Linking sequences of exposure to residential (dis)advantage, individual socioeconomic status, and health

Abstract

Life course theories suggest that the relationship between residential (dis)advantage and health is best understood by examining the ordering and duration of cumulative exposures across the life course. This study employs sequence and cluster analysis on two decades of residential histories linked to the Survey of the Health of Wisconsin to define typologies of exposure to residential (dis)advantage and use these typologies to predict self-rated fair/poor health. Exposure to residential (dis)advantage is mostly stable across the adult life course and greater disadvantage predicts fair/poor health. Longitudinal exposures to residential (dis)advantage shape health independently of and in tandem with individual-level resources.

Keywords: Health; Life course; Residential disadvantage; Socioeconomic status.

Fig. 1.

Twenty most frequent sequences of exposure to residential (dis)advantage for each typology. Note: “Year 20” (last year observation) represents the year that an individual entered SHOW; “missing” indicates years without any residential history; Yearly neighborhood exposure is measured by census-tract level percent of non-white residents, percent renter-occupied units, median household income, percent of the civilian population in the labor force over 16 who are unemployed, percent of the population 25 years or older with less than a high school degree, and percent of occupied housing units without a car.

Fig. 2.

Predicted probability of fair/poor health by typology. Note: Red star indicates a significant difference (p < 0.05) in odds of fair/poor health compared to the Majority Advantaged typology. Predicted probabilities are computed from Models 2 and 3 in Table 2.

Fig. 3.

Average marginal effects of college degree and total household income on probability of fair/poor health across typologies. Note: Average Marginal Effects represent a change in log-odds of fair/poor health between having no college degree (ref.) to having a college a degree (left panel) and a change in log-odds of fair/poor health between a household income of <$25,000 (ref.) compared to belonging in the other income groups. Average Marginal Effects are calculated from Models 1 and 2 in Appendix D.

Fig. 4.

Predicted probability of fair/poor health by typology across total household income and difference in log-odds of fair/poor health by typology across total household income. Note: Average Marginal Effects represent a difference in log-odds of fair/poor health between the Majority Advantaged and other typologies across household income categories. Confidence intervals (right panel) not overlapping with 0 indicate a significant difference in fair/poor health compared to the Majority Advantaged typology. Values in both figures are calculated from Model 2 in Appendix D.