Optimal assessment of protective behavioral strategies among college drinkers: An item response theory analysis.

Abstract

College student alcohol use and related consequences continue to warrant significant concern. Extant research demonstrates protective behavioral strategies (PBS; self-regulatory strategies that can be employed before, during, or after drinking to prevent intoxication or negative consequences) have promise for preventative interventions. Variations in conceptualization and measurement of the construct limit generalization of PBS research. To advance generalization of PBS research, there is a need for a brief, comprehensive, content valid, PBS measure that demonstrates equivalence in measurement across demographically diverse college students. The present study aimed to develop a psychometrically robust measure of PBS using item response theory (IRT) to address measurement and methodological issues including identifying optimal response anchors and items that represent the full range of the PBS construct for use with college men and women from different racial backgrounds. Participants were 503 college students enrolled in a midsized university in the Southern United States. IRT and differential item functioning (DIF) analyses of 68 PBS items extracted from 6 existing PBS measures resulted in a 20-item instrument, the Protective Drinking Practices Scale, with items that functioned equivalently for White and Black/African American college men and women. The measure also demonstrated good internal and external validity. Widespread use of this measure will help enhance the comparability of findings in PBS research, allowing for more targeted and impactful research on PBS as a mechanism of change. Suggestions for future research are provided.

Psychometric validation of the Protective Drinking Practices Scale in college students across the United States.

“…The 20-item PDPS (Martin et al, 2020) was used to evaluate participants’ use of protective behavioral strategies. Participants were asked to rate how often they used specific protective strategies in the past 30 days when drinking or partying on a scale from 1 ( never ) to 6 ( always ).…”

“…To evaluate the PDPS in the present sample, we first calculated correlations between PDPS scores and other alcohol-related measures to evaluate construct validity. Next, we replicated the item response theory (IRT) analysis from the original development sample (Martin et al, 2020) with the current sample. We then evaluated whether the PDPS items functioned similarly in the current sample compared to the previous sample from the Martin et al (2020) study.…”

“…Next, we replicated the item response theory (IRT) analysis from the original development sample (Martin et al, 2020) with the current sample. We then evaluated whether the PDPS items functioned similarly in the current sample compared to the previous sample from the Martin et al (2020) study. Finally, following suggestions from Martin et al (2020), we evaluated a three-item version of the scale.…”

“…Although the PDPS shows promise, the development and initial evaluation of the measure were conducted with one sample from one university and thus is in its infancy. Further investigation of the measure is needed to attempt to replicate Martin et al (2020) findings in more diverse groups than the development sample and further evaluate the psychometric properties of the measure. Thus, the purpose of this study was (a) to test the PDPS in a nationally collected sample of college student drinkers and (b) to compare the functioning of the PDPS in the present sample to that of the original sample in Martin et al (2020) work.…”

“…Further investigation of the measure is needed to attempt to replicate Martin et al (2020) findings in more diverse groups than the development sample and further evaluate the psychometric properties of the measure. Thus, the purpose of this study was (a) to test the PDPS in a nationally collected sample of college student drinkers and (b) to compare the functioning of the PDPS in the present sample to that of the original sample in Martin et al (2020) work. Specifically, we examined whether the functioning of the PDPS items in the current sample and the Martin et al (2020) sample was comparable, explore the PDPS factor structure, and evaluate the convergent validity of the PDPS with common alcohol use measures.…”