School self-evaluation instruments and cognitive validity. Do items capture what they intend to?

Abstract

ABSTRACTSchool self-evaluation (SSE) often makes use of questionnaires in order to sketch a picture of the school. How respondents cognitively process questionnaire items determines the validity of SSE results. Still, one readily assumes that respondents interpret and answer items as intended by the instrument developer (referred to as cognitive validity), but it remains unclear whether they do. This study tested an exemplary SSE instrument by focusing on the extent to which SSE results are cognitively valid, and on the extent to which differences in cognitive validity can be attributed to respondents and/or items. Cognitive interviews with 20 participants made respondents’ answering processes manifest. Results show that, overall, fewer than 50% of respondents’ processes of interpreting and elaborating on items are cognitively valid. Cross-classified multilevel analyses indicate that various hierarchical levels, respondents and items, are significant in explaining differences in cognitive validity, but not for all stages of the answering process.