 -0.9 0.7 0.3 -3.0 -2.2 -0.8 -1.0 -0.1 -1.0 -0.9 -1.0 0.0 29 0.4 -0.8 -0.9 0.1 -0.9 0.5 -0.9 0.7 0.3 0.1 0.7 0.4 0.2 -0.1 0.1 0.2 0.2 0.0 30 1.5 1.4 1.2 1.2 1.3 1.5 1.3 -1.1 0.3 1.1 0.7 1.7 1.4 1.1 1.3 1.4 1.5 1.1

Note: zcoop = z-score of the ‘stakeholders’ cooperation’ item, zfinc = z-score of the ‘financial resources’ item, zresc = z-score of the ‘safety resources’ item, zhmnr = z-score of the ‘human resources’ item, ztrng = z-score of the ‘training’ item, zrisk = z-score of the ‘risk assessment’ item, zfdbk = z-score of the ‘feedback’ item, znobm = z-score of the ‘no-blame approach’ item, zhskp = z-score of the ‘housekeeping’ item, zdocu = z-score of the ‘safety documentation’ item, zbnmk = z-score of the ‘benchmarking system’ item, zjstf = z-score of the ‘job satisfaction’ item, zswbh = z-score of the ‘safe work behaviour’ item, zacci = z-score of the ‘number of accidents’ item, zcstm = z-score of the ‘customers’ expectations’ item, zimge = z-score of the ‘industrial image’ item, zmrle = z-score of the ‘workforce morale’ item, zcost = z-score of the ‘cost of accidents’ item

A S

yste

m D

ynam

ics

App

roac

h to

Con

stru

ctio

n Sa

fety

Cul

ture

22

6

case

zc

oop

zfin

c zr

esc

zhm

nr

ztrn

g zr

isk

zfdb

k zn

obm

zh

skp

zdoc

u zb

nmk

zjst

f zs

wbh

za

cci

zcst

m

zim

ge

zmrl

e zc

ost

31

0.4

-0.8

-0

.9

0.1

0.2

0.5

0.2

-1.1

0.

3 0.

1 -1

.3

0.4

0.2

-0.1

-1

.0

0.2

-2.2

0.

0 32

0.

4 0.

3 0.

1 0.

1 0.

2 -0

.5

0.2

0.7

0.3

0.1

0.7

0.4

0.2

-0.1

0.

1 0.

2 0.

2 1.

1 33

-1

.8

-0.8

-0

.9

-1.1

-0

.9

-1.4

-0

.9

-0.2

-0

.9

-0.9

-1

.3

-0.8

-1

.0

-1.4

-1

.0

0.2

0.2

0.0

34

0.4

-0.8

-0

.9

0.1

0.2

-0.5

0.

2 -0

.2

0.3

0.1

0.7

0.4

0.2

-1.4

-1

.0

0.2

0.2

0.0

35

-2.8

-1

.8

-0.9

-2

.2

-0.9

-1

.4

-0.9

-2

.0

0.3

-0.9

-1

.3

-3.3

-1

.0

1.1

0.1

0.2

0.2

0.0

36

1.5

1.4

1.2

1.2

1.3

0.5

1.3

1.6

1.6

1.1

-2.2

0.

4 1.

4 1.

1 1.

3 1.

4 0.

2 -3

.3

37

0.4

-0.8

1.

2 -1

.1

0.2

-1.4

-2

.0

0.7

0.3

-0.9

-1

.3

0.4

-2.3

-0

.1

1.3

0.2

-1.0

-1

.1

38

0.4

1.4

1.2

1.2

1.3

0.5

-0.9

0.

7 0.

3 0.

1 -0

.3

0.4

1.4

1.1

0.1

0.2

0.2

0.0

39

0.4

0.3

0.1

0.1

0.2

0.5

1.3

0.7

0.3

0.1

0.7

0.4

0.2

1.1

0.1

1.4

0.2

1.1

40

0.4

0.3

0.1

0.1

0.2

0.5

0.2

0.7

0.3

0.1

0.7

0.4

0.2

-0.1

0.

1 0.

2 0.

2 0.

0 41

0.

4 -0

.8

0.1

-1.1

-0

.9

-0.5

0.

2 0.

7 -0

.9

0.1

0.7

0.4

0.2

-0.1

-1

.0

0.2

-1.0

0.

0 42

0.

4 0.

3 0.

1 0.

1 0.

2 0.

5 0.

2 0.

7 0.

3 0.

1 0.

7 0.

4 0.

2 -0

.1

0.1

0.2

0.2

0.0

43

1.5

1.4

0.1

1.2

1.3

1.5

1.3

-0.2

0.

3 0.

1 -0

.3

0.4

1.4

1.1

1.3

0.2

1.5

0.0

44

1.5

0.3

0.1

1.2

0.2

1.5

1.3

0.7

1.6

1.1

1.6

0.4

1.4

1.1

1.3

1.4

1.5

1.1

45

-0.7

-0

.8

0.1

-1.1

0.

2 -0

.5

0.2

-0.2

0.

3 0.

1 0.

7 0.

4 0.

2 -0

.1

0.1

0.2

-1.0

-1

.1

46

0.4

1.4

1.2

0.1

1.3

1.5

1.3

0.7

0.3

1.1

0.7

0.4

0.2

-1.4

0.

1 1.

4 0.

2 -1

.1

47

0.4

0.3

1.2

1.2

1.3

-0.5

-0

.9

0.7

0.3

1.1

0.7

0.4

0.2

-0.1

1.

3 1.

4 1.

5 1.

1 48

0.

4 0.

3 0.

1 0.

1 0.

2 0.

5 0.

2 0.

7 -2

.2

0.1

0.7

0.4

0.2

-0.1

0.

1 0.

2 0.

2 0.

0 49

1.

5 0.

3 1.

2 1.

2 1.

3 1.

5 1.

3 0.

7 0.

3 1.

1 0.

7 0.

4 1.

4 1.

1 1.

3 1.

4 1.

5 1.

1 50

0.

4 0.

3 0.

1 0.

1 0.

2 -0

.5

0.2

-1.1

0.

3 0.

1 -0

.3

0.4

0.2

-0.1

0.

1 0.

2 0.

2 0.

0 51

0.

4 0.

3 0.

1 0.

1 0.

2 0.

5 0.

2 0.

7 0.

3 0.

1 0.

7 0.

4 0.

2 -0

.1

0.1

0.2

0.2

0.0

52

1.5

1.4

1.2

1.2

1.3

1.5

1.3

1.6

0.3

1.1

0.7

0.4

0.2

-0.1

0.

1 0.

2 0.

2 1.

1 53

0.

4 1.

4 1.

2 1.

2 1.

3 0.

5 1.

3 0.

7 1.

6 1.

1 0.

7 0.

4 1.

4 1.

1 1.

3 0.

2 1.

5 1.

1 54

0.

4 0.

3 0.

1 0.

1 0.

2 0.

5 0.

2 0.

7 0.

3 0.

1 0.

7 0.

4 0.

2 -0

.1

0.1

0.2

0.2

0.0

55

0.4

1.4

1.2

1.2

1.3

1.5

1.3

-0.2

0.

3 1.

1 1.

6 1.

7 1.

4 1.

1 1.

3 1.

4 1.

5 1.

1 56

0.

4 1.

4 1.

2 1.

2 0.

2 0.

5 1.

3 1.

6 0.

3 1.

1 0.

7 0.

4 0.

2 -0

.1

0.1

0.2

0.2

0.0

57

0.4

0.3

0.1

0.1

0.2

0.5

0.2

0.7

0.3

0.1

0.7

0.4

0.2

-0.1

0.

1 0.

2 0.

2 0.

0 58

0.

4 0.

3 0.

1 0.

1 0.

2 0.

5 0.

2 0.

7 0.

3 0.

1 0.

7 0.

4 0.

2 -0

.1

0.1

0.2

0.2

0.0

59

-0.7

0.

3 0.

1 0.

1 -0

.9

-0.5

0.

2 0.

7 0.

3 0.

1 0.

7 0.

4 0.

2 -0

.1

-1.0

0.

2 -1

.0

-1.1

60

-1

.8

-1.8

-0

.9

-2.2

1.

3 -1

.4

-0.9

-2

.0

-2.2

-2

.0

-0.3

-3

.3

-2.3

-1

.4

-2.2

-2

.1

-1.0

-1

.1

Not

e: z

coop

= z

-sco

re o

f th

e ‘s

take

hold

ers’

coo

pera

tion’

ite

m, z

finc

= z

-sco

re o

f th

e ‘f

inan

cial

res

ourc

es’

item

, zre

sc =

z-s

core

of

the

‘saf

ety

reso

urce

s’ i

tem

, zhm

nr =

z-s

core

of

the

‘hum

an r

esou

rces

’ ite

m, z

trng

= z

-sco

re o

f th

e ‘t

rain

ing’

item

, zri

sk =

z-s

core

of

the

‘ris

k as

sess

men

t’ it

em, z

fdbk

= z

-sco

re o

f th

e ‘f

eedb

ack’

item

, zno

bm =

z-s

core

of

the

‘no-

blam

e ap

proa

ch’

item

, zhs

kp =

z-s

core

of

the

‘hou

seke

epin

g’ it

em, z

docu

= z

-sco

re o

f th

e ‘s

afet

y do

cum

enta

tion’

item

, zbn

mk

= z-

scor

e of

the

‘ben

chm

arki

ng s

yste

m’

item

, zjs

tf =

z-sc

ore

of

the

‘job

satis

fact

ion’

item

, zsw

bh =

z-s

core

of

the

‘saf

e w

ork

beha

viou

r’ it

em, z

acci

= z

-sco

re o

f th

e ‘n

umbe

r of

acc

iden

ts’

item

, zcs

tm =

z-s

core

of

the

‘cus

tom

ers’

exp

ecta

tions

’ ite

m,

zim

ge =

z-s

core

of t

he ‘i

ndus

tria

l im

age’

item

, zm

rle

= z-

scor

e of

the

‘wor

kfor

ce m

oral

e’ it

em, z

cost

= z

-sco

re o

f the

‘cos

t of a

ccid

ents

’ ite

m

A S

yste

m D

ynam

ics

App

roac

h to

Con

stru

ctio

n Sa

fety

Cul

ture

22

7

case

zc

oop

zfin

c zr

esc

zhm

nr

ztrn

g zr

isk

zfdb

k zn

obm

zh

skp

zdoc

u zb

nmk

zjst

f zs

wbh

za

cci

zcst

m

zim

ge

zmrl

e zc

ost

61

-0.7

0.

3 0.

1 -1

.1

-2.0

-1

.4

0.2

-0.2

-2

.2

0.1

-1.3

-0

.8

0.2

-0.1

-1

.0

-0.9

-1

.0

0.0

62

-1.8

-0

.8

-0.9

-1

.1

0.2

0.5

-2.0

-0

.2

-0.9

-0

.9

-1.3

-0

.8

-1.0

-0

.1

0.1

0.2

0.2

-1.1

63

-0

.7

-0.8

0.

1 0.

1 -0

.9

0.5

-0.9

0.

7 -0

.9

-2.0

-0

.3

-0.8

-1

.0

-0.1

0.

1 0.

2 0.

2 0.

0 64

1.

5 0.

3 0.

1 0.

1 0.

2 -0

.5

0.2

-1.1

0.

3 0.

1 -0

.3

-0.8

0.

2 -0

.1

-1.0

0.

2 0.

2 0.

0 65

0.

4 0.

3 0.

1 0.

1 0.

2 0.

5 0.

2 0.

7 0.

3 0.

1 0.

7 0.

4 0.

2 -0

.1

0.1

0.2

0.2

0.0

66

1.5

1.4

1.2

1.2

1.3

1.5

1.3

1.6

1.6

1.1

-0.3

1.

7 1.

4 1.

1 1.

3 -0

.9

0.2

1.1

67

0.4

-0.8

0.

1 -1

.1

0.2

0.5

0.2

-0.2

0.

3 0.

1 0.

7 0.

4 0.

2 -0

.1

0.1

0.2

0.2

0.0

68

0.4

1.4

1.2

1.2

0.2

-0.5

0.

2 1.

6 1.

6 0.

1 -0

.3

0.4

0.2

-0.1

0.

1 0.

2 0.

2 1.

1 69

0.

4 0.

3 1.

2 0.

1 0.

2 0.

5 0.

2 0.

7 0.

3 0.

1 0.

7 0.

4 0.

2 -0

.1

0.1

0.2

0.2

0.0

70

0.4

0.3

0.1

0.1

0.2

0.5

0.2

-0.2

-0

.9

0.1

0.7

-0.8

0.

2 -0

.1

0.1

0.2

0.2

0.0

71

-0.7

-0

.8

-2.0

0.

1 -0

.9

-1.4

0.

2 -0

.2

-0.9

-2

.0

-0.3

-2

.0

-2.3

-1

.4

-2.2

0.

2 0.

2 -2

.2

72

0.4

0.3

0.1

0.1

. .

-0.9

-0

.2

. -0

.9

-0.3

.

-1.0

-1

.4

0.1

. -1

.0

0.0

73

1.5

. .

0.1

-0.9

0.

5 0.

2 1.

6 -0

.9

-3.0

-2

.2

-0.8

0.

2 1.

1 0.

1 0.

2 0.

2 1.

1 74

0.

4 0.

3 0.

1 0.

1 0.

2 0.

5 0.

2 -0

.2

1.6

0.1

0.7

0.4

0.2

1.1

1.3

1.4

1.5

1.1

75

. .

1.2

1.2

1.3

. 1.

3 .

1.6

1.1

1.6

. 0.

2 1.

1 0.

1 .

. 0.

0 76

-2

.8

-2.9

-3

.1

-3.3

-3

.1

-2.4

-3

.1

-2.0

.

-3.0

-2

.2

. -3

.5

-3.9

-3

.4

-3.2

-3

.4

-3.3

77

0.

4 0.

3 1.

2 1.

2 1.

3 0.

5 1.

3 0.

7 1.

6 1.

1 0.

7 0.

4 0.

2 1.

1 0.

1 .

1.5

1.1

78

. -0

.8

0.1

0.1

-0.9

-0

.5

0.2

1.6

0.3

1.1

0.7

-0.8

0.

2 -1

.4

. .

-1.0

0.

0 79

0.

4 -0

.8

-0.9

0.

1 -0

.9

0.5

0.2

0.7

-0.9

0.

1 0.

7 0.

4 0.

2 -0

.1

1.3

0.2

0.2

1.1

80

-0.7

0.

3 0.

1 0.

1 .

0.5

0.2

. .

. 0.

7 0.

4 0.

2 -0

.1

0.1

0.2

0.2

1.1

81

1.5

1.4

1.2

1.2

1.3

0.5

1.3

-0.2

0.

3 1.

1 -0

.3

0.4

1.4

1.1

1.3

1.4

1.5

1.1

82

1.5

1.4

1.2

1.2

1.3

1.5

1.3

1.6

1.6

1.1

0.7

1.7

1.4

1.1

1.3

1.4

1.5

1.1

83

-0.7

-0

.8

1.2

1.2

-0.9

0.

5 1.

3 0.

7 1.

6 1.

1 -0

.3

-0.8

0.

2 1.

1 -1

.0

-0.9

0.

2 1.

1 84

0.

4 -0

.8

1.2

-1.1

-0

.9

-0.5

0.

2 -2

.0

0.3

1.1

-1.3

-0

.8

0.2

-0.1

-1

.0

0.2

-1.0

1.

1 85

0.

4 0.

3 0.

1 0.

1 1.

3 0.

5 0.

2 -2

.0

0.3

1.1

0.7

0.4

0.2

1.1

1.3

0.2

-1.0

1.

1 86

-0

.7

0.3

0.1

0.1

-0.9

-0

.5

0.2

0.7

0.3

0.1

0.7

0.4

0.2

-0.1

0.

1 0.

2 0.

2 0.

0 87

1.

5 1.

4 1.

2 1.

2 1.

3 1.

5 1.

3 1.

6 0.

3 1.

1 1.

6 1.

7 0.

2 -0

.1

0.1

0.2

0.2

1.1

88

1.5

1.4

1.2

1.2

1.3

1.5

1.3

-2.0

1.

6 1.

1 1.

6 1.

7 1.

4 1.

1 1.

3 1.

4 1.

5 1.

1 89

0.

4 0.

3 1.

2 1.

2 1.

3 0.

5 1.

3 1.

6 1.

6 1.

1 0.

7 -0

.8

1.4

1.1

0.1

0.2

0.2

1.1

90

0.4

0.3

0.1

1.2

1.3

0.5

0.2

0.7

0.3

0.1

0.7

0.4

0.2

1.1

0.1

0.2

0.2

0.0

Not

e: z

coop

= z

-sco

re o

f th

e ‘s

take

hold

ers’

coo

pera

tion’

ite

m, z

finc

= z

-sco

re o

f th

e ‘f

inan

cial

res

ourc

es’

item

, zre

sc =

z-s

core

of

the

‘saf

ety

reso

urce

s’ i

tem

, zhm

nr =

z-s

core

of

the

‘hum

an r

esou

rces

’ ite

m, z

trng

= z

-sco

re o

f th

e ‘t

rain

ing’

item

, zri

sk =

z-s

core

of

the

‘ris

k as

sess

men

t’ it

em, z

fdbk

= z

-sco

re o

f th

e ‘f

eedb

ack’

item

, zno

bm =

z-s

core

of

the

‘no-

blam

e ap

proa

ch’

item

, zhs

kp =

z-s

core

of

the

‘hou

seke

epin

g’ it

em, z

docu

= z

-sco

re o

f th

e ‘s

afet

y do

cum

enta

tion’

item

, zbn

mk

= z-

scor

e of

the

‘ben

chm

arki

ng s

yste

m’

item

, zjs

tf =

z-sc

ore

of

the

‘job

satis

fact

ion’

item

, zsw

bh =

z-s

core

of

the

‘saf

e w

ork

beha

viou

r’ it

em, z

acci

= z

-sco

re o

f th

e ‘n

umbe

r of

acc

iden

ts’

item

, zcs

tm =

z-s

core

of

the

‘cus

tom

ers’

exp

ecta

tions

’ ite

m,

zim

ge =

z-s

core

of t

he ‘i

ndus

tria

l im

age’

item

, zm

rle

= z-

scor

e of

the

‘wor

kfor

ce m

oral

e’ it

em, z

cost

= z

-sco

re o

f the

‘cos

t of a

ccid

ents

’ ite

m


228

case zcoop zfinc zresc zhmnr ztrng zrisk zfdbk znobm zhskp zdocu zbnmk zjstf zswbh zacci zcstm zimge zmrle zcost 91 0.4 1.4 0.1 0.1 0.2 0.5 1.3 -1.1 0.3 0.1 0.7 0.4 0.2 1.1 0.1 0.2 0.2 0.0 92 0.4 -0.8 -0.9 0.1 0.2 -1.4 -0.9 -0.2 -0.9 0.1 -1.3 -0.8 0.2 -0.1 -1.0 -0.9 -1.0 -1.1 93 -0.7 -0.8 -2.0 0.1 -0.9 -1.4 -0.9 -1.1 -0.9 0.1 -0.3 -0.8 -1.0 -1.4 -1.0 -0.9 0.2 0.0 94 0.4 -1.8 -2.0 1.2 -0.9 -2.4 0.2 -2.0 0.3 0.1 -2.2 1.7 1.4 1.1 1.3 0.2 0.2 -2.2 95 -0.7 -0.8 -2.0 -1.1 -0.9 -1.4 -0.9 -1.1 -0.9 -0.9 -1.3 0.4 0.2 -0.1 0.1 -0.9 0.2 0.0 96 -1.8 -0.8 -0.9 -3.3 -0.9 -1.4 -2.0 -1.1 -2.2 -2.0 -1.3 -0.8 -1.0 -1.4 0.1 -2.1 -1.0 1.1 97 0.4 -0.8 -2.0 0.1 -0.9 -1.4 -2.0 -2.0 -0.9 0.1 -1.3 0.4 -1.0 -0.1 0.1 -0.9 0.2 -1.1 98 -1.8 -1.8 -0.9 -2.2 1.3 -1.4 -0.9 -2.0 -2.2 -2.0 -0.3 -3.3 -2.3 -1.4 -2.2 -2.1 -1.0 -1.1 99 -0.7 -0.8 -2.0 0.1 -0.9 -1.4 -0.9 -1.1 -0.9 0.1 -0.3 -0.8 -1.0 -1.4 -1.0 -0.9 0.2 0.0

100 -0.7 -0.8 0.1 -1.1 0.2 -0.5 0.2 -0.2 0.3 0.1 0.7 0.4 0.2 -0.1 0.1 0.2 -1.0 -1.1 101 -0.7 0.3 0.1 -1.1 -2.0 -1.4 0.2 -0.2 -2.2 0.1 -1.3 -0.8 0.2 -0.1 -1.0 -0.9 -1.0 0.0 102 0.4 -0.8 0.1 -1.1 -0.9 -0.5 0.2 0.7 -0.9 0.1 0.7 0.4 0.2 -0.1 -1.0 0.2 -1.0 0.0 103 0.4 0.3 1.2 1.2 1.3 -0.5 -0.9 0.7 0.3 1.1 0.7 0.4 0.2 -0.1 1.3 1.4 1.5 1.1 104 0.4 0.3 0.1 0.1 0.2 0.5 0.2 0.7 -2.2 0.1 0.7 0.4 0.2 -0.1 0.1 0.2 0.2 0.0 105 0.4 1.4 1.2 0.1 -0.9 0.5 1.3 -0.2 0.3 1.1 0.7 0.4 1.4 -0.1 0.1 0.2 0.2 0.0 106 -0.7 -0.8 -2.0 -1.1 -0.9 -1.4 -0.9 -1.1 -0.9 -0.9 -1.3 0.4 0.2 -0.1 0.1 -0.9 0.2 0.0 107 1.5 1.4 1.2 1.2 1.3 1.5 1.3 -1.1 0.3 1.1 0.7 1.7 1.4 1.1 1.3 1.4 1.5 1.1 108 -0.7 0.3 -0.9 0.1 0.2 0.5 0.2 0.7 -0.9 0.1 -0.3 0.4 0.2 -1.4 -1.0 0.2 0.2 0.0 109 0.4 -1.8 -2.0 1.2 -0.9 -2.4 0.2 -2.0 0.3 0.1 -2.2 1.7 1.4 1.1 1.3 0.2 0.2 -2.2 110 0.4 1.4 0.1 0.1 0.2 0.5 0.2 -0.2 0.3 0.1 0.7 0.4 0.2 1.1 0.1 0.2 1.5 1.1 111 0.4 0.3 0.1 1.2 0.2 -0.5 0.2 -0.2 0.3 0.1 0.7 0.4 0.2 -0.1 0.1 0.2 0.2 0.0 112 -1.8 -0.8 -0.9 0.1 -0.9 -1.4 -2.0 -0.2 -0.9 -0.9 -1.3 -0.8 -1.0 -0.1 -1.0 -2.1 0.2 0.0 113 -0.7 -0.8 -0.9 -1.1 -2.0 -1.4 -2.0 -2.0 -0.9 -0.9 -2.2 -2.0 -2.3 -2.6 -1.0 -2.1 -2.2 -2.2 114 -1.8 -0.8 0.1 -1.1 -2.0 0.5 -2.0 -1.1 -2.2 -2.0 -1.3 -2.0 -2.3 -2.6 -2.2 -2.1 -2.2 -2.2 115 -1.8 -0.8 0.1 -1.1 -2.0 0.5 -2.0 -1.1 -2.2 -2.0 -1.3 -2.0 -2.3 -2.6 -2.2 -2.1 -2.2 -2.2 Note: zcoop = z-score of the ‘stakeholders’ cooperation’ item, zfinc = z-score of the ‘financial resources’ item, zresc = z-score of the ‘safety resources’ item, zhmnr = z-score of the ‘human resources’ item, ztrng = z-score of the ‘training’ item, zrisk = z-score of the ‘risk assessment’ item, zfdbk = z-score of the ‘feedback’ item, znobm = z-score of the ‘no-blame approach’ item, zhskp = z-score of the ‘housekeeping’ item, zdocu = z-score of the ‘safety documentation’ item, zbnmk = z-score of the ‘benchmarking system’ item, zjstf = z-score of the ‘job satisfaction’ item, zswbh = z-score of the ‘safe work behaviour’ item, zacci = z-score of the ‘number of accidents’ item, zcstm = z-score of the ‘customers’ expectations’ item, zimge = z-score of the ‘industrial image’ item, zmrle = z-score of the ‘workforce morale’ item, zcost = z-score of the ‘cost of accidents’ item


229

AAppppeennddiixx 44

MMeeaassuurreemmeenntt MMooddeell RReessuullttss


230

NOTES FOR MODEL (Default model)

Computation of degrees of freedom (Default model)

Number of distinct sample moments: 324

Number of distinct parameters to be estimated: 87

Degrees of freedom (324 - 87): 237

Result (Default model)

Minimum was achieved

Chi-square = 390.173

Degrees of freedom = 237

Probability level = .000

Estimates (Group number 1 - Default model)

Scalar Estimates (Group number 1 - Default model)

Maximum Likelihood Estimates


231

Regression Weights: (Group number 1 - Default model)

Item Estimate S.E. C.R. P Label

Safety initiatives <--- Pol 1.00

Safety standards <--- Pol 0.93 0.09 10.41 ***

Number of accidents <--- Goals 1.03 0.17 6.21 ***

Industrial image <--- Goals 1.05 0.19 5.65 ***

Workforce morale <--- Goals 0.96 0.16 5.83 ***

Cost of accidents <--- Goals 1.00

Leading by example <--- Lds 0.82 0.12 6.84 ***

Accountability <--- Lds 1.00

Communication <--- Lds 1.07 0.11 9.37 ***

Commitment <--- Lds 0.87 0.10 8.51 ***

Safety awareness <--- Pol 1.05 0.11 9.95 ***

Safety and productivity alignment <--- Lds 0.88 0.12 7.15 ***

Workers’ relationships <--- Pol 0.64 0.09 6.85 ***

Human resources <--- Ppl 1.00

Stakeholders’ cooperation <--- Ppl 1.00 0.08 12.00 ***

Safety responsibilities <--- Ppl 0.44 0.06 7.08 ***

Financial resources <--- Prs 1.00

Safety resources <--- Prs 1.00 0.08 13.44 ***

Workers’ involvement <--- Prs 0.47 0.07 7.07 ***

Training <--- Prs 0.61 0.07 8.89 ***

Benchmarking system <--- Pro 0.85 0.11 7.82 ***

Safety integration in business goals <--- Pro 0.92 0.10 9.28 ***

Feedback <--- Pro 0.89 0.10 9.33 ***

Safety documentation <--- Pro 1.00


232

Standardized Regression Weights: (Group number 1 - Default model)

Item Estimate

Safety initiatives <--- Pol .868

Safety standards <--- Pol .787

Number of accidents <--- Goals .742

Industrial image <--- Goals .651

Workforce morale <--- Goals .679

Cost of accidents <--- Goals .603

Leading by example <--- Lds .613

Accountability <--- Lds .837

Communication <--- Lds .782

Commitment <--- Lds .727

Safety awareness <--- Pol .766

Safety and productivity alignment <--- Lds .635

Workers’ relationships <--- Pol .588

Human resources <--- Ppl .848

Stakeholders’ cooperation <--- Ppl .874

Safety responsibilities <--- Ppl .609

Financial resources <--- Prs .898

Safety resources <--- Prs .869

Workers’ involvement <--- Prs .589

Training <--- Prs .691

Benchmarking system <--- Pro .673

Safety integration in business goals <--- Pro .766

Feedback <--- Pro .769

Safety documentation <--- Pro .813


233

Intercepts: (Group number 1 - Default model)

Item Estimate S.E. C.R. P Label

Safety initiatives 3.82 0.08 48.52 ***

Safety standards 3.80 0.08 47.36 ***

Safety and productivity alignment 3.77 0.09 43.85 ***

Safety awareness 3.64 0.09 38.87 ***

Safety integration in business goals 3.78 0.09 44.30 ***

Safety responsibilities 3.76 0.06 58.98 ***

Workers’ involvement 3.73 0.07 50.82 ***

Workers’ relationships 3.72 0.08 49.84 ***

Human resources 3.72 0.11 35.38 ***

Safety resources 3.70 0.11 34.74 ***

Financial resources 3.55 0.10 34.52 ***

Stakeholders’ cooperation 3.50 0.10 34.31 ***

Feedback 3.78 0.08 45.76 ***

Training 3.79 0.08 46.54 ***

Number of accidents 4.10 0.07 57.75 ***

Industrial image 3.74 0.08 45.59 ***

Workforce morale 3.77 0.07 52.35 ***

Cost of accidents 3.97 0.09 46.86 ***

Leading by example 4.23 0.08 50.97 ***

Accountability 4.02 0.07 54.22 ***

Communication 3.65 0.09 43.11 ***

Commitment 4.16 0.07 56.41 ***

Benchmarking system 3.34 0.09 36.83 ***

Safety documentation 3.91 0.09 44.55 ***


234

Covariances: (Group number 1 - Default model)

Estimate S.E. C.R. P Label

Pol <--> Lds .412 .071 5.829 ***

Lds <--> Ppl .421 .085 4.977 ***

Lds <--> Prs .465 .087 5.333 ***

Pro <--> Lds .395 .073 5.391 ***

Goals <--> Lds .261 .059 4.410 ***

Pol <--> Ppl .490 .093 5.252 ***

Pol <--> Prs .565 .098 5.777 ***

Pro <--> Pol .533 .087 6.127 ***

Pol <--> Goals .316 .068 4.673 ***

Ppl <--> Prs .901 .141 6.377 ***

Pro <--> Ppl .603 .107 5.626 ***

Goals <--> Ppl .478 .097 4.927 ***

Pro <--> Prs .664 .111 5.980 ***

Goals <--> Prs .475 .096 4.949 ***

Pro <--> Goals .353 .075 4.707 ***

Correlations: (Group number 1 - Default model)

Estimate Pol <--> Lds .855 Lds <--> Ppl .668

Lds <--> Prs .713

Pro <--> Lds .783

Goals <--> Lds .724

Pol <--> Ppl .706

Pol <--> Prs .786

Pro <--> Pol .960

Pol <--> Goals .796

Ppl <--> Prs .960

Pro <--> Ppl .831

Goals <--> Ppl .922

Pro <--> Prs .884

Goals <--> Prs .884

Pro <--> Goals .850


235

Variances: (Group number 1 - Default model)


Pol .531 .093 5.707 ***

Goals .297 .088 3.369 ***

Lds .438 .083 5.292 ***

Ppl .906 .165 5.504 ***

Prs .971 .160 6.084 ***

Pro .581 .112 5.164 ***

e8 .174 .032 5.524 ***

e7 .279 .043 6.540 ***

e6 .502 .072 6.932 ***

e5 .412 .062 6.680 ***

e9 .343 .051 6.716 ***

e11 .291 .040 7.210 ***

e13 .401 .055 7.290 ***

e14 .416 .057 7.237 ***

e20 .355 .060 5.911 ***

e19 .316 .053 5.937 ***

e18 .233 .044 5.297 ***

e17 .280 .052 5.368 ***

e23 .318 .047 6.696 ***

e21 .395 .056 7.098 ***

e30 .258 .041 6.210 ***

e32 .442 .065 6.802 ***

e33 .319 .048 6.663 ***

e34 .519 .074 6.981 ***

e4 .489 .070 7.001 ***

e3 .188 .035 5.366 ***

e2 .318 .052 6.085 ***

e1 .291 .045 6.513 ***

e27 .513 .072 7.092 ***

e26 .299 .047 6.348 ***


236

Squared Multiple Correlations: (Group number 1 - Default model)

Estimate



Commitment .529

Communication .612

Accountability .700


Cost of accidents .364

Workforce morale .461

Industrial image .423

Number of accidents .551

Training .478

Feedback .592














237

MODIFICATION INDICES (Group number 1 - Default model)


M.I. Par Change

e26 <--> Lds 6.805 -.066

e4 <--> Goals 4.990 .056

e4 <--> e26 4.799 -.087

e33 <--> Lds 5.023 .058

e32 <--> e1 6.810 -.098

e30 <--> e27 8.256 -.108

e21 <--> e27 4.431 .094

e23 <--> Ppl 5.288 .064

e23 <--> e34 4.272 .085

e18 <--> Lds 4.741 .053

e19 <--> Lds 6.205 -.067

e19 <--> e33 6.141 -.086

e19 <--> e30 4.876 .070

e20 <--> e23 4.663 .079

e14 <--> e17 4.486 .078

e13 <--> e21 6.667 .101

e11 <--> e13 5.968 .081

e9 <--> Ppl 4.076 -.058

e9 <--> Lds 9.345 .081

e9 <--> e3 8.482 .084

e5 <--> e27 7.203 .125

e5 <--> e14 7.143 .111

e6 <--> e26 5.214 -.092

e6 <--> e3 7.911 -.096

e7 <--> Goals 4.205 .040

e7 <--> e30 6.123 .071

e7 <--> e19 4.521 .070

e7 <--> e5 4.177 -.072

e8 <--> e7 9.689 .076


238


M.I. Par Change


M.I. Par Change

Safety responsibilities <--- Pol 6.088 .181

Means: (Group number 1 - Default model)

M.I. Par Change


M.I. Par Change

MODEL FIT SUMMARY

CMIN

Model NPAR CMIN DF P CMIN/DF

Default model 87 390.173 237 .000 1.646

Saturated model 324 .000 0

Independence model 48 2022.641 276 .000 7.328

Baseline Comparisons

Model NFI Delta1

RFI rho1

IFI Delta2

TLI rho2 CFI

Default model .807 .775 .914 .898 .912

Saturated model 1.000 1.000 1.000

Independence model .000 .000 .000 .000 .000


239

Parsimony-Adjusted Measures

Model PRATIO PNFI PCFI

Default model .859 .693 .783

Saturated model .000 .000 .000

Independence model 1.000 .000 .000

NCP

Model NCP LO 90 HI 90

Default model 153.173 102.862 211.385


Independence model 1746.641 1607.495 1893.216

FMIN

Model FMIN F0 LO 90 HI 90

Default model 3.423 1.344 .902 1.854

Saturated model .000 .000 .000 .000

Independence model 17.742 15.321 14.101 16.607

RMSEA

Model RMSEA LO 90 HI 90 PCLOSE

Default model .075 .062 .088 .002

Independence model .236 .226 .245 .000

AIC

Model AIC BCC

Default model 564.173 613.050

Saturated model 648.000 830.022

Independence model 2118.641 2145.607


240

ECVI

Model ECVI LO 90 HI 90 MECVI

Default model 4.949 4.508 5.460 5.378

Saturated model 5.684 5.684 5.684 7.281


HOELTER

Model HOELTER .05

HOELTER .01

Default model 81 85

Independence model 18 19


241

AAppppeennddiixx 55

SSttrruuccttuurraall MMooddeell RReessuullttss


242

NOTES FOR MODEL (Default model)

Computation of degrees of freedom (Default model)

Number of distinct sample moments: 299

Number of distinct parameters to be estimated: 77

Degrees of freedom (299 - 77): 222

Result (Default model)

Minimum was achieved

Chi-square = 373.328

Degrees of freedom = 222

Probability level = .000

Estimates (Group number 1 - Default model)

Scalar Estimates (Group number 1 - Default model)

Maximum Likelihood Estimates


243



Ppl <--- Lds .910 .148 6.151 ***

Prs <--- Lds .236 .118 2.003 .045

Prs <--- Ppl .893 .100 8.969 ***

Pol <--- Lds .647 .124 5.220 ***

Pol <--- Prs .268 .076 3.535 ***

Pro <--- Ppl .356 .068 5.259 ***

Pro <--- Pol .619 .094 6.589 ***

Goals <--- Pro .662 .108 6.147 ***

Safety initiatives <--- Pol 1.000

Safety standards <--- Pol .929 .087 10.660 ***

Industrial image <--- Goals 1.045 .190 5.509 ***

Workforce morale <--- Goals .978 .170 5.763 ***

Cost of accidents <--- Goals 1.000

Accountability <--- Lds 1.000

Communication <--- Lds 1.071 .114 9.404 ***

Commitment <--- Lds .863 .101 8.512 ***

Safety awareness <--- Pol 1.030 .105 9.816 ***

Safety and productivity alignment <--- Lds .855 .123 6.924 ***

Workers’ relationships <--- Pol .635 .093 6.819 ***

Human resources <--- Ppl 1.000

Stakeholders’ cooperation <--- Ppl .996 .085 11.775 ***

Safety responsibilities <--- Ppl .442 .061 7.193 ***

Financial resources <--- Prs 1.000

Safety resources <--- Prs .996 .074 13.500 ***

Training <--- Prs .606 .068 8.889 ***

Safety integration in business goals <--- Pro .927 .105 8.863 ***

Feedback <--- Pro .928 .100 9.257 ***

Safety documentation <--- Pro 1.000

Number of accidents <--- Goals 1.041 .171 6.074 ***

Benchmarking system <--- Pro .838 .115 7.271 ***

Workers’ involvement <--- Prs .465 .066 7.045 ***


244

Standardized Regression Weights: (Group number 1 - Default model)

Estimate

Ppl <--- Lds .637

Prs <--- Lds .159

Prs <--- Ppl .858

Pol <--- Lds .586

Pol <--- Prs .361

Pro <--- Ppl .458

Pro <--- Pol .615

Goals <--- Pro .904

Safety initiatives <--- Pol .875

Safety standards <--- Pol .798

Industrial image <--- Goals .645

Workforce morale <--- Goals .687

Cost of accidents <--- Goals .599

Accountability <--- Lds .842

Communication <--- Lds .789

Commitment <--- Lds .731

Safety awareness <--- Pol .758

Safety and productivity alignment <--- Lds .621

Workers’ relationships <--- Pol .586

Human resources <--- Ppl .847

Stakeholders’ cooperation <--- Ppl .869

Safety responsibilities <--- Ppl .618

Financial resources <--- Prs .902

Safety resources <--- Prs .868

Training <--- Prs .690

Safety integration in business goals <--- Pro .752

Feedback <--- Pro .778

Safety documentation <--- Pro .788

Number of accidents <--- Goals .743

Benchmarking system <--- Pro .640

Workers’ involvement <--- Prs .587


245



Safety initiatives 3.817 .079 48.516 ***

Safety standards 3.800 .080 47.360 ***

Safety and productivity alignment 3.765 .086 43.850 ***

Safety awareness 3.635 .094 38.866 ***

Safety integration in business goals 3.783 .085 44.289 ***

Safety responsibilities 3.757 .064 58.980 ***

Workers’ relationships 3.722 .075 49.843 ***

Human resources 3.722 .105 35.377 ***

Safety resources 3.696 .106 34.742 ***

Financial resources 3.548 .103 34.519 ***

Stakeholders’ cooperation 3.504 .102 34.312 ***

Feedback 3.783 .083 45.747 ***

Training 3.791 .081 46.543 ***

Industrial image 3.739 .082 45.584 ***

Workforce morale 3.774 .072 52.342 ***

Cost of accidents 3.965 .085 46.854 ***

Accountability 4.017 .074 54.223 ***

Communication 3.652 .085 43.111 ***

Commitment 4.157 .074 56.408 ***

Safety documentation 3.913 .088 44.536 ***

Number of accidents 4.096 .071 57.731 ***

Benchmarking system 3.339 .091 36.819 ***

Workers’ involvement 3.730 .073 50.819 ***


246



Lds .443 .084 5.305 ***

ee2 .538 .112 4.801 ***

ee3 .063 .040 1.561 .118

ee1 .123 .033 3.740 ***

ee4 .003 .017 .205 .838

ee5 .053 .025 2.116 .034

e8 .165 .031 5.340 ***

e7 .267 .042 6.399 ***

e6 .517 .074 6.935 ***

e5 .424 .064 6.667 ***

e9 .361 .051 7.025 ***

e11 .286 .040 7.171 ***

e14 .417 .058 7.214 ***

e20 .357 .060 5.906 ***

e19 .318 .053 5.943 ***

e18 .224 .043 5.186 ***

e17 .291 .053 5.514 ***

e23 .308 .045 6.917 ***

e21 .396 .056 7.093 ***

e32 .448 .067 6.725 ***

e33 .313 .048 6.495 ***

e34 .524 .076 6.912 ***

e3 .182 .036 5.094 ***

e2 .309 .052 5.895 ***

e1 .288 .045 6.410 ***

e26 .333 .049 6.863 ***

e30 .257 .042 6.046 ***

e27 .554 .076 7.288 ***

e13 .403 .055 7.288 ***


247

Squared Multiple Correlations: (Group number 1 - Default model)

Estimate

Ppl .406

Prs .936

Pol .773

Pro .994

Goals .818



Number of accidents .552


Commitment .534

Communication .622

Accountability .709

Cost of accidents .358

Workforce morale .473

Industrial image .416

Training .476

Feedback .605













248

MODIFICATION INDICES (Group number 1 - Default model)


M.I. Par Change ee5 <--> ee2 4.439 .063 ee5 <--> ee1 4.034 -.034

e13 <--> ee1 9.077 .085

e27 <--> ee1 6.662 .085

e27 <--> e13 4.760 .100

e30 <--> e13 4.339 -.069

e30 <--> e27 11.985 -.135

e26 <--> ee1 4.469 .055

e26 <--> e13 5.879 .088

e26 <--> e27 4.941 .094

e1 <--> e30 4.696 .064

e34 <--> ee2 5.979 .142

e33 <--> e2 4.124 .070

e32 <--> e1 6.572 -.097

e21 <--> e13 6.840 .102

e21 <--> e27 4.781 .100

e23 <--> e34 4.596 .087

e17 <--> ee1 4.552 -.057

e17 <--> e13 4.340 -.077

e17 <--> e32 4.410 .085

e18 <--> ee1 8.593 -.070

e18 <--> e26 4.963 -.069

e19 <--> e30 4.789 .071

e19 <--> e33 6.775 -.090

e20 <--> e13 4.068 -.081

e20 <--> e27 4.968 -.104

e14 <--> ee3 4.070 -.064

e14 <--> e17 4.509 .081

e11 <--> ee1 13.404 .087

e11 <--> e13 5.710 .079

e9 <--> ee2 6.823 -.125

e9 <--> e3 7.679 .081

e9 <--> e17 4.922 -.079


249

Covariances: (Group number 1 - Default model) (Cont.)

M.I. Par Change e5 <--> e27 8.034 .138 e5 <--> e14 7.668 .117

e5 <--> e9 4.601 .085

e6 <--> e26 6.102 -.103

e6 <--> e2 4.405 .090

e6 <--> e3 6.688 -.089

e7 <--> e30 5.637 .068

e7 <--> e19 4.832 .072

e7 <--> e5 4.567 -.076

e8 <--> ee5 4.138 -.035

e8 <--> e7 7.310 .065


M.I. Par Change


M.I. Par Change

Safety responsibilities <--- Pol 5.437 .169

Means: (Group number 1 - Default model)

M.I. Par Change


M.I. Par Change


250

MODEL FIT SUMMARY

CMIN

Model NPAR CMIN DF P CMIN/DF

Default model 77 373.328 222 .000 1.682

Saturated model 299 .000 0

Independence model 46 1944.668 253 .000 7.686

Baseline Comparisons

Model NFI Delta1

RFI rho1

IFI Delta2

TLI rho2 CFI

Default model .808 .781 .912 .898 .911

Saturated model 1.000 1.000 1.000

Independence model .000 .000 .000 .000 .000

Parsimony-Adjusted Measures

Model PRATIO PNFI PCFI

Default model .877 .709 .799


Independence model 1.000 .000 .000

NCP

Model NCP LO 90 HI 90

Default model 151.328 101.949 208.593


Independence model 1691.668 1555.060 1835.701

FMIN

Model FMIN F0 LO 90 HI 90

Default model 3.275 1.327 .894 1.830

Saturated model .000 .000 .000 .000



251

RMSEA

Model RMSEA LO 90 HI 90 PCLOSE

Default model .077 .063 .091 .001

Independence model .242 .232 .252 .000

AIC

Model AIC BCC BIC CAIC

Default model 527.328 568.395

Saturated model 598.000 757.467

Independence model 2036.668 2061.202

ECVI

Model ECVI LO 90 HI 90 MECVI

Default model 4.626 4.193 5.128 4.986

Saturated model 5.246 5.246 5.246 6.644


HOELTER

Model HOELTER .05

HOELTER .01

Default model 79 84

Independence model 18 19


252

�

AAppppeennddiixx 66

SSDD EEqquuaattiioonnss ooff BBaassee RRuunn SSiimmuullaattiioonn


253

GOALS(t) = GOALS(t - dt) + (rgoals)*dt

INIT GOALS = 68

Inflows:

rgoals = used_pro*DF_goals_pro

LEADERSHIP(t) = LEADERSHIP(t - dt) + (rlds)*dt

INIT LEADERSHIP = 0

Inflows:

rlds = ((used_lds + ggoals)*rldsf) + (glds*plds)

PARTNERSHIPS_&_RESOURCES(t) = PARTNERSHIPS_&_RESOURCES(t - dt)

+ (rprs)*dt

INIT PARTNERSHIPS_&_RESOURCES = 0

Inflows:

rprs = (used_lds*DF_prs_lds) + (used_ppl*DF_prs_ppl) + (gprs*pprs)

PEOPLE(t) = PEOPLE(t - dt) + (rppl)*dt

INIT PEOPLE = 0

Inflows:

rppl = (used_lds*DF_ppl_lds) + (gppl*pppl)

POLICY_&_STRATEGY(t) = POLICY_&_STRATEGY(t - dt) + (rpol)*dt

INIT POLICY_&_STRATEGY = 0

Inflows:

rpol = (used_lds*DF_pol_lds) + (used_prs*DF_pol_prs) + (gpol*ppol)


254

PROCESSES(t) = PROCESSES(t - dt) + (rpro)*dt

INIT PROCESSES = 0

Inflows:

rpro = (used_ppl*DF_pro_ppl) + (used_pol*DF_pro_pol) + (gpro*ppro)

ENABLERS = used_lds + used_pol + used_ppl + used_prs + used_pro

CSC_INDEX = ENABLERS + used_goals

Co_lds_pol = 0.59

Co_lds_ppl = 0.64

Co_lds_prs = 0.16

Co_pol_pro = 0.62

Co_ppl_pro = 0.46

Co_ppl_prs = 0.86

Co_pro_goals = 0.90

Co_prs_pol = 0.36

DF_goals_pro = (ggoals*Co_pro_goals)/100

DF_pol_lds = (gpol*Co_lds_pol)/100

DF_pol_prs = (gpol*Co_prs_pol)/100

DF_ppl_lds = (gppl*Co_lds_ppl)/100

DF_pro_pol = (gpro*Co_pol_pro)/100

DF_pro_ppl = (gpro*Co_ppl_pro)/100

DF_prs_lds = (gprs*Co_lds_prs)/100

DF_prs_ppl = (gprs*Co_ppl_prs)/100


255

goals = 500

dlds = 100

doll = 80

dppl = 90

dero = 140

dors = 90

ggoals = IF(CSC_INDEX <= 200)THEN(100 - used_goals)ELSE(IF(CSC_INDEX <=

400)THEN(200 - used_goals)ELSE(IF(CSC_INDEX <= 600)THEN(300 - used_goals)

ELSE(IF(CSC_INDEX <= 800)THEN(400 - used_goals)ELSE(500 - used_goals))))

glds = dlds - used_lds

gpol = doll - used_pol

gppl = dppl - used_ppl

gpro = dpro - used_pro

gprs = dprs - used_prs

plds = 0

ppol = 0

pppl = 0

ppro = 0

pprs = 0

used_goals = MIN(GOALS,dgoals)

used_lds = MIN(LEADERSHIP,dlds)

used_pol = MIN(POLICY_&_STRATEGY,dpol)

used_ppl = MIN(PEOPLE,dppl)

used_pro = MIN(PROCESSES,dpro)


256

used_prs = MIN(PARTNERSHIPS_&_RESOURCES,dprs)

rldsf = 0.08

�


257

AAppppeennddiixx 77

LLiinneeaarr RReeggrreessssiioonn RReessuullttss


258

REGRESSION

Descriptive Statistics

Mean Std. Deviation N

Enablers 377.3158 68.14106 114

Goals 392.7632 68.05806 114

Correlations

Enablers Goals

Pearson Correlation Enablers

Goals

1.000

.787

.787

1.000

Sig. (1-tailed) Enablers

Goals

.

.000

.000

.

N Enablers

Goals

114

114

114

114

Variables Entered/Removedb

Model Variables Entered Variables Removed Method

1 Goalsa Enter

a. All requested variables entered. b. Dependent variable: Enablers

Model Summary

Model R R Square Adjusted R Square Std. Error of the Estimate

1 .787 a .619 .615 42.25812

a. Predictors: (Constant), Goals

ANOVAb

Model Sum of Squares df Mean Square F Sig.

1 Regression

Residual

Total

324678.3

200003.8

524682.1

1

112

113

324678.322

1785.748

181.816 .000 a

a. Predictors: (Constant), Goals. b. Dependent variable: Enablers


259

Coefficients a

Model Unstandardized Coefficients Standardized Coefficients t Sig.

B Std. Error Beta

1 (Constant)

Goals

67.974

.788

23.280

.058

.787

2.920

13.484

.004

.000

a. Dependent variable: Enablers


260

AAppppeennddiixx 88

SSeennssiittiivviittyy AAnnaallyysseess RReessuullttss wwhheenn tthhee IInniittiiaall VVaalluuee ooff

eeaacchh EEnnaabblleerr iiss CChhaannggeedd


261

Sensitivity results of the ‘used_pol’ value when the initial values of Pol are changed

Page 3

1.00 5.00 9.00 13.00 17.00

Years

1:

1:

1:

0

40

80

used pol score: 1 - 2 - 3 - 4 -

1

1

1

1

2

2

2

2

3

3

33

4

4

4 4

Note: Numbers 1, 2, 3 and 4 represent the initial value of Policy and Strategy of zero, 20, 40, and 60 points, respectively

Sensitivity results of the CSC index when the initial values of Pol are changed

Page 8

1.00 5.00 9.00 13.00 17.00

Years

1:

1:

1:

0

500

1000


1

1

1

1

2

2

2

2

3

3

3

3

4

4

44

Note: Numbers 1, 2, 3 and 4 represent the initial value of Policy and Strategy of zero, 20, 40, and 60 points, respectively


262

Sensitivity results of the ‘used_ppl’ value when the initial values of Ppl are changed

Page 4

1.00 5.00 9.00 13.00 17.00

Years

1:

1:

1:

0

45

90

used ppl score: 1 - 2 - 3 - 4 -

1

1

1

1

2

2

2

2

3

3

33

4

4

4 4

Note: Numbers 1, 2, 3 and 4 represent the initial value of People of zero, 22.5, 45, and 67.5 points, respectively

Sensitivity results of the CSC index when the initial values of Ppl are changed

Page 8

1.00 5.00 9.00 13.00 17.00

Years

1:

1:

1:

0

500

1000


1

1

1

1

2

2

2

2

3

3

33

4

44

4

Note: Numbers 1, 2, 3 and 4 represent the initial value of People of zero, 22.5, 45, and 67.5 points, respectively


263

Sensitivity results of the ‘used_prs’ value when the initial values of Prs are changed

Page 5

1.00 5.00 9.00 13.00 17.00

Years

1:

1:

1:

0

45

90

used prs score: 1 - 2 - 3 - 4 -

1

1

1

1

2

2

2

2

3

3

3 3

4

4

4 4

Note: Numbers 1, 2, 3 and 4 represent the initial value of Partnerships and Resources of zero, 22.5, 45, and 67.5 points, respectively

Sensitivity results of the CSC index when the initial values of Prs are changed

Page 8

1.00 5.00 9.00 13.00 17.00

Years

1:

1:

1:

0

500

1000


1

1

1

1

2

2

2

2

3

3

33

4

4

44

Note: Numbers 1, 2, 3 and 4 represent the initial value of Partnerships and Resources of zero, 22.5, 45, and 67.5 points, respectively


264

Sensitivity results of the ‘used_pro’ value when the initial values of Pro are

changed

Page 7

1.00 5.00 9.00 13.00 17.00

Years

1:

1:

1:

0

70

140

used pro score: 1 - 2 - 3 - 4 -

1

1

1

1

2

2

2

2

3

3

33

4

4

4 4

Note: Numbers 1, 2, 3 and 4 represent the initial value of Processes of zero, 35, 70, and 105 points, respectively

Sensitivity results of the CSC index when the initial values of Pro are changed

Page 8

1.00 5.00 9.00 13.00 17.00

Years

1:

1:

1:

0

500

1000


1

1

1

1

2

2

2

2

3

3

3

3

4

4

4

4

Note: Numbers 1, 2, 3 and 4 represent the initial value of Processes of zero, 35, 70, and 105 points, respectively


265

AAppppeennddiixx 99

SSeennssiittiivviittyy AAnnaallyysseess RReessuullttss wwhheenn tthhee EExxttrraa EEffffoorrtt GGiivveenn ttoo

IImmpprroovvee eeaacchh EEnnaabblleerr iiss CChhaannggeedd


266

Sensitivity results of the ‘used_pol’ value when the ‘ppol’ values are changed

Page 3

1.00 5.00 9.00 13.00 17.00

Years

1:

1:

1:

0

40

80

used pol score: 1 - 2 - 3 - 4 -

1

1

1

1

2

2

22

3

3

3 3

4

44 4

Note: Numbers 1, 2, 3 and 4 represent ‘ppol’ of zero, 0.1, 0.2, and 0.3, respectively

Sensitivity results of the CSC index when the ‘ppol’ values are changed

Page 8

1.00 5.00 9.00 13.00 17.00

Years

1:

1:

1:

0

500

1000


1

1

1

1

2

2

2

2

3

3

3

3

4

4

44

Note: Numbers 1, 2, 3 and 4 represent ‘ppol’ of zero, 0.1, 0.2, and 0.3, respectively


267

Sensitivity results of the ‘used_ppl’ value when the ‘pppl’ values are changed

Page 4

1.00 5.00 9.00 13.00 17.00

Years

1:

1:

1:

0

45

90

used ppl score: 1 - 2 - 3 - 4 -

1

1

1

1

2

2

2

2

3

3

3 3

4

44 4

Note: Numbers 1, 2, 3 and 4 represent ‘pppl’ of zero, 0.1, 0.2, and 0.3, respectively

Sensitivity results of CSC index when ‘pppl’ values are changed

Page 8

1.00 5.00 9.00 13.00 17.00

Years

1:

1:

1:

0

500

1000


1

1

1

1

2

2

2

2

3

3

33

4

4

44

Note: Numbers 1, 2, 3 and 4 represent ‘pppl’ of zero, 0.1, 0.2, and 0.3, respectively


268

Sensitivity results of the ‘used_prs’ value when the ‘pprs’ values are changed

Page 5

1.00 5.00 9.00 13.00 17.00

Years

1:

1:

1:

0

45

90

used prs score: 1 - 2 - 3 - 4 -

1

1

1

1

2

2

22

3

3

3 3

4

44 4

Note: Numbers 1, 2, 3 and 4 represent ‘pprs’ of zero, 0.1, 0.2, and 0.3, respectively

Sensitivity results of the CSC index when the ‘pprs’ values are changed

Page 8

1.00 5.00 9.00 13.00 17.00

Years

1:

1:

1:

0

500

1000


1

1

1

1

2

2

2

2

3

3

3

3

4

4

44

Note: Numbers 1, 2, 3 and 4 represent ‘pprs’ of zero, 0.1, 0.2, and 0.3, respectively


269

Sensitivity results of the ‘used_pro’ value when the ‘ppro’ values are changed

Page 7

1.00 5.00 9.00 13.00 17.00

Years

1:

1:

1:

0

70

140

used pro score: 1 - 2 - 3 - 4 -

1

1

1

1

2

2

22

3

3

3 3

4

44 4

Note: Numbers 1, 2, 3 and 4 represent ‘ppro’ of zero, 0.1, 0.2, and 0.3, respectively

Sensitivity results of the CSC index when the ‘ppro’ values are changed

Page 8

1.00 5.00 9.00 13.00 17.00

Years

1:

1:

1:

0

500

1000


1

1

1

1

2

2

2

2

3

3

3

3

4

4

4

4

Note: Numbers 1, 2, 3 and 4 represent ‘ppro’ of zero, 0.1, 0.2, and 0.3, respectively


270

AAppppeennddiixx 1100

SSDD EEqquuaattiioonnss ooff OOrrggaanniizzaattiioonn ‘‘AA’’


271


INIT GOALS = 129

Inflows:



INIT LEADERSHIP = 20

Inflows:



+ (rprs)*dt

INIT PARTNERSHIPS_&_RESOURCES = 18

Inflows:



INIT PEOPLE = 43.2

Inflows:



INIT POLICY_&_STRATEGY = 19.2

Inflows:



272


INIT PROCESSES = 37.3

Inflows:




Co_lds_pol = 0.59

Co_lds_ppl = 0.64

Co_lds_prs = 0.16

Co_pol_pro = 0.62

Co_ppl_pro = 0.46

Co_ppl_prs = 0.86

Co_pro_goals = 0.90

Co_prs_pol = 0.36










273

dgoals = 500

dlds = 100

dpol = 80

dppl = 90

dpro = 140

dprs = 90





gpol = dpol - used_pol




plds = 0

ppol = 0

pppl = 0

ppro = 0

pprs = 0







274


rldsf = 0.08


275


SSDD EEqquuaattiioonnss ooff OOrrggaanniizzaattiioonn ‘‘BB’’


276


INIT GOALS = 200

Inflows:



INIT LEADERSHIP = 85

Inflows:



+ (rprs)*dt

INIT PARTNERSHIPS_&_RESOURCES = 40.5

Inflows:



INIT PEOPLE = 43.2

Inflows:




Inflows:



277


INIT PROCESSES = 56

Inflows:




Co_lds_pol = 0.59

Co_lds_ppl = 0.64

Co_lds_prs = 0.16

Co_pol_pro = 0.62

Co_ppl_pro = 0.46

Co_ppl_prs = 0.86

Co_pro_goals = 0.90

Co_prs_pol = 0.36










278

dgoals = 500

dlds = 100

dpol = 80

dppl = 90

dpro = 140

dprs = 90









plds = 0

ppol = 0

pppl = 0

ppro = 0

pprs = 0







279


rldsf = 0.08


280


SSDD EEqquuaattiioonnss ooff tthhee CCyycclliiccaall SSttyyllee ooff SSaaffeettyy MMaannaaggeemmeenntt


281

GOALS(t) = GOALS(t - dt) + (rgoals - rgoals2)*dt

INIT GOALS = 381.64

Inflows:


Outflows:

rgoals2 = IF (rlds2 > 0) OR (rlds3 > 0) THEN (rgoals + (rgoals*Co_lds_pol*

Co_pol_pro*Co_pro_goals) + (rgoals*Co_lds_ppl*Co_ppl_pro*Co_pro_goals)) ELSE

(0)

LEADERSHIP(t) = LEADERSHIP(t - dt) + (rlds - rlds2 - rlds3)*dt

INIT LEADERSHIP = 60.37

Inflows:

rlds = IF (CSC_INDEX > dCSC_INDEX) OR ((800 < CSC_INDEX < dCSC_INDEX)

AND (slope < 0)) THEN (0) ELSE ((used_lds + ggoals)*rldsf)

Outflows:

rlds2 = IF (CSC_INDEX < 800) OR ((800 < CSC_INDEX < dCSC_INDEX) AND

(slope > 0)) THEN (0) ELSE ((glds + ggoals)*rldsf)

rlds3 = IF (CSC_INDEX > dCSC_INDEX) AND (rlds = 0) AND (rlds2 = 0) THEN

((glds + ggoals)*rldsf) ELSE (0)


+ (rprs - rprs2)*dt

INIT PARTNERSHIPS_&_RESOURCES = 82.23

Inflows:

rprs = (used_lds*DF_prs_lds) + (used_ppl*DF_prs_ppl)

Outflows:

rprs2 = IF (rlds2 > 0) OR (rlds3 > 0) THEN (rprs + (rprs*Co_lds_prs) +

(rprs*Co_lds_ppl *Co_ppl_prs)) ELSE (0)


282

PEOPLE(t) = PEOPLE(t - dt) + (rppl - rppl2)*dt

INIT PEOPLE = 68.07

Inflows:

rppl = (used_lds*DF_ppl_lds)

Outflows:

rppl2 = IF (rlds2 > 0) OR (rlds3 > 0) THEN (rppl + (rppl*Co_lds_ppl)) ELSE (0)

POLICY_&_STRATEGY(t) = POLICY_&_STRATEGY(t - dt) + (rpol - rpol2)*dt


Inflows:

rpol = (used_lds*DF_pol_lds) + (used_prs*DF_pol_prs)

Outflows:

rpol2 = IF (rlds2 > 0) OR (rlds3 > 0) THEN (rpol + (rpol*Co_lds_pol) +

(rpol*Co_lds_prs *Co_prs_pol)) ELSE (0)

PROCESSES(t) = PROCESSES(t - dt) + (rpro - rpro2)*dt

INIT PROCESSES = 132.44

Inflows:

rpro = (used_ppl*DF_pro_ppl) + (used_pol*DF_pro_pol)

Outflows:

rpro2 = IF (rlds2 > 0) OR (rlds3 > 0) THEN (rpro + (rpro*Co_lds_pol*Co_pol_pro) +

(rpro*Co_lds_ppl*Co_ppl_pro)) ELSE (0)

desired_CSC_INDEX(t) = desired_CSC_INDEX(t - dt) + (CSC_flow)*dt

INIT desired_CSC_INDEX = 950


283

Inflows:

CSC_flow = PULSE (10,(slope < 0) AND (CSC_INDEX >= desired_CSC_INDEX),

(CSC_INDEX >= desired_CSC_INDEX))



Co_lds_pol = 0.59

Co_lds_ppl = 0.64

Co_lds_prs = 0.16

Co_pol_pro = 0.62

Co_ppl_pro = 0.46

Co_ppl_prs = 0.86

Co_pro_goals = 0.90

Co_prs_pol = 0.36










284

dgoals = 500

dlds = 100

dpol = 80

dppl = 90

dpro = 140

dprs = 90









dCSC_INDEX = MIN(desired_CSC_INDEX,1000)







rldsf = 0.08

slope = DERIVN(CSC_INDEX,1)


285

RREEFFEERREENNCCEESS

ACSNI, 1993. Study group on human factors. Third report: organizing for safety.

Advisory Committee on the Safety of Nuclear Installations, London: HMSO.

Ahmed, A.M., Yang, J.B. and Dale, B.G., 2003. Self-assessment methodology: the

route to business excellence. The quality management journal, 10(1), 43-57.

Ahmed, S.M., Tang, P., Azhar, S. and Ahmad, I., 2004. An evaluation of safety

measures in the Hong Kong construction industry based on total quality management

principals [online]. Available from: http://www.fiu.edu/~sazha002/ research/tqmpaper.

pdf [Accessed 9 September 2004].

Aksorn, T., and Hadikusumo, B.H.W., 2004. Investigating unsafe acts and decision to

err factors of Thai construction workers. CIB international symposium on globalisation

and construction, meeting the challenges, reaping the benefits, 17-19, November 2004,

Thailand.

Aksorn, T. and Hadikusumo, B.H.W., 2006. Critical success factors of safety programs

implementation in Thai construction projects. In: D. Fang, R.M. Choudhry and J.W.

Hinze, eds. Proceedings of the CIB W99 2006 international conference on global unity

for safety and health in construction, 28-30 June 2006, Beijing, China. Beijing:

Tsinghua University Press, 328-336.

Aksorn, T. and Hadikusumo, B.H.W., 2007. Critical success factors influencing safety

program performance in Thai construction projects. Safety science, In press, Corrected

proof, Available online 6 August 2007.


286

Ali, T.H., 2006. Influence of national culture on construction safety climate in Pakistan.

Thesis (PhD), School of Engineering, Griffith University, Australia.

Arboleda, A., Morrow, P.C., Crum, M.R., and Shelley II, M.C., 2003. Management

practices as antecedents of safety culture within the trucking industry: similarities and

differences by hierarchical level. Journal of safety research, 34, 189-197.

Australian Bureau of Statistics, 1993. Australian and New Zealand standard industrial

classification 1993 ANZSIC, Australia, Cat No. 1292.0.

Barnes, J., Gaskin, J., Park, T., Saengdaeng, S. and Wilson, A., 2005. System dynamics

and its use in an organization [online]. Available from: http://www.mngt.waikato.ac.nz/

depts/mnss/courses/511/Ass1pdf/99Gp1.pdf [Accessed 14 March 2005].

Berg, K.S., 2006. Synergy for success effective construction safety and health

management. In: D. Fang, R.M. Choudhry and J.W. Hinze, eds. Proceedings of the CIB

W99 2006 international conference on global unity for safety and health in

construction, 28-30 June 2006, Beijing, China. Beijing: Tsinghua University Press, 34-

37.

Blockley, D., 1995. Process re-engineering for safety. Proceedings of risk engineering

and management in civil, mechanical and structural engineering. ICE, London, 51-66.

Boonrod, C., Kittiwisit, C., and Laokhongthavorn, L., 1998. Feasibility study of TIS

18000 for construction industry. Paper presented at the 6thnational civil engineering

conference, 10-12 May, 1998, Petchaburi, Thailand, 1-6.

Byrne, B.M., 2001. Structural equation modelling with AMOS: basic concepts,

applications and programming. Mahwah, NJ: Lawrence Erlbaum Associations, Inc.


287

Camison, C., 1996. Total quality management in hospitality: an application of the

EFQM model. Tourism management, 17(3), 191-201.

Caravatta, M., 1997. Conducting an organizational self-assessment using the 1997

Baldrige award criteria. Quality progress, 30(10), 87-91.

Chritamara, S. and Ogunlana, S.O., 2002. System dynamic modelling of design and

build construction projects. Construction innovation, 2, 269-295.

Clarke, S., 1999. Perceptions of organizational safety: implications for the development

of safety culture. Journal of organizational behaviour, 20(2), 185-198.

Clissold, G., 2004. Understanding safety performance using safety climate and

psychological climate. Working Paper 65/04, Department of Management, Monash

University, Australia.

Coakes, S.J. and Steed, L.G., 2003. SPSS analysis without anguish version 11.0 for

Windows. Australia: John Wiley & Sons.

Cohen, A., 1977. Factors in successful occupational safety programs. Journal of safety

research, 9, 168-178.

Cohen, J.M., 2002. Measuring safety performance in construction. Occupational

hazards, 64(6), 41-44.

Cooper, M.D., 2000. Towards a model of safety culture. Safety science, 36, 111-136.


288

Cooper, M.D., 2002. Safety culture: a model for understanding and quantifying a

difficult concept. Professional safety, 47(6), 30-35.

Cortada, J.W. and Woods, J.A., 1994. The Quality year 1994. New York: McGraw-Hill.

Curran, P.J., West, S.G., and Finch, J.F., 1996. The robustness of test statistics to

nonnormality and specification error in confirmatory factor analysis. Psychological

methods, 1(1), 16-29.

Dale, B.G. and Smith, M., 1997. Spectrum of quality management implementation grid:

development and use. Managing service quality, 7(6), 307-311.

Damodaran, R., 2006. Safety in construction organizations [online]. Property bytes: the

Indian real estate blog. Available from: http://propertybytes.com/?p=193 [Accessed 20

February 2007].

Dane, F., 1990. Research methods. California, U.S.A.: Brooks/Cole Publishing

Company.

Dedobbeleer, N. and Beland, F., 1991. A safety climate measure for construction sites.

Journal of safety research, 22, 97-103.

Dias, L.M.A. and Coble, R.J., 1996. Implementation of safety and health on

construction sites. Rotterdam, Netherlands: A.A. Balkema.

Dunlap, S., 2004. The role of management: influencing a complete safety culture

[online]. Available from: http://www.geaps.com/proceedings/2004/Dunlap.cfm

[Accessed 2 July 2004].


289

Duynisveld, J.L., 1999. A dynamic model of the N flows on a dairy farm. Thesis

(Master). Nova Scotia Agricultural College, Truro, Nova Scotia, Canada.

Eberlein, B., 2007. System dynamics software info [online]. Available from:

http://www.vensim.com/sdmail/sdsoft.html [Accessed 14 March 2007].

EFQM, 1998. Self-assessment: guidelines for companies, Brussels, Belgium: The

European Foundation for Quality Management, EFQM.

EFQM, 2000. Introducing excellence, Brussels, Belgium: The European Foundation for

Quality Management, EFQM.

Embassy of Denmark, Bangkok, 2006. Establishment of a regional workers’ institute

for occupational health, safety, and environment project [online]. Available from:

http://www.ambbangkok.um.dk/en/menu/DevelopmentCooperation/ProgrammeCompon

ents/UrbanEnvironmentalManagement/EstablishmentofaRegionalWorkersInstituteforOc

cupationalHealthSafetyandEnvironmentProject.htm [Accessed 26 September 2007].

Erdos, P. L., 1970. Professional mail surveys. New York: McGraw-Hill.

Eskildsen, J.K. and Dahlgaard, J.J., 2000. A causal model for employee satisfaction.

Total quality management, 11(8), 1081-1094.

Fang, D., Chen, Y. and Wong, L., 2006. Safety climate in construction industry: a case

study in Hong Kong. Journal of construction engineering and management, June, 573-

584.

Flannery, J.A., 2001. Safety culture and its measurement in aviation. Thesis (Master of

Aviation Management), University of Newcastle, Australia.


290

Flin, R. and Mearns, K.J., 1999. Assessing the state of organizational safety: culture or

climate? Current psychology, 18(1), 13-17.

Ford, A., 1999. Modelling the environment: an introduction to system dynamic

modelling of environment systems, USA: Island Press.

Forrester, J.W., 1961. Industrial dynamics, Cambridge, MA: MIT Press.

Forrester, J.W., 1985. Industrial dynamics, Cambridge, MA: MIT Press.

Forrester, J.W. and Senge, P.M., 1980. Tests for building confidence in system dynamic

models. TIMS studies in management science, 14, 209-228.

Fung, I.W.H., Tam, C.M., Tung, K.C.F. and Man, D.S.K., 2005. Safety cultural

divergences among management, supervisory and worker groups in Hong Kong

construction industry. International journal of project management, 23(7), 495-572.

Gadd, S. and Collins, A.M., 2002. Safety culture: a review of the literature. Report by

the Health & Safety Laboratory, UK: Sheffield University.

Garson, G.D., 2006. Structural equation modelling [online]. Available from:

http://www2.chass.ncsu.edu/garson/pa765/structur.htm [Accessed 4 August 2006].

Geller, E.S., 2000. 10 leadership qualities for a total safety culture. Professional safety,

45(5), 38-41.

Geller, E.S., 2001. A total safety culture: from a corporate achievement to a global

vision. Behaviour and social issues. 11(1), 18-20.


291

Gibb, A.G.F. and Foster, M., 1996. Safety motivation: evaluation of incentive schemes.

In: L.M.A. Dias and R.J. Coble, eds. Proceedings of the first international conference of

CIB working commission, 4-7 September 1996, Portugal. Rotterdam. Netherlands: A.A.

Balkema, 405-415.

Gillen, M., Baltz, D., Gassel, M., Kirsch, L., and Vaccaro, D., 2002. Perceived safety

climate, job demands, and co-worker support among union and non-union injured

construction workers. Journal of safety research, 33, 33-51.

Glendon, A.I. and Litherland, D.K., 2001. Safety climate factors, group differences and

safety behaviour in road construction. Safety science, 39, 157-188.

Goetsch, D.L., 2003. Construction safety and health, New Jersey: Pearson Education,

Inc.

Grote, G. and Kunzler, C., 2000. Diagnosis of safety culture in safety management

audits. Safety science, 34, 131-150.

Guldenmund, F.W., 2000. The nature of safety culture: a review of theory and research.

Safety science, 34(1-3), 215-257.

Hair, J.F., Anderson, R.E., Tatham, R.L. and Black, W.C., 1998. Multivariate data

analysis, 5th edition, U.S.A.: Prentice-Hall International.

Hale, A.R., 2000. Culture’ s confusions. Safety science, 34(1-3), 1-14.

Harvey, J., Erdos, G., Bolam, H., Cox, M.A.A., Kennedy, J.N.P. and Gregory, D.T.,

2002. An analysis of safety culture attitudes in a highly regulated environment. Work &

stress, 16(1), 18-36.


292

Hinze, J.W. and Raboud, P., 1988. Safety on large building construction projects.

Journal of construction engineering and management, 114(2), 286-293.

Ho, J.K.L. and Zeta, K.C., 2004. Cultural factors and their significance to the Hong

Kong construction industry [online]. Available from: http://www.ic.ployu.edu.hk/

esh/kb/culture/Ho&Zeta.pdf [Accessed 20 September 2004].

Hofstede, G., 1980. Framework for describing cultural differences [online]. Available

from: http://www.changecommblog.com/2006/09/26/framework-for-describing-

cultural-differences/ [Accessed 17 January 2008].

HSC, 2003. Health and safety statistics highlights 2002/03, HSE Books, Health and

Safety Commission, Sudbury, Suffolk.

Hsu, C.H., 2002. A structural equation modelling analysis of transformational

leadership, organizational culture and organizational effectiveness in Taiwanese

sport/fitness organizations. Thesis (PhD), United Stated Sports Academy, Alabama,

USA.

Hudson, P., 2001. Safety management and safety culture: the long and winding road.

Presented to CASA, September, 10, 2001, Canberra.

Huy, V.V., 2002. A multiple perspectives approach to organizational problem

formulation: two case studies. Thesis (PhD), Information and Operations Management,

Texas A&M University, USA.

ICAO, 1992. Human factors digest no.10: human factors, management and

organization. Montreal, Canada: International Civil Aviation Organization, ICAO.


293

INEEL, 2004. Idaho national engineering and environmental laboratory voluntary

protection program: total safety culture [online]. Available from: http://www.inel.gov/

vpp/total-safety-culture.shtml [Accessed 20 December 2004].

International Labour Organization, 2005. Thailand – occupational safety and health in

the construction industry [online]. Available from:http://www.ilo.org/public/english/

region/asro/bangkok/download/background/osh/conth05.pdf [Accessed 13 August

2007].

Ithink, 2003. Software reference guide: Ithink 8 technical documentation. USA: Isee

systems, Inc.

Jaafari, A., 1996. Human factors in the Australian construction industry: towards total

quality management. Australian journal of management, 21(2), 159-186.

Jackson, S.L., 2003. Research methods and statistics a critical thinking approach.

U.S.A.: Thomson Wadsworth.

Jaselskis, E.J., Anderson, S.D. and Russell, J.S., 1996. Strategies for achieving

excellence in construction safety performance. Journal of construction engineering and

management, March, 61-70.

Jones, M., 2007. Risks to high consequence [online]. Available at: http://www.awe.

co.uk/Images/Discovery%2013%20article%2002_tcm6-4325.pdf [Accessed 8

September 2007].

Joreskog, K. and Sorbom, D., 1993. LISREL 8: structural equation modelling with the

SIMPLIS command language. Chicago: SSI Inc.


294

Kartam, N., 1997. Integrating safety and health performance into construction CPM.


Kartam, N.A., Flood, I. and Koushki, P., 2000. Construction safety in Kuwait: issues,

procedures, problems and recommendations. Safety science, 36, 163-184.

Khanna, V.K., Vrat, P., Shankar, R., and Sahay, B.S., 2004. Managing the transition

phases in the TQM journey: a system dynamics approach. International journal of

quality & reliability management, 21(5), 518-544.

Kline, R.B., 2005. Principles and practice of structural equation modelling, 2nd edition,

New York: Guilford Press.

Kristensen, K. and Juhl, H.J., 1999. Beyond the bottom line-measuring stakeholder

value. In: B., Edvardsson, and A., Gustafsson, eds. The Nordic School of Quality

Management. Studentlitteratur: Lund.

Kumar, R., 2005. Research methodology: a step-by-step guide for beginners, 2nd

edition, London: Thousand Oaks.

Lamotte, G. and Carter, G., 2000. Are the balanced scorecard and the EFQM

excellence model mutually exclusive or do they work together to bring added value to a

company? Prepared for the EFQM common interest day, March 17, 2000, UK:

Balanced Scorecard Collaborative Europe.

Langford, D., Rowlinson, S. and Sawacha, E., 2000. Safety behaviour and safety

management: its influence on the attitudes of workers in the UK construction industry.

Engineering, construction and architectural management, 7(2), 133-140.


295

Lardner, R., Fleming, M. and Joyner, P., 2001. Towards a mature safety culture.

IChemE symposium series, 148, 635-642.

Lee, T., 1998. Assessment of safety culture at a nuclear reprocessing plant. Work &

stress, 12(3), 217-237.

Lingard, H. and Blismas, N., 2006. Building a safety culture: the importance of shared

mental models in the Australian construction industry. In: D. Fang, R.M. Choudhry and

J.W. Hinze, eds. Proceedings of the CIB W99 2006 international conference on global

unity for safety and health in construction, 28-30 June 2006, Beijing, China. Beijing:

Tsinghua University Press, 201-208.

Little, A.D., 2002. Improving safety culture in the construction industry. A workshop

for senior management in construction contracting and client companies. Cambridge:

University Press.

Love, P.E.D., Mandal, P., Smith, J. and Li, H., 2000. Modelling the dynamics of design

error induced rework in construction. Construction management and economics, 18,

567-574.

Maloney, W.F., 2003. Employee involvement, consultation and information sharing in

health and safety in construction. Report submitted on the work performed under

engineering physical science research Council. University of Kentucky and Glasgow

Caledonian University.

Mbuya, E. and Lema, N.M., 2004. Towards development of a framework for integration

of safety and quality management techniques in construction project delivery process

[online]. Available from: http://buildnet.csir.co.za/w107/Authors/Accepted% 20Papers/

012p%20-%20final.doc [Accessed 5 May 2004].


296

McBurney, D.H., 1994. Research methods. Pacific Grove, California, U.S.A.:

Brooks/Cole Publishing Company.

McDougall, M., 2004. Developing a positive safety culture. Environment, Health and

Safety Office, University of California, San Diego, U.S.A.

McLucas, A.C., 2005. System dynamics applications: a modular approach to modelling

complex world behaviour. Canberra, Australia: Argos Press.

Melville, S. and Goddard, W., 1996. Research methodology an introduction for science

& engineering students. Kenwyn: Juta & Co., Ltd.

Miller, N.G., 1990. A methodology for introducing technology into organizations.

Thesis (PhD), University of New South Wales, Australia.

Mohamed, S., 1999. Empirical investigation of construction safety management

activities and performance in Australia. Safety science, 33, 129-142.

Mohamed, S., 2002. Safety climate in construction site environments. Journal of

construction engineering and management, 128(5), 375-384.

Mohamed, S., 2003. Scorecard approach to benchmarking organizational safety culture

in construction. Journal of construction engineering and management, 129(1), 80-88.

Molenaar, K., Brown, H., Caile, S. and Smith, R., 2002. Corporate culture: a study of

firms with outstanding construction safety. Professional safety, July, 18-27.


297

Morecroft, J.D.W., 1988. System dynamics and microworlds for policymakers.

European journal of operational research, 35, 301-320.

Morgan, G.A. and Griego, O.V., 1998. Easy use and interpretation of SPSS for

Windows. USA: Lawrence Erlbaum Associates.

Mueller, R.O. and Hancock, G.R., 2004. Factor analysis and latent structure,

confirmatory. International Encyclopaedia of the Social & Behavioural Sciences, 5239-

5244.

Niskanen, T., 1994. Safety climate in the road administration. Safety science, 17, 237-

255.

NIST, 1993. 1993 award criteria, Malcolm Baldrige National Quality Award.

Gaithersburg, MD: National Institute of Standards and Technology, NIST.

Northouse, P.G., 1997. Leadership: theory and practice. Thousand Oaks, Calif: Sage

Publications.

NOSHC, 2005. Regulation impact statement, National Occupational Health and Safety

Commission, Sydney, Australia.

NPS Risk Management Division, 2006. Occupational safety and health overview for

NPS employees [online]. Available from: http://www.nps.gov/training/tel/Guides/

OSH_emp_ pguide_2006_0605.pdf [Accessed 3 August 2007].

O’ Dea, A. and Flin, R., 2001. Site manager and safety leadership in the offshore oil and

gas industry. Safety science, 37, 39-57.


298

Ogunlana, S., Promkuntong, K., and Jearkjirm, V., 1996. Construction delays in a fast-

growing economy: comparing Thailand with other economies. International journal of

project management, 14(1), 37-45.

Ogunlana, S., Siddiqui, Z., Yisa, S., and Olomolaiye, P., 2002. Factors and procedures

used in matching project managers to construction projects in Bangkok. International

journal of project management, 20(5), 385-400.

Oklahoma Department of Labour, 1998. Essential elements of an effective safety &

health program [online]. Available from: http://www.state.ok.us/~okdol/peosh/stee.pdf

[Accessed 29 January 2007].

Olcott, J.W., 1997. Characteristics of safety cultures. Flight Safety Foundation

Corporate Aviation Safety Seminar, Phoenix, AZ, May 1, 1997.

Otley, D.T., 1999. Performance management: a framework for management control

system research, Management accounting research, 10, 363-382.

Packendorff, J., 1995. Inquiring into the temporary organization: new directions for

project management research. Scandinavian journal of management, 11(4), 319-333.

Pallant, J., 2005. SPSS survival manual: a step by step guide to data analysis using

SPSS for windows (version 12), New South Wales: Allen & Unwin.

Pannirselvam, G.P. and Ferguson, L.A., 2001. A study of the relationships between the

Baldrige categories. International journal of quality & reliability management, 18(1),

14-34.


299

Pasman, H.J., 2000. Risk informed resource allocation policy: safety can save costs.

Journal of hazardous materials, 71, 375-394.

Paul, P.S. and Maiti, J., 2007. The role of behavioural factors on safety management in

underground mines. Safety science, 45(4), 449-471.

Pipitsupaphol, T. and Watanabe, T., 2000. Identification of root causes of labour

accidents in the construction industry. Proceedings of the 4th Asia Pacific Structural

Engineering and Construction Conference, Malaysia, pp. 93-98.

Potter, D.L., 2003. Research report organizational culture and safety: integrating for a

safe workplace [online]. Available from: http://www.debpotter.com/admin/files/files/

Organization%20safety%20and%20culture.pdf [Accessed 14 July 2004].

Ratikin, S.R., 2001. Software verification and validation for practitioners and

managers. Norwood : Artech House.

Rodrigues, A.G. and Bowers, J., 1996. The role of system dynamics in project

management. International journal of project management, 14(4), 213-220.

Rodrigues, A.G. and Williams, T.M., 1998. System dynamics in project management:

assessing the impacts of client behaviour on project performance. Journal of the

operational research society, 49(1), 2-15.

Rosenfeld, Y., Rozenfeld, O., Sacks, R., and Baum, H., 2006. Efficient and timely use

of safety resources in construction. In: D. Fang, R.M. Choudhry and J.W. Hinze, eds.

Proceedings of the CIB W99 2006 international conference on global unity for safety

and health in construction, 28-30 June 2006, Beijing, China. Beijing: Tsinghua

University Press, 290-297.


300

Saeed, K. and Brooke, K., 1996. Contract design for profitability in macro-engineering

projects. System dynamics review, 12(3), 235-246.

Sarshar, M., Haigh, R., Finnemore, M., Aouad, G., Barrett, P., Baldry, M. and Sexton,

M., 2000. SPICE: a business process diagnostic tool for construction projects.

Engineering, construction and architectural management, 7(3), 241-250.

Sawacha, E., Naoum, S. and Fong, D., 1999. Factors affecting safety performance on

construction sites. International journal of project management, 17(5), 309-315.

Selvanathan, S. and Selvanathan, S., 2005. Sampling methods and survey &

questionnaire design. RHD statistics and research design support (STARDS) unit,

Griffith University, Gold Coast, Australia.

Seo, D.C., Torabi, M.R., Blair, E.H., and Ellis, N.T., 2004. A cross-validation of safety

climate scale using confirmatory factor analytic approach. Journal of safety research,

35, 427-445.

Sheffield Hallam University, 2003. Linking the EFQM excellence model to other

management models and tools. Centre for Integral Excellence, Sheffield Hallam

University, Howard, Sheffield, UK.

Simonovic, S., 2005. System simulation [online]. The University of Western Ontario,

Canada. Available from: http://www.engga.uwo.ca/research/iclr/simonovic/ES566/

Lecture4.pdf [Accessed 22 March 2005].

Siu, O., Phillips, D.R. and Leung, T., 2004. Safety climate and safety performance

among construction workers in Hong Kong: the role of psychological strains as

mediators. Accident analysis and prevention, 36, 359-366.


301

Smith, G.R. and Roth, R.D., 1991. Safety programs and the construction manager.


Sorensen, J.N., 2002. Safety culture: a survey of the state-of-the-art. Reliability

engineering and system safety, 76, 189-204.

Speirs, F. and Johnson, C.W., 2002. Safety culture in the face of industrial change: a

case study from the UK rail industry. Research Report, University of Glasgow,

Glasgow, Scotland, May 29, 2002.

Sterman, J.D., 1992. System dynamic modelling for project management. Cambridge:

MIT.

Suraji, A., Duff, R., and Peckitt, S.J., 2001. Development of causal model of

construction accident causation. Journal of construction engineering and management,

July/ August, 337-344.

Tabachnick, B.G. and Fidell, L.S., 2007. Using Multivariate Statistics. 5th edition, USA:

Pearson Education, Inc.

Tam, C.M., Zeng, S.X. and Deng, Z.M., 2004. Identifying elements of poor

construction safety management in China. Safety science, 42(7), 569-586.

Tang, S.L., De Saram D.D., Wang, Z.M. and Zhang, T.Q., 2003. Costs of construction

accidents in social and humanity context. Proceedings of the 9th east-pacific conference

on structural engineering & construction, Indonesia. Paper number: 1433.

Tang, Y.H. and Ogunlana, S.O., 2003a. Modelling the dynamic performance of a

construction organization. Construction management and economics, 21, 127-136.


302

Tang, Y.H. and Ogunlana, S.O., 2003b. Selecting superior performance improvement

policies. Construction management and economics, 21, 247-256.

Taylor, R.H., 2003. Managing risks through developing a strong safety culture. Nuclear

energy, 42(6), 341-346.

Teo, A.L. and Fang, D., 2006. Measurement of safety climate in construction industry:

studies in Singapore and Hong Kong. In: D. Fang, R.M. Choudhry and J.W. Hinze, eds.

Proceedings of the CIB W99 2006 international conference on global unity for safety

and health in construction, 28-30 June 2006, Beijing, China. Beijing: Tsinghua

University Press, 157-164.

Teo, E.A.L., Ling, F.Y.Y. and Chong, A.F.W., 2005. Framework for project managers

to manage construction safety. International journal of project management, 23(4), 329-

341.

The UAE Ministry of Labour and Social Affairs, 2001. The reality of occupational

health and safety in UAE 2000-2001. Dubai, UAE: The Department of Research and

Studies, Ministry of Labour and Social Affairs.

Tummala, V.M.R. and Tang, C.L., 1995. Strategic quality management, Malcolm

Baldrige and European quality awards and ISO 9000 certification: core concepts and

comparative analysis. 1994-95 Annual Issue of IIE, Hong Kong.

Turner, B.A., 1991. The development of a safety culture. Chemistry and industry, 7,

241-247.


303

Ullman, J.B., 2001. Structural equation modelling. In: B.G. Tabachnick, and L.S D.

Fidell. Using Multivariate Statistics. 5th edition. Needham Heights, MA: Allyn &

Bacon, 157-164.

Vecchio-Sadus, A.M. and Griffiths, S., 2004. Marketing strategies for enhancing safety

culture. Safety science, 42(2), 601-619.

Vennix, J.A.M., 1996. Group model-building: facilitating team learning using system

dynamics England: John Wiley & Sons.

Ventana Systems, Inc., 2001. System behaviour and causal loop diagrams [online].

Available from: http://www.public.asu.edu/~kirkwood/sysdyn/SDIntro/ch-1.pdf

[Accessed 24 July 2007].

Wallace, A. and Neal, A., 2000. A report on safety in the Queensland meat industry.

Report prepared for the Meat Industry Advisory Group and the Australian Meat

Industry Employees Union, Australia.

Wangniwetkul, K., 2007. Use of personal safety equipment in construction [online].

Available from: http://www.eit.or.th/article/data/01040026.pdf [Accessed 24 September

2007].

Williamson, A.M., Feyer, A., Cairns, D. and Biancotti, D., 1997. The development of a

measure of safety climate: the role of safety perceptions and attitudes. Safety

science, 25(1-3), 15-27.

Wong, P.S.P. and Cheung, S.O., 2005. Structural equation model of trust and partnering

success. Journal of management in engineering, April, 70-80.


304

Wongrassamee, S., Gardiner, P.D. and Simmons, J.E.L., 2003. Performance

measurement tools: the balanced scorecard and the EFQM excellence model. Measuring

business excellence, 7(1), 14-29.

Wright, M.S., Brabazon, P., Tipping, A. and Talwalkar, M. 1999. Development of a

business excellence model of safety culture: safety culture improvement matrix. London:

Entec UK Ltd.

Zohar, D., 1980. Safety climate in industrial organizations: theoretical and applied

implications. Journal of applied psychology, 65(1), 96-102.