Tools for assessing the psychometric adequacy of latent variables in conservation research.

Abstract

Conservation psychology has a history of measuring variables that cannot be seen (e.g., values, attitudes, norms). Such latent variables are critical drivers of human action and are often measured using responses to survey questions. Tools for establishing the psychometric adequacy of unobservable, latent variables has been a century-long pursuit and challenge for quantitative psychologists and statisticians. Fundamental questions at the heart of this challenge include is what is claimed to be measured (validity) being measured and is measurement consistent (reliability)? We examined common methods used to establish the validity and reliability of psychometric instruments. Through a case study of anglers in Texas, we investigated the protocols and metrics used to evaluate the measurement of latent variables. The indicators we tested (identity, awareness of consequences, ascription of responsibility, and personal norms) validly and reliably assessed latent variables. Our findings also illustrated decision protocols (e.g., discriminant validity, convergent validity, internal consistency) involved in assessing the psychometric adequacy of latent variable indicators. The ability to correctly identify significant relationships among unobserved variables and their influence on human action is directly tied to the adequacy of measurement. In an era of instability and change that threatens social-ecological systems worldwide, the need for accuracy and precision in conservation social science has never been greater. Research that employs flawed measures has potential to lead to erroneous conclusions and undermine conservation and biodiversity protection.