Revising the BIS/BAS Scale to study development: Measurement invariance and normative effects of age and sex from childhood through adulthood.

Authors

David Pagliaccio, Program in Neuroscience, Washington University in St. Louis
Katherine R Luking, Program in Neuroscience, Washington University in St. Louis
Andrey P Anokhin, Department of Psychiatry, Washington University in St. Louis
Ian H Gotlib, Department of Psychology, Stanford University
Elizabeth P Hayden, Department of Psychology, Western UniversityFollow
Thomas M Olino, Department of Psychology, Temple University
Chun-Zi Peng, Department of Psychiatry, Washington University in St. Louis
Greg Hajcak, Department of Psychology, Stony Brook University
Deanna M Barch, Department of Psychology, Washington University in St. Louis

Document Type

Article

Publication Date

4-1-2016

Journal

Psychological assessment

Volume

28

Issue

4

First Page

429

Last Page

442

URL with Digital Object Identifier

10.1037/pas0000186

Abstract

Carver and White's (1994) Behavioral Inhibition System/Behavioral Activation System (BIS/BAS) Scales have been useful tools for studying individual differences in reward-punishment sensitivity; however, their factor structure and invariance across development have not been well tested. In the current study, we examined the factor structure of the BIS/BAS Scales across 5 age groups: 6- to 10-year-old children (N = 229), 11- to 13-year-old early adolescents (N = 311), 14- to 16-year-old late adolescents (N = 353), 18- to 22-year-old young adults (N = 844), and 30- to 45-year-old adults (N = 471). Given poor fit of the standard 4-factor model (BIS, Reward Responsivity, Drive, Fun Seeking) in the literature, we conducted exploratory factor analyses in half of the participants and identified problematic items across age groups. The 4-factor model showed poor fit in our sample, whereas removing the BAS Fun Seeking subscale and problematic items from the remaining subscales improved fit in confirmatory factor analyses conducted with the second half of the participants. The revised model showed strict invariance across age groups and by sex, indicating consistent factor structure, item loadings, thresholds, and unique or residual variances. Additionally, in our cross-sectional data, we observed nonlinear relations between age and subscale scores, where scores tended to be higher in young adulthood than in childhood and later adulthood. Furthermore, sex differences emerged across development; adolescent and adult females had higher BIS scores than males in this age range, whereas sex differences were not observed in childhood. These differences may help us to understand the rise in internalizing psychopathology in adolescence, particularly in females. Future developmental studies are warranted to examine the impact of rewording problematic items.