SMABS 2004 Jena University
SMABS 2004 Home Organization About Jena Sponsors Links Imprint / Contact SMABS Home

European Association of Methodology

Department of methodology and evaluation research

Jena University

Contributions: Abstract

How invariant are measures of human values? Results from a population survey

Holger Steinmetz Peter Schmidt
Andrea Tina-Booh Sigrid Wieczorek
University of Giessen

Whenever groups (e.g. gender, age or even cultures) are to be compared in terms of one or more constructs or correlations, researchers assume that the measurement parameters (i.e. factor loadings, measurement errors etc.) are the same in the groups.

This measurement invariance is seen as an important prerequisite for the comparisons (see Meredith, 1993). If not met, violations of these assumptions (e.g. unequal factor loadings) can lead to ambiguous or even wrong conclusions. For instance, unequal factor loadings can mean that one is measuring different constructs in the groups. In the same way, unequal item intercepts can lead to ambiguous mean comparisions. Thus, at least partial measurement invariance has to be proven before groups can be compared.

In our presentation we show results from a population survey with 1,600 subjects. Using this sample we tested the human value theory of Shalom Schwartz for measurement invariance regarding the sociodemographical groups of gender, age and education. This was done using the confirmatory factor analytic framework where measurement invariance can be successively tested via cross-groups equality constraints. Our analyses also included estimation of latent means.

The theory of Schwartz and its measurement has become a widely acknowledged standard in the research of human values. It postulates the cross-cultural existence of ten human values. These are achievement, hedonism, self-direction, benevolence, conformity, security, stimulation, power, tradition, and universalism. We will provide evidence that - with few exceptions - measurement invariance could be established.