European Survey Research AssociationEuropean Survey Research Association
 
Home About us Membership Conferences Journal Minutes Contact

Conferences

Conferences


Composite Scores 2008


Miniconference on Composite Scores (Barcelona, 13–15 February 2008)

In comparative research often means and relationships of variables are compared across countries without much attention to the problem of the equivalence of the measurement instruments across countries. Methodologists have drawn the attention to the problem of comparability of measurement instruments across countries. Special attention is given in this context to concepts by postulation or constructs measured by several indicators. Typical examples are concepts like Political efficacy, Political Trust, Social trust, Human values but also more objective variables like Political participation, Media use etc.

A fundamental distinction is made between concepts with formative indicators and concepts with reflective indicators. Formative indicators define the concept by postulation while reflective indicators are determined by the latent variable (constructs) behind these indicators. Most of the time composite scores are computed as indices for these concepts by postulation and the means and relationships of these indices are compared across countries without attention to the equivalence problem.

In the methodological research equivalence of the measures has been studied by specification of different requirement for invariance of the measurement models across countries for different purposes. However it turns out that these requirements are still open for debate. One does not make a distinction between cognitive processes (the interpretation of the questions) and response processes (measurement error) , the way the invariance has been tested is also rather questionable, statistical significance is not the same as substantive relevance and not enough attention has been paid to the choice of the weights for the computation of the indices or composite scores.

In this conference we have discussed the different steps in the process to construct indices for comparative research on the basis of empirical data of the ESS. Especially the index construction for the concepts of the core questionnaire of the ESS have been elaborated.

The following issues were discussed for all concepts:

  • Operationalization by measurement models
  • Estimation of measurement error
  • Testing Cross cultural comparison
  • Optimal and equal weights
  • Quality of the index
  • External validity
  • Comparison of means and relationships across counties
  • Relationships between cross countries and cross time analysis

The following papers have been presented:

Willem Saris (ESADE):

Quality requirements for composite scores for comparative Research (download this presentation)
Measurement models (download this presentation)

Eldad Davidov (U. of Mannheim):

A Cross-Country and Cross-Time
Comparison of the Human Values. Measurements with the Second Round of the European Social Survey (download this presentation)

Desiree Knoppen (ESADE):

Schwartz Human Values with special attention to operationalization (download this presentation)

Cristiana Martini (U. of Modena):

Measurement of income: an IRT model

Jaak Billiet (KUL):

Index for religiosity (download this presentation)

Akos Munnich (U. of Debrecen):

Comments on some measurement models (download this presentation)
The role of the reference point and orientation parameter in subjective value judgment (download this presentation)

Daniel Oberski (ESADE):

A different way of analysing complex multiple group models like split ballot MTMM with an application to political trust

William van der Veld (RU):

Composite scores for social Trust: Special attention to testing cross cultural comparability (download this presentation)
Model Evaluation with JRule (download this presentation)
An Alternative Invariance Test (download this presentation)

Patrick Sturgis (U. of Surrey):

Assessing the Validity of Generalized Trust Questions: What kind of trust are we measuring?

Filomena Maggino (U. of Florence):

Choice of subjective weights for subjective variables (download this presentation)

Lluis Coromina (ESADE):

Composite scores based on Media use: special attention to Quality of Composite Scores and External Validation (download this presentation)

Besides the mentioned presenters the following people participated in the conference:
ESADE/ Universitat Ramon Llull: Joan Manuel Batista, Wiebke Weber
Universitat Pompeu Fabra: Clara Riba, Albert Satorra, Nick Longford
Universitat de Girona: Germa Coenders
University of Giessen: Peter Schmidt
European Social Survey: Eric Harrison