These tools will no longer be maintained as of December 31, 2024. Archived website can be found here. PubMed4Hh GitHub repository can be found here. Contact NLM Customer Service if you have questions.


PUBMED FOR HANDHELDS

Search MEDLINE/PubMed


  • Title: Social determinants of alcohol and cigarette use by race/ethnicity: Can we ignore measurement issues?
    Author: Lopez-Vergara HI, Rosales R, Scheuermann TS, Nollen NL, Leventhal AM, Ahluwalia JS.
    Journal: Psychol Assess; 2020 Nov; 32(11):1075-1086. PubMed ID: 32924524.
    Abstract:
    Psychometric critiques of cross-cultural research emphasize testing whether instruments measure the same construct across cultural groups. We tested for measurement invariance (by race/ethnicity) of instruments used to evaluate the relationship between alcohol and tobacco use with perceived discrimination and socioeconomic status (SES). Tests of psychometric equivalence across race/ethnicity focused on: the latent organization of constructs (configural invariance); if observed indicators have equal factor loadings or "true score" variance (metric invariance); and whether manifest indicators change uniformly contingent on change in the latent variable (scalar invariance). A cross-sectional survey of 2,376 cigarette smokers (794 Black, 786 Latinx, 796 White; mean age = 43 [SD = 12]; 58% female) was recruited via an online research panel. Discrimination was indicated by self-report; SES was indicated by self-reported education, employment, income, and the "SES Ladder;" alcohol use was indicated by frequency and typical quantity of drinking, and frequency of heavy drinking; tobacco use was indicated by frequency of smoking, cigarettes per smoking day, and time to first cigarette. All instruments demonstrated configural invariance; either full metric invariance (alcohol and discrimination) or partial metric invariance (tobacco and SES); and all constructs demonstrated partial scalar invariance. Results support psychometric critiques; for example, all of the SES indicators violated assumptions of classical measurement theory for valid between group comparisons. All of our instruments displayed some degree of systematic bias in measurement across race/ethnicity. Studies testing ethnic/racial differences may need to move beyond classical measurement theory, and may benefit from using statistical approaches that can test for (and model) bias in measurement. (PsycInfo Database Record (c) 2020 APA, all rights reserved).
    [Abstract] [Full Text] [Related] [New Search]