Abstract
There is a consistent but poorly understood finding that self-report and behavioral measures of impulsivity are weakly correlated or uncorrelated. There are many possible explanations for this observation, including differences in how these instruments are administered and scored. The present study examined the utility of alternative scoring algorithms for self-report measures that aim to identify participants' peak impulsivity (or self-control), informed by estimates of item difficulty from Item Response Theory (IRT) analyses. College students were administered self-report questionnaires (Zuckerman Sensation Seeking Scale [ZSS], Barratt Impulsiveness Scale [BIS-11], behavioral measures related to risk-taking and impulsivity (Balloon Analog Risk Task [BART], Experiential Discounting Task [EDT]), and the substance use module of a clinical interview (past-six-month alcohol and marijuana use). IRT analyses were conducted on self-report measures to estimate item difficulty. Scoring algorithms ranked items by difficulty and scored items based on consecutive items endorsed or denied. A maximal scoring algorithm increased the concordance between the BIS-11 and BART (r = 0.08 vs. −0.07), but there was no evidence of increased incremental validity for substance-use. Findings suggest that methodological factors may help explain the poor concordance of self-report and behavioral measures of impulsivity, but the magnitude of these correlations remained quite small. Further, alternative scoring algorithms were correlated with substance use but did not explain any variance that was distinct from typical algorithms. Future directions are discussed for elucidating the discrepancy between self-report and behavioral measures of impulsivity-related constructs, such as using large self-report item pools to develop computer adaptive tests.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have