Het probleem van meetinvariantie bij het vergelijken van subgroepen op basis van somscores
Arne De Boeck, Wim Hardyns, Lieven J. R. Pauwels · Tijdschrift voor Criminologie · 2014
The problem of measurement invariance when using sum scores to make group comparisons. A case study on the measurement of avoidance behavior Using summated scale scores to make group comparisons is only meaningful if one can assume that the scale measures attribute in the same way in each of the groups involved in the comparison. This assumption is called measurement invariance. This contribution discusses the use of modern statistical measurement models to test this assumption and to verify the consequences of a possible violation for the comparison of group means. In the empirical part of the contribution the authors illustrate their account by examining whether a scale assessing avoidance behavior – commonly used in fear of crime research – is invariant across gender and age groups.