Scales are often used to measure different aspects of mental health within social surveys. Following this some form of group comparison is common, such as comparisons in means or prevalence's by gender, ethnicity, or country. However, meaningful comparisons between groups can only happen when the assumption of measurement invariance (also known as measurement equivalence) holds. Measurement invariance suggests that the measure functions in the same way within different groups, i.e it is working in the same way and measuring the same construct with the same degree of accuracy within those groups.

Here I use data from several European countries to assess whether the assumption of measurement invariance is met. I show the extent to which the scales that measure depressive mood (CES-D), self esteem (Rosenberg) and antisocial behaviour perform differentially within different countries.