A Test of the Social Causation Hypothesis and the SES Gradient: The Utility of Wealth, Self Reported Health, and Unmeasured Heterogeneity

Michael L. Spittel, National Center for Health Statistics (NCHS), CDC

This study examines the “social causation hypothesis” of several conventional SES measures and health. To control for alternative interpretations, such as “health selection” and “reverse causality” this study employs: measures which are not as sensitive to changes in health, such as wealth; self reported health measures; and event history methods that have been developed to control unmeasured heterogeneity. All results are net of the standard measures of occupation, income, and education. The data used for this study come from the PSID. Results indicate that income, wealth, and employment status are all important factors in predicting mortality net of self reported health. This contributes to the SES and health literature by utilizing all known approaches to mitigate directionality problems. Although this does preclude the potential effect of health on status, it does suggest that certain measures of SES are strongly associated with health.