Recorded presentation

Paula A. Braveman, MD, MPH, Catherine Cubbin, PhD, Susan Egerter, PhD, and Kristen S. Marchi, MPH. Center on Social Disparities in Health, University of California, San Francisco, 500 Parnassus Avenue, MU-3 East, Box 0900, San Francisco, CA 94143-0900, 415-502-8565, ccubbin@itsa.ucsf.edu

Background: “Socioeconomic status/position” (“SES”) appears throughout health literature but is rarely defined. Diverse measures are used, generally without justification. Often a single domain is measured – in the U.S., typically income or education.

Purpose: To inform how SES is conceptualized and measured in health research

Methods: Analyses of population-based data (National Household Interview Survey, National Health and Nutrition Examination Survey, Youth Behavioral Risk Factor Survey, and California’s Maternal and Infant Health Assessment) focused on several outcomes (self-reported health, all-cause mortality, risk-taking behaviors, asthma, blood lead, and birth outcomes). Overall and within racial/ethnic subgroups (African American/Black, European American/White, Latino/Hispanic/Mexican-American), we examine: correlations among different SES measures; and the associations between each outcome and different SES measures, using logistic regression to estimate odds ratios for (a) different SES measures and (b) race/ethnicity, with and without adjustment for different SES measures.

Results to date: Education and income are correlated modestly overall and weakly within particular subgroups. Racial/ethnic and socioeconomic disparities can vary with both the domain (income or education) and specification (continuous, categorical, how categorized) of the SES measure. Conceptual considerations and the literature suggest that similar findings will apply to many if not all of the outcomes to be examined.

Conclusions: Education is not a proxy for income, particularly in populations that are ethnically diverse. Multiple measures are needed to reflect the multi-dimensional nature of SES. Researchers should select SES measures based on likely explanatory pathways, not mechanically. Literature claiming to have “controlled for SES” despite limited SES measurement should be reassessed.

Learning Objectives: At the conclusion of the session, participants (audience) will

• Be aware of the problematic nature of current standard approaches to measuring and conceptualizing SES
• Understand that SES is multi-dimensional and cannot be measured by any single variable
• Know that education is not an acceptable proxy for income, and that income is not an acceptable proxy for wealth, particularly in populations that are ethnically diverse
• Know that researchers may reach different conclusions based on the SES measures they use
• Know that they should select SES measures based on thinking through plausible explanatory pathways from their independent variables to the dependent variables of interest, rather than in a one-size-fits-all fashion

Presenting author's disclosure statement:
I do not have any significant financial interest/arrangement or affiliation with any organization/institution whose products or services are being discussed in this session.

Recorded presentation