11/14/2024 Oliver Whilhelm: Exhausting the Utility of Psychological Measures through Indicator Sampling
CCN Brown Bag
Oliver Whilhelm, PhD
Professor of Psychological Assessment
Ulm University
Date: Thursday, November 14, 2024
Time: 12:10PM-1PM
Location: 316 Wilson Hall
Exhausting the Utility of Psychological Measures through Indicator Sampling
Data reduction is commonly used to condense information inherent in measures of human abilities, traits, and behaviors. Requirements of these techniques are often not met. They show a lack of unidimensionality and unwanted residual correlations of indicators after controlling for latent variables. In this presentation we illustrate how this lack of data-quality can be a bliss and not a curse. For measures of knowledge we illustrate the benefits of domain coverage as opposed to maximizing consistency in measures. Additionally, we show how item residuals contribute meaningfully to predict outcomes. For measures of self-reported personality traits, we show that minimizing factor saturation can yield better outcome prediction than striving for simple structure.
Questions? Contact Isabel Gauthier