Abstract

Reactivity takes place when being measured or categorized affects a subject’s attitudes and behaviour to such an extent that it affects results in their (subsequent) measurement or categorization. Think, for example, of higher education institutions hiring temporary staff with the sole purpose of receiving a higher score on performance indicators. Or consider the effects that the act of filling out a quality of life survey may have on how a respondent ranks their quality of life compared to others. Although researchers and methodologists commonly think of reactivity as a source of measurement error, this presentation shows that some quite extreme reactive changes may be legitimate, in the sense that the measurement results before and after the shift are both accurate to the measurand. While the presentation starts from recent considerations in the psychometric literature on response shift, it shows that these considerations cannot be extended easily to social scientific phenomena. To fill this gap in the literature, the presentation presents a tailored inventory of the different types of reactivity in social scientific measurement, as well as a conceptual framework for distinguishing which type of reactivity is legitimate, based on specific properties we believe the measurand to have.