Differences in Depressive Symptom Presentation between Latino and non-Latino White Youths on the Center for Epidemiologic Studies Depression Scale (CES-D): A Moderated Nonlinear Factor Analysis (MNLFA) Approach

Public Deposited
Analytics

Downloadable Content

Download PDF
Request Version for Screen Reader
Creator
  • Salcedo, Stephanie
    • College of Arts and Sciences, Department of Psychology and Neuroscience
Abstract
  • Most mental health measures have been validated with English speaking and majority White samples (Guillemin et al., 1993; Vega & Rumbaut, 1991). Despite guidelines recommending formal cultural adaptation to ensure that scale content still measures what it intended and provide accurate group comparisons, such adaption is rare (Borsboom, 2006; Chen, 2008). Latinos are the largest minority in the US, making it imperative to understand how cultural factors can influence mood symptom endorsement and conceptualization. This study examined the Center for Epidemiologic Studies Depression Scale (CES-D) for differential item functioning (DIF). Secondary analyses of Latino (n=3,208) and non-Latino White youths (n=9,919) from the National Longitudinal Study of Adolescent Health (Add Health). Moderated nonlinear factor analysis (MNLFA) examined ethnicity DIF, while accounting for other covariates (gender, age, parental education, acculturation). An exploratory aim compared Latino subgroups separately (Mexicans, Puerto Ricans, and Cubans) to assess subgroup differences in the item intercepts and factor loadings. In the initial model (only incorporating ethnicity), at similar levels of depression, Latinos were more likely to score higher on being bothered by things, feeling unable to shake the blues, talking less, and feeling like a failure. However, Latinos were less likely to endorse having sleep problems, feeling like others were unfriendly, having crying spells, or difficulty getting started on things. There were also significant differences in the factor loadings for the blues and crying items. In the second model incorporating covariates (e.g., gender, age, parental education), nine items showed Latino-White DIF (ps<.05), but two items (being bothered by things, crying spells) no longer showed ethnicity DIF. However, graphical analyses showed that DIF effects on the overall measure appear small and unlikely to affect total scores between groups. Given that the magnitude of DIF observed on the CES-D between Latinos and non-Latinos in this sample was small, the CES-D measured depressive symptoms similarly between Latino and White youths. Using MNLFA to examine DIF provided a more comprehensive picture of how demographic characteristics influenced symptom reporting. MNLFA modeled covariates simultaneously when assessing DIF, avoiding potentially confounding that might otherwise cloud our understanding of cultural differences.
Date of publication
Keyword
DOI
Resource type
Advisor
  • Youngstrom, Eric A
  • Hussong, Andrea M
  • Bardone-Cone, Anna
  • Curran, Patrick J
  • Jones, Deborah J
Degree
  • Doctor of Philosophy
Degree granting institution
  • University of North Carolina at Chapel Hill Graduate School
Graduation year
  • 2020
Language

Relations

Parents:

This work has no parents.

Items

Thumbnail Title Date Uploaded Visibility Actions
file details Salcedo_unc_0153D_19581.pdf 2020-08-17 Public Download