March 6, 2014

Journalists Should Embrace Complexity in Food Desert Research

Karina Christiansen

Karina Christiansen

CLF-Lerner Fellow

Center for a Livable Future

RRS994_DSC_6625-smeading journalist Heather Tirado Gilligan’s recently published article, “Food Deserts Aren’t the Problem: Getting fresh fruits and vegetables doesn’t make poor people healthier,” in Slate was an exercise in frustration. The author argues that a recent study published in Health Affairs demonstrates the essential failure of supermarket redevelopment policies to improve health in low-income “food desert” neighborhoods. In Ms. Gilligan’s article, you can find the typical errors that academics are familiar with when research is translated in the popular press – and to be fair, journalists are not trained to be research scientists (no more than research scientists are trained to be journalists!). But what is galling about the Slate article was not just the author’s inability to judge the quality of the research she’s reporting on, or unfamiliarity with how the latest article fits into the wider field of knowledge, but how transparently adversarial it was and how that obscured the validity of some of her later conclusions. The way I see it, this article exposes the pressure on news media to attract viewers through click-generating headlines promising a takedown of conventional wisdom or academic theories. What happened to nuance?

Academic research (to some of our dismay) is not quite so powerful to confirm or reject our understanding of how the world works in any one study. The Slate article, and a slew of related pieces from other media outlets, reports on the findings from a study that assessed how the development of a new grocery store impacted the perceptions, diet and weight of neighboring residents. It is an important (but not perfect) contribution in the field of “food desert” literature – where research focuses on how food stores differ by neighborhood and its impact on health. Despite how it is portrayed in the popular media, within academic circles there is an intense discourse on just what it means to have low access to healthy food, what shapes people’s food shopping and eating behaviors, and how the environments in which we live, work, and play impact our diet, weight, and health. The answers to these questions may seem esoteric, but they matter to real people whose lives and welfare could be improved by promoting evidence-based and equitable public policy.

So what did this study contribute to our knowledge, and (as importantly) what are its limitations? Anne Palmer, program director at the Center for a Livable Future, has already posted her take on the study and its coverage (here) where she offers a thorough review of the study methods and findings. Suffice it to say here that within six months to one year post opening, there is evidence that residents near the new supermarket have an improved perception of healthy food, but have not increased consumption of fruits and vegetables or lost weight. The time period of follow-up matters here. Should we expect these changes within a year, just because a new store is opened? The Slate article makes a basic fact-checking error when it reported that the follow-up was four years after the new supermarket opened. Despite numerous tweets, Slate has not corrected this simple, but crucial, error.

What is a reporter to do with these results, which on face value appear disappointing? The Slate article claimed this proves supermarkets do not improve the health of neighborhood residents. Of course, from one study, no matter if its findings are in the expected direction or not, we cannot jump to such a conclusion! We see that in less than one year post opening, area residents have not overwhelmingly made the switch to the new store as their main place to shop for groceries (about 25 percent switched). I did not see a separate analysis of the shopping behaviors of those residents who did make the switch (perhaps the sample was too small to detect anything but a large change in behavior), but it would have been interesting to explore. One could imagine many reasons why people have not switched en masse to the new store – maybe the prices are too high, or the store has not marketed itself adequately to the surrounding neighborhood, or it is just not enough time for people to change habits – but what this finding indicates is that if the intervention is to be successful, merely opening a store is not enough to change where and how people shop and eat. The study’s authors suggest complementary interventions to get shoppers into the store and making dietary decisions that could improve health (total health, not just weight on the scale). This is perhaps a less exciting conclusion than exposing that supermarket development has always been a pipedream, and we now have definitive proof.

The next important question is how do these findings fit into what has already been published to date? Here Ms. Gilligan commits a cardinal sin known to castigated undergraduates and unprepared grad students alike – defending her thesis only with studies that corroborate her argument and ignoring ones that refute or complicate it. I could imagine (and I have seen) papers that make a contradictory argument to hers, using only sources that bolster the viewpoint that there is a strong relationship between neighborhood food stores and dietary health. To be clear: this would also be problematic!

Here is what we know. This is the third quasi-experimental study of the impact of a supermarket opening on diet and health. The other two come from the United Kingdom and present mixed results. The first study, which lacked a comparison group, found positive, but modest impacts on diet (Wrigley 2003); the second found no evidence for improved diet, and while it had a comparison group the response rate was extremely low – only about 15 percent of people initially recruited agreed to take part in the study (Cummins 2005). Ms. Gilligan also cites a longitudinal study that found proximity to a supermarket to be unrelated to consumption of fruits and vegetables, contrary to expected findings (Boone-Heinonen 2011). This study relied on fixed effects models; when the authors used random-effects modeling they found associations between supermarket and diet in the expected direction (Lehmann 2012). Also problematic is Ms. Gilligan’s use of USDA data that demonstrates there are more grocery stores in low-income areas compared with wealthier areas. Let’s think this through. What we would call bodegas or corner stores are abundant in low-income urban neighborhoods, not so much in wealthier suburbs. The presence of many food outlets does not mean the presence of healthy, high quality, or affordable foods! This is the problem of aggregating data: come visit me in Baltimore, or Brooklyn, or any number of post-industrial cities and I will show you what this data means on the street level.

There are many troubling aspects of the Slate piece – the easily fact checked mistakes, the headline aimed at provocation and ad revenue generating clicks, the misunderstanding and therefore distorting of scientific evidence to support a pre-established point of view – but what I find most troubling is that the conclusion of the piece is really quite important for food policy advocates: poverty is a fundamental cause of poor health. However, the author does a disservice to her own argument by not recognizing the complexity and interdependence of pathways by which poverty actually produces ill health. How much of poor health is explained by environments, or behaviors, or the interaction of both?

In identifying harmful environments and how they constrain personal behavior, health disparities researchers are moving beyond a model that suggests the poor are primarily responsible for their own high rates of ill health through “lifestyle” decisions. We see that life is much more complicated.  People have agency, and do make their own choices, but not everyone has the same power and resources to eat the most nutritious foods, to live in the healthiest neighborhoods, to send their kids to the highest preforming schools.

Food policy researchers are trying to figure out what makes a neighborhood healthier, because we can promote those features through policy. Even if we think the presence of supermarkets makes neighborhoods “healthier,” academics and even most advocates recognize the complexity in the relationship between neighborhood, food, and health, and do not expect supermarkets to be a silver bullet to end obesity. But healthy retail, properly done, could begin to level the playing field so that Americans can access high-quality food everywhere, not just in white or wealthy neighborhoods. This is an issue of social justice, not just “obesity research.” It is another question entirely if supermarket development is the best way to go about achieving this goal of improved and equitable access. There is and ought to be a healthy debate on this topic.

Though it may not seem it, I agree with the author that supermarket redevelopment is a fundamentally conservative approach and I would argue that the marriage of public and private funds should not go unexamined or unchallenged. What could be some unintended consequences of supermarket development on the independent retail sector? Can healthy food retail development promote small business or co-operative ownership, especially by people of color and women? How can city policymakers support the community self-determination work already happening to promote equitable food access in marginalized neighborhoods? These are essential questions, and I applaud Ms. Gilligan for this line of analysis in her article.

Where the author goes wrong is after knocking down the supermarket redevelopment straw man, to propose an equally simplistic explanation by which poverty produces ill health. For her, it is cumulative stress – allostatic load. For the record, this is an important and emerging field of inquiry in public health. Work has been done, for example, on how racism directly impacts health and health disparities through this mechanism. However, the real world is much more complicated than this sole pathway for health disparities – for example, let’s ask how the food environment contributes to cumulative stress. A simple example: when the stores you can walk to only serve you in a small, dirty, enclosed room through a turnstyle surrounded by bullet proof glass, what message does that reinforce about safety in your neighborhood?

Neighborhood poverty, cumulative stressors, and the food retail landscape in urban communities are intimately tied to one another in a system (a system that involves myriad other components of the social world). Addressing any single component of it will always mean our work is incomplete, and I agree that can be unsatisfying. What we need is a way to move forward despite these limitations by doing research that suggests potential policies, prioritizing those policies based on our values of equality and justice, and returning again to evaluate how well we have done and where we can improve. It is an iterative process, and one that should not be shot down by any one study, any one research agenda. It is a discourse of people identifying how the world is, and how the world ought to be.

<< Read Anne Palmer’s The Positive Change Created by New Supermarkets in Food Deserts


Boone-Heinonen J, Gordon-Larsen P, Kiefe CI, Shikany JM, Lewis CE, Popkin BM. Fast food restaurants and food stores: longitudinal associations with diet in young to middle-aged adults: the CARDIA study.  Arch Intern Med. 2011;171(13):1162-1170.

Cummins, S., Petticrew, M., Higgins, C., Findlay, A., & Sparks, L.  (2005).  Large scale food retailing as an intervention for diet and health: Quasi-experimental evaluation of a natural experiment.  Journal of Epidemiology and Community Health, 59, 1035-1040.

Lehmann Y, Morland K, Godbold J. Supermarkets: Components of Causality for Healthy Diets. Arch Intern Med.2012;172(2):195-197.

Wrigley N, Warm D. Deprivation, diet, and food-retail access: findings from the Leeds ‘food deserts’ study. Environ. Plann. A, 35 (2003), pp. 151–188.

One Comment

  1. Pingback: Food News Fast - March 10, 2014 | Food Politic

Leave a Comment

Your email address will not be published. Required fields are marked *