This article was downloaded by: [University of Western Ontario]On: 07 October 2014, At: 01:58Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954 Registeredoffice: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Laterality: Asymmetries of Body, Brainand CognitionPublication details, including instructions for authors andsubscription information:http://www.tandfonline.com/loi/plat20

Using hand performance measures topredict handednessSusan Brown a , Eric Roy a , Linda Rohr a & Pamela Bryden ba University of Waterloo, Canadab Wilfrid Laurier University, ON, CanadaPublished online: 21 Sep 2010.

To cite this article: Susan Brown , Eric Roy , Linda Rohr & Pamela Bryden (2006) Using handperformance measures to predict handedness, Laterality: Asymmetries of Body, Brain and Cognition,11:1, 1-14, DOI: 10.1080/1357650054200000440

To link to this article: http://dx.doi.org/10.1080/1357650054200000440

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the“Content”) contained in the publications on our platform. However, Taylor & Francis,our agents, and our licensors make no representations or warranties whatsoever as tothe accuracy, completeness, or suitability for any purpose of the Content. Any opinionsand views expressed in this publication are the opinions and views of the authors,and are not the views of or endorsed by Taylor & Francis. The accuracy of the Contentshould not be relied upon and should be independently verified with primary sourcesof information. Taylor and Francis shall not be liable for any losses, actions, claims,proceedings, demands, costs, expenses, damages, and other liabilities whatsoever orhowsoever caused arising directly or indirectly in connection with, in relation to or arisingout of the use of the Content.

This article may be used for research, teaching, and private study purposes. Anysubstantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

Using hand performance measures to predict

handedness

Susan G. Brown, Eric A. Roy, and Linda E. Rohr

University of Waterloo, Canada

Pamela J. Bryden

Wilfrid Laurier University, ON, Canada

Handedness is defined by the individual's preference to use one hand pre-dominately for unimanual tasks and the ability to perform these tasks more effi-ciently with one hand (Corey, Hurley, & Foundas, 2001). It is important to useperformance variables to measure handedness because they are more objectivethan traditional hand preference questionnaires (Bryden, Pryde, & Roy, 2000a).The current study develops a predictive model of handedness as measured by theWaterloo Handedness Questionnaire (WHQ) using several performance indicatorsof handedness. A total of 120 individuals (60 right-handers and 60 left-handers)were asked to complete four performance-based tasks: the Grooved Pegboard(GP), the Annett pegboard (AP), finger tapping (FT), and grip strength (GS) aswell as an observational measure of preference, the Wathand Box Test (WBT).Backward linear regression analysis showed that the Wathand Box measure andthe laterality quotients for several performance measures (GP place, AP, and FT)combined to act as the most accurate predictors of hand preference. The predictivemodel of handedness developed is as follows: WHQ = 72.760 ± 0.667(GP place)+ 0.809(FT) + 0.234(WBT) ± 0.748(AP) with an explained variance of 0.836.These results illustrate, as Corey et al. (2001) suggested, that the best predictivemodel of handedness combines preference measures and several performancemeasures that tap into different elements of motor performance. By developing thismodel, it is possible to get an accurate measure of handedness using objectivemeasures.

Handedness is one of the most obvious asymmetries of human behaviour, and

arises from individuals using one hand more often than the other for unimanual

activities (Corey, Hurley, & Foundas, 2001). The relationship between hand

preference, hand performance, and hemispheric asymmetries is crucial to

Address correspondence to Susan G. Brown, Department of Kinesiology, University of Waterloo,

200 University Avenue W, Waterloo, ON N2L 3G1, Canada. Email: sgbrown@ahsmail.uwaterloo.ca

This work was supported by grants from NSERC (E.A.R. and P.J.B.) and the Heart and Stroke

Foundation of Ontario (E.A.R.).

# 2006 Psychology Press Ltd

http://www.psypress.com/laterality DOI:10.1080/1357650054200000440

LATERALITY, 2006, 11 (1), 1±14

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 0

1:58

07

Oct

ober

201

4

understanding the neural systems that underlie human behavioural asymmetries

(Corey et al., 2001). Specifically, an understanding of the relationship between

hand preference and hand performance will undoubtedly help in examining the

relationship between handedness and cerebral organisation, especially with

respect to language functions, and may ultimately help define the risk of

acquiring specific language disorders that occur more frequently in left-handers

than right-handers, such as aphasia (Bryden, Bulman-Fleming, & MacDonald,

1996).

Both hand preference measures and hand performance measures can be used

to categorise individuals into handedness groups. Preference measures provide a

way of subjectively measuring handedness for unimanual activities, whereas

performance measures afford a more objective measurement of handedness

(Bryden, Pryde, & Roy, 2000b). It is possible to link the subjective preference

measures with the objective performance measures by developing predictive

models of handedness.

The distributions of preference and performance data are different. Whereas

preference measures typically exhibit a bimodal distribution with two handed-

ness groups, performance measures such as peg-moving tasks tend to be dis-

tributed unimodally (Annett, 2002). As a result, there are two identifiable groups

when examining preference data, but not when examining performance data.

Therefore, finding a link between preference and performance data is

challenging, unless the factors that underlie handedness groups are the same for

preference and performance measures (Corey et al., 2001).

Questionnaires offer a convenient means of dividing individuals into hand-

edness groups, as they are easier to administer than performance-based measures

(Bryden et al., 2000b). However, despite the time advantage, measuring hand-

edness with preference measures is not always ideal due to the inherent sub-

jectivity of the task; they rely on the reader's interpretation of the question as

well as ability to imagine oneself performing the particular task (Bryden et al.,

1996). Due to their subjectivity, questionnaires are particularly unreliable when

administered to special populations such as the elderly or children, because

individuals in these populations may have difficulty remembering which hand

they would use to perform certain tasks, and may have difficulty making

appropriate judgement calls about which hand is used in certain circumstances

(Bryden et al., 2000a). In contrast, performance measures have an important

objectivity but require both a greater amount of time and increased resources to

administer compared to preference measures. Despite the objectivity advantage,

preference measures have traditionally been used to divide individuals into

handedness groups due to the ease of administration (Peters, 1998).

Because there are so many different underlying components of hand per-

formance (proximal versus distal musculature, and fine versus gross control), a

combination of performance measures, each emphasising a different aspect of

performance, is important in creating an accurate predictive model of handed-

2 BROWN ET AL.

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 0

1:58

07

Oct

ober

201

4

ness (Corey et al., 2001). Some performance measures, such as writing and

throwing a ball, are highly specialised and could be influenced by the indivi-

dual's experience in performing the task, thus skewing the outcome distribution

in favour of the preferred hand (Peters, 1998). To ensure that the performance

tasks are measuring true hand differences, they must be specialised, but not so

much that those differences are due to the individual's experience in performing

the task (Peters, 1998). For example, a task such as writing or sewing would be

specialised enough to show large hand performance differencesÐhowever, the

fact that the participant would routinely perform these tasks with one hand

would skew the results and add a component of differential experience between

the hands to the task performance. Tasks that are less specialised and so meet the

criteria mentioned above include peg-moving tasks as well as finger-tapping

tasks.

The current study attempts to examine in more detail the relationship between

measures of hand preference and measures of hand performance. By establishing

how preference and performance measures are related, it might be possible to

develop effective preference-based measures of handedness that use many of the

same skills as do performance-based measures. In this way, it would be possible

to remove some of the subjectivity of preference measures by basing specific

items on the questionnaire on those performance measures that have been linked

to true lateralisation effects, or developing more robust observational measures

of preference to measure handedness.

METHOD

Participants

Data were collected on 120 volunteers between the ages of 18 and 25. Partici-

pants were not recruited on the basis of gender, but handedness was important:

60 participants were self-proclaimed right-handers (43 females, 17 males) and

60 participants were self-proclaimed left-handers (38 females, 22 males). Rather

than use a sample representative of the population with approximately 90%

right-handers and 10% left-handers, an equal number of right-handed and left-

handed individuals were tested in order to be able to increase the power to find

the relationship between hand preference and hand performance. This paper

does not attempt to predict the probability of being either right- or left-handed

based on test scores, and therefore using an equal number of right- and left-

handers was appropriate. Inclusion in either the right or left hand preference

group was determined with respect to the participant's stated hand preference.

The majority of subjects were recruited from introductory psychology and

psychomotor behaviour classes. All subjects were free from neurological

damage and had corrected to normal vision. The study was approved by the

Office of Research Ethics, and all participants gave informed consent before

beginning the study.

PERFORMANCE MEASURES OF HANDEDNESS 3

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 0

1:58

07

Oct

ober

201

4

Apparatus and procedures

The preference and performance tasks used in this study are standard measures

used in research on manual asymmetries, and included a preference measure,

performance measures, and an observational measure of preference (the Wat-

hand Box Test). Hand preference was evaluated using the Waterloo Handedness

Questionnaire and the Wathand Box Test. Hand performance indicators inclu-

ded: two pegboard tasks, a finger-tapping task, and a measure of grip strength.

Participants completed the tasks in a randomised order, however each partici-

pant completed all trials of a single task before moving on to the subsequent

task.

Waterloo Handedness Questionnaire (WHQ). The Waterloo Handedness

Questionnaire was the primary measure of hand preference in the study.

Participants were presented with 20 questions, asking them to indicate which

hand they would use to perform a series of unimanual activities (such as using a

hammer, writing). Some of the items on the questionnaire reflect skilled

performance (i.e., writing) whereas other items reflect relatively unskilled

activities (i.e., opening a drawer). Five possible responses were offered for each

question, allowing the participant to rate the frequency with which they would

use a particular hand for each activity using a 5-point scale (i.e., ``always use the

left hand'', ``usually use the left hand'', ``use both hands equally often'',

``usually use the right hand'', ``always use the right hand''). Each of these

responses was scored as 72, 71, 0, 1, 2 respectively, and the dependent

handedness measure was calculated as the total composite score of these

individual responses. Therefore, right-handed individuals yielded positive scores

on the WHQ whereas left-handed individuals yielded negative scores. For a list

of the questions included in the WHQ, please refer to the Appendix.

Wathand Box Test (WBT). Participants were asked to complete several

unimanual tasks such as lifting a cupboard door, using a toy hammer, placing

rings on hooks, and tossing a ball. The researcher recorded which hand they used

to perform these activities. A laterality quotient was calculated by subtracting

the number of tasks performed with the left hand from the number of tasks

performed with the right hand and dividing by the total number of tasks. This

laterality quotient was renamed the ``Wathand Box Score'', and was used in the

statistical analysis.

Previous research has shown that the WBT is an accurate performance-based

measure of preference and it has been used to measure hand preference in

special populations, including children, where traditional questionnaire-based

measures of preference would not be appropriate (Bryden et al., 2000b).

4 BROWN ET AL.

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 0

1:58

07

Oct

ober

201

4

Grooved Pegboard (GP). The GP task provided a measure of manipulative

skill and visuomotor control. The first portion of the GP place phase of the GP

task was completed according to the standard instructions provided in the

Lafayette Instrument Instruction/Owner's Manual for the 32025 Grooved

Pegboard (Lafayette Instruments, 1989). Participants were asked to complete

two trials with each hand in which they were required to move a set of 25 pegs

from a large receptacle to a set of 25 holes. This was called the ``place phase''.

The pegs were irregularly shaped and would therefore only fit into the holes in

one orientation. These pegs were approximately the same size as the holes and

therefore the task required significant skill to complete. In addition to

completing the place phase, participants were required to complete two trials

using an alternative method of administration in which they were required to

remove the pegs from the holes and place them back into the receptacle,

beginning at the bottom of the board, on the contralateral side to the hand being

used (Bryden & Roy, 2005). This was called the ``replace phase''. For both the

place and replace phases, participants first completed each set of trials with their

dominant hand, and in each case, a dropped peg resulted in the trial being

repeated.

The dependent variable for this task was the time required for each compo-

nent to be completed. For the place phase, timing began when the first peg was

placed in the hole and was stopped with the last peg was placed in the hole, and

for the replace phase, timing began when the first peg was placed in the

receptacle and was stopped when the last peg was placed in the receptacle. In

each phase, the average time across the two trials was used in calculating the

laterality quotient.

Annett pegboard (AP). Similar to the GP task, the AP task provided a

measure of skill and speed. All participants were asked to move 10 small dowels

from one horizontal row of holes to a parallel row (located closer to the subject)

as rapidly as possible. Performance was measured by the amount of time

required to complete the peg-moving task. Three trials were completed with

each hand and the times were averaged for each hand. The hand with which

participants began was randomised between subjects, and participants alternated

between performing the task with the starting hand and the other hand.

Participants were requested to begin with their hand on the table and were

provided with a cue to begin the task. Participants were required to move the

pegs, one at a time, beginning on the side of the pegboard contralateral to the

hand being used to perform the task. Timing ended when the last peg was placed

in the hole. As with the Grooved Pegboard task, if the participant dropped a peg,

the trial was repeated.

Finger tapping (FT). The FT task provided a measure of distal muscular

control and coordination. Participants were required to place their hand palm

PERFORMANCE MEASURES OF HANDEDNESS 5

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 0

1:58

07

Oct

ober

201

4

down on the table with their index finger extended and on a key. The other

fingers were extended and resting on the table. Participants were asked to tap the

key as many times as possible in a 10-second period, using only the index finger.

Participants completed five trials with the preferred hand, followed by five trials

with the non-preferred hand. For each hand, a rest period of 30 seconds was

offered following the third tapping trial, in order to avoid fatigue. Performance

was measured by the number of taps performed in the 10-second period, and

averaged across repeated trials.

Grip strength (GS). Each participant was required to execute three

maximum grip strength efforts with each hand, while standing with the arm

straight, holding a dynamometer. The participant performed three trials with

each hand, beginning with the preferred hand. Performance was measured in

KgW for each effort and was averaged across all trials.

Data analysis

For the Grooved Pegboard (GP), Annett pegboard (AP), finger-tapping (FT), and

grip strength (GS) tasks described above, performance was expressed as a

laterality quotient, i.e., (Rperformance ± Lperformance) / (Rperformance + Lperformance).

Therefore, for the pegboard tasks, a positive score was indicative of left hand

superiority whereas a negative score was indicative of right hand superiority. For

all other performance measures, a positive score is indicative of right hand

superiority.

RESULTS

In an attempt to remove outlying data points, the data were examined to ensure

that all data points fell within three standard deviations of the mean for each

measure of preference or performance prior to data analysis. Three data points

were removed because they did not meet this criterion. Table 1 presents the

modified means and standard deviations (according to hand preference and

gender) following the exclusion of the outlying values. Subsequent data analysis

was completed based on the trimmed sample.

Upon inspection of the data divided in this manner, it appears as though the

two handedness groups showed similar variability across each of the perfor-

mance variables. While two of the variables (WBT and GS) appeared to show a

fairly large difference in variability (WBT due to the nature of the way in which

the task is scored, and GS due to the gross nature of the task itself) the difference

in variability on any of the performance tasks was not statistically significant (as

calculated by Fisher's F-test for the equality of variances). Despite the observed

equality of variance, the left-handers were less strongly biased to the left than

the right-handers were biased to the right. Table 2 presents data about the

concordance of hand advantage (based on outcomes of performance tasks) for

6 BROWN ET AL.

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 0

1:58

07

Oct

ober

201

4

each of the performance-based measures with the individual's hand preference

(as measured by the WHQ). In this table, the number of participants in each

hand preference group (and the corresponding proportion of participants in that

category) as well as the means for that portion of the sample are reported. It

appears as though the only task that showed equal concordance between stated

hand preference and hand advantage on the performance task between the two

handedness groups was the Annett pegboard task. The hand advantage on all

other performance measures showed higher concordance with stated hand

preference in the group of right-handers than in the group of left-handers.

Is hand performance related to hand preference asmeasured by the WHQ?

Examination of the correlation matrix (Table 3) indicates that all performance

variables were significantly correlated with the composite score of the Waterloo

Handedness Questionnaire (called the ``handedness score''), except the replace

phase of the Grooved Pegboard task. This follows previous research, reinforcing

TABLE 1Modified means and standard deviations for all performance variables according to

gender and hand preference

Right-handers Left-handers

Gender Variable N Mean SD N Mean SD

Female Grooved Pegboard ± place 43 75.89 4.93 37 4.02 4.71

Grooved Pegboard ± remove 43 72.04 3.86 37 70.37 3.38

Finger tapping 43 4.46 4.49 37 70.70 3.86

Grip strength 43 6.05 6.00 38 70.93 7.07

Wathand Box Test 43 70.70 35.55 38 741.05 46.25

Annett pegboard 43 73.57 3.65 38 3.27 3.55

Male Grooved Pegboard ± place 17 74.23 4.11 22 3.03 5.02

Grooved Pegboard ± remove 17 71.45 3.034 22 72.41 4.61

Finger tapping 17 3.95 3.50 22 70.88 4.38

Grip strength 17 2.96 5.19 22 70.59 5.13

Wathand Box Test 17 69.41 43.08 22 728.18 37.37

Annett pegboard 17 72.12 4.99 22 1.88 3.39

Total Grooved Pegboard ± place 60 75.42 4.74 59 3.65 4.81

Sample Grooved Pegboard ± remove 60 71.87 3.63 59 71.13 3.97

Finger tapping 60 4.32 4.21 59 70.77 4.02

Grip strength 60 5.17 5.91 60 70.80 6.38

Wathand Box Test 60 70.33 37.46 60 736.33 43.33

Annett pegboard 60 73.16 4.09 60 2.76 3.53

PERFORMANCE MEASURES OF HANDEDNESS 7

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 0

1:58

07

Oct

ober

201

4

the concept that the replace phase of the Grooved Pegboard is a relatively

unskilled activity that reveals only small differences between the preferred and

the non-preferred hands (Bryden & Roy, 1999). The Wathand Box Test had the

highest correlation with the handedness score of the Waterloo Handedness

Questionnaire, suggesting that it is an accurate predictor of hand preference

because they are both measures of hand preference. A similar result was seen in

our previous work (Brown, Roy, Rohr, Snider & Bryden, 2004).

Despite relatively strong correlations between the hand preference item and

the hand performance tasks, the intercorrelations between the hand performance

tasks themselves are quite weak. This may be due to the fact that each of the

TABLE 2Concordance between stated hand preference and hand advantage on performance

tasks

Hand preference as determined by WHQ

Hand advantage as determined by Right Left

performance measures

N Mean N Mean

Right Grooved Pegboard ± place 53

(88%)

76.44 7

(12%)

2.30

Grooved Pegboard ± remove 43

(72%)

73.51 17

(28%)

2.29

Finger tapping 52

(88%)

5.43 7

(12%)

73.34

Grip strength 49

(83%)

7.18 10

(17%)

74.14

Wathand Box Test 55

(93%)

78.18 4

(7%)

720.00

Annett pegboard 47

(78%)

74.73 13

(22%)

2.53

Left Grooved Pegboard ± place 13

(22%)

72.93 46

(78%)

5.51

Grooved Pegboard ± remove 35

(59%)

73.70 24

(41%)

2.62

Finger tapping 22

(38%)

3.38 36

(62%)

73.32

Grip strength 29

(51%)

4.29 28

(49%)

76.17

Wathand Box Test 7

(14%)

34.29 43

(86%)

756.28

Annett pegboard 13

(22%)

72.26 47

(78%)

4.14

Percentage scores refer to percentage in each hand preference group.

8 BROWN ET AL.

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 0

1:58

07

Oct

ober

201

4

performance tasks taps into a different aspect of motor control (i.e., proximal

control vs manipulative abilities, etc.) and these different aspects of motor

control may not be strongly correlated with each other. Because the inter-

correlations between hand performance tasks are weak, a combination of the

performance tasks is required to appropriately predict hand preference; no one

hand performance task alone would serve this purpose.

What is the best way to predict handedness (asmeasured by the WHQ)?

Backward linear regression analysis was completed using all variables that were

correlated significantly with the Waterloo Handedness Questionnaire. The first

model included all variables and produced a predictive equation as follows (with

an explained variance term of 0.837):

TABLE 3Correlation matrix (all performance and preference measures)

WHQ GP ± P GP ± R FT GS WBT AP

WHQ 1 70.734**

0.000

119

70.175

0.057

119

0.574**

0.000

119

0.498**

0.000

120

0.877**

0.000

120

70.680**

0.000

120

GP ± P 1

119

0.309*

0.001

119

70.417**

0.000

119

70.389**

0.000

119

70.670**

0.000

119

0.665**

0.000

119

GP ± R 1

119

70.081

0.386

118

70.151

0.101

119

70.173

0.059

119

0.326*

0.000

119

FT 1

119

0.264**

0.004

119

0.492**

0.000

119

70.379**

0.000

119

GS 1

120

0.509**

0.000

120

70.411**

0.000

120

WBT 1

120

70.609*

0.000

120

AP 1

120

Bold = Pearson correlation; Italics = Significance (2-tailed); Plain text = N.

PERFORMANCE MEASURES OF HANDEDNESS 9

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 0

1:58

07

Oct

ober

201

4

WHQ = 72.942 7 0.665(GP place) + 0.808(FT) + 0.133(GS)

+ 0.228(WBT) 7 0.719(AP) (1)

In this model, the grip strength variable did not contribute significantly (p =

.427), and therefore, a second model was computed excluding the grip strength

variable (this time with an explained variance term of .836, adjusted R2 = .830):

WHQ = 72.760 7 0.667(GP place) + 0.809(FT) + 0.234(WBT) 7 0.748(AP) (2)

The change in explained variance between the first and the second model was

not significant (p = .427), illustrating that the second model was the most

parsimonious predictive model of handedness when WHQ is the dependent

variable. Corey et al. (2001) also found that removing the grip strength variable

from a regression model predicting preference from a series of performance

variables (including finger-tapping and pegboard tasks) did not significantly

reduce the level of explained variance in the model. Therefore, EQ2 is a more

parsimonious model than EQ1, using fewer variables to make as accurate a

prediction as EQ1.

Do the performance variables relate in the sameway to the WHQ and the WBT?

Both the questionnaire and the Wathand Box Test are effective tools for

uncovering performance differences between the hands, and Table 2 shows that

both of these variables are highly correlated with performance on the Annett

pegboard task, which has been widely used as a standard tool for exposing

differential performance between the hands. Because the questionnaire and the

Wathand Box Test are highly correlated with the Annett pegboard scores as

well as with each other, it was important to know whether one of these vari-

ables was statistically more significant at measuring hand preference than the

other. If the performance variables predict the Wathand Box Test score as well

as they predict the questionnaire score, it is possible that the Wathand Box Test

could replace the questionnaire as a performance-based measure of preference

and therefore remove some of the subjectivity associated with preference

measures. In order to test this, two new regression models were developed

using only the performance variables (i.e., Grooved Pegboard, finger tapping,

grip strength, and Annett pegboard) to predict the composite scores from the

two preference measures (the Wathand Box Test and the questionnaire). These

equations differ from EQ2 in that they do not use the questionnaire score as a

predictor variable.

The equations are reported using the unstandardised coefficients, but the

standardised coefficients are reported in Table 4 for future discussion.

Equation 3 used the performance measures to predict the WBT score (with an

explained variance term of .575, adjusted R2 = .559) and was as follows:

10 BROWN ET AL.

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 0

1:58

07

Oct

ober

201

4

WBT = 4.635 ± 3.729(GP place) + 2.655(FT) + 2.265(GS) ± 2.546(AP) (3)

Equation 4 used the performance measures to predict the WHQ score (with

an explained variance term of .689, adjusted R2 = .678) and was as follows:

WHQ = ±1.886 ± 1.515(GP place) + 1.413(FT) + 0.649(GS) ± 1.300(AP) (4)

Examination of the plots of the unstandardised residual values from the

resultant regression equations showed no evidence of deviation from normality,

indicating that the regression equations did not violate the assumption of nor-

mality, despite the preference measures being bimodally distributed. In order to

determine if the difference in explained variance between EQ3 and EQ4 was

statistically significant, the unstandardised residuals from each model were

compared using a chi-square analysis. The greater the distribution of residual

values, the less effective the model at determining a linear relationship between

the response and predictor variables. The results of the chi-square analysis were

significant, meaning that the regression model predicting the WHQ score (EQ4)

accounted for significantly more of the variance in the data. Therefore, per-

formance variables alone are better able to predict the composite score on the

WHQ than the composite score on the WBT.

DISCUSSION

The aim of this study was to examine the relationship between measures of hand

preference and measures of hand performance. By establishing this relationship,

it might be possible to develop effective tools to measure hand preference that

would be easy to administer and that involve elements of performance to remove

some of the subjectivity that is inherent in preference-based tasks.

Our previous work (Brown et al., 2004) found that the Wathand Box Test had

the highest correlation with the handedness score of the Waterloo Handedness

Questionnaire, and was the best performance-based predictor of hand preference

of all the variables scrutinised. Because the Wathand Box Test measures hand

TABLE 4Unstandardised and standardised coefficients for EQ3 and EQ4

Unstandardised coefficients Standardised coefficients

Performance

variable

GP

(place)

FT GS AP GP

(place)

FT GS AP

EQ3 73.729 2.655 2.265 72.546 70.373 0.194 0.236 70.187

EQ4 71.515 1.413 0.649 71.300 70.391 0.266 0.174 70.246

PERFORMANCE MEASURES OF HANDEDNESS 11

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 0

1:58

07

Oct

ober

201

4

preference based on the performance of individual tasks, it is not surprising that

it shows the highest correlation with the traditional preference measure. Simi-

larly, the current work has shown that the Wathand Box Test was indeed the best

predictor of hand preference: it had the highest correlation with the handedness

score of the Waterloo Handedness Questionnaire, lending further support to the

proposition that the Wathand Box Test is a robust alternative to traditional

preference measures such as questionnaires, and may prove more accurate when

working with special populations such as children (Bryden et al., 2000b).

Despite the high correlation between the WBT and the WHQ, the current

analysis has shown that performance-based measures more accurately predict

handedness as measured by the WHQ than they predict handedness as measured

by the WBT. Comparing the standardised coefficients for each of these

regression equations reveals the relative weightings of the performance-based

tasks in each of the equations. It appears that EQ4 (using performance measures

to predict the score on the WHQ test) has greater emphasis on the Grooved

Pegboard, finger-tapping, and Annett pegboard tasks (revealed by higher stan-

dardised coefficients for each of these variables), whereas EQ3 (using perfor-

mance measures to predict the score on the WBT test) emphasises the grip

strength task (revealed by a higher standardised coefficient for this variable). It

is interesting that the performance-based tasks requiring the most skill are more

heavily weighted on the questionnaire task, whereas grip strength, which is the

least skilled performance task due to its lack of manipulative or visuomotor

components, is more heavily weighted on the Wathand Box Test. The fact that

the questionnaire examines many skilled elements while the grip strength

measure is not a skilled activity might explain why the grip strength measure did

not contribute significantly to the equation predicting the questionnaire score

using all of the performance measures (including the Wathand Box Test) (EQ1).

Although the Wathand Box Test does involve skilled activities such as turning a

key in a small lock and using a screwdriver, the questionnaire presents more

skilled activities such as using a needle to sew, using tweezers, etc. Future

analysis examining the individual components of each of the performance-based

tasks will help to explain the relationship between the skilled and unskilled tasks

included on the questionnaire and in the Wathand Box Test, and will serve to

assist in developing preference-based tools that will incorporate both skilled and

unskilled activities. As Corey et al. (2001) suggest, it is difficult to establish a

link between preference and performance measures unless the factors that

underlie these measures are the same for handedness groups and it seems as

though the relative lack of the Wathand Box Test to incorporate skilled activities

as compared with the questionnaire is the reason why traditional performance-

based measures better predict the score on the questionnaire than the score on

the Wathand Box Test.

Peters (1998) suggests that in order to accurately develop a predictive model

of handedness, the performance tasks used must emphasise many different facets

12 BROWN ET AL.

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 0

1:58

07

Oct

ober

201

4

of hand performance. Indeed, our predictive model of handedness combined

several types of performance. As has been stated earlier, the Wathand Box Test

serves as an observational measure of preference. Each of the performance tasks

used focuses on a slightly different aspect of hand performance. For instance,

while finger tapping requires good control of the distal musculature, the Annett

pegboard requires that the participant have good control of proximal muscles as

well as distal muscles, and the Grooved Pegboard not only requires strong skill

of both proximal and distal musculature, but it also places high importance on

the use of vision to place the pegs in the appropriate orientation. This result also

supports the work by Corey et al. (2001) by incorporating a combination of

preference- and performance-based measures into the predictive model of

handedness.

In conclusion, the current study developed a predictive model of handedness

that accurately predicted handedness by incorporating both preference and

performance measures of handedness, and including consideration of the dif-

ferent components involved with movement control. Rather than simply linking

the laterality quotients from the performance-based measures with the pre-

ference-based scores, future analysis will study the link between the components

of hand performance such as visuomotor requirements, fine motor control, and

strength contributions of the performance tasks, how these components are

related to the two measures of hand preference, and whether or not gender

differences exist. By examining the components of the standard performance-

based tasks, it will be possible to develop effective performance-based measures

of preference that incorporate movement components from all domains.

Manuscript received 23 November 2004

Revised manuscript received 3 February 2005

PrEview proof published online 21 June 2005

REFERENCES

Annett, M. (2002). Handedness and brain asymmetry: The right shift theory. Hove, UK: Psychology

Press.

Brown, S. G., Roy, E. A., Rohr, L. E., Snider, B. R., & Bryden, P. J. (2004). Preference and

performance measures of handedness. Brain and Cognition, 55, 283±285.

Bryden, M. P., Bulman-Fleming, M. B., & MacDonald, V. (1996). The measurement of handedness

and its relation to neuropsychological issues. In D. Elliott & E. A. Roy (Eds.), Manual asym-

metries in motor performance (pp. 57±81). Boca Raton, FL: CRC Press.

Bryden, P. J., Pryde, K. M., & Roy, E. A. (2000a). A performance measure of the degree of hand

preference. Brain and Cognition, 44, 402±414.

Bryden, P. J., Pryde, K. M., & Roy, E. A. (2000b). A developmental analysis of the relationship

between hand preference and performance: II. A performance-based method of measuring hand

preference in children. Brain and Cognition, 43(1±3), 60±64.

Bryden, P. J., & Roy, E. A. (1999). Spatial task demands affect the extent of manual asymmetries.

Laterality, 4(1), 27±37.

PERFORMANCE MEASURES OF HANDEDNESS 13

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 0

1:58

07

Oct

ober

201

4

Bryden, P. J., & Roy, E. A. (2005). A new method of administering the Grooved Pegboard Test:

Performance as a function of handedness and sex. Brain and Cognition, 58, 258±268.

Corey, D. M., Hurley, M. M., & Foundas, A. L. (2001). Right and left handedness defined: A

multivariate approach using hand preference and hand performance measures. Neuropsychiatry,

Neuropsychology and Behavioural Neurology, 14, 144±152.

Lafayette Instrument. (1989). Instruction manual for the 32025 Grooved Pegboard Test. Lafayette

IN: Lafayette Instrument.

Peters, M. (1998). Description and validation of a flexible and broadly usable handedness ques-

tionnaire. Laterality, 3(1), 77±96.

APPENDIX

Questions included in the Waterloo HandednessQuestionnaire

Each of the questions below offers five possible responses (as described in the Apparatus and

procedures section of this paper): RA (right always), RU (right usually), EQ (equal), LU (left

usually), and LA (left always).

1. Which hand would you use to spin a top?

2. With which hand would you hold a paintbrush to paint a wall?

3. Which hand would you use to pick up a book?

4. With which hand would you use a spoon to eat soup?

5. Which hand would you use to flip pancakes?

6. Which hand would you use to pick up a piece of paper?

7. Which hand would you use to draw a picture?

8. Which hand would you use to insert and turn a key in a lock?

9. Which hand would you use to insert a plug into an electrical outlet?

10. Which hand would you use to throw a ball?

11. In which hand would you hold a needle while sewing?

12. Which hand would you use to turn on a light switch?

13. With which hand would you use the eraser at the end of a pencil?

14. Which hand would you use to saw a piece of wood with a hand saw?

15. Which hand would you use to open a drawer?

16. Which hand would you turn a doorknob with?

17. Which hand would you use to hammer a nail?

18. With which hand would you use a pair of tweezers?

19. Which hand do you use for writing?

20. Which hand would you turn the dial of a combination lock with?

21. Is there any reason (e.g. injury) why you have changed your hand preference for any of the above

activities?

YES NO (circle one) Explain.

22. Have you ever been given special training or encouragement to use a particular hand for certain

activities?

YES NO (circle one) Explain.

14 BROWN ET AL.

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 0

1:58

07

Oct

ober

201

4