RejoinderCausal inference and the relevance of social epidemiology
Section snippets
On Professor Diez-Roux's comments
Diez-Roux makes several important points about my paper. We both appreciate that the inferential obstacles faced in neighborhood effect research are ubiquitous in both observational epidemiology and sociology, at least. Her comment that authors too often inappropriately employ the word “effect” (I would add “impact,” “influence,” and “result”) is spot on, and remains a driving force behind my effort. We also agree that the principal modeling issue is “selection”, and that propensity score
On Professor Subramanain's comments
Subramanian passionately defends the multilevel model for estimating neighborhood effects with observational data. He is certain that existing neighborhood effect estimates are valid parameters of how social forces/factors impact health and that such work should continue unfettered. We share fewer points of agreement.
Subramanian's first important criticism is that I mixed-up issues of design and analysis; that the multilevel model is equally appropriate for analyzing both observational and
Final remarks
Diez-Roux, Subramanian and I agree that social epidemiology is an emergent field with enormous potential to improve the public health, for it is abundantly clear that social forces and group dynamics affect health. My effort here was aimed at helping us (1) confront causal inference and (2) produce policy-relevant (i.e., actionable) findings. I tried to motivate and enhance both theory and methods for a practicable social epidemiology. My concern with the otherwise insightful comments is that
References (13)
- et al.
The measurement of SES in health researchCurrent practice and steps toward a new approach
Social Science & Medicine
(2003) - et al.
Does the state you live in make a difference? Multilevel analysis of self-rated health in the US
Social Science & Medicine
(2001) - et al.
Social capital and community governance
Journal of Economic Perspectives
(2002) Foundations of social theory
(1990)- et al.
Socioeconomic status and injury mortalityIndividual and neighbourhood determinants
Journal of Epidemiology and Community Health
(2000) - et al.
Neighborhood of residence and incidence of coronary heart disease
New England Journal of Medicine
(2001)
Cited by (20)
Neighborhood effects in a behavioral randomized controlled trial
2014, Health and PlaceCitation Excerpt :Namely, he claimed that identifying an independent neighborhood effect on a health outcome was impossible given current methodologies (i.e. multilevel modeling of observational data). While the significance of Oakes’ critique has been debated, (Diez Roux, 2004; Subramanian, 2004; Oakes, 2004b) the field has generally responded favorably with a more cautious approach to making causal claims about neighborhood effects. At the same time, while there is great interest in the design and testing of randomized controlled trials (RCTs) aimed at modifying health behaviors, scant attention has been paid to understanding how Oakes’s arguments pertain to causal inference in the context of RCTs.
Food availability/convenience and obesity
2014, Advances in NutritionUsing matched areas to explore international differences in population health
2011, Social Science and MedicineCitation Excerpt :In other literature however, the debate continues. Oakes for example, (Oakes, 2004a, 2004b, 2006, 2009) has been a vocal critic of the multilevel approach to examining neighbourhood influences on health pursued by some researchers (Subramanian, Jones, Kaddour, & Krieger, 2009a, 2009b). Oakes’ solution is to call for the use of experimental designs, and randomised community trials in particular.
The Relevance of Social Epidemiology in HIV/AIDS and Drug Abuse Research
2007, American Journal of Preventive MedicineCitation Excerpt :Questions of its relevance, like the endless discussions of its definition, are no longer relevant. In fact, epidemiology, in its logistic lethargy, has been reenergized by the issues that “social” epidemiology brings to the fore: methods—the need to deal with different logical types of data in a single framework;6–10 theory—the need for some guiding thoughts that provide testable hypotheses and logical syntheses;11,12 and philosophy—the meaning and role of causality in epidemiologic thinking.13–19 In other words, social epidemiology differs from traditional epidemiology only because it is harder.
Behavioral science at the crossroads in public health: Extending horizons, envisioning the future
2006, Social Science and MedicineRobust estimation of the causal effect of time-varying neighborhood factors on health outcomes
2020, Statistics in Medicine