Psychometric properties and Longitudinal Measurement Invariance of the Spanish version of the Alcohol Expectancies Questionnaire Short Form among young adult binge drinkers

• Published in: The American journal of drug and alcohol abuse 2024-09, Vol.50 (5), p.631-642
• Main Authors: González-Ponce, Bella M, Pilatti, Angelina, Rivarola Montejano, Gabriela, Bravo, Adrian J, Fernández Calderón, Fermín
• Format: Article
• Language: English
• Subjects: Adolescent; Adult; Alcohol abuse; Alcohol use; Binge Drinking - psychology; Cognitive impairment; Confirmatory factor analysis; Disinhibition; Drinking behavior; Emotional states; Factor Analysis, Statistical; Female; Humans; Longitudinal Studies; Male; Measurement; Measures; Motivation; Predictive validity; Protective factors; Psychometrics; Quantitative psychology; Questionnaires; Reliability; Reproducibility of Results; Social facilitation; Spain; Spanish language; Suitability; Surveys and Questionnaires - standards; Time; Validity; Young Adult; Young adults
• ISSN: 0095-2990; 1097-9891; 1097-9891
• DOI: 10.1080/00952990.2024.2377256

Longitudinal Measurement Invariance (LMI) is critically important to evaluate changes in alcohol expectancies over time. However, past research has not explored the longitudinal properties of the Spanish Expectancy Questionnaire Short Form (EQ-SF). To examine the reliability, sources of validity (structural, invariance across sex, and concurrent validity), and LMI of the Spanish EQ-SF among young adults who engage in binge drinking. Participants (  = 279; 48.4% female) completed the EQ-SF and, two months later, completed it again along with measures of alcohol use, drinking motives, and protective behavioral strategies (PBS). We performed confirmatory factor analysis for structural validity and measurement invariance analysis for longitudinal and sex stability. The eight-factor intercorrelated model (i.e. social facilitation, fun, sexual disinhibition, tension reduction, antisocial effects, negative emotional states, negative physical effects, and cognitive impairments) provided the best fit ( (df) = 497.29(224), = .962, = .064, = .049). This model was invariant across sex and time. Reliability coefficients (Ordinal alpha) for each dimension were consistently strong at both time points (from .72 to .93 at T1 and .73 to .91 at T2). Positive alcohol expectancies at baseline were positively related to alcohol use and drinking motives and negatively related to PBS at follow-up, demonstrating predictive validity. Our results support the temporal invariance of the EQ-SF scores among Spanish young adults who engage in binge drinking. The evidence supports the suitability of this measure for accurately assessing changes in alcohol expectancies over time in interventions aimed at preventing binge drinking in young adults.