Increasing Theoretic Precision in Economic Inequality Research

Background and Purpose: Research on income inequality as a determinant of health outcomes in social work and related disciplines has been plagued by a conceptual inconsistency. As the literature on population-level income disparities producing negative health effects has progressed, a theoretic problem has emerged wherein studies have failed to distinguish between various conceptions of income itself. Two primary conceptions of income predominate. In the first case income is seen primarily as a proxy variable for access to materials which produce health (e.g., clean drinking water, sufficient nutrition, appropriate housing). In the second case income is conceptualized as a proxy variable for status, including notions of material purchasing power, but seeing health effects as driven substantially by non-material factors as well.

Methods: I conducted an extensive (n=~100) critical review (Grant & Booth, 2009) of the very large interdisciplinary academic literature on income inequality and health. Critical reviews seek to examine both a representative breadth of articles on the relevant topic as well as the highest impact research. Through this I aimed to investigate the extent of this conceptual inconsistency problem, better understand it, and to gain some clarifying insights. I reviewed articles for both their theoretical conceptualization of income and the associated methodological approach that was employed.

Results: Research on income inequality and health tends to take what I have termed either the material conception or the status conception view of income as a proxy variable. This difference has generated substantial methodological as well as conceptual inconsistency. Some research includes individual level income as a control variable while assessing the role of population level income inequality (material conception), theorizing that income inequality should show some effect above and beyond the material effect of individual income. Contrastingly, other researchers choose not to model individual income as a control (status conception) for fear of double controlling the individual effect out of the population level effect. The divide between research employing a status versus material conception is wide. The conclusions drawn when looking at identical statistical results may differ substantially.

Conclusions and Implications: Controlling for individual income with compositional effect testing in mind as compared to not doing so with contextual effects in mind may clearly lead to substantially different conclusions about the same population given the same raw data (e.g., Kondo et al., 2009; Pickett & Wilkinson, 2015). Based on this review I make the following suggestions. 1) In the short term, research on income inequality would benefit, at minimum, from explicitly identifying the conceptualization and modeling approach regarding income. 2) Given our lack of conclusive knowledge in this research domain, performing sensitivity analyses to alternate approaches is advisable. 3) Finally, researchers should work toward unification of the material versus status conception approach. Without these and other steps, the body of literature seems likely to remain at odds with itself, finding conflicting results and failing to make progress on consensus knowledge-building regarding and effect pathway identification.