Food Assistance and Nutrition Research Small Grants Program

United StatesDepartment ofAgriculture

EconomicResearchService

Food Assistanceand NutritionResearch ReportNumber 38

Food Assistance andFood Assistance andNutrition ResearchNutrition ResearchSmall Grants ProgramSmall Grants ProgramExecutive SummariesExecutive Summariesof 2002 Research Grantsof 2002 Research Grants

Food Assistance & Nutrition Research Program

Laura Tiehen

Food Assistance and Nutrition Research Small Grants Program: ExecutiveSummaries of 2002 Research Grants. By Laura Tiehen (editor). Food and RuralEconomics Division, Economic Research Service, U.S. Department of Agriculture,Food Assistance and Nutrition Research Report No. 38.

Abstract

This report summarizes research findings from the Food Assistance and NutritionResearch Small Grants Program. The Economic Research Service created the pro-gram in 1998 to stimulate new and innovative research on food assistance andnutrition issues and to broaden the participation of social science scholars in theseissues. The report includes summaries of the research projects that were awarded1-year grants in summer and fall 2001. The results of these research projects werepresented at the October 2002 Small Grants Program conference. The projectsfocus on food insecurity and hunger, nutritional status and diet quality, Federalfood assistance program participation, and the role of private-sector organizationsin the provision of food assistance. Some projects focus on specific populations,such as people living in the rural South and those living on American Indianreservations.

Keywords: Food assistance, nutrition, vulnerable populations, food security, foodinsecurity, hunger, hungry, food assistance, food spending, well-being, Food StampProgram, food stamps, National School Lunch Program, WIC, Food Assistanceand Nutrition Research Program

1800 M Street, NWWashington, DC 20036-5831 November 2003

ii Food Assistance and Nutrition Research Small Grants Program / FANRR-38 Economic Research Service/USDA

Contents

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .iv

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1

Executive SummariesFood Insecurity and Hunger

Does Household Food Insecurity Affect Cognitive and SocialDevelopment of Kindergartners? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4

Ame Stormer and Gail G. Harrison

Hunger, Food Insecurity, and Child Obesity . . . . . . . . . . . . . . . . . . . . . . .5Marilyn Townsend and Hugo Melgar-Quiñonez

A Study of Older Adults on the Waiting List for Home-DeliveredMeals in North Carolina . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6

Mary Anne P. Salmon and Jessalyn Gooden

Impact of Home-Delivered Meals on the Nutritional Statusand Food Security of the Elderly in New York State . . . . . . . . . . . . . . . .7

Edward A. Frongillo and Wendy S. Wolfe

The Role of the Private Sector in Providing Food AssistancePrivate Food Assistance in the Deep South: Agency Profiles andDirectors’ Perceptions of Needs and Opportunities UnderCharitable Choice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9

Suzie T. Cashwell, John P. Bartkowski, Patricia A. Duffy,Joseph J. Molnar, Vanessa Casanova, and Marina Irimia-Vladu

Making Ends Meet: An Examination of TANF and Former TANF Food Pantry Users in Virginia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10

Ann Nichols-Casebolt

The Arkansas Child and Adult Care Food Program: A Study of Factors Associated With Program Participation . . . . . . . . . . . . . . . . .11

Richard A. Huddleston and Yetunde A. Shobo

Dynamics and Life Course Patterns of Food AssistanceProgram Participation

Estimating the Probabilities and Patterns of Food StampUse Across the Life Course . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13

Mark R. Rank and Thomas A. Hirschl

Food Stamp Program Participation Dynamics in U.S. Counties and States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14

Stephan J. Goetz, Anil Rupasingha, and Julie N. Zimmerman

The Dynamics of Prenatal WIC Participation . . . . . . . . . . . . . . . . . . . . .15Christopher A. Swann

Economic Research Service/USDA Food Assistance and Nutrition Research Small Grants Program / FANRR-38 iii

Nutritional Status and Diet QualityDevelopment of a Diet Quality Index for Preschool Childrenand Its Application in Examining Dietary Trends in theUnited States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16

Anna Maria Siega-Riz and Sibylle Kranz

Heat or Eat? Cold Weather Shocks and Nutrition in PoorAmerican Families . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17

Jayanta Bhattacharya, Thomas DeLeire, Steven Haider,and Janet Currie

Nutrition Assessment and Education for Keweenaw Bay Ojibwa . . . . .18Deborah Parrish

Food Security and Food Assistance Programs in the Community Context

Community Attitudes Toward Traditional Tohono O’odham Foods . . . .19Daniel Lopez, Tristan Reader, and Paul Buseck

Contextual Determinants of Food Security in Southern Hispanicand African-American Neighborhoods . . . . . . . . . . . . . . . . . . . . . . . . . .20

Margaret E. Bentley, Sonya Jones, and Janice Dodds

Welfare Reform and Food Assistance ProgramsAccess to Food Assistance Programs Among NorthernCheyenne Families . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .21

Judith Davis, Rita Hiwalker, Carol Ward and Eric Dahlin

Food Stamp Program Participation of Refugees and Immigrants:Measurement Error Correction for Immigrant Status . . . . . . . . . . . . . . .23

Chris Bollinger and Paul Hagstrom

Food Stamp Receipt by Families With Noncitizen HouseholdHeads in Rural Texas Counties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .24

Steve White, Xiuhong You, Steve H. Murdock, and Tami Swenson

iv Food Assistance and Nutrition Research Small Grants Program / FANRR-38 Economic Research Service/USDA

Preface

The Economic Research Service’s (ERS) Food Assistance and Nutrition ResearchProgram offers grants to social science scholars to stimulate new and innovativeresearch on food and nutrition assistance and to broaden partication in research onthese issues. To administer the program, ERS partners with five academic institu-tions and research institutes that competitively award grants for 1-year researchprojects. Most grants are for $20,000 to $40,000. The Small Grants Program seeksto give junior scholars an opportunity to gain experience in conducting research onfood security and food and nutrition assistance programs and to encourage moresenior scholars to apply their skills and knowledge in these areas as well.

This report presents summaries of the research findings from the fourth set ofsmall grants, which were awarded in summer and fall 2001. Preliminary findingswere presented at a conference at ERS in Washington, DC, on October 17 and 18,2002, and the research projects were completed in December 2002. More informa-tion about the Small Grants Program partners, as well as many of the completedresearch papers themselves, is on the websites of the administering institutionslisted below:

Institute for Research on Poverty, University of Wisconsin-Madison

Focus: The effects of food assistance programs on food security, income security,and other indicators of well-being among low-income individuals and families.Web address: www.ssc.wisc.edu/irp/smgrants/smhome.htm

The Joint Center for Poverty Research, University of Chicago andNorthwestern University

Focus: Interactions between food assistance programs and other welfare programs,and the effects of the macroeconomy on the need for food assistance, the level ofparticipation, and costs of food assistance programs.Web address: www.jcpr.org/usdarfp.html

The American Indian Studies Program, University of Arizona

Focus: The relationship between food assistance programs on reservations andfamily poverty.Web address: info-center.ccit.arizona.edu/~aisp

The Department of Nutrition at the University of California, Davis

Focus: The impact of food assistance programs on nutritional risk indicators(anthropometric, biochemical, clinical, and dietary), food purchasing practices, andfood insecurity. Web address: nutrition.ucdavis.edu/usdaers.html

Southern Rural Development Center, Mississippi State University

Focus: Food assistance research issues for rural people, families, and communitiesin the South.Web address: srdc.msstate.edu/focusareas/health/fa/food.htm

Economic Research Service/USDA Food Assistance and Nutrition Research Small Grants Program / FANRR-38 1

Introduction

Federal food and nutrition assistance programs form acrucial component of the social safety net in theUnited States. Unlike many other social programs,food assistance programs provide benefits and haveeligibility requirements that are uniform nationwide.The Food Stamp Program (FSP)—the largest Federalfood assistance program—is, with few exceptions,available to all Americans whose income and assetsfall below certain levels. Although the other foodassistance programs are generally targeted to specificdemographic groups, together the 15 Federal foodassistance programs reach an estimated 1 in 5Americans at some point each year. The U.S.Department of Agriculture (USDA), the Federalagency charged with administering almost all of theFederal food and nutrition assistance programs, has aparticular interest in monitoring program effectivenessand contributing to the policy goal of a healthy, well-nourished population.

The purpose of the Small Grants Program is to stimu-late new research on food and nutrition policy issuesand to broaden the participation of social sciencescholars in the research effort. Grant recipients comefrom a number of disciplines and employ a variety ofapproaches in their research. They include economists,sociologists, nutritionists, anthropologists, and publichealth professionals. Some conduct exploratoryresearch using ethnographic methods to examineunderlying factors influencing program participationand outcomes. Others use descriptive statistics to char-acterize the populations of interest. Still others usestatistical models to analyze individuals’ responses topolicy changes. All the methods employed contribute

to a growing body of knowledge of the food needs,coping behaviors, and food program outcomes of low-income families and individuals.

Small Grants Program Partners

ERS created partnerships with five academic institu-tions and research institutes to administer the SmallGrants Program. Partner institutions have the advan-tage of being prominent members of the researchcommunity and being closer to the particular regionaland State environments that influence programdelivery and outcomes. ERS chose two of the fivepartner institutions for their experience in conductingpolicy-relevant poverty research at the national leveland their ability to attract prominent scholars from avariety of social science disciplines to work on povertyand hunger issues. One of these is the Institute forResearch on Poverty (IRP) at the University ofWisconsin-Madison. IRP has a distinguished history ofresearch and policy evaluation, including previousinvolvement in administering small research grantsfunded by USDA’s Food and Nutrition Service. Thesecond partner is the Joint Center for Poverty Research(JCPR) at the University of Chicago and NorthwesternUniversity. JCPR was established in 1996 with a grantfrom the U.S. Department of Health and HumanServices (HHS) to conduct and fund research and toadvise Federal policymakers on issues of poverty. Itssmall grants program with HHS and the CensusBureau served as a model for the ERS Small GrantsProgram.

ERS chose the remaining three of the five partnerinstitutions for their ability to direct research of policyinterest to USDA, either on a particular subset of foodassistance and nutrition issues or on a particular

Food Assistance and Nutrition Research

Small Grants Program

Executive Summaries of 2002 Research Grants

Laura Tiehen, Editor

2 Food Assistance and Nutrition Research Small Grants Program / FANRR-38 Economic Research Service/USDA

subpopulation of those eligible for food and nutritionassistance. Among these, the Department of Nutritionof the University of California at Davis brought to theSmall Grants Program its expertise in nutrition educa-tion design and evaluation. A core group of facultyfocuses its research efforts on identifying meaningfulapproaches to the design and evaluation of nutritioneducation for ethnically diverse, low-income familiesserved by a variety of food assistance programs. Thegroup views multidisciplinary research as critical toeffectively monitoring the outcomes of nutritionprograms.

The Southern Rural Development Center (SRDC) waschosen to administer small grants for its ability andcommitment to conduct research on the problems ofthe rural poor in the South and its particular commit-ment to study the effects of welfare reform on thispopulation. USDA has special ties to the SRDCbecause of the land-grant status of its member institu-tions. The South is also of particular interest to USDAbecause of the large proportion of rural poor and ruralAfrican-Americans who reside in the region.

American Indian families living on reservations are asignificant component of the low-income rural popula-tion in many of the Western and Plains States. ERSchose the University of Arizona’s American IndianStudies Program (AISP) to administer small grants forresearch on the food assistance and nutrition needs andproblems of American Indians. AISP is the home ofthe only doctoral program in American Indian Studiesin the country. The program maintains close ties to thetribal colleges, which were given land-grant status byCongress in 1994.

Research Overview

The research projects completed in 2002 cover sixbroad topic areas.

1. Food Insecurity and Hunger. The recent develop-ment of a Federal measure of food security makes itpossible to monitor food-related material hardship inU.S. households and its relationship to program partic-ipation and other outcomes. Stormer and Harrison exam-ined the association between household food securityand the cognitive performance and social behavior ofchildren entering kindergarten. Townsend and Melgar-Quiñonez assessed the relationship between householdfood security and the prevalence of overweight andobesity in children. Two research projects examined

food security among elderly households. Salmon andGooden documented the food security status of elderlywho were placed on a waiting list to receive home-delivered meals in North Carolina, while Frongillo andWolfe used longitudinal data to examine the effect ofreceiving home-delivered meals on food securityamong elderly people in New York.

2. The Role of the Private Sector in Providing FoodAssistance. Although Federal programs provide thebulk of food assistance in the United States, manyhouseholds rely on private charitable organizations intheir communities to help them meet their food needs.Cashwell and coauthors assessed the knowledge of thedirectors of private, faith-based emergency foodproviders in the South about the charitable choiceprovision of the 1996 welfare reform legislation thatallows them to compete for government funds toadminister social service programs. Nichols-Casebolt,focusing on households that get food from privatecharitable organizations, examined how use of Federalfood and cash assistance by food pantry users inVirginia has changed since passage of the 1996welfare reform legislation. Child care providers areanother type of private-sector organization that mayplay a role in providing food assistance to low-incomehouseholds. Eligible child care providers can partici-pate in the Child and Adult Care Food Program(CACFP), a Federal program that provides reimburse-ment for meals served to children and adults.Huddleston and Shobo assessed factors that influenceparticipation in the CACFP by child care providers inArkansas.

3. Dynamics and Life Course Patterns of FoodAssistance Program Participation. Food assistanceprograms serve a dynamic population of low-incomeAmericans. Many people enter and leave the programseach month, which calls for research that examinesprogram participation over time. Rank and Hirschlused 30 years of data to assess the extent to whichmembers of low-income households depend on theFood Stamp Program (FSP) over the course of theirlives. Goetz and coauthors assessed factorscontributing to FSP caseload decline during the late1990s, with a focus on the role of county-level factors.Swann analyzed factors that influence a pregnantwoman’s decision to participate in the SpecialSupplemental Nutrition Program for Women, Infants,and Children (WIC) and the length of time she partici-pates in the program.


4. Nutritional Status and Diet Quality. Reliablemeasures of nutritional outcomes are needed to assesswhether food assistance programs improve the nutri-tional well-being of low-income households. Siega-Rizand Kranz designed a tool to assess the quality ofAmerican preschoolers’ diets. Bhattacharya and coau-thors examined changes in the nutritional well-beingof low-income households during cold-weather periodsas part of their investigation of the “heat or eat”phenomenon—when families cut back on food expen-ditures. Parrish examined the nutritional content of thefood available to the Keweenaw Bay Ojibwa residentsof the L’Anse Reservation in northern Michigan andthe eating and exercise patterns of a sample of elderson the reservation.

5. Food Security and Food Assistance Programs inthe Community Context. A household’s participationin food assistance programs and its level of food secu-rity are likely to be influenced by the characteristics ofthe community in which it lives. Lopez and coauthorsexamined the availability of traditional foods on theTohono O’odham Reservation and assessed how recep-tive food assistance recipients on the reservation wouldbe to eating traditional foods if they were available.Bentley and coauthors, in partnership with a sample ofresidents of predominately minority neighborhoods inDurham, NC, conducted an analysis of the communityfactors related to household food security.

6. Welfare Reform and Food Assistance Programs.The 1996 welfare reform legislation introduced majorchanges to the cash welfare system in the U.S. Thesechanges are likely to have affected FSP participationbecause, historically, many cash welfare recipients alsoreceived food stamps. Davis and coauthors examinedthe use of food assistance among Northern Cheyennefamilies in the post welfare reform environment. Thewelfare reform law also eliminated eligibility of mostlegal immigrants to receive food stamps.1 Bollingerand Hagstrom examined the distinct patterns of FSPparticipation of refugees, who were exempted from theeligibility restrictions, and nonrefugee immigrants.White and coauthors examined whether factors otherthan the eligibility restrictions contribute to the declinein FSP participation among noncitizen households inTexas.

1Subsequent legislation in 1997 allowed legal immigrants whowere disabled, elderly, or children living in the United States inAugust 1996 to regain eligibility for the Food Stamp Program.More recently, the 2002 Farm Act restored food stamp eligibilityto legal noncitizens who have lived in the United States for at least5 years or who are children or disabled, regardless of how longthey have lived in the United States.


Does Household Food Insecurity AffectCognitive and Social Development ofKindergartners?

Ame Stormer and Gail G. Harrison, University ofCalifornia, Los Angeles

Contact:Gail G. Harrison, Professor and ChairUCLA School of Public HealthDepartment of Community Health Sciences14-171 Warren Hall900 Veteran AvenueLos Angeles, CA 90095Phone: [email protected]

Grant awarded by the Institute for Research onPoverty, University of Wisconsin-Madison

This study explored the relationships between house-hold food insecurity and the cognitive performanceand social behavior in U.S. children entering kinder-garten. Earlier studies have found that hunger is asso-ciated with poor school performance, such as moreschool absences, tardiness, and increased probabilityof repeating grades. There is also some evidence ofcompromised social and emotional functioning amongadolescents in food-insecure households. The authorsfocused on children entering kindergarten in order toexamine cumulative childhood development prior toschooling. In addition, readiness for school is apowerful predictor of later success and developmentfor children.

The authors used data from the 1998 Early ChildhoodLongitudinal Study of Kindergartners (ECLS-K). TheECLS-K is a nationally representative cluster sampleof approximately 20,000 children in both public and

private schools. The survey includes the 18 questionsused to construct the Federal measure of householdfood security. The authors examined the links betweenhousehold food insecurity, teacher- and parent-reportedsocial skills, teacher-reported cognitive ability, and anindependent direct assessment of children’s cognitiveabilities in math, reading, and general knowledge.These measures constitute an unusually extensive eval-uation of children’s cognitive and social skills.Additionally, the authors investigated the associationsbetween household food insecurity and children’sheight and weight when they entered kindergarten.

The authors used factor analysis to reduce the cogni-tive and social data to five summary measures, andmultiple linear regression analysis to examine theprediction of these measures by a large number ofpotential independent variables, including sociodemo-graphic characteristics, school and home environmentcharacteristics, parental and teacher characteristics,day care and preschool experience, and household andchild participation in Federal assistance programs.

The study results show that household-level food inse-curity is not a significant independent predictor ofcognitive performance, whether assessed by teachersor by an independent observer. However, food insecu-rity is significantly and negatively related to parents’rating of their children’s emotional state and socialinteraction skills and to teachers’ rating of children’ssocial skills. Food insecurity is not a significant inde-pendent predictor of short stature, overweight, orunderweight among kindergarten children. Physicalactivity is a significant predicator of overweight statusamong children. The amount of time spent watchingTV on weekends is positively associated with beingoverweight, while the teacher-rated activity levelduring free play is negatively associated with beingoverweight.

Executive Summaries

Food Insecurity and Hunger


Hunger, Food Insecurity, and Child Obesity

Marilyn Townsend and Hugo Melgar-Quiñonez,University of California, Davis

Contact:Marilyn TownsendUniversity of California, DavisDepartment of Nutrition 3150D Meyer HallOne Shields AvenueDavis, CA 95616-8669Phone: [email protected]

Grant awarded by the Department of Nutrition,University of California, Davis

While the prevalence of obesity is increasing amongchildren of all socioeconomic groups, obesity is mostcommon among children of low socioeconomic status.At the same time, many low-income households withchildren report that they have difficulty in meetingtheir food needs. The development of a Federalmeasure of food security has made it possible toconsistently monitor the difficulties that householdsexperience in meeting their food needs. In 2001, about13 million children lived in food-insecure households,in which, according to the Federal definition, avail-ability of food was limited or uncertain because offinancial constraints. Recent studies have found a posi-tive association between food-related material hardshipand overweight among U.S. women. This studyconsidered whether a positive association betweenoverweight and food-related material hardship alsoexists among children.

The authors examined the relationship between ameasure of food-related material hardship and childoverweight for Mexican-American, non-HispanicBlack and non-Hispanic White boys and girls ages 2-19. They combined several years of data from theContinuing Survey of Food Intakes for Individuals(CSFII) to examine a nationally representative sampleof 6,473 children. The data provide information on thereported height and weight for each child. The authorsadjusted the values of height and weight to account forthe error typically found in self-reported measures, andused the adjusted values to calculate an indicator ofwhether the child was overweight or at risk ofbecoming overweight.

Surveyed households reported whether they had (1)enough of the kinds of food they wanted to eat, (2)enough but not always the kinds of food they wantedto eat, (3) sometimes not enough to eat, or (4) oftennot enough to eat. The authors categorized householdsthat indicated that they had enough of the kinds offood they wanted to eat as food secure and all otherhouseholds as food insecure.2 Note that this measureof food security differs from the Federal measure justdescribed, which is assessed by a series of 18 ques-tions about a household’s difficulties in meeting itsfood needs due to financial constraints.

The authors found that 12 percent of children are over-weight, and another 16 percent are at risk of becomingoverweight. A child’s risk of becoming or being over-weight increases as their dietary energy intake, satu-rated fat intake, or time spent watching televisionincreases, and decreases as their household incomerelative to the poverty line increases.

About three-fourths of children live in householdscategorized as food secure, according to the study’sdefinition. The authors separated children into four agegroups and found that food insecurity is positivelyassociated with overweight and risk of overweightamong children ages 12-15 and children ages 16-19.They also separated children according to their raceand ethnicity and found no significant relationshipbetween overweight and food insecurity among non-Hispanic White children of any age. However, theyfound a positive association between overweight andfood insecurity for several age groups of Mexican-American and non-Hispanic Black children. Foodinsecurity is positively associated with overweightand risk of overweight for non-Hispanic black chil-dren ages 12-15 and Mexican-American childrenages 6-11.

The study results show that older minority childrenwho are food insecure are more likely to be over-weight than those who are food secure. The authorsnoted that corroboration of this finding may guide thedevelopment of education interventions that accom-pany food assistance programs, such as the FoodStamp Program, available to food-insecure families.

2A number of other research studies have used the responses tothis survey question to categorize people as food insufficient ifthey report that they sometimes or often do not have enough to eat,or food sufficient otherwise.


A Study of Older Adults on the Waiting List for Home-Delivered Meals in North Carolina

Mary Anne P. Salmon, The University of NorthCarolina at Chapel Hill

Jessalyn Gooden, The University of North Carolina atGreensboro

Contact:Mary Anne P. Salmon, Clinical Associate ProfessorThe University of North CarolinaSchool of Social WorkCenter for Aging Research and EducationCB# 3550, 301 Pittsboro StreetChapel Hill, NC 27599-3550Phone: [email protected]

Grant awarded by the Southern Rural DevelopmentCenter, Mississippi State University

The Nutrition Services Incentive Program (NSIP) is aFederal program that provides incentives for Stateagencies to provide meals to older adults. The programis administered by the U.S. Department of Health andHuman Services (HHS), and States may obtaincommodity foods from USDA.3 The NSIP providesfunding and commodities for the provision of home-delivered meals to adults over age 60 in a number ofStates. The NSIP does not allow State agencies to setincome limits for the receipt of home-delivered meals.However, the NSIP is not an entitlement program, anda number of applicants may be placed on a waiting listif the State does not have sufficient resources toprovide meals for all applicants. This study examineda population of older adults who have been placed ona waiting list for home-delivered meals in five countiesin North Carolina to assess their functional ability,nutritional status, and strategies to obtain food.

The authors conducted telephone interviews with 110people who had been assessed as eligible and placed onthe waiting list for home-delivered meals in one AreaAgency on Aging (AAA) Region in North Carolina.

They found that those on the waiting list were aboutthe same age as those who began receiving home-delivered meals in 2001, about as likely to be female,and somewhat more likely to be African-American andto live alone. Almost half of those on the waiting listlived in households with incomes below the povertyline. Survey respondents were asked whether theyneeded assistance with nine standard activities of dailyliving, such as getting dressed and transferring frombed to chair. About 20 percent reported that they didnot require the help of another person for any of thenine activities. Those on the waiting list were lesslikely to report needing help getting dressed and trans-ferring from bed to chair than those who beganreceiving home-delivered meals in 2001.

Almost all of those on the waiting list were at highnutritional risk, based on an index used by the HHSAdministration on Aging. More than a third reportedthat they had neither eaten fruit nor drunk fruit juicethe day before the interview, over 40 percent had eatenno nonstarchy vegetables, less than 15 percent hadeaten neither fruits nor vegetables, and over 40 percenthad neither drunk milk nor eaten calcium-rich prod-ucts. While many on the waiting list were at highnutritional risk, very few were underweight. In fact,more than half were overweight.

Those on the waiting list to receive home-deliveredmeals relied on a variety of sources for their meals.Seven in 10 reported that they had someone bringthem prepared meals, which accounted for an averageof 3.2 of their meals each week. Almost two-thirds hadsomeone bring them groceries, which provided aweekly average of almost 16 meals. The respondent’sadult children were the most likely to bring preparedmeals or groceries. Almost 10 percent of respondentsdid not receive prepared meals or groceries or the helpof someone coming to their home to cook for them.Almost half of the survey respondents reported thatthey did not always have enough money (or foodstamps) for food. However, only 15 percent reportedthat they were receiving food stamps.

The study results indicate that almost all of those onthe waiting list to receive home-delivered meals are athigh nutritional risk. However, there is some variationin their access to an informal support network to helpprovide them with meals, which suggests a need tofocus resources on applicants who do not haveinformal supports.

3Until 2003, the program was known as the Nutrition Programfor the Elderly and administered at the Federal level by USDA.


Impact of Home-Delivered Meals on theNutritional Status and Food Security ofthe Elderly in New York State

Edward A. Frongillo and Wendy S. Wolfe,Cornell University

Contact:Edward A. Frongillo, Associate Professor

and DirectorCornell UniversityDivision of Nutritional SciencesProgram in International NutritionB17 Savage HallIthaca, NY 14853Phone: [email protected]


The Home-Delivered Meals (HDM) component of theNutrition Services Incentive Program provides mealsto noninstitutionalized frail elderly persons who are nolonger able to obtain an adequate diet without assis-tance.4 This study examined whether participation inthe HDM program improves nutritional outcomes,characterized elders who are most likely to benefit fromthe program, and identified a number of nutritionalindicators that can be used in program evaluation.

Several studies have found benefits associated withHDM. For example, a recent national evaluation foundthat participants in both HDM and Congregate Mealshad a higher average daily intake of nutrients than amatched comparison group of nonparticipants.However, due to methodological limitations, thesestudies could not determine whether participation inthe HDM program was responsible for the improvedoutcomes. The authors addressed the methodologicallimitations of previous studies by comparing thosewho received HDM with a nonrandomized controlgroup and by using longitudinal data on the nutritionaloutcomes of both the control and treatment groupmembers. The authors also examined outcomes otherthan nutrient intakes, such as food security and dietarypatterns, and information on measured, rather thanself-reported, height and weight.

This collaborative study analyzed data collected by theNew York State Office for the Aging in 1999 in threecounties representative of Upstate New York. Allelders referred for aging services over a 5-monthperiod received a standard assessment and were askedto participate in the study. Of the 212 people whoagreed to participate, 171 began receiving HDM and41 began receiving other services but not HDM. Boththe HDM recipients and nonrecipients completed abaseline survey and two followup surveys, at 6 monthsand 12 months after the baseline survey. The surveyscontained a 24-hour dietary recall and questions usedto construct the Federal measure of food security. Anumber of elders in the initial sample did not completethe followup surveys, primarily because they stoppedparticipating in HDM because they no longer neededit, moved, or died. A total of 99 elders completed the6-month followup survey, and 67 elders completed the12-month followup survey. The authors used estima-tion techniques that suggest that the bias resultingfrom the loss of sample members is minimal.

The study found that respondents’ dietary intakesbefore receipt of HDM were lower than recommendedfor many nutrients and food groups. After receipt ofmeals through the HDM program, however, partici-pants reported eating more vegetables, a greatervariety of fruits and vegetables, and more betacarotene, vitamin E, and magnesium. In addition, theprevalence of food insecurity among elders in thesample decreased from 23 percent to 13 percent. Theanalysis shows that the positive effects of the HDMare greater among males and elders who live alone.Participants with poorer initial nutritional status alsoshow greater improvement, regardless of demographiccharacteristics.

The analysis compared HDM participants whose dietswere assessed in the followup surveys on days thatthey had eaten an HDM meal with participants whosediets were assessed on days that they had not eaten anHDM meal. Participants who had eaten an HDM mealon assessment day showed greater improvement overtime in 25 of 27 indicators of nutrient intake anddietary patterns compared with participants who hadnot eaten an HDM meal on assessment. The authorsfound similar results when they examined nutrientdensity rather than total nutrient intake.

These results provide evidence that the HDM programimproves the nutritional well-being of the elderly. Thefact that some applicants are placed on a waiting list

4Until 2003, the program was known as the Nutrition Programfor the Elderly and administered at the Federal level by USDA.


shows there is unmet interest in participating in theHDM program. The positive effects of the program arean important consideration in the decision to allocatelimited resources to social programs. Information onthose who are likely to benefit most from the program

can be used to target program resources more effec-tively. In addition, the new nutritional indicators iden-tified in the study can be used to measure the effect ofthe HDM and other food assistance and nutritionprograms.


Private Food Assistance in the DeepSouth: Agency Profiles and Directors’Perceptions of Needs and OpportunitiesUnder Charitable Choice

Suzie T. Cashwell, Western Kentucky University

John P. Bartkowski, Mississippi State University

Patricia A. Duffy, Joseph J. Molnar, Vanessa Casanova,and Marina Irimia-Vladu, Auburn University

Contact:Suzie T. CashwellWestern Kentucky UniversitySocial Work Program1 Big Red WayBowling Green, KY 42101Phone: [email protected]


Emergency food providers are an important source offood assistance to low-income households in manycommunities. The direct providers of emergency foodassistance are often private faith-based organizations,which are the focus of the charitable choice provisionof 1996 welfare reform legislation. The charitablechoice provision enables government agencies to fundthe service programs of faith-based organizations andhas the potential to significantly alter the fundingpossibilities available to emergency food providers.

This study examined the types of local communityagencies that provide food assistance to the poor in theAlabama-Mississippi area and information about theagencies’ directors. The authors collected primary datafrom a random sample of approximately 230 foodpantry directors in Alabama and Mississippi. Thesurvey provided information on the demographic char-acteristics of food pantry directors and their attitudesabout poverty and food pantry use. The survey alsocontained questions designed to assess how knowl-edgeable the directors are about the charitable choiceprovision (that is, their familiarity with the generalpolicy contours, specific legal provisions, and imple-mentation status provision) and how receptive they areto receiving government funds.

The study found that about three-quarters of the surveyedfood pantries are affiliated with a religious congregation.About two-fifths of faith-based food pantries are affiliatedwith religious congregations with fewer than 100members, and another two-fifths are affiliated with reli-gious congregations with 100-500 members. About 70percent of the food pantries reported that they primarilyserve rural populations. Three-fifths of the food pantrydirectors are White, and about 65 percent are female.Almost 80 percent have some education beyond highschool, and about 40 percent reported an annual house-hold income of $50,000 or more. More than two-thirdsreported that they attend church at least once per week.

The survey provides information on the perceptions offood pantry directors regarding pantry users and theirneed for food assistance. Most of the directorsreported that food pantry use is related to low wages insome industries or to sickness or physical disability,and that most food pantry users who are able to workare trying to find jobs. However, almost one-third offood pantry directors believed that too many peopleusing food pantries should be working, and almost halfbelieved that many people getting food are not honestabout their needs.

The study also gauged food pantry directors’ aware-ness of the charitable choice provision and their open-ness to accepting government funding. A third of thesurveyed food pantry directors reported that they donot currently receive government funds, while slightlymore than half indicated that they would be willing toapply for government funds in the future. Most of thedirectors were generally aware of the legal responsibil-ities associated with receipt of government funding,but were also unsure of the general policies relating tothe charitable choice provision, such as the process toapply for funds and the extent to which charitablechoice has been implemented in the U.S.

The authors noted that additional research in other partsof the U.S. will help determine if there are broaderregional variations or systematic rural-urban differencesin the knowledge about the charitable choice provisionamong community-level organizations. However, thestudy results suggest that there is a need to educateorganizations that can potentially benefit from thecharitable choice provision of the 1996 welfare reformlegislation so that they can make a reasoned choiceabout the new opportunities available to them.

The Role of the Private Sector in Providing Food Assistance


Making Ends Meet: An Examination ofTANF and Former TANF Food PantryUsers in Virginia

Ann Nichols-CaseboltVirginia Commonwealth UniversitySchool of Social Work1001 W. Franklin StreetRichmond, VA [email protected]

Grant awarded by the Joint Center for PovertyResearch, University of Chicago and NorthwesternUniversity

Since passage of the 1996 welfare reform legislation,questions have been asked about whether those whohave stopped using Federal cash assistance and foodstamps have achieved self-sufficiency, or whether theyhave instead come to rely on assistance from privateorganizations. Nichols-Casebolt conducted a series ofstatewide surveys of Virginia food pantry users from1997 to 2001. She examined the characteristics ofVirginia families with children who sought assistancein food pantries and described changes over time intheir receipt of cash assistance and food stamps. Shealso collected information on the material well-beingof food pantry users, including their food securitystatus in 2001.

The study found that demographic characteristics offood pantry users changed little between 1997 and2001. In contrast, the share of food pantry usersreceiving cash assistance or food stamps fell. About 42percent of food pantry users were also receiving foodstamps at the time of the 1997 survey, while about 30percent of food pantry users were also receiving foodstamps in 2001. The author also found that over 35

percent of food pantry users who had recently left theTemporary Assistance for Needy Families (TANF)indicated that they had also stopped receiving foodstamps within the last 6 months. After statisticallycontrolling for a number of other factors that influencefood stamp receipt, the author found that TANFleavers are significantly less likely than TANF recipi-ents to receive food stamps. Food pantry users who aremost likely to be at economic risk—including theunemployed, single parents, parents with young chil-dren, and those with low education levels—are mostlikely to receive food stamps.

The study also documented the material hardships ofmany households who seek services at food pantries.Over 80 percent of the families were food insecure andover 25 percent lost telephone service at some time inthe past 6 months. About 15 percent had been recentlyforced to change their living arrangements. The authorexamined the factors associated with the food securitystatus of food pantry users. The results indicate thathigher household income is associated with greaterfood security. After household income and demo-graphic characteristics are controlled for, the receipt ofeither cash assistance or food stamps has no statisti-cally significant effect on the food security status offood pantry users.

The author concluded that private food assistanceplays a strong role in meeting the food needs of somelow-income families. She noted that the decrease inthe Food Stamp Program (FSP) participation ratesuggested that food pantries are promising locationsfor FSP outreach efforts. Given the strong positiverelationship between household income and food secu-rity, the author suggests that policies that focus onemployment and work supports are important elementsto improve food security in this population.


The Arkansas Child and Adult Care FoodProgram: A Study of Factors AssociatedWith Program Participation

Richard A. Huddleston and Yetunde A. Shobo,Arkansas Advocates for Children and Families

Contact:Richard A. Huddleston, Research DirectorArkansas Advocates for Children and Families523 S. Louisiana, Suite 700Little Rock, AR 72201Phone: 501-371-9678


The Child and Adult Care Food Program (CACFP) is aFederal food assistance program that provides reim-bursement for meals provided to children and adults ineligible child and adult care programs. To receivereimbursements, child care providers must sign up forCACFP and meet program eligibility criteria. Accessto the CACFP program by the target populationdepends on the availability of child care providers andon providers’ decisions to participate in the program.Two major types of child care providers are familychild care homes and child care centers.5 In Arkansas,family child care homes are licensed to serve amaximum of 16 children at a time, while child carecenters are typically licensed to serve a larger numberof children. This study examines the geographic varia-tion in child care availability in Arkansas and thefactors that affect child care providers’ participation inthe CACFP.

The 1996 welfare reform legislation introduced a newreimbursement structure for family child care homesinto the CACFP. The reimbursement rates that apply tohomes in low-income areas or homes operated by low-income people are similar to those that applied beforethe 1996 legislation, but reimbursement rates to allother homes are at least 40 percent lower.6 The changein reimbursement rates may have reduced CACFPparticipation among family child care homes servingchildren from middle-income families.

The authors used 2002 State administrative data onchild care licensing and CACFP participation toanalyze county-level child care provision and CACFPparticipation in Arkansas. They also used 2000 Censusdata to construct measures of potential need for childcare and for food assistance. In addition, they surveyedover 900 licensed child care providers in 2002 tocollect information on the factors that influence theirdecisions to participate in the CACFP.

The study found that almost 70 percent of family daycare homes in Arkansas participate in the CACFP,comprising 70 percent of CACFP providers. Child carecenters make up 13 percent of CACFP providers inArkansas. The analysis shows that the availability ofchild care providers, particularly those that participatein the CACFP varies significantly across regions ofArkansas. Family day care homes are concentrated inthe delta region, which is located in southwestArkansas and has the highest concentration of povertyin the State. The authors constructed a measure ofchild care capacity in a county, equal to the totalnumber of children that child care providers in thecounty are licensed to serve. They found that childcare capacity as a share of the child population (age 5and younger) is higher in the counties of the deltaregion than in the counties in the rest of the State.However, because of the high rate of poverty in theregion, the child care capacity of CACFP providers asa share of the population of poor children is lower inthe delta region than in the rest of the State.

The authors used logit regression analysis to examinethe effect of provider characteristics and perceptionson CACFP participation by providers. The results ofthe regression analysis indicate that family day carehomes are more likely than child care centers to partic-ipate in the CACFP and that nonprofit child careproviders are more likely to participate than for-profitproviders. The longer the provider’s hours of operationper week, the more likely it is to participate in theCACFP.

The results of the child care providers survey indicatethat 90 percent of child care providers are satisfiedwith CACFP administration. However, about halfreported that the reimbursement rates are not adequateto cover costs, and about one-fourth reported that thepaperwork requirements are too high. Child careproviders not participating in the CACFP were askedabout the barriers to participation. Almost 40 percentreported that they do not know about the program, and

5Other CACFP providers include infant and toddler care,school-age care, and sick care programs, which are not analyzed inthis study.

6For these other homes, higher reimbursements remained avail-able for meals served to individual low-income children.


one-fourth reported that they do not know how toapply or that the application process is too difficult.

The authors concluded that many of the providers whodo not participate in the CACFP are not eligible toreceive large amounts of meal reimbursements becausethey typically serve few low-income children.

However, the survey results suggest that some poten-tially eligible providers do not participate because theyare not familiar with the program or are overwhelmedby the application process. These findings suggest thatintervention strategies may be developed to encouragegreater participation by eligible non-CACFP providers.


Estimating the Probabilities and Patternsof Food Stamp Use Across the Life Course

Mark R. Rank, Washington University

Thomas A. Hirschl, Cornell University

Contact:Mark R. RankWashington UniversityGeorge Warren Brown School of Social WorkSt. Louis, MO 63130Phone: [email protected]

Grant awarded by the Joint Center for PovertyResearch, University of Chicago and Northwestern University

The Food Stamp Program (FSP) is the largest U.S.food assistance program. With some exceptions, theFSP is available to all households for which incomeand assets fall below certain levels. Information isalready available on the extent to which householdsrely on the FSP during a given year. However, muchless is known about how households use food stampsover the course of many years. Rank and Hirschl esti-mated the lifetime probabilities and patterns of foodstamp use for the U.S. population, using a life tableprocedure. This approach provides empirical evidenceon the range and scope of the FSP in the lives ofAmericans.

The authors used the Panel Study of Income Dynamics(PSID), a longitudinal survey of a representativesample of U.S. individuals and their families. Theymerged 30 waves of data from 1968 to 1997 to build aseries of life tables detailing the cumulative probabili-ties of participating in the FSP. They examine twoperiods of the life course—childhood (ages 1-20) andworking-age adulthood (ages 20-65). Within each ofthese periods, they estimated the overall likelihood offood stamp use, the total and consecutive number ofyears that food stamps are accessed, and the effectsthat race, education, gender, and marital status haveupon the likelihood of using the FSP.

The results indicate that food stamp use is quitecommon during both childhood and working-ageadulthood: 49 percent of American children receivefood stamps at some point by the time they reach age20, and 51 percent of American adults participate inthe FSP sometime between the ages of 20 and 65.Furthermore, once a household uses food stamps, it isquite likely to use the program again. Two-thirds ofchildren who receive food stamps will do so in at leastone additional year, while three-quarters of adults whouse the program will do so more than once.

The results further indicate that food stamp use acrossthe life course tends to occur over relatively shortperiods of time. For example, while half of all childrenreceive food stamps some time before age 20, only 1in 10 will do so in 5 consecutive years. These findingsare consistent with earlier work examining the lifecourse patterns of poverty, as well as the bulk ofresearch examining the dynamics of poverty andwelfare use. Although some households use foodstamps for long periods of time, most food stamp usersrely on the FSP to provide short-term assistance.

In this study, race, education, and marital statusprofoundly affect the probability that a person will usefood stamps during his or her lifetime. BlackAmericans, people who have not graduated from highschool, and children residing in nonmarried house-holds have a high probability of using food stampsover the course of their lives. For example, 90 percentof Black children use food stamps at some pointduring their childhood compared with 37 percent ofWhite children.

Study results indicate that a wide segment of theAmerican population uses FSP some time during theirlives. However, because different households partici-pate in different years, more households participate atsome point over a period of several years than partici-pate in any 1 year. While roughly half of Americanchildren and half of working-age adults participate inat least 1 year, most FSP participants use the programfor short-term assistance.

Dynamics and Life Course Patterns ofFood Assistance Program Participation


Food Stamp Program ParticipationDynamics in U.S. Counties and States

Stephan J. Goetz and Anil Rupasingha,The Pennsylvania State University

Julie N. Zimmerman, University of Kentucky

Contact:Stephan J. GoetzThe Pennsylvania State UniversityThe Northeast Regional Center for Rural DevelopmentDepartment of Agricultural Economics

and Rural Sociology7E Armsby BuildingUniversity Park, PA 16802-5602Phone: [email protected]


The Food Stamp Program (FSP) and cash assistancecaseloads have fallen dramatically since the mid-1990s, and the rate of caseload decline differs acrossStates and counties. Prior research on cash assistanceand FSP caseload declines focused on the effect ofmacroeconomic changes, such as changes in State-level unemployment rates. This study examined thefactors associated with caseload declines, accountingfor economic conditions and demographic characteris-tics at the county level, and for employment conditionsin the labor markets that are most likely to employformer welfare recipients.

In their analysis, the authors used county-level data oneconomic conditions, which provided more reliable

information than State-level data about the employ-ment prospects of former welfare recipients. In addi-tion, they accounted for employment conditions in thefoodservice and retail sectors, in which former welfarerecipients are most likely to find jobs.

The authors used estimation techniques that accountedfor the spatial clusters of program participation, aswell as the direct effect that welfare reform has onlabor market conditions. The authors found that highercounty-level earnings per capita were associated with amore rapid decline in per capita FSP spending between1995 and 1999, while county-level retail employmentgrowth did not appear to have an effect on the declinein spending. Reductions in food stamp payments percapita were lower in rural counties than in suburbancounties. In counties with proportionally more foreign-born and African-American residents, per capita FSPspending fell more quickly. In counties with propor-tionally more single female-headed households, percapita FSP spending fell more slowly. A greaternumber of vehicles per household and more bus serv-ices per person in a county were associated with amore rapid decline in FSP spending.

The study results suggest that both individual- andcommunity-level factors play a role in explainingchanges in FSP participation over time. Further, theway that community-level factors are measured is crit-ical, as is controlling for the spatial clustering ofprogram participation. Finally, per capita FSPspending falls less rapidly in rural counties than insuburban counties, suggesting a systematic differencein the processes that affect food stamp use in rural andsuburban counties. The authors noted that the studyfindings can be used in forecasting fiscal outlays aseconomic conditions change and measuring the effec-tiveness of the FSP.


The Dynamics of Prenatal WIC Participation

Christopher A. Swann, Assistant ProfessorState University of New York at Stony BrookDepartment of Economics, SBS S637Stony Brook, NY 11794Phone: [email protected]


The Special Supplemental Nutrition Program forWomen, Infants, and Children (WIC) provides foodvouchers, nutritional counseling, and health care refer-rals to low-income pregnant and breastfeeding womenand their young children. This study analyzed thecharacteristics that affect the probability of WICparticipation over the course of a woman’s pregnancy.Understanding the program rules and other factors thataffect the timing and duration of WIC participation canhelp to ensure that the program is designed to targetresources most effectively.

Although many studies have examined the effect of WICparticipation on various health outcomes, relativelylittle research has focused on the decision to partici-pate in WIC and none have modeled the decision ofwhen to participate. One study provided descriptiveinformation about the month at which participationbegins and the number of months of participation, anda number of other studies explored the role of earlyversus late participation in WIC on birth outcomes.Another recent study, as part of an analysis of healthoutcomes associated with participation, estimated asimple equation to explain WIC participation.

In this study, the author used data from the 1988National Maternal and Infant Health Survey, a nation-ally representative sample of women who experienceda live birth or infant death in 1988, and informationabout State WIC policies from the 1988 Survey ofWIC Program Characteristics. The WIC program datainclude information on State-level program rules,which varied significantly across the States in 1988.The State-level differences helped the author to assesswhich program characteristics are important determi-nants of early participation in WIC.

The study used both parametric and nonparametrichazard rate analysis to examine the relationshipsamong recipient characteristics, WIC program rules,and the timing of WIC participation. These techniquesrelate maternal, family, and program characteristics tothe likelihood that an eligible woman begins partici-pating in WIC during each month of pregnancy, giventhat she has not participated in the previous months ofpregnancy.

The analysis shows that WIC participation is higheramong women who have low education levels, areHispanic, have low income (even among those who areeligible), and participate in other welfare programs.The probability of participating in WIC increasesduring the first 4 months of pregnancy and decreasesthereafter. Women who have participated in WICduring a previous pregnancy are three times morelikely to participate in WIC than women who have notparticipated in the past. The author noted, however,that this result should be interpreted with cautionbecause of limitations on the information in the surveyabout previous WIC participation.

The study also found that State-level program charac-teristics are important determinants of WIC participa-tion. Women in States that allow applicants toself-report their income when determining eligibilityare 30 percent more likely to participate in WIC thanwomen in States that require applicants to providedocumentation of their income. In addition, the State-level policy to offer adjunctive eligibility for WIC toMedicaid and cash assistance recipients is associatedwith a 10- to 20-percent increase in the likelihood ofWIC participation.

Recent policy changes have mandated documentationof income for eligibility determination and haveexpanded adjunctive eligibility. Given the estimates ofthis study, these policies have opposite effects onparticipation. Restricting the ability of applicants toself-declare income decreases the probability of WICparticipation, and mandating adjunctive eligibilityincreases it. The author simulated the effects of thesepolicy changes and found that the net effect is a 1-percentage-point increase in the probability ofWIC participation over the course of a 40-weekpregnancy.


Development of a Diet Quality Index forPreschool Children and Its Applicationin Examining Dietary Trends in theUnited States

Anna Maria Siega-Riz, The University of NorthCarolina at Chapel Hill

Sibylle Kranz, The Pennsylvania State University

Contact:Sibylle Kranz, Assistant ProfessorThe Pennsylvania State UniversityDepartment of Nutritional Sciences252 Henderson South BuildingUniversity Park, PA 16802Phone: [email protected]


The prevalence of childhood obesity has increased sig-nificantly in the past three decades. Obesity carries anumber of negative medical consequences, includingcardiovascular problems and diabetes. In addition, thedietary intake behavior of adults is strongly associatedwith their dietary intake behavior as children. The authorsdesigned a tool to assess total diet quality in Americanpreschoolers: the children’s diet quality index (C-DQI).The C-DQI can be used to measure the effectiveness ofchild nutrition programs in improving overall diet qualityin children and to help target child nutrition programs topopulation groups at greatest risk of poor diet quality.

The authors used dietary and sociodemographic data forover 5,000 children ages 2-5 from the 1994-96 and 1998Continuing Survey of Food Intake in Individuals (CSFII).They designed the index using information on commonnutritional problems in the preschool population anddietary intake recommendations by such organizationsas the American Academy of Pediatrics, AmericanDietetic Association, and the National Academy ofSciences. The authors selected eight components of theindex that incorporate information on the recommendedconsumption levels of added sugar, total fat, saturatedfat, fruit, vegetables, grains, fruit juice, and iron.

The study found that the C-DQI values range from 16 to70 points within the population, with an average of 46

out of 70 points. The authors tested the index’s ability todistinguish between different levels of diet quality andfound that better diet quality (a more healthful diet)within each of the components of the index is signifi-cantly associated with a higher overall C-DQI score.The analysis indicates that overall diet quality is betterfor boys than for girls and for children living in metro-politan areas than for those living in nonmetropolitanareas. Diet quality in low-income households is betterfor Hispanic children than for non-Hispanic Black orWhite children and for children who attend day care orpreschool than for children who do not attend. Averagediet quality declines as children get older.

The authors also examined the trend in children’s dietquality over time by comparing C-DQI total and compo-nent scores among respondents to the National FoodConsumption Survey 1977-79, the CSFII 1989-91, andthe CSFII 1994-96, 1998. They found that overall dietquality improved somewhat during the study period, butthat an increase in consumption of added sugar, excessjuice, and excess dairy and a decrease in iron consump-tion caused declines in some components of diet quality.Fat consumption, as measured in grams per day,remained stable between 1977 and the mid- to late 1990s,while fat consumption as a percentage of total caloriesdropped 4 percent. The average number of fruit servingsconsumed almost doubled between the early 1990s andthe mid- to late 1990s, with most of the rise attributableto an increase in fruit juice consumption. The authorsnoted that the intake of fat and saturated fat has notdecreased since 1977, despite public health messagesabout the importance of reducing fat consumption.

The study results indicate that children’s diet quality isinfluenced by sociodemographic characteristics of chil-dren and their families. Thus, public health messagescould be targeted specifically to population groups atgreatest risk for poor diet quality. Federal programsdesigned to enhance children’s diet quality, such as theSpecial Supplemental Nutrition Program for Women,Infants, and Children (WIC) and Head Start, targetchildren at high risk for poor diet quality. The C-DQIcould be used to measure the impact of Federalprograms, such as WIC, on overall diet quality and on thelevel of diet quality in the individual components of theindex. Using the C-DQI as an outcome measure couldhelp to inform policymakers in the design and implemen-tation of nutrition programs that will most effectivelyimprove children’s overall diet quality.

Nutritional Status and Diet Quality


Heat or Eat? Cold Weather Shocks andNutrition in Poor American Families

Jayanta Bhattacharya, Stanford University Medical School

Thomas DeLeire, University of Chicago

Steven Haider, Michigan State University

Janet Currie, University of California, Los Angeles

Contact:Jayanta BhattacharyaStanford Medical SchoolCenter for Primary Care and

Outcomes Research117 Encina CommonsStanford CA [email protected]


Poor American families with children may have tomake difficult tradeoffs when they face high heatingcosts in cold weather. This study investigates whetherpoor American families spend less on food and reducethe amount and nutritional value of the food they eatduring these cold periods.

One study found that the diets of poor American chil-dren are inadequate during winter. Two studies ofBritish children, in contrast, failed to identify any rela-tionship between excess winter mortality and depriva-tion. Economists have also examined nutritional resourcesharing among members of poor families. Studies exam-ining the extent to which poor families on food stampswill reduce their food consumption toward the end of abenefit month conclude that food consumption in poorfamilies is potentially vulnerable to financial strains butthat the parents are able to protect their children fromthe adverse effects of these strains to some extent.

The authors measured patterns of household spendingon food and home fuel and patterns of nutritional well-being at the individual level. They used expendituredata from the Consumer Expenditure Survey (CEX)for 1980-98. The data on nutritional well-being arefrom the National Health and Nutrition ExaminationSurvey (NHANES) for 1988-94.

The authors examined spending in four categories—foodeaten inside the home, food eaten away from home,clothing, and home fuel—in unseasonably cold orwarm months. Because changes in spending over thecourse of a year by richer families are less constrainedby financial resources, the authors used these familiesas a comparison group for the spending changes inpoor families. In the analysis of nutritional outcomes,they compared the change in nutritional outcomesbetween summer and winter. Specifically, theycompared the change in nutritional outcomes sepa-rately by age (children vs. adults) and income level(rich vs. poor). They define poor families as thosewhose incomes are below 150 percent of the povertylevel and rich families as those whose incomes aremore than 300 percent of the poverty level.

The study found that poor families spend less on foodin months with unusually low temperatures. Both poorand rich families spend more on heating. While thedollar increase in heating expenditures for a poorfamily is less than that for a rich family, the change isa larger share of the poor family’s budget. In addition,both adults and children in poor families reduce theircaloric intake during the winter. Caloric intake doesnot differ significantly between summer and winter foreither adults or children in rich families.

The results suggest that poor American families withchildren spend more on home fuel at the cost ofspending on food and nutritional well-being. Parents inpoor households are not fully able to protect their chil-dren from the effects of cold weather shocks. Bothchildren and adults in poor families eat less foodduring the winter.


Nutrition Assessment and Education forKeweenaw Bay Ojibwa

Debra ParrishKeweenaw Bay Ojibwa Community College409 Superior AvenueBaraga, MI 49908Phone: [email protected]

Grant awarded by the American Indian StudiesProgram, University of Arizona

Keweenaw Bay Ojibwa Community College estab-lished the Nutrition Assessment and Education projectto assess the nutritional needs of the Ojibwa peopleand to examine ways to address these nutritional needswhile maintaining the traditional nutrition practices ofthe Ojibwa people. The author initiated the project, incollaboration with other tribal organizations and busi-nesses, in response to the high risk and prevalence ofdiabetes, heart disease, and other nutrition-relatedhealth problems among the Ojibwa people.

The author focused on the members of the KeweenawBay Ojibwa community living on or near the L’AnseReservation in northern Michigan on Keweenaw Bayof Lake Superior. About 860 of the 3,550 members ofthe Keweenaw Bay Ojibwa tribal community live onthe reservation. This study provides information on theinitial year of the nutrition project, in which the authorconducted a primary data collection by analyzing thenutrition content of the food available on the L’AnseReservation and surveying 40 elders living on or adja-cent to the reservation about their health status andfood consumption practices.

The author collected information on the food availableat the three restaurants on the L’Anse Reservation andassessed the nutrient content of the food. While therestaurants offer some low-calorie meals, many of themeals are high in fat and calories and lack fruits andvegetables. None of the restaurants serve traditionalOjibwa foods, such as wild rice, fish, wild game, andseasonal fruits and vegetables. The author alsocollected information on the food served at the ElderlyNutrition Program and the Head Start program. These

programs do not serve traditional Ojibwa food regu-larly, but do serve it when they receive donations fromlocal fishermen, hunters, or gardeners. Even then, thefood is often fried rather than prepared with traditionallow-fat cooking methods.

Preliminary results of the survey of elders indicate thathealth problems are much more prevalent among thispopulation than among the elderly U.S. populationoverall. Almost half of the Ojibwa elders whoresponded to the survey are obese, 35 percent havediabetes, and almost 40 percent have high bloodpressure.

Most of the elders who responded to the survey (85percent) reported getting some exercise each week.About two-thirds walk as their primary exercise. Overhalf of the respondents eat fast food only once permonth, but almost one-fourth eat fast food at leastthree times per week. The Elderly Nutrition Program,which provides light breakfasts and lunches Mondaythrough Friday to adults age 55 and over, is an impor-tant source of food for elders on the reservation. About80 percent of the survey respondents participate in theprogram, and 25 percent eat at least three meals perweek through the program.

Less than one-third of the respondents eat traditionalOjibwa food once per week or less. However, overone-half reported that they would like to eat traditionalOjibwa food at least once per week, and over one-fourth reported that they would like to eat it at leastonce per day. The most frequently reported barrier toeating Ojibwa food is that it is difficult to get.

In the second year of the nutrition project, the authorwill continue the primary data collection from Ojibwaelders and begin a similar survey of children on thereservation. The analysis of these data could provideimportant insights into the factors related to healthyfood consumption practices and activity levels byOjibwa elders and children. To encourage the Ojibwapeople to eat traditional food, the nutrition projectplans to produce a cookbook of traditional Ojibwafood and to encourage the local restaurants andfeeding programs to incorporate this food in theirmenus.


Community Attitudes Toward TraditionalTohono O’odham Foods

Daniel Lopez, Tohono O’odham Community College

Tristan Reader, Tohono O’odham Community Action

Paul Buseck, University of California, Davis

Contact:Daniel Lopez, InstructorTohono O’odham Community CollegeP.O. Box 3129Sells, AZ 85634Phone: 520-383-8401


The Tohono O’odham people have the highest rate ofdiabetes among Native American tribes. About 50percent of Tohono O’odham adults have adult-onsetdiabetes compared with 4-6 percent of the overall U.S.population. A number of studies have shown that thecomponents of many traditional Tohono O’odhamfoods, such as tepary beans, cholla cactus buds, andwild spinach, help regulate blood sugar and reduce theincidence and effects of diabetes. In this study, theauthors examined the prevalence of traditional TohonoO’odham food in the diets of tribal members andassessed the level of interest among tribal members inincorporating more traditional food into their diets.

The authors surveyed primary caregivers in 128 house-holds, which contain an estimated 625 householdmembers. Almost 75 percent of respondents arefemale, and about 60 percent are under age 36. Almost20 percent of respondents suffer from diabetes, andalmost 80 percent reported that at least one familymember has diabetes.

The study found that less than one-fourth of surveyrespondents often eat traditional food. Respondentswere asked about a number of possible obstacles toeating more traditional food. They reported that thelimited availability of traditional food and lack of timeto prepare it are the two main obstacles to eating tradi-tional Tohono O’odham food. Over 60 percent wouldeat traditional food often if it were available. About 13percent reported that they do not know what the tradi-tional foods are, and less than 1 percent reported thatthey do not like the taste.

The survey shows that 60 percent of the respondentsknow that eating many traditional foods helps preventdiabetes, but just 53 percent know that eating thesefoods can help keep diabetics healthier.

The study found that about 27 percent of respondentsreceive food stamps, 7 percent receive USDAcommodities through the Food Distribution Programon Indian Reservations (FDPIR), and 32 percentparticipate in the Special Supplemental NutritionProgram for Women, Infants, and Children (WIC).Almost 90 percent of surveyed FDPIR participantswould like to receive traditional food as part of theircommodity package, and almost 90 percent ofsurveyed WIC participants would like to receivesupplemental coupons to purchase traditional foods, ifthese foods were available. Almost half of surveyedFood Stamp Program participants would definitely buytraditional foods with their food stamps, and another30 percent would probably buy them. Although theseresults are based on responses to hypothetical ques-tions, they suggest that these Federal food assistanceprogram participants are receptive to incorporatingtraditional food into their diets. The authors maderecommendations for how Federal food assistanceprograms could help to encourage the consumption ofhealthy, traditional Tohono O’odham food.

Food Security and Food Assistance Programs in the Community Context


Contextual Determinants of Food Security in Southern Hispanic andAfrican-American Neighborhoods

Margaret E. Bentley, Sonya Jones, and Janice Dodds,The University of North Carolina at Chapel Hill

Contact:Margaret E. BentleyThe University of North CarolinaDepartment of Nutrition 4115-C McGavran-Greenberg, CB 7400 Chapel Hill, NC 27599-7400 Phone: [email protected]


Since 1995, when the Federal Government beganmonitoring food insecurity, African-American andHispanic households have had consistently higher ratesof food insecurity than the overall U.S. population.This study examines whether community-level factorsin predominantly minority neighborhoods can help toexplain the higher rates of food insecurity amongminority households.

Previous research has found that accessibility of foodvaries with the level of residential racial segregationand that supermarket access is associated with averagedietary quality in a community. The authors conducteda community-based participatory study, in partnershipwith residents of predominantly minority neighbor-hoods, to identify the most important communityfactors that relate to household food insecurity.

The authors held focus groups and photography sessionswith residents of six predominantly minority neighbor-hoods in Durham, NC. They used two innovative

methods to gain an “insider’s perspective” of commu-nity factors related to food insecurity. First, all partici-pants were asked to use cameras to identify and recordthe aspects of their communities that they thoughtwere important in supporting their efforts to eathealthful, adequate diets. Second, participantsanalyzed data from the focus groups to identify thesalient themes among members of their neighborhoodresearch group. In addition, 12 members of the neigh-borhood research groups participated in individualinterviews to explore further their personal experienceswith food insecurity.

Participants in the study identified dietary quality astheir primary food-related concern. They expressedconcerns about five aspects of dietary quality: taste,freshness, balance, variety, and nutrient density offoods. They identified community factors that influ-ence their diet quality, including access to fast foodrestaurants and food stores, the National School Lunchand School Breakfast programs, Special SupplementalNutrition Program for Women, Infants, and Children(WIC), and the Food Stamp Program. Hispanic partici-pants expressed concern about their children’s dietaryhabits related to their greater acculturation to “fastfoods.” However, many participants recognized theimportance of fast food restaurants as communitycenters and as the only source of safe playgrounds insome minority neighborhoods.

This study identified community factors, such asaccessibility of retail food and food assistanceprograms, related to food insecurity in predomi-nantly minority neighborhoods in Durham, NC. Theauthors noted that quantitative analysis is necessaryto test the relationship between community character-istics and household food insecurity. The qualitativeresearch could help to identify possible community-level policies that might help reduce household foodinsecurity.


Access to Food Assistance ProgramsAmong Northern Cheyenne Families

Judith Davis and Rita Hiwalker, Dull KnifeMemorial College

Carol Ward and Eric Dahlin, Brigham Young University

Contact:Judith Davis, Dean of Academic AffairsDull Knife Memorial CollegeBox 98, 1 College DriveLame Deer, MT 59043Phone: 406-477-6215, [email protected]


Federal food assistance programs are an integral partof the social safety net on the Northern CheyenneReservation in southeastern Montana. In 2000, almosthalf of Northern Cheyenne families with children livedin poverty, and in 2002, the unemployment rate for theNorthern Cheyenne Nation was 71 percent. Previousresearch has shown that about one-third of NorthernCheyenne Reservation residents use food stamps andone-third receive USDA commodities through the FoodDistribution Program on Indian Reservations (FDPIR).7

This study examined the use of food assistance by twopopulation groups that are particularly vulnerable tothe adverse economic conditions on the NorthernCheyenne Reservation: seasonal workers and familiesin which the adults face high barriers to employment.

The authors used administrative data to documentchanges in the use of Federal food and cash assistancesince the mid-1990s in Rosebud County, MT, wheremuch of the Northern Cheyenne Reservation islocated. In addition, the authors interviewed 32 reser-vation residents who are either seasonal workers orindividuals who face high barriers to employment. Theinterviews provided detailed accounts of individualand family experiences with the use of Federal food

assistance programs. The authors compared the infor-mation collected from seasonal workers with data theyhad collected from other reservation residents in anearlier study.

The study found that the average monthly number offood stamp recipients fell almost 30 percent between1996 and 2000 and then rose 8 percent between 2000and 2001. The decline in the use of food assistance isparticularly striking because economic conditions forthe Northern Cheyenne Reservation, unlike for most ofthe rest of the country, did not improve during the1990s. The number of households that receivedcommodities through the FDPIR increased slightlybetween July 1998 and July 1999. However, the mostrecent program data show that the average number ofhouseholds receiving commodities decreased almost16 percent between 2000 and 2002.

Most survey respondents are seasonal workers, whomost often cited fire-fighting as their seasonal job.Seasonal workers also reported doing construction work,selling firewood, babysitting, cleaning houses, orworking at local schools as a school aide, cook, or busdriver. The single-parent families in the survey are muchmore aware of the food assistance programs that may beavailable to them than other types of families. Single-parent families with children are most likely to beeligible for and receive cash assistance, and their partici-pation in the cash assistance program is likely to makethem more aware of other Federal assistance programs.

The seasonal workers who do access food assistanceprograms reported that the FSP and FDPIR are impor-tant in helping them meet their food needs. FSP partic-ipants appreciate the flexibility that they have topurchase the kinds of food that they want but notedthat the high food prices on the reservation make itdifficult to stretch their food stamp benefits.Commodity recipients like the amount and kinds offoods they receive. Many reported that the monthlycommodity package provides them with more foodthan they could buy with food stamps and that theapplication process for FDPIR is simpler than for theFSP. A number of seasonal workers also reported diffi-culty in establishing their eligibility for the FSP whentheir employment ends. However, whether this is dueto actual eligibility restrictions or to misinformationabout the program’s eligibility requirements is notclear.

Welfare Reform and Food Assistance Programs

7FDPIR primarily provides commodity foods to low-incomehouseholds living on Indian reservations. Many households partic-ipate in the FDPIR as an alternative to the Food Stamp Programbecause they do not have easy access to food stamp offices orauthorized food stores.


The authors collected information that allowed them toassess the level of food insecurity, stress, and healthproblems among survey respondents. They found thatseasonal workers experience levels of food insecurityand nutritional risk that are almost as high as thoseexperienced by the unemployed. Seasonal workersreported higher levels of stress than the unemployed.While seasonal workers reported that they are able toprovide for the needs of their families during the partof the year that they are employed, they do notconsider the income they receive to be sufficient to last

the remainder of the year. The limitations of the localeconomy prevent most seasonal workers from findingalternative employment.

The cultural norm on the Northern Cheyenne Reservationis to share food and help family members when they arein need. However, because poverty and unemploymentrates are so high, this dependence on extended familystretches the resources of most families very thin.Therefore, Federal food assistance programs representan important source of support for these families.


Food Stamp Program Participation ofRefugees and Immigrants: MeasurementError Correction for Immigrant Status

Chris BollingerUniversity of KentuckyDepartment of EconomicsLexington, KY [email protected]

Paul HagstromHamilton CollegeDepartment of EconomicsClinton, NY [email protected]


After two decades of increasing participation in cashand noncash public assistance programs by immigranthouseholds, the Personal Responsibility and WorkOpportunity Reconciliation Act (PRWORA) of 1996drastically altered the availability of Federal publicassistance to legal immigrants. Immigrants who werenot yet naturalized by 1996 or who entered the countryafter August 1996 became ineligible for Federal bene-fits, although States had the option to provide themwith cash assistance or Medicaid benefits.8 Refugees,however, were given a 5-year exemption from theeligibility restrictions on Federal benefits that appliedto other legal immigrants. Despite the exemption,since 1996, the participation rate of refugees in publicassistance programs, such as the Food Stamp Program(FSP), has fallen at least as fast as for other foreign-born residents. FSP participation of refugees fell 37percent between 1994 and 1997. During the same

period, participation in the FSP dropped 30 percent forimmigrants and 21 percent for native-born citizens.

The authors used the Current Population Survey toestimate the impact of refugee status on FSP participa-tion from 1994 to 2001 and estimated the effect ofPRWORA on FSP participation of refugees. Theycorrected for errors in the measurement of refugee statusand for the misreporting of FSP participation, whichallowed them to get consistent estimates of the effectof refugee status and PRWORA on FSP participation.

The study found that refugees and nonrefugee immi-grants have distinct patterns of FSP participation.Refugees are more likely than other immigrants to usefood stamps near the time of their arrival in the U.S.However, the FSP participation rate of refugeesdeclines with the number of years since their arrival inthe U.S., whereas this decline does not occur amongnonrefugee immigrants. The FSP participation rate ofrefugees is more sensitive to the economic climatethan that of other immigrants or of U.S. citizens. Theauthors also found differences in program use by citi-zenship status. Immigrants who opt for citizenship aremore likely to participate in welfare programs thanthose who do not.

Even though FSP participation fell 37 percent between1994 and 1997, welfare reform does not appear tohave had the unintended consequence of reducing FSPparticipation among refugees. The study resultssuggest that food stamp use among refugees is prima-rily explained by their response to favorable economicconditions, rather than to welfare reform.

The authors found that the usual approach to meas-uring refugee status leads to a substantial underesti-mate of the effect of refugee status on participation inthe FSP. Additionally, the failure to account forresponse error in program participation causes anunderestimate of the effects of all variables onparticipation. The methods used in this study can beapplied in research on program participation, tocorrect for these measurement problems and toensure that research findings will be useful to policyanalysis.

8Subsequent legislation in 1997 allowed legal immigrants whowere disabled, elderly, or children living in the United States inAugust 1996 to regain eligibility for the Food Stamp Program.More recently, the 2002 Farm Act restored food stamp eligibility tolegal noncitizens who have lived in the United States for at least 5years or who are children or disabled, regardless of how long theyhave lived in the United States.


Food Stamp Receipt by Families WithNoncitizen Household Heads in RuralTexas Counties

Steve White, Xiuhong You, and Steve H. Murdock,Texas A&M University

Tami Swenson, University of Minnesota

Contact:Steve White, Assistant Research ScientistTexas A&M UniversityDepartment of Rural SociologyCollege Station, TX 77843-2125Phone: [email protected]


The 1996 welfare reform legislation eliminated theeligibility of most legal immigrants to receive foodstamps, although it did make exceptions based on alegal immigrant’s refugee status, work history, or U.S.veteran status. Subsequent legislation in 1997 allowedlegal immigrants who were disabled, elderly, or chil-dren living in the United States in August 1996 toregain eligibility for the Food Stamp Program (FSP).More recently, the 2002 Farm Act restored food stampeligibility to legal noncitizens who have lived in theUnited States continuously for at least 5 years or whoare children or disabled, regardless of how long theyhave lived in the United States.

In 2000, noncitizens made up almost 10 percent of thepopulation in Texas. Noncitizens were more than twiceas likely as citizens to live in poor households.Although noncitizens in Texas are concentrated inmetropolitan areas, about 5 percent of people innonmetropolitan counties are noncitizens. Because ofthe large noncitizen population in Texas, the elimina-tion of the eligibility of most noncitizens to receivefood stamps would be expected to have a large impacton the State’s FSP caseload. This study examined thedecline in the number of households in Texas headedby a noncitizen that participated in the FSP between1995 and 2001 and the factors that contribute to thedecline. It also compared the decline in the noncitizenFSP caseload in metropolitan counties with the declinein nonmetropolitan counties.

The study found that the number of FSP households inTexas declined 45 percent between 1995 and 2001,while the number of FSP households headed by anoncitizen declined 72 percent. The proportion ofelderly household heads among noncitizen FSP house-holds increased from 10 percent in 1995 to almost 25percent in 2001, a trend that is consistent with therestoration of FSP eligibility to elderly noncitizens.

The authors used monthly FSP administrative caseloaddata to examine the factors associated with the declinein the noncitizen FSP caseload. While the eligibilityrestrictions contribute to much of the decline, theauthors find that other factors, such as demographiccharacteristics and program changes, also contribute tothe decline. Noncitizen FSP households are less likelyto leave the program when the household is larger, andwhen the head is older, female, and has low levels ofeducation and income. The study results indicate thatthese demographic characteristics have a slightlystronger effect on citizen FSP households than onnoncitizen FSP households. Noncitizen FSP house-holds that live in a nonmetropolitan county are lesslikely to leave the FSP than those that live in metro-politan counties, and metropolitan status has a strongereffect on noncitizen FSP households than it does oncitizen FSP households. The authors also examinedwhether the frequency with which households mustrecertify their eligibility for the FSP has a differentialeffect on noncitizen FSP households. Regardless ofcitizenship status, the more frequently FSP householdsmust recertify eligibility, the more likely they are toexit the program. However, the effect of a morefrequent recertification policy is stronger on citizenFSP households than on noncitizen FSP households.

The study results indicate that restrictions placed on FSPeligibility of noncitizens are responsible for a largeshare of the decline in their use of food stamps, but thatother factors, such as increasing income, also contributedto the decline. This finding implies that noncitizensrespond to changes in economic conditions in wayssimilar to citizens. One notable difference is that resi-dence in a nonmetropolitan county reduces the proba-bility that a noncitizen FSP household leaves theprogram by more than it reduces the probability that acitizen FSP household leaves the program. This suggeststhat research focus on whether county-level programadministration or economic conditions have differenteffects on citizen and noncitizen FSP households.

Food Assistance and Nutrition Research Small Grants Program

Documents

Transcript of Food Assistance and Nutrition Research Small Grants Program