4.2.3 data quality, composite indicators and aggregation 1 data quality, composite indicators and...

17
1 4.2.3 Data Quality, Composite Indicators and Aggregation DATA QUALITY, COMPOSITE INDICATORS AND AGGREGATION UPA Package 4, Module 2

Upload: ira-poole

Post on 18-Dec-2015

223 views

Category:

Documents


1 download

TRANSCRIPT

14.2.3 Data Quality, Composite Indicators and Aggregation

DATA QUALITY, COMPOSITE INDICATORS AND AGGREGATION

UPA Package 4, Module 2

24.2.3 Data Quality, Composite Indicators and Aggregation

Data Quality, Composite Indicators and Aggregation

• Data errors • Less is more; benefit-cost of data• Data cleaning• Composite indicators• Introduction exercise 4.2.3 Exploring Data sets• Introduction exercise 4.2.4 Aggregation

34.2.3 Data Quality, Composite Indicators and Aggregation

Data Errors

• Biased data• Outliers, error or extreme value• Sample too small• Too much precision or regularity (too good to be true)• Missing values• Inconsistencies

44.2.3 Data Quality, Composite Indicators and Aggregation

Less is more; Benefit and Cost of Data

• Quality (full coverage and maintenance)

• Quantity (many variables but missing values and outdated)

Physical Characteristics

of a building

Ownership Characteristics

of a building

54.2.3 Data Quality, Composite Indicators and Aggregation

Benefit and Cost of Data

• Data Benefit and Costs

Strategy and clear objectives of developing databases

Data (and functionalities) requirement study

Data benefit, the value of information and quantification– costs reduction, effectiveness/priorities of (public)

resource allocation– transparency, awareness, involvement

Data costs high (acquisition, editing, conversion, updates, maintenance)

64.2.3 Data Quality, Composite Indicators and Aggregation

Benefit and Cost of Data

Primary and secondary data, data sharing• Primary, ad-hoc, single use of data, (too) expensive • Secondary matching with requirement for poverty studies• Combination of existing data and samples• Data collection embedded into institutional settings, from data

projects to data processes

74.2.3 Data Quality, Composite Indicators and Aggregation

Composite Indicators

• Poverty without reliable income data• Slums• Composite Indicator

Human Development Indicator, Poverty Index• Proxy indicators (consumption / income)

84.2.3 Data Quality, Composite Indicators and Aggregation

Composite Indicators

94.2.3 Data Quality, Composite Indicators and Aggregation

Aggregation

• Aggregate cases into a single summary case• Break variable defines a group and create one case

e.g. neighborhood• Aggregate functions

Summary, fractions

104.2.3 Data Quality, Composite Indicators and Aggregation

Small Area Statistics

• Limited (existing) data, limited funds for data collection• Sample survey and auxiliary data sets (+ analytical skills) =

small area statistics• Developing a model to identify the relationship between the

survey and the auxiliary data more reliable estimates can be made and the possibilities to extrapolate to areas not covered by a household survey

114.2.3 Data Quality, Composite Indicators and Aggregation

Introduction Exercise 4.2.3

Exploring Datasets

Classifying interval data (number of foreigners, income, family size) into meaningful groups (e.g. low income, medium income, high income).

Create cross tables and analyze relationships between these ordinal data sets.

124.2.3 Data Quality, Composite Indicators and Aggregation

Introduction Exercise 4.3.2

Count

incomecl Total

Low Medium High Very High

housecl Low 37 50 25 0 112

Medium 51 177 182 0 410

High 1 1 8 6 16

Total 89 228 215 6 538

Symmetric Measures

.309 .044 7.517 .000 c

.288 .042 6.956 .000 c

538

Pearson's RInterval by IntervalSpearman CorrelationOrdinal by Ordinal

N of Valid Cases

ValueAsymp.

Std. Error a Approx. T

b Approx. Sig.

Not assuming the null hypothesis.a. Using the asymptotic standard error assuming the null hypothesis.b. Based on normal approximation.c.

Cross table (mean income x mean house value)

Municipalities in the Netherlands

134.2.3 Data Quality, Composite Indicators and Aggregation

Introduction Exercise 4.2.3

Aggregation

Central Bureau of Statistics of The Netherlands three main spatial units:Municipality (n=538)Districts (n=2382)Neighbourhoods (n=10737) Aggregation, summarizing data, why and what Spatially homogenous versus heterogeneous variablesWhich statistics to use (mean or other statistical figures)Simple and weighted aggregates

144.2.3 Data Quality, Composite Indicators and Aggregation

Introduction Exercise 4.3.2

154.2.3 Data Quality, Composite Indicators and Aggregation 144.2.3 Data Quality, Composite Indicators and Aggregation

Introduction Exercise 4.3.2

164.2.3 Data Quality, Composite Indicators and Aggregation

Introduction Exercise 4.2.3

174.2.3 Data Quality, Composite Indicators and Aggregation

Introduction Exercise 4.2.3