Elina Vessonen: Concept Formation in Psychometrics
Psychometrics is one of the main approaches to measurement in social sciences. Psychometric measures i) are meant to track unobservable attributes (e.g. intelligence and personality), ii) typically take the form of questionnaires or tests, and iii) are validated in light of statistical tests of properties such as reliability and model-fit. In this paper I evaluate these measures from the perspective of concept formation. I argue that although psychometric measures are typically treated as if they track non-operationally characterized, quantitative concepts, the usual validation process does not ensure that psychometric measures in fact track such concepts. I'll explain why such a mismatch between "concepts-assumed" and "concepts-validated" is problematic, and discuss ways to resolve the mismatch.
Psychometrics is one of the main approaches to measurement in social sciences. Psychometric measures i) are meant to track unobservable attributes (e.g. intelligence and personality), ii) typically take the form of questionnaires or tests, and iii) are validated in light of statistical tests of properties such as reliability and model-fit. In this paper I evaluate these measures from the perspective of concept formation. I argue that although psychometric measures are typically treated as if they track non-operationally characterized, quantitative concepts, the usual validation process does not ensure that psychometric measures in fact track such concepts. I'll explain why such a mismatch between "concepts-assumed" and "concepts-validated" is problematic, and discuss ways to resolve the mismatch.