 

 

 

 

 

mm

 

:73 ,.
.mwﬁ

.
WTPJM at 3 a
‘ l ‘I l .V"..‘.

I . ‘ .
-1. 3. . .3.
. 5..., ...\ . .. 2.2:
.Mwﬂunﬁuvﬁ‘ .WmeciHnﬂ
. Mm . -
a w:

2. 5.333%

E“

lav.

. tfnlzﬂh.
A {1%. ‘ . 53L
-IP..i{\1.!.I|.-.9

¢' \k

'1)

 

I..lBR”'ARY
Michigan State
University

 

 

 

This is to certify that the
dissertation entitled

FOOD DESERTS? PURCHASING PATTERNS AND
PERCEPTIONS OF FOOD INSECURE AND SECURE
HOUSEHOLDS IN A MIDWEST URBAN NEIGHBORHOOD

presented by

BRIAN JASEN THOMAS

has been accepted towards fulﬁllment
of the requirements for the

 

 

 

 

PhD. degree in Sociology
1 a? C f/
Major Professor’s Signature
(1 / 7 / e ’71
/ l
' Date I

MSU is an aﬁinnative-action, equal-opportunity employer

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

MAR 2 4 2009

 

011209

 

 

 

 

 

 

 

 

 

 

 

 

 

6/07 p:IClRC/DateDue.indd-p.1

 

FOOD DESERTS? PURCHASING PATTERNS AND PERCEPTIONS OF FOOD
INSECURE AND SECURE HOUSEHOLDS IN A MIDWEST URBAN
NEIGHBORHOOD
By

Brian J asen Thomas

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Sociology

2007

ABSTRACT
FOOD DESERTS? PURCHASING PATTERNS AND PERCEPTION S OF FOOD
INSECURE AND SECURE HOUSEHOLDS IN A MIDWEST URBAN
NEIGHBORHOOD
By
Brian J asen Thomas

Over the last decade new attention has been paid to the issue of food security in the
United States. Food security represents a useful lens through which to examine the issue
of agency since food is difﬁcult for a large sector of the population to access, even
though it is considered a basic human right. By examining two classes of households,
food secure and food insecure, I shed light on some factors that lead to the relative ability
of households to successfully and reliably obtain food. I focus on the perception and
behavior of consumers related to decisions to shop, or not to shop, at food retailers.
Some theories of consumer behavior tend to focus either on class-related cultural
elements which determine taste while others focus on structural elements of the food
system which force a limited selection onto different social groups. While it is assumed
that culture affects taste to some extent, results from this study suggest that structural
elements of the food system and economic differences among households have a larger
inﬂuence on store choice than cultural preferences. In fact, both food secure and insecure
individuals indicated similar sets of criteria that they used in determining store choices.
However, food insecure households were more likely to shop at deep discounters and
more likely to travel farther to obtain food. These results suggest that structural elements
such as food retailer locations may lead to increased transaction costs for food insecure

households when compared to food secure households. While food insecure households

do have limited agency in terms of procuring food, their options are more limited and, in
many cases, more costly than those of food secure households. Therefore, the structure
of the conventional food system, in particular as it relates to physical distribution of food
retailers, most closely ﬁts the interests of those who already have few, if any, problems

reliably purchasing healthy food.

Copyright by
BRIAN JASEN THOMAS
2007

ACKNOWLEDGEMENTS

I am grateful to everyone who has provided me with guidance and support throughout my
graduate experience. I would like to thank my advisor, Dr. Toby TenEyck, for his insight
and support. Dr. TenEyck’s commitment to graduate students and meaningful
professional advice is one of the primary reasons I have succeeded in completing this
project. I would like to thank my committee, Dr. Laura Delind, Dr. Geoffry Habron, and
Dr. Dan J affee for their direction and contributions. I would especially like to thank Dr.
Delind for connecting me with the Allen Neighborhood Center and instilling in me a
dedication to research oriented towards positive social change. I would also like to thank
Dr. Alan Rudy for giving me an appreciation for theory and guiding me through much of
my graduate career and Dr. Kim Chung for her help with the qualitative portion of this
dissertation. The people at the Allen Neighborhood Center, especially Joan Nelson and
Hollie Hamel, not only introduced me to the people of the Allen Neighborhood, but they
also showed me the difference that can be made by bringing people together.

Most of all, I would like to thank my wife, Holli Wallace. She was patient when
necessary, assertive when appropriate, and comforting when needed. She taught me
about people and showed me the importance of helping those who need help the most. I
would like to thank my mother, Shuk—chong Thomas, and my father, Frederick Thomas,
for teaching me the importance of education and providing me with the opportunities that
have allowed me to reach this place. I would also like to thank my brother, David
Thomas, whose accomplishments have always inspired me. Finally, I would like to thank
my mother-in-law, Dr. Denise Reiling, whose professional and personal advice has been

essential to negotiating the complicated world of academia.

DEDICATION S

This dissertation is dedicated to my son, Elliott James Thomas, who has given

new purpose to everything I do.

vi

TABLE OF CONTENTS

LIST OF TABLES ................................................................................... ix
LIST OF FIGURES ................................................................................. x
CHAPTER ONE: INTRODUCTION
Food and the Consumer ..................................................................... 1
Research Questions .......................................................................... 9
Signiﬁcance of Study ...................................................................... 11
Organization of Dissertation .............................................................. 13

CHAPTER TWO: CONSUMER BEHAVIOR AND FOOD

Theories of Consumer Behavior ......................................................... 15
The Sociology of Agriculture and the Consumer ...................................... 19
Agency and the Consumer ................................................................ 30
Space, Agency, and Structure of the Food System .................................... 35

Space and the Structure of the Food System .................................. 40

From Supercenters to Food Deserts ............................................. 42
Conclusion .................................................................................. 50

CHAPTER THREE: STUDY SITE AND GENERAL METHODOLOGY

Study Site .................................................................................. 54
General Methodology .................................................................... 59
Qualitative Methods ...................................................................... 62
Quantitative Methods ..................................................................... 64
Survey Mode ...................................................................... 64
Sampling ........................................................................... 67
Description of Sample ........................................................... 67
Deﬁnition of Terms ....................................................................... 74

CHAPTER FOUR: CAUSES OF FOOD INSECURITY

The Problem of Hunger and Food Insecurity....................... ................... 78

Analysis and Results ...................................................................... 89
Variables ........................................................................... 90
Results .............................................................................. 92

Conclusions ................................................................................. 99

CHAPTER FIVE: PERCEPTION AND BEHAVIOR OF FOOD SECURE AND
INSECURE HOUSEHOLDS

Agency, Taste, and Class ................................................................ 104
Qualitative Results: Tastes and Purchasing Strategies .............................. 110
Analysis ........................................................................... 110

Results ............................................................................ l 10

vii

Quantitative Results: Preferences and Behaviors .................................... 116

Analysis ........................................................................... 1 16
Results: Food Retailer Perception ............................................. 117
Results: Food Purchasing Patterns ............................................ 124
Conclusions ............................................................................... 130

CHAPTER SIX: SPACE AND FOOD DESERTS

Analysis and Results ..................................................................... 137
Perceptions of Space ............................................................ 137
Qualitative Results ...................................................... 138

Quantitative Results .................................................... 139

Behaviors in Space ............................................................... 140
Conclusions ................................................................................ 148

CHAPTER SEVEN: CONCLUSION

Agency Revisited ......................................................................... 152
Resistance Movements and Alternative Food Sources .............................. 159
Limitations of this Research and Future Research ................................... 168
Recommendations ........................................................................ 169
Final Words ................................................................................ 173
APPENDICES ..................................................................................... 176
BIBLIOGRAPHY ................................................................................. 179

viii

LIST OF TABLES

Table 1: Data Collection Techniques and Relationship to Research Questions and
Organization of Dissertation ....................................................................... 62

Table 2: Summary of variables used to predict food security status ......................... 91

Table 3: T-test and chi-square results for each dependent variable in relation to food

security status ........................................................................................ 93
Table 4: Cross-tabulation of household income and food security status .................. 94
Table 5: Cross-tabulation of household education and food security status ............... 94
Table 6: Logistic regression model of food insecurity status based on all variables ....... 95

Table 7: Logistic regression model of food insecurity status based on a reduced set of
variables .............................................................................................. 96

Table 8: Logistic regression model of food insecurity status based on poverty status and
simpliﬁed education variables ..................................................................... 97

Table 9: Ranking of the relative inﬂuence of store characteristics on the decision to shop
at the primary food retailer ............................. 123

Table 10: Average distance in miles from food secure and food insecure households to
different food retailers ............................................................................. 144

ix

LIST OF FIGURES

Figure 1: Determinants of consumptive practices .............................................. 26
Figure 2: Location of study site in relation to the state of Michigan ........................ 56
Figure 3: Median income for the study site and surrounding census blocks ............... 57

Figure 4: Response rates as a percent of the sampling frame in each census block in the

study site ............................................................................................. 69
Figure 5: Household income distribution of the study sample and 48912 zip code ........ 70
Figure 6: Highest level of education of someone in the household for the study sample

and population education for 48912 zip code ................................................... 71
Figure 7: Food security/insecurity as related to agency and structure as determinants of

consumptive practices .............................................................................. 79
Figure 8: Food security levels in the study site, Michigan, and United States .............. 89

Figure 9: Store characteristics that inﬂuenced respondents’ decisions to shop there. ....120

Figure 10: Reasons for shopping at most frequently shopped at store by food secure and
food insecure households ......................................................................... 122

Figure 11: Stores where food was purchased in the last 30 days by food secure and food
insecure households ............................................................................... 124

Figure 12: Store shopped at most frequently in the past 30 days by food secure and
insecure households ............................................................................... 125

Figure 13: Store shopped at second most frequently in the past 30 days by food secure
and insecure households .......................................................................... 127

Figure 14: Distribution of convenience and party stores, supermarkets, and
supercenters within and around the study site ........................................ - ......... 136

CHAPTER ONE

INTRODUCTION

Food and the Consumer

For the past few decades, social scientists have been interested in issues of food
availability. The ability of a society to produce food is critical to its development
(Diamond 1997) and intricately related to basic survival and environmental quality
(Ehrlich, Ehrlich, and Daily 1993). Access to food has been examined as a factor related
to both fertility and mortality (Basu 1992). Authors such as Smil (1994) have argued that
food sufﬁciency needs to be understood as more than an issue of simple calories, but also
of nutritional quality and social and environmental differences among populations. Some
efforts to identify the impacts of different nutritional food regimes have focused on
developing countries. For instance, Jenkins and Scanlan (2001) found increases in food
security in many developing countries from 1970 to 1990. However, the increases were
not equally distributed globally and negative change was identiﬁed in areas such as sub-
Saharan Africa. The United States, despite its global political power and economic
productivity, is not immune from problems of food availability. In 2005, 11% of the US.
was still food insecure (Nord, Andrews, and Carlson 2006). Moreover, research indicates
that food insecurity has very real social, psychological, and physiological impacts
(Hamelin, Habicht, and Beaudry 1999; Morland, Wing, and Roux 2002; Olson 1999;
Rose 1999; Travers 1996; Vozoris and Tarasuk 2003). These impacts are not just caused

by a lack of food, but also by the quality of food that is consumed.

The question that needs to be addressed in the United States, as elsewhere, is what
factors are leading to differential food consumption patterns by some populations? Why
do some people suffer from a lack of food or only have low quality foods to eat?
Needless to say, the inability to pay for food is a primary obstacle to food access and lack
of income is an important factor in assessing food security. People do not, however,
simply buy less food when they have less money. Rather, there are a complex series of
decisions that inﬂuence the quality, quantity, source, and timing of food procurement
(Travers 1996). The transaction in the checkout line is only one of a set of negotiations
which takes place when people are making purchasing decisions.

Several authors have argued that the problem is not only whether people have
sufﬁcient money for food, but also the relative difﬁculty for some groups in getting to the
store to buy food (Blanchard and Lyson 2002; Caraher, Dixon, and Car-Hill 1998;
Cummins and Macintyre 1999; Furey, Strugnell, and McIlveen 2001; Lang and Caraher
1998). After all, price is only part of the cost of food. Transaction costs refer to all of the
other costs associated with getting food, including traveling to the store, negotiating the
parking lot, spending time cutting coupons and looking for sales, and carrying food.
Household characteristics such as physical disabilities, lack of an automobile, or small
children may increase the transaction costs for accessing food from different locations.
Furthermore, neighborhood characteristics can inﬂuence the distance that people must
travel to purchase food and inﬂuence the price they pay for various commodities (Alwitt
and Donley 1997; Hall 1983; Kaufman, MacDonald, Lutz, and Smallwood 1997).

In the past few decades the ﬁeld of sociology has begun to theorize about the role

of consumers within the food system (Fine 2002; Lockie 2000; Lockie and Kitto 2000;

Marsden, Flyn, and Harrison 2000). New questions have arisen regarding the ability of
consumers to inﬂuence the content and structure of the food system. For example
research on alternative agriculture has provided evidence to suggest that the structure of
the food system is very much inﬂuenced by consumers (Allen and Kovach 2000;
Raynolds 2000); however, scholars working largely within the sociology of agriculture
are still struggling to understand the nature and extent of that inﬂuence (Goodman and
DuPuis 2002). This new attention to consumers in the food system has developed, in
part, because of the recent development of a sociology of consumption, which has begun
to examine the cultural, economic, political, and social forces that inﬂuence consumption.
Building on classical work by scholars such as Veblen (1959) and more recent work by
scholars such as Bourdieu (1984), work in this area has examined important questions
regarding how consumption has, and has not, changed in the past few decades. Given the
complex nature of the process of consumption, it will be important that theories of
consumers in the food system also incorporate research from other disciplines. For
instance, the ﬁeld of geography has recently sought to understand the behavior of
consumers within the context of a retail environment through the concept of food deserts
(Furey et al. 1999; Lang and Caraher 1998).

The scholarly understanding of the role of the consumer in the food system has
very real implications for those populations in the United States who are suffering either
from a lack of food or reliable access to high-quality food. While consumer movements
around food safety and quality appear to be inﬂuencing the activities of food companies,
such as by limiting the sale of genetically modiﬁed organisms in some countries, it is not

clear whether these movements are beneﬁting all segments of the population. The high

levels of food insecurity in the United States certainly suggest that simplistic notions of
consumer agency must be cast aside for a more nuanced understanding of the variety of
ways in which different consumers inﬂuence and are inﬂuenced by the larger food
system.

Hays (1994) explains that traditional sociological understandings of structure and
agency are problematic because they fail to acknowledge that there are various social
structures which dictate individual interests. Agency is reduced to the extent to which
individuals can get what they want and structure represents static structures that either
facilitate or prevent those wants from being met. But, as Hays suggests, if agency
represents the extent to which people can get what, we must keep in mind that there were
social structures related to class, gender, ethnicity, economics, politics, education, and
family that also inﬂuence individual interests. There is no such thing as agency that is
separate from some social structure.

This does not mean that agency and structure lose all sociological meaning.
Rather, it suggests the importance of speciﬁcity. Agency does not stand in opposition to
all social structures; however, it can be a measure of the extent to which various social
structures,.which are manifest in individual tastes, coincide with a speciﬁc social
structure. This is how agency is approached in this study. In this case, the social
structure we are interested in is the conventional food retail system in our study site.
Agency, therefore, can be measured by examining the extent to which households are
able to obtain what they want from this food system without turning to other social
structures to procure food. In other words, agency is a measure of the extent to which

consumers can easily get what they want from the area food stores.

Even narrowed in this fashion, the relationship between consumers and food
retailers is important as this is the primary vector through which consumers engage the
food system. Research into food security and food deserts has helped to highlight the
relationship between consumption and the retail environment, but the nature of this
relationship is not always clear. This study examines the situation of consumers in an
inner city neighborhood in Lansing, Michigan. Within a distance of ﬁve miles from the
neighborhood, there are nearly 100 food retailers. These food retailers represent a cross
section of the industry: 45 are convenience or party stores, 17 are small grocers, 11 are
specialty stores, and 22 are large supermarkets. This variety and density of retailers in a
medium sized urban area appears to stand in opposition to much of the current food
desert research which would suggest that given the high density of food retailers, access
should not be a signiﬁcant problem for neighborhood residents. In spite of this apparent
abundance of retailers, however, this study found food insecurity twice that of the state.
These high levels of food insecurity indicate that, despite this high density of food
retailers, signiﬁcant sectors of the population are having difﬁculty getting what they want
from the food system.

In 2003, a local neighborhood center went door to door and conducted a survey in
the study site that included several food related questions (Thomas 2004). In addition to
ﬁnding food insecurity levels much higher than the state of Michigan, 7% of the
respondents indicated that transportation to food retailers was a problem and over 9%
indicated that getting to the store was either difﬁcult of very difﬁcult. Also, nearly 30%
of those who were food insecure indicated that they also had a disability. While this

survey was conducted as part of a community outreach effort and therefore not randomly

sampled, it nonetheless indicates that food security is an issue for many residents and
barriers related to disabilities and transportation may be part of the problem.

Each of these barriers indicates that the ability of some individuals to purchase
food is not simply an issue of income sufﬁciency, but also inﬂuenced by a combination
of proximity to stores and ability to negotiate that distance. For instance, someone with a
relatively high income, but a physical disability, may have difﬁculty traveling to some
stores or even traversing the parking lot or large area of a large supermarket. At the same
time, someone with a low income might be able to negotiate their resource limitations by
driving across town to discount grocery stores. Income matters, but it is not the only
factor that inﬂuences food insecurity. Similarly, barriers such as distance cannot be
measured through binary assessments of “too far” or “close enough.” It is important to
consider the characteristics of individual households in order to understand what “too far”
and “close enough” means for different groups of people. As I will discuss later in this
study, qualitative research in this neighborhood indicates that distance to food sources is
often perceived and negotiated by food insecure individuals through very complex food
purchasing strategies, such as shopping at multiple stores to seek sale items and driving
out of the neighborhood to discount grocers. Food-insecure individuals engage in a
diverse set of activities to balance limited incomes and transportation opportunities
against the structure of the retail environment.

The trends evident in Lansing indicate that a signiﬁcant proportion of the
population in the area is struggling to obtain food. The results of the 2003 survey also
indicated that a number of people without automobiles live outside a convenient walking

distance (deﬁned as V2 a mile) from a major food retailer. Much of the food desert

 

literature suggests that for disadvantaged consumers simply making it to the store
represents a signiﬁcant obstacle (Furey et al. 2001; Whelan 2002) and data from this
study indicates that many people in Lansing are encountering these same obstacles.

While identifying obstacles to reliable access to food is itself an important
contribution of food desert literature to broader sociological thought, it is only recently
that researchers in this area have begun to look closely at how people behave within a
food desert, as well as how disadvantaged consumers perceive and make decisions within
their retail environment (Whelan, Wrigley, Warm, and Cannings 2002). Moving from
understanding the distribution of food retailers in relation to different populations and
towards understanding how different people perceive and negotiate that environment
represents an important direction for food desert literature. Consumer decision-making
regarding store selection involves a complex set of economic and non-economic factors
(Bell, Ho, and Tang 1998; Moore 1989). Without explicit incorporation of consumer
perception of food retail options and how that determines behavior, food desert literature
risks digressing to simplistic spatial models of food availability and access. Furthermore,
this provides food insecure populations with a voice in regards to the structure of the food
system.

Food desert research has made the important point that the physical structure of
the conventional food system makes it easier for some sectors of the population to buy
food. However, current applications of food desert methodologies have been insufﬁcient
to explain the high levels of food insecurity in the study site. This stems, in part, from
the privileging of the physical structure of the food system in determining whether or not

a household will be food insecure and, in part, from the failure to acknowledge either the

existence of other social structures or the agency of people in the context of the food
system or other social structures. Distance, after all, has different meaning for different
people. Surely, it may represent a barrier, but it may represent one that can be overcome
through other resources such as a kind neighbor. At the same time, living one hundred
feet from a grocery store does not matter-if you do not have sufﬁcient money to buy food
there. However, a trip to a relatively distant dollar store might allow one to stretch a
limited set of ﬁnancial resources. The point is not that distance, as a characteristic of
food system structure, is of little relevance. Rather, it points to the importance of a more
nuanced understanding of how it matters to different people.

The next step for food desert research is to reconcile the agency of disadvantaged
consumers without dismissing the initial, and still pertinent, point that the structure of the
food system does inﬂuence household retailer choice and, consequently, consumption
habits. Food insecure households and disadvantaged consumers do not completely lack
agency, even in the context of a physical environment where food retailers are located
relatively far from people’s homes. People are creative and innovative; they are able to
devise solutions to transportation limitations and to call upon friends, neighbors, and
relatives for assistance. Economic limitations can be negotiated through trips to discount
grocers, dollar stores, or other activities.

Yet, the agency that is evident in these strategies and activities is, in many ways,
constrained. Certainly, these activities and strategies can put food on the table, but they
require additional time identifying sales and coupons, traveling to discount grocers, or
shopping at multiple stores. There may even be additional costs that are incurred as the

up front cost of repairing an automobile is traded against the cost of purchasing an

overpriced box of macaroni and cheese from the convenience store that is within walking
distance. Furthermore, relying on friends and family members for food or rides to the
grocery store requires turning to social structures that are not part of the conventional
food system.

Building from the sociology of agriculture and food desert research, this leads us
to a fundamental question: To what extent is food retailer choice an issue of personal
preference or one of structural necessity? This question is at the heart of this study
because it provides insight into the reasons that people are differentially served by the
conventional food system. Furthermore, it provides insight into the reasons that different
people appear to have differing levels of agency in the context of the structure of the food
system. Based on a combination of qualitative and quantitative data I will make the point
in this study that food secure and food insecure individuals often share similar
perceptions of the food retail environment. However, due to a limited ability to negotiate
the retail environment, food insecure individuals are forced to develop complex
purchasing strategies that prioritize cost. In short, I will address the question: to what
extent do differing levels of agency relate to different perceptions of and behaviors

related to food purchasing?

Research Questions
Six major research questions are addressed in this research. For clarity, these
questions will be examined separately in different chapters.
1. What types of households are more likely to be food insecure or food insecure

with hunger? This question is central to research on food security and important

to identifying populations that are particularly vulnerable to problems related to
food access. While, without surprise, evidence indicates that household income is
one factor related to food security, it appears that other factors may also have an
inﬂuence on food security status (Rose 1998).

. Do food secure and food insecure households perceive food retailers diﬁ‘erently?
Major consumption theories, such as those posed by Bourdieu (1984), suggest
that social class may inﬂuence how people perceive their environment. Class
inﬂuences the ability of people to interpret different elements of their social
worlds. This question examines whether food security status is related to
perception of food retailers.

. Do food secure and food insecure households use diﬁerent criteria for selecting
food retailers? Regardless of whether or not retailers are perceived to be the
same or different, it is possible that different criteria are used to actually select
food retailers. For instance, shoppers in both food secure and food insecure
households may perceive Wal-Mart as the place to get low-cost bulk items, but
only the food insecure households may see this as a factor important to their
decision whether or not to shop there. This question addresses the decision
making criteria used by people in the neighborhood. What factors go into
peoples’ decisions to shop at individual stores?

. Do food secure and food insecure households shop at diﬁ'erent types of food
retailers? It is important to understand the consequences of perception on actual

behavior. In particular, since shopping behavior may be inﬂuenced by factors that

10

 

respondents may or may not be aware of, it is important that actual shopping
behavior of respondents be examined in addition to perception of retailers.

. Are food secure and food insecure households located at diﬁ‘erent distances from
food retailers? Food desert research suggests that distance from food retailers
inﬂuences food security status. This research question tests the accuracy of this
proposition in the study neighborhood. If distance really inﬂuences food security
status, food insecure households should be located farther from food retailers than
food secure households.

Does distance inﬂuence shopping behaviors of food secure and food insecure
households? Of course, food security status is only one measure of food
purchasing patterns and this question examines the possible inﬂuence of distance
on other shopping behaviors. In other words, is proximity a factor that inﬂuences

the likelihood of households to purchase food from different sources?

Signiﬁcance of Study

This study has a number of implications for scholarship as well as those engaged

in work to alleviate food insecurity. First, this research will shed light on the role of the

consumer in the food system. Current research, in particular that stemming from

agrifood studies, is still working to “theorize the consumer” as an agent in the modern

food system. Such efforts have been limited by a tendency to treat consumers as a

homogenous group. This research begins to alleviate some of the problems related to this

homogenous treatment by examining a wider spectrum of consumers and drawing

explicit divisions between groups of consumers.

ll

 

Another debate that has emerged among scholars examining consumer behavior
relates to the extent to which choice is dictated by culturally derived preferences and to
what extent choice it is dictated by the availability of goods and services. By looking at
consumer preferences and behaviors among different groups of people, this research will
shed light on the relative inﬂuence of individual preferences and availability on choice.
In much of the literature these discussions have related to item selection; however,
recognizing that there are a number of decisions involved in procuring commodities, it is
important that other decisions, such as the decision regarding retailer choice, also be
examined.

This research will extend scholarship that has recently begun to view hunger as
part of the structure of the food system, rather than only a problem for individuals or
households. Speciﬁcally, food desert research suggests that the physical structure of the
food system creates barriers for certain populations who ﬁnd distance a barrier, or limit,
to procuring food. This research will look explicitly at the relationship among distance,
food security status, and shopping behavior. Given conﬂicting evidence on the impact of
distance, this research will take an important step by looking directly at individual
household behavior as it relates to distance from different types of food retailers.

Finally, this research will point to some reasons why households may be food
insecure. This information can be of vital use to organizations that are engaged in
activities to alleviate food security problems. A number of non-proﬁt and governmental _
agencies are currently engaged in activities to make healthy food more accessible in the

neighborhood that was studied, and it is hoped that the information derived from this

12

dissertation can be used to develop new programs and guide existing programs that will

help alleviate food insecurity.

Organization of Dissertation

The next chapter in this dissertation examines the theoretical framework that has
deﬁned this research with particular emphasis on the issue of consumer agency in the
food system. Fundamentally, this study questions the extent to which food insecure
households have control over what they eat. It is important, therefore, that I start by
examining the meaning of agency in the food system. In Chapter Three I outline the
basic methods used to examine the food purchasing patterns of households in the study
site. I collected both qualitative and quantitative data through a combination of a survey
and interviews. Additional data on food retailers was also collected and assessed using 3
Geographic Information System. In this chapter, I also provide an overview and rationale
for the study site. Given the range of analytic techniques used in this research, the
speciﬁc ways in which data were analyzed is outlined in each chapter before reporting the
results of the analysis. In Chapter Four, I examine basic questions regarding the causes
of food insecurity. In this section, the factors related to food insecurity in the
neighborhood are examined and compared to other studies looking at the causes of food
security. In Chapter Five I take a closer look at the differences and similarities between
food secure and food insecure households by comparing the perception and behavior of
households in relation to food shopping. Utilizing a spatial analysis of survey
respondents and food retailers in Chapter Six, 1 shed light on the role that the physical

structure of the food retail environment plays on the issue of food security. Finally, in

13

Chapter Seven, I will revisit the issue of agency as examined in Chapter Two and
demonstrate how the physical structure of the food system plays a role in food security

levels.

14

CHAPT ER TWO

CONSUMER BEHAVIOR AND FOOD

Theories of Consumer Behavior

In many ways, the act of consuming food is exemplary of the relationship
between individual decisions and social structures. At one level, eating is an act that is
inﬂuenced by individual tastes and nutritional requirements. On the other hand, eating
represents the expression of a set of socially constructed symbols, the manifestation of
culturally determined tastes, and decisions constrained by a variety of social structures.
We eat what we want, but our wants have themselves social roots and are constrained by
the foods that we can obtain. Fundamentally, therefore, decisions around food offer
potential insight into some very basic sociological questions regarding structure and
agency. To what extent is an individual decision about what to eat (or in the case of this
study, where to shop for food) dictated by their individual wants and desires and to what
extent is the decision dictated by social structures which dictate food price, availability,
and desirability.

Interestingly, despite the potential of food to shed light on issues central to
sociology, Beardsworth and Keil (1997) note that, historically, sociology has not paid
attention to it. They suggest that “it may well be that the very taken-for-granted nature of
eating has rendered this activity, and the complex of other activities and relationships
which cluster around it, relatively ‘invisible’ to sociologists (2).” While this may have

been the case in the past, rapid growth in the sociological literature around the area of

15

food points to a newfound awareness of the topic. Not only does the food sector
represent a signiﬁcant part of the economy with food and beverage store sales reaching
nearly $520 billion dollars in 2005 (US Census), but it also represents the theoretically
important conﬂuence of two important, albeit relatively novel, sets of sociological
literature: the sociology of agriculture and the sociology of consumption. In the case of
the former, Goodman and Dupuis (2002: 8) state “Consumption as a theoretical category
has ‘arrived’ in agro-food studies.” In the case of the latter, Warde (1997: 22) argues that
“for a sociology of consumption, food is a most instructive critical case study.” While it
would seem that the interest of these two literatures on a single topic demonstrates the
emergence of a sociology of food, a cohesive understanding of the role of the consumer
in the food system has yet to be fully developed. Regardless, it is worth reviewing
existing sociological theories of consumer behavior to understand the current state of
sociological theories of food consumption. To what extent can consumer behavior
around food be understood within the context of existing theories of general consumer
behavior?

Certainly there have been some very inﬂuential theoretical works that have sought
to understand consumer behavior from a sociological perspective. Campbell (2000)
provides a summary of the predominant theories that claim to explain why and how we
consume. He breaks these theories down to Veblenesque, instinctivism, marginal utility
theory, and manipulation. Veblenesque theories are based, to varying extents, on
Veblen’s classic work Theory of the Leisure Class (1959) which introduced the concept
of “conspicuous consumption” to everyday language. For Veblen, consumption was a

cultural act that reinforced social class differences. In particular, unproductive

16

consumption (i.e. luxuries) gained cultural value among the upper class which was used
to distinguish them from other social classes. More recently, Bourdieu (1984) built upon
the cultural role of consumption as an act of differentiation. He stated “consumption is . .
. a stage in a process of communication, that is, an act of deciphering, decoding, which
presupposes practical or explicit mastery of a cipher or code (2).” The inﬂuence of these
theories can be seen in studies such as that by Peterson and Kern (1996) which suggests
that omnivorous consumptive tastes have begun to deﬁne upper social classes.

Despite the popularity of this approach, Campbell (2000) remains critical of
Veblen’s theories noting a failure to distinguish between societies of close-knit
communities of peers and socially mobile, fairly open societies. Most importantly,
according to Campbell (2000:66), “Veblen does not provide a basis for distinguishing
traditional from modern consumer behavior, and hence does not account for the
insatiability and desire for novelty which is such a crucial hallmark of the latter.”

Campbell (2000) states that Instinctivism is largely based on humanistic
psychology’s concept of a hierarchy of needs and the idea that wants stem from latent
desire. In the context of the hierarchy, individuals ﬁrst need the basics of life, such as
food, water, and shelter. After those needs have been met, humans will pursue wants at
higher levels in the hierarchy. According to Campbell (2006:56), “The postulation of
such a hierarchy, is obviously an attempt to cope with the problem of the non-universality
of ‘instinctivism’ desires, that is, that the fact that the whole of humanity does not display
one common set of consumer wants.” Explanations of consumer behavior along this
thread argue that traditional'people were too preoccupied with obtaining the wants at the

base of the hierarchy and it has only been with the broader satisfaction of these needs that

17

new desires have become manifest. Campbell concludes that these explanations are
dubious at best. He points out that human beings will often “override the imperatives of
biological urges for the sake of a ‘higher-order’ need such as love or self-respect (56).”
In understanding consumer food choices, Campbell’s critique is apt. McIntyre et al.
(2003) for instance found that low-income mothers would give up food for their children.
This provides one example of how the biological imperative of food may take second
place to the social need to provide for one’s children.

Marginal utility theories are based upon the idea that consumers pursue items and
activities that maximize their utility. While Campbell concedes that this perspective has
provided some theoretical utility, it does not constitute a theory of consumer behavior. In
particular, he notes the failure of theories in this category to successfully explain the
nature or origins of wants and tastes. Furthermore, the failure of this perspective to
explain “irrational” consumer behavior, such as excessive consumption, limits its utility.
Given the recent growth of health problems related to excessive food consumption, it
would seem that this line of theorizing has only limited utility when applied to food.
Rappoport (2003) provides several accounts of “irrational” consumer decisions that
derive from cultural background rather than nutritional utility. Both the refusal of some
foods, such as preserved ﬁsh, to embracing other foods such as cat, highlights the limited
applicability of marginal utility theory to food consumption.

Manipulation theories of consumer behavior are based on the idea that wants are.
created by the actions of outside agencies. In short, people want things because other
groups, generally advertising agencies, have created the desire. Campbell (2000:57)

explains “this perspective derives from what has been called the ‘hypodermic’ model of

18

the workings of the mass media, one which implies that the various media of modern
society—such as ﬁlm, television, and newspapers—each function like a hypodermic
needle to inject a given message into their audience.” From this perspective, consumers
play only a passive role. Wants and desires are generated by others. This set of theories
may carry some weight in explaining consumer behavior. The growth of food related
advertising certainly indicates that food companies see this as a possibility and there is
evidence that at least children are being inﬂuenced by food advertisements (Harford et al.
2004). Still, the colossal failure of heavily marketed products, such as New Coke in the
1980’s indicate that simplistic notions of passive consumers may be overstated.

This typology of consumer behavior provides us with a background for
understanding current theories related to food consumption. As I will discuss in the next
section, despite the breadth of theories stemming from the consumption literature,
explanations for food consumption have been largely oriented towards marginal utility

theories and manipulation theories.

The Sociology of Agriculture and the Consumer

In the literature on food I see the greatest theoretical overlap between the
sociology of consumption and the sociology of agriculture in the latter two of Campbell’s
(2000) categories, marginal utility theories and manipulation theories. In the sociology of
agriculture, there has been a historic emphasis on consumption either. as a simplistic set
of desires from a rational group of people or as an activity that is determined by outside
agents. As Lockie (2000: 279) explains, “until recently, most accounts of consumption

within the agri-food studies dichotomized production and consumption and dealt with

19

consumption either as a set of practices manipulated by capital and the state in the
interest of capital accumulation or as the simple agglomeration of individually free and
rational choices.” This dichotomy continues as an inﬂuential part of sociology of
agriculture accounts of consumer behavior and mirrors, albeit not perfectly, the
dichotomy represented by Campbell’s. typologies of marginal utility theories at one end
and manipulation theories at the other. In terms of agency and the food system,
consumers are either very inﬂuential or lack any ability to inﬂuence the structure or
content of the system. Interestingly, while past studies have been more oriented towards
production and a lack of consumer agency, recent studies have moved eloser to the other
end of the spectrum and attributed consumers with a greater amount of agency (Lockie
2000.)

This has likely been inﬂuenced, in part, by the idea of reﬂexive modernity
discussed by Beck (1992). Sociology of agriculture scholars have begun to recognize
that consumers are active and reﬂexive — both interested in and able to inﬂuence the
structure of the food system. This reﬂexivity is apparent in the recent, but rapidly
growing concern of consumers with issues of food safety. Busch (1997) suggests that the
recent growth in concern with food safety has only developed in the past 50 years.
Beardsworth and Keil (1997) also note the growth in concern with food safety that
characterizes modern society. They attribute some of this concern to the erosion of
traditional ways of managing the food paradox (it must be eaten, but eating can also
cause harm). They state “many of the structural and ideological features of the modern
food system have the effect of raising the visibility of the paradoxical nature of food and

eating (160).” Similarly, Senauer et al. (1991) note the growth in consumer concern with

20

food safety and the health consequences of eating. 'At times concern with food safety is
punctuated by food scares such as outbreaks of salmonella, E. coli, or bovine spongiform
encephalopathy (mad cow). The growth in these concerns has been elaborated in other
studies (Green, Draper, and Dowler 2003; N ygard and Storstad 1998; Sapp and Bird
2003).

This reﬂexivity is not simply a matter of concern with risk. Zukin and Maguire
(2004) found trends towards reﬂexivity in their review of recent consumer research.
They contend that much of this literature shows people actively engaged in consumption
in ways that contribute to their sense of identity. In addition, they suggest that many
people think of consumption as a creative activity for self expression. While this is not to
say that elements of this satisfaction are not in part the result of marketing, it does
suggest that people are to some level consciously engaged in consumptive practices.

Associated with this growing consumer reﬂexivity is a willingness and ability of
consumers to use their consumptive practices to promote change in other areas. Drawing
from geography, Crewe (2001) notes that consumers increasingly are not only engaged in
reﬂexive consumption, but that this reﬂexive consumption has increasingly political
implications. She states that “there is evidence of a shift in focus during the early 2000's
towards consumption as a site of political action through explorations into the
connections between commodity culture, self-identity, citizenship and political
participation (632).” This builds upon Zukin’s conclusions that consumption has become
an element of identity to suggest that it is also one of social change. Miller (1995) also
takes the position that consumers have a very active and powerful role in shaping society.

He states, "Consumerism as an activist movement remains one of the most powerful

21

critical points in assessing the consequences of capitalism (40)." For Miller the power of
consumers is one of the major driving forces of society.

This growing consumer reﬂexivity is intricately related to recent changes in the
agrifood system. Industrial agriculture has created environmental and health problems
that have become difﬁcult to entirely ignore. Highly publicized environmental problems
associated with agrichemicals, such as DDT, highlighted the risks associated with the
modern food system. Such chemicals were an integral part of Fordist production
strategies. At the same time, medical developments around the relationship between diet
and diseases such as cancer, heart disease, diabetes, and obesity heightened concern
about the types of food consumed. Interestingly, several of these health problems have
ties to meat, processed grain, corn syrup, and other food products that tended to easily ﬁt
into the rationalized mass production of Fordism. These health concerns were further
inﬂuenced by the growing role of the government in promoting health (Nestle 2002).

According to Beardsworth and Keil (1997: 148), the growth in consumer concern
with the relationship between health and food can also be traced to the social movements
of the 19603 where health food was promoted as a sort of “alternative cuisine.” By the
19803 and ‘90s these movements had grown and found legitimacy in the mainstream in
the form of heightened concern with the health consequences of food, though I also
recognize the efforts of other social movements, such as the environmental or labor
movement, in making important contributions to publicizing issues related to food
(Beardsworth and Keil .1997.) In the case of both environmental and health concerns,
growing consumer reﬂexivity can be seen as closely related to problems that resulted

from conventional agricultural production patterns.

22

Perhaps the most obvious recent example of peoples’ efforts to alter the structure
of the food system through their role as consumers can be seen in the Fair Trade
Movement, within which Fair Trade coffee represents the largest product. This
movement seeks-to create new linkages between producers in developing countries and
consumers in developed countries. While these linkages are in part economic in that
commodities are being exchanged, they also carry with them certain values related to
social justice. Third parties, such as Transfair USA, certify and label products as Fair
Trade based on a set of social standards. Consumers in developed countries spend their
money on Fair Trade products because they believe these products are produced in a
more socially just fashion. Some advocates of the movement would even suggest that the
more money that is spent on Fair Trade the more socially just the world will be. While
Fair Trade is still a very small market, its recent growth and inclusion into large
transnational corporations such as Starbucks suggests it is one way in which consumers
are having an inﬂuence on the food system (Renard 2003).

While there appears to be evidence that scholars are giving greater credence to the
role of consumers in the food system, theorization of consumption remains a highly
contested area (Fine, Heasman, and Wright 1996; Goodman and DuPuis 2002; Lockie
2000). Part of the difﬁculty stems from the traditionally production oriented nature of
agricultural studies in the past thirty years. The historic orientation of the sociology of
agriculture on production has led to an orientation towards a manipulation explanation of

consumer behavior. Such an emphasis has made acknowledgement of the active role of
the consumer in agrifood systems difﬁcult. A production/consumption dichotomy still

exists that gives agency to producers and ignores the agency of consumers. Fine’s

23

Systems of Provision (Fine et al. 1996) represents one attempt to move beyond a simple
production/consumption dichotomy. However, as Guthman (2002) and Lockie and Kitto
(2000) point cut, production is still typically treated as the driving force behind consumer
demand. Goodman and Dupuis (2002) and Lockie (2000) have sought to apply actor
network theory to the puzzle of consumer behavior and while such attempts are
promising, they generally treat consumer demand as either given or the result of
production. Other scholars, such as Appadurai (1986) and Miller (2000)'have looked at
consumption as an issue of material culture; however, as Goodman and Dupuis (2002)
caution, such perspectives should not preclude the inﬂuence of production. Bringing
together these divergent perspectives of consumption represents a signiﬁcant obstacle to
a meaningful understanding of the inﬂuence of consumers on and within the agrifood
system.

Warde (1997) frames current debates around food consumption as involving four
different trajectories: individual diversity, market segmentation, massiﬁcation, and
structural division. These trajectories represent different explanatory directions in the
literature on food consumption. The individual diversity trajectory, according to Warde,
stems directly from Beck’s risk society (1992) and encompasses some of the recent
agrifood studies interested in food safety and focused on growing consumer reﬂexivity
and anxiety. The market segmentation trajectory suggests that there is an increasing
differentiation of distinctive lifestyles that are evolving out of the era of mass

consumption. Warde points to analyses of post-Fordist trends in economic production
8 uch as Harvey (1989) as representative of this thread. Literature in this area suggests

that the food system is deﬁned by greater nichiﬁcation and greater diversiﬁcation.

24

Directly opposite this thread, massiﬁcation studies have a cultural orientation and focus
increasingly on ubiquitous cultural elements, perhaps best represented by Ritzer’s (1996)
McDonaldization thesis. Finally, literature focused on structural elements argues that
social divisions are becoming more intense. Warde notes that these divisions are not
simply based on socio-economic class, but also nationality, ethnicity, race, and gender.
Perhaps the most widely read recent work along this trajectory would be Bourdieu
(1984).

While Warde (1997) drew largely from the sociology of consumption in
constructing this framework, I would offer that these four trajectories can also be used to
situate literature examining consumer agency and the agrifood system. Understood in
this manner I am able to apply axes representative of two major tensions that exist within
current literature on consumer behavior and fundamental to sociology: that of structure
versus agency and that of production versus consumption (Figure 1). Framed in this way,
I see individual diversity and massiﬁcation as representing literature related to
consumption with the former allowing for greater individual agency than the latter. At
the same time, studies along the market segmentation and structural division trajectories
tend to situate the locus of power at the end of production with varying degrees of

inﬂuence allocated to the predominant social structure.

25

AGENCY

segmentation d'VGISW
-5 59999.??? __________________ Q QN§IJMEIIQN
structural massification
dwrsron

STRUCTURE

_--_-_----...-_..________..4------—-..._..___..---__-_.._

Figure 1. Determinants of consumptive practices
While this slightly altered perspective of the literature is not a substantial change
from that offered by Warde, it does provide a cohesive way of bringing literature from
the sociology of consumption together with recent work from the sociology of
agriculture. In Goodman and Dupuis’s recent review of consumption and agrifood
studies they note that many studies based in rural sociology remain “critically impaired
by the continued reliance on production-centered theoretical frameworks (2005:5).” For
Goodman and Dupuis the trend for work in this area has been to place explanations on
the left side of Figure 2. These authors highlight Marx-inspired work along the lines of
Friedland’s (1984) commodity systems analysis as being a major theoretical thread
attempting to bring consumption into agrifood studies. More recently, Fine’s (2002)
systems of provision represents another attempt to reconcile power of production and that
of consumption. While both represent meaningful theoretical developments in their

expansion of agrifood studies “beyond the farm gate,” both have been clearly situated at

26

the production end of the spectrum by placing power predominantly in the hands of
producers (or at least in the hands of actors at that end of the commodity chain such as
processors). Dixon (1999) builds upon Friedland’s commodity system analysis and
attempts to incorporate consideration of consumption to a greater degree with her cultural
economy model for agrifood studies. I would suggest that this perspective comes closest
to escaping the production orientation of past models by explicitly incorporating the
conditions of consumption that are so inﬂuential on consumption decisions.
Interestingly, the growth in scholarly interest in consumers and food has led to a
growth at the other end of the production/consumption spectrum. That is, there has
developed a turn towards looking only at culturally derived changes to explain the
structure of the food system. Goodman and Dupuis (2000) attribute this to the cultural
“turn” that has taken place more broadly in the social sciences. One of the most
inﬂuential studies noted by Goodman and Dupuis is Ritzer’s (1996) McDonaldization of
Society. Studies in this category focus on commodities as elements of material culture.
Consumption is seen as a way of creating and recreating cultural values and identities.
Irnportantly, for our discussion here, studies in this area have little, if any, consideration
for the inﬂuence of production. This, for Goodman and Dupuis (2000: 14), represents a
sticking point as they caution against “overshooting the cultural ‘turn.’”
But is there no way of reconciling the taste and preferences at the site of the
consumer while giving appropriate consideration to the inﬂuence of producers? In order
to reconcile culturally derived tastes of consumers with the power of producers, we must
look beyond simplistic conceptualizations of culture. Bourdieu ( 1984) provides us with

one means of reconciling production and consumption in his efforts to bring together

27

economic and cultural elements of society. As Bourdieu (1984: 6) famously states
“[t]aste classiﬁes, and it classiﬁes the classiﬁer.” For Bourdieu, tastes are socialized,
largely by social class, and expression of tastes further serves to identify social class in
society. Taste then, can be understood as something that stems from the class divisions
caused by a capitalist economic system, but also as the cultural reiﬁcation of this system.
Culture then becomes a system for understanding the symbols imbedded in commodity
tastes related to class.

Of course, we are not necessarily aware of how tastes have been dictated by our
class. Bourdieu offers the concept of habitus as a way of relating taste as dictated by
class as well as manifestations of class. Swartz (1997: 117) explains Bourdieu’s habitus
stating “If on one level habitus can be read as a way of conceptualizing culture as
practice, on another level it associates practice with habit.” Our tastes are both what we
are used to as well as the means through which we decode or apply meaning to the
objects in our world.

Bringing together culture and economic systems as they relate to tastes is
important as it allows us to incorporate culture into understanding tastes without losing
the inﬂuence of producers in the food system. Bourdieu is speciﬁcally concerned with
meanings that are attached to material goods, such as food. Consequently, we must
concede that cultural tastes and preferences as class manifestations are also limited by the
availability of material goods. For those in disadvantaged classes this can mean “a taste

for what they are anyway condemned to (Bourdieu 1984: 507).” Wright, Nancarrow, and
Kwok (2001) go even further to note the relative abundance and scarcity of foods can

inﬂuence the meaning that classes apply to food. They note, for example, the shift in

28

perception of ideal body weight. At one point, being heavy was considered a mark of
high class, but is now considered more symbolic of the poor.

Food producers and retailers have a great deal of control over the relative
abundance and price of food in different locations and therefore they have some control
over the taste and preferences for different types of food. As Wright et al. (2001: 335)
explain “Supermarkets stock food which reﬂect the tastes of their consumers whilst
simultaneously stimulating their taste buds in new directions.” Much in the same way
that DeBeers’ near monopolistic control over diamond production has had an inﬂuence
on the class related cultural meanings associated with diamonds, it seems reasonable to
note that producers are able to inﬂuence preferences by changing prices and relative
availability. Organic produce, for instance, carriers a price premium. This both limits its
availability among lower classes as well as creating cultural meaning associated with
desirable elements of higher classes.

Certainly, as Bourdieu and others suggest, there is a relationship between taste
and class. However, the relationship between the two is more complicated than simple
high-brow and low-brow tastes. As Peterson and Kern (1996) note, there has been an
increase in consumption activities that appear to transcend class boundaries. They argue
that music traditionally associated with “low brow” tastes, such as bluegrass, is now
being consumed by upper class groups. What is the nature of the relationship between
class and tastes as it pertains to food? In terms of understanding the causes of food
insecurity it seems a stretch to suggest that people “prefer” not to have sufﬁcient healthy

food to eat. In order to also understand the causes and consequences of food insecurity,

29

we must consider not only tastes and preferences, but also agency and the extent to which

households are able to negotiate the structure of the food system.

Agency and the Consumer

I applaud the efforts of scholars to reconcile the consumption/production debate
in such a manner as to understand how different actors are linked along the spectrum.
However, I think that the project of “theorizing the consumer” in agrifood systems is
hindered by a failure to fully acknowledge the second axis—that of structure versus
agency. My modiﬁcation of Warde’s (1997) explanations of changing consumption
overlaps along the horizontal axis with existing analyses of agrifood studies (Figure 2).
However, the addition of the vertical axis highlights an issue rarely considered in
agrifood studies: the fact that actors, even in the same part of the production/consumption
spectrum, have different amounts of agency in terms of operating in, and enacting change
within, the conventional agrifood system. A small farmer and Monsanto might both exist
at the producer end of the production/consumption spectrum, but both do not have equal
inﬂuence on the structure of the food system. The same can be said of consumers.

What does it mean to say that consumers have different amounts of agency? In
order to answer this question, it is worth giving some consideration to what is often
meant by structure. Despite its prominence in sociological literature, it is a somewhat
elusive and problematic concept. Sewell (1992), for example, notes that structures in
social science literature often appear as things that are impervious to human agency and

which exist outside of humans. As he explains, this approach to structure is problematic

as it is too rigid in its determinism and makes social change difﬁcult to understand. This

30

is not to say that social structures do not exist, as individual behavior is inevitably shaped
by the resources and ideas that are set before us. Sewell suggests that structure is better
understood as two-dimensional. One dimension involves a depth dimension (schema)
and the other involves a power (resource) dimension. The depth dimension refers to the
extent to which the ideology related to a structure pervades a society. The power
dimension refers to the extent to which resources are allocated based on structure.

These two dimensions can be easily applied toward an understanding of the
structure of the food system. On one side the modem food system carries a signiﬁcant
amount of depth and, through advertising and other vectors, has promoted a set of tastes
that inﬂuence not only what people eat, but more generally what people consider to be
worthwhile qualities of food. Interviews with both food secure and insecure individuals
that were conducted as part of this study revealed widespread understanding that organic
food was “better” than other food, even if respondents’ understanding of organic varied
widely. Similarly, many respondents noted the strong desire, especially among children,
for food from McDonalds, regardless of an awareness of the nutritional problems
associated with fast food. In both examples the food system strongly inﬂuences symbolic
associations with food. In the case of organic, the association of “organic” with “higher
quality” took place even without an understanding of the organic certiﬁcation process.
Similarly, McDonalds was seen as desirable and good, even when respondents were
aware that many McDonalds products were not healthy.

In terms of the second dimension of structure -- power -- the food system has a
signiﬁcant inﬂuence on the material allocation of food. One’s ability to obtain a

particular type of food is certainly limited by its availability where food is typically

31

purchased. While additional options might be available such as getting food through the
mail or driving farther to obtain it, those activities require greater resources. Even when
desired food is available in an area, some individuals might not have sufﬁcient resources
to gain access. While in a sense the food system, as with other largely economic
structures, constrains or enables people’s ability to act based on how much money they
have, it is important that we understand resources as including, but not limited to,
ﬁnancial ones. Gardening, food sharing, and food assistance all represent ways that food
can be procured that require resources, but not necessarily money.

Hays (1994) provides us with a way of relating this multidimensional
understanding of structure to the issue of agency. Importantly, Hays makes the point that
agency is not the same thing as freewill or randomness. He states:

Social structures are simultaneously constraining and enabling: although

structural constrains absolutely preclude the possibility of making certain

choices, they also provide the basis of human thought and action, and

therefore offer the very possibility of human choice. A sociological

understanding of agency, then, does not confuse it with individualism, .

subjectivity, randomness, absolute freedom, or action in general, but

recognizes it as embracing social choices that occur within structurally

deﬁned limits among structurally provided alternatives (65).

Such a conceptualization of agency is useful in understanding the role of consumers in
inﬂuencing the food system because it avoids simplistic explanations whereby some
individuals are completely constrained by a structure and others are completely
empowered by it. Rather, agency becomes an indicator of the extent to which consumers

are able to alter one social structure to coincide with another structure. In the case of this

study, it allows us to examine the extent to which the structure of the conventional food

32

system coincides with the social structures which dictate food purchasing preferences of
consumers.

. Food security status represents a useful indicator of agency since it allows us to
examine this concept in a broad fashion to the structure of the conventional food system.
After all, it seems reasonable to assume that nearly everyone in a population would like
reliable access to quality food. Most people prefer to be food secure. Admittedly, food
security status is partially a subjective measure leaving it up to the respondent to
determine what constitutes a “balanced meal.” Even questions related to whether a
respondent had “enough money” for food require a subjective assessment on the part of
the respondent. The subjectivity of the food security questions is useful because it allows
the respondent to bring to bear all of the social structures (class, gender, ethnicity,
education, etc.) which inﬂuence taste and preferences into a measure of the adequacy of
food access. Food insecurity, therefore, becomes an indicator of constrained agency
within the conventional food system because the ability of food insecure households to
obtain food in preferred quantities over time, without sacriﬁcing resources that they
would rather allocate elsewhere, is constrained by the existing structure of the food
system.

This does not mean that those who are food insecure completely lack agenCy or,
in many cases, the ability to access food through other means. Rather it indicates that the
food system, as represented by food retailers, is a social structure that food insecure
households are unable to inﬂuence to ﬁt the other social structures that inﬂuence their
food purchasing and consumption preferences. This represents a deviation from many

agrifood studies which look at how consumers, as a general category, are able to

33

inﬂuence the structure of the food system. Consumer agency is often perceived through
conventional lines as consumers engage with increasingly concentrated food retailers
(Marsden, Flynn, and Harrison 1997). In some cases agency is seen in terms of new
consumer demands for products such as organic produce. In other cases, consumer
demand for food that was produced in a fashion to meet certain social standards has led to
the development of altemative‘networks such as the case of Fair Trade. Finally, the
growth of locally based alternative marketing arrangements, such as farmers’ markets and
Community Supported Agriculture, is commonly associated with consumer interest in
supporting local growers. With few exceptions, e. g. Guthman (2002), accounts of
consumer agency in the agrifood literature treats consumers as homogenous. There is
little, or no, consideration for the fact that some consumers are (often because of resource
constraints) unable to reorient the food system to coincide with their interests. After all,
if all people were able to alter the structure of the food system in a similar fashion, the
approximately 10% without reliable access would disappear.

It is important to note that the focus of this study on agency in the context of the
conventional food system, as represented by food retailers, does not suggest the denial of
the existence of agency in other social contexts. In other words, to have or not to have
agency in the conventional food system does not preclude other forms of agency in other
contexts. Even food insecure households have at least some opportunities for political
activity that could, for instance, inﬂuence zoning laws that determine where food retailers
are located. Place based relationships with neighbors that lead to the exchange of garden
produce could help alleviate food insecurity problems while taking place external to the

conventional food system. As a more tangible example, the non-proﬁt community center

34

located in the neighborhood has recently built a greenhouse in an underutilized
neighborhood park to help neighborhood residents learn to grow food on their own.
These all represent ways in which the agency of neighborhood residents has been
manifest in the context of social structures other then that of the conventional food
system. Still, everyone in the neighborhood engages with the conventional food system

and on a regular basis must negotiate their agency within the structure of this system.

Space, Agency, and Structure of the Food System Structure

What does it mean to talk about the structure and agency within the food system?
While there are a variety of characteristics of the conventional food system, recently,
sociologists and geographers have begun to examine the physical structure of the food
system as manifest in retailer locations. The spatial distribution of food retailers provides
an excellent vector through which to understand the intersection between structure of the
food system and the agency of consumers. On one level, people have no inﬂuence on the
distribution and array of food retailers in their area (a few communities have been able to
keep Wal—Mart out of their neighborhoods, but these are exceptions to the rule). It
represents an element of their environment that they must inevitably negotiate despite
their relative inability to inﬂuence its structure. At the same time, the distribution of food
retailers in an area is strongly inﬂuenced by the aggregate characteristics of the
population. Food retailers make location decisions based on physical characteristics,

such as the availability of a building and sufﬁcient land for a parking lot, as well as

35

population characteristics, such as the purchasing power of a neighborhood.1 After all,
stores that are unable to sell sufﬁcient products will eventually go out of business.
Therefore, space is an element of social structure, but nonetheless one that is partially
inﬂuenced by aggregations of individuals. The difﬁcult task is in understanding how
these two elements intersect. In order to do this, we must ﬁrst examine the sociological
implications of space.

Consideration of space in sociological literature has both a long and short history.
It is long in the sense that sociologists have been working in space regardless of subject
matter, as sociology studies almost invariably take place somewhere. Furthermore, space
has often played an extremely signiﬁcant role in both‘ bounding the ﬁeld and guiding
research directions in subﬁelds such as rural and urban sociology. At the same time,
space has only recently been recognized as having a signiﬁcant and deﬁnable role in the
larger sociological enterprise (Gieryn 2000; Tickamyer 2000). This new recognition of
the importance of space and place has come about both by a new awareness of the role of
places in processes such as globalization (Swyngedouw 1997) as well as critical
developments in subﬁelds that have led to questions regarding the content of those ﬁelds.
This has been particularly notable in the case of urban (Fischer 1975) and rural (Lobao
and Rogelio 2002) sociology.

Gieryn’s (2000) seminal paper on place in sociology represents a useful point for
understanding the role of space and place in sociology. In this paper, Gieryn states that
place has three features: geographic location, material form, and investment with

meaning and value. A place must have a unique location, some compilation of objects on

 

' Purchasing power refers to a combination of neighborhood income and populations. High population
density areas with low average incomes may have less purchasing power than low population density areas
with very high incomes.

36

or associated with it, and meaning attached to the place. For Gieryn, space is simply
abstract geometries—place, on the other hand, is imbued with meaning. This distinction
between space as meaningless and place as a location of meaning is echoed by Gans
(2002). Rather than “space” as the socially-void arena, however, Gans suggests a
distinction between “natural space” and “social space.” The former is simply “air over
dirt” (Gans 2002: 379) and immediately becomes social space as soon as people use it.

The meaning that is attached to space (i.e. place) has real and meaningful impact
on the subjective experience of distance. Consider that while a store may be very close in
terms of distance (e. g. ﬁve blocks), a variety of factors such as poor road conditions, a
lack of sidewalks, a “dangerous” neighborhood, might make the trip to that store much
more difﬁcult. At the same time, even though a particular store is many miles away it
may be located on the way to one’s workplace, in a particularly pleasant area, located
conveniently close to other stores, or on a reliable bus route. All of these factors might
lead to a perception that a store is in close proximity regardless of physical distance to
one’s place of residence. In terms of understanding the role of distance on consumer
behavior, therefore, it stands to reason that understanding people’s perception of distance
will provide important insight into how they negotiate the spaces around them.

Is an understanding of place as a subjective measure compatible with objective
measures of space? Tickamyer (2000) argues that space can be conceptualized in three
ways: as place, relational units, and scale. Place refers to the particular locale or setting.
Relational units is a term which refers to the organization of ideas about places. Scale
addresses the size of the units to be compared. She states that these can be viewed as

“context, cause, or outcome for other social processes (806).” While her

37

conceptualization does move beyond a simple dualism, it really offers more of a
descriptive framework than theoretical understanding of the role that space does and does
not play vis-a-vis society.

Agnew (1989) suggests that some of the tensions that exist around place stem
from disciplinary differences around how it is perceived. He suggests that economists
and economic geographers have emphasized location (physical distribution),
microsociologists and humanistic geographers have been concerned with locale as the
site of daily action, and anthropologists and cultural geographers have focuses on the
sense of place or identiﬁcation with place (Agnew 1989:2). In many cases, these
different perspectives are seen as mutually incompatible rather than complementary.
Agnew further demonstrates how conceptualizations of space have been closely tied to
conceptualizations of community. Speciﬁcally, he traces how critiques of modernization
and capitalist development have often focused on a shift towards a “placeless” society
and the erosion of a sense of community (Agnew 1989).

Gieryn (2002:342) draws a parallel between sociological studies of space and
studies of science. He notes the tension that exists in science and technology studies
around the role of “reality” in determining social processes and cautions that a failure to
recognize the inﬂuence of space risks “unbridled social constructivism.” Similarly,
Zukin (2002) is very critical of the relatively simplistic perception of space
conceptualized, in particular by Gans. Zukin states that by emphasizing space as the site
upon which action takes place and use as the primary means of socializing space, Gans
“fails to raise the larger issue of how and why spaces are constituted for critical social

purposes (Zukin 2002: 345).” Zukin goes a step farther than Gieryn by questioning the

38

existence of any socially-void space, whether it is “air over dirt” (Gans) or space without
meaning (Gieryn). Zukin (2002: 346) warns “There is no neutral vantage point and only
partial objectivity. Whether built or only portrayed, moreover, a landscape obscures as
much as it reveals. To study only the use of space prevents us from dealing with what is
suppressed by a landscape, forcefully reshaped, or willfully forgotten.”

While I think that Zukin is correct to emphasize the importance of understanding
all spaces as socially constructed, the nature and structure of that social construction is
not without limits. Materiality still matters and some of the limits to how place is
perceived are physical. Measures of physical distance (space) still matter. This can be
understood on a practical level—there is, after all, only a limited distance that people can
practically travel in order to purchase food from a desired store or to obtain the desired
selection of food. One cannot simply perceive a store to be close enough without the
time and resources to negotiate the necessary distance. Therefore, while it may be
impossible to fully understand space without consideration of the meaning that people
associate with it, that does not obscure the complimentary task of objective
measurements of space.

Reconciling subjective distance (place) and objective distance (space) is
fundamentally related to our central question of food retailer perception and choice. First
of all, in both interviews and the survey, respondents indicated that distance was an
important factor inﬂuencing store choice. The question is the extent to which distance
matters. Answering this question requires both a subjective component (i.e. assessing the
extent to which perceived distance is seen as an important characteristic of store choice)

as well as an objective component (i.e. assessing the extent to which physical distance

39

inﬂuences store choice). By understanding both these components a fuller understanding
of the relationship between space and consumption can be gained. Fortunately, recent
scholarship provides us with new ways of understanding the relationship between space
and consumption. First, however, we need a basic understanding of how food retail

spaces have evolved in the recent past.

Space and the Structure of the Food System

The evolution of food retailers has followed much the same expansion and
concentration as other retail segments. In the ﬁrst half of the 20th century, small grocers
were the standard food retailer. Selection was limited and the shopping experience was
very service oriented with the grocer selecting items for shoppers. By the middle of the
century, supermarkets rose to prominence, replacing the smaller grocers with a retail
format that allowed consumer to browse and select items themselves. Writing during this
period, Zimmerman (1941: 402) estimates that from the 19303 to 19403 50% of
consumers had switched from a “the long established habit of service, credit and delivery
to that of self-service, cash and carry.” Not only did supermarkets provide a wide
selection of items, Zimmerman explains, but they also had a signiﬁcant impact on the
space of the cities and towns in which they located. In contrast to small grocers, which
were typically located in downtown or city center areas, he states, “The character of the
supermarket makes central location not important. In fact, many of the ﬂourishing
supermarkets are on the fringe of towns, where rentals are low, but where they can attract
the motor trade (404).” Zimmerman concluded that these stores would expand in size to

approximately 10,000 square feet in the Midwest and begin to incorporate more and more

40

product lines such as lunchrooms, soda fountains, drug departments, notions, novelties,
and magazines (409).

Zimmerman was correct that food retailers would get bigger and diversify, only to
a much greater extent than be anticipated. In the 19803, food retailing went through
another signiﬁcant shift from large supermarkets to superstores and hyperrnarkets. No
longer focusing only on food, these stores offer an even wider variety of food and
nonfood items in stores even larger than seen before. Kaufman (1995) describes the shift
through the 19803 and early 19903. He notes that ﬂoor space in supermarkets grew
during this period from 23,000 square feet to 35,000 square feet. At the same time, the
variety of items available also expanded from an estimated 14,000 items in 1980 to
25,000 items in 1993 (Kaufman 1995:26). While conventional grocery stores continued
to be inﬂuential in the 19903 capturing 50% of grocery sales this had fallen from 80% in
1980. This growth in superstores was inﬂuenced in no small part by the growth of Wal-
Mart’s supercenters during this period and by 2003 one out of every ﬁve dollars spent on
food in the United States was spent at Wal-Mart, with sales greater than the other top four
supermarket chains (Tillotson 2005).

Wrigley (2002) traces many of these changes to a consolidation and acquisition
wave that swept across the United States food retailing industry in the 19903. A decline
in power of anti-trust legislation in the 19803 was a contributing factor while the
development of information technologies that facilitated centralized control also helped
the process. Wrigley also notes that the rapid expansion of Wal-Mart’s supercenters in
the late 19803 and into the 19903 also put pressure on other food retailers to “get big” in

order to compete. Wal-Mart’s supercenters led to more frequent shopper visits and,

41

consequently, greater proﬁts. The success of this model, which was actually pioneered
by Meijer and Fred Meyer, put pressure on other food retailer to do something similar.
The expansion in ﬂoor space has had some clear implications for the location of
stores. Much as Zimmerman noted the movement of food retailing to urban fringes in
order to obtain land necessary for larger stores, so too have superstores been developed
increasingly farther from city and town centers in order to ﬁnd sufﬁcient space to locate
these stores. The issue is not only one of sufﬁcient land, but also convenience. Even
when sufﬁcient land is available within a city’s limits, obtaining that land often requires
negotiating with 25 to 30 landowners, each with a different set of interests and
willingness to sell. This can translate into taking two to three times longer to develop a

store within city limits than in the suburbs (Ehrenhalt 2006).

From Supercenters to Food Deserts

Of course, the growth of large food retailers does not represent a critical element
of the structure of the food system because of where they are located. They are also a
critical element because of where they are not. And lack of food retailers, or particular
types of food retailers, has become part of a new wave of interest in food security that
looks, not only at individuals and households, but also at the entire community. Moving
away from calorie focused notions of food security that were the norm in the 19603, there
is new consideration that includes how the physical structure of the food system may be
limiting access to some populations and leading towards greater food insecurity. As
Anderson and Cook (1999) note, in the past ten years the concept of food security has

shifted almost entirely towards concern with reliable access, affordability, and quality.

42

This shift took place through cross fertilization with areas such as agriculture and the
environment, nutrition, and community development. This changing conceptualization of
food security was important because it opened up discourse around food and poverty to
include a wider range of considerations for political, economic, and environmental
factors that may be contributing to food insecurity.

This change in orientation is highlighted by the evolution of the community food
security movement, which is working to develop communities where “all community
residents obtain a safe, culturally acceptable, nutritionally adequate diet through a
sustainable food system that maximizes community self-reliance and social justice.”2
Notably, this movement includes concern for the nutritional quality and cultural
appropriateness of diets as well as the environmental and sOcial impacts of food
production practices. Rather than simply focusing on the ability of individuals to access
sufﬁcient calories or macronutrients, there is broader consideration both for the needs of
individuals as well as indirect social and environmental impacts of production. Some
scholars have noted that the concept of community food security is problematically vague
(Anderson and Cook 1999; Pelletier, Kraak, McCullum, Uusitalo, and Rich 1999).
However, this is as much a consequence of its novelty as it is of its wider goals.

The expansion of food security concerns can be seen as part of the growing
alternative agriculture movement. This movement has evolved out of increasing concern
with the impacts of the conventional food system on the environment, health, and
communities. This includes, among other things, elements such as organic agriculture,
local food, Fair Trade, and other efforts to reembed food production and consumption in

something more than market relations (Hendrickson and Heffeman 2002). Community

 

2 www.foodsecurity.org

43

food security is one element of this movement that has focused on issues of hunger and
equityias these pertain to the food system (Koc and Dahlberg 1999). One element of this,
of course, is access to food and, in particular, the role of space in deﬁning access.
Drawing from widespread concern with suburban sprawl, urban ﬂight, and food
retailer concentration, the concept of food deserts was developed in the 19903 by a
working group for the Low Income Project Team of the Department of Health in the
United Kingdom (Lang and Caraher 1998). The concept was meant to address issues
related to the closing of food retailers in inner cities and the associated growth of large
supermarkets in suburban areas—often beyond convenient access of low-income, inner
city residents. In the decade since its inception, the concept of a food desert has sparked
off a series of related research (Cummins and Macintyre 1999; Donkin, Dowler,
Stevenson, and Turner 2000; Morland, Wing, Roux, and Poole 2002; Whelan, Wrigley,
Warm, and Cannings 2002). Much of this research has involved the use of GIS to
examine the relationship between food retailer distribution and low-income populations.
In some cases, consumer surveys and price indexes have been incorporated into the
research. Additionally the concept has been applied to rural areas (Blanchard and Lyson
2002; Furey et al. 2001). Although the coining of the phrase marked a new recognition
of the role of physical access in food security, the inequalities inherent in food retailers
had been examined to a limited extent in previous decades (Guy 1985; Hall 1983).
Furey et al. (2001: 447) deﬁne a food desert as “an area where people do not have
easy access to healthy, fresh foods, particularly if they are poor and have limited
mobility.” In contrast to traditional conceptualizations of food security, the emphasis is

not on the characteristics of households, but rather the physical distribution of food

retailers. An important critique underlying much of the literature is that the food system
is structured in such a fashion as to negatively impact vulnerable populations. While this
structuring is deliberate as big-box retailers position themselves in suburbs where large
tracts of land are available and populations have sufﬁcient purchasing power to support
these retailers, the consequence is that inner-city populations are at a greater distance
from certain food retailers, making food purchasing difﬁcult for some. The point is that
the food deserts literature stresses deﬁciencies in the modern food system rather than
deﬁciencies of food insecure individuals.

Studies have been conducted that support the hypothesis that physical access to
food is unequal. Moreland et al. (2002) examined the location of food retailers in 1
Mississippi, North Carolina, Maryland, and Minnesota as compared to US Census data.
They found large numbers of supermarkets located in wealthy, predominantly white areas
and relatively few in poor, black neighborhoods. Interestingly, they also found higher
concentrations of places that sold alcohol in the poor, black neighborhoods. A similar
study was conducted in New York (Hall 1983) which found that consumers in
neighborhoods with high concentrations of black and elderly people face higher prices,
lower levels of quality and cleanliness, and less variety of prices, brands, and sizes. More
recently, Blanchard and Lyson (2002) found that poor rural residents, when compared
with higher income residents, were often located disproportionately far from large
supermarkets.

Alwitt and Donely (1997) found that poor zip code areas in Chicago have fewer
and smaller retail outlets overall than nonpoor areas, including fewer supermarkets,

banks, and large drug stores. However, they note that “it may be misleading to claim that

45

poor areas are underserved, relative to nonpoor areas, unless one controls for differential
levels of purchasing power. Even though there may be fewer stores of various types in
poor neighborhoods, there may be or may not be a reasonable number given the sales
expected from those neighborhoods (143).” After controlling for purchasing power, they
found no difference in number of supermarkets. This highlights the point that the actions
of the retailers make economic sense while at the same time generating social problems
related to food access. Furey et al. (2001) also found mixed results in their study.
Although prices for some products were higher in small grocers than in large
supermarkets, the differences were often only minor. Furthermore, most of the people in
their survey were satisﬁed with the availability of food. However, much in line with
other studies, segments of the population that had special needs, such as elderly
populations, were still very concerned with physical access.

The point that different populations have different needs is further elaborated on
by Whelan et al. (2002). They used a qualitative approach based on focus groups to
assess a potential food desert in the UK. They found that physical access was a concern
for many of their subjects; however, the level of concern often depended on the type of
household. They conclude that access is not uniform across household groups. For
instance, mothers with younger children were more motivated by cost while elderly
subjects were more concerned about physical access due to personal mobility issues.
Donkin et al. (1999) sought to develop a culturally sensitive price index by drawing on
lists of foods commonly purchased by four major ethnic groups in London. Prices and
availability of these foods was then collected from 210 food outlets in the area. They

found that culturally appropriate and reasonably priced food did appear to be physically

46

accessible to the groups examined. They note, however, that even when food is
reasonable priced (that is, no more expensive than other places) and physically
accessible, this does not mean that all people can afford it.

Fundamentally, the issues addressed by food desert research relate to the
relationship between the physical retail environment and consumer behavior. One of the
most interesting studies in this area was conducted by Wrigley et al. (2002), which
involved a “natural experiment” when a large food retailer opened after a study of
consumption patterns in a food desert. After the opening, the researchers found a
signiﬁcant upward shift in fruit and vegetable consumption amongst those with the
poorest diets. Furthermore, many local residents shifted away from using limited
range/budget stores as a source of fruits and vegetables.

Despite this evidence, it is important to note that food deserts do not exist in all
low income areas. For instance, Cummins and Macintyre (2002b) actually found a
higher concentration of food retailers in the low-income areas of Glasgow. In many
cases, food prices were also lower in low-income areas. It may appear, therefore, that
while low-income populations are particularly vulnerable to food deserts and some inner
cities may represent areas of higher cost for food retailers, impoverished neighborhoods
and food deserts do not inevitably go hand in hand. Cummins and Macintyre, for
example, found that the types of food that drove down the average cost of food in the
low-income areas was often high-fat and high-sugar foods. Thus, while the food desert
concept might be a problematic metaphor, it still retains some utility as a lens to

understand the spatial distribution of different types and prices of food.

47

Despite the popularity and rhetorical power of the concept of the food desert, its
analytical utility has often been limited or reduced to factoids (Cummins 2002a). This, in
part, has stemmed from the 500 meter radius that was set by the UK Department of
Health as a “convenient” walking distance and which has been used in several studies
(Clarke, Eyre, and Guy 2002). This simplistic proximity indicator fails to take into
account the situation of elderly or disabled individuals who may be unable to carry
groceries 500 meters or people with cars who may think little of traveling ten or more
miles to shop for food. Furthermore, other elements of the landscape such as a dangerous
neighborhood or a highway or road without sidewalks may create barriers not apparent in
typical food desert research methods. Regularly traveled routes, such as between work
and home, may be perceived as close, regardless of actual distance.

Other problems also exist. Blanchard and Lyson’s (2002) study suffers from a
high level of spatial aggregation. While operating at the zip code level may be necessary
given data availability (something the authors admit), this can lead to problematic
assessments of the relationship between a population and retailers. Such analyses can
lead to the promotion of solutions, such as recruiting a large supermarket, which may
actually undermine existing food retailers and reduce total access (Clarke et al. 2002).

There have been other problems with studies that fail to actually obtain a
statistically signiﬁcant measure of the consumption patterns of people at various
proximities to food retailers. Rather, assessments are done either qualitatively (Whelan et
al. 2002) or conclusions are drawn from census data (Alwitt and Donley 1997; Donkin et
al. 2000; Morland et al. 2002). Simply because there are few grocery stores in low-

income neighborhoods does not lead to the conclusion that these populations must be

48

either having difﬁculty accessing food or compromising the quality of the food that they
eat. It is important that assessments of food retailers be compared against actual surveys
of consumer behavior. Otherwise, unreasonable generalizations regarding relationships
between retailer distribution and consumer behavior may be drawn based on class-based
generalizations and vacant data on actual consumer behavior. It also overlooks the
strategies that people have developed to purchase food, even when it is not readily
available in their neighborhood.

Despite these problems, the food desert concept represents an interesting turn in
assessing consumer behavior. Rather than attributing behavior to either independent
action or the dictates of capital, food desert literature recognizes that different consumers
have various levels of ability to engage in independent action and different levels of
accountability to the interests of capital. Unfortunately, these levels of agency vis-a-vis
the food system have not been closely examined. Rather, consumers are lumped into two
different categories: disadvantaged consumers and everyone else. The former category
lacks agency, even to travel 500 meters, while the latter feels few structural constraints on
purchasing decisions.

While the existence of food deserts in inner cities is not inevitable, a growing
body of research demonstrates that the structure of the environment can have a real
impact on the food that people select. French, Story, and Jeffery (2001) review a wide
variety of social changes that they argue have contributed to nutrition problems in the
US. Among other changes, they note the increased convenience of food with poor
nutritional quality which has led to additional health problems. Sobal and Wansink

(2007) argue that the spatial relationships inﬂuence food consumption in unconscious

49

ways. They note how the distribution of food items on a plate, on a table, or in a kitchen,
can inﬂuence the types and amounts of food that people consume. In part, the new
interest in the relationship between the environment and nutrition stems from concern
with obesity rates and related health problems in the US. In addition, the relationship has
implications for food security. After all, environments increasing the intake of some
foods are likely to reduce the consuming of others. Similarly, Moreland, Wing, and Roux
(2002) used food surveys to assess risk of atherosclerosis on people in different
communities. They compared this risk against the number of food retailers in each
neighborhood. They found that there was a relationship between the two with density of
grocery stores being related to fruit and vegetable consumption. Interestingly, the authors
found the inﬂuence of retailer density on African—American consumption to be greater
than White American consumption.

Space therefore represents one area in which the structure of the food system
overlaps with the social structures that dictate household food demands. At the same
time, recent developments in food desert literature suggest that the extent of overlap
varies, quite literally, by how far people live from a food retailer. It appears therefore
that the agency of different groups in relation to the structure of the food system may be

inﬂuenced by the physical location of food retailers.

Conclusion
Within a capitalist system, it is assumed that consumers with greater ﬁnancial
resources will have greater ability to access certain products than other consumers.

People with access to reliable transportation will have a wider range of stores from which

50

they can shop. People with children may ﬁnd it difﬁcult to dedicate the time to ﬁnd the
range of products at the price they desire, even if available in the immediate vicinity.
People with physical disabilities may struggle to obtain certain types of food, even when
they may be nutritionally beneﬁcial. Of course, we should not acknowledge such
constraints without also considering the rational, creative ways in which people negotiate
and overcome the obstacles they face. Financial limitations may be minimized by
strategies such as coupon use, sale seeking, or purchasing of bulk items. Time
constraints may be minimized by shopping only at a single store with a wide selection of
food and nonfood items. Concerns with managing small children may be mitigated by
shopping at a single store known to be friendly to children. Structures, including the
structure of the food system, can at times constrain and at times can be overcome.
Unfortunately, recognizing the differential agencies of consumers vis-a-vis the food
system is something that has received very little attention in either the sociology of
consumption or agrifood studies.

Addressing this issue of agency is important because the new focus on consumer
reﬂexivity and agency has paid insufﬁcient attention to the differential agencies among
segments of the population. If growing consumer concern with food safety and health is
driving segmentation of the food industry into areas such as organic, we must also
account for the fact that it is only a relatively afﬂuent sector of the population that is
affecting such a change. For the sociology of consumption, important questions remain
as to whether the cultural value of commodities is still associated with class position. As
Warde et al. (1999: 107) state, “[t]he question of whether there is a cultural hierarchy

reﬂecting class position and how it might have been affected by the variegation of

51

cultural items in circulation has exercised many scholars recently.” While consumption
is still inﬂuenced by availability and a variety of other factors related to class, evidence
exists supporting the idea that class is declining as a factor inﬂuencing tastes (Peterson
and Kern 1996). At the same time, however, it is clear that social class has signiﬁcant
inﬂuence on the types of food that people consume (Rose, Gunderson, and Oliveira
1998). The question is not, of course, a binary one and a clear understanding of how
much social class inﬂuences consumption warrants greater empirical research.

The impetus for research such as this stems from the fact that the intricacies of
food consumption patterns need to be understood in order to understand how economic
stratiﬁcation is created and recreated through consumption patterns. Unreliable access to
nutritious food can have a detrimental impact on populations. While in the US there is
more concern with chronic health problems such as diabetes, cancer, and heart disease
than relatively acute problems related to starvation. The growing rate of food insecurity
in Michigan -- and nationally -- points to a signiﬁcant social problem (Nord et al. 2006).
It is well understood that food insecurity can have lasting physical and psychological
problems (Vozoris and Tarasuk 2003). However, it is not well understood why people of
different economic strata are engaging differently with the agrifood system.

This study will challenge and extend the literature on the sociology of food by
building upon the question of the role of consumer agency in purchasing decisions,
focusing on differential abilities in navigating and inﬂuencing the conventional agrifood
system. Speciﬁcally, this study examines differential consumer agencies in relation to
food retailers. Since, in the US, people’s primary engagement with the food system is

through the purchasing of food at grocery stores and supermarkets, this represents a

52

poignant locus for examination. Furthermore, the ways in which characteristics of the
agrifood system, such as food retailer distribution, relate to individual behavior are not
well understood.

Looking speciﬁcally at how consumers make decisions about where to shop and
when and how frequently to shop is important because current research is struggling to
negotiate the relationship between food system structure and differing agencies of
consumers. This study will demonstrate how and why different groups’ engagements
with the food system may be related to a relative amount of constrained agency. It will
show how the conventional food system limits the choices and power of food insecure
households. At the same time, this study will also show that having constrained agency
is not the same as lacking agency. Many households manage to successfully obtain food
through conventional means. However, the costs associated with simply getting food
may not be the same for all groups. For instance, time spent with children may be
sacriﬁced in order to pour over ads and coupons to maximize shopping returns. Discount
grocers may sell cheap food, but require large amounts of gas to get there. Convenience
stores may be available to those without an automobile, but prices may be higher. The
point is that a complete understanding of consumer agency vis-a-vis the food system
requires an understanding of the act of consumption that recognizes the heterogeneity of

consumers .

53

CHAPTER THREE

STUDY SITE AND GENERAL METHODOLOGY

For four years, I have been involved in projects taking place in a neighborhood
within the city of Lansing, mainly through involvement with the Allen Neighborhood
Center. These projects have involved analysis of food security survey data, mapping and
spatial assessment of neighborhood food retailers, examination of area census data, and
qualitative interviews of neighborhood residents. In 2003, the neighborhood center
added a series of food related questions to a door-to-door health survey (Thomas 2004).
While this survey was conducted as part of a community organizing effort, and therefore
not randomly sampled, the results indicated that access to food was a problem for a
signiﬁcant portion of the neighborhood. The results of this initial survey helped stimulate
greater interest among neighborhood activists in examining issues of food.
Unfortunately, due to its lack of scientiﬁc rigor as well as the small number of food
related questions that were asked, it provided only brief insight into the factors that were
contributing to food insecurity and the food purchasing patterns of households in the
neighborhood. This dissertation was constructed to help answer the many questions that

were left unanswered in this initial survey.
Study Site

The study site examined in this dissertation is an urban neighborhood located in

Lansing, Michigan. Lansing is the capital of Michigan and hasa total population in 2000

54

of around 120,000 (US Census). Downtown Lansing is dominated by state government
buildings and other ofﬁces. This area is surrounding by several residential areas mixed
with commercial activity. Lansing is home to a community college and a law school.
Located adjacent to Lansing is East Lansing the home of Michigan State University, a
large land-grant university with over 45,000 students. As with many Michigan cities,
Lansing has struggled with the declining American automotive industry; the city had an
unemployment rate of 7.6% in 2005 (US Census). Median household income for the city
was around $34,000 in 2005, below the national median of $46,000 (US Census).

The neighborhood examined in this study is an urban area located on the east side
of the city. The land area of this neighborhood is approximately 4.8 square miles. While
the neighborhood is primarily residential, there are some small businesses scattered
throughout the area and signiﬁcant commercial activity to the north and between two
major roads on its eastern edge. The boundaries used in this study are the 48912 zip
code. This area corresponds with the service area of the Allen Neighborhood Center
(ANC), a non-proﬁt community group established in 1996 and evolving out of the
Eastside Neighborhood Organization (ENO), a 30 year-old volunteer neighborhood
advocacy organization. Many people in the area identify as being part of the same
neighborhood, in part due to the activities of the END and ANC so it was felt these were
reasonable boundaries to use. The 48912 zip code corresponds approximately to 17

census blocks on the east side of Lansing.

55

 

Figure 2: Location of the study site in relation to the state of Michigan

The 2000 US. Census provided a sociodemographic proﬁle of the neighborhood.
The total population of the study area is 18,583 with 8,740 housing units. The average
household has 2.27 people and the average family has 3.09 people. There is some racial
and ethnic diversity in the neighborhood. Seventy-three percent of the population
identiﬁes as white while 13.3% identify as African American. Just over ten percent
(10.7%) of the neighborhood is Hispanic or Latino/Latina. There is also a signiﬁcant
foreign born population (8.8%).

Even in this relatively small geographic area, a signiﬁcant amount of economic
differentiation is evident. In particular, populations in the northern part of the

neighborhood tend to have signiﬁcantly higher incomes than populations in the southern

56

area. This is evident in the median income of individual census blocks which range from
$14,069 to $57,768 in the richest (Figure 3). Largely in the southern part of the

neighborhood, signiﬁcant poverty exists and, in aggregate, 18.5% of the people in the

neighborhood live below the poverty line.

    

Study site boundaries

Median Income

I: 0- 15.000

15.001 - 30.000

- 30.001 - 45.000
- 45.001. 00.000

Figure 3: Median income for study site and surrounding census blocks

 

 

 

As noted previously, there are nearly 100 food retailers located either within or in
close proximity to the neighborhood and InfoUSA data provide a picture as to the relative
inﬂuence of each of these retailers on the food system. The largest number of which

consisted of convenience or party stores (37%), with sales volumes most often below 3

57

million dollars per year. The only grocery store located centrally within the
neighborhood is Apple Market, a relatively small grocery store with just over 8 million
dollars in annual sales volume. In contrast, the two nearby Kroger stores, where a large
portion of the neighborhood shops, have sales ranging from 12 to 16 million dollars per
year. While sales data for the closest Meijer was not available, other Meijer stores in the
area typically have in excess of 100 million dollars per year in sales volume.3

Not only did a 2003 survey indicate neighborhood food insecurity levels well
above state and national levels, the structure of the neighborhood also showed some
promise as a case study for theoretical insights regarding the inﬂuence of the structure of
the food system on different groups. The current situation of the neighborhood represents
a location that intersects among contemporary trends in retailing, consumption, and the
agrifood system. From a research perspective, the built environment of the neighborhood
provides some interesting opportunities to examine spatial inequality—in particular the
concept of the food desert as measured by proximity to food retailers.

Focusing on a single urban area allows for greater depth of analysis than a study
examining a larger geographic region. Understanding consumer perception and behavior
of the retail food system requires both an assessment of consumers as well as an
assessment of the structure of the food system. By focusing on a population in a single
area, the amount of variation in the retail environment will be limited compared to
examining a population across a wide geographic distribution. It will therefore be
possible to use geographic information system (GIS) software to relate survey responses

to the physical distribution of the retail environment. Such integration is greatly

 

3 Meijer is a Michigan based retailer that has adopted the supercenter retailing format that provides a full
grocery store and department store in the same location. Meijer has 170 locations throughout the Midwest
United States.

58

facilitated by limiting the geographic focus of the region and is a typical method used in
food desert research. The retail environment is, essentially, held constant so that other
household variables (income, education, car ownership, disability status, etc.) can be
examined. While some researchers have examined larger geographic regions (e. g.
Blanchard and Lyson 2002), this is only possible by using aggregated data, such as from
counties or census blocks. Use of aggregated data can only indicate the types of areas
where food retailers are located—it does not assess the actual behavior or perceptions of
consumers which may or may not involve traveling to the closest store. This study goes a
step farther by collecting data at the household level and relating that to the structure of
the food system in the neighborhood. In this way, we gain insight into the actual

relationship between retailer choice and the retail environment.

General Methodology

. Both qualitative and quantitative methods were used in this study. I felt that the
use of multiple methods was the most appropriate means to answer my research
questions. First, the complexity of perceptions and decisions surrounding food retailers is
better understood by examining the issue from a variety of angles. As Wansink and
Sobal (2007) point out there are literally hundreds of decisions that are related to the food
we obtain and eat On a daily basis. These decisions relate to decisions such as when, how
much, where, and with whom we will consume food. Similarly, it is reasonable to
suggest that the range of decisions regarding when, how frequently, where, and with
whom one goes shopping is equally as complicated. Qualitative methods provide a depth

of understanding of the varied strategies and perceptions that people have of the food

59

system. At the same time, quantitative methods are useful for assessing how widespread
various activities and perceptions are within the neighborhood. Newman et al. (2003)
recommend the use of mixed methods for the assessment of complex phenomena.

The use of mixed methods also allows for the triangulation of data and an
indicator of the internal validity of my results. By examining the issue of food
purchasing from multiple perspectives, overlapping conclusions are more likely to be
valid. For instance, if both survey and interview data indicate that food quality is a factor
inﬂuencing store selection, this indicates a greater probability that each respective
method is a valid measure for what is taking place in the neighborhood.

The qualitative data had limited generalizability due tothe relatively small sample
size (N=16) and the fact that sampling frame from which respondents were drawn did not
consist of the entire neighborhood (details of data collection are addressed later).
However, the interviews did provide depth of information about food purchasing
patterns, especially of disadvantaged residents, that would not have been possible in a
self-administered survey. This data provided an important basis for the development of
the survey instrument, which allowed for the testing of extent of certain behaviors and
perceptions in the neighborhood. Since the survey involved randomly sampling from the
entire neighborhood, I was able to generalize the results to the entire neighborhood,
something which was not possible had I drawn only from qualitative data.

The changing structure of food retailers has led to the growth of large food
retailers in suburban areas and the decline in small grocers. At the same time, there has
been a decline in small grocers and the growth in convenience stores in inner city areas.

This has placed different segments of urban populations in different proximities from

60

different types of retailers. This has the potential of making food purchasing particularly
problematic for disadvantaged inner city populations for whom traversing distance may
be difﬁcult due to limited mobility. One of my goals in this research is to understand the
role that distance from food retailers has, as an independent variable, on both the
perception and behavior of households. In order to obtain an objective measure of
distance, I incorporated spatial data for analysis using GIS. As the inﬂuence of space on
behaviors is, in some cases, unconscious (Sobal and Wansink 2007), it was necessary to
measure space objectively in addition to assessing how respondents felt space inﬂuenced
their decisions.

Details of how the data were collected are discussed in this section. Analyses are
discussed in subsequent chapters. This breakdown was necessary to avoid confusion
related to the wide variety of analytic techniques that were used in this study. Table 1
provides a breakdown of how different types of data were incorporated into the study.
Included in this table is also a breakdown of how different research questions are

addressed in subsequent chapters.

61

Table I: Data Collection Techniques and Relationship to Research Questions and
Organization of Dissertation

Research Questions

 

 

What types Do food

of secure and Do food

households Do food food secure and Are food

are more secure and insecure food secure and

likely to be food households insecure food

food insecure use households insecure

insecure or households different shop at households Does

food perceive criteria for different different distance

insecure food selecting types of distances influence
Data with retailers food food from food shopping
Source hunger? differently? retailers? retailers? retailers? behaviors?
Semi- CHAPTER CHAPTER
structured FIVE FIVE
interviews

‘ CHAPTER CHAPTER CHAPTER CHAPTER CHAPTER CHAPTER
Survey FOUR FIVE FIVE FIVE SIX SIX
GIS CHAPTER CHAPTER
Mapping SIX SIX
Qualitative Methods

To understand the shopping experience of neighborhood residents, a series of 16

semi-structured, qualitative interviews were conducted in 2004 and 2005. Additional

interviews were also conducted with members of a local neighborhood organization that

is interested in food and hunger problems. Interviewees were identiﬁed through the

neighborhood organization through their participating in two existing programs being

offered by the neighborhood organization, a Community Supported Agriculture (CSA)

drop-off point and a free bread program.4 CSA members were contacted directly by

phone. Additional neighborhood residents were recruited for interviews at the

neighborhood center on Fridays, the day in which a free bread program is held. On this

day, a clipboard announcing the study was put at the front desk along with other materials

 

‘ Interviews were conducted as part of a larger study on the potential of the Community Supported
Agriculture farm to increase fresh produce availability to low-income neighborhood residents. ‘

62

giving information about events in the community. Interested parties signed up and were
later contacted by phone to schedule interviews to be held in the neighborhood center.

A total of 16 interviews were conducted with neighborhood residents and an
additional two interviews were conducted with key informants who were involved with
the neighborhood center. Each interview lasted between 45 and 90 minutes and
examined where, how, and why different food retailers were selected. The demographic
characteristics of the respondents included a variety Of age groups, from mid 203 to late
603, different racial categories; Hispanic, African American and Caucasian. The majority
of interviewees were female. They included single people, single parents, married
couples with children and married couples without children, as well as families living
with extended family members. Almost a third of the interviewees were either disabled
or suffered from a chronic illness. Six were unemployed.

Minimal demographic information was collected as part of the interviews.
However, the overrepresentation of women (81%), families with dependent children
(81%), and unemployed individuals (38%) suggests that this was not a sample
representative of the neighborhood. At least two respondents noted that they had
physical disabilities and four either had no working car or were unable to drive
themselves to a store for other reasons. Furthermore, since interviewees were all selected
from programs associated with the neighborhood center, this led to a bias towards
respondents who had an interest in food programs. The type of program from which
interviewees were selected helped ensure that a diverse group was interviewed. I selected
one set of respondents from people accessing a free bread program while I selected

another set of respondents from those participating in a Community Supported

63

Agriculture Farm. The former group consisted of resource limited individuals while the
latter group consisted of individuals with greater resources. Since the interviews were, in
part, conducted as background data for survey construction and to understand the diverse
behaviors and perceptions related to food purchasing, with a particular focus on

disadvantaged consumers, I felt the breadth of backgrounds met the needs of the study.

Quantitative Methods
Survey mode

In order to understand the relationship between food security and shopping
behaviors, a survey of the Lansing eastside neighborhood was conducted using a self-
administered mail survey. Previous research in one area of the neighborhood had
indicated potentially high levels of food insecurity which has been related, in part, to
limited access to grocery stores (Thomas 2004). While the earlier study also indicated
that food insecure individuals were more likely to shop at discount or “big box” food
retailers, the role of income in determining the purchasing patterns of households was not
identiﬁed. Furthermore, the limited geographic area, lack of random sampling, and lack
of income data collected in the 2003 study limited generalizability and analysis. By
expanding the geographic scope, as well as speciﬁcally addressing issues related to
income and purchasing patterns, the data collected by this study will be generalized to the
entire area served by the local neighborhood center and may be useful to studies
examining similar neighborhoods.

Use of a survey to examine food security and purchasing patterns is consistent

with other research into food security and food deserts. The USDA uses telephone

surveys in its national assessment of food security (Nord et al. 2006), but has also
provided methodologies for conducting food security assessments through other survey
modes (Cohen 2002). The Community Food Security Coalition also discusses the
potential of using various survey modes to determine food security levels (Pothukuchi,
Hugh, Burton, and Fisher 2002). Wrigley et al. (2002) successfully used a combination
of a self-administered survey and face-to-face interviews in order to assess food
consumption patterns in an urban area in Britain.

To increase generalizability, it is important that this study be comparable with the
national USDA telephone data and there exists the potential of a mode effect due to using
a self-administered survey instead of a telephone survey (Dillman 2000). However, it is
likely that these effects will be minimal and may, in fact, reduce the error in the self-
administered survey in comparison to the USDA survey. First, there is ample evidence
that self—administered surveys reduce responses that are based simply on social
desirability (Aquilino 1994). While food security questions do not involve information
of a very personal manner, some respondents may feel social pressure to hide or
exaggerate the extent of their food deprivation. Therefore, their answers to a self-
administered survey may be more accurate. Dillman (2000) also notes that acquiescence
to a telephone interviewer and primacy/recency effects may also inﬂuence responses,
which is less likely in a self-administered survey. Finally, there is evidence that
adaptations of the USDA metrics are often just as accurate as the conventional USDA
measurements (Keenan, Olson, Hersey, and Parmer 2001).

The survey itself consisted of ﬁve mailings in a manner similar to that

recommended by Dillman’s (2000) Tailored Design Method. Pre-survey letters have

65

been shown to increase response rates of mailed surveys (Dillman 2000; Linsky 1975;
Yammarino, Skinner, and Childers 1991). First a pre-survey letter was sent informing
respondents that they had been selected to participate in the survey. Next, approximately
a week after the pre-survey letter, the survey itself was mailed. This mailing led to
approximately 150 responses. The third mailing consisted of a thank-you/reminder
postcard, which was sent out approximately one week after the survey was ﬁrst mailed.
This postcard thanked people who had already responded to the survey, as well as
reminded those who had yet to return surveys. Another 50 responses were received after
the postcards were sent. After two more weeks, a replacement survey was sent to non-
respondents. The replacement survey was very effective in increasing response rates and
led to approximately 100 additional responses. A ﬁnal mailing of thank-you/reminder
postcards was also provided two weeks after the replacement survey was sent; however
only a few surveys were received after this postcard was sent. One respondent who had
difﬁculty reading the instrument was surveyed by phone.

The survey took an estimated 15 minutes to complete. It was anticipated that
keeping the survey brief would increase response rates. A self-addressed, stamped
envelope was included with both the survey and replacement survey. In both cases, ﬁrst-
class stamps were used on the return envelopes as a means of increasing response rates
(Armstrong and Luske 1987). If the name of the respondent was known the envelope
was personalized; however, when the name of the respondent was unknown, the survey
was addressed to “Current Resident.” The level of analysis of the survey was the
household, so instructions on the survey asked that the individual primarily in charge of

purchasing food was to complete the survey.

66

Sampling

The sampling frame for households was constructed based on a list of properties
purchased from Accudata Inc. for the zip code of 48912. Accudata Inc. is a marketing
company that provides residential mailing lists, based on data obtained from the United
States Postal Service. While on site enumeration is commonly regarded as one of the
most comprehensive ways to obtain a complete sampling frame, the use of purchased,
residential mailing lists was chosen to save time and resources. Assessments of the
accuracy of residential mailing lists compared to on-site enumeration has found sampling
frames from the two sources to largely coincide (Iannoaccacchione, Jennifer, and Redden
2003). The list from Accudata Inc. contained 7,939 addresses. According to the 2000
US Census there were 8,740 households in the neighborhood. While the Accudata Inc.
list left approximately 800 addresses unaccounted for, this was still felt to be a
sufﬁciently large sampling frame for the purposes of this study. From the sampling
frame a random sample was drawn of 1,100 addresses. Each questionnaire had an
identiﬁcation code placed on it that correlated with the sample in order to track which
surveys were returned, and to ensure that replacement questionnaires were sent to

appropriate addresses.

Description of Sample
A total of 302 surveys were returned for a response rate (of 27%. Approximately
8% of the sample was returned due to vacancies, insufﬁcient or incorrect addresses, or

refusal to participate. Of these, the vast majority were returned because the residence

67

was vacant with only a few returned because of an explicit refusal to participate. It is
possible that the high percent of vacancies was an effect of conducting the survey in the
summer. As noted below, many of those in the sample appeared to be college students,
and it is likely that many of these residents were not living in rental housing during the
summer.

In order to assess the spatial representativeness of the sampling frame, it was
broken into census blocks. The randomly drawn sample included an average of 12% of
the households from each census block with a standard deviation of 3%. I felt this meant
that the sampling frame offered a reasonably even amount of spatial representation once
normalized for areas of higher population density. The spatiality of response rates was
also examined using census blocks as the unit of comparison. The average response rate
of all the census blocks was 25% with a standard deviation of 9.5%. Response rates
ranged from 14% to 52%. Pearson’s r was used to check exarrrine whether or not the
nature of the surrounding area could have inﬂuenced response rates by testing a
correlation between median income and response rate for census blocks. N o statistically
signiﬁcant correlation was found. Furthermore, since only one of the 16 census blocks
exhibited a response rate of >40%, I believe this was a statistical outlier and the spatial
distribution of response rates was relatively consistent with population distributions.

Figure 4 demonstrates that the spatial distribution of response rates.

68

 

 

 

 

 

 

Response Rate
1 -10
11 -20
21 -30
31 -40

40 or m ore

 

 

IIIDD

 

Figure 4: Response rates as a percent of the sampling frame in each census block in
study site

The socioeconomic diversity of the region was reﬂected in the sample (Figure 5).
Nearly 29% of the households surveyed had total incomes less than $20,000 per year
while 28% indicated household incomes of greater than $80,000 per year. The income
distribution of the sample matched the income levels reported by the US Census
reasonably closely. Only in the case of incomes ranging from $40,001 to $80,000 was
there some discrepancy. The sample included an under-representation of households

with incomes ranging from $40,001 to $60,000. On the other hand, households with

69

incomes of $60,001 to $80,000 were over-represented. This may be partially due to the
fact that a true comparison between US Census data and sample data is not possible since
Census uses $75,000 as the breaking point between categories while my survey used
$80,000. Still, it appears likely that these income groups were not accurately represented

in the sample.

 

30
25
20
15 Emil—e7
P481312 I
23031:.
10 *
5 .

 

 

 

 

 

   

 

$0 to $10,001 to $20,001 to $40,001 to $60,001 to $80,001 or
$10,000 $20,000 $40,000 $60,000 $80,000‘ more“

*48912 Zip code data actually represents a range of $60,000 to $75,000.

**48912 Zip code data actually represents a range of $75, 000 or more.

Figure 5: Household income distribution of the study sample and 48912 zip code
Despite the fact that the income distribution of the sample closely reﬂected the

population, this was not the case with education. The sample had a clear over

representation of households with higher education with higher response rates for people

with bachelors, graduate or professional degrees (Figure 6). In fact, over 28% of the

sample indicated that they had a graduate or professional degree, which was much higher

than the 12.6% that US Census reported for the zip code. The sample appeared to be

70

more representative of those with some college or bachelors degrees, although there was
still a higher response from those with bachelor’s degrees (25.4%) when compared to the
Census data (19.5%). At the same time, there was an under-representation of households
with high school degrees or less. Only 3.1% of the sample had less than a high school
degree and 12.2% had only a high school degree compared to 16.1% and 21.8%
respectively for the zip code. Notably, US census collects education data at the
individual level while this survey asked for the highest education level of anyone in the
household. This may account for some of the discrepancy since many households consist
of a primary income provider with a higher level of education than other people in the
household. Still, research indicates that surveys may be vulnerable to bias those with
higher education (Figure 6) and I believe that this was also a factor in this survey.

35
30
25

20

l348912 I

I'Sample 7
l .
I Zip Code J:

101

 

 

 

 

 

 

No high school High school Some college Bachelors Graduate or
professional

Figure 6: Highest level of education of someone in the household for the study sample
and population education for 48912 zip code

71

Over 30% of the households in the sample were renters. This was surprising
since, according to the 2000 US Census, 47% of the households in the area were rented.
However, once the percent of surveys returned due to vacancies are taken into account
these numbers come closer to matching. This likely supports the idea that the
neighborhood is very much inﬂuenced by college students who might not be present
during the summer when the survey was distributed.

Household size appeared to be relatively small with the average household
containing only 2.2 people. Just over 20% of the households in the survey contained
children with the average household with children containing an average of 2 adults and 2
children. The US Census reports 24% of the households in the neighborhood contain
children so the sample appeared to accurately represent families with children. Four
percent of the sample consisted of households with a single adult and children. It is
likely that the sample under represents the number of single-parent households, which the
US Census reports as 12.9% of the households in the zip code.

The median age of the person who most frequently obtained food for the
household was 51. While this suggests an older population, there was a notable spike in
the percentage of respondents between the ages of 21 and 30. Over 18% of the sample
fell within this age range. Other than this, the age distribution followed a roughly
normal distribution around the average age of 49. Elderly populations, which are often
more vulnerable to food insecurity, are also represented (just over 9% of the sample
reported being 71 or older).

Both automobile ownership and whether anyone in the household had a physical

disability were assessed as these characteristics tend to be associated with food

72

insecurity. Just over 19% of the sample households had one or more individuals with a
physical disability while 8.5% of the sample indicated that they did not own a working
automobile. According to the US Census, 19.9% of the households in the neighborhood
had a person with a disability while 11.9% of the neighborhood did not own an
automobile. Consequently, the sample was felt to be largely representative on these
factors.

While the response rate of the survey was slightly lower than the anticipated value
of 35%, the response of 302 suggests that the sample is representative of the population at
a conﬁdence interval of +/-5.5% at the 95% conﬁdence level. Furthermore, comparison
of the sample against US Census data in terms of income, household size, children, age,
disability status, and automobile ownership suggest that the characteristics of the sample
are very close to that of the population. In terms of education level and home ownership,
the sample appeared to be better educated and less likely to rent than the general
population. In the case of the former, this may be due to the effect of education on the
likelihood of respondents to appreciate, and consequently respond to, the survey. The
survey came from Michigan State University so it may be that students of this university
were more likely to respond. The relatively low instance of renters who were part of the
sample is more difﬁcult to explain; however, it may be an artifact caused by conducting
the survey in the summer. While these concerns must be considered when generalizing
the results of this study, the sample was felt to be reasonably representative of the

population.

73

Definition of Terms

The following deﬁnitions are used throughout the dissertation:

Food secure—The United States Department of Agriculture (USDA) is the most

inﬂuential agency involved in assessing food security in the United States.
According to the USDA, food secure households are those that “have consistent,
dependable access to enough food for active, healthy living (Nord, Andrews,
Carlson 2006: iv). Use of the term food secure in this study is based explicitly on
USDA deﬁnitions. In order to assess national food security levels, the USDA
conducts an annual survey using a twelve question assessment tool (Appendix A).
In this study, a six question modiﬁed version of the USDA measure was used
(Appendix B: Questions 30 through 35). Blumberg, Bialostoksky, Hamilton, and
Briefel (1999) have demonstrated that this shortened version is a valid measure,
although it fails to detect some elements of food insecurity, speciﬁcally hunger
levels in children. Food security status is deﬁned by the number of food insecure
conditions and behaviors exhibited by the household. If a household has no or
only one or two food insecure conditions (e. g. only infrequently could not afford
balanced meals), the household is deﬁned as food secure.

Food insecure—Based on the USDA criteria, if a household reports three or more

 

food insecure conditions they are considered to be food insecure. Examples of
food insecure conditions include running out of food, being unable to afford
balanced meals, cutting portion sizes, or feeling hungry at least once in the past 12
months. There are two levels of food insecurity. The ﬁrst is low food security

and the second is very low food security. The threshold of criteria varies with

74

whether there are children in the household. Households with low food security
have multiple indicators of food access problems, but few or no indicators of
reduced food intake. Households with very low food security have multiple
indicators of food access problems and disrupted eating patterns due to inadequate
resources to obtain food (Nord et al. 2006). Prior to 2005, the USDA used the
phrase “food insecure” to refer to households that have “low food security” and,
the phrase “food insecure with hunger” to refer to households that have “very low
food security” (criteria remained unchanged). Given the prevalence of this
nomenclature in the literature, the pre-2005 language is used in this dissertation.
Also, in this dissertation, as well as most USDA reporting, references to food
insecure households refer to both categories combined.

Disadvantaged consumer—Income is a signiﬁcant factor inﬂuencing the ability of
a household to access food; however, scholars have recently begun to consider
other factors that may limit food access, such as a lack of automobile ownership
or a physical disability. The term disadvantaged consumer is meant to capture a
wider variety of factors that may limit access to food. This study will draw from
deﬁnitions commonly used in food desert literature. Whelan et al. (2002) state
that ‘disadvantaged consumers’ experience constrained food choice as a function
of low income and restricted mobility.” “Disadvantage” is a matter of both
income and restricted mobility. Restricted mobility may stem from a lack of
access to transportation, personal disabilities, or other factors.

Food Desert—The concept of a food desert evolved out of research into poverty

and food retailers that was conducted, largely in Britain, starting in the 19903.

75

This research was driven in part from concern about the nutritional causes of
health problems suffered by low-income populations and in part by the concern
with the decline of inner cities and the growth of large food retailers in the
outskirts of urban areas. These two concerns were brought together within the
concept of a food desert, which is an area where few, if any, food retailers are
located. In practice, the concept of a food desert has been operationalized in
different manners, including, most frequently, the use of 500 meter “walking
distance” buffers around food retailers. A deeper examination of the meaning of
the concept will be explored in Chapter Six. For purposes of discussion, a food
desert simply refers to an area that is either devoid of food retailers or where
selection of food retailers might be limited, for instance an area that consists only
of convenience stores.

Food Retailer CLassiﬁchtion—In order to avoid a simplistic assessment of the
structure of the food system in relation to food retailers, a typology of different
food retailers was used in this study. This typology is similar to the
classiﬁcations used in other studies (Kaufman et al. 2003). The initial basis for
the classiﬁcation was Standard Industrial Codes (SICs), commonly used by the
federal government and obtained from a list from InfoUSA, a business
information company. Convenience or party stores are stores, such as Quality
Dairy, that only sell a limited number of food items, focus on convenience, and
are relatively small. Liquor stores that also sell food were included in this
category. Specialty stores include stores like health food stores or ethnic grocers

thatcarry food with a particular orientation. Since there were relatively few

76

specialty stores in the area, this category actually cOnsisted of several SICs. While
standard SICs only note grocers as a single category, it was felt that the variety of
types of grocers should also be taken into account. Most importantly, it was felt
necessary to distinguish between food retailers that specialize only in food and
those that also sell a wide variety of non-food items. Supermarkets refer to the
former type of stores with a focus only on fOod. Kroger and Apple Market are
two major supermarkets in the area in the neighborhood. Supercenters refer to the
latter type of store, which carries the same selection as supermarkets and has a

wide selection of nonfood items. Meijer and Super Wal-Mart are supercenters.

77

CHAPTER FOUR

CAUSES OF FOOD INSECURITY

The Problem of Hunger and Food Insecurity

In Chapter Two I drew from Hayes (1994) to explain that, sociologically, agency
can be understood as the extent to which people are able to alter one social structure (in
this case, the commercial food system) to coincide with another structure (perceived food
requirements). This perspective is useful since it allows us to consider agency in context
without ignoring the varied social structures (e. g. ethnic background, gender, class, etc.)
which inﬂuence individual tastes. Agency, in the context of the commercial food system,
becomes an issue of the extent to which this social structure coincides with the demands
and interests of different people. In most of the US the food retail system is the primary
means through which people obtain food. Poignantly, in the US, a country with nearly
4,000 calories available per person (Putnam, Allshouse, and Kantor 2002), the fact that
one in ten people do not have reliable access to sufﬁcient food indicates a clear problem
(Nord et al. 2006). It also suggests that the conventional food retailing system does not
meet the needs of at least 10% of the population. This failure suggests that this is a social
structure within which a large subgroup has only limited agency.

In a very real sense those who have unreliable access to nutritious food can be
seen as representing segments of the population with relatively limited agency vis-é-vis
the structure of the food system. The theoretical framework presented in Chapter Two

becomes very useful in situating problems with food acquisition in a society where

78

consumers are increasingly reﬂexive and inﬂuential (Figure 7). At the top of Figure 7 are
food secure individuals who, by virtue of their ability to reliably obtain food, have agency
within the structure of the food system and at the bottom are food insecure individuals

who are largely constrained by this same system.

 

 

 

 

 

 

 

 

> i
0 I
z .
L” I
52:
market I individual
segmentation I d1versrty
I Food secure populations
-35999939"! _________ "993.31! MPI'Q')‘.
structural I massification
lelSIOfl w I
g I Food insecure populations
1— I
0 l
D I
(I I
I— 1
W I

Figure 7: Food security/insecurity as related to agency and structure as determinants of
consumptive practices

In this section I will discuss the traditional conceptualizations of hunger and food
insecurity. The two terms are often used interchangeably, although hunger is generally
understood as a more extreme form of food insecurity. My purpose is to trace some of
the roots of food insecurity and hunger as concepts in order to understand some of their
commonly accepted meanings. Furthermore, by describing some of the current research
in the area I will demonstrate limits to common conceptualizations. After this I will
examine the results of survey data examining the causes of food insecurity within my

study site. Finally, I will discuss how my ﬁndings relate to conventional

79

conceptualizations of food security and highlight the limitations of assessments of food
security that do not also take into account the structural context of the food system.

Anderson and Cook (1999) provide a history of the term food security, noting that
concerns with “food security” originated during the 19603 and 19703 in reference to
international development. The focus of this research was on food supplies and
availability and the unit of analysis was often national in scale. By the 19803, there was a
shift towards household and individual food access and a long-term perspective on food
availability. Concern with food security had broadened from simple access within a
country to encompassing individual and household characteristics which impact people’s
long term access to food. By the 19903 the concept of food security had shifted almost
entirely towards concern with reliable access to food, affordability of food, and quality of
food. The concept of food security was further solidiﬁed in 1995 when the USDA began
funding programs to alleviate food insecurity and began conducting national surveys of
food spending, food access and adequacy, and sources of food assistance.

Frongillo (1999) assessed the validity of the USDA 12 question food security
assessment measure. Previous food security measures that were questionnaire based have
been criticized for not being accurate representations of people’s real experiences.
Further concern existed with whether the USDA measure was consistent with other
theoretical understandings of food security. Frongillo assessed validity based on six
criteria and a review of research that had been conducted using USDA measures in
combination with qualitative and other research methods. Frongillo concluded that the
USDA measure was consistent with the experience of households, consistent with other

food security measures, precise, dependable, and accurate measures of food security and

80

hunger. This validation is important since the USDA, as well as this study, utilize
respondent perceptions of having adequate food to assess levels of food security. As
Frongillo found, this type of food security measure is associated with health, nutrition,
and children’s development that supports the USDA conceptualized link with sufﬁciency
for active, healthy lives.

The speciﬁc USDA deﬁnition of food security is “access by all people at all times
to enough food for an active, healthy life (Nord et al. 2006).” This conceptualization of
food security involves two dimensions. The ﬁrst dimension involves the consistency
over time with which households are able to obtain the food that they feel is necessary for
the nutrition of everyone in the household. The second dimension involves hunger, or a
measure of whether households have had to signiﬁcantly alter their eating practices due
to a lack of resources. Only in severe cases of food insecurity is hunger experienced.

As Nord et al. (2006: 4) notes two-thirds of all food insecure households in 2005
were able to avoid substantial reductions or disruptions in food intake by “relying on a
few basic foods and reducing variety in their diets.” Even when sufﬁcient calories are
consumed, food insecurity can be a signiﬁcant problem for some households. For
instance, Blisard, Stewart, and Jolliffe (2004) found a signiﬁcant difference in the
expenditures of high and low income households on fruits and vegetables. They found
that, on average, low-income households spent $3.59 per capita per week on fruits and
vegetables while higher income households spent $5.02 (2004). This suggests that while
hunger might be avoided by many food insecure households, it is done by compromising
the nutritional quality of their diets. In short, hunger is only one part of the problem of

food insecurity.

81

Olson (1999) looked speciﬁcally at food security in women of childbearing age
and school children and found that for the women there was a correlation between
household level food insecurity and a higher body mass index, showing that obesity, not
famine, appeared to be correlated with food insecurity, even when income and education
were held constant. This echoes other research demonstrating the relationship between
poverty and obesity, which is particularly a problem in industrialized countries. Olson
(1999) also found that children in food insecure households suffered from compromised
psychosocial functioning related to behaviors such as stealing, ﬁghting, and not getting
along well with others.

Hamelin et al. (1999) have also examined the consequences of food insecurity.
First, they note physical impairments such as hunger, depletion, and illness. Next they
describe psychological suffering such as stress and constraints arising from the need to go
against held norms and values. Finally, the authors describe sociofarrrilial perturbations
that include modiﬁcation of eating patterns and ritual, disruption of household dynamics,
and distorted means of food acquisition and management. The authors demonstrate that
the impacts of food insecurity are not simply related to physiological health, but also
psychological and social, leading to problems such as theft or poaching.

More focused research on single mothers with children in Canada highlights how
management of food insecurity can have detrimental health impacts. McIntyre et al.
(2003) found that single women often compromised their nutrition in order to ensure the
health of their children. Parents were found to reduce calorie intake as well as the

nutritional quality of the foods they consumed. This was a particular problem during

82

times when food availability was limited. In these cases, children’s food intake was
maintained at the expense of the parent.

The implications of this research are that food insecurity, even when it is not
associated with hunger, can have a signiﬁcant detrimental impact on people and
households. In some cases the consequences are very direct and related to nutritional
deﬁciencies in food insecure households, but in other cases the consequences are less
direct, such as the anxiety felt by people in households with irregular access to food. In
some cases, the consequences of food insecurity even spill over to communities and
schools through behavioral problems such as children struggling with the social norms of
a community (e.g., breaking laws and misbehaving in school).

Based on the 1995 Current Population Survey, Rose (1999) found that 17% of
households with incomes less than 50% of the poverty level were affected by some form
of hunger. They also found that food insufﬁciency tended to be lower for high school
graduates, homeowners, small households, white households, and high income
households. Importantly, the authors found that, while there was a relationship between
income and food security, it was not a one to one relationship. In short, while many of
the socioeconomic characteristics of food insecurity are related to income, this is not
sufﬁcient to explain food insecurity. In fact, half of the people in the 1995 USDA survey
who experienced hunger had incomes above the poverty line (Rose 1999).

Rose’s conclusions are meaningful in that they indicate that food security, while
strongly inﬂuenced by income, is also inﬂuenced by a variety of additional factors.
Disability status, family status, children, transportation ownership, and race have all been

implicated as factors that contribute to food security status. Speciﬁcally, households with

83

one or more people with physical disabilities, single-parents, large numbers of children,
no automobile, or some racial minorities tend to have higher instances of food insecurity.
The relationship is not necessarily direct and households may have the resources to
negotiate the difﬁculties in food purchasing related, for instance, to a lack of automobile
ownership. However, since these households need to dedicate either ﬁnancial or
temporal resources to riding the bus or waiting for a neighbor to go to the store, this leads
to a de facto increase in food purchasing costs by households who might least be able to
bear those transaction costs.

Unfortunately, this increased cost is generally seen as a household or individual
problem and not a social problem. Much as poverty itself is seen as an individual
problem, a lack of adequate access to food is also seen as an individual problem, sharing
the! same kind of attribution errors. Consequently, solutions to inadequate food have
focused on the individual or household through programs oriented towards making food
available for free or at a reduced cost or providing nutrition education about how to make
do with less. While such programs have utility in reducing some of the impacts of food
insecurity, they tend to place the burden of alleviating the problems associated with food
insecurity either on social services or on the impacted individuals and households. Class
stratiﬁcation, gender inequality, and other larger social structures are either not
implicated or given only passing consideration, despite evidence that disadvantaged
groups of all sorts suffer disproportionately from food insecurity.

This uncritical conceptualization of food security and hunger is equally evident in
strategies to alleviate hunger. Interestingly, this can be traced to efforts in the late 18003

to reform working class eating habits. Levenstein (1988) describes how in the 18703

84

reformers such as Boston businessman Edward Atkinson saw within the working class
diet a means to increase the living standards of the poor without increasing wages and
violating the laissez-faire canon of the time. Drawing from the “New Nutrition”
perspective, which began to see food as constituted of basic elemental nutrients, these
reformers attempted to educate poor populations to cook and eat more efﬁciently. Rather
than examining the social, economic and political factors that may have been limiting
access to healthy foods, they sought to teach the poor how to live healthy on minimum
wages. Much as it is commonly understood today, hunger and food security were seen as
an individual’s problems to be solved through education.

DeVault and Pitts (1984) describe how hunger has been conceived in relation to
the evolution of the Food Stamp Program. They note two potential ways in which hunger
can be deﬁned—an issue of income or an issue of food. The Food Stamp Program,
perhaps the most inﬂuential federal food provisioning program, deﬁnes the issue of
hunger as a problem of food rather than income. The implications of this are that social
inequity as a social problem is left unaddressed and the number of solutions to hunger are
limited simply to household or individual level food provisioning.

Poppendiek (2000) provides a more contemporary look at how hunger and food
security are conceptualized. In particular, she outlines how modern efforts to alleviate
hunger often serve more as “moral relief” for progressive and socially minded middle
class advocates. Because hunger is seen as a problem for individuals and households it
becomes one that can be alleviated through charity. This can have dire consequences, as
she argues that “[f]ood programs not only make the well fed feel better, they reassure us

that no one will starve, even if the nation ends welfare and cuts gaping holes in the food

85

stamp safety net (196).” The central point is that hunger needs to be understood as a state
of inequality—not as simply an issue of people who do not have enough to eat.

It is clear from this discussion of conceptualizations of food security and hunger
that these problems are still largely seen as a matter of household or individual
circumstances, and there is relatively little questioning of the causal factors involved. For
instance, while it is analytically useful to understand who is vulnerable to food insecurity
(Rose 1999), the unique impacts on particular populations such as women and school
children (Olson 1999), or the nutritional and social consequences of food insecurity more
broadly (Hamelin et al. 1999) are rarely addressed. This means that there is little light
shed on why particular people are food insecure in the ﬁrst place. In short, traditional
views of food security and hunger in the literature have conceptualized these issues as
conditions limited to individuals and households by limiting the focus to this level of
sOciety. I do not mean to dismiss the importance of this research, for indeed I think it is
critical to an expanded understanding of food security; however, we must also look for
factors that may be inﬂuencing populations differentially to truly understand, and not just
describe, problems related to food insecurity.

Traditional conceptualizations of hunger and food security can be seen as akin to
the “individualist” poverty explanations described by Cotter (2002:536-537), who notes
that accounts of poverty are often explained by either “individualist” accounts that focus
on deﬁciencies of populations or “structuralist” accounts that view poverty as a macro
social problem related to the degree of economic and social opportunity. While in many
cases these are seen as mutually exclusive, I would agree with Cotter that more integrated

approaches are much more insightful. Regardless, in the case of traditional theories of

86

food security, there is a clear emphasis on “individualist” accounts. This is evident in a
research emphasis that is often limited to examining food insecure households as well as
the strategies formulated to address problems with food security.

The problem, as authors such as Travers (1996) argue, is that the decisions that
people make in regards to their food purchasing and consumption patterns are deeply
embedded within a variety of social constructs outside of their control. These include
constructs of class and gender; the bureaucratic and political organization of food, health,
and welfare; and the commercial organization of food, health, and welfare. Travers
(1996:552) is very critical of research that fails to incorporate the perspective of the
disadvantaged and ﬁnds that eating is highly contingent on “the oppressive nature of
corporate control over food, of welfare policies grounded in assumptions that social
problems are caused by individual inadequacies, and of professional discourse that
'blames the victim' for their failures to follow nutrition recommendations.”
Understanding the extent to which consumers have agency in light of existing social
structures is a difﬁcult matter and one which warrants greater empirical investigation.

It is important that food insecurity be understood as a social problem related to the
structure of the food system since it then supports the idea that food insecurity stems
from constraints to agency of particular groups within the food system. Certainly social
services, such as food banks, help people get another meal and that itself is a worthy
cause; however, these social services tend to exist alongside the conventional food
system and do not enhance the agency of food insecure individuals within the context of
that structure. In the next section, I will examine the characteristics of food insecure

households within the study site. My results were largely consistent with other research

87

on the characteristics of food insecure households which, as I will explain at the end of
this chapter, indicates that agency, as measured by food security, is indicative of a

broader lack of agency in relation to other social structures.

Analysis and Results
The USDA classiﬁes food security into three levels—food secure, low food
security, and very low food security. As nOted previously, these categories are
synonymous with the categories of food secure, food insecure, and food insecure with
hunger. The latter nomenclature is used in this study. Nord et al. (2006) explain that
households classiﬁed as having low food security have reported multiple
indications of food access problems, but typically have reported few, if any,
indications of reduced food intake. Households classiﬁed as having very low food
security have reported multiple indications of reduced food intake and disrupted
eating patterns due to inadequate resources for food.
The same USDA criteria were applied to the sample developed for this study. Low and
very low food security levels were found among 21.5% of households in this study
(Figure 8), nearly dOuble the national average of 11% and well above state averages of
11.5% (Nord et al. 2006). While such levels are not necessarily surprising given the high
levels of poverty evident in some areas of the neighborhood, as well as results from
previous surveys, it nonetheless represents a disturbing trend. Perhaps even more
worrisome was that 9.6% of the sample indicated very low food security levels. This was

nearly three times national levels of 3.3% and more than twice state levels of4. 1% (Nord

et al. 2006).

88

Percent of total households

 

25

HIT“

 

 

 

I Food insecure households I
PHouseholds with very low food segurityj

 

 

  

 

(1‘25: .T:

 

Study site

 

 

 

Michigan United States
Region

Figure 8: Food security levels in the study site, Michigan, and United States

In order to examine the factors that contribute to food security, two sets of

analysis were conducted. First I used chi-square and t-tests to examine the individual

relationships between each variable and food security status. Second, I used logistic

regression to examining the inﬂuence of all the variables together on food insecurity.

Much of the literature on food security focuses on factors that contribute to food security

status and, while the USDA measure used in this study also looked at levels of hunger,

this measure is functionally a subcategory of the food insecure category. Consequently,

binary logistic regression focusing on the distinction between food security and insecurity

was chosen as the most appropriate statistical test. Similar to ordinary-least-squares

multiple regression, logistic regression evaluates particular effects while controlling for

possible confounding variables. Use of logistic regression therefore allows us to

89

determine which variables have a statistically signiﬁcant inﬂuence on food security status

as well as the strength of the inﬂuence of each variable.

Variables

The dependent variable consisted of a simple binary measure of food secure or
food insecure. Independent variables for the initial model were identiﬁed from a review
of the literature on factors contributing to food security. These included the interval level
variables of distance to the closest grocery store, distance to the closest food retailer, age
of primary food shopper, number of children in the household, and number of people in
the household. Distances were calculated as straight line distances based on GIS
analysis. The details of how this was done are discussed in Chapter Six.

Ordinal and nominal variables in the initial model included education level,
' physical disability status, home ownership, working automobile ownership, and total
household income. Since data were collected at the household level and the primary
shopper may not be the one with the highest education, education was assessed by asking
respondents the highest level of education of anyone in the household. The education
variable consisted of ﬁve categories, including less than high school degree, high school
degree, some college, bachelor’s degree, or graduate or professional degree. As will be
noted later, for the third logistic regression model the variable for education was recoded
into a binomial variable with bachelor’s degree or more and some college or less. Home
ownership was assessed by asking whether the household rented or owned the home.
Automobile ownership was determined by asking about the number of functioning

automobiles. This was then recoded into a binominal variable measuring whether any

90

functioning automobile was owned by the household. Income was collected as an ordinal
set of data at $10,000 increments below $20,000 and at $20,000 increments above
$20,000 (this distinction was done to allow for the identiﬁcation of poverty status for
small households). Income consisted of six total categories. All variables were recoded
into dummy variables for use in the model.

A new variable for poverty status was created based on 2007 Federal Poverty
Guidelines (Federal Register 2007.) This variable is based on income levels and
household size. Since income data were collected as an ordinal rather than interval
variable, poverty status could only be identiﬁed for all households that had incomes less
than $10,000 (the poverty level for households of one person is $10,210) and households
of four or more people that had incomes less than $20,000 (the poverty level for
household of four people is $20,650). While it is likely that households in poverty were
excluded from this, it was decided that a conservative estimate was more appropriate for

the regression model. Fifteen percent (N=42) of the sample was identiﬁed as being in

 

 

 

 

 

 

poverty.
Table 2: SummaQ (j variables used to predict food security status
Interval Variables Average
Distance to closest food retailer 0.29 miles
Distance to closest grocery store 0.76 miles
People in household 2.17
Number of children 0.4
_Age 48.92
Ordinal Variables Median
Income $20,001 to $40,000 per year
Education Bachelor’s Degree
Nominal Variables Percentege
Rent their home 30.20%
Do not own a car 8.50%
Single parents 4.10%
In poverty 15.10%
Have a disability 19.20%

 

 

 

91

Results

The ﬁrst set of tests examined the hypothesis that there was a difference between
food secure and food insecure households for each of the twelve variables using chi-
square or t-tests, as applicable to the appropriate level of measurement. T-test results
indicate a statistically signiﬁcant difference between the distance of food insecure and
food secure households to the closest grocery store. Similar results were found in terms
of distance to the closest food retailer. This is consistent with current literature on food
deserts and will be examined more closely in Chapter Six. For the remaining interval
variables, total people in household, number of children, and age, no statistically
signiﬁcant differences were found. This suggests no difference between food secure and
food insecure households in terms of household size or age of primary shopper.

Chi—square revealed statistically signiﬁcant relationships between income and
food security and education and food security (Table 3). I also conducted post-hoe
testing using Cramer’s V to determine the strength of association and found that both
income and education had a moderate association with food security status, with
education appearing to have a slightly stronger, but still moderate, association. All other
variables, which the literature suggested would inﬂuence food security status, were also
found to have a statistically signiﬁcant inﬂuence and are summarized in Table 3. I
measured the level of association of binomial Variables with phi. Based on phi, poverty
status appeared to have the strongest association with food security status among the

nominal variables examined, although it was only of moderate strength. Home renting

92

also had a moderate inﬂuence, while car ownership, single parent status, and disability
status demonstrated a relatively weak inﬂuence on food security status.

Table 3: T-test and chi-square results for each dependent variable in relation to food

 

 

 

 

 

 

 

security status
Interval Variables t
Distance to closest food retailer -3.02*"
Distance to closest grocery
store 2.191 *
People in household -.445
Number of children -1.807
_Age 2.996
Ordinal Variables Chi-square Cramer's V
Income 35.017" 0.345
Education 58.539" _ 0.459
Nominal Variables Chi-square phi
Rent their home 28.324“ 0.31
Do not own a car 14.278“ 0.22
Single parents 14.560" 0.224
In poverty 31.479" 0.337
Have a disability 12.087“ 0.204

 

 

*p<0.05 **p<0.01

The nature of the relationship between each variable and food security status was
largely consistent with other research in this area. Households that are rented, do not
have working automobiles, are headed by single parents, include people with physical
disabilities, or are in poverty are more likely to be food insecure. Similarly, the higher
the household income and the more education that someone in the household has, the less
likely the household is to be food insecure. Cross-tabulation of income and food security
status reveals that income was particularly inﬂuential for households with incomes less
than $60,000 (Table 4). For households with incomes above $60,000, food insecurity
was only an issue for less than 3% of the households. In the case of education,
households with more education appeared less likely to be food insecure (Table 5).
There appeared to be a large percentage (35.9%) of households with “some college”

education being food insecure. Of households with bachelor’s degrees, only 16% were

93

food insecure while food insecurity dropped to only 3% for households ’with graduate or

professional degrees.

Table 4: C ross-tabulation of household income and food security status

 

 

 

 

 

 

 

 

 

 

 

What was rour total household income for the last year?
$10,001 $20,001 $40,001 $60,001
$1 0,000 to to to to $80,001
or less $20,000 $40,000 $60,000 $80,000 or more Total
Food
secure 14 27 57 38 40 37 213
37.8% 62.8% 75.0% 88.4% 100.0% 94.9% 76.6%
Food
insecure 23 16 19 5 0 2 65
62.2% 37.2% 25.0% 1 1 .6% 0.0% 5.1% 23.4%

 

 

Table 5: Cross tabulation of household education and food security status

 

 

 

 

 

 

 

 

 

 

What is the highest education of anyone in your household?
Less
than Graduate or
high High Some Bachelors Professional
school school colleg @ree degee Total
Food
secure 5 23 59 63 80 230
55.6% 63.9% 64.1% 84.0% 96.4% 78.0%
Food
insecure 4 1 3 33 1 2 3 65
44.4% 36.1% 35.9% 16.0% 3.6% 22.0%

 

 

 

After testing each of the variables individually, I developed three logistic

 

regression models to examine the inﬂuence of each of the variables while holding the
others constant. This was important since it allowed me to hold other variables constant
and determine the inﬂuence of each variable. In each model I tested the null hypothesis
that the variables together did not have an inﬂuence on the likelihood of predicting that a
household would be food insecure.

The initial model incorporated all twelve of the variables discussed above. This
model correctly predicted 84% of the cases and was statistically signiﬁcant (p<.001).

Coefﬁcients, standard error, p values, and odds ratios are presented in Table 6. Results

94

indicate that age, some college, and renting a home had a statistically signiﬁcant
relationship with food insecurity (p<0.05). A one year increase in age is linked with a
3.6% decrease in the odds of a household being food insecure. Therefore, the older the
primary shopper in a household, the less likely the household is to be food insecure. In
terms of education, those households with some college education were over nine times
more likely to be food insecure than a household with an education level of the reference
group of a graduate or professional degree. For education, households with incomes of
$10,000 or less were over ﬁfty two times more likely to be food insecure than a
household with an income of over $80,000 per year. In this model which holds all other
variables constant, renting a home increased the probability of a household being food
insecure by 2.4 times that of home owners. None of the other variables signiﬁcantly
contributed to the model.

Table 6: Logistic regression model of food insecurity status based on all variables

 

 

 

Variables Coefficient S.E. P Value Odds Ratio
Distance to closest food retailer 0.500 0.465 0.283 1.648
Distance to closest grocery store —0.091 1.482 0.951 0.913
People in household 0.008 0.233 0.973 1.008
Number of children . 0.174 0.316 0.583 1.189
Age 0036 0.014 0.012 0.964'
Income-less than $10,000 3.964 1.738 0.023 52.685‘
Income-$10,001 to $20,000 0.763 1.015 0.452 2.144
Income-$20,001 to $40,000 0.920 0.937 0.326 2.510
Income-$40,001 to $60,000 -0.170 1.052 0.871 0.843
Income-$60,001 to $80,000 -18.699 6170.198 0.998 0.000
Education-less than high school 0.420 1.156 0.716 1.522
Education-high school 1.108 0.884 0.210 3.029
Education-some college 2.205 0.718 0.002 9.067“
Education-bachelor's degree 0.614 0.783 0.433 1.848
Rent their home 0.876 0.446 0.050 2.401 *
Do not own a car 0.368 0.621 0.553 1.445
Single parents 1.608 0.937 0.086 4.992
In poverty -2.029 1.401 0.148 0.131
Have a disability 0.954 0.540 0.077 2.596
Constant -2.698 1 .372 0.049 0.067

 

*p<0.05 **p<0.01

95

 

Based on these results, I reﬁned the model by removing variables that clearly had
no statistically signiﬁcant inﬂuence. Since disability status and single-parent status were
nearly signiﬁcant, these variables were retained. Distance to retailers, number of people
in the household, number of children, car ownership, and poverty status were dropped.
This second model is presented in Table 7. This model successfully predicted 82.5% of
the cases and was also signiﬁcant (p<0.001). The results of the second model were very
similar to the ﬁrst model, with the exception that disability status appeared as a
statistically signiﬁcant contributing factor. In this model, having someone in the
household with a physical disability actually increased the probability of food insecurity
status 3.6 times compared to those households without someone with a disability.

Table 7: Logistic regression model of food insecurity status based on a reduced set of
variables ‘

 

 

 

Variable Coefficient S.E. P Value Odds Ratio
Age 0036 0.013 0.006 0.965"
Income-less than $10,000 1.963 0.952 0.039 7.123*
Income-$10,001 to $20,000 0.819 0.931 0.379 2.268
Income—$20,001 to $40,000 0.937 0.877 0.285 2.552
Income-$40,001 to $60,000 0145 0.996 0.885 0.865
Income-$60,001 to $80,000 -18.765 6195.105 0.998 0.000
Education-less than high school 0.727 1.074 0.498 2.069
Education-high school 1.537 0.820 0.061 4.653
Education-some college 2.196 0.688 0.001 8.985"
Education-bachelor's degree 0.698 0.742 0.347 2.010
Rent their home 0.644 0.407 0.113 1.905
Single parents 1.398 0.761 0.066 4.046
Have a disability 1.286 0.480 0.007 3.618“
Constant -2.248 1 .054 0.033 0.106

 

*p<0.05 **p<0.01

 

I was concerned that the multiple categories involved in the income and education
variables may have been creating a problem for the ﬁrst two models. Therefore, I created
a third model, similar to the second model, except using poverty status in place of income

and turning education into a single dummy variable of households with no college degree

96

completed and households with a college degree completed. This model predicted a
similar percentage of the cases as‘the other two models at 82.9%. In this model, all but
home renting appeared as a statistically signiﬁcant predictor of food insecurity status
(Table 8). Again age had a minor, but signiﬁcant, negative inﬂuence with a one year
increase in age linked to a 3% decrease in the probability of being food insecure. Being
in poverty meant that one was nearly three times more likely to be food insecure than a
household that is not in poverty. Having a four year college degree or more meant a
decrease in the probability of being food insecure by nearly 80% while single parents
were nearly ﬁve times more likely to be food insecure than other households. As with
the second model, having at least one person in the household with a physical disability
meant that household was nearly four times more likely to be food insecure.

Table 8: Logistic regression model of food insecurity status based on poverty status and
simpliﬁed education variables

 

 

 

Variable Coefficient S.E. P Value Odds Ratio
Age 0030 0.012 0.014 0.970'

In poverty 1.069 0.430 0.013 2912*

Four year college degree -1.570 0.383 0.000 0.208“
Rent their home 0.686 0.368 0.062 1.986
Single parents 1.584 0.751 0.035 4877*
Have a disability 1.336 0.444 0.003 3.803"
Constant -0.058 0.655 0.929 0.944

 

 

*p<0.05 **p<0.01

The results of these three models were largely, although not entirely, consistent
with existing literature on food insecurity. Nord et al. (2006), in the yearly USDA report
on food security in the US, reported disproportionately high levels of food insecurity
among households with incomes below the poverty line, households with single parents
and children, and Black and Hispanic households. Race data were not collected as part of
this survey; however the results support the inﬂuence of poverty status on food security

status. Single parent status also appeared to predict food insecurity. The ﬁndings of this

97

study also support other research involving USDA food security data such as that by
Rose (1999), who found that while income served as one of several signiﬁcant factors
predicting food insecurity.

Previous research has suggested that single parent status increased the probability
of food insecurity (Olson, Rauschenbach, Frongillo, and Kendall 1997). Similarly, my
research indicated that single-parents were more likely to be food insecure than other
households. Day to day activities require a different set of negotiations when small
children are involved. Yet, food must be available to eat. The relationship between
single-parent status and hunger indicates that the structure of the food system is not one
that readily coincides with single parents. This trend was further supported by comments
by interviewees regarding the difﬁculty of shopping with young children. As one
respondent explained in response to planning meals before going shopping, “Since he
was born it’s kind of got off that. I still every now and then try to plan them ahead... if I
can get him to cooperate, I’ll do it.”

Qualitative research by Wolfe et aI. (2003) indicates that age can play a
' signiﬁcant role in food security. Their research indicates that, while money plays a
signiﬁcant role in the ability of households to reliably obtain food, other factors related to
age, such as the ability to obtain transportation or prepare food, are also very inﬂuential.
Data from this survey found a relationship between age and food security status (p<0.05).
However, age appeared to have a slight negative relationship with food insecurity. This
suggests that young households are slightly more likely to be food insecure, though this
might also be related to a lack of connections with the surrounding neighborhood.

Younger residents, and especially college students who are likely to be transitory, may be

98

less likely to make connections with neighbors that could be exchanged for food.
Notably, however, the inﬂuence of age was not as strong compared to that of other
variables, such as income and education.

The lack of inﬂuence of transportation on either general food insecurity or food
insecurity with hunger came as a surprise. Studies such as that by Caraher et al. (1998)
and Bromely and Thomas (1993) suggest that a lack of transportation is a critical factor
related to food insecurity. Interviews also indicated that transportation was a limiting
factor for many food insecure households. While a lack of transportation may limit the
frequency or location of shopping trips, it may not limit these practices to the extent that
food accessibility is severely limited.

Despite the lack of evidence in support of transportation as a contributing factor,
there was evidence that physical access to food may be a problem. The issue may be one
of people with physical disabilities negotiating the shopping process and not simply
getting transportation to the store. In the second and third models, households with
someone with a physical disability were nearly four times more likely to be food
insecure. This strongly supports the idea of a “disadvantaged” consumer who ﬁnds food
access problematic for reasons other than income . Furthermore, it suggests that
differential access to the food system is not simply an issue of whether one has sufﬁcient

ﬁnancial resources.

Conclusions

Travers (1996) makes the argument that nutritional inequities in society are best

understood by taking into consideration the social contexts within which purchasing and

99

consumption decisions are made. Consumption decisions are made within a variety of
contexts, each inﬂuenced by different, but overlapping, sets of social structures. For
example, at the household level gendered divisions of labor inﬂuence food selection and
preparation. The household level is itself embedded within the commercial organization
of the food system and the bureaucratic and political organization of food and health.
These systems are themselves embedded within structures of class and gender.

Traver’s conceptualization of consumption deviates from traditional
conceptualizations which focus on hunger as a problem only of a household or individual.
In traditional, deﬁnitively non-sociological understandings of hunger, the problem is
reduced to a lack of food availability at the household or individual level. Solutions,
consequently, are focused on simple food provision. In contrast, Travers sees hunger as
the result of inequality derived from social structures. This conceptualization ﬁts well
with a sociological understanding of food security which seeks to understand hunger and
food insecurity as caused by larger social structures.

The results of my assessment of causes for food security and insecurity indicate
income and education were very strong predictors of food security status. Perhaps it
comes as no surprise that the less money one has the more likely one is to be food
insecure. More broadly, however, these two factors, which are commonly associated
with socioeconomic class, indicate that social stratiﬁcation in the US is having a real, and
detrimental, impact, on the ability of large groups of people to meet their basic needs
(Nord et al. 2006). The problem is not simply that many households do not have enough
money to have reliable access to food. Other issues related to social stratiﬁcation appear

to be inﬂuencing food security also.

100

It comes as no surprise that education had a signiﬁcant inﬂuence on food security
status. Even controlling for all other variables including income, households with at least
one person who had a bachelor’s degree or more were 80% less likely to be food
insecure. This suggests that households with people of higher education may be ﬁnding
it easier to reliably obtain food, even if household incomes are low. I would note,
however, that this does not necessarily mean that people in well educated households
simply “know more” in terms of how to manage ﬁnances or purchase food on a small
budget. In fact, interviews conducted as part of the qualitative portion of this study, and
discussed in the next chapter, suggest that food insecure households are very strategic
about food purchasing. This suggests that people in well educated households may be
drawing from other resources related to their position in a higher socioeconomic class.
Thirty respondents indicated on an open-ended question that they received additional
food from family or friends with several listing “mom” or “my parents.” Twenty-seven
of those 30 had some college, a bachelor’s degree, or a graduate or professional degree.
While the small numbers and the fact this is drawn from an open-ended question limit
generalizations, this does suggest that socioeconomic class may have an inﬂuence beyond
its relationship to income.

Consistent with research into disadvantaged consumers, disability status proved to
be a strong predictor of food security status (19% of the sample households included at
least one person with a physical disability). This, combined with the lack of inﬂuence of
automobile ownership, indicates that physical access may be misrepresented in current
literature on food security. Speciﬁcally, physical access may be less an issue of mode of

transportation to the store and more a question of individual physical ability to negotiate

101

the shopping process. Again, this was consistent with some of the interviews, which will
be discussed in the next chapter. Notably, respondents with physical disabilities stated
that loading and unloading groceries either into their automobile or into their houses were
often more of a problem than was driving to the store.

This neighborhood is not entirely representative of either the rest of the state or
the nation in terms of food security levels. However, much can still be learned from the
neighborhood as an example of an economically diverse area where food insecurity is a
signiﬁcant problem for many residents. It comes as no surprise that elements generally
associated with socioeconomic class, such as income and education levels, were strong
predictors of food security levels. In addition, other factors proved to be even more
important in some models, speciﬁcally disability status and single-parent status.

Economic class stood out as the most inﬂuential factor inﬂuencing food security
although there was some indication that gender has a role, at least in relation to single-
parent households. Food security, however, is not only about class. Disability status, in
particular, points to the fact that our conceptualizations of the problems associated with
food insecurity should not be reduced to individual poverty. These results, while useful
in understanding neighborhood food insecurity patterns, leave us with many unanswered
questions as to why these different groups may be more vulnerable to food insecurity.
The original question regarding why different groups have different agency vis-a-vis the
conventional food system remains unanswered.

As Travers (1996) points out, the grocery store represents the most common
intersection between consumption at the household level and the commercial food

system, which inﬂuences food availability for the majority of the population. At this

102

point, let us look closer at the factors that may be contributing to the differences in ability
to purchase food exhibited by food secure and food insecure households. In the next
chapter I will examine the differences in perception and behavior of food retailers by
food secure and food insecure households. After all, a full and productive understanding
of the causes of food insecurity requires more than a demographic proﬁle of the
characteristics of food insecure households. An understanding of why decisions are
made that lead these households to food insecurity is equally important. Or, as it relates
to our larger question of agency and structure within the food system, to what extent does
food security stem from the decisions households make and to what extent does it stem
from an inability of households to make decisions? While seemingly tautological, this
question lets us look more closely at the ways in which the food system helps and hinders

certain behaviors.

103

CHAPTER FIVE
PERCEPTION AND BEHAVIOR OF FOOD SECURE AND INSECURE

HOUSEHOLDS

Agency, Taste, and Class

Measures of food security may provide us with some insight into the differing
levels of agency that exist between food secure and insecure households. It is important
that we avoid simplistic perspectives which assume that food insecure households lack
agency. In this section, I will examine the different ways in which food insecure
households mobilize various components of agency to cope with differing, and
sometimes changing, abilities to purchase food. Food insecure and secure households
clearly differ in the ability to obtain food from the conventional food system—the
question is how and why do they differ? My comparison will involve two dimensions 1)
perceptional differences in food sources and 2) behavioral differences in terms of food
purchasing strategies. Based on a combination of qualitative and quantitative data I will
make the point that food secure and food insecure individuals often share similar
perceptions of the food retail environment. However, due to a limited ability to negotiate
the retail environment, food insecure individuals are forced to develop complex
purchasing strategies that prioritize cost. In short, I will address the question: to what
extent do differing levels of agency relate to different perceptions of and behaviors

related to food purchasing?

104

First, it should be understood that the process of procuring food is complicated.
In fact, Wansink and Sobal (2007) estimate that people make well over 200 decisions
regarding the how much, when, what and who of food consumption on a daily basis.
Wansink and Sobal (2007: 108) note that these decisions are based in part on personal
norms or patterns, but that “the norms can also be altered on any given occasion by the
environmental cues around that person.” While, the vast majority of food decisions are
not consciously made, this does not detract from the importance of understanding the
nature and reasons for the decisions that people make.

Some scholars such as Greder and Brotherson (2002) and Kaufman et al. (1997)
have focused speciﬁcally on the food decisions of food insecure families and the ways in
which households negotiate resource scarcity. These authors identiﬁed ﬁve major coping
strategies of low income households: (1) relying on others; (2) adjusting resources; (3)
reducing food consumption; (4) making trade-offs; and (5) acquiring nutrition and.
shopping knowledge and skills. Not only does this demonstrate the complexity of
decisions regarding food purchasing, but it also demonstrates the inﬂuence of strategic
and conscious decisions among food insecure households. Speciﬁcally, food insecure
households are often able to devise strategies to reduce or eliminate some of the problems
related to the difﬁculty of procuring food.

In a similar vein, Kaufman et al. (1997) found that despite the fact that low-
income populations often faced higher prices, they were able to adapt by buying more
economical and lower quality foods such as hamburger instead of steak. Item selection,
consequently, becomes an important way in which cost can be'negotiated. Based on

household survey data they ﬁnd their highest income group spent $48 on ﬁsh and seafood

105

per person while the lowest income group only spent $26 (Kaufman et al. 1997: 9).
Similar trends were found for fresh fruits and vegetables, with high income groups
spending much more than lower income groups. Of course, the implications of this are
that while item selection may be a useful way of negotiating cost, there are likely some
nutritional consequences of eating less fresh fruits, vegetables, and seafood and more
hamburger.

In many cases food insecure individuals will go to great lengths to ensure that
they and others in their household have sufﬁcient food. Based on focus groups, Kempson
et al. (2003) found that many of these strategies were unknown, even to practitioners
working to educate resource limited populations. They cited often unacknowledged
strategies such as participating in research projects, selling blood, avoiding having guests
over for meals or snacks, going to church dinners, cooking with other people, getting
food from work, and even committing a crime to go to jail. While on one level these
strategies are indicative of the high level of desperation that many of these people feel, it
also shows that food insecure households have devised some creative and strategic ways
to negotiate food insecurity.

Many of the strategies noted by researchers relate to obtaining food outside of the
conventional food system. Obtaining food stamps, starting a garden, going to a food
bank, or even borrowing food from friends or family represent strategies that are external
to the conventional food system. This suggests that while resource limited horiseholds
are often able to obtain sufﬁcient calories, they are often doing so through sources
external to the mainstream vector through which people access the food system —- the

food retailer. While this research points to the importance of alternative food sources for

106

disadvantaged consumers, it does not address questions related to why these consumers
do not access all of their food from retailers. Under what circumstances and why do
disadvantaged consumers turn to alternative sources to obtain food? To what extent is
their behavior an issue of culture (i.e. disadvantaged consumers learn there are numerous
ways to obtain food)? To what extent is their behavior due to the costs associated with
accessing food from conventional sources (i.e. disadvantaged consumers would like to
purchase food from retailers, but lack the ﬁnancial resources)?

Disadvantaged consumers are resource limited in some fashion, but at the same
time they demonstrate a highly rational, and often complicated, set of strategies to obtain
sufﬁcient calories within the bounds of these limitations. Unfortunately, people working
to alleviate problems related to food insecurity do not always recognize that food
insecure households might be doing their best to obtain food and that the real problem is
related to the food system. Stokols (1992) makes the point that most activities to promote
healthy nutrition have been focused on strategies that relate to educating people to
modify their behaviors. Modiﬁcation of behaviors and the education of disadvantaged
consumers on strategies to more efﬁciently, and nutritiously, obtain and prepare food are
seen as solutions to the problems associated with food insecurity. Such efforts suggest a
prevalence of the perception that the problem among food insecure households is cultural
or attitudinal and that modiﬁcation of that culture will help them make “better” choices
that will alleviate the problems of food insecurity. Despite this, it is not clear whether
cultural preferences are really driving the choices of either food secure or food insecure

households.

107

While research speciﬁcally examining the cultural factors that contribute to food
retailer choice is hard to ﬁnd, cultural explanations of consumption decisions nright be
adapted from individual food preferences. For instance, drawing on Bourdieu’s work,
lower class households might lack the symbols to interpret the haute cuisine available at
some food retailers while higher class households might be drawn to them for reasons not
understood by lower class households. Bourdieu (1984) argues that this stems from the
fact that food tastes are socialized from birth leading to different tastes among the class
hierarchy. This is meaningful because it suggests that the item selection by food insecure
households might be a matter of socially constructed taste, in addition to econorrric
necessity. Caraher et al. (1998) note a tendency among low income households to select
items based on cost instead of quality. Potentially, this is indicative of differences in the
socially constructed tastes of households of different classes.

Of course, the issue of taste preferences is complicated by the issue of
availability. As Wright, Nancarrow, and Kwok (2001) describe, globalization has led to
the proliferation of a variety of products within most supermarkets in developed
countries. In many ways, preferences that have been limited by availability are now
being challenged by the widespread distribution of various products such as hummus,
sushi, and kiwi around the world. Even for food insecure households, supermarkets offer
a wide variety of items with relatively low prices. Wright et al. (2001) still argue that the
culture in which one is raised, of which social class is a part, strongly inﬂuences food
tastes. The issue of availability offers a further complication to devising ways to think

about food preferences.

108

This research indicates that cultural preferences have a role in determining food
choice, which suggests that cultural preferences have a role in determining food retailer
choice. In order to test this proposition it is important that I examine food retailer
perception as well as choice. The reason for this is that perception provides an indicator
as to whether the symbols that food insecure and food secure households associate with
individual food retailers are similar. Additionally, exanrining perception allows us to
compare food retailer “tastes” between the two groups.

Importantly, perception is not the same as behavior and one’s tastes say nothing
of what one cats (or in this case, where one shops). Therefore, the actual behavior of
households is also examined in order to see whether actual store selection is similar
between the two groups. The extent to which perception and behavior coincide, therefore
provides us with some indication of the extent of, and reasons for, differential agency
among these two groups.

As Hays (1994) notes, culture itself is part of a social structure. Agency is a
measure of the extent to which one social structure, such as those related to cultural
preferences, is able to alter another social structure, such as that of the food system, to
match speciﬁc needs. In short, the extent to which the structure of the food system
coincides with the preferences of a group indicates the amount of agency that group has
vis-a-vis the second system. It is possible, of course, that the food system is catering to
the preferences of both groups. However, the literature on this topic indicates that it is
much more likely that food insecure households are ﬁnding themselves lirrrited in their

ability to select the food retailer which would provide them with food they would prefer.

109

Qualitative Results: Tastes and Purchasing Strategies
Analysis

In order to understand the food purchasing strategies of these consumers, a series
of semi-structured, qualitative interviews were conducted with food secure and food
insecure residents of the neighborhoods A total of seven food secure neighborhood
residents were interviewed and nine interviews were conducted with food insecure
individuals. Data were analyzed for emergent themes and codes were developed.
Coding was conducted using NVivo Software. Displays were developed to reveal
patterns among food secure and food insecure respondents with respect to food
preferences, particularly among vegetables, meal patterns, shopping behavior, and coping
strategies. While the small sample size clearly precludes generalizations of strategies to
the entire neighborhood, the in-depth analysis of their food purchasing strategies provides
important insight into the perceptions of the food retail environment and the different

strategies that both types of households have developed to negotiate the food system.

Results

First, patterns of purchasing vegetables varied greatly between groups. While just
over half of the food secure respondents indicated that they did eat fresh vegetables on a
regular basis, consumption of fresh vegetables appeared to be somewhat limited in food
insecure households. In particular, respondents indicated that even if they were interested

in eating vegetables their male spouses and/or children tended to prefer meat and/or

 

5 As noted above, these interviews were conducted as part of a larger study on a Community Supported
Agriculture farm that was distributing produce in the neighborhood. While no formal assessment was
made of food security status, individuals who were interviewed who were seeking free bread through the
neighborhood center were determined to be food insecure. Individuals who were members of the CSA
farm were determined to be food secure. ‘

110

dislike vegetables. One respondent described how s/he was able to overcome the
resistance of his/her children to vegetables, “Put some hard boiled egg on top of it. Give
them spinach with a hard boiled egg on top and they will eat their spinach. As long as I
can ﬁgure out a way'to make it, you know, I’ve cooked it up. So when I’ve had a
problem with them eating carrots then I started candying them.” Unfortunately, the
respondent’s efforts to change the taste preference of her children was often thwarted by
the spouse, “They learned how to not like liver because their dad won’t eat it. If he had
eaten it like it was McDonald’s hamburgers...”

Problems with the taste preference of family members were echoed by other
respondents. One stated, “[I]t’s convenient for me to make it, and he likes it. So I just
make what he likes right now. He really rarely will eat fruit or vegetables right now.” In
many cases preferences toward vegetables varied greatly within the family. One
respondent noted “[m]y daughter’s kind of choosy, but other than that everyone else is
ﬁne” and another said that, “[t]he boys, they are so picky. But not the girls. They eat
everything.”

Respondents were required to negotiate these taste preferences within a set of
ﬁnancial constraints that further complicated food purchasing patterns. Speciﬁcally,
most respondents described relatively complex food acquisition strategies that involved
activities such as sale seeking, bulk buying and storage, food assistance, coupon use,
multiple store visits, buying seasonally, taking condiments and crackers from restaurants
and reviewing of ads, and buying food after paying bills.

Selection of items based on sales was extremely important in determining which

items were purchased—“buy. . .canned vegetables when they are on sale or frozen

111

vegetables, whatever is on sale.” Another respondent noted that “I just go in and
whatever’s on sale, that’s what I’m going to have. Like I ﬁnd chickens for 99 cents a
pound, that’s a real good deal, and these chickens are two-ﬁfty a pound, I went and got
big chickens. So then we eat chicken.” while one simply stated “it’s whatever they have
on sale.”

Sales served not only as a means to save money, but also as an opportunity to
stock up on items that could be used throughout the year. In two cases, respondents
purchased or acquired fresh vegetables that they would then preserve through canning or
freezing. One respondent described how rain checks could be used to further extend the
beneﬁts of sale items:

They were on special for like 75 cents a piece. Sol went out there every

day and I picked up four for three dollars. Well that was all of the ﬁrst

week. And then the next week I went in and picked up four more; the next

day four more. Anything they was out of I got a rain check on. So I have

twelve different cans, you know, items up there and I’ve got like 10 rain

checks yet and they’re good forever.

It appeared that the purchasing of meat was one item commonly bulk purchased. “I’ll go
spend $100 and then just buy meat. Lots of meat,” stated one respondent. Another
respondent explained that

I order a side of beef and get me a quarter of a hog, and that elirrrinates my

meat cost. And then if they got sausage or something minor, and I can

afford it, I add to that. Like if hamburger or something’s on sale for a

reasonable price, or them country style ribs, I add to what I got.

Even with all of these strategies, however, respondents still described problems in
meeting all of their household needs. One respondent explained “I only have X amount

of dollars to spend. I can’t go over. So you have to choose from this to that, from

deodorant to a gallon of milk. What are you going to buy?” Another respondent was

112

very conscious of the seasonality of costs and planned his/her consumption accordingly,
saying that

Because I try to stretch that meat to get to where the snow is gone. . .cause you

gotta deal with them high utility bills, you cannot buy food and think you gonna

stay warm,-at the same time. So the best plan is by the end of August, September,

October, get the freezer packed, as packed as you can.

One respondent described a complex process of budgeting monthly disability
checks. Each month upon receiving the check, this person would ﬁrst leave early in the
morning to stand in line at the Salvation Army to get a number to receive food assistance.
After receiving a number s/he would go to the bank for assistance writing checks to pay
bills due to an eye sight problem. In some cases, if the disability check came after a bill
was due, an additional trip had to be made to pay overdue bills. After this, the individual
would have to return to the Salvation Army to wait several hours in line, hope that there
would be food s/he could eat due to dietary limitations related to allergies and diabetes.
After ﬁnding out what food was available from the Salvation Army, the ﬁnal step would
be to drive to a discount grocery store and sometimes dollar stores to get food for the rest
of the month. Not only does each of these steps require a great deal of waiting in line,
but it can be physically demanding, in particular for someone with physical disabilities.

Difﬁculties related to physical disabilities were echoed by other respondents. One
described problems with managing a menu that would help her chronically ill daughter
while another described how the physical pain caused by unloading groceries required
that s/he often leave part or all the groceries in the automobile until assistance was

available. Disabilities of individuals and household members consequently served as a

limitation to the types of food items that could be reasonably selected.

113

Obstacles related to the physical acquisition of food were further exacerbated by
transportation, including an unreliable automobile, or no automobile at all. One
respondent did not have a driver’s license and relied on rides from family members. Two
respondents stated that they did not currently own functioning vehicles and therefore had
to either carry groceries on public transportation or rely on the schedules and assistance
of friends or family members. A lack of automobile ownership serves to limit both the
quantity and variety of foods that can be selected at any one time. After all, it is only
possible to carry so many gallons of milk and pounds of fresh vegetables onto a bus.
Furthermore, as one respondent mentioned, relying on rides to the store from friends
often required going wherever that friend wanted and reduced sale seeking activities.
While this respondent was aware of sales that would save the household money and had a
set of preferred retailers, without a working car the trips to the grocer was dependent on
where and when the neighbor shopped.

It would be a mistake to conclude, however, that quality was not something that
was of interest to food insecure households. Consider responses to questions of organic
A food—a term that is commonly perceived as higher quality food. Nearly all of the food
secure respondents indicated an interest in organic food and most had purchased it at least
occasionally in the recent past. Notably, food secure respondents were recruited from a
group who had already indicated interest in organic, fresh produce through participation
in a Community Supported Agriculture Farm. On the other hand, only two food insecure
respondents indicated any interest in purchasing organic vegetables and one had bought
organic. However, 90% of the food insecure respondents indicated concern with the

nutritional quality of the food that they purchased and two respondents, who were not

114

interested in organic food, explained that they had found other ways of obtaining food
with the same qualities. One food insecure respondent reported blanching vegetables to
remove the pesticides. This person suggested that it provided a more economical way of
obtaining fresh vegetables of similar quality as organic. This indicates that, at least in the
case of organic food as a measure of food quality, behaviors might have differed between
groups, but at the same time both groups had a similar Set of interests in the qualities
associated with organic.

These results point to some similarities and differences and food procurement
patterns of food secure and insecure households. One notable difference is related to how
issues of food quality were balanced against cost. Food secure individuals indicated a
high level of interest in high quality food, with fresh—sometimes organic—vegetables a
clear priority. In many cases, these individuals stated that they dedicated signiﬁcant
resources to obtain these foods, both in terms of paying extra money or driving to
specialty stores to obtain the desired products. Interestingly, many of these individuals
maintained a high level of interest in saving money or saving time while shOpping. One
individual even mentioned a tendency to always buy sale items, even though it was clear
that he had no ﬁnancial reason to do 30.

Perhaps it comes as no surprise that food insecure individuals placed a much
greater priority on food cost than on quality. It was interesting, however, to hear about
the relatively complicated, well-devised strategies that individuals had created to
minimize costs using available resources. These strategies were often time consuming,
reliant on non-perishable products, and closely tied to schedules of store sales, looking

after small children, work, food assistance distributions, and release of public assistance

115

monies which limited the ability to deviate from existing patterns. A closer examination
of the interviews will show how these patterns have emerged and are maintained.

These results also show that there is some similarity in perception of food
selection criteria between food secure and insecure individuals. It also indicates that, in
the context of limited resources, these perceptions are leading to a vastly different set of
behaviors, though cost and food quality were of interest to both groups. However, how
the two elements were balanced differed noticeably between the two groups. Given the
limitations of the sample size, generalizations from this data are limited. Therefore, in
order to gain a better perspective of the relationship between perception of food retailers
and shopping behaviors, and the prevalence of these behaviors, it is necessary to turn to

the neighborhood survey.

Quantitative Results: Preferences and Behaviors
Analysis

Qualitative analyses indicate that even food insecure individuals have agency to
negotiate within the food system and, to some extent, minimize the nutritional impact of
food insecurity. However, the transaction costs of this negotiation were often high in
terms of time, driving distances, making multiple trips, compromising food quality and
selection, and other ways. To expand these ﬁndings, I turn to the survey data to examine
whether the strategies and perceptions of the food system exhibited by interviewees were,
in fact, representative of the neighborhood. In particular, I sought to identify the
existence of differences between behaviors and perceptions of food secure and food

insecure households. Two sets of data were collected through the survey that pertained to

116

this research question. The ﬁrst set of data related to the reasons that various stores were
selected and the second set of data related to the frequency of shopping trips made to area
grocery stores. The former set of data was used to assess how different food retailers
were perceived by food secure and insecure individuals while the latter data set was used
to assess behavior patterns. A series of hypotheses were developed based on the results

of the qualitative analysis in addition to literature in the area of food security and hunger.

Results: Food Retailer Perception

The ﬁrst set of hypotheses all stated that the perception of eleven different food
retailers based on twelve different criteria would be different for food secure and food
insecure households. Chi-square analysis enabled comparison of these individuals based
on reasons stores were selected among those who have shopped at each store in the past
30 days. Eleven speciﬁc stores were examined, as well as convenience stores and health
stores. This represented all major food retailers within six miles of the center of the
neighborhood. The list of stores was compared against interviews to ensure that the
majority of potential food retailers were included in the list. While an additional open-
ended question was included to ensure comprehensive data collection, the list of thirteen
sources encompassed the vast majority of places where respondents shopped with the
only exceptions being farmers markets and dollar stores (only 5% showed at dollar
stores and 6% shopped at area farmers markets). Twelve criteria were used to examine
perceptions of food retailers, including sale item availability, everyday low prices,
availability of bulk items, availability of specialty items, food selection, selection of

nonfood items, food quality, proximity to home, proximity to work, proximity to other

117

stores, familiarity with store, information from media sources, and other reasons. Criteria
were generated from thematic coding of interviews.

Did food insecure and food secure households perceive food retailers as having
the same characteristics? Results indicated very few differences in how these individuals
perceived each retailer. Differences were only noted in ﬁve of the thirteen food retailers
examined and only for a small number of selection criteria. Stores with signiﬁcant
differences were Meijer, Sam’s Club, Shoprite, Wal-Mart, and convenience stores.
Interestingly, in one case the results contradicted the qualitative ﬁndings regarding the
priorities that food insecure individuals placed on cost. In the case of Meijer, food secure
individuals were more likely to shop there due to low prices. Notably, however, the
difference was slight though statistically signiﬁcant. There was some indication that food
secure individuals were more inﬂuenced by non-cost related factors and, with Meijer,
food secure individuals stated that they were more likely to shop there due to quality and
selection. Similarly, food secure individuals were more likely to shop at ShopRite due to
the availability of specialty items. Only in the case of convenience stores were food
insecure individuals more likely to shop due to the availability of sale items. At the same
time, familiarity was a signiﬁcant factor for food insecure individuals in the case of
convenience stores and Wal-Mart. Contrary to what was expected there was not a
statistically signiﬁcant difference in the emphasis that the two groups placed on distance
as a selection criteria.

Overwhelmingly, however, it appeared that food secure and food insecure
populations perceived food retailers as sharing similar characteristics, whether the stores

examined were supermarkets, grocery stores, discount grocers, membership based stores,

118

health stores, or convenience stores. Out of a total of 156 different store characteristics
examined, only eight statistically signiﬁcant differences were noted and none of these
were particularly large. This suggests that there were very few differences in how food
secure and food insecure households perceived food retailers. Admittedly, due to
relatively small numbers of people who shopped at some of the stores, it was difﬁcult to
draw deﬁnite conclusions. Nonetheless, the similarity in perceptions demonstrated by
both groups supports the conclusions drawn from the qualitative analysis—that both food
secure and food insecure groups share similar perceptions. In this case, it is clear that this
similarity of perception is not simply in terms of similar concerns with food cost and
quality, but also a shared perception of the characteristics of food retailers.

Given the similarities in perception between the two groups it is worth assessing
how respondents perceived retailers. To do this, store selection criteria were broken
down into four categories: cost related attributes, quality and selection related attributes,
distance attributes, and other attributes. Cost related factors included sales and everyday
low prices. Food quality and selection items included the availability of bulk items,
specialty items, food selection, and food quality. Distance characteristics related to
proximity to home, work, other stores, or bus stops. Other factors included familiarity
with store, information from media, and other factors. Each retailer was then ranked
based on the proportion of the sample that had shopped at each store in the past 30 days
and stated that cost, quality or selection, distance, or other factors inﬂuenced their

decision.

119

 

  

CSStonvenlence
ores

I

     

 

 

- I
W
_

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Health Food
Stores
Country
Markett I I I
Wal-Mart I I I I I
1 l I
l i
H°"°°ks W I I
i F I
Apple Market . _ - ._ -_- _1_ _._ _
I Other
Sho rite WW. El Distance
p I Food Quality/Selection
El Cost __ _
Sam's Club I
I

 

 

 

 

 

 

 

. . I I
W. I
- I 1 I

I 1
. .
' I

7///A
—

 

 

I

' i
Save-a-Iot . I I
i

I

I

 

      
     
 

   

 

 

 

 

 

 

 

 

 

1 Meijer
, I i . I l
Kroger WW I
_
I I
|
l

or s 10 15 20 as 30 35 4o 45 50
Percent of respondents who shop at each store

 

Figure 9. Store characteristics that inﬂuenced respondents’ decisions to shop there

It is clear from the chart that cost and food quality and selection were two of the

most inﬂuential factors related to store selection (Figure 9). Interestingly, the major

supermarkets, Kroger and Meijer, attracted consumers primarily on cost, but also based

on perceived food quality, selection, distance, and other factors. On the other hand, for

deep discounters such as Aldi and Save-a-Lot, and Sam’s Club (a membership based bulk

120

store), cost was much more important. For the health food stores and Horrock’s (a
farmers market/grocer), food quality and selection were the top attributes. Finally, for
convenience stores and Apple Market (a small grocer closest to most residents),
proximity was the predominantly deciding factor.

In order to better understand these differences in retailer perception, I also
examined the selection criteria that were applied to the store most frequently visited
(Figure 10). Chi-square analysis of these factors identiﬁed two selection criteria with
statistically signiﬁcant differences between the two groups. Surprisingly, food secure
households were actually more likely than food insecure households to select a primary
food retailer based on everyday low prices. Less surprisingly, food secure households
were also more likely to select a primary food retailer based on quality. While not
statistically signiﬁcant, there did appear to be a possible relationship between food
security status and food selection (p<.10) and familiarity with the store (p<.10) with these
being more important to food secure households. No signiﬁcant differences were found
for cost, proximity, or other criteria. These results indicate that even in the case of
primary store choice, the criteria that food secure and food insecure households use in
store selection is relatively similar. Cost related factors such as sales and low prices were
very important. Other factors such as proximity to home, familiarity with store, food

selection, and food quality also carried some signiﬁcance.

121

 

Other

Newspaper, TV, or radio

 

 

 

Proximity to bus stop

Proximity to other stores F—IJ I I
. 1

Proximity to home I I

Proximity to work FIJ - . -
I I D Food Insecure

Food quality‘ I Food secure

7

i
l
I

 

 

 

 

I
l
Familiarity
I
I

 

 

 

 

 

  
 
     
 

 

 

 

 

Nonfood selection

 

 

Food selection

 

 

Specialty Items

Bulk Items

 

 

 

 

Low Prices‘

 

 

Sales
0 10 20 30 40 50 60 70 80 90 100
Percent of respondents
* Indicates a statistically signiﬁcant difference between groups (p<0.05)
Figure 10. Reasons for shopping at most frequently shopped at store by food secure and
food insecure households
A ranking of the reasons that food secure and insecure households shop at their
primary location reveals some difference in the relative inﬂuence of different store
characteristics (Table 9). For instance, sales appeared-to be more important than every
day low prices for food insecure populations with these characteristics being ranked
number two and number six respectively. On the other hand, for food secure households,
every day low prices was the most inﬂuential factor while sales appeared as the fourth
most inﬂuential factor. The differences, however, were not great. Notably, food quality

was the sixth and ﬁfth most inﬂuential characteristic for food insecure and food secure

households respectively. Food selection, nonfood item selection, and specialty items also

122

appeared with similar rankings for both groups. In contrast to comments made during I
interviews regarding the importance of store ﬂiers, food secure households appeared
more likely to use media as a reason to shop at their primary retailer. Less surprisingly,
proximity to a bus stop was more likely to be a factor for food insecure households;
however, as previously noted, it was not particularly important for either group.

Table 9: Ranking of the relative inﬂuence of store characteristics on the decision to shop
at the primary food retailer

 

 

Ranking
Food secure Food insecure

1 Low prices Proximity to home

2 Proximity to home Sales

3 Familiarity Familiarity

4 Sales Food selection

5 Food selection Food quality

6 Food quality Low prices

7 Nonfood selection Nonfood selection

8 Specialty Items Specialty Items

9 Newspaper, TV, or radio Proximity to work
10 Proximity to work Proximity to other stores
11 8qu Items Newspaper, TV, or radio
12 Proximity to other stores Other
13 Other Proximity to bus stop
14 Proximity to bus stop Bulk Items

Where does that leave us? We now have evidence that food secure and food
insecure individuals have similar concerns with food cost and quality. We also know that
they have a similar set of perceptions of the food retail environment. Already, the
differing set of food purchasing strategies identiﬁed by the interviews indicates that food
insecure individuals are only marginally, and through complex strategies, able to meet
their perceived consumptive needs, relying on more than just ﬁnances to navigate the
food system. Both groups use complex patterns of shopping, but for one group this

seems to be a necessity. Does this stand out in the quantitative data?

123

Results: Food Purchasing Patterns

While the qualitative results indicated limited but important differences between
the food purchasing patterns of food secure and food insecure households, the differences
were less pronounced in the survey results. Part of this might be due to the relative
insensitivity of this research method'to complex food purchasing as indicated in the
interviews. Also, since interviews were conducted with food insecure individuals who
were already obtaining food from a free bread program, it is likely that this group had

very low food security.

 

Country Markett ,

 

Save-a-lot"

I
I
I
Aldi l

 

L&L

 

Shoprite* ,- _, _ ___.
' I D Food Insecure
. ! ' Food Secure

I“I 7 **** e
I i

I
1

Health Food Stores

Horrocks

 

 

 

 

Kroger

 

 

 

 

. L ' ‘ I
O 10 20 30 40 50 60 70 80 90 100
Percent of respondents

 

* Indicates a statistically signiﬁcant difference between groups (p<0.05)
Figure I 1: Stores where food was purchased in the last 30 days by food secure and food
insecure households

124

In terms of store selection, the vast majority of people from both groups obtained
food from Meijer and Kroger (Figure 11), large supermarkets with closest proximity to
the neighborhood, and were reported as the stores most frequently visited in the past 30
days. Notably, convenience stores, of which there are many in the area appeared to be a
signiﬁcant source of food for both groups with over 50% of each purchasing food from
one of these sources in the past 30 days. Apple Market, a small grocer that is located in
the middle of the neighborhood, was the third most visited food retailer by food secure

respondents.

 

._.—.ﬂ

L&L I
Horrocks
Save-a-Iot E
WaI-Mart
I-

I0 Food insecure

I

I

I

Shop Rite l
I ;
l I In Food secure J
‘ A-“ Jpn“
I
I
I
I

I
I
I
I
Convenience Stores 1

Sam's Club

 

I

Health Food Stores

Apple Market

Kroger H.
Meijer H

7* f

 

 

 

 

 

Percent of respondents

Figure 12: Store shopped at most frequently in the past 30 days by food secure and
insecure households

An examination of the stores that respondents identiﬁed as their most frequently
used food source paints a similar picture of purchasing patterns (Figure 12). Again,

Meijer and Kroger appear as the staple food source with other food retailers declining

125

rapidly in importance. Notably, food secure individuals appear to rely more heavily on
Meijer than food insecure individuals with 49% of them reporting shopping at Meijer
most frequently and only 33% of food insecure individuals reporting the same behavior.
The remaining food insecure individuals appeared to be relying more heavily on Kroger
and Wal-Mart to a minor extent, and convenience stores to a much greater extent. Only
3% of food secure individuals reported that convenience stores were the most frequently
used opposed to over 11% of food insecure individuals.

Respondents were also asked to identify one other store most frequented in the
past 30 days (Figure 13). Differences in store choice become even more apparent in
these results. In addition to continued reliance on convenience stores, food insecure
individuals are also relying on “deep discounters” such as Wal-Mart, Aldi, and Save-a-
lot. While the difference is not as signiﬁcant as expected, these trends seem to indicate
that similarities in group behavior in terms of general shopping behavior (i.e. stores
where food was purchased at any point in the past 30 days), may mask the extent to
which each group relies on the stores in question. Examination of primary and secondary
retailer choices indicates that food insecure individuals are relying more on convenience
stores and deep discounters to obtain food, even though these sources do not replace

reliance on large supermarkets.

126

 

 

L&L 5 I I I
Horrocks . I I
I I

I

 

 

 

 

 

I
Save-a-lot I I
. . I I I
' I
Wal-Mart F : I I I I
ShOp Rite .- ' I I 015033 insecure
I 3 Food secure
I
I

I
I
Convenience Stores I
I

_ Sam's Club

 

Health Food Stores I

. I
Apple Market I ; I

 

 

 

 

 

 

Kroger HI I
. I I
Meijer
0 1 0 20 30 40 50 60

Percent of respondents

Figure 13: Store shopped at second most frequently in the past 30 days by food secure
and insecure households

The tendency of food insecure individuals to rely on deep discounters is further
supported by a chi-square analysis of the stores shopped at during the last 30 days.
Statistically signiﬁcant relationships were evident between food security levels and
shopping at Save-a-lot, Shoprite, Apple Market, and Wal—Mart. Food secure individuals
were more likely to shop at Shoprite and Apple Market while food insecure individuals
were more likely to shop at Save-a-lot and Wal-Mart. Shoprite is a local grocer that
tends to carry high cost specialty items, such as wine and cheese, so it is possible that this
contributes to food secure shopping. However, the causes for the difference in Apple
Market shopping is less clear. Both Save-a—lot and Wal-Mart represent retailers that

emphasize low prices.

127

The high percentage of the sample that shopped at Kroger and Meijer can be
explained, in part, by the number of these retailers that are within a ten mile radius of the
study site. Respondents reported shopping at ﬁve different Kroger stores and four
different Meijer stores. Therefore, the most frequently shopped at store was
disaggregated into available locations throughout the city (other stores with multiple
locations reported included Wal-Mart, Sam’s Club, health stores, and convenience
stores). Distinguishing between stores of the same chain was important since
interviewees discussed times preferences between stores of the same chain noting things
like differences in item selection and food quality. At the same time, one might imagine
that consumers would simply shop at the store that was closest to their home or work.
Re-analysis of the relationship between food security status and store selection with
stores at different locations revealed no statistically signiﬁcant relationship, showing that
there did not appear to be a relationship between food security status and the speciﬁc
location chosen to shop. However, in the case of Meijer the relationship was nearly
signiﬁcant (p<.10). Food insecure individuals appeared to have a greater propensity to
shop at Meijer locations that were not the closest to their homes.

Given that many interview respondents noted the importance of going to multiple
stores speciﬁcally for sale items or because particular items (such as canned goods or
meat) were perceived as cheaper and/or of higher quality, the number of stores visited in
a given month, as well as frequency of visits, was also assessed. No difference was
found between the average number of trips that each group made to go buy food (average
for both groups was around 15 food shopping trips per month). Surprisingly, there was

no apparent difference in the number of stores that each group visited during a given

128

month with both groups averaging 4.3 different stores in the past 30 days. Given the
emphasis in the interview on pursuit of sales and low prices at different stores, this issue
was further addressed by comparing the number of stores each group selected based on
the thirteen store selection criteria. No signiﬁcant differences were found between
groups.

Differences in shopping frequency did become apparent in comparing how
frequently each group shopped at the thirteen stores examined. Differences, bewever,
were only statistically signiﬁcant in the case of Meijer, Apple Market, and Wal-Mart.
While the difference was not very large with any of the stores, the orientation of food
insecure individuals towards stores that were seen as low~cost retailers continued to be
evident. Food secure individuals shopped more frequently at Meijer and Apple Market
while food insecure individuals shopped more frequently at Wal-Mart.

There was not a statistically signiﬁcant difference between the number of
different stores that food secure and food insecure shopped at or the amount of time spent
cooking, preparing, and preserving food. There was a difference in the number of hours
spent shopping for food (p<.05); however, in contrast to statements made during
interviews, it was the food secure individuals who spent slightly more time shopping than
food insecure individuals.

While the food that was purchased and consumed was not explicitly part of this
study, one question did involve the amount of fresh produce consumed on a weekly basis.
In other studies, fresh produce consumption has been used as a proxy for measuring
‘healthy’ food consumption (Wrigley et al. 2002), representing at least a crude indicator

of how food purchasing practices might be inﬂuencing the health of consumers. Notably,

129

there was a statistically signiﬁcant relationship between food security status and the
amount of fresh produce consumed in the past week with food secure individuals more

likely to have eaten ﬁve or more serving of fresh produce in the past seven days.

Conclusions

Both quantitative and qualitative data indicate a signiﬁcant similarity in food
shopping perceptions between food secure and insecure populations. In the interviews, it
came as no surprise that food insecure individuals voiced signiﬁcant concern with costs;
however, the differences in perception of the food retail system between food secure and
insecure individuals were not as pronounced as expected and cost and food quality were
an issue for both groups. Differences were more apparent in how the two groups
balanced cost and quality. Food insecure individuals did not select stores and items
based only on cost, but rather negotiated cost against quality. Sometimes this negotiation
was only possible when dealing with items or stores of perceived equal cost; however,
quality remained an important part in decision making. Similarly, despite growing
theoretical claims that quality was becoming more important than price, most food secure
individuals also noted the necessity of balancing cost against quality. Convenience and
distance also appeared to be issues of concern for both groups.

Despite similarities in perception, the shopping behaviors of the two groups were
notably different. Speciﬁcally, there were differences in the ways in which cost and
quality were negotiated. For instance, both groups noted being inﬂuenced by sale items,
though food insecure individuals were much more likely to go to a store speciﬁcally

because of a sale item. Notable was the fact that food secure individuals only stated three

130

strategies for food cost reduction—reviewing ads, sale seeking and bulk buying. In
sharp contrast, food insecure individuals described a range of cost reduction strategies
including reviewing ads, sale seeking, bulk buying, shopping at multiple stores, shopping
at non-traditional outlets such as dollar stores, preserving food, buying seasonal items,
using coupons, taking free samples and condiments, and drawing on food assistance.

The similarity in perception demonstrated by both groups would seem to indicate
that the cultural inﬂuence of social class on food purchasing perceptions may not be as
pronounced as one might expect from the work of Veblen or Bourdieu. For both of these
theorists, the process of consumption involved the association of cultural values based on
social class. While there is ample evidence that food secure and food insecure
individuals come from different social classes, data collected in this study indicate that, at
least in terms of perception of food retailers, values associated with retailers tends to be
similar.

Does this represent a broader social breakdown in class difference in
consumption? This is hardly the case since, as discussed in the previous chapter, class, at
least as evident through income and education, appeared as one of the most signiﬁcant
factors inﬂuencing food security status. Rather it points to the fact that, at least in terms
of food shopping, class has relatively little inﬂuence on store perception. That being
said, class does have an inﬂuence on food security and food purchasing behaviors.

Analysis of the different shopping behaviors of food secure and insecure
individuals indicates that food system structure clearly places constraints on the shopping
decisions made by both food secure and insecure individuals, with the latter

demonstrating a limited ability to negotiate ﬁnancial constraints, but often doing so at the

131

expense of meeting desires for food quality. Despite these apparent similarities between
the two groups, examination of shopping behavior patterns showed more signiﬁcant
differences. While both groups did more of their food purchasing from supermarkets,
food insecure individuals had a greater propensity to also purchase food from low cost
retailers such as Wal-Mart or Save-a-lot or from sources such as convenience stores that
do not necessarily have lower prices, but are used due to perceived convenience (location
and proximity).

The implications of this difference in behavior are signiﬁcant when related to the
difference in perception of food retailers. For instance, Save-a-lot and Wal-Mart were
perceived as desirable primarily because of their low prices, not because of food quality
or variety. Yet a disproportionate number of food insecure individuals tend to frequent
these stores. This suggests that the decision that food insecure individuals are making to
shop at these sources involves the all too necessary choice of forgoing desired food
quality or selection in order to meet a budget. While this likely comes as no great
surprise, it represents that any breakdown in class differences in consumption is still very
much limited to basic ﬁnancial differences.

So, while our two groups demonstrate some convergence on an ideological level,
there is also divergence at the behavioral level. In a sense, this points to a materialist
explanation for food insecurity and hunger rather than an ideological one. Food
insecurity does not appear to stem from cultural preferences that are part of class culture,
but instead food insecure people share similar values as those who are food secure. The
difference comes when we look at the abilities to act on those values. The implications

of this are signiﬁcant for practitioners who seek solutions to nutritional problems

132

associated with poverty through education. Based on this study, the beneﬁts of
educational programs that seek to alter how those who are food insecure perceive and
negotiate their food retail environment are likely to have only limited success. Food
insecure individuals already demonstrate rational actions to both minimize food costs
while trying to maximize food quality and reliability of access. The problem is that they
have only limited resources to work with, which limits their range of options.

How do we know that the behavior of food insecure households represents
structural constraints by the food system? In Chapter Four, I demonstrated that food
insecurity is the result of larger structures of inequality, primarily class. In the ﬁrst part
of this chapter, I demonstrated that cultural differences, at least in relation to perception
of food retailers, did not differ signiﬁcantly between food secure and food insecure
households. Finally, I showed that, despite the similarity of perception, the two groups
differed in their shopping behaviors. The divergence in behaviors given a conﬂuence in
perception indicates that there is a non-cultural factor that is limiting the behavior of the
food insecure group. What is limiting the behavior of the food insecure group? Is that
limiting factor an element of the structure of the conventional food system?

In the next chapter I will answer this question by examining one characteristic of
the food system related to problems with food access which has begun to receive
attention by some scholars—space. As I have explained, the evolution of the modern
food system is characterized by increasingly large food retailers. In order to ﬁnd space
for these large stores, food retailers have situated stores at the periphery of cities. The
growth of these large retailers has met the needs of an increasingly automobile oriented

society as well as the expansion of suburbs. The continued movement of these stores to

133

the outer fringes of residential areas, and in some cases the decline of stores still located
in inner cities, has led to greater spatial inequality in terms of resources required to access
food retailers. Now, as greater economies of scale are found in superstores, certain
segments of the population must either make due with limited selection (and at times,
higher prices), from small grocers in close proximity or must allocate scare ﬁnancial and

time resources to leaving the neighborhood to ﬁnd food.

134

CHAPTER SIX

SPACE AND FOOD DESERTS

In the last chapter I demonstrated that food insecure and food secure households
perceive food retailers very similarly. Based on twelve different retailer criteria which
were applied to eleven different retailers, I found very little difference in the perceptions
of the two groups. Furthermore, qualitative data suggested that food insecure and secure
households have many of the same desires for healthy, affordable, high quality food.
Despite these shared perceptions, households often deviated in terms of the food retailers
that were selected (Sam’s, Walmart, Apple, Sav a Lot). In this chapter, I will examine
the role of space, speciﬁcally distance, in determining the differing behaviors of food
secure and insecure households. As discussed in Chapter Two, space has started to
receive more attention in sociology and, particularly with the evolution of the food desert
concept in the 1990’s, many scholars have suggested that distance has a signiﬁcant role
in determining food insecurity. The distribution of food retailers has certainly changed
over time as the commercial food retail system has evolved. I would now like to examine
the extent to which spatial distribution, as a characteristic of the commercial food system,
inﬂuences the behaviors of food secure and insecure households.

Earlier I highlighted the fact that modern food retailing is increasingly dominated
by large retailers, often with a supercenter format. These supercenters, and even large
supermarkets, are typically located in suburban areas or urban edges where large tracks of
land are readily available and afﬂuent consumers are in close proximity. This has led to

the decline in small grocers which once served inner cities. A similar spatial pattern of

135

food retailers is evident in the neighborhood examined in this study (Figure 14).

‘ l

Food retailers
e Supercenter
a Supermarket
0 Convenience or Party Store
’ Roads
I Study site

 

Figure 14: Distribution of convenience and party stores, supermarkets, and
supercenters within and around the study site

The food retailers located within the neighborhood consist primarily of small
convenience stores and one, relatively small, conventional supermarket in the middle.
An additional supermarket lies at its eastern edge. A Kroger lies to its north east. The
closest supercenters to the neighborhood is a Sam’s Club, which is a membership based
bulk buy store and a Meijer, both located at the north side of the neighborhood. Notably,
according to the US Census, the northern part of the neighborhood has signiﬁcantly
higher median incomes than the southern parts. It is also worth pointing out that the
western edge of the neighborhood is closest to the center of the city and also the area

farthest from supermarkets and supercenters.

136

These patterns indicate a food retail system with a spatial distribution similar to
the national trends discussed in Chapter Two. Despite the residential characteristic of the
neighborhood, there are many food retailers located within the boundaries of this area.
However, these tend to consist of small convenience or specialty stores. The farther you
get from the residential areas, and consequently the more land is readily available for
development, the larger the food retailers. So at the edges of the neighborhood you get
larger supermarkets, and ﬁnally, to the northwest in the least developed area is the closest
supercenter.

Space has both a subjective and an objective dimension. By subjective, I refer to
the extent to which people perceive distance as inﬂuencing their food retailer choices.

By objective, I refer to the extent to which household decisions and circumstances are
related to actual distance from various types of food retailers. Examining both
dimensions is important in reconciling objective distance as a measure of the structure of
the food system and subjective distance as a measure of household agency in the context

of that structure.

Analysis and Results
Perceptions of Space

In order to assess the subjective element of space in relation to store choice, I
focused on a series of survey questions that related to the inﬂuence of distance on store
choice. Speciﬁcally, I looked at the extent to which respondents indicated that distance
from their home, work, bus stops, and other stores dictated where they purchased food.

This would indicate the extent to which distance was seen as relevant to choice of store.

137

While one could argue that any store where food is purchased is “close enough,” it is
important to understand whether and how distance is consciously considered in an area
with many different food retailers. The question is not whether distance matters, but why

and to what extent it matters.

Qualitative Results

Eight of the nine food insecure respondents I interviewed indicated that distance
was a factor that inﬂuenced at least some of their store choices. Distance was an even
bigger problem for two of the respondents who either did not have an automobile at the
time of the interview or were having signiﬁcant mechanical problems with their cars. As
one respondent explained, “Mostly it’s because it’s on the bus line. I can get there. Or
walk, even.” One interviewee did not have a driver’s license, which required that s/he
rely on rides from family members to get to the store. While s/he had several older
children that made this relatively easy, being the primary food preparer made distance to '
stores a meaningful factor. Even for those with access to reliable transportation, distance-
was often a concern that determined where they went shopping and limiting access to
some of their preferred food retailers. One person explained, “Kroger’s would usually be
the third place, and Wal-Mart would be the last place. Like I said, but that’s because of
convenience. Wal-Mart’s way out in Charlotte.”

Food secure individuals also voiced concern with the inﬂuence of distance on
their store choices. This was not as big an issue, as only three of the seven indicated that
distance was a primary factor. While none of these individuals faced being without an

automobile, the time required to travel to different stores had to be carefully negotiated

138

against other interests, such as selection. One respondent indicated that s/he was at times
unable to get vegetables from the preferred source due to its distance: “I mean that would
be our preferred place to get the vegetables and stuff, it’s just the convenience sometimes
of Shop-rite, or if you need other things too; that are more paper products and ice cream
and things like that.”

While interviews indicated that distance was an issue, if not a problem, for both
food secure and insecure households, it appeared to be more of an issue for food insecure
households. First is the simple issue of having transportation to get to the store. Second
is the concern of getting to the store that is preferred. Finally, several food insecure
individuals indicated that seeking out discount stores or going to several different stores
for select sale items were important cost management strategies. Both of these strategies
were limited by factors related to distance. Food secure households, on the other hand,
simply referred to distance as an issue of convenience that inﬂuenced how they spent

their time or the selection of foods that they were able to get on short notice.

Quantitative Results

As part of the survey, respondents were asked whether distance to home, work,
other stores, and bus stops inﬂuenced their decisions to shop at a particular location.
Distance to home was clearly the most important thing respondents considered (89% of
the sample indicated that distance to home inﬂuenced their decision to shop at least one
store in the past month). Distance to work was the second most important (30%),
followed by distance to other stores (24%). Distance of stores to bus stops inﬂuenced

store choice for less than 6% of the sample in the last 30 days.

139

The importance of distance to home was also evident in primary store choice
(75% of respondents indicated that this was one of the factors behind their decision to
obtain food from their primary retailer), while distance to work and other stores was only
relevant for 17% and 16% of the sample respectively, and distance to bus stops was
important to 5% of the sample. Interestingly, despite this apparent emphasis on selecting
a primary retailer based on distance to home, less than 9% of the sample actually shopped
at the food retailer that was closest to their home. This indicates that proximity is clearly
being balanced against other factors and is not the only concern inﬂuencing primary store
choice.

A t-test comparison between food insecure and food secure households revealed
no statistically signiﬁcant difference related to distance to home, work, other stores, or
bus stops on their decision to shop at their primary food retailer. These results came as a
surprise as it was anticipated that food insecure households, which suffer from some level
of resource scarcity such as the time and gas that are required to travel greater distances.
Given the number of respondents who indicated that they did not have an automobile, it

came as even more of a surprise to ﬁnd no difference between the two groups.

Behaviors in space

Given the discrepancy between apparent differences in the perceived importance
of distance in interviews and the perceived importance of distance in the survey, it is
worth considering space from an objective perspective. One of the underlying
assumptions of the food desert literature is that physical distance, regardless of subjective

perception, is a factor that contributes to food security and makes it difﬁcult for

140

disadvantaged consumers to purchase food. In order to assess distance as an objective
measure, I used GIS (ESRI ArcView 3.1a) to map the location of survey respondents in
relation to neighborhood food retailers. Since addresses were listed on the survey form, it
was possible to geocode6 nearly all of the respondents. In three cases, respondents
removed the address label on the returned survey so coding was not possible. Geocoding
was conducted using an automated service provided by Tele Atlas North America, Inc.
This service was unable to match coordinates for 11 of the addresses, leavinga total of
289 responses which were spatially analyzed.

Since I was interested in the distance between consumers and food retailers, a list
was compiled of all food retailers within a ﬁve mile radius of the neighborhood. Names
and addresses of retailers were collected from two intemet business databases:
www.yellowpages.com and www.mapsonus.com. Additional retailers outside the ﬁve-
mile radius were added to the list if they were mentioned either in the qualitative
interviews or listed in surveys. The websites for large retailers such as Kroger, Meijer,
L&L, and other supermarkets were also checked to insure all locations were included in
the list. Businesses were categorized into supermarkets, grocery stores,
convenience/party stores, specialty stores, and other. Supermarkets included all large
chain food retailers such as Kroger and Meijer. Grocery stores included smaller,
generally independent food retailers. Convenience and party stores consisted of stores
I such as Quality Dairy, liquor stores, and gas station based food retailers where food is
available but selection is extremely limited. Specialty stores included both health food

stores and stores specializing in particular ethnic foods. Only one store was not included

 

6 Geocoding refers to the process through which a street address is converted into a speciﬁc geographic
location. Most commonly this involves associating the address with a set of latitude and longitude
coordinates.

141

in the above categories—Sam’s Club, a membership based warehouse store. Later
examination of retailer patterns led to a distinction being drawn between supermarkets,
which primarily sold food, and supercenters, which sold a combination of food and other
items. The location of all stores were geocoded using the same process as the surveys.

Two sets of spatial analyses were conducted. First, the distance between
respondents and the closest food retailer were examined. This distance was assessed
using an ArcView extension that automatically calculated the distance from each survey
respondent to the nearest food retailer (Jenness 2004b). Subsequent comparisons were
also made for the distance between survey respondents and the closest of each type of
food retailer. Next, the distance between respondents and the food retailer most
frequently used was assessed. In this case, a different ArcView extension was used
which matched the most frequently shopped at store to the location of the survey
respondent (Jenness 2004a).

In both cases, straight line distances were measured, not actual driving distances.
Therefore, distances do not represent actual distance one would have to travel to a food
retailer. This. straight line measurement is similar to the methods used to assess food
deserts which involve the creation of buffers surrounding retailers to look for residential
areas that are not close to any food retailers. In this case, however, I was able to quantify
actual distances, rather than only identifying gaps.

In general I found that survey respondents all live in very close proximity to food
retailers. The average distance from a food retailer to respondents was less than a third of
a mile. Some residents lived much closer and no respondents lived more than % of a mile

from a food retailer. Many of the stores in close proximity to respondents were not large

142

supermarkets. For 49% of the sample, the closest place was a convenience or party store,
while for 32% the closest food retailer was a grocer (this can be accounted for by the
presence of a single small grocer located in the center of the neighborhood), and 7.2%
were closest to two small ethnic grocers. Notably, the distance that respondents did
travel to shop for food was over twice the necessary distance with the average commute
to the primary food retailer being about two miles.

One of the central arguments of the food desert literature is that distance from a
food retailer increases the likelihood that a household will be food insecure, at least for
disadvantaged consumers. A t-test revealed a statistically signiﬁcant difference (p<.05)
between the distance from food insecure and food secure households and the closest food
retailer (Table 10). The difference was very small with food insecure households located
only 0.05 miles farther from the closest retailer than food secure households.

Since the vast majority of food retailers in the neighborhood and surrounding area
were convenience or party stores and the majority of respondents indicated that
supermarkets were their primary food retailer, additional comparisons were made
between the average distances to supermarkets and convenience stores (Table 10).

Again, a t-test revealed a statistically signiﬁcant difference in the distance to both
supermarkets (p<.01) and convenience stores (p<.01). Food insecure households were
located an average of 0.08 miles closer to convenience stores and 0.22 miles farther from
the nearest supermarket. While these straight line distances appear rather small, they
translate into much larger distances that people must travel following roads and sidewalks

to reach the nearest food retailer.

143

Table 10. Average distance in miles from food secure and food insecure households to
different food retailers

 

 

Food Food

secure Insecure
Distance to the most frequently used food retailer“ 2.04 2.25
Distance to the closest supermarket' 0.71 0.93
Distance to the closest convenience or party store‘ 0.38 0.30
Distance to the closest food retailer" 0.30 0.25

 

* Indicates a statistically signiﬁcant difference between groups (p<0.05)

Additional analysis was conducted using one-way ANOVA in order to assess
whether distance was related to food security with hunger in addition to just food
insecurity. In this case, relationships were found with proximity to convenience stores
and supermarkets (p<.01), but not to the closest food retailer. Post-hoe testing was
conducted using the Least Signiﬁcant Difference (LSD) test. Post—hoe testing of the
distances to supermarkets revealed signiﬁcant differences between those who were food
secure and those who were food secure with hunger, but not in other cases. This
indicates that distance makes a particular difference for those suffering from very low
food security. Post-hoe testing of distances to convenience and party stores revealed no
difference between food insecure and food insecure with hunger respondents, but
difference between both levels of food insecure and those that were food secure.

The inﬂuence of convenience stores was further apparent in the distances from
households to convenience stores. A t-test revealed a statistically signiﬁcant difference
between the distance from food secure and food insecure households. In this case, food
insecure households tend to be located, on average, closer to convenience stores than
food secure households. This suggests that convenience stores are more likely to be
located in areas of high food insecurity. Despite the high density of convenience stores in
the area, relatively few respondents identiﬁed this type of store as being a primary

retailer. It is reasonable to suggest that few consumers are able to rely solely on the

144

selection available from these retailers. Therefore, additional spatial analysis was
conducted looking speciﬁcally at supermarkets.

Notably, only four different retailers were represented in a spatial analysis of
nearby supermarkets. For 72% of the sample, Apple Market was the closest supermarket,
followed by two different Kroger stores (16% -- eastern border; 7% northeast of the study
site), and a Meijer (5%) Less than 1% of the sample was located in close proximity to a
Shop Rite located in the southeast of the study site. There was no statistically signiﬁcant
difference between food security status and proximity to any of these retailers. This was
somewhat unexpected due to census data that indicate higher income and education levels
at the northern edge of the study site. However, due to the central location of the Apple
Market, food security levels as related to proximity to different supermarkets might not
be readily apparent since this single store may have obscured the inﬂuence of distance
that might be apparent across a larger space.

The next set of analyses involved comparing the distances from respondents’
homes to the food retailer where food was most frequently purchased. This revealed two
trends. First of all, there was a statistically signiﬁcant difference between the two groups
in terms of distance to their primary food retailers with food insecure households being
located, on average, farther than food secure households (the difference was 0.93 miles
for food insecure respondents and 0.7 miles for secure respondents). More notably, food
insecure individuals appeared to travel farther than those who were food secure to their
primary retailer. In this case, food insecure households were located an average of 2.3
miles from their most frequently used food retailer while food secure households were

located only 1.8 miles away.

145

This difference in behavior is particularly notable since analysis of survey data
indicated no difference between the two groups’ spatial perception as measured by how
the groups thought about distance from home into their shopping decisions. This
difference in behavior, while no doubt inﬂuenced in part by the fact that they lived on
average slightly farther from available food retailers, is nonetheless interesting given that
there was no apparent difference between the groups in terms of how distance from home
inﬂuenced primary store choice. Again, this seems to reiterate the fact that perception of
distance as a factor inﬂuencing store choice was the same for each group; however, how
each group negotiated the distance to stores differed with food insecure households
traveling farther.

This suggests that cultural preferences or perceptions related to class are not
inﬂuencing the importance of distance on store choice. At the same time, food insecure
households are driving farther in order to access their preferred stores and therefore
suffering an additional cost due to the location of food retailers. If agency represents a
measure of the extent to which a structure serves the preferences of a group, it appears
that, at least in terms of physical distribution, food secure households are being better
served. This is despite survey results that indicate a similar set of interests between the
two groups.

While a t-test revealed no statistically signiﬁcant difference between the average
distances from primary food retailer to home for both groups, there was a moderate
negative correlation (-.212) between the proximity to the closest food retailer and the
distance traveled to the primary food retailer. This suggests that the closer a consumer

lived to a food retailer, the farther that they were likely to travel to a retailer. While these

146

results contradict the trends suggested by the food desert literature, this can be explained
in part by the confounding factor of the convenience and party stores. In fact, the closest
food retailer to over half of the sample consisted of a convenience or party store. With 62
different convenience or party stores in the surrounding area and ten stores located within
the neighborhood this comes as no surprise.

To control for the inﬂuence of convenience stores, correlations between distance
to supermarkets and distance to primary retailer were also assessed. There is a moderate
positive correlation (.186) between distance from a supermarket and the distance traveled
to the primary food retailer. Interestingly, when the sample was divided into food secure
and foodinsecure groups, the correlation continued to be statistically signiﬁcant.
However, the distance traveled by food secure households to the primary food retailer
was inﬂuenced more by proximity to a supermarket (.146) than food insecure households
(.298). This suggests a few things. First, there is a relationship between space and food
security when looking only at supermarkets. Second, for food secure and food insecure
households, distance inﬂuences store choice; however, for food insecure households,
proximity has a much stronger inﬂuence.

Despite this apparent relationship between distance and supermarket selection, it
is important to keep in mind that most residents do not shop at the supermarket that is
closest to them. In fact, only 15% of the sample did most of their shopping at the closest
supermarket. Furthermore, food insecure households were no more likely than food
secure households to shop at the closest supermarket. This suggests that while distance
does play a signiﬁcant role in food retailer selection, it is not necessarily the force that is

driving households, food insecure or otherwise, to the closest supermarket. In other

147

words, distance to food retailers inﬂuences where people shop, but being the closest

retailer may be insufﬁcient to cause people to shop there.

Conclusions

By utilizing GIS to examine the spatial distribution of consumer responses, I was
able to look at the role that space plays on consumer behavior and perception. In the
previous chapter it was demonstrated that food secure and food insecure households tend
to have similar perceptions of food retailers, though behavior in relation to food retailers
differed, and it is not clear at this point what factors may be inﬂuencing store choice.
Changes in the structure of food retailers has led to growth in areas increasingly far from
residential areas and food desert research suggests that this has led to problems among
disadvantaged populations in accessing food. The spatial distribution of food retailers in
this study site has similar patterns.

Does this mean that distance limits the behavior of food insecure households?
The results of this study demonstrate trends very similar to those found in the previous
chapter. Both food secure and insecure households indicated similar perceptions of
space. Speciﬁcally, neither group was more likely to indicate that distance from home,
work, nor bus stops inﬂuenced their store choice. This was a bit of a contrast with
comments made during interviews that suggested that distance may be a factor limiting
shopping habits, particularly within food insecure households. Furthermore, research has
suggested that not owning a car can be a signiﬁcantly limiting factor to food purchasing,
which suggests that distance may be more of an issue for some segments of the

population. It is not clear from this research why this might be the case. Possible

148

explanations might include public transportation, the availability of convenience stores,
or the ability to share transportation with other households. Even though it was not a
factor predictive of food insecurity, it was expected that it would at least have an
inﬂuence on store choice.

Further analysis indicated that of those without cars only 26% actually shop at the
store closest to them. While this is higher than the sample percentage of 15%, it is still
not as high as food desert research might suggest. Even the distance that households
without cars were located from their most frequently usedfood source did not appear to
differ signiﬁcantly from that of food secure households. Those without cars were on
average 2.2 miles from their primary food retailer, which is very close to the aggregate
value for food insecure populations of 2.3 miles. This suggests that the inﬂuence of
having a car, even in a society so oriented around the automobile as the US, might not be
as great as previously thought. Even those without cars who were interviewed noted
several strategies to get to the store, primarily relying on rides from other people or using
the bus, soit is possible that households without cars are simply ﬁnding other ways to get
to the store. Proximity to a bus stop might matter more than whether you need to ride on
the bus for another few minutes and/or a few more miles to your preferred store.

It is possible that while distance was equally important for both groups, the way
in which distance was important may differ in a fashion not detected by the survey. For
instance, during one interview a food secure individual explained that “100%” of his/her
shopping was based on distance. However, it was not because s/he lacked transportation,
but rather because of the time required to travel farther. Jabs and Devine (2006) cite time

as a signiﬁcant factor that inﬂuences food consumption practices. These authors also

149

note that time can be an even more signiﬁcant constraint for low-income households who
are unable to “buy time” by paying others to prepare food, clean, or engage in other
activities in exchange. Several food insecure individuals mentioned during interviews
that distance did inﬂuence their store choice, especially in the context of transportation
problems or cost saving strategies that required trips to several different stores.
Therefore, while survey results indicated that the importance of distance on store choice
was the same for. each group, it seems likely that the reasons for, and consequences of,
distance differed between groups, even if it was not apparent from the survey.

This is further supported by results that suggest a relationship between food
security status and proximity to food retailers. Food insecure households were located
slightly farther from supermarkets while slightly closer to convenience stores. The
farther one was located from a supermarket and the closer to a convenience store, the
farther a respondent was likely to go to shop.

Perhaps it comes as no surprise that distance to a supermarket correlates with
distance traveled to the most frequently used store. However, when we also consider that
food insecure households tend to be located farther from supermarkets we see clear
overlaps between the structure of the food system and the differing agency of consumers.
Is distance the only factor that contributes to food security? Certainly not, and when
distance to supermarkets and distance to nearest food retailer were added to the logistic
regression models, they were not found to be statistically signiﬁcant predictors of food
security. However, the point of the food desert literature has never been to suggest that

distance from food retailers is the only factor leading to food insecurity. Rather, as our

150

results also suggest, the point is that distance can create an additional burden on segments
of the population that are already struggling to obtain food.

Distance to a supermarket itself might not be enough to dictate whether a
household is food secure or insecure. However, this does not mean that distance fails to
inﬂuence food purchasing. Our results indicate that it might inﬂuence whether a
household is able to manage that insecurity without going hungry. Research into the
elements of the food system that weigh disproportionately on disadvantaged segments of
the population provides an important insight into the structural constraints that many
people face. Food desert research has certainly pointed to the negative impacts of a food _
system that is migrating farther and farther from city centers. It will be important that
this research not limit itself solely to issues of distance, however, and begin also to look
for other factors that may be constraining choice or creating a burden on those struggling

to obtain food for themselves and their families.

151

CHAPTER SEVEN

CONCLUSION

Agency Revisited

Hunger and food are both problematic in the US, despite the fact that there are
nearly 4,000 calories of food available per person (Putnam et al. 2002). Theproblem is
not an issue of quantity, but one of distribution. Increasingly chronic health problems
from heart disease and cancer to diabetes and obesity are being related to nutrition.

While in the past, food consumption decisions have been seen as individual matters,
recent developments in ﬁelds such as community food security and food deserts have
begun to examine the relationship between hunger and nutritional inequity and the
structure of the modern food system. While income plays a critical role, food desert
scholars have suggested that other characteristics of the food system, in particular the
spatial distribution of food retailers, have placed a disproportionate burden on those who
are already struggling to reliably obtain nutritious food.

This research project does more than simply identifying the factors that contribute
to food insecurity and hunger. At its foundation, it examines the extent to which the
modern food system is structured in a fashion that precludes certain segments of the
population. Some scholars hail the era of postmodernism and the reﬂexive consumer.
But many have forgotten that the same market forces that have, at least to some extent,
empowered consumers has certainly not empowered all. The structure of the food system

is inﬂuenced by the changing tastes and interests of buyers—hence the rise of organic

152

food and 100 calorie snack-packs to meet new environmental and health concerns.
However, is this really based on the interests of all consumers?

This is a question with which sociologists are still struggling to answer. To what
extent do consumers inﬂuence the structure of the system, and to what extent are
consumers simply selecting from what has been made available to them? I suggested in
Chapter Two that to understand the position of the consumer vis-a-vis the structure of the
food system requires an appreciation for the varied levels of agency among consumers.
Since we exist at the intersection of a variety of social structures that inﬂuence our tastes
and desires, the issue of differential agency really addresses the extent to which
consumers are able to align the structure of the food system with their interests as related
to other social structures.

Food security is not only a meaningful measure of a household’s ability to
purchase food—it also represents a measure of agency vis-a-vis the food system. When a
household begins skipping meals, cutting portion sizes, or simply going hungry on a
regular basis, it seems realistic to conclude that household’s interests are not aligned with
the structure of the food system. Certainly money plays a large part in this, and a food
system oriented primarily along market lines is going to exclude those with insufﬁcient
funds. But is this the only way that the food system excludes people?

' The expansion of food retailers from small, downtown grocery stores to today’s
suburban superstores has also involved a gradual “migration” of food retailers to the
edges of communities where large tracks of land are cheap and easy to obtain. At worst
this has involved the closing of small grocers in inner cities and at best it has simply

meant the wide selection and low prices of supermarkets and superstores has never

153

developed in urban areas. Urban ﬂight and the common availability of automobiles have
meant this movement has had little impact on the ability of retailers to attract sufﬁcient
consumers to remain economically viable. On the other hand, it has certainly changed
the physical structure of the retail landscape. In the case of the neighborhood examined
in this study, it has led to one area with a heavy concentration of convenience and party
stores, but relatively far from supermarkets and superstores in an urban setting.

The inﬂuence of distance on food security status was not apparent from this
research. This suggests the concept of food deserts requires careful consideration and
should not be assessed based only on retailer location, without consideration for actual
consumer behavior and different types of food retailers. Nonetheless, distance did appear
to be related to food purchasing patterns. Perhaps most notable was the fact that distance
from a supermarket appeared to correlate with whether a household was food insecure or
food insecure with hunger. Does this mean that the increased distance created the
hunger? It would be premature to draw such simplistic conclusions. However,
qualitative interviews as well as survey results indicated that distance was something that
respondents considered when deciding where to shop and, even if it was not the primary
factor contributing to a household’s food security status, it certainly represents an
additional burden on a household with limited resources.

Korczynski and Ott (2006:912) offer a meaningful, and appropriate, metaphor for
the question of agency and structure vis-a-vis the food system. They offer the menu as a
heuristic device stating that “the menu operates as a key form of mediation between
enchanting myths of individual sovereignty and structures of power. It acts as a buffer

between the two in such a way as to allow their simultaneous co—existence.” Certainly

154

the idea of a menu offers equal insight into how the structure of the food retail system
constrains households while at the same time not completely disempowering them. The
structure of the food system presents itself as a menu with a different set of items and a
different set of costs for various people depending on their resources and position in
relation to food retailers.

Even food insecure households have some choice in their selection from the menu
offered by food retailers. Qualitative interviews highlighted the complex strategies that
many food insecure households utilized in order to obtain food. Even individuals without
cars were able to negotiate rides with friends and family members or use public
transportation. A dollar can certainly be stretched through the use of coupons, seeking
out sales, or other cost-saving strategies. While such strategies are useful in that they
prevent, or at least reduce, problems due to insufﬁcient food, I would note that they still
require households to balance one resource in such a fashion that limits their selection. If
it is not on sale, on clearance, being sold at a store where a friend is already going, or just
plain inexpensive, the item is not on the menu. One interviewee stated, “So you have to
choose from this to that, from deodorant to a gallon of milk. What are you going to
buy?” Even when hunger is successfully avoided, this does not mean that there are not
costs involved.

Similarly, distance represents different cost to different people. The movement of
large food retailers to the suburbs or farther has not necessarily put them out of reach of
disadvantaged consumers. However, it might act as another factor that limits what is
available on the menu. For a disadvantaged household the ﬁnancial cost of going farther

for food must be balanced against the price of the items on sale. Several interview

155

respondents stated that they were willing to drive to relatively distant discount stores or
visit multiple stores for lower priced items. Time spent driving is time that could have
been spent clipping coupons, preserving food, or engaged in other activities. Money
spent on gas or a bus ticket is also money not being spent on food. The consequence of
living farther from supermarkets might not be felt directly through food insecurity, but it
acts as another factor that must be negotiated against other scarce resources. Survey
results indicated that distance was the primary reason people go to convenience stores
and nearly four times as many food insecure than food secure households indicated that
they shopped at convenience stores more frequently than any other store.

Keep in mind that food insecure households tend to be located farther from
supermarkets than food secure households. This does not suggest that increased distance
leads to food insecurity. Only a longitudinal study would be able to determine that
causation, and even the data point to a variety of other factors, such as income, education,
and disability status, which are more likely to inﬂuence food insecurity. Nonetheless, the
difference in distance from supermarkets points to the fact that food insecure households
may have an increased resource strain placed on them due to the spatial distribution of
food retailers.

It is worth reiterating that the apparent difference in behavior of food secure and
food insecure households did not appear to stem from cultural differences. This point is
important as, building on Bourdieu, a great deal of scholars point to class culture as
inﬂuencing taste. While this may be at work in some circumstances, class culture did not
appear to dictate the store choices of respondents in this neighborhood—even though

incomes ranged from over $80,000 per year to below $10,000. In fact, respondents

156

appeared to perceive stores similarly in terms of cost, selection, quality, or other criteria.
Both food secure and food insecure individuals who were interviewed indicated the need
to balance food quality against cost. For those who were food insecure, quality was often
sacriﬁced for cost. This trend has been found in other areas (Caraher et al. 1998).

Food insecure households are not powerless in the context of the food system.
Nor, however, are they able to inﬂuence its structure to meet their needs to the same
extent as others. This does not stem from a lack of interest in being able to afford high
quality food as quite the opposite was apparent in this study. Lack of knowledge was not
a signiﬁcant problem. The problem is simply that the price of the food is too high and, as
our assessment of spatial relations indicates, the distance required to get that food only
drives the price higher. The total cost that food insecure households must pay to
purchase food appears to be higher than food secure households, if not in real terms than
deﬁnitely in relative ones. Consider the greater distance from primary food retailers to
food insecure households and the effective, but time and resource costly, strategies of
sale seeking and driving out of the way to discount grocers.

The lack of apparent inﬂuence of class culture on food retailer selection points to
a rather singular explanation as to why, in Figure 5, food insecure households are situated
at the bottom of the chart, with less apparent agency, and food secure households are
situated at the top, with greater apparent agency. The conventional food system caters to
these, relatively afﬂuent, consumers. While it makes rational economic sense for the
corporations, the consequences of this on disadvantaged consumers is often
unconsidered. In many cases, the food system simply diversiﬁes to provide consumers

with a wide variety of products that meet the demands of a variety of individuals. So we

157

do see the growth of organic, specialty, and health foods in response to an increasingly
reﬂexive population. Of course, these consumers may also beneﬁt indirectly as large
superstores make economically rational decisions that also happen to lead to locating
stores in convenient proximity to food secure households.

At the same time, however, we see food insecure households having relatively
more access to the massiﬁed elements of the food system. Consider the lack of food
selection in a convenience store or discount grocer and this points to the different ways in
which the food system treats different segments of the population. Highly processed,
mass produced products with lengthy shelf-lives tend to be the norm. 1 have yet to see
hand-crafted goat cheese or greenhouse tomatoes at a typical party store.

This kind of market behavior is economically rational for food retailers and I
would not suggest that selling goat cheese at a 7-11 would somehow alleviate food
insecurity. However, it was apparent that food insecure households did have a genuine
interest in conveniently obtaining nutritious food at a cost they could afford. This is not
something that they are ﬁnding at convenience stores and it is something that requires the
use of scarce resources to obtain elsewhere. And, of course, given the purchasing power
of most food insecure households compared to food secure households, providing
convenient sources for affordable, nutritious food will always take second place to
overpriced packages of dehydrated macaroni and cheese—a product readily available in
most convenience stores.

In short, it has been indicated that there is a failure of the conventional food
system to readily provide sufﬁcient, affordable food to large segments of the

neighborhood. This failure does not stem from a lack of demand due to a difference in

158

cultural preferences related to class. This failure stems from the unequal distribution of
resources that not only leaves a segment of the population struggling to make ends meet,
but also has them paying a higher price for their food (either literally or by requiring them
to travel farther to obtain the food that they need and want). Ironically, the very segment
of the population least able to cope with the burden of the health consequences of
nutritional inequity is the same one for whom avoiding nutritional inequity requires
overcoming the most obstacles.

There is, however, hope. There have been new efforts to change food availability
in the neighborhood which, in many cases, involve drawing upon resources and social
structures that are not part of the conventional food system. This highlights that limited
agency in one area does not preclude agency in other arenas. Generated from the same
part of the neighborhood where food problems are most pronounced is a small
neighborhood center that has for the last few years begun to address the problem of_ food
access. As I will discuss in the next section, this neighborhood center has managed to tap
into national social movements, as well as state level resources, in order to initiate
programs to alleviate food security issues. Much as the structure of the food system has
created a problem associated with space, this neighborhood center has turned to space,

and place, as part of the solution to the problem.

Resistance Movements and Alternative Food Sources
Limited or constrained agency in the context of the conventional food system
does not inevitably lead to food insecurity. In this neighborhood, a local center made up

of area residents, some of whom are food insecure, has been working to make this the

159

case by increasing the availability of healthy food in the neighborhood. Increased
concern that the neighborhood food system was not adequately meeting the needs of all
residents stimulated the Allen Neighborhood Center to begin organizing around the topic
of food and in the past several years they have expanded their activities to make food a
part of many of their activities.

This group currently has a spectrum of food related programs, some of which are
market oriented and some which rely on donations. On a weekly basis, free bread is
available to neighborhood residents and, when available, other food, such as fresh
vegetables, is also provided. This program is critical for many residents, as it not only
provides food, but it puts disadvantaged residents in contact with other resources
available at the neighborhood center, such as health education programs. Center staff
estimates that between 80 to over 100 families take advantage of this program on a
weekly basis. For the past few years, the center has worked with a group of children to
grow food in a community garden, which creates a new source of fresh vegetable within
the neighborhood and teaches children important skills. Recently, some children have
also sold the produce they have grown in the neighborhood’s farmers’ market.

The farmers’ market, which operates in a parking lot adjacent to the neighborhood
center, was started three years ago with only a handful of farmers. The market was
initiated by center staff who were aware that fresh produce was not readily available in
the neighborhood. Bringing fresh produce directly into the neighborhood was done to
reduce the distance people had to travel to obtain food. At the same time, by holding a
farmers’ market where food was purchased directly from farmers, the cost of the produce

was held down. Further innovations, such as creating a system to accept food stamps,

160

have made the farmers’ market even more accessible to a variety of households. After
only a couple years of operation, they estimate 500 people were served each week during
the 2006 season. More recently, the center has obtained funds and the use of a local park
to develop an urban greenhouse project. They anticipate using the greenhouse as an
education site as well as a location to help neighborhood residents grow food all year.
Construction of this greenhouse was completed in the summer of 2007.

The efforts of the neighborhood center point to potential solutions to the problem
of food insecurity. If food insecure residents are not being fully served by the
conventional food system, it seems logical that they should seek and create other social
structures to meet their needs. Just because a group lacks agency in the context of one
structure (the conventional food system) does not preclude their agency in the context of
other structures (the alternative agriculture movement). In some cases these solutions
involve creating new connections with other people who feel underserved by the food
system—which is the case with the farmers’ market that is helping small farmers
maintain their economic viability in an era of large agribusiness dominance. In other
cases, solutions involve simply shifting community resources to help those who are at
greatest risk—such as the free bread program. Finally, the solutions might simply
empower residents to circumvent the conventional food system entirely—such as with
the children’s garden program and the urban greenhouse.

In fact, the success of the neighborhood center in creating alternative sources for
food has hinged signiﬁcantly on its ability to align its activities away from the
conventional food system, which is relatively resistant to change, including changes

meant to support food insecure households, though many of the large supermarkets have

161

bought into various aspects of alternative food systems (Belasco 1993). The center,
however, was able to use the alternative food movement in strategic ’ways. Despite its
history of providing free food to low-income residents and engagement in health outreach
programs, the neighborhood center had relatively little experience working in the area of
food activism when it began expanding its food related activities. The rapid growth of
programs at the center took place in large part because it was able to access political,
expert, and economic resources that were related to an associated social movement.
Entering into such relationships was based on the ability of the center to successfully
negotiate its own interests with those of the larger social movement. By drawingon the
rhetoric of the larger social movement, the center has been able to mobilize resources and
gain legitimacy in new ways.

The alternative agriculture movement is only loosely deﬁned as a social
movement and represents a set of values related to sustainable agriculture, supporting
small and family farmers, localizing production, organic production, fair trade labeling,
and community food security. In general, activities by organizations in this movement
have involved farmers’ markets, Community Supported Agriculture groups (CSA), food
policy councils, food system advocacy groups, and community gardening activities and
efforts (Koc and Dahlberg 1999: 113). The movement is largely deﬁned by its resistance
to problems associated with the conventional food system related to issues of food quality
and safety, environmental degradation, and community decline. Community food
security represents one element of this alternative agriculture movement.

Rochon (1998) has argued that a true social movement pursues both political and

social change as part of its agenda, and the success of the movement can be recognized

162

through changing cultural values. Certainly the success of the alternative agriculture
movement in enacting political change is apparent in policies such as the USDA organic
program (imperfect as it may be) as well as the development of USDA supported
community food security program. The growth of organic and local food in both
conventional food retail outlets as well as outlets such as community supported farms and
farmers markets suggests that the movement is making inroads along social lines also.
Rochon stated that cultural values were more difﬁcult to change; however, by building on
the successes of other social movements such as the environmental movement and in the
context of other social changes, we can see new levels of concern with the structure of
the agrifood system.

While it seems clear that the alternative agricultural movement is inﬂuencing a
certain level of social change, to what extent is the center part of this movement?
Certainly, the activities that the center has been engaged in are reﬂective of those
promoted by the movement. The center has held a farmers’ market, hosted a CSA,
facilitated a youth group community garden, sponsored a harvest festival, and is working
to create an urban greenhouse in a nearby city park. The center has worked with local,
organic farmers to develop ways to bring fresh vegetables into the city in a way that
beneﬁts both farmers and neighborhood residents. The center has also been developing
programs to increase the self-sufﬁciency of neighborhood residents by looking for ways
that they can grow food in their own neighborhood. All of these activities speak to a high
degree of collaboration between the center and the alternative agriculture movement.

Understanding the activities of the center in the context of the larger alternative

agriculture movement is important. This is because the activities of the center are to a

163

signiﬁcant extent shaped by the trajectory of the larger movement. Farmers’ markets and
CSAs, for example, have already gained legitimacy in the alternative food movement, as
well as in the broader public arena. Engaging in these activities is more likely to bring
social legitimacy to the center. This legitimacy provides the center with access to the
entire repertoire of research and rationales that the movement has already generated.
Such research and rhetoric has proved useful in grant writing and meeting with city
ofﬁcials, where center staff members have drawn from the pre—existing set of arguments
generated by the alternative agriculture movement in order to make their own case. In
addition to facilitating resource acquisition in other areas, the alternative agriculture
movement has itself served as a resource. Other programs and organizations have
provided the center with program ideas as well as information about how to implement
particular programs.

Despite this beneﬁcial association with the alternative agricultural movement, the
center is not simply another element of the movement, as the agendas are not the same.
Drawing from Rochon (1998), I would suggest that despite the ideological alliance
between the center and the alternative agriculture movement, they each stem from a
different set of critical communities and, consequently, operate with a different set of
concerns and identiﬁed problems. The alternative agriculture movement is driven to
remedy concerns with the conventional agrifood systems. The center, on the other hand,
is driven to remedy place speciﬁc concerns with the neighborhood. This is evident from
their website, which states that the center “serves as a hub for neighborhood education
and capacity building. We offer activities that promote the health, safety, and stability of

families and neighborhoods on the Eastside of Lansing.” In short, the mission of the

164

organization is focused on meeting neighborhood needs—not on promoting widespread
social change. While the alternative agriculture movement represents a fairly typical
example of a New Social Movement in its concern with lifestyle based issues and goals
of social change (Pichardo 1997), the center represents an example of a place-based
movement7 that is very much grounded in the people and place where it operates.

What does this mean? Well, certainly the alternative agriculture movement
contains within it a set of priorities related to re-embeding people in place, and while the
movement seeks to prioritize place in an abstract sense, the center prioritizes a specific
place. In the case of food related issues, the interests of the center in representing the
community overlap heavily with that of the alternative agriculture movement and that is
one of the reasons such an alliance has been formed. By looking at some of the
dissimilarities, we can shed some light on where the place-based movement of the center
ends and the alternative agriculture social movement begins.

Perhaps the best example of incongruence between the needs of the center and the
alternative agriculture movement is apparent in the CSA the center has hosted for the past
two years. Despite initial hopes that the CSA would help provide residents with
convenient access to fresh vegetables, the high cost of shares has meant that each year
only a small handful of residents have participated in the program. For the most part the
center has served to beneﬁt afﬂuent CSA members who live outside of the neighborhood,
but ﬁnd the center a convenient pickup location. At the same time, center resources are
supplementing the running of the CSA by providing space and staff time towards

coordinating share drop off. While association with the CSA has helped further the

 

7 In this case, I am using the term place-based movement to reference collective action by a group working
in the interests of a (virtual or proximal) community.

165

legitimacy of the organization as well as promoted center activities to those outside of the
neighborhood, some concern remains as to the actual beneﬁts of the CSA to the residents
whom the center serves.

The place—based orientation of the center is further apparent in the extent to which
community organizing and community involvement take place as part of the
organization. In particular in the past year, the center has been actively working to
involve community residents in the planning and implementation of food programs
through the formation of guidance committees related to the farmers’ market, CSA, and
other food related activities. Even prior to the formation of these committees, the center
held several public meetings with the community and conducted a neighborhood wide
survey so residents could provide feedback for the center’s programs. Again, while the
alternative agriculture movement involves a certain amount of grassroots activism, it is
largely oriented towards working to mainstream its values. The center, on the other hand,
looks to the people in the neighborhood for the values that guide its programs. The
alternative agriculture movement contains within it consideration for place as a value.
However, my point here is that place, and the speciﬁcities of place, play a central role to
the center and this distinguishes it as a place based movement which is closely associated
with, but not entirely encompassed by, the larger social movement.

The distinction between the alternative agriculture movement and the center as a
place based movement helps in furthering our understanding of the ability of place based
movements to inﬂuence social change at a broader level. Escobar (2001) has suggested
that the role of place based movements on inﬂuencing larger structures -- globalization in

particular -- has remained largely unacknowledged in the literature. He suggests that

166

these place based movements represent a powerful way through which people are able to
defend and construct their own sense of place and remain embedded in place.
Understanding how these place-based movements operate, in terms of their
collaborations and conﬂicts with other movements, will help us to understand their role in
social change.

Even from the results of this research it appears that the activities of the center
may be starting to have an impact on the residents of the neighborhood. Nearly six
percent of the sample indicated that they obtained food from the neighborhood farmers’
market. Furthermore, both food secure and insecure interviewees spoke very favorably
of the farmers’ market and felt that it was something that was easily accessible, even to
households on limited incomes. As one respondent stated “Oh I loved it the couple of
times I’ve been here, to mingle with the people in the neighborhood and the people that
bring their stuff in and see the fresh different things and just talk about stuff.”

In many ways, the success of the neighborhood center’s programs lies in more
than simply making healthy food more accessible to residents. The center has managed
to take the space (literally and metaphorically) that has been left by the growth of large
supermarkets at its fringes and turn it into something that is ﬁlled with people and
activities that are genuinely oriented towards community needs and not simply retailer
proﬁt. An empty parking lot becomes a farmers’ market and an unused park has been
turned into a greenhouse. Perhaps even more important, neighborhood residents are
ﬁnding a locus to collectively reconsider and alter the structure of the food system in

their neighborhood.

167

Limitations of this Research and Future Research

Food retailer selection is only one decision amidst a series that households make
regarding food consumption. Since it is the primary vector through which people interact
with the food system, stores represent an important site for understanding the relationship
between consumer interests and the structure of the food system. That being said, a
complete understanding of the differential agency among consumers requires additional
research into the other food related decisions that consumers make. For instance,
selection of sale items was not onlya factor that inﬂuenced decisions about what store to
go to, it also served to impact what items were selected. Item selection, and the reasons
why particular items are selected, remains an important topic that warrants further
explanation. Obesity is a problem, particularly among low-income populations, and item
selection due to economic constraints appears to be one potential cause (Darmon,
Ferguson, and Briend 2003).

Also unexamined in this study was the relative cost of items from various
retailers. Donkin et a1. (2000), for instance, examined price disparities among different
food retailers as a means to determine spatial inequality of food access. Data from this
research indicate that convenience stores represent an important food source, especially
for food insecure households. Since data on the prices of food was not collected,
obtaining food from convenience stores as opposed to traditional food retailers was not
assessed from an economic standpoint. It seems likely that there is an additional cost
associated with shopping at convenience stores, and the amount of this additional cost
would be another useful measure of how different consumers are impacted within the

food system.

168

Similarly, food selection among retailers was not assessed. Not all food is created
equal and foods that are traditionally considered healthy, such as fresh fruits and
vegetables, are not available in the same quantities, varieties, or qualities among the
various retailers in this study. Important questions remain regarding the inﬂuence of food
selection on the perception and behaviors of consumers. This would provide an
additional way of assessing the structure of the food system.

Finally, this research fails to take into account the inﬂuence of sources of food
other than retailers. For food insecure households, non-traditional food sources were
often a critical part of their food purchasing strategies. A truly complete assessment of
the community food system should incorporate food from food banks, churches, friends
and families, and other retailers such as dollar stores. In addition, food secure households
are obtaining food from restaurants or growing food in gardens so the role of these
sources in people’s food purchasing patterns should be assessed. Researchers are
beginning to examine other food sources, such as fast food restaurants, in order to assess

their distribution in relation to different populations (Block, Scribner, and DeSalvo 2004).

Recommendations

The results of this research point to some very real potential solutions to the
problems associated with food security in the neighborhood. While household income
was a signiﬁcant factor related to food insecurity, other factors such as education, single-
parent status, and physical disabilities were also contributing factors. Furthermore, the
ways in which neighborhood residents interact with the food system suggests that some

solutions, such as direct marketing arrangements with local producers, may have more

169

potential than recruiting more conventional food retailers into the region. The following

are six recommendations stemming from my experiences and the results of this study.

1.

2.

Create more and better employment opportunities for neighborhood residents.
Fundamentally, food insecurity remains a problem in the neighborhood because
many residents have very low incomes. In all logistic regression models
households with low incomes were much more likely to be food insecure than
households with higher incomes. In fact, one model indicated that households
with incomes of less than $10,000 were over 50 times more likely to be food
insecure than households with $80,000 or more. This suggests that efforts, such
as the recent increase in the minimum wage, may have a positive impact on food
insecure households. While providing even better employment opportunities in
the area would likely decrease food insecurity levels, this can be a time-
consuming and resource intensive endeavor. Furthermore, I was able to identify
several non-economic variables that were very strong predictors of food insecurity
status.

Develop linkages with area colleges 313d universities to increase educational
opportunities and complete college degrees. Results from the survey indicate that
people with only some college were nearly nine times more likely to be food
insecure than someone with a professional or graduate degree. At the same time,
having a four year degree led to a signiﬁcant decrease in the likelihood that a
household would be food insecure. Therefore, in order to decrease food

insecurity levels in the area it would be useful to assist people who have started

170

but not ﬁnished college. It would also help to provide greater access to post-
secondary education to those who have no college education.

. Target single-parents and households with physical disabilities. Interview
respondents described the difﬁculties of obtaining food when they or someone in
their household has a physical disability. Single—parents expressed the difﬁculties
of negotiating work, family, and school against shopping. Modelsshowed that
single-parents are nearly ﬁve times more likely to be food insecure than
households with two-parents and households without children. At the same time,
households with at least one person with a physical disability are nearly four
times more likely to be food insecure. In fact, single-parent status and physical
disabilities were even stronger predictors of food insecurity than poverty status.
This highlights the particular problems of two subpopulations, which would also
beneﬁt from programs that target them speciﬁcally.

. IncmLse healthy food Mbility to reduce tra_nsaction costs—Even when a food
retailer is within a few miles, this does not mean that there are no transaction costs
involved. Especially for households which are already relying on time intensive
food purchasing strategies such as shopping at several stores or limited by child or
adult care responsibilities, the cost of traveling even a small distance will be
greater than for others, and the most conveniently available food may not be the
healthiest. While actual food selection was not assessed in this study, spatial
assessment of convenience store locations and comments by interviewees makes
it clear that the selection of nutritious food varies signiﬁcantly among stores.

Bringing healthy food into the neighborhood makes it cheaper for the segments of

171

the population that need it most. Given the importance of cost to food insecure
populations, these new food sources should ideally be cheaper and more
convenient than existing food sources.

. Coordinate activities with the food patterns of food insecure households—It is
important that new food resources ﬁt into existing purchasing patterns, especially
for food insecure households. While many households are successfully avoiding
hunger through complicated food purchasing strategies, this may leave them with
little ﬂexibility in terms of adapting to new sources. This appeared to be the case
with the Community Supported Agriculture Farm, which operated out of the
neighborhood center for a few years. Food insecure households did not have the
ﬂexibility to adapt to the high upfront cost, tight pickup schedule, and inﬂexible
selection that was required for CSA membership. On the other hand, the farmers
market, in particular by going out of its way to accept food stamps, ﬁts more

closely with existing purchasing patterns.

. Enhance alternative food networks—This research suggests that several of the

food access problems confronted by households stem from the structure of the
conventional food system. It is possible that recruiting a deep discount grocer,
such as Sav-a-lot, would help decrease food insecurity. At the same time, many
larger supermarkets are resistant to moving into inner cities and may have
difﬁculty obtaining land even if willing to make the move. It may, therefore, be
useful to look to more novel forms of food retailing. While efforts such as
encouraging convenience stores to carry more fresh produce show some promise

and should be continued, alternative food systems also show potential in helping

172

provide nutritious food more equitably. Direct marketing arrangements with local
food producers or the enhancement of community food production both represent
potential ways in which the logic that drives larger food retailers might be
avoided. Direct marketing arrangements, if done correctly, also help with keeping
the money spent on food within the local community, which may enhance

employment opportunities in the future.

Final Words

A scholarly understanding of the role of consumers in the agrifood system would
be greatly enhanced through a differentiated understanding of consumers and this
research indicates real differences in the amount of agency that consumers have. In the
context of the conventional food system, some groups have greater levels of agency then
others. The food insecure/food secure distinction offers a poignant example of the way in
which these different levels of agency are manifest when two groups are differentially
served by a social system as inﬂuential as the conventional food system. This has very
real consequences in terms of health and nutrition and the only way these problems can
truly be addressed is by acknowledging the structures that generate the inequality. In an
even broader sense, this points to the ways in which capitalist systems, of which the food
system is deﬁnitely an example, create and maintain inequality.

On a pragmatic level, this research also points to the potential effectiveness of
systems, or environmentally based, solutions to food. While the degree to which
neighborhood residents were inﬂuenced by the distribution of food retailers varied, it

appeared that food insecure households were partially constrained by the distances to

173

various types of food retailers. At the very least, the varied distances represented higher
transaction costs for accessing certain types of foods. The negotiation of these costs can
be a ﬁnancially, physically, or psychologically difﬁcult for many disadvantaged
consumers to bear.

The efforts by the neighborhood center are encouraging and demonstrate that
space can be a solution as much as it is a problem. Hendrickson and Heffeman (2002)
argue that alternative food systems have the potential to ﬁll the interstitial spaces that
larger elements of the food system are unable to adapt to. After all, the reason there is no
super Wal-Mart in the neighborhood is because it would be too expensive to obtain
sufﬁcient land and the low purchasing power of the neighborhood would make such
efforts unproﬁtable. What keeps Wal-Mart out, though, creates a space for local farmers
to sell kale, cucumbers, and kohlrabi. The success of local initiatives such as the farmers’
market suggests that unconventional retail arrangements have potential.

The theoretical framework of this research has focused on agency vis-a-vis the
conventional food system. This framework has maintained explicit recognition of the
fact that a plethora of other social structures exist that inﬂuence the tastes and interest of
neighborhood residents. We must also recognize that food retailers are only one social
structure through which residents obtain food. Even food insecure households have
agency in a variety of other social contexts such as political and place-based. The
success of place-based food programs promoted by the Allen Neighborhood Center
indicate that these structures have the potential to alleviate some of the inequities
perpetuated by the conventional food system. In fact, the success of these programs

indicates that theses structure may even be more accessible than the conventional food

174

system to certain segments of the population. For instance, the farmers market is
spatially the closest location to obtain fresh produce for the parts of the neighborhood
with the lowest median incomes. For many people, it is within easy walking distance.
As new, and alternative, food networks are brought into the neighborhood it
seems very likely that some of the problems associated with food access can be
alleviated. However, most pe0ple in the neighborhood buy their food from food retailers
and, as this research indicates, the structure of food retailers places an undue burden on
those who are often the least able to bear the cost. Furthermore, the failure of many
households to bear this cost without compromising their health and/or turning to
alternative sources for food indicates that it is only by reconsidering the structure of this

system that greater widespread equity in food distribution will be accomplished.

175

A endix A: USDA uestions Used To Assess the Food Securi of Households

1. “We worried whether our food would run out before we got money to buy more.” Was
that often, sometimes, or never true for you in the last 12 months?

2. “The food that we bought just didn’t last and we didn’t have money to get more.” Was
that often, sometimes, or never true for you in the last 12 months?

3. “We couldn’t afford to eat balanced meals.” Was that often, sometimes, or never true
for you in the last 12 months?

4. In the last 12 months, did you or other adults in the household ever cut the size of your
meals or skip meals because there wasn’t enough money for food? (Yes/No)

5. (If yes to Question 4) How often did this happen—almost every month, some months
but not every month, or in only 1 or 2 months?

6. In the last 12 months, did you ever eat less than you felt you should because there
wasn’t enough money for food? (Yes/No)

7. In the last 12 months, were you ever hungry, but didn’t eat, because you couldn’t
afford enough food? (Yes/No)

8. In the last 12 months, did you lose weight because you didn’t have enough money for
food? (Yes/No) '

9. In the last 12 months, did you or other adults in your household ever not eat for a
whole day because there wasn’t enough money for food? (Yes/No)

10. (If yes to Question 9) How often did this happen—almost every month, some months
but not every month, or in only 1 or 2 months?

(Questions 11—18 are asked only if the household included children age 0-18)
11. “We relied on only a few kinds of low-cost food to feed our children because we were
running out of money to buy food.” Was that often, sometimes, or never true for you in

the last 12 months?

12. “We couldn’t feed our children a balanced meal, because we couldn’t afford that.” I
Was that often, sometimes, or never true for you in the last 12 months?

13. “The children were not eating enough because we just couldn’t afford enough food.”
Was that often, sometimes, or never true for you in the last 12 months?

14. In the last 12 months, did you ever cut the size of any of the children’s meals because
there wasn’t enough money for food? (Yes/No)

176

15. In the last 12 months, were the children ever hungry but you just couldn’t afford more
food? (Yes/No)

16. In the last 12 months, did any of the children ever skip a meal because there wasn’t
enough money for food? (Yes/No)

17. (If yes to Question 16) How often did this happen—almost every month, some
months but not every month, or in only 1 or 2 months?

18. In the last 12 months, did any of the children ever not eat for a whole day because
there wasn’t enough money for food? (Yes/No)

177

Appendix B: Themes and sub-themes applied to interviews

 

Theme

Sub-theme

 

Food sources

Food sources

 

Factors influencing store selection

 

Influence of cost on source selection

 

Satisfaction with food from current sources

 

Food procurement

Factors influenMopping trip timing

 

Strategies and planning involved in shopping

 

Food consumption
values

Cost concerns

 

Taste concerns

 

Health concerns

 

Environmental concerns

 

Agricultural concerns

Social group concerns

 

Other concerns

 

Meal patterns

 

Vgcgtable consumption

 

Attitude toward organic

 

Misc.

How food fits into the household budget

 

Coﬂgﬂrategies for food insecurity

 

Food acquisition
strategy

 

 

What drives their food patterns?

 

178

 

BIBLIOGRAPHY

Agnew, J .A. 1989. "Devaluation of place in social science." Pp. 10—29 in The Power of
Place: Bringing Together Geographical and Sociological Imaginations, edited by
J. A. Agnew and J. S. Duncan. Boston: Unwin Hyman.

Allen, Patricia and Martin Kovach. 2000. "The Capitalist Composition of Organic: The
Potential of Markets in Fulﬁlling the Promise of Organic Agriculture."
Agriculture and Human Values 17:221-232.

Alwitt, Linda F. and Thomas D. Donley. 1997. "Retail Stores in Poor Urban
Neighborhoods." The Journal of Consumer Affairs 31: 139-164.

Anderson, Molly D. and John T. Cook. 1999. "Community food security: Practice in
need of theory?" Agriculture and Human Values 16:141—150.

Appadurai, Arjun. 1986. "Introduction: commodities and the politics of value." Pp. 3-63
in The Social Life of Things, edited by A. Appadurai. New York: Cambridge
University Press.

Aquilino, W.S. 1994. "Interview mode effects in surveys of drug and alcohol use: A ﬁeld
experiment." Public Opinion Quarterly 58:210-240.

Armstrong, J .S. and E.J. Luske. 1987. "Return Postage in Mail Surveys: A Meta-
Analysis." Public Opinion Quarterly 58:233-248.

Basu, Alaka Malwade. 1992. Culture, the Status of Women, and Demographic Behavior.
Oxford, England: Clarendon Press.

Beardsworth, Alan and Teresa Keil. 1997. Sociology on the Menu: An Invitation to the
Study of Food and Society. London: Routledge.

Beck, Ulrich. 1992. Risk Society: Towards a New Modernity. Thousand Oaks, CA:
SAGE Publications.

Belasco, Warren. 1993. Appetite for Change (updated version). Ithaca,
NY: Cornell University Press.

Bickel, Gary, Mark Nord, Christopher Price, William L. Hamilton, and John T. Cook.
2000. "Guide to Measuring Household Food Security." US. Department of
Agriculture, Food and Nutrition Service, Alexandria, VA.

Blanchard, Troy and Thomas Lyson. 2002. "Access to Low Cost Groceries in
Nonmetropolitan Counties: Large Retailers and the Creation of Food Deserts." in
Measuring Rural Diversity Conference Proceedings. Washington, DC: Economic
Research Service, US Department of Agriculture.

179

Blisard, Noel, Hayden Stewart, and Dean Jolliffe. 2004. "Low-Income Households'
Expenditures on Fruits and Vegetables." USDA Economic Research Service.

Block, Jason P., Richard A. Scribner, and Karen B. DeSalvo. 2004. "Fast Food,
Race/Ethnicity, and Income: A Geographic Analysis." American Journal of
Preventative Medicine 27:211—217.

Blumberg, Stephen J ., Karil Bialostosky, William L. Hamilton, and Ronette R. Briefel.
1999. "The Effectiveness of a Short Form of the Household Food Security Scale."

American Journal of Public Health 89: 1231-1234.

Bourdieu, P. 1984. Distinction: A Social Critique of the Judgment of Taste. Translated by
R. Nice. Cambridge, MA: Harvard University Press.

Bromley, Rosemary and Colin Thomas. 1993. "The Retail Revolution, the Carless
Shopper and Disadvantaged Consumer." Transactions of the Institute of British
Geographers, New Series 18:222-236.

Busch, Lawrence. 1997. "Grades and Standards and the Social Construction of Safe
Food." The Social Construction of Safe Food, Trondheim, Norway.

Campbell, Colin. 2000. "The Romantic Ethic and the Spirit of Modern Consumerism."
Pp. 48-72 in The Consumer Society Reader, edited by M. J. Lee. Oxford:
Blackwell Publishing.

Caraher, M., P. Dixon, and R. Car-Hill. 1998. "Access to health foods: Part I. Barriers to
accessing healthy foods: Differentiations by gender, social class, income and
mode of transport." Health Education Journal: 191-201.

Clarke, Graham, Heather Eyre, and Cliff Guy. 2002. "Deriving Indicators of Access to
Food Retail Provisions in British Cities: Studies of Cardiff, Leeds and Bradfor ."
Urban Studies 39:2041-2060.

Cohen, Barbara. 2002. "Community Food Security Assessment Toolkit." Economic
Research Service, US Department of Agriculture, Washington, DC.

Cotter, David A. 2002. "Poor People in Poor Places: Local Opportunity Structures and
Household Poverty." Rural Sociology 67:534-555.

Crewe, Louise. 2001. "The Besieged Body: Geographies of Retailing and Consumption."
Progress in Human Geography 25:629-640.

Cummins, Steven and Sally Macintyre. 1999. "The location of food stores in urban areas:
a case study of Glasgow." British Food Journal. 101:545-553.

180

Cummins, Steven and Sally Macintyre. 2002a. ""Food deserts"--evidence and assumption
in health policy making." British Medical Journal 325:436-438.

Cummins, Ronnie and Sally Macintyre. 2002b. "A Systematic Study of an Urban
Foodscape: The Price and Availability of Food in Greater Glasgow." Urban
Studies 39:2115-2130.

Dannon, Nicole, Elaine Ferguson, and Andre Briend. 2003. "Do economic constraints
encourage the selection of energy dense diets?" Appetite 41:315-322.

DeVault, Marjorie and James P. Pitts. 1984. "Surplus and Scarcity: Hunger and the
Origins of the Food Stamp Program." Social Problems 31:545-557.

Diamond, Jared. 1997. Guns, Germs, and Steel: The Fates of Human Societies. New
York: W.W. Norton and Company.

Dillman, Don A. 2000. Mail and Internet Surveys: The Tailored Design Method. New
York: John Wiley & Sons, Inc.

Dixon, Jane. 1999. "A cultural economy model for studying food systems." Agriculture
and Human Values 16:151-160.

Donkin, A.J.M., Elizabeth A. Dowler, Simon J. Stevenson, and Sheila Turner. 2000.
"Mapping access to food in a deprived area: The development of price and
availability indices." Public Health Nutrition 3:33-38.

Donkin, Angela J .M., Elizabeth A. Dowler, Simon J. Stevenson, and Sheila Turner. 1999.
"Mapping access to food at a local level." British food journal. 101.

Ehrenhalt, Alan. 2006. "The Grocery Gap." Governing 19:9-10.

Ehrlich, Paul R., Anne H. Ehrilich, and Gretchen Daily. 1993. "Food security, population,
and environment." Population and Development Review 19:1-32.

Escobar, A. 2001. "Culture sits in places: reﬂections on globalism and subaltem
strategies of localization." Political Geography 20: 139-174.

2007. Federal Register 72:15, 3147—3148
Fine, Ben. 2002. The world of consumption. London: Routledge.

Fine, Ben, Michael Heasman, and Judith Wright. 1996. Consumption in the age of
afﬂuence: the world of food. New York: Routledge.

Fischer, CS. 1975. "The study of urban community and personality.” Annual Review of
Sociology 1:67-89.

181

French, Simone A., Mary Story, and Robert Jeffery. 2001. "Environmental Inﬂuences on
Eating and Physical Activity." Annual Review of Public Health 22:309-335.

Friedland, W.H. 1984. "Commodity systems analysis: an approach to the sociology of
agriculture." Research in Rural Sociology and Agriculture 1:221-236.

Frongillo Jr., Edward A. 1999. "Validation of Measures of Food Insecurity and Hunger. "
Journal of Nutritionz5068-5098. ‘

Furey, Sinead, Heather McIlveen, and Christopher Strugnell. 1999. "Food Deserts: An
Issue of Social J ustice." M/C: A Journal of Media and Culture 2.

Gans, Herbert. 2002. "The Sociology of Space: A Use-Centered View." City &
Community 1:329-339. '

Gieryn, Thomas F. 2000. "A Space for Place in Sociology." Annual Review of
Sociologyz463-496.

—. 2002. "Give Place a Chance: Reply to Gans." City & Community 1:341-343.

Goodman, David and E. Melanie DuPuis. 2002. "Knowing Food and Growing Food:
Beyond the Production-Consumption Debate in the Sociology of Agriculture."
Sociologia Ruralis 42:5-22.

Greder, Kimberly and Mary Jane Brotherson. 2002. "Food security and low-income
families: Research to inform policy and programs." Journal of Family and
Consumer Sciences 94:40-47.

Green, Judith M., Alizon K. Draper, and Elizabeth A. Dowler. 2003. "Short Cuts to
Safety: Risk and 'Rules of Thumb' in Accounts of Food Choice." Health, Risk &
Society 5:33-52.

Guthman, Julie. 2002. "Commodiﬁed Meanings, Meaningful Commodities: Re-thinking
Production-Consumption Links through the Organic System of Provision."
Sociologia Ruralis 42:296-31 1. .

Guy, Clifford. 1985. "The Food and Grocery Shopping Behavior of Disadvantaged
Consumers: Some Results from the Cardiff Consumer Panel." Transactions of the

Institute of British Geographers, New Series 10:181-190.

Hall, Bruce. 1983. "Neighborhood Differences in Retail Food Stores: Income Versus
Race and Age of Population." Economic Geography 59:282-295.

182

Hamelin, Anne-Marie, Jean-Pierre Habicht, and Micheline Beaudry. 1999. "Food
Insecurity: Consequences for the Household and Broader Social Implications."
Journal of Nutrition 129:5258-5288.

Halford, Jason, Jane Gillespie, Victoria Brown, Eleanor Pontin, and Terence Dovey.

2004. "Effect of television advertisement for foods on food consumption in
children." Appetite 42:221-225.

Harvey, David. 1989. The condition of postmodemity: an enquiry into the origins of
cultural change. New York: Blackwell.

Hays, Sharon. 1994. "Structure and Agency and the Sticky Problem of Culture."
Sociological Theory 12:57-72.

Hendrickson, Mary K. and William D. Heffeman. 2002. "Opening Spaces through
Relocalization: Locating Potential Resistance in the Weaknesses of the Global
Food System." Sociologia Ruralis 42:347-369.

Iannoaccacchione, Vincent, Staab Jennifer, and David Redden. 2003. "Evaluating the Use
of Residential Mailing Addresses in a Metropolitan Household Survey." Public
Opinion Quarterly 67:202-210.

Jabs, Jennifer and Carol M. Devine. 2006. "Time scarcity and food choices: An
overview." Appetite 47:196-204.

J enness, J. 2004a. Distance and bearing between matched features (distbyid.avx)
extension for ArcView 3.x, v. 2. Jenness Enterprises. Available at:
http://www.jennessentcom/gview/distance bv id.htm.

Jenness, J. 2004b. Nearest features (nearfeatavx) extension for ArcView 3.x, v. 3.8a.
Jenness Enterprises. Available at:
http://www.jennessenLoom/grcview/nearest featureshtm.

Kaufman, Phil. 1995. "Fewer But Larger Supermarkets." Food Review 18:26-30.

Kaufman, Phillip R., James M. MacDonald, Steve M. Lutz, and David M. Smallwood.
1997. "Do the Poor Pay More for Food? Item Selection and Price Difference
Affect Low-Income Household Food Costs." Food and Rural Economics
Division, Economic Research Service, US Department of Agriculture,
Washington, DC.

Keenan, Debra Palmer, Christine Olson, James C. Hersey, and Sondra M. Parmer. 2001.
"Measures of food insecurity/security." Journal of Nutrition Education and

Behavior 33:849-858.

Kempson, Kathryn, Debra Palmer Keenan, Puneeta Sonya Sadani, and Audrey Adler.
2003. "Maintaining Food Sufﬁciency: Coping Strategies Identiﬁed by Limited-

183

Resource Individuals versus Nutrition Educators." Journal of Nutrition Education
and Behavior 35:179.

Koc, Mustafa and Kenneth Dahlberg. 1999. "The restructuring of food systems: Trends,
research, and policy issues." Agriculture and Human Values 16:109-116.

Korczynski, Marek and Ursula Ott. 2006. "The Menu in Society: Mediating Structures of
Power and Enchanting Myths of Individual Sovereignty." Sociology 40:911-928.

Lang, T. and M. Caraher. 1998. "Access to healthy foods: Part 11. Food poverty and
shopping deserts: What are the implications for health promotion policy and
practice?" Health Education Joumal:202-211.

Levenstein, Harvey. 1988. Revolution at the table: the transformation of the American
Diet. Oxford: Oxford University Press.

Linsky, Arnold S. 1975. "Stimulating Responses to Mailed Questionnaires: A Review."
Public Opinion Quarterly 39:82-10]. '

Lobao, Linda and Saenz Rogelio. 2002. "Spatial Inequality and Diversity as an Emerging
Research Area." Rural Sociology 67:497-511.

Lockie, Stewart. 2000. "'The Invisible Mouth': Mobilizing 'the Consumer' in Food
Production-Consumption Networks." Sociologia Ruralis 42:278-294.

Lockie, Stewart and S. Kitto. 2000. "Beyond the farm gate: production-consumption
networks and agri-food research." Sociologia Ruralis 40:3-19.

Marsden, T., A. Flynn, and M. Harrison. 1997. "Retailing, regulation and food
consumption: The public interest in a privatized world?" Agribusiness 13:211-
226.

Marsden, Terry, T. Flyn, and M. Harrison. 2000. Consuming interests: the social
provision of foods. London: UCL Press.

Miller, D. 1995. "Consumption as the Vanguard of History." in Acknowledging
Consumption. New York: Routledge. '

—. 2000. "Object Domains, Ideology and Interests." Pp. 106-124 in The Consumer
Society Reader, edited by M. J. Lee. Oxford: Blackwell Publishing.

McIntyre, Lynn, N. Theresa Glanville, Kim D. Raine, Jutta B. Dayle, Bonnie Anderson,
and Noreen Battaglia. 2003. "Do low—income lone mothers compromise their
nutrition to feed their children?" Canadian Medical Association Journal 168:686-
691.

184

Morland, Kimberly, Steve Wing, Ana Diez Roux, and Charles Poole. 2002.
"Neighborhood Characteristics Associated with the Location of Food Stores and
Food Service Places." American Journal of Preventative Medicine 22:23-29.

Morland, Kimberly, Steve Wing, and Ana Diez Roux. 2002. "The Contextual Effect of
the Local Food Environment on Residents' Diets: The Atherosclerosis Risk in
Communities Study." American Journal of Public Health 92: 1761-1767.

Nestle, Marion. 2002. Food politics: how the food industry inﬂuences nutrition and
health. Berkeley: University of California Press.

Newman, Isadore, Carolyn S. Ridenour, Carole Newman, and George DeMarco. 2003.
"A Typology of Research Purposes and Its Relationship to Mixed Methods." Pp.
167-188 in Handbook of Mixed Methods in Social and Behavioral Research,
edited by A. Tashakkori and C. Teddle. Thousand Oaks: SAGE Publications.

Nord, Mark, Margaret Andrews, and Steven Carlson. 2006. "Household Food Security in
the United States, 2005." Food and Rural Economics Division, Economic
Research Service, US Department of Agriculture, Washington DC.

N ygard, Berit and Oddveig Storstad. 1998. "De-globalization of Food Markets?
Consumer Perceptions of Safe Food: The Case of Norway." Sociologia Ruralis
38:35-53.

Olson, Christine M. 1999. "Nutrition and Health Outcomes Associated with Food
Insecurity and Hunger. " Journal of Nutrition 129:521S-524S.

Olson, Christine M., Barbara S. Rauschenbach, Edward A. Frongillo Jr., and Anne
Kendall. 1997. "Factors contributing to household food insecurity in a rural

upstate New York Community." Family Economics and Nutrition Review 10:2-
17.

Pelletier, David L., Vivica Kraak, Christine McCullum, Ulla Uusitalo, and Robert Rich.
1999. "Community food security: Salience and participation at community level."
Agriculture and Human Values 16:401—419.

Peterson, Richard A. and Roger M. Kern. 1996. "Changing Highbrow Taste: From Snob
to Omnivore." American Sociological Review 61:900-907.

Pichardo, Nelson. 1997. "New Social Movements: A Critical Review." Annual Review of
Sociology 23:411-430.

Poppendieck, Janet. 2000. "Want Amid Plenty: From Hunger to Inequality." Pp. 189-202
in Hungry for Proﬁt: The Agribusiness Threat to Farmers, Food, and the
Environment, edited by F. Magdoff, J. B. Foster, and F. H. Buttel. New York:
Monthly Review Press.

185

Pothukuchi, Kami, Joseph Hugh, Hannah Burton, and Andy Fisher. 2002. What's
Cooking in Your Food System? A Guide to Community Food Assessment.
Venice, CA: Community Food Security Coalition.

Putnam, Judy, Jane Allshouse, and Linda Scott Kantor. 2002. "US. Per Capita Food
Supply Trends: More Calories, Reﬁned Carbohydrates, and Fats." Food Review
25:2-15.

Rappoport, Leon. 2003. How We Eat: Appetite, Culture, and the Psychology of Food.
Toronto: ECW Press.

Raynolds, Laura T. 2000. "Re-embedding global agriculture: The international organic
and fair trade movements." Agriculture and Human Values 17:297-309.

Renard, MC. 2003. "Fair trade: quality, market, and conventions." Journal of Rural
Studies 19:87-96.

Ritzer, G. 1996. The McDonaldization of Society. Thousand Oaks, CA: Pine Forge Press.

Rochon, Thomas R. 1998. Culture Moves. Princeton: Princeton University Press.

Roos, Eva, Eero Lahelma, Virtanen Mikko, Prattala Ritva, and Pirjo Pietinen. 1998.

"Gender, Socioeconomic Status and Family Status as Determinants of Food
Behaviour." Social Science and Medicine 46: 1519-1529.

Rose, David. 1999. "Economic Determinants and Dietary Consequences of Food
Insecurity in the United States." Journal of Nutrition 129:5178-520S.

Rose, Donald, Craig Gunderson, and Victor Oliveira. 1998. "Socio-Economic
Determinants of Food Insecurity in the United States: Evidence from the SIPP and
CSFII Datasets." Food and Rural Economics Division, Economic Research
Service USDA, Washington DC.

Sapp, Stephen G. and Sharon R. Bird. 2003. "The Effects of Social Trust on Consumer
Perceptions of Food Safety." Social Behavior and Personality 31:413-421.

Senauer, Ben, Elaine Asp, and Jean Kinsey. 1991. Food trends and the changing
consumer. St. Paul: Eagan Press.

Sewell, William H. 1992. "A Theory of Structure: Duality, Agency, and Transformation."
The American Journal of Sociology 98: 1-29.

Sobal, Jeffery and Brian Wansink. 2007. "Kitchenscapes, Tablescapes, Platescapes, and

Foodscapes: Inﬂuences of Microscale Built Environments on Food Intakes."
Environment and Behavior 39:124-142.

186

Smil, Vaclav. 1994. "How many people can the earth feed?" Population and
Development Review 20:255-292.

Stokols, Daniel. 1992. "Establishing and Maintaining Healthy Environments: Toward a
Social Ecology of Health Promotion." American Psychologist 47:6-22.

Swartz, David. 1997. Culture and Power: The Sociology of Pierre Bourdieu. Chicago:
The University of Chicago Press.

Swyngedouw, E. 1997. "Neither global nor local: 'Glocalization' and the politics of
scale." Pp. 137-166 in Spaces of Globalization, edited by K. Cox. New York:
Guilford Press.

Thomas, Brian. 2004. "The relationship among household characteristics, geographic
space, and food security in the Allen Neighborhood." Allen Neighborhood
Center, Lansing, MI.

Tickamyer, Ann R. 2000. "Space Matters! Spatial Inequality in Future Sociology."
Contemporary Sociology 29:805-813.

Tillotson, James E. 2005. "Wal-Mart and Our Food." Nutrition Today 40:234-237.

Travers, Kim D. 1996. "The Social Organization of Nutritional Inequities." Social
Science and Medicine 43:543-553.

Veblen, T. 1959. The Theory of the Leisure Class. New York: Viking.

Vozoris, Nicholas T. and Valerie S. Tarasuk. 2003. "Household Food Insufﬁciency Is
Associated with Poorer Health." Journal of Nutrition: 120-126.

Warde, Alan. 1997. Consumption, Food, and Taste. Thousand Oaks, CA: SAGE
Publications, Inc.

Warde, Alan, Lydia Martens, and Wendy Olsen. 1999. "Consumption and the problem of
variety: cultural omnivorousness, social distinction and dining out." Sociology
33: 105-127.

Wansink, Brian and Jeffery Sobal. 2007. "Mindless Eating: The 200 Daily Food
Decisions We Overlook." Environment and Behavior 39:106-123.

Whelan, Amanda, Neil Wrigley, Daniel Warm, and Elizabeth Cannings. 2002. "Life in a
'Food Desert'." Urban Studies 39:2083-2100.

187

Wolfe, Wendy S., Edward A. Frongillo, and Pascale Valois. 2003. "Understanding the
experience of food insecurity by elders suggests ways to improve its
measurement." Journal of Nutrition 133:2762-2769.

Wright, Len Tiu, Clive Nancarrow, and Pamela M.H. Kwok. 2001. "Food taste
preferences and cultural inﬂuences on consumption." British Food Joumalz348-
357.

Wrigley, Neil. 2001. "The consolidation wave in US food retailing: A European
perspective." Agribusiness 17:489.

Wrigley, Neil, Daniel Warm, Barrie Margetts, and Amanda Whelan. 2002. "Assessing
the Impact of Improved Retail Access on Diet in a 'Food Desert': A Preliminary
Report." Urban Studies 39:2061-2082.

Yammarino, Francis J ., Steven J. Skinner, and Terry L. Childers. 1991. "Understanding
Mail Survey Response Behavior: A Meta-Analysis." Public Opinion Quarterly
55.

Zimmerman, M.M. 1941. "The Supermarket and the Changing Retail Structure." Journal
of Marketing 5:402—409.

Zukin, Sharon. 2002. "What’s Space Got to Do with It?" City & Community 1:345-348.

Zukin, Sharon and Jennifer Smith Maguire. 2004. "Consumers and Consumption."
Annual Review of Sociology 30: 173-197.

188

 

IIIIIIIIIIIIIIIII