Affiliation:
1. University of Vienna, Department of Methods in the Social Sciences, Vienna, Austria
Abstract
Assessing measurement invariance (MI) is an important cornerstone in
establishing equivalence of instruments and comparability of constructs.
However, a common concern is that respondent differences in acquiescence
response style (ARS) behavior could entail a lack of MI for the measured
constructs. This study investigates if and how ARS impacts MI and the level
of MI achieved. Data from two representative samples and two popular short
Big Five personality scales were analyzed to study hypothesized ARS
differences among educational groups. Multiple-group factor analysis and the
random intercept method for controlling ARS are used to investigate MI with
and without controlling for ARS. Results suggest that, contrary to
expectations, controlling for ARS had little impact on conclusions regarding
the level of MI of the instruments. Thus, the results suggest that testing MI
is not an appropriate means for detecting ARS differences per se.
Implications and further research areas are discussed.
Publisher
National Library of Serbia
Cited by
7 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献