• TABLE OF CONTENTS
HIDE
 Copyright
 Main














Title: Impact of the food stamp and expanded food and nutrition education programs on food expenditure and nutrient intake of low income rural Florida househ
CITATION THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00026901/00001
 Material Information
Title: Impact of the food stamp and expanded food and nutrition education programs on food expenditure and nutrient intake of low income rural Florida househ
Physical Description: Book
Creator: Neenan, P. H.
Publisher: Agriculture Experimental Stations, Institute of Food and Agricultural Sciences, University of Florida
Publication Date: 1978
Copyright Date: 1978
 Record Information
Bibliographic ID: UF00026901
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: aca1949 - LTUF
05346572 - OCLC
000363374 - AlephBibNum

Table of Contents
    Copyright
        Historic note
    Main
        Page ii
        Page iii
        Page iv
        Page v
        Page vi
        Page vii
        Page viii
        Page ix
        Page x
        Page xi
        Page xii
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
        Page 75
        Page 76
        Page 77
        Page 78
        Page 79
        Page 81
Full Text





HISTORIC NOTE



The publications in this collection do
not reflect current scientific knowledge
or recommendations. These texts
represent the historic publishing
record of the Institute for Food and
Agricultural Sciences and should be
used only to trace the historic work of
the Institute and its staff. Current IFAS
research may be found on the
Electronic Data Information Source
(EDIS)

site maintained by the Florida
Cooperative Extension Service.






Copyright 2005, Board of Trustees, University
of Florida







December 1978 Bulletin 805



Impact of the Food Stamp and
Expanded Food and Nutrition Education
Programs on Food Expenditure and
Nutrient Intake of Low Income
Rural Florida Households

P. H. Neenan and C. G. Davis

















Agricultural Experimental Stations
Institute of Food and Agricultural Sciences
University of Florida, Gainesville
F. A. Wood, Dean for Research












Impact of the Food Stamp and
Expanded Food and Nutrition Education
Programs on Food Expenditure and
Nutrient Intake of Low Income
Rural Florida Households











Impact of the Food Stamp and
Expanded Food and Nutrition Education
Programs on Food Expenditure and
Nutrient Intake of Low Income
Rural Florida Households


P. H. Neenan and C. G. Davis
Food and Resource Economics Department
University of Florida, Gainesville








Contributing project to Florida Agricultural Experiment Station
Project AS-01629-"Incidence and Causes of Rural Poverty and
Economic Benefits of Poverty Programs"








AUTHORS
P. H. Neenan was formerly Research Assistant, Food and Resource Eco-
nomics Department, University of Florida. C. G. Davis is an Associate Pro-
fessor, Food and Resource Economics Department, University of Florida.
















CONTENTS


ACKNOWLEDGMENTS ................................. x
ABSTRACT ....................... .................... xi

CHAPTER
I INTRODUCTION ................................... 1
Statement of the Problem ............................ ... 1
Objectives .................. ........ .... ............. 2
Data Selection ............ ........................... 2

II BACKGROUND INFORMATION ....................... 4
The Food Stamp Program .................................... 4
The Expanded Food and Nutrition Education Program ........ 7
Eco-nutrition Studies ................................... 7

III THEORETICAL MODEL OF CONSUMER DEMAND ...... 10
Traditional Demand Theory .............................. 10
Lancaster's Consumer Goods Characteristics Approach ........ 10
Becker's Household Production Function Approach ........... 11
Socioeconomic Determinant Approach ...................... 11
Theoretical Framework for Evaluating Food Stamp
Participation ........................................ 14
Theoretical Model Specification ............................ 16
Data Base and Sampling Procedure ........................ 18

IV STATISTICAL MODELS .............................. 22
Total Food Expenditure Model ............. .............. 22
Food Group Expenditure Models .......................... 23
Nutrient Adequacy Models ............................... 25

V ANALYSIS AND RESULTS ........................... 28
Total Food Expenditure ................................. 28
Food Stamp Participants ............................. 28
Eligible Nonparticipants ............................... 31
Food Group Expenditures .................... .......... 32
Meat and Protein Product Expenditures (FGE1) ........... 33
Dairy Product Expenditures (FGE2) ..................... 35
Fruit and Vegetable Expenditures (FGE3) ................ 36
Bread and Grain Expenditures (FGE4) ................... 37

v








Nutrient Adequacy Ratios ................................ 38
Protein Adequacy (NAR1) ........................... 38
Calcium Adequacy (NAR2) ............................ 40
Iron Adequacy (NAR3) ............................... 41
Vitamin A Adequacy (NAR4) .......................... 42
Vitamin C Adequacy (NARS) .......................... 43

VI POLICY IMPLICATIONS, SUMMARY, AND CONCLUSIONS 44
Policy Implications ............. ....................... 44
Summary and Conclusions ................................ 54

GLOSSARY OF TECHNICAL TERMS ........................ 58
APPENDIX ............................................... 59
REFERENCES ............................................. 77









































vi

















LIST OF TABLES


Table Page
2.1 Food stamp coupon allotments and purchase requirements, by
family size and net income, January 1976 ................... 5
2.2 Proposed maximum net monthly income standards for food stamp
participants, by poverty levels, United States, 1976 ............ 7
2.3 Selected characteristics, Expanded Food and Nutrition
Education Program participants, Polk County and State of
Florida, December 1975 ................................. 8
3.1 Recommended Dietary Allowances, United States,
revised 1973 ........................................... 13
5.1 Group means for selected variables, food stamp participants
and eligible nonparticipants, Polk County, Florida, 1976 ....... 29
5.2 Statistical summary equation [4.1], GLS coefficients, dependent
variable, total food expenditure by food stamp participants,
TFEFs, Polk County, Florida, 1976 ........................ 30
5.3 Statistical summary equation [4.2], OLS coefficients, dependent
variable, total food expenditure by eligible nonparticipants,
TFEnon,,S, Polk County, Florida, 1976 ..................... 31
5.4 Group means for monthly food group expenditures, by race and
program participation group, Polk County, Florida, 1976 ...... 33
5.5 Selected group means, by race and program participation
group, Polk County, Florida, 1976 ......................... 34
5.6 Group means for nutrient adequacy ratios, by program
participation group, Polk County, Florida, 1976 .............. 39
6.1 Alternative policy specifications, total food expenditure,
Polk County, Florida .................................... 45
6.2 Projected monthly total food expenditures, by alternative
food policy, by four person, rural nonfarm, nonwhite
households, Polk County, Florida ......................... 45
6.3 Alternative policy specifications, nutrient adequacy ratios ...... 48
6.4 Summary of MPEj and MPEg, for food group expenditures, by
program participation group, Polk County, Florida, 1976 ....... 56









vii

















APPENDIX TABLES


Table Page
A-I Statistical summary equations [4.3, 4.4, 4.5, 4.6], OLS
coefficients, dependent variable, food group expenditures meat
and protein products, FGE1, Polk County, Florida, 1976 ..... 61
A-2 Statistical summary equations [4.3, 4.4, 4.5, 4.6], OLS
coefficients, dependent variable, food group expenditures dairy
products, FGE2, Polk County, Florida, 1976 ................ 63
A-3 Statistical summary equations [4.3, 4.4, 4.5, 4.6], OLS
coefficients, dependent variable, food group expenditures fruit
and vegetable products, FGE3, Polk County, Florida, 1976 .... 64
A-4 Statistical summary equations [4.3, 4.4, 4.5, 4.6], OLS
coefficients, dependent variable, food group expenditures, bread
and grain products, FGE4, Polk County, Florida, 1976 ....... 65
A-5 Statistical summary equations [4.7, 4.8, 4.9, 4.10], OLS
coefficients, dependent variable, nutrient adequacy ratio
protein, NAR1, Polk County, Florida, 1976 ................. 67
A-6 Statistical summary equations [4.7, 4.8, 4.9, 4.10], OLS
coefficients, dependent variable, nutrient adequacy ratio
calcium, NAR2, Polk County, Florida, 1976 ................ 68
A-7 Statistical summary equations [4.7, 4.8, 4.9, 4.10], OLS
coefficients, dependent variable, nutrient adequacy ratio
iron, NAR3, Polk County, Florida, 1976 ................... 70
A-8 Statistical summary equations [4.7, 4.8, 4.9, 4.10], OLS
coefficients, dependent variable, nutrient adequacy ratio
vitamin A, NAR4, Polk County, Florida, 1976 .............. 72
A-9 Statistical summary equations [4.7, 4.8, 4.9, 4.10], OLS
coefficients, dependent variable, nutrient adequacy ratio
vitamin C, NAR5, Polk County, Florida, 1976 .............. 74













viii

















LIST OF FIGURES


Figure Page
1 Hypothetical income, consumption, and budget constraint
relationships, four person Food Stamp Program households .... 15
2 Effect of alternative policies on income, consumption,
and budget constraint relationships ........................ 46
3 Projected nutrient adequacy protein, NARI, with
alternative policies, Polk County, Florida ............. ... .. 49
4 Projected nutrient adequacy calcium, NAR2, with
alternative policies, Polk County, Florida ................... 50
5 Projected nutrient adequacy iron, NAR3, with
alternative policies, Polk County, Florida ................... 51
6 Projected nutrient adequacy vitamin A, NAR4, with
alternative policies, Polk County, Florida ................... 52
7 Projected nutrient adequacy vitamin C, NAR5, with
alternative policies, Polk County, Florida ............... .. 53



























ix

















ACKNOWLEDGMENTS


Appreciation is extended to L. Polopolus, R. D. Emerson, L. H. Myers,
W. W. McPherson, and S. Lane for important technical contributions. Special
appreciation is extended to J. Cameron, Extension Home Economist, Polk
County, for assistance with data collection. Appreciation is also extended to
anonymous Florida Agricultural Experiment Station reviewers for their con-
structive comments. The authors, however, bear responsibility for any errors
or deficiencies in the report.



































x















ABSTRACT

Impact of the Food Stamp and Expanded Food and
Nutrition Education Programs on Food Expenditure and
Nutrient Intake of Low Income Rural Florida Households
P. H. Neenan and C. G. Davis


This report examines the socioeconomic determinants of food expen-
diture levels for Food Stamp Program (FSP) participants and eligible
nonparticipants in a rural area of Florida. It also examines food nutrient
intake responses associated with participation in the Food Stamp Pro-
gram and/or the Expanded Food and Nutrition Education Program
(EFNEP). Responses were assessed in terms of impact on four alterna-
tive food group expenditure levels for meat and protein, dairy product,
fruit and vegetable, and bread and grain product; and their associated
nutrient intake levels for protein, vitamin A, vitamin C, calcium, and
iron.
The study utilized data obtained from 1976 records of the Expanded
Food and Nutrition Education Program (EFNEP) participants in Polk
County, Florida. The sample was stratified into food program partici-
pating and nonparticipating groups.
Results of regression analysis suggested that income and family size
explained significant variation in total food expenditures among both
food stamp participants and eligible nonparticipants. The Marginal Pro-
pensity to Expend money income (MPEI) for food stamp households,
when calculated at group means, was 0.06. For every one dollar increase
in money income, total food expenditures increased by six cents. In
contrast, the MPE, for eligible nonparticipants was 0.135. The bonus
coefficient and coefficients for interactions between bonus value and both
income and family size were also significant for FSP households. The
food stamp bonus value was more efficient in increasing food expendi-
tures at larger family size and lower income levels. For FSP participants,
group mean income and family size, the Marginal Propensity to Expend
from bonus (MPEB) was 0.45. The positive sign of the bonus-family
size interaction term suggested that as family size increased, an increas-

xi








ing proportion of the discretionary income freed by Food Stamp Pro-
gram participation was spent to purchase additional food.
The food stamp bonus was also positively related to increased expen-
ditures for grain, fruit, vegetable, dairy, and meat products. However,
the bonus response for meat products was clearly the largest coefficient.
Increases in bonus value were not significantly related to increases in
nutrient adequacy. The first year of participation in the Expanded Food
and Nutrition Education Program was effective in increasing the nutrient
adequacy ratios of vitamin A, vitamin C, and iron.
Results indicated that participation in the FSP did increase purchas-
ing power. However, results also suggested that participation in both the
Food Stamp Program and the Expanded Food and Nutrition Education
Program may have been a more effective means of improving the nutri-
tional status of low income households.





























KEY WORDS
Food Stamp Program, Expanded Food and Nutrition Education Program,
food expenditures, low income households, Nutrient Adequacy Ratio, bonus
value, nutritional status, policy impact.

xii












Chapter I
INTRODUCTION


With the tendency for proliferation of federal antipoverty programs
to subsidize food consumption of low income households and the in-
creased emphasis on program accountability, it has become important to
identify not only the degree of increased purchasing opportunity, but
also modifications in nutrient intake patterns in participating households.
The Food Stamp Program (FSP) and the Expanded Food and Nutrition
Education Program (EFNEP) are two examples of national food pro-
grams designed to improve the nutritional status of low income house-
holds.
The Food Stamp Program operates under the assumption that mal-
nutrition and undernutrition are partially income-induced [32]. It has
been noted that persons living at or near the designated poverty income
thresholds or levels are less able to purchase nutritionally adequate diets
than higher income households of similar composition [1, 5, 23]. In con-
trast to the income supplementation mechanism of the FSP, the Ex-
panded Food and Nutrition Education Program (EFNEP) is an educa-
tional program. The program was designed to improve the nutritional
knowledge, food economy expertise, and food preparation skills of low
income households. The program utilizes indigenous paraprofessional
nutrition aides working on a one-to-one basis with program participants.
A basic assumption of the EFNEP is that improved knowledge of the
nutrient composition of foods, coupled with food economy principles,
will result in long run improvement in nutrient intake of program par-
ticipants [10, 18].


STATEMENT OF THE PROBLEM
Despite some attempts to evaluate the distributional effects of the
Food Stamp Program, most program evaluations have tended to focus
on coupon production, distribution, program monitoring, and fraud con-
trol. A limited number of studies [17, 21, 33], have examined changes
in food expenditures, changes in the composition of the purchased com-
modity bundle, or the modification of nutritional status of participating
households. Also, relatively little research has been done on the joint

1








effects of income supplement programs, (such as food stamps) and the
level of nutrition education on the food purchasing behavior and nutri-
tional status of participating households [10]. This study was undertaken
in an attempt to fill the knowledge gap in some of these areas. It is hoped
that relevant findings will be incorporated into future antipoverty pro-
grams and food and nutrition policies at the state and national levels.


OBJECTIVES
The general objective of this report is to identify consumer behavioral
responses associated with participation in the Food Stamp Program
(FSP) and/or the Expanded Food and Nutrition Education Program
(EFNEP). The specific objectives are to:
(1) Identify factors affecting total food expenditure levels for FSP
participants and eligible nonparticipants.
(2) Determine if food stamp supplementation affects food group ex-
penditures differently than would direct income supplement.
(3) Determine the extent of food stamp supplementation and/or
participation in the Expanded Food and Nutrition Education
Program on consumption of protein, vitamin A, vitamin C, cal-
cium, and iron.


DATA SELECTION
Two federal antipoverty programs were considered in this analysis. It
was hypothesized that participation in the Food Stamp Program (FSP)
and/or the Expanded Food and Nutrition Education Program (EFNEP)
would affect total household food expenditure levels, food group expen-
ditures, and intake of protein, vitamin A, vitamin C, iron, and calcium.
Federal regulations precluded identification of food stamp participants
and related income, demographic, and socioeconomic information from
Food Stamp Program administrative files. Individual household informa-
tion regarding EFNEP participants was also unavailable at the statewide
level. A smaller sample area, Polk County, Florida, was selected as the
study area for this analysis. Polk County is relatively representative of
predominantly rural, low income areas of the state of Florida. Accord-
ing to 1970 census data, over 39 percent of the county's population was
classified as rural, and over 15 percent was classified below the desig-
nated poverty levels [30].
Data concerning household income, food expenditures, socioeconomic
variables, food program status, and food intake during a 24-hour period
were collected by means of survey records at six month intervals for
households participating in EFNEP. Program survey records collected

2








during the spring of 1976 were selected for use in this study. Four pro-
gram participation stratifications were designated. Group one, (FS
EFNEP), consisted of households that participated in both the Food
Stamp Program and the Expanded Food and Nutrition Education Pro-
gram. Group two, (FS non-EFNEP), consisted of households that used
food stamps but had not participated in EFNEP. Group three, (non-FS
EFNEP), households that did not use food stamps but were EFNEP
participants. Group four, (non-FS non-EFNEP), did not participate in
either the FSP or EFNEP.
Two explanatory notes regarding data selection are necessary. House-
holds designated as non-EFNEP households, group two (FS non-
EFNEP) and group four (non-FS non-EFNEP), were classified as such
because at the time of the survey they had not yet received any nutrition
education training. At the first contact between an EFNEP aide and a
prospective participating household, an initial 24-hour food intake report
and family data record were completed. These initial contact records
were the source of the designated non-EFNEP data. Households selected
for the non-food stamp strata, group three (non-FS EFNEP) and group
four (non-FS non-EFNEP), were all eligible by income standards for
food stamp participation. Households with incomes above food stamp
eligibility thresholds were not considered in this analysis.
Regression analysis was used to estimate the relative impact of tradi-
tional economic and food program variables on total food expenditures,
food group expenditures, and nutrient adequacy ratios for protein, vita-
min A, vitamin C, calcium, and iron. Variables external to the consumer
decision making process, such as income, age, or ethnic background,
were included in the analysis. Differences in food purchasing behavior
may also depend on variations in the cognitive or learning levels. There-
fore, participation in the Expanded Food and Nutrition Education Pro-
gram was included to account for differences in the length and type of
nutrition education.















3














Chapter II

BACKGROUND INFORMATION


THE FOOD STAMP PROGRAM
The first Food Stamp Program was implemented during the adminis-
tration of Franklin D. Roosevelt but was discontinued early in World
SWar II [27]. In 1961, through an executive order, John F. Kennedy
reintroduced the Food Stamp Program into selected pilot counties as an
alternative to direct surplus commodity distribution. The intent of the
"original program was to provide sufficient purchasing power for low in-
come households to select and purchase foods providing nutritionally
adequate diets within normal trade channels. The Food Stamp Act of
1964 (P. L. 88-525, August 31, 1964) mandated two program objec-
tives [321:

(1) to safeguard the health and well being of the nation's population
and raise levels of nutrition among low income households.
(2) to promote the distribution in a beneficial manner of our agricul-
tural abundance that will strengthen the agricultural economy.

During initial program years, national priorities directed program em-
phasis towards the utilization of surplus agricultural commodities, with
less emphasis on the health objective. However, a growing national
awareness of domestic hunger and malnutrition resulted in an effort to
improve the nutritional status of low income households. Eligibility cri-
teria were liberalized and program benefits increased. In 1973, when
the program became mandatory in all counties, the Commodity Distribu-
tion Program (CDP) was almost entirely phased out.
The basic participation unit of the Food Stamp Program is the house-
hold. A household may consist of any person or group of persons who
purchase, store, and prepare food. They must have cooking facilities.
Program eligibility is based on net household income, total assets, and
household size. Eligible households purchase coupons which are used in
retail food outlets. Households of equal size receive coupons of equal
purchasing value, but the cash purchase requirement varies with net in-
come. Table 2.1 lists coupon allotments and purchase requirements
effective January 1, 1976. The bonus value is defined as the difference

4








between the purchase requirement and the coupon allotment value. For
Food Stamp Program purposes, the Thrifty Food Plan replaced the
Economy Food Plan in 1975. The Thrifty Food Plan has its own inde-
pendent market basket of goods based on preferences of households
purchasing at the level of the Low Cost Plan in the 1965-66 Household


TABLE 2.1.-FOOD STAMP COUPON ALLOTMENTS AND PURCHASE REQUIRE-
MENTS, BY FAMILY SIZE AND NET INCOME, JANUARY 1976.

48 States and D.C. Number of persons in household

1 2 3 4 5 6 7
Monthly coupon allotment

$50 $92 $130 $166 $198 $236 $262
Monthly net income Monthly purchase requirement

-- - - - --dollars - - -
$ 0to 19.99 0 0 0 0 0 0 0
20 to 29.99 1 1 0 0 0 0 0
30 to 39.99 4 4 4 4 5 5 5
40 to 49.99 6 7 7 7 8 8 8
50 to 59.99 8 10 10 10 11 11 12
60 to 69.99 10 12 13 13 14 14 15
70 to 79.99 12 15 16 16 17 17 18
80 to 89.99 14 18 19 19 20 21 21
90 to 99.99 16 21 21 22 23 24 25
100 to 109.99 18 23 24 25 26 27 28
110 to 119.99 21 26 27 28 29 31 32
120 to 129.99 24 29 30 31 33 34 35
130 to 139.99 27 32 33 34 36 37 38
140 to 149.99 30 35 36 37 39 40 41
150 to 169.99 33 38 40 41 42 43 44
170 to 179.99 38 44 46 47 48 48 50
190 to 209.99 38 50 52 53 54 55 56
210 to 229.99 40 56 58 59 60 61 62
230 to 249.99 62 64 65 66 67 68
250 to 269.99 68 70 71 72 73 74
270 to 289.99 72 76 77 78 79 80
290 to 309.99 72 82 83 84 85 86
310 to 329.99 88 89 90 91 92
330 to 359.99 94 95 95 97 98
360 to 389.99 102 104 105 106 107
390 to 419.99 111 113 114 115 116
420 to 449.99 112 122 123 124 125
450 to 479.99 131 132 133 134
480 to 509.99 140 141 142 143

5








TABLE 2.1.-continued

48 States and D.C. Number of persons in household

1 2 3 4 5 6 7
Monthly coupon allotment
$50 $92 $130 $166 $198 $236 $262
Monthly net income Monthly purchase requirement
- - - dollars - - -
$510 to 539.99 142 150 151 152
540 to 599.99 142 159 160 161
570 to 599.99 168 169 170
600 to 629.99 170 178 179
630 to 659.99 170 187 188
660 to 689.99 170 196 197
690 to 719.99 204 206
720 to 749.99 204 215
750 to 779.99 204 224
780 to 809.99 204 224
810 to 839.99 226
840 to 869.99 226

"aA revised coupon allotment and purchase requirement schedule will become
effective July 1, 1977. [Federal Register, 1977, Vol. 42, (85)].
Source: [Federal Register, 1976, Vol. 41 (90)].


Food Consumption Survey. It allows food purchases equal to about 80
percent of the average level of spending for food for use at home by all
U.S. consumers.
Early in 1976, the U.S. Department of Agriculture proposed several
/structural changes in the FSP. Proposals included a $100 standard de-
duction per household in place of the individualized schedule. An addi-
tional $25 deduction was proposed for each household member over 65
years of age. The official poverty income levels were proposed as pro-
gram eligibility cutoffs (Table 2.2). The purchase requirement of each
household was to be standardized at 30 percent of net income. Adverse
public reaction and a court injunction sought by over half of the states,
several church groups, and labor unions delayed implementation of the
proposed changes.
In March of 1977, President Carter also proposed structural changes
in the FSP. In a similar action to the 1976 proposal, it was suggested
that itemized deductions be replaced by a standard deduction of $75-
$80 per household. However, in contrast to the standardized 30 percent
purchase requirement of the 1976 proposal, no cash purchase require-
ment would be levied. The bonus value would be determined by sub-

6








tracing 30 percent of net income from the current coupon allotments.
Each household would receive coupons equal in value to the bonus
value.


TABLE 2.2.-PROPOSED MAXIMUM NET MONTHLY INCOME STANDARDS FOR
FOOD STAMP PARTICIPANTS, BY POVERTY LEVELS,
UNITED STATES, 1976.

Household Maximum net monthly
size income standards

-dollars-
1 233
2 308
3 383
4 458
5 533
6 608
7 683
8 758
Each additional person +75

Source: [Federal Register, 1976, Vol. 41 (90)].

THE EXPANDED FOOD AND NUTRITION EDUCATION PROGRAM

A different type of federal antipoverty effort was established with a
$10 million grant in November 1968. The Expanded Food and Nutrition
Education Program (EFNEP) was established to help families living in
or near poverty-especially those with young children-to acquire
knowledge, skills, and changes in behavior to achieve adequate diets
providing normal nutrition [18]. The Extension Service of the U.S. De-
partment of Agriculture, in cooperation with State Cooperative Exten-
sion Services, was charged with program implementation. The relatively
new approach utilizes paraprofessional indigenous program aides, work-
ing under the supervision of Extension Service home economists. Pro-
gram aides provide nutrition education and teach food purchasing econ-
omy, meal planning, and preparation on a one-to-one basis with program
homemakers and youths. Table 2.3 shows selected characteristics of
EFNEP participants during December 1975 for the study area, Polk
County, and the State of Florida.


ECO-NUTRITION STUDIES
During a recent symposium conducted by the American Dietetic As-
sociation entitled "Food: The Realities of Economics, Politics, and
Population," Paarlberg [26] noted that the inherent problem in national

7








TABLE 2.3.-SELECTED CHARACTERISTICS, EXPANDED FOOD AND NUTRITION
EDUCATION PROGRAM PARTICIPANTS, POLK COUNTY AND
STATE OF FLORIDA, DECEMBER 1975.

Characteristic Polk County State of Florida

--- --- percent - --
Residence
Urban 56.9 72.3
Rural 43.1 27.7
Racial or ethnic group
White 21.1 27.3
Black 77.2 68.3
Spanish .7 4.0
Other 1.0 .4
Food Stamp Participation 27.1 52.0
Homemaker age
Under 24 years 16.9 19.7
25 to 55 48.7 63.8
Over 55 34.4 16.5
Homemaker education
Less than grade 8 57.4 63.9
Over grade 8 42.6 36.1

Source: [Compiled from annual report, Polk County, EFNEP, 1975].

food and nutrition policy was that there was a multiplicity of conflicting
objectives to be maximized. Overall improvement in the level of nutri-
tion, elimination of income-related malnutrition, freedom of choice in
food selection, and efficient food production, processing, and merchan-
dising are all examples. As indicated earlier, a relatively small number
of studies have attempted to identify factors which affect food expendi-
tures and the nutritional status of the household within an economic
framework. A brief review of some of the more relevant findings for this
study follows.
Madden and Yoder [21] drew the following conclusions from a 1972
study of the distributional impact of the Food Stamp and Commodity
Distribution Programs in rural Pennsylvania:
(1) Low income family diets were most deficient in vitamin A and
calcium and least deficient in protein and phosphorous.
(2) Food stamp participation enhanced the diets of families only if
the family had not received income for a period longer than two
weeks. In this case, iron and thiamine intakes were higher.
(3) When families had received income within the previous two
week period, impact of the Food Stamp Program on the Mean
Adequacy Ratio (MAR)' was insignificant.

"See Glossary of Technical Terms for definition of MARs and NARs.

S8







(4) The amount of the food stamp bonus was not significantly related
to changes in the MAR.
(5) Nutritional efficiency with which additional food dollars were
spent declined as the families diets improved.
(6) Total value of foods purchased did not increase significantly with
participation in the Food Stamp Program.
(7) No significant difference in dietary adequacy could be attributed
to number of nutrition aide visits. However, the authors did note
that the sample size was too small for definitive conclusions for
this aspect.
In a 1974 California study, Lane [17] found that diets of participants
in the Food Stamp Program were nutritionally superior to participants in
the Commodity Distribution Program and comparable nonfood program
participating low income families. Significant improvements in the NAR
occurred for calories, calcium, protein, thiamine, and riboflavin for food
stamp participants. Ethnicity and urbanity were consistently significant
variables affecting the Nutrient Achievement Ratios. Therefore, changes
in dietary status associated with food program status differed according
to ethnic group, residence, size of family, and family composition.
Adrian's 1974 study [1] utilized data collected in the 1965-66 House-
hold Food Consumption Survey to evaluate household nutrient consump-
tion throughout the income spectrum. He found that, with the exception
of carbohydrate and vegetable protein, income elasticities increased
throughout the lower income levels until the $8,000 annual income level
was reached and then declined at higher income levels. Two additional
interesting findings were related to socioeconomic factors. First, for
households sampled in the South, black families consumed significantly
less of all nutrients than did white households, with the exception of
vitamin A, vitamin C, and thiamine. Second, families of homemakers
with high school education consumed significantly less carbohydrate and
thiamine, but more vitamin A and vitamin C than did families where the
homemaker had less than grade school education.
Swanberg and Shipley [29] reported in a 1975 study on the nutritional
status of rural families in Colombia. They concluded that certain nutrient
deficiencies may be reduced by increasing incomes and consequently
food expenditures of rural Colombian families. However, other types of
nutritional deficiencies were not responsive to income supplementation.
They concluded that for deficiencies of vitamin A, calcium, and ribo-
flavin, nutrition education designed to modify food consumption patterns
could be an alternative solution.






9













Chapter III
THEORETICAL MODEL OF
CONSUMER DEMAND


TRADITIONAL DEMAND THEORY
Traditional consumer demand theory revolves around the existence
of a continuous utility function. Given the traditional axioms of utility
theory, the consumer maximizes his utility function within the limitations
of a budget constraint.
While specific functions are unobservable, utility theory provides a
conceptual framework within which observed consumer behavior in the
market can be described and forecasts made for future behavior. Esti-
mation of demand functions by (a) specification of the utility function,
(b) imposition of the theoretical constraints to specified relationships,
(c) the total differential approach, or (d) separability of utility func-
tion, are current approaches in applied analysis.
Separability of the utility categories is a treatment pioneered by Strotz
[28], Gorman [13], and Houthakker [14]. Commodities were grouped
into several subsets of close complements or substitutes, but the groups
are not necessarily assumed independent of changes in other groups. A
necessary condition for grouping is that the utility function is not changed
by grouping. Strotz [28], Gorman [13], and Blackorby, et al., [4] dem-
onstrate that the necessary and sufficient conditions for the existence of
group price indices and subsequent budget allocation by these indices
and income require either strongly separable groups or weakly separable
homethetic functions.


LANCASTER'S CONSUMER GOODS CHARACTERISTICS APPROACH
In the Lancaster approach to consumer theory [16], characteristics of
goods directly provide utility while market goods are the means of sup-
plying characteristics. The utility function is defined in characteristics
space rather than in goods space. Characteristics are related to goods
by a linear technology matrix. Different characteristics can contribute to
different levels of satisfaction or utility. Therefore, the traditional limita-
tion of a single value utility function is overcome, as the Lancaster ap-

10








proach allows a more general multidimensional function. Consumption
technology is considered as a concave production transformation sur-
face, and acts as one of the utility maximization constraints. This extends
the framework in which consumer choice is made, as the effects of adver-
tising and product changes can be considered as changing consumption
technology rather than affecting the utility function.
BECKER'S HOUSEHOLD PRODUCTION FUNCTION APPROACH
In Becker's household production model [3], goods purchased through
the market system are not the arguments of the utility function. Rather,
these goods are considered inputs in the production of the commodities
which do directly enter the preference structure. He argues that this
formulation minimizes use of explanatory variables such as residence
location or other socioeconomic proxies traditionally included to account
for intra-sample variation.
An example of household production might be the production of fam-
ily health. Medical services alone do not provide direct utility. Rather, a
combination of diet, medical services, heredity, and a number of other
variables are conceptualized as inputs in the production of various levels
of health. In the production function framework [3], the level of educa-
tion would be introduced as an input in the health production function.
Becker argues that an educated person may be able to produce a given
level of health more efficiently with relatively less medical care because
of greater knowledge of the role of nutrition or the benefits of exercise.
More generally, education enters the demand function not as a demand
shifter, but rather as a factor which alters the efficiency of health pro-
duction.
SOCIOECONOMIC DETERMINANT APPROACH
Despite attempts to explain variation among consuming units by the
efficiency concept of Becker or the consumption technology framework
of Lancaster, additional unexplained consumption variation exists be-
tween households and over time within the same household.
The ability of a household to obtain foods in sufficient variety and
quantity to provide nutrients to maintain normal health is determined
in part by household income, price of food and prices of other commodi-
ties necessary to fulfill the remaining "primary needs."2 Availability of
foods from sources outside the retail market will expand the household's
access to food. By definition, participation in a food aid program such as

2According to Maslow [22] there is an orderly sequence in which human
needs are fulfilled. Food, clothing, and shelter are the primary elements of
the need hierarchy and are fulfilled before other needs such as position se-
curity, peer recognition, or self-fulfillment.

11








Food Stamps, School Breakfast, School Lunch, or Congregate Meals for
the Elderly will increase the household's access to food.
In the traditional demand model, it is assumed that all households
face the same level of market prices. In reality, not only may prices vary
among types of retail establishments, but the set of marketing alterna-
tives may differ for specified groups of households. Regional variations
are not relevant in this study, since all observations are within the same
geographic area. Urbanization, however, is considered since such a vari-
able may pick up access to nonmarketed foods, i.e., those grown in home
gardens or on the farm. Market prices faced by households of differing
residence location may differ because of variations in the size and prox-
imity of grocery stores. Larger chain grocery stores may be able to offer
lower prices and larger selection because of economies of scale in pur-
chasing, distribution, and management. Urban residents would tend to
benefit not only by the lower prices but also by proximity, as transpor-
tation costs could be minimal or an automobile entirely unnecessary for
food purchasing. In contrast, residents in rural farm or nonfarm areas
would either patronize smaller owner-managed local stores or need ac-
cess to transportation to travel to the lower priced urban chain stores.
In the case of the latter, the differences in prices between urban and
rural residents would be equal to the difference in transportation costs.
The employment status of the homemaker is another factor respon-
sible for variation in food consumption behavior between households.
A spectrum of food related services can be purchased in conjunction
with the basic commodity. For example, for additional costs, whole
fryers can be cut into quarters, or sold in packages of thighs or breasts.
The chicken can also be cooked and ready for consumption at the time
of purchase. The employed homemaker has a different value of time for
food preparation than the unemployed homemaker. If the value of the
employed homemaker's time, represented by the wage rate, is greater
than the cost of purchasing food related services, food expenditures by
these households would be expected to be greater than those of house-
holds with unemployed homemakers.
Two related situations may also affect at-home food expenditures. The
working homemaker may eat fewer meals at home, thereby decreasing
at-home food expenditures. In a parallel fashion, children participating
in the school breakfast or lunch program, or retired senior citizens eat-
ing at a Congregate Meals for the Elderly site, would be expected to
decrease the amount of expenditure for at-home consumption.
Family size and composition are expected to affect both the quantity
of at-home food expenditures and the composition of the purchased
bundle of goods. When income and food expenditures are expressed on
a per capital basis, larger family per capital expenditures may differ from
those of smaller families, largely as a result of economies of scale in food

12








purchase. Lack of storage space or spoilage rates may prevent smaller
households from purchasing in larger quantities. Economies of size dif-
ferences between households can be represented by a family size variable.
However, family composition may also affect the quantity and quality
of the purchased food bundle. Individual differences in food and nutrient
requirements are based on age, sex, height, weight, and activity levels.
The Recommended Dietary Allowances (RDA), 1973 revision, estab-
lishes nutrient levels adequate to maintain health in essentially all healthy
persons. The allowances are not minimum requirements, but rather, are
designed to allow for individual variation. An adaptation of the RDA
is given in Table 3.1. The table illustrates how different quantities and
types of foods would be required for households containing two pre-
school age children versus two teenage boys.


TABLE 3.1.-RECOMMENDED DIETARY ALLOWANCES, UNITED STATES,
REVISED 1973.

Vitamin Vitamin
Protein A C Calcium Iron
Years (g) (IU) (mg) (mg) (mg)

Infants 0-6 mo. kgx2.2 1400 35 360 10
6-12 mo. kgx2.0 2000 35 540 15
Children 1-3 23 2000 40 800 15
4-6 30 2500 40 800 10
7-10 36 3300 40 800 10
Males 11-14 44 5000 45 1200 18
15-18 54 5000 45 1200 18
19-22 52 5000 45 800 10
23-50 56 5000 45 800 10
51+ 56 5000 45 800 16
Females 11-14 44 4000 45 1200 18
15-18 48 4000 45 1200 18
19-22 46 4000 45 800 18
23-50 46 4000 45 800 18
51+ 46 4000 45 800 18
Pregnancy +30 5000 60 1200 18+
Lactation +20 6000 60 1200 18

Source: [Adapted from National Research Council, National Academy of Sciences.
1974].


A number of problems arise when generalizing the behavior of the
individual consumer to that of many households. In any given household,
no single individual consumes or even purchases all food items. How-
ever, for at-home consumption, Lewin [19] has suggested the gatekeeper
theory. Food expenditures are attributed to one person, generally the

13








female homemaker. Decisions regarding food purchases may be affected
by specific family members' desires, but the homemaker is the ultimate
decision maker. This person purchases and prepares food for the family.
Thus, through knowledge of different types of foods and methods of
preparation, and knowledge of nutrition, family food habits are formed.
Therefore, the age, education, and motivation of this "gatekeeper" will
be the key factors in the establishment of food habits for the entire
household.
Food provides much more than just the biogenic needs of most peo-
ple. It is a manifestation of culture, a means of demonstrating affection
within the family unit, or a means of attaining social stature. Since food
habits are developed over a lifetime, older persons may be more resistant
to change. This may mean that older persons are less likely to purchase
newer products and food related services [6].
Education of the homemaker is often used to account for variation in
consumption of the purchased food bundle. Educational level may accu-
rately represent exposure to information about health and nutrition, but
in no way measures health attitudes or motivations. Effects of specific
types of educational program, such as the Expanded Food and Nutrition
Education Program (EFNEP), may be considered as a refinement of the
traditional educational variable. Health motivation, or a perception of
the relationship between nutrition and health, may affect food efficiency
and the consumption of the commodity bundle.

THEORETICAL FRAMEWORK FOR EVALUATING FOOD STAMP
PARTICIPATION
Analysis of consumer behavior in this study has conceptually incorpo-
rated two stage budget maximization, Becker's concept of the household
production function [3], and the characteristics approach to demand
theory [16]. It is hypothesized that household food expenditures are a
function of income, prices, food aid status, and a set of socio-economic
variables that condition tastes and preferences. The consumer is assumed
to maximize utility within the limitations of a budget constraint, with or
without food.stamps. However, the Food Stamp Program acts as an in-
kind income supplement and therefore affects initial group budget allo-
cations. Food expenditure decisions of food stamp households may have
three general outcomes. Recall that FSP supplementation requires a pur-
chase requirement based on income levels. The value of the bonus cou-
pons, i.e., face value of the food stamp coupons minus the purchase
requirement, is the real income supplement. The income constraint be-
comes a kinked constraint as represented in Figure 1. Figure 1 is an
hypothetical example of a four person household, with monthly income
of $200. The FSP coupon allotment for a family of this size is $166. A

14








purchase requirement of $60 is necessary to obtain $166 worth of food
stamp coupons [11]. Since the coupons can only be used to purchase
food, the new income constraint is represented by the kinked ACIG
rather than LG, which would be the constraint if a cash income supple-
ment is used. AP represents the food stamp purchase requirement. Vec-
tors a,, as and aS are income-consumption relationships and represent
the relative proportions of income that would be spent on food and non-
food items under three alternative preference situations.
Vector ai represents a household that normally spends less for food
than the food stamp purchase requirement. Depending on the preference
structure, the household may or may not choose to participate in the
FSP. Theoretically, households represented by indifference curves f1 and
#2 would choose to participate since higher levels of utility can be ob-
tained with participation. However, income-consumption vectors which



306 L


a2

E



o 200

C

140 P
"0 / N N
a D\ U' \ (r3
x 100
H
0)


F G
0 100 200 306

Dollar expenditures for food items

Figure 1.-Hypothetical income, consumption, and budget constraints rela-
tionships, four person food stamp program households.

15








cross AF at points closer to A will be less likely to participate in the
FSP. It is possible to have indifference curves that do not cross CI. In
these cases, higher levels of utility would not be reached with FSP par-
ticipation.
Vector as represents a household that normally spends more for food
than the food stamp purchase requirement, but less than the coupon
allotment. The difference between the amount usually spent for food and
the purchase requirement is freed or discretionary income. The preferred
position of J is unobtainable, since it is outside the feasible set of possi-
bilities. The dashed segment of LI indicate that these points are not at-
tainable under the two alternative budget regimes under consideration.
At point I, the food stamp coupon allotment, the entire discretionary
income is spent on nonfood items. The income effect is ie Io.'ii. sile for
portion DM, while MI is the in-kind food stamp effect. This solution is
not unique. The family may choose to spend an additional -irt;iii'olthe--
discretionary income for food, represedtted-by-t-he portion of the income
constraint between point I and point .V This third part of the subsidiza-
tion effect could be the result of a relative price decline for food com-
modities in relation to the price of nonfood items. Households normally
spending a smaller budget share for food than FS allotment would re-
alize a decline in their average budget share for food as real income
remain unchanged. Households with higher propensities to consume,
such as those normally spending more for food than food stamp allot-
Sment, will not necessarily realize a relative price effect. Higher levels of
utility can be reached with no more than the income effect, as expansion
from point E to H on vector as

THEORETICAL MODEL SPECIFICATION
Additional information regarding differences in food expenditures with
respect to FSP participation was desired in this study, so estimation by
the two stage maximization procedure was insufficient. While the two
stage procedure was not used operationally, it did permit conceptual
decomposition of the regression coefficient for bonus value into weighted
sums of effects attributable to the in-kind provision of the FSP and those
attributable to changes in relative prices. For this study, the general form
of the Engel type total food expenditure relationship was written:
[3.1] TFE = f(I, B, SE,)
where TFE = total food expenditures/household/month
I = household income/month
B = bonus value of food stamp supplement


16








SE, = socioeconomic variables to measure family size, composition,
ethnicity, urbanity, etc.

Empirical results of this type of relationship should provide insights
into the first set of study objectives regarding variation in the effect of
food stamp supplementation as differences occur in household income
level, household size, level of bonus supplementation, and family com-
position.
The second set of objectives sought to determine if food stamp sup-
plementation affected expenditures for groups of foods differently, i.e.,
expenditures for meat and protein foods versus dairy goods or fruits and
vegetables or grain products. Since no relative price difference between
food groups occurred with food stamp supplementation, the expected
response was parallel to the income response for each group. However,
nutrition education may have also significantly affected the food group
budget allocation. Therefore, to provide empirically useful answers to
these questions, the general form was written:
[3.2] FGE,, = f(I, B, SEr, MP)
where FGE, = food group expenditures
I, B, SE, are defined as in equation [3.1]
MP = months of participation in the Expanded Food and Nutrition
Education Program.

The third set of objectives sought to answer the question as to whether
participation in the Food Stamp and/or Expanded Food and Nutrition
Education Program affected consumption of five selected nutrients. The
consumer goods characteristics model enabled consideration of nutrient
composition as one of the characteristics of food. As previously dis-
cussed, food represents a number of characteristics to different persons,
but cross sectional analysis of the variation in nutrient intake levels
allows comparisons of nutrient utility demand between households. Nu-
trient consumption was defined in terms of a Nutrient Adequacy Ratio
(NAR). The NAR was defined for each nutrient as:

nutrient consumption/day
SRecommended Dietary Allowance (RDA)
and the functional relationship defined as:

[3.3] NAR, = f(I, B, FS, LC, NE, SEr, TFE)
NAR's were calculated for protein, iron, calcium, vitamin A, and vitamin
C. Nutrient adequacy was a function of the availability of food to the
household as measured by income (I), food program status (B), house-


17








hold (FS), and household composition (LC). Nutrition education (NE)
and other socioeconomic variables (SEr) were included to account for
additional variation. Two stage budget maximization implies that de-
mand for a commodity within its own grouping, in this case a charac-
teristic, depends on prices of goods within the grouping and the budget
share allotted to the group. Therefore, total food expenditure (TFE)
was included as an explanatory variable.

DATA BASE AND SAMPLING PROCEDURE
Food expenditure information is a normal part of each EFNEP six
month record. Actual food expenditure information was available for
nonparticipating households. In group 1 (FS EFNEP) and group 2 (FS
non-EFNEP), the food stamp purchase requirement is normally recorded
in place of total food expenditures. For FSP households, actual food ex-
penditure equals the value of the food stamp coupon allotment plus any
additional cash expenditure for food. An accurate estimate of the value
of food purchased by FSP households (groups 1 and 2) was obtained
through an additional questionnaire administered to these households at
Sthe time of the spring 1976 food recall. Food stamp families were asked
Show much, if any, additional money was spent for food above the nor-
Smal food stamp coupon allotment [25]. Most food stamp households
"provided specific expenditure information. If additional monies beyond
Sthe food stamp coupon allotment were spent, an estimate of the dollar
value was provided. However, a few households responded that addi-
tional money was spent on food beyond the food stamp coupon allot-
ment, but no accurate estimate of the magnitude of this amount was
available.
No distinguishing characteristics could be attributed to either the food
stamp respondents providing specific expenditure information or the food
stamp respondents that could not estimate additional expenditures. De-
spite differing amounts of available information, and since all households
were sampled from the same population, it was decided to pool the food
stamp household observations. From FSP households that provided spe-
cific additional expenditure information, budget analysis indicated that
these households spent additional funds equivalent, on the average, to 12
percent of the food stamp coupon allotment. It was assumed that house-
holds indicating unspecified expenditures in excess of the coupon allot-
ment could be represented by an average food expenditure proxy value
equal to 1.12 times the food stamp allotment.
Pooling data with different quantities of information introduces a
form of heteroscedasticity in the error term, and the general assumption:
E (uu') = o-2

18








is violated. The basic model can be represented:
[3.4] y = Xp + u
The group for which specific expenditure information was available was
represented as:
[3.5] yl = XiP + u,
Group 2, or the group for which less information was available was rep-
resented as:
[3.6 y2 = X2 + U2
The relationship between the true expenditure (y2) and the observed
expenditure y*2 for this group can be represented as:
1
[3.7] y*2 = 1.1Y2 +

where y*2 is equal to the food stamp allotment and e represents the
measurement error in the dependent variable y, which was present in
group 2 but not group 1 of the sample. Equating [3.6] and [3.7]:
[3.8] 1.12y*2 = X2f + u2 + 1.126
The joint empirical model was:
[3.9] y* = XP + v


where y* [= Y1
1.12 y*2


and v = ,1
u2 + 1.12 e
















19









The variance-covariance matrix was defined:

[3.10] S (vv') = .

0 O12 0




0 <2, 0

0 -12 0


0 oa22

0 22

where o(,2 = Var (ul)

and o22 = Var (u2 + 1.12 E)

The diagonal matrix A was defined:

[3.11] A= 1



0 1 0
(T1






1
(T2




"t"






20
"'2



20








Premultiplying [3.9] by A gives:
[3.12] Ay* =AXpf + Av
The variance-covariance matrix for the standardized disturbance term,
Av, was:
[3.13] E(Avv'A') = AYA' = I
A two step procedure was used to estimate the food stamp total food
expenditure model [4.1]. Step one generated estimates of the standard
errors, frf and -2, used to create the A matrix. Dependent and indepen-
dent variables were then premultiplied by the estimate of A to standard-
ize the variance. Generalized least squares (GLS) estimates obtained in
this manner are considered consistent estimates of p when using data for
which different quantities of information are available [15].
Since complete food expenditure information was available for the
entire subsample of eligible nonparticipating food stamp households,
equation [4.2] was estimated using ordinary least squares (OLS) proce-
dure.
It should be noted that some experimentation was done with the func-
tional form of both income and family size variables. It has been sug-
gested that the natural logarithm of income may be used to represent a
satiation level, as food expenditures increase at a decreasing rate with
increasing income. Similarly, the logarithm of family size may measure
economies of size [33]. Feaster and Perkins [10] have used an income
squared term which gave improved fit to explain food expenditures. In
this study the linear form of equation variables provided the best statisti-
cal fit.




















21
















Chapter IV

STATISTICAL MODELS


Three theoretical models, developed in Chapter III, are represented
by equations [3.1], [3.2], and [3.3]. Before the equations are estimated
using cross sectional data, the variables are specified. In addition, socio-
economic factors likely to account for variability among households are
also defined. In all equations, the samples are stratified by program par-
ticipation.

TOTAL FOOD EXPENDITURE MODEL
The total food expenditure model is stratified into two groups, food
stamp participants and eligible nonparticipating households. The func-
tional form of the food expenditure model for food stamp households
was estimated:
[4.1] TFEF, = ao + aCHI + o2B + a3FS + a4F + a- A + apW
+ a7Y + asH + a9BHI + aloBFS + P'LC + T'E
+ T'R + o'S + w
where:
underlined variables and coefficients represent vectors.
TFEF, = total household food expenditure/month for food stamp
households
HI = household income/month, including the sum of earnings for
all household members, welfare payments, pensions and so-
cial security
B = bonus stamp value
FS = number of persons in household
LC = vector of 0-1 dummy variables representing life cycle3 fam-
ily composition

3Duval [7] explains that the majority of households follow a sequential de-
velopmental pattern. This decomposition of families by the age of the oldest
child is one way to predict sibling groupings and act as a proxy for family
composition.

27







LC1 = 1 for beginning couple, no children
LC2 = 1 for oldest child birth to 6 years
LC3 = 1 for oldest child 7 to 13 years
LC4 = 1 for oldest child 14 to 20 years
LC5 = 1 for first child gone until last one leaves
LC6 = 1 for empty nest or retirement couple
E = vector of 0-1 dummy variables for ethnic background
El = 1 if white
E2 = 1 if nonwhite
F = 1 if head of household is female
A = number of household members regularly eating away-from-
home meals, for example, in School Lunch or Congregate
Meals for the Elderly
W = 1 if homemaker is employed
Y = age of homemaker
R = vector of 0-1 dummy variables for residence location
R1 = 1 if rural nonfarm
R2 = 1 if urban
S = vector of 0-1 dummy variables representing highest level of
education completed by homemaker
S1 = 1 if less than grade 9 education
S2 = 1 if grades 9-12 education
H = 1 if homemaker indicated a perception of a special health
need
BHI = interaction term between income and bonus stamp level
BFS = interaction term between family size and bonus stamp level
w = disturbance term
The functional form of the total food expenditure model estimated for
the eligible but nonparticipating FSP sample (TFEo,,ns) is as specified
in [4.1], but excludes variables measuring bonus value (B) and the
bonus interactions (BHI and BFS). Thus:
[4.2] TFEn, ,s- = [4.1], except for deleted variables as noted above.
The omitted categories of the dummy variables appearing in the intercept
of equations [4.1] and [4.2] were life cycle 1, white, male head of house-
hold, unemployed homemaker, rural nonfarm, less than grade 9 educa-
tion, and no perceived health need.

FOOD GROUP EXPENDITURE MODELS
It was suggested earlier that two stage budget maximization may pro-
vide a conceptual framework for the estimation of total food expendi-
tures. An alternative way of explaining variation in household food
expenditure is possible. Groups of foods with similar characteristics,

23








such as all vegetables, all meat products, all dairy products, or all grain
products, can be considered as weakly separable groups. They are not
necessarily subsets of the major grouping of food, and estimation of
variation among households proceeds in the same manner as for total
food expenditures.
Four sets of equations, one for each food group expenditure, were
estimated for each participation grouping. The first set of equations for
group 1, the FS EFNEP sample, was specified:
[4.3] FGE,, = ao + aHI + a2B + as FS + a4F + a5A
+ a6W + a-CY + a8H + a9BHI + aioBFS
+ allMP + 3'LC + T'E + T'R + O'S + 4'LNE + w
Previously undefined variables include:
FGE,,, = food group expenditures
m= 1-4
1 = meat and protein expenditures
2 = dairy product expenditures
3 = fruit and vegetable expenditures
4 = bread and grain product expenditures
MP = months of participation in the Expanded Food and Nutri-
tion Education Program
LNE = number of food demonstrations by EFNEP aides
LNE1 = number of demonstrations with protein foods
LNE2 = number of demonstrations with dairy products
LNE3 = number of demonstrations with fruits and vegetables
LNE4 = number of demonstrations with grain products
The LNE variable chosen corresponds to each particular food group ex-
penditure equation.
The equation estimated for the second participation grouping, FS non-
EFNEP, was:
[4.4] FGE, = [4.3], except for the deleted MP variable and vector
4'LNE
The estimaton for group 3, non-FS EFNEP, was specified:

[4.5] FGE, = [4.3], except for the deleted aB, aBHI and aBFS
variables.
Group 4, non-FS non-EFNEP, is the final participation stratum. The
functional form of the equation was specified:
[4.6] FGE, = [4.3], except for deletion of the aB, aBHI, aBFS,
aMP variables and vector 4'LNE

24








All variables are as defined previously, and all equations are estimated
using ordinary least squares with continuous and dummy variables. Var-
ious functional forms of the income, family size, and months in EFNEP
variables were used if improvement occurred in overall explanation of
variation.


NUTRIENT ADEQUACY MODELS
Five nutrient adequacy equations were estimated for each of the four
program participation stratifications. In the characteristics approach to
demand theory [16], the characteristics are the arguments of the utility
function. Therefore, the dependent variable NAR, was defined where:

R nutrient intake/homemaker/day
NAR, =
S Recommended Dietary Allowance (RDA)
i= 1-5
1 = protein
2 = calcium
3 = iron
4 = vitamin A
5 = vitamin C

If the NAR value exceeded 200 percent, it was truncated to 2.0 as
done by Madden and Yoder [21]. Variation in demand or variation in
redundant consumption of a nutrient beyond twice the recommended al-
lowance was not a point of interest in this study. It is also possible that
the few excessive consumption would mask variation among lesser in-
takes. Madden and Yoder [21] also noted that truncation at the value
of 2.0 made the distribution more nearly normal and the variance more
uniform.
Variation in demand or intake of a particular nutrient is determined
by the traditional economic variables of income, prices, and factors
which influence tastes and preferences. Cross sectional data were used,
so prices were assumed constant among all households. There was no
a priori reason to omit particular explanatory variables from any specific
nutrient adequacy equation. Economic theory provided little guidance
in the precise selection of functional form. It was assumed that as the
household's access to food increased, as measured by increased income,
bonus food stamps, or food expenditures, nutrient adequacy either in-
creased at a decreasing rate or peaked and declined. This may have
occurred as households had sufficient resources to substitute foods that
satisfied other wants than the biogenic or primary needs. Therefore, a
number of functional forms were compared for each equation.

25








As with the food group expenditure equations, the sample was strati-
fied into four program participation groupings. Separate NAR equations
were estimated for each group. The basic linear model estimated for the
FS EFNEP group was:
[4.7] NAR, = ao + a1HI + a2B + a3FS + a4F + asA
+ a0W + a7Y + a8H + aYBHI + ao1BFS
+ anlTFE + ai2MP + p'LC + T'E + T'R
-- I
+ e'S + p'LNE + w

The equation estimated for the second participation grouping, FS non-
EFNEP was:

[4.8] NARi = [4.7], except for the deleted aMP variable and vector
O'LNE
The estimation for group 3, non-FS EFNEP, was specified:

[4.9] NAR, = [4.7], except for the deleted aB, aBHI, and aBFS
variables.
Group 4, non-FS non-EFNEP, is the final participation stratum. The
equation was specified:

[4.10] NAR, = [4.7], except for the deleted aB, aBHI, aBFS,
aMP variables and vector p'LNE

where all variables have been previously defined.
One explanatory note is required for interpretation of the nutrient
adequacy variables. True nutritional status must be determined from
biochemical measures of nutrients and metabolites in the blood, urine,
and tissues. Factors which interfere with internal ingestion, absorption,
or utilization of nutrients may also affect nutritional status. Such factors
were not the focus of this study.
Food consumption surveys have been widely used as proxies for nutri-
tional status. Data used in this study were collected by Expanded Food
and Nutrition Education Program aides as part of the regular six month
family food recall record. In this method, each homemaker is asked to
name types and quantities of all foods and beverages eaten in the pre-
vious 24-hour period. Because intake on any particular day may be un-
representative of the homemaker's diet, this method cannot be used to
evaluate individual status. However, when sampling groups of individ-
uals, abnormal consumption effects are generally balanced out and the
24-hour recall results can be used in group comparisons [35]. Madden
et al. [20] note that group differences measured by 24-hour food recalls
may be understated, as subjects tend to over-report small quantities eaten
and under-report large quantities for calories, protein, and vitamin A.

26








In elderly subjects studied, with the exception of calories, group nutrient
means for recalled and actual intake were not statistically different.
Intakes for each of the five included nutrients were calculated for
each household so that the distribution of intakes within each grouping
was visible. The types and quantities of food consumed over the 24-hour
period were converted into nutrient composition and summed. Since
consumption of the homemaker was meant to act as a proxy for house-
hold consumption, all intakes were divided by the RDA standard for the
22-55 year old woman, rather than slightly differing actual requirements
for the younger or older homemakers.






































27














Chapter V

ANALYSIS AND RESULTS


TOTAL FOOD EXPENDITURE
It was suggested earlier that participation in the Food Stamp Program
would affect household food expenditures. Since food stamp supplemen-
tation provides in-kind income to participating households, it was also
suggested that all participating households would experience the income
effect, and certain households would also experience the price effect.
Comparison of group means presented in Table 5.1 indicated that group
differences did exist for the sample population. Total cash expenditure
for food stamps and additional food was significantly less for food stamp
households than total cash food expenditure for nonparticipating house-
holds. Food stamp participants spent a monthly average of $14.14 per
person for food stamp coupons and any purchased food for at-home
consumption in excess of the food stamp coupon allotment. By compari-
son, eligible nonparticipating households spent $33.22 per person for
food. The total value of purchased food also differed significantly be-
tween food stamp households and nonparticipating households. Total
value of purchased food averaged $195.94 per month for food stamp
households versus $125.92 for nonparticipants. Results of regression
analysis of total food expenditures for FSP participants and eligible non-
participants appear in Tables 5.2 and 5.3, respectively.
FOOD STAMP PARTICIPANTS
Income, bonus value and family size explained significant variation in
total food expenditures of FSP households. To the authors knowledge,
no prior study had found significant interaction between the value of
bonus stamps and either income or family size. In this study both inter-
actions were significant. The income response was:
DTFE
[5.1]a = .199 .0011 B
DHI
while the bonus value response was:
?TFE
[5.2] = .518 .0011 HI + .0507 FS

28








Evaluated at group means, the Marginal Propensity to Expend from
money income (MPEI)4 was 0.06 (standard error = .02). For every
one dollar increase in money income, total food expenditures increased
by six cents. The MPE1 increased with a reduction in the value of bonus
stamps and decreased with an increase in bonus value.


TABLE 5.1.-GROUP MEANS FOR SELECTED VARIABLES, FOOD STAMP
PARTICIPANTS AND ELIGIBLE NONPARTICIPANTS, POLK COUNTY,
FLORIDA, 1976.

Food stamp Eligible
Variable participants nonparticipants

Sample size 123 196
Family size 5.1 3.79

- dollars monthly - -
Money income 299.16 349.79
Real income 423.01a 349.79
Food stamp purchase requirement 64.47 -
Cash food expenditure 7.42b 125.92
Total cash expenditures 72.09 125.92
Food stamp bonus 123.85
Total value of purchased food 195.94e 125.92
Total cash expenditure, per capital 14.14 33.22
Value of purchased food, per capital 38.41 33.22

alncludes bonus value of food stamps.
bExpenditures made in excess of the food stamp purchase requirement.
eThe test statistic for the two sample t-test with unequal variances is 6.386.
dThe test statistic for the two sample t-test with unequal variances is 8.306.
eFood stamp purchase requirement + bonus value + cash food expenditure.


Evaluated at sample means, the Marginal Propensity to Expend from
bonus stamps (MPEB) was 0.45 (standard error = .07). The positive
sign of the bonus-family size interaction term suggested that as family
size increased, an increasing proportion of the discretionary income freed
by FSP participation was spent to purchase additional food. The negative
sign of the bonus-income interaction indicated that the MPE1 was larg-
est at low income levels and declined with increasing income (Table
5.2).
The full impact of a change in bonus value was measured not only
through the change in the MPEI (3TFE/1B), but also through the

"4While the traditional terminology used is Marginal Propensity to Con-
sume (MPC), the use of expenditure data in this study necessitated a corre-
sponding terminology modification.

29








TABLE 5.2.-STATISTICAL SUMMARY EQUATION [4.1], GLS COEFFICIENTS,
DEPENDENT VARIABLE, TOTAL FOOD EXPENDITURE BY FOOD STAMP
PARTICIPANTS, TFEFs, POLK COUNTY, FLORIDA, 1976.

Cell count Regression Standard
Variableab n = 123 coefficient error

Intercept 16.320 35.53
Income (HI) 0.199 0.053
Bonus value (B) 0.517 0.126
Family size (FS) 12.378 4.032
Life cycle:
LC2 12 -10.103 31.57
LC3 32 -16.820 31.54
LC4 48 -11.775 32.32
LC5 13 3.045 32.62
LC6 14 -33.764 31.42
Ethnicity:
Nonwhite (E2) 95 5.596 7.009
Female head of household (F) 75 1.776 6.440
Meals away from home 0.190 2.091
Employed homemaker (W) 23 1.264 6.688
Homemaker age (Y) 0.317 0.235
Residence:
Urban (R2) 81 7.553 5.381
Schooling:
GR 9-12 (S2) 73 0.028 5.628
Health need (H) 11 1.729 9.224
Interactions:
BHI -0.00109 0.00037
BFS 0.05068 0.01983

"aComplete variable definition can be found on pp. 22-25.
bA generalized least squares procedure was used to standardize the variance of
equation [4.1], so the R2 statistic is inappropriate.

change in the MPE, (3MPE1,/B). Although the sign of the bonus-in-
come interaction term was negative, if interaction variables were mea-
sured at group means, a one dollar increase in bonus value resulted in
overall increases in total food expenditure for households with monthly
incomes up to $700.00.
The coefficient of the family size variable also explained significant
variation among FSP households. The total food expenditure effect with
respect to family size was:

DTFE
[5.3] TFE 12.378 + .05068 B
3FS --

At mean bonus value, the family size effect was 18.61 (standard error
= 2.67). For every additional person in a FSP household, an additional
$18.61 per month was spent for food.

30








No other variable explained significant variation among FSP house-
holds.


ELIGIBLE NONPARTICIPANTS
The income response (MPEz) for eligible nonparticipants was 0.135,
or for every one dollar increase in money income, total food expenditures
increased by $.135 (Table 5.3). The magnitude of this response was
larger than that of the FSP sample (MPE, = .06). This result appeared
reasonable, since by income standards, households in this sample were
eligible to participate in the Food Stamp Program, but had chosen not to
do so. Reasons for nonparticipation were not evaluated in this study.
However, it seemed reasonable to expect that a number of households
in this group may have desired additional food for their families but did
not participate in the FSP for a number of reasons. Some of these rea-
sons may have included lack of transportation, lack of knowledge, in-

TABLE 5.3.-STATISTICAL SUMMARY EQUATION [4.2], OLS COEFFICIENTS,
DEPENDENT VARIABLE, TOTAL FOOD EXPENDITURE BY ELIGIBLE
NONPARTICIPANTS, TFEnon.s,, POLK COUNTY, FLORIDA, 1976.

Cell count Regression Standard
Variable n = 196 coefficient error

Intercept 63.380 17.900
Income (HI) 0.135 0.023
Family size (FS) 7.817 2.144
Life cycle:
LC2 47 -10.228 12.530
LC3 39 1.118 13.520
LC4 34 17.119 14.690
LC5 15 0.215 16.200
LC6 52 5.102 14.480
Ethnicity:
Nonwhite (E2) 159 3.734 6.779
Female head of household (F) 56 12.693 6.301
Meals away from home (A) 3.393 2.492
Employed homemaker (W) 34 0.229 6.931
Homemaker age (Y) 0.087 0.250
Residence:
Urban (R2) 85 8.222 5.410
Schooling:
GR 9-12 (S2) 109 5.780 5.692
Health need (H) 19 8.031 8.039

R2 = .6064
F17, 179 = 17.234

"aComplete variable definition can be found on pp. 22-25.

31








ability to obtain the required purchase requirement, unwillingness to
spend the time necessary to fill out the required certification form, or the
stigma attached to "welfare."
The family size coefficient of 7.18 was significantly different from zero
for eligible nonparticipants. This was considerably smaller than the FSP
participant coefficient of 12.38 (Table 5.2). One could assume that
variation in family size coefficients between groups was a result of FSP
participation. However, to make such an assumption, it would be neces-
sary to assume that both group samples were random selections of the
same general population, and that no program selectivity or sample bias
existed. In light of this assumption, interpretation of the family size co-
efficient as a FSP effect would have to be made with caution.
Nonparticipating female-headed households spent, on the average,
$12.69 less per month than male-headed households. In the FSP sample
there was no expenditure difference between female and male-headed
households. Since the incidence of poverty tends to be higher among
female-headed households than among male-headed households [31],
the FSP may have been operating as an equalizing factor between these
two household categories.
No other explanatory variable was consistently significant in explain-
ing total food expenditure variation among nonparticipating households.
However, the signs and magnitudes of the life cycle variables (LC2-
LC6) suggested certain trends in this sample. The coefficient for eligible
nonparticipant families with the oldest child 6 years or under was nega-
tive. This result seemed plausible, since a household including small
children would tend to consume less food than a family composed pri-
marily of young adults. Food expenditure was at a maximum for house-
holds in which the oldest child was between 13 and 20 years of age
(Table 5.3).

FooD GROUP EXPENDITURES
Two stage budget maximization suggests an initial optimal allocation
of income to budgeting groups for food, housing, etc. The second stage
allocation proceeds in proportion to the prices of individual goods within
the budgeting group. It is possible for the food expenditure budgeting
group to undergo one additional maximization step. The primary budget-
ing category can be allocated to food group expenditures. It is of interest,
from a policy standpoint, to determine the effect of food stamp supple-
mentation and nutrition education on household food group expenditures.
Table 5.4 presents group means of food group expenditures and per-
centages of total food expenditure by program participation group and
race. Table 5.5 presents selected group means by program participation
group and race.

32








MEAT AND PROTEIN PRODUCT EXPENDITURES (FGE1)

A complete statistical summary appears in Appendix Table A-i. As
expected, income and bonus value variables are positively related to
protein product and meat expenditures. With respect to protein expendi-
tures, at group means, the MPE1 for the FS EFNEP group was 0.06
(standard error = .038), compared to an MPE1 of 0.128 (standard
error = .048) in FS non-EFNEP households. The MPE, for non-FS
EFNEP households was 0.062, while the MPE, of the non-FS EFNEP
sample was 0.117. It was interesting that the two groups which partici-
pated in EFNEP had about one-half the Marginal Propensities to Ex-
pend from income than did non-EFNEP households. One possible inter-


TABLE 5.4.-GROUP MEANS FOR MONTHLY FOOD GROUP EXPENDITURES,
BY RACE AND PROGRAM PARTICIPATION GROUP, POLK COUNTY,
FLORIDA, 1976.

Food group expenditures
FGE1 FGE3 FGE4
(Meats and FGE2 (Fruit and (Bread and
Program protein (Dairy vegetable grain
status products) products) products) products)
- - - dollars - - -
White
FS EFNEP 60.62 20.30 41.62 32.69
(38.4)b (12.2) (25.3) (16.8)
FS non-EFNEP 88.25 40.38 34.62 30.25
(41.7) (20.0) (16.5) (14.9)
non-FS EFNEP 64.86 25.14 28.86 22.00
(43.6) (17.5) (17.9) (13.7)
non-FS non-EFNEP 59.40 20.90 25.10 18.80
(40.1) (17.1) (18.4) (15.3)
Nonwhite
FS EFNEP 80.27 34.60 33.30 23.14
(42.7) (17.9) (18.9) (12.1)
FS non-EFNEP 64.19 22.12 33.30 30.07
(38.7) (13.7) (18.9) (17.3)
non-FS EFNEP 44.36 13.72 22.54 14.15
(43.3) (13.1) (22.8) (12.8)
non-FS non-EFNEP 54.85 17.30 23.61 18.67
(43.1) (13.6) (19.6) (15.1)

"aA complete program participation group definition can be found on page 3.
'Figures appearing in parentheses indicate the percentage of total food expendi-
ture allocated to the particular food group. Summation of percentages may not
equal 100 percent as miscellaneous group expenditures are not included.

33

















TABLE 5.5.-SELECTED GROUP MEANS, BY RACE AND PROGRAM PARTICIPATION GROUP, POLK COUNTY, FLORIDA, 1976.

Program status

FS EFNEPb FS non-EFNEP non-FS EFNEP non-FS non-EFNEP

white nonwhite white nonwhite white nonwhite white nonwhite
Variableb n = 13 n = 37 n = 8 n = 26 n = 7 n = 66 n = 10 n = 13

HI 213.73 314.88 347.75 261.94 423.23 216.29 371.48 307.22
FS 4.46 5.16 5.50 4.65 4.14 3.06 4.80 4.52
B 114.23 117.38 141.50 106.69 -
A 2.23 3.00 2.25 2.00 .85 1.27 1.20 1.60
MP 19.69 16.32 13.57 16.76 -

"A complete program participation group definition can be found on p. 3.
hA complete definition of variables can be found on pp. 22-25.








pretation was that this was an improvement in food economy for par-
ticipating EFNEP households.
The sign of the bonus-income interaction term was negative for FS
EFNEP households. At mean income, the MPE, for FS EFNEP house-
holds was 0.335 (standard error = .226). Discretionary income, freed
by the food stamp bonus, was used to increase meat and protein product
expenditures for households with monthly incomes below $538. At in-
comes above this level, the freed income was spent for other goods. The
sign of the bonus family size interaction term was negative for FS non-
EFNEP households. Similarly, the MPE, for FS non-EFNEP house-
holds of four persons was 0.328 (standard error = .119). For all food
stamp households, about $0.33 of each bonus dollar were spent for addi-
tional meat and protein product. As family size increased, the MPE,
declined, and less of the bonus dollar was spent for meat. This reflected
the effect of greater fixed expenditures for larger households.
The variable measuring months of participation in EFNEP, MP, did
not explain variation within the FS EFNEP sample. However, in the
non-FS EFNEP sample, meat expenditures increased slightly throughout
the first 14 months of EFNEP participation.


DAIRY PRODUCT EXPENDITURES (FGE2)

A complete statistical summary appears in Appendix Table A-2.
Dairy product expenditures were consistently positively related to income
levels. Evaluated at a monthly income of $300, the MPEI of the FS
EFNEP group was 0.014 (standard error = .008). The FS non-EFNEP
group MPE, was 0.0328 (standard error = .017), compared to 0.027
(standard error = .008), for the non-FS EFNEP group. The magnitude
of the non-FS non-EFNEP group (MPE, = .112) was considerably
larger.
In neither of the two food stamp samples was the bonus response,
MPEB, significantly different from zero. One interpretation is that none
of the income freed by food stamp participation was used to increase
dairy product expenditures.
Variables measuring EFNEP participation explained considerable
variation among both EFNEP participation groups. In the FS EFNEP
sample, the coefficients of months in program (MP) and months in pro-
gram squared (MPSQ) indicated that dairy product expenditures de-
clined throughout the first 17 months of EFNEP participation. The num-
ber of dairy product lessons (LNE2), explained considerable variation
among the non-FS sample. For this participation group, every lesson
demonstrating dairy products could be expected to decrease monthly
dairy expenditures by $2.72.

35








One additional point should be made regarding intra-sample variation.
Dairy product expenditures of nonwhite households were significantly
lower than those of white households for three or four program group-
ings. A portion of this difference may have been related to lactose intol-
erance common in many nonwhite populations [2]. Many nonwhite per-
sons have a deficient supply of lactase, the enzyme necessary for the
digestion of lactose, the sugar found in milk. Ingestion of dairy products
in excess of digestive capabilities causes flatulence, stomach distension,
or diarrhea. Therefore, lower dairy product expenditure was not surpris-
ing, but did have implications for nutrition educators.

FRUIT AND VEGETABLE EXPENDITURES (FGE3)
Complete regression results for fruit and vegetable expenditures ap-
pear in Appendix Table A-3. Fruit and vegetable expenditures were, in
general, income responsive. For the FS EFNEP sample, however, expen-
ditures were not responsive to increases in income. At a monthly income
level of $300 the MPEi of the FS non-EFNEP sample was 0.029 (stan-
dard error = .014). Similarly, the MPEI of the FS non-EFNEP sample
was 0.043 (standard error = .020) and the MPE, of the non-FS non-
EFNEP was 0.053 (standard error = .022).
Bonus food stamps were effective in increasing fruit and vegetable
expenditures for both the FS EFNEP and FS non-EFNEP samples. The
bonus response (MPEB) of the FS EFNEP sample was positive and
proportional to family size. The MPE, for a family of four was 0.095
(standard error = .044). For every additional bonus dollar, nine cents
was spent on additional fruit and vegetable expenditures. The bonus re-
sponse of the FS non-EFNEP sample, 0.220, was significantly larger.
Family size was a significant indicator of fruit and vegetable expendi-
tures for both the FS EFNEP and non-FS EFNEP samples. As men-
tioned earlier, bonus value and family size interacted positively in the
FS EFNEP sample. At mean bonus value and family size of four, the
marginal increase in fruit and vegetable expenditures due to family size,
3FGE3/1FS, was $8.37. In the non-FS EFNEP sample, the response
for a family size of four was $10.84 per month.
The EFNEP participation variables (MP and LNE3) did not affect
fruit and vegetable expenditures for these samples.
Two additional factors were of interest. In the FS non-EFNEP sam-
ple, nonwhite fruit and vegetable responses were significantly higher
($17.43) than those of white households. No expenditure difference
occurred between racial groups in the FS EFNEP sample. Data were not
available to distinguish quantities purchased. Therefore, it is only pos-
sible to speculate as to reasons for expenditure differences. Food stamp
supplementation may have encouraged nonwhite households to purchase

36








significantly larger quantities of fruits and vegetables, proportionally
more food related services, or more likely, some combination of both.
One possible reason that the nonwhite section of the FS EFNEP sample
did not exhibit significantly higher expenditures than the white popula-
tion, may have been that the nonwhite section was more susceptible to
food economy instruction of the EFNEP paraprofessionals.
The final observation with respect to fruit and vegetable expenditures
was related to the felt-health need variable (H). Recall that a dummy
variable was used to indicate participants expressing a special health
status, for example, pregnancy or weight control. Persons expressing a
felt-health need in both the FS EFNEP and non-FS EFNEP samples
spent significantly more for fruits and vegetables than those not express-
ing such a need. The results can be interpreted in two ways. Either
EFNEP participation was redefining the objective perception of health
for participating households, or persons chose to participate in EFNEP
because of compatible health ideologies.


BREAD AND GRAIN EXPENDITURES (FGE4)
"Detailed results for bread and grain expenditures appear in Appendix
Table A-4. The MPEI values for FS EFNEP and FS non-EFNEP house-
holds were not significantly different from zero. Results of this analysis
suggested that an increase in income would not significantly affect bread
expenditures for these groups. This contrasts to a MPEI of 0.056 (stan-
dard error = .015) for non-FS non-EFNEP households. Both responses
were calculated at a monthly income level of $300. In the non-FS
EFNEP and non-FS non-EFNEP groups, the most descriptive func-
tional form for income included an income squared (HISQ) term in the
analysis. This suggested that expenditures increased until a certain income
level was reached, and then declined. In the non-FS EFNEP sample,
grain expenditures increased with increasing income until monthly in-
comes of $573 were reached. The maximum monthly income was lower
in the non-FS non-EFNEP sample. Grain expenditures began to decline
with increasing income above monthly income levels of $350.
The bonus response (MPEn) of the FS EFNEP sample was posi-
tively related to family size (BFS = .052, standard error = .016). The
bonus response of a four person household was 0.208. The bonus re-
sponse of the FS non-EFNEP sample was not related to family size.
The MPE, for this sample was 0.229.
The effect of family size on bread and grain expenditures was mani-
fested in the form of different variables across samples. As mentioned
previously, the FS EFNEP bonus effect depended on family size. In the
FS non-EFNEP sample, the life cycle variables (LC2-LC6), explained
significant expenditure variation. As expected, families with younger

37








children or older retired members spent less on bread and grain products
than did families with teenagers. All life cycle categories spent more
than the standard category of two adults. In the non-FS EFNEP sam-
ple, the life cycle variables also explained variation. As in the FS non-
EFNEP sample, LC5 expenditures were higher than those of two adults
(LC1) and all other categories. However, all other life cycle categories
appeared to spend less for bread and grain products than the two-person
household. The increase in bread and grain products was directly pro-
portional to family size in the non-FS non-EFNEP sample. In this par-
ticipation group, for every additional household member, monthly bread
expenditures increased by $1.53.
Participation in EFNEP had contrasting results between the FS
EFNEP and non-FS EFNEP samples. In the FS EFNEP sample, expen-
ditures for bread and grain products declined throughout the first 13
months of EFNEP participation. However, in the non-FS EFNEP sam-
ple, expenditures increased with EFNEP participation. If adequate re-
sources were available in the food stamp households to provide sufficient
food, the EFNEP aides could proceed to teach these homemakers how
to use these resources efficiently. In the non-FS EFNEP households, food
economy may have been sacrificed in order to achieve improved nutrient
consumption.

NUTRIENT ADEQUACY RATIOS
The empirical results presented up to this point suggest that both the
Food Stamp Program and the Expanded Food and Nutrition Education
Program influenced food consumption. Both programs share the common
objective of improving the quality of low income households' diets.
Table 5.6 presents mean Nutrient Adequacy Ratios (NARs) by program
participation groups. For each program participation group, it also pre-
sents the percentage of diets below the Recommended Dietary Allow-
ance (RDA), and those below 66 percent of the RDA. Since the RDA
is an allowance, rather than the minimum requirement for a nutrient,
the 66th percentile is often used to differentiate between adequate or
poor diets. Mean values for protein adequacy were the highest, as no
household fell below the 66th percentile in three of the four program
groups. Mean values for iron consumption were the lowest. The overall
mean value for iron was less than six-tenths of the RDA.

PROTEIN ADEQUACY (NAR1)
Media attention has been focused on the protein deficit of developing
countries. It is commonly believed that protein deficiency is the most
serious domestic nutritional deficiency. While protein deficiency may be

38








a problem in many developing nations, it is not the most predominant
nutritional deficiency in the low income population of the United States
[5]. Sample protein adequacy means from Polk County intakes indicated
that sample means were similar to the national trend. Protein adequacy
sample means were all significantly greater than 100 percent of the RDA.
The intake mean for the non-FS non-EFNEP group was significantly
less than intakes in the FS EFNEP, FS non-EFNEP, or non-FS EFNEP
groups. However, non-FS non-EFNEP intake was still greater than one
and one-third times the RDA for protein.

TABLE 5.6.-GROUP MEANS FOR NUTRIENT ADEQUACY RATIO, BY PROGRAM
PARTICIPATION GROUP, POLK COUNTY, FLORIDA, 1976.

Nutrient Program status
Nutrient
adequacy FS FS non-FS non-FS
ratio EFNEP non-EFNEP EFNEP non-EFNEP
(NAR) n = 49 n = 34 n = 73 n = 71

NAR1 (Protein)
Mean 1.52 1.51 1.58 1.36
"% below RDA 6.1 18.6 10.9 21.1
"% below 66% RDA 0.0 0.0 0.0 8.5
NAR2 (Calcium)
Mean 0.88 0.70 0.98 0.58
% below RDA 65.3 69.8 52.1 83.1
% below 66%RDA 36.7 48.8 31.5 64.8
NAR3 (Iron)
Mean 0.63 0.60 0.62 0.54
"% below RDA 95.9 97.7 95.9 98.6
"% below 66% RDA 63.3 65.1 65.7 76.1
NAR4 (Vitamin A)
Mean 1.29 0.89 1.25 0.84
% below RDA 36.7 60.5 43.8 66.2
% below 66% RDA 26.5 51.1 23.3 54.9
NAR5 (Vitamin C)
Mean 1.49 1.15 1.49 0.96
"% below RDA 18.4 44.2 28.7 57.8
"% below 66% RDA 8.1 39.5 17.8 46.5


Protein adequacy regression results appear in Appendix Table A-5.
No consistent relationship existed between the level of income and ade-
quacy of protein. Food expenditures were more consistent indicators of
protein adequacy in three of four samples. In three cases, protein ade-
quacy declined with increasing food expenditures, and rose after a
certain level of expenditure was attained. For the FS non-EFNEP and

39







non-FS non-EFNEP groups, protein adequacy increased after monthly
food expenditures reached $191 and $151, respectively. For the FS non-
EFNEP group, protein adequacy did not increase until food expenditures
rose above $270 per month.
The food stamp bonus value was consistently positive, but of relatively
small magnitude in both the FS EFNEP and FS non-EFNEP samples.
The EFNEP variables did not explain significant variation in the FS
EFNEP sample. However, in the non-FS EFNEP sample, the number
of food demonstrations in which protein or meat products were used,
(LNE1), was positively related to increases in protein adequacy.
Additional variation was explained by the variable measuring meals
eaten away-from-home (A). The A coefficient was positive in participa-
tion groups, 1, 2, and 3-FS EFNEP, FS non-EFNEP, and non-FS
EFNEP, respectively. This was difficult to interpret within an economic
framework. However, the positive sign may have suggested something
about the economic characteristics of away-from-home food consump-
tion. A typical meal at a fast food outlet might include a hamburger or
fried chicken, both foods rich in protein. A greater variety of foods was
available for at-home consumption, so choice may have been less re-
stricted to protein-rich foods.
A final note should be made regarding protein consumption. As pre-
viously noted, each household nutrient intake was truncated at 200
percent of the RDA. Since the NARs for protein were so high, less
variation among households was present than for other nutrients. The
effect of income and bonus stamp supplementation may have been
masked by the truncation process.

CALCIUM ADEQUACY (NAR2)
Calcium adequacy regression results appear in Appendix Table A-6.
Only in the non-FS non-EFNEP group did income explain a significant
amount of variation in calcium adequacy. The relationship was negative
until a monthly income level of $370 was reached. After this point, cal-
cium adequacy increased with income.
It was of interest to observe the bonus coefficient for the FS EFNEP
and FS non-EFNEP samples. The FS non-EFNEP coefficient was not
significantly different from zero. An increase in bonus value did not
result in increases in calcium adequacy for this group. However, the
bonus response for FS EFNEP households was negative. For every addi-
tional bonus dollar, calcium adequacy (NAR2), declined by 0.0056. It
was possible that as increased funds for food expenditure became avail-
able via bonus stamps, sodas and other popular, but low calcium bever-
ages, were substituted for milk products.

40







Neither months of participation in EFNEP (MP), nor lessons with
dairy products (LNE2), had an impact on calcium adequacy for the FS
EFNEP sample. However, the nutrition education program did signifi-
cantly influence calcium adequacy levels of non-FS EFNEP households.
For this sample, the NAR2 was increased during the first 14 months of
participation. Each food preparation lesson utilizing dairy products also
increased calcium adequacy levels for non-FS EFNEP households by
0.2819 (standard error = .0541). Nutrition education appeared to be
effective in increasing calcium adequacy for non-food stamp households,
but had no effect on households that received stamps.
One additional point should be made regarding calcium adequacy
levels. As previously noted, nonwhite households spent significantly less
for dairy products than did white households. Although dairy products
were a major source of calcium, calcium adequacy levels of nonwhite
households were not significantly different from those of white house-
holds in any group, other than the FS non-EFNEP sample. This sug-
gested that nonwhite households may have achieved adequate calcium
levels with lower expenditures than did white participants in the FS
EFNEP and non-FS EFNEP samples.

IRON ADEQUACY (NAR3)
Iron adequacy regression results appear in Appendix Table A-7. Iron
adequacy levels were responsive to income in all participation groupings.
FS EFNEP households demonstrated a positive income response up to
income levels of $735. The income response of the FS non-EFNEP
sample was positive only up to income levels of $225. The negative
bonus-income interaction was also significant in both food stamp sam-
ples. The income response of the non-FS non-EFNEP household was
the reverse of both food stamp samples. In this group, iron adequacy
declined until an income level of $400 was reached. After this point,
iron adequacy also increased.
The bonus effect was positive for the FS EFNEP households up to
monthly income levels of $350. This suggested that an increase in bonus
value was effective in improving iron adequacy levels for FS EFNEP
households with monthly incomes up to $350. An increase in bonus
value did not significantly affect iron adequacy of households in the FS
non-EFNEP sample.
Months of participation in EFNEP was positively related to increased
iron levels in both the FS EFNEP and non-FS EFNEP samples. The
variable measuring the number of lessons in which grain products were
demonstrated (LNE4), explained significant variation among the FS
EFNEP sample. For this sample, these lessons were more effective in
improving iron adequacy at low income levels as indicated by the nega-

41







tive sign of the interaction term (HILNE4). Grain product lessons were
not effective in increasing iron levels for non-food stamp households.
There was no difference in iron adequacy between white and nonwhite
FS EFNEP households. However, iron adequacy of nonwhites was sig-
nificantly lower than that of whites in both the FS non-EFNEP and
non-FS EFNEP households. This suggested that nonwhite households
may have achieved higher iron levels with joint participation in the FSP
and EFNEP, than with participation in either program separately.

VITAMIN A ADEQUACY (NAR4)
Regression results for vitamin A adequacy appear in Appendix Table
A-8. Across program participation groups, vitamin A adequacy did not
respond consistently to changes in either income or family size.
For the FS EFNEP sample, as the value of bonus stamps increased,
vitamin A adequacy declined. The positive sign of the interaction term,
BHI, indicated that the decline was of a smaller magnitude as income
rose. The bonus effect became positive only for households with monthly
incomes greater than $950. In the FS non-EFNEP sample, the bonus
value coefficient was not significantly different from zero.
The educational level of the homemaker accounted for significant
variation within the FS EFNEP and FS non-EFNEP samples. Home-
makers completing some high school education registered higher levels
of vitamin A adequacy than did homemakers with eighth grade or less
education. No differences could be attributed to educational status of the
homemaker for the non-FS EFNEP or non-FS non-EFNEP groups. This
suggested that more highly educated homemakers may have been able
to use income supplementation, made available by the FSP, in a more
nutritionally efficient manner than did homemakers completing fewer
years of schooling.
Nutrition education significantly affected the vitamin A adequacy of
both the FS EFNEP and non-FS EFNEP groups. In both groups, vita-
min A adequacy increased throughout the first 12 to 14 months of
EFNEP participation. The number of fruit or vegetable preparation
demonstrations (LNE3) was also positively related to increased vitamin
A adequacy, up to certain income levels. An income-vegetable lesson
interaction term was significant for both groups. At an income level of
$300 per month, each lesson in vegetable preparation increased vitamin
A adequacy in the FS EFNEP sample by a factor of 0.185 (standard
error = .094). In this sample, vegetable lessons were effective in increas-
ing vitamin A levels for households with monthly income levels up to
$503. Vegetable lessons were effective in increasing vitamin A adequacy
throughout a smaller income range for the non-FS EFNEP sample. Posi-
tive effects were achieved only within households with monthly incomes

42








of less than $285. These results suggested that nutrition education of this
type was more effective with joint participation in the FSP and EFNEP.


VITAMIN C ADEQUACY (NAR5)
Vitamin C regression results appear in Appendix Table A-9. Vitamin
C is a water soluble nutrient which must be consumed daily. The require-
ment is easily supplied from a number of fruit or vegetable sources.
However, 39.5 percent of all FS non-EFNEP households and 46.5 of all
non-FS non-EFNEP households consumed less than two-thirds of the
daily requirement for vitamin C on a day of the food recall. The FS
EFNEP and non-FS EFNEP samples consumed significantly higher
levels of vitamin C, with only 8.1 percent and 17.8 percent, respectively,
consuming less than two-thirds of the RDA (Table 5.6).
For food stamp participants in the FS EFNEP and FS non-EFNEP
samples, the income effect explained less variation in vitamin C adequacy
than did the bonus and family size effects. The bonus-family size inter-
action term was positive in both samples. This suggested that bonus
stamps were more effective in improving vitamin C adequacy as family
size increased. The bonus response (MPEB) for a family of four was
0.0087 and 0.0158 for the FS EFNEP and FS non-EFNEP samples,
respectively. At a family size of six, the responses increased to 0.0130
and 0.0238, respectively.
Family size also explained variation in all samples. In the FS EFNEP
sample, the family size effect was negative at bonus values less than
$195. In the FS non-EFNEP sample, at mean income, the family size
response was positive at bonus values above $121. The family size re-
sponse was also negative for the non-FS EFNEP sample at monthly
income levels below $405, and became positive at higher income levels.
One explanation for the observed family size responses is that citrus
products may have been considered luxury goods at low income levels.
Thus, an increase in either money income or bonus food stamps may
have been used to purchase other necessities first. Larger families may
have had proportionally more fixed expenditures. Therefore, higher levels
of income or bonus stamps were necessary to promote positive increases
in vitamin C adequacy.
Participation in EFNEP improved vitamin C adequacy for both the
FS EFNEP and non-FS EFNEP households. The magnitude of the
FS EFNEP coefficient (LNMP = 0.376, standard error = .147), was
slightly larger than the non-FS EFNEP coefficient (LNMP = 0.306,
standard error = .133). The number of fruit and vegetable preparation
lessons did not explain variation in vitamin C adequacy. This seemed
plausible, since few vitamin C-rich foods, such as orange juice, required
special preparation.

43













Chapter VI

POLICY IMPLICATIONS, SUMMARY,
AND CONCLUSIONS


POLICY IMPLICATIONS

Results of this study suggested that participation in the Food Stamp
Program and the Expanded Food and Nutrition Education Program af-
fected total food expenditure levels, food group expenditures, and nu-
trient adequacy of participating households. Results of the regression
analyses were used to compare various policy alternatives. In such com-
parisons, variables such as family size, ethnicity, urbanity, and schooling
were held constant for each alternative policy. Source and amount of
food program supplementation and months of EFNEP education consti-
tuted the matrix of policy alternatives that varied in the sets of compara-
tive analysis. The set of constant family characteristics were not meant
to be a representative sample of the low income population. For exam-
ple, while this type of analysis would not allow one to project actual
values for the entire low income population, it did provide a useful
method of comparing the effect of differing levels of the policy variables.
To compare how total food expenditures changed with various levels
and types of income supplementation, four policy alternatives were for-
mulated. A summary of alternative policy levels appears in Table 6.1.
In policy alternative A, a cash supplement replaced the coupon allot-
ment of the then existing FSP. Policy B represented the FSP, using


TABLE 6.1.-ALTERNATIVE POLICY SPECIFICATIONS, TOTAL FOOD
EXPENDITURE, POLK COUNTY, FLORIDA.

Amount of Source of
Policy supplementation supplementation
dollars -
A 123 Cash
B 123 FSP bonus
C 91 FSP bonus
D 0

44








sample mean bonus value of $123, as the program was structured at the
time of the study. Deductions for this household would equal $150.
Since the value of the FSP coupon allotment was not a variable in
the models of this study, exact expenditure levels could not be projected
for the FSP proposed by President Carter in the 1977 Welfare Reform
Proposal. The models could, however, measure differences attributable
to changes in bonus value. Policy C incorporated the then newly pro-
posed standard deduction and calculated the bonus value by subtracting
30 percent of net income from the existing coupon allotment. Subtrac-
tion of the proposed $80 standard deduction resulted in a net income of
$250. The bonus value under Policy C was equal to $91. This bonus
level was considerably lower than the sample level of $123, but was
greater than the existing program bonus value for a family with no in-
come deductions ($71). For comparison purposes, Policy D represented
those persons who chose not to participate in a food aid program. Each
policy alternative projected expenditures for a family of four persons,
at a monthly income level of $330. Projected differences in total food
expenditures, or total value of purchased food, resulting from alternative
food policies appear in Table 6.2.


TABLE 6.2.-PROJECTED MONTHLY TOTAL FOOD EXPENDITURES, BY ALTERNA-
TIVE FOOD POLICY, BY FOUR PERSON, RURAL NONFARM,
NONWHITE HOUSEHOLDS, POLK COUNTY, FLORIDA

Total food
Policy expenditures

dollars -
A 150.81
B 174.50
C 162.74
D 131.78



These projections indicated that nonwhite, four person families not
participating in the FSP, with monthly incomes of $300, could be ex-
pected to spend $131.78 per month for food. In contrast, the same family
participating in the FSP, as structured at the time of the study, under
Policy B, would spend $174.50 monthly for food. In a reduced bonus
value situation, Policy C, this family would be expected to spend $162.74
per month. If a cash supplement equal to the then existing bonus value
($123) was offered, Policy A, the family would spend $150.81.
A graphic example will demonstrate the effect of alternative policies
on various income-consumption patterns (Fig. 2).

45










L



E
I / 1K




V / V
A





o
H

E

F G

Dollar expenditures for food items


Figure 2.-Effect of alternative policies on income, consumption, and budget
constraint relationships.

The preference schedule represented by vector as would be equally
affected by all policies. With a coupon or cash supplement, consumption
would increase from point E to point H. Households represented by vec-
tor Q2 would increase from point D to point J. Elimination of the pur-
chase requirement, but retainment of the bonus value coupons, Policy
C, would result in a new income constraint of AMG. The desired posi-
tion of point J would be no longer attainable. Consumption on vector as
would increase from point D to point M. Food stamp policy, as struc-
tured at the time of the study, with a purchase requirement, Policy B,
would be represented by a budget constraint of ACIG. Consumption
would increase from point D to point I. For households with preference
schedules represented by vector a2, Policy B would increase total food
expenditure the most, Policy C less, Policy A the least.

46








Families with preference schedules represented by al would also re-
spond differently to each alternative policy. The possibility existed that
certain households would not participate in the FSP as implemented in
Policy B. With elimination of the purchase requirement, higher levels
of utility would be possible for all a, households. Therefore, it could
be expected that Policy C would induce greater participation than Policy
B. For al households, participation in the FSP as it existed, Policy B,
would have increased food expenditures from point B to poiht I. Policy
C participation would have increased expenditures from point B to point
M. A cash income supplement would have resulted in increased expendi-
tures from point B to point K.
One of the stated objectives of the Food Stamp Program is provision
of adequate income supplementation to increase food expenditures by
low income households. Food stamp policy, as it existed at the time of
the study, required a cash purchase requirement and provided a coupon
allotment equal in value to the purchase requirement plus a bonus sub-
sidy: Of the policies studied, this policy (Policy B) appeared to be the
most effective method of increasing total food expenditures. Elimination
of the cash purchase requirement and provision of a coupon allotment
equal in value to the bonus subsidy (Policy C) was less effective in in-
creasing total food expenditures. This policy, however, probably would
encourage food aid program participation for families with lower Mar-
ginal Propensities to Consume food. A cash income grant (Policy A)
appeared to be the least effective method of increasing food expenditures
for low income households.-
One of the common goals of both the Food Stamp Program and the
Expanded Food and Nutrition Education Program is to improve the
nutritional quality of low income households' diets. In light of current
interest in reorganizing welfare policy and, in particular, the structure
of the Food Stamp Program, it was of interest to examine the effect of
alternative policies on the nutrient adequacy ratios.
Results of regression analyses can be used to project nutrient ade-
quacy ratios in alternative policy situations. Again, the absolute numbers
obtained could not be interpreted as a projection for the entire low
income population. For the nutrient adequacy ratios, alternative policy
levels included not only amount and source of income supplementation,
but also months of participation in the Expanded Food and Nutrition
Education Program. A complete specification of alternative policy com-
binations appears in Table 6.3. The rationale behind the supplementation
levels of $123 and $91 has been described earlier.
Histograms depicting projected nutrient adequacy ratios achieved with
alternative policies appear in Figures 3-7. Each figure shows projected
NAR values for a particular nutrient for each policy situation. Each
policy level is keyed to specifications given in Table 6.3.

47








TABLE 6.3.-ALTERNATIVE POLICY SPECIFICATION, NUTRIENT
ADEQUACY RATIOS

Amount of Source of Months of EFNEP
Policy supplementation supplementation participation
dollars -
A 123 FSP bonus 12
B 123 FSP bonus 18
C 91 FSP bonus 12
D 91 FSP bonus 18
E 123 FSP bonus 0
F 91 FSP bonus 0
G 0 -12
H 0 18
I 123 Cash 12
J 123 Cash 18
K 91 Cash 12
L 91 Cash 18
M 0 0
N 123 Cash 0
O 91 Cash 0


Projected NAR values for protein adequacy (Fig. 3) and vitamin C
adequacy (Fig. 7) were greater than 100 percent of the RDA for all
policy alternatives. A larger proportion of sample households had pro-
tein intakes in this range (Table 5.10). However, Table 5.10 also points
out that, although sample mean NARs for vitamin C were in the pro-
jected range, vitamin C consumption by a large percentage of selected
households fell below 100 percent and 66 percent of the RDA. There-
fore, one could and should not conclude from the projected NARs that
all sample households consumed adequate quantities of this nutrient.
Variations in vitamin A (Fig. 6) and iron (Fig. 5) adequacy more
clearly reflected differences attributable to alternative food policies. Vita-
min A adequacy registered the most dramatic difference. The projected
NAR for vitamin A by persons not participating in a food aid program
or EFNEP (Policy M), was .01. Cash income supplements of $123 and
$91 increased the NAR4 only to 0.08 and 0.04, respectively. In con-
trast, participation in either the Food Stamp Program alone (Policies E
and F) or EFNEP alone (Policies G and H) increased vitamin A ade-
quacy to near or above 100 percent of the RDA. Joint program partici-
pation produced the highest NAR levels (Policies A, B, C, and D).
Projections for iron adequacy and calcium adequacy fell below the
RDA in all but two cases. In each case, the highest projected NAR
values occurred with joint participation in the FSP and EFNEP (Policies
C and D). The lowest NAR values occurred for households that did not

48







3.56

3 3.33


"I r 1.63 1.64 1.63
152 1 "53
1.48 1.48 1.47 1.47

1.35 1.36
.35 .36 1.32 1.33




NAR1 --.- -- -- - -- -- -- 100 Percent RDA
Protein











A B C D E F G H I J K L M N O

Policy
Figure 3.-Projected nutrient adequacy protein, NAR1, with alternative policies, Polk County, Florida.



















1.07
1.01 .94 .92 .93 .91
NAR2 --... -- -- ------ 100 Percent RDA
Calcium .89 .79 .77
r83! .79 .77
.74 ._
.66


.40 .39






A B C D E F G H I J K L M N O

Policy




Figure 4.-Projected nutrient adequacy calcium, NAR2, with alternative policies, Polk County, Florida.



















NAR3 ____ _________ 100 Percent RDA
Iron


.63 .62
.57 .56 .54
.46 .45 49 .45 .48



.29 .19 .18



A B C D E F G H I J K L M N O
Policy




Figure 5.-Projected nutrient adequacy iron, NAR3, with alternative policies, Polk County, Florida.














1.83
"1.75

1.47
1.39 1.39 1.37
1.30
1 27 1.25 1.22



NAR4 .99 .94
NAR4 ._ 100 Percent RDA
Vitamin A









.08
.01 .08 .04

A B C D E F G H I J K L M N O
Policy

Figure 6.-Projected nutrient adequacy vitamin A, NAR4, with alternative policies, Polk County, Florida.















1.49 1.51 1.51
1.48
1.41 1.38 1.39 1.38
1.33 1.35 1.32 1.281.38
1.28
1.18 1.20

1.06
NAR5 .- 100 Percent RDA
Vitamin C









A B C D E F G H I J K L M N O

Policy


Figure 7.-Projected nutrient adequacy vitamin C, NAR5, with alternative policies, Polk County, Florida.








participate in either program, or received only a cash income supplement
(Policies M, N, and 0).
Nutritional status cannot be determined by the level of intake of one
or two nutrients. However, intake surveys have been useful in identify-
ing persons at nutritional risk. The Ten State Nutrition Survey [34] iden-
tified the most serious intake deficiencies to be iron, calcium, vitamin A,
and vitamin C. This finding paralleled mean intake levels for the sample
population, and projected values in the nutrient adequacy simulations.
Therefore, identification of changes in the nutrient adequacy of these
nutrients occurring with differing policy levels may be considered one
critical aspect of future food and nutrition policy.
Policy generalizations are summarized as follows. Projected nutrient
adequacy ratios indicated that reduction in the FSP bonus from $123 to
$91 did not result in concurrent reduction in the NARs. In contrast, the
NARs for calcium, vitamin A, and vitamin C increased. Joint participa-
tion in the FSP and EFNEP resulted in superior NARs for calcium,
iron, and vitamin A. However, participation in EFNEP with no food
program supplementation, or EFNEP with a cash income supplement,
promoted nutrient adequacy ratios near or above those calculated with
joint FSP and EFNEP participation. Nutrient adequacy did not gener-
ally increase with 12 to 18 months of EFNEP participation. A cash
income supplement alone resulted in inadequate nutrient adequacy ra-
tios for calcium, iron, and vitamin A.

SUMMARY AND CONCLUSIONS
Income, food expenditure, socioeconomic variables, food program
status, and 24-hour dietary recall information are collected regularly by
EFNEP nutrition aides in Polk County, Florida. These data were used
in regression analyses with total food expenditures, food group expendi-
tures, and the nutrient adequacy ratios as dependent variables. Results
of the regression analyses were then used to simulate alternative policy
situations with different sources and levels of income, Food Stamp Pro-
gram supplementation and alternative levels of participation in the Ex-
panded Food and Nutrition Education Program.
In this study, food stamp households spent an average $14.14 per
person, per month, for food. By comparison, eligible nonparticipating
households spent $33.22 per person, per month, for food. Total value
of purchased food differed between food stamp and nonparticipating
households by $195.94 versus $125.92, respectively.
For food stamp households, the Marginal Propensity to Expend from
money income (MPE,) was 0.06. It was negatively related to the value
of bonus stamps. The MPEB, at group means, was 0.45. A positive
bonus-family size interaction suggested that as family size increased, a

54







larger proportion of the discretionary income freed by FSP participation
was spent to purchase additional food. A negative bonus-income inter-
action indicated that the MPEB was largest at low income levels and
declined with increasing income. At mean bonus value, total food ex-
penditures increased by $18.61 for each additional person in food stamp
households.
The MPEj for nonparticipating households was 0.135. Family size
also explained significant variation in total food expenditures as each
additional person increased expenditures by $7.18.
Female-headed households spent significantly less than did male-
headed households in the nonparticipating sample. However, no expen-
diture difference of this type existed among FSP households.
Food group expenditures were also affected by participation not only
in the Food Stamp Program, but also by participation in the Expanded
Food and Nutrition Education Program. Table 6.4 summarizes the Mar-
ginal Propensities to Expend from money income and the Marginal
Propensities to Expend from bonus for all participation strata. It should
be noted that some of the MPEI's and the MPEB's were functions of
family size, income level, or bonus value. In these cases, the values were
calculated at monthly income of $330, bonus value $123, or family size
of four.
The Food Stamp Program appeared to be effective in increasing ex-
penditures for all food groups, except dairy products. Income coefficients
for the FS EFNEP sample were consistently slightly smaller than the
coefficients for the FS non-EFNEP sample. While the question of pro-
gram selectivity bias exists, EFNEP participation may have been one
reason for smaller income responses of food stamp households. EFNEP
seemed to work at two levels. Expenditures of eligible nonparticipating
FSP households tended to increase slightly, or decrease at a slower rate
than FS EFNEP households. Eligible nonparticipating FSP households
spent at lower levels than participating FSP households. EFNEP par-
ticipation may have encouraged nonparticipating FSP households to in-
crease food expenditures in order to improve nutritional levels. Con-
versely, FSP homemakers who concurrently participated in EFNEP,
had sufficient purchasing power to provide adequate quantities of food.
In this case, EFNEP aides could not only teach nutrition to improve
nutritional levels, but could also improve the efficiency with which the
household used bonus resources.
Nutrient adequacy of protein, calcium, iron, vitamin A, and vitamin
C were not consistently related to either income or bonus value. Food
expenditures were the most consistent indicator of protein adequacy in
three of four groups. For the FS EFNEP and non-FS EFNEP samples,
protein adequacy increased after food expenditures increased above $270
per month.

55








TABLE 6.4.--SUMMARY OF MPEI AND MPEB, FOR FOOD GROUP EXPENDITURES,
BY PROGRAM PARTICIPATION GROUP, POLK COUNT, FLORIDA,
1976

Marginal propensity Marginal propensity
to expend money to expend bonus
FGE income (MPEI) (MPEB)
Mean value Mean value
coefficients S.E. coefficients S.E.
Meat and protein
expenditures (FGE1)
FS EFNEP .060 (.038) .335 (.226)
FS non-EFNEP .128 (.048) .328 (.119)
non-FS EFNEP .062 (.035)
non-FS non-EFNEP .117 (.055)
Dairy expenditures (FGE2)
FSEFNEP .014 (.008) -.065 (.113)
FS non-EFNEP .033 (.017) .053 (.062)
non-FS EFNEP .027 (.008)
non-FS non-EFNEP .112 (.082)
Fruit and vegetable
expenditures (FGE3)
FS EFNEP -.015 (.021) .095 (.044)
FS non-EFNEP .029 (.014) .220 (.055)
non-FS EFNEP .043 (.020)
non-FS non-EFNEP .053 (.022)
Bread and grain
expenditures (FGE4)
FSEFNEP .013 (.026) .208 (.064)
FS non-EFNEP .039 (.030) .229 (.075)
non-FS EFNEP .056 (.029)
non-FS non-EFNEP .027 (.017)

Calcium adequacy decreased with increases in bonus value in FS
EFNEP households. Calcium adequacy increased during the first 14
months of EFNEP participation for non-FS EFNEP households. Each
dairy product lesson also increased calcium adequacy by 0.28 for this
group.
Iron adequacy responded positively to income in three of four partici-
pation groupings. The bonus response of FS EFNEP households was
positive up to income levels of $350. However, the bonus response of
FS non-EFNEP households was not significantly different from zero.
Months of participation in EFNEP were significantly related to increased
iron adequacy levels for both the FS EFNEP and non-FS EFNEP groups.
Bread and grain product lessons were effective in increasing the iron
adequacy ratio for food stamp households, but not for nonparticipating
families.

56








Vitamin A did not consistently respond to changes in income. As the
value of bonus stamps increased, vitamin A adequacy declined for the
FS EFNEP sample. The bonus coefficient was not significantly different
from zero for FS non-EFNEP households. EFNEP education increased
vitamin A adequacy throughout 12-14 months of program participation.
Vegetable lessons were also effective in increasing vitamin A adequacy
up to income levels of $503 per month for FS EFNEP families.
Vitamin C adequacy was positively related to income levels in non-
food stamp households. However, the income response explained less
variation than did the bonus and family size effects for food stamp fam-
ilies. The bonus-family size interaction term was positive for both the
FS EFNEP and FS non-EFNEP samples. In the FS EFNEP sample,
the family size coefficient was negative at bonus values less than $195.
In the FS non-EFNEP sample, the family size response was positive at
bonus values above $121. EFNEP participation improved vitamin C
adequacy for both the FS EFNEP and non-FS EFNEP groups.
Participation in the Food Stamp Program and the Expanded Food
and Nutrition Education Program affected total food expenditures, food
group expenditures and nutrient adequacy of protein, calcium, iron,
vitamin A, and vitamin C. The source and amount of income supple-
mentation and length of time of participation in the Expanded Food and
Nutrition Education Program also affected expenditures and nutrient
adequacy. If the two main goals of the Food Stamp Program remain
food expenditure supplementation and improvement of the nutritional
quality of low income households' diets, the choice of food policy is
likely to significantly affect both goals.
It would be desirable to have a stratified national low income data
base from which to test policy implications. Specific coefficients or actual
numbers of a case study of the type reported in this study should be
extrapolated to the national population with a great deal of caution.
However, a study of this nature can be used to identify the direction of
general relationships and the relative impact of alternative policy mea-
sures.
Cost-benefit information is necessary to determine economically fea-
sible policy variable combinations such as cost of education program
versus stamps. Additional information with respect to multiple program
participation and the joint effect on consumption is desirable. Neverthe-
less, information obtained from this study is a first step in the identifica-
tion of characteristics of food program participants and evaluation of
the effects of proposed alternative policy modifications on expenditures
and nutrient adequacy.





57
















GLOSSARY OF TECHNICAL TERMS


BONUS VALUE-the difference between the food stamp purchase requirement
and the face or redemption value of the food stamp coupons.
COMMODITY DISTRIBUTION PROGRAM (CDP)-the federal food aid program
in which selected commodities were distributed free of charge to low
income households.
EXPANDED FOOD AND NUTRITION EDUCATION PROGRAM (EFNEP)--a federal
antipoverty program administered by the Cooperative Extension Service.
Paraprofessional nutrition aides teach nutrition education, food econ-
omy, meal planning and preparation.
FOOD STAMP PROGRAM (FSP)-the federal food aid program in which house-
hold food purchasing power is increased through redemption of food
stamp coupons at retail food outlets.
MALNUTRITION-the state of impaired functional ability or deficient struc-
tural integrity or development resulting from inadequate supply of essen-
tial nutrients and calories to the body tissue.
MEAN ADEQUACY RATIO (MAR)-measure developed by Madden and Yo-
der. The MAR is the simple unweighted mean of ten nutrient adequacy
ratios, each truncated at a maximum value of 100 percent.
NUTRIENT ADEQUACY RATIO (NAR)-measure used by Madden and Yoder.
The NAR is the percentage of the Recommended Dietary Allowance
met by the family's diet for a particular nutrient.
NUTRIENT ACHIEVEMENT RATIO (NAR)-measure used by Lane, equal to
the NAR of Madden and Yoder.
NUTRITIONAL RISK-persons with nutrient requirements greater than the pop-
ulation at large. Examples include persons with anemia, persons with
inadequate growth patterns such as underweight or obesity, pregnancy
or low birth weight infants (less than 2500 grams).
POVERTY THRESHOLDS (levels)-a set of 124 thresholds arranged in a four
dimensional matrix consisting of family size, cross-classified by presence
and number of family members 18 years old, sex of head, and farm and
nonfarm residence. The one-and two-person families are further differ-
entiated by age of head (under 65 years old, 65 years and over). The
total family income of each family is tested against the appropriate pov-
erty threshold to determine the poverty status of the family.
PURCHASE REQUIREMENT-sliding scale cash payment required to obtain food
stamp allotment.





58












APPENDIX











TABLE A-l1.-STATISTICAL SUMMARY EQUATIONS [4.3, 4.4, 4.5, 4.6],
OLS COEFFICIENTS, DEPENDENT VARIABLE, FOOD GROUP
EXPENDITURES MEAT AND PROTEIN PRODUCTS, FGE1,
POLK COUNTY, FLORIDA, 1976.

Program status

FS non- non-FS non-FS
FS EFNEP EFNEP EFNEP non-EFNEP
Variable n = 50 n = 34 n = 73 n = 43

HI 0.2338 0.1277 0.1541 0.1170
(0.089)h (0.048) (0.061) (0.055)
HISQ --0.000153
(0.000079)
FS -6.0864 10.9340 3.5534
(6.616) (7.229) (2.472)
LNFS 10.9750
(6.622)
BFS 0.0081 -0.0828
(0.033) (0.042)
B 0.7550 0.6590
(0.207) (0.216)
BHI -0.0014
(0.001)
E2 -0.4468 0.0561 -4.9718 10.5760
(10.620) (9.985) (6.661) (10.750)
S2 -2.5481 -3.6485 -7.1349 -1.2527
(10.170) (9.980) (3.895) (9.294)
R2 21.5340 -13.4810 -16.6460 1.6099
(10.200) (9.895) (4.801) (9.091)
A 4.3317 -0.4646 0.8441 -2.6870
(3.676) (3.877) (2.165) (3.301)
W -2.2569 --0.2430 13.5710 15.1850
(11.080) (10.730) (5.184) (13.710)
Y -0.3834 0.6257 0.2101 -0.3903
(0.341) (0.637) (0.174) (0.426)
F -5.2609 -10.3730 -9.6584 6.3791
(8.507) (12.700) (3.960) (11.360)
H -0.2436 -19.7870 5.1952 23.6840
(15.000) (11.380) (7.665) (12.300)
LC2 -38.9450 -7.6433 -6.1079
LC3 -40.8530 1.0570 8.2611
LC4 -29.5800 -11.2430 6.9623
LC5 -53.5150 8.3088 2.1710
LC6 -43.0410 -0.1245 -5.3149
Group FStat. (0.985) (1.630) (0.870)

61








TABLE A-l.-continued

Program status

FS non- non-FS non-FS
FS EFNEP EFNEP EFNEP non-EFNEP
Variable n = 50 n = 34 n = 73 n = 43

MP 0.4414 -1.6470
(0.506) (0.982)
LNMP 23.1610
(14.450)
Intercept -21.310 -2.027 -22.990 3.694
(31.3) (41.3) (32.7) (33.3)

R2 .7147 .8641 .8080 .6430
F 6.262 5.986 11.963 3.242

aA complete definition of variables can be found on pp. 22-25.
bStandard errors appear in parentheses.


TABLE A-2.--STATISTICAL SUMMARY EQUATIONS [4.3, 4.4, 4.5, 4.6],
OLS COEFFICIENTS, DEPENDENT VARIABLE, FOOD GROUP
EXPENDITURES DAIRY PRODUCTS, FGE2, POLK COUNTY,
FLORIDA, 1976.

Program status

FS non- non-FS non-FS
FS EFNEP EFNEP EFNEP non-EFNEP
Variable" n = 50 n = 34 n = 73 n = 43

HI 0.1764
(0.096)b
HISQ -0.000216
(0.000148)
LNHI 4.3306 9.8692 8.2983
(2.847) (5.463) (2.655)
FS 8.1319 3.3109 -0.5904
(3.118) (1.902) (1.116)
LNFS 7.0896
(3.056)
B -0.0652 0.0533
(0.113) (0.062)
E2 -23.0920 -11.7990 -6.9045 1.3193
(7.793) (5.759) (2.591) (4.861)
S2 -14.9210 -3.7631 -1.2954 4.2826
(8.270) (5.934) (1.963) (4.557)

62








TABLE A-2.-continued.

Program status

FS non- non-FS non-FS
FS EFNEP EFNEP EFNEP non-EFNEP
Variable n = 50 n = 34 n = 73 n = 43

R2 -0.6684 1.2424 1.7690 8.9059
(7.223) (5.779) (2.096) (4.142)
A -3.6608 -2.1797 0.0274 0.7396
(2.675) (1.750) (1.075) (1.491)
W 7.741 -0.7602 0.4467 -2.9545
(7.903) (6.193) (2.599) (6.213)
Y 0.0752 0.0679 0.0376 -0.1353
(0.237) (0.346) (0.091) (0.193)
F 1.8903 0.1333 -1.6650 1.3229
(5.628) (5.620) (2.107) (5.140)
H -4.4840 8.8351 -0.8647 4.0254
(12.780) (6.153) (3.375) (5.557)
LC2 7.6085 7.2727 6.0458
LC3 6.7350 6.2126 4.3838
LC4 -3.5863 7.0671 11.2950
LC5 -1.8402 7.6590 5.1816
LC6 3.8303 8.1345 10.5270
Group F Stat. (0.810) (0.291) (0.396)
LNMP 0.9735
MP -3.5779 (1.568)
(1.925)
MPSQ 0.1056
(0.057)
LNE2 1.4126 -2.7174
(15.420) (1.769)
BLNE2 0.1334
(0.086)
HILNE2 -0.0466
(0.030)
Intercept 24.28 -41.16 -43.25 -21.20
(20.4) (29.6) (17.6) (21.4)

R2 .6278 .7552 .5946 .4180
F 3.478 3.277 6.298 1.167

"aA complete definition of variables appears on pp. 22-25.
bStandard errors appear in parentheses.



63








TABLE A-3.-STATISTICAL SUMMARY EQUATIONS [4.3, 4.4, 4.5, 4.6],
OLS COEFFICIENTS, DEPENDENT VARIABLE, FOOD GROUP
EXPENDITURES FRUIT AND VEGETABLE PRODUCTS, FGE,
POLK COUNTY, FLORIDA, 1976.

Program status
FS non- non-FS non-FS
FS EFNEP EFNEP EFNEP non-EFNEP
Variable" n = 50 n = 34 n = 73 n = 43

HI -0.0146 0.1095
(0.021)" (0.040)
HISQ -0.000104
(0.000047)
LNHI 8.6493 15.8690
(4.796) (7.240)
FS -0.9183 2.7102 1.4012
(1.670) (1.337) (1.234)
LNFS 21.8340
(7.980)
BFS 0.0238
(0.010)
B -0.1017 0.2198
(0.103) (0.055)
E2 5.3690 17.4290 0.9942 2.4072
(5.057) (5.056) (3.670) (5.448)
S2 -4.0231 -2.6334 -6.7936 0.6948
(4.957) (5.210) (2.397) (4.934)
R2 -3.1580 1.2464 1.7349 -8.4573
(4.954) (5.073) (3.023) (4.643)
A -2.8473 -2.8608 1.0518 -1.3647
(1.785) (1.537) (1.371) (1.663)
W 3.8995 --7.5777 -5.0081 -5.2396
(5.365) (5.437) (3.229) (6.983)
Y 0.0859 0.1705 --0.0102 0.3141
(0.170) (0.304) (0.108) (0.216)
F 9.3656 -1.7626 -1.8427 -0.9886
(4.027) (4.934) (2.365) (5.635)
H 11.692 -7.5806 9.2146 2.939
(7.172) (5.402) (4.665) (6.252)
LC2 15.7310 -20.0920 -0.1450
LC3 10.8490 -22.2530 -4.5881
LC4 27.5660 -15.2740 -6.8162
LC5 9.6920 -24.9190 -9.6621
LC6 -1.9440 -20.1560 -5.8254


64








TABLE A-3.-continued

Program status

FS non- non-FS non-FS
FS EFNEP EFNEP EFNEP non-EFNEP
Variable n = 50 n = 34 n = 73 n = 43

Group F. Stat. (2.38) (1.94) (0.49)
MP 0.0234 0.1936
(0.246) (0.308)
Intercept 1.324 -57.21 12.20 -77.23
(14.5) (26.0) (12.5) (43.3)

R2 .7174 .8479 .6924 .4567
F 7.034 5.925 6.399 1.513

"aA complete definition of variables appears on pp. 22-25.
bStandard errors appear in parentheses.



TABLE A-4.-STATISTICAL SUMMARY EQUATIONS [4.3, 4.4, 4.5, 4.6],
OLS COEFFICIENTS, DEPENDENT VARIABLE, FOOD GROUP
EXPENDITURES BREAD AND GRAIN PRODUCTS, FGE4,
POLK COUNTY, FLORIDA, 1976.

Program status

FS non- non-FS non-FS
FS EFNEP EFNEP EFNEP non-EFNEP
Variable n = 50 n = 34 n = 73 n = 43

HI 0.0129 0.0392 0.1159 0.1871
(0.026)b (0.030) (0.034) (0.068)
HISQ -0.000101 -0.000267
(0.000044) (0.000106)
FS -3.2337 -1.0518 -0.2018 1.5345
(3.503) (2.377) (1.305) (0.794)
BFS 0.0520
(0.016)
B -0.1082 0.2294
(0.115) (0.075)
E2 1.3706 6.3232 3.9775 -5.2218
(6.180) (6.855) (4.150) (3.458)
S2 -3.9770 -3.1866 -4.6821 -0.8518
(6.256) (6.318) (2.313) (3.241)
R2 -3.2531 -1.068 0.7328 -2.1930
(6.227) (3.146) (2.886) (2.947)

65








TABLE A-4.-continued

Program status

FS non- non-FS non-FS
FS EFNEP EFNEP EFNEP non-EFNEP
Variable n = 50 n = 34 n = 73 n = 43

A -3.9833 -2.7535 1.1795 -1.9218
(2.252) (2.002) (1.369) (1.061)
W 6.4113 0.1358 -2.6652 8.4744
(6.871) (7.630) (3.062) (4.420)
Y -0.3712 -0.7014 -0.1621 0.2183
(0.204) (0.393) (0.102) (0.137)
F 8.4483 11.6130 0.0706 3.155
(5.091) (7.104) (2.398) (3.656)
H -3.3260 -9.7140 1.0570 -10.8840
(9.403) (7.648) (4.547) (3.953)
LC2 1.8777 -13.4520 -9.6517
LC3 5.7269 -8.8046 -7.8191
LC4 19.5370 -5.7296 -7.0594
LC5 31.6580 1.5896 -9.2627
LC6 20.1320 -10.2180 -5.1174
Group FStat. (2.775) (2.470) (0.610)
MP 2.6663
(1.389)
LNMP -35.0790 6.3289
(19.470) (2.196)
LNE4 0.2822 4.9158
(2.320) (2.751)
Intercept 84.17 5.23 -11.51 -9.33
(36.6) (20.8) (13.3) (15.2)

R2 .6818 .7626 .6897 .5972
F 4.856 3.855 6.137 2.409

"aA complete definition of variables appears on pp. 22-25.
"bStandard errors appear in parentheses.











66








TABLE A-5.--STATISTICAL SUMMARY EQUATIONS [4.7, 4.8, 4.9, 4.10],
OLS COEFFICIENTS, DEPENDENT VARIABLE, NUTRIENT
ADEQUACY RATIO PROTEIN, NAR1, POLK COUNTY,
FLORIDA, 1976.

Program status
FS non- non-FS non-FS
FS EFNEP EFNEP EFNEP non-EFNEP
Variable n = 49 n = 43 n = 73 n = 71

HI 0.0100 0.000056
(0.006)b (0.000861)
HISQ -0.000018
(0.000009)
S-f HI 0.0828 0.3235
(0.057) (0.205)
SFS -0.1174 0.0498 -0.0546 0.0206
(0.084) (0.184) (0.059) (0.042)
B 0.008 0.0071
(0.002) (0.005)
R2 0.0740 -0.1773 0.2000 -0.0238
(0.158) (0.186) (0.117) (0.111)
E2 -0.1636 -0.1550 -0.1998 0.1711
(0.151) (0.241) (0.187) (0.155)
S2 -0.0948 0.2107 -0.0378 0.0640
(0.124) (0.177) (0.118) (0.121)
A 0.1079 0.1521 0.1021 -0.0138
(0.087) (0.080) (0.057) (0.051)
W 0.0130 0.2157 -0.1202 0.2951
(0.138) (0.195) (0.134) (0.203)
Y 0.0053 -0.0201 -0.0013 0.0144
(0.006) (0.017) (0.005) (0.007)
F 0.0681 -0.1846 -0.0120 -0.3679
(0.118) (0.271) (0.110) (0.153)
LC3 -0.4400 -0.4288 0.3894 -0.2290
LC4 -0.6753 -0.2793 -0.5468 -0.1657
LC5 -0.4513 -0.4041 -0.3493 -0.1195
LC6 -0.4699 0.3701 -0.1371 -0.3406
Group F Stat. (0.772) (0.695) (2.284) (0.648)
FE 0.0006 -0.0381 -0.0076 -0.0162
(0.002) (0.015) (0.006) (0.006)
FESQ 0.000072 0.000038 0.000054
(0.000033) (0.000021) (0.000017)
MP -0.0349
(0.028)

67









TABLE A-5.-continued

Program status

FS non- non-FS non-FS
FS EFNEP EFNEP EFNEP non-EFNEP
Variable n = 49 n = 43 n = 73 n = 71

LNMP 0.0267 0.5291
(0.098) (0.404)
LNE1 -0.0491 0.1274
(0.043) (0.071)
CME 0.0445 -0.6691 0.018 --0.3083
(0.038) (0.630) (0.022) (0.431)
Intercept 1.3910 4.2500 -0.3024 1.I':'"
(0.493) (1.330) (1.320) (0.542)

R2 .4711 .5022 .5139 .3822
F 1.247 1.009 2.618 1.822

"aA complete definition of variables appears on pp. 22-25.
bStandard errors appear in parentheses.

TABLE A-6.-STATISTICAL SUMMARY EQUATIONS [4.7, 4.8, 4.9, 4.10],
OLS COEFFICIENTS, DEPENDENT VARIABLE, NUTRIENT
ADEQUACY RATIO CALCIUM, NAR2, POLK COUNTY,
FLORIDA, 1976.

Program status

FS non- non-FS non-FS
FS EFNEP EFNEP EFNEP non-EFNEP
Variable' n = 49 n = 43 n = 73 n = 71

HI -0.000564 0.0075 -0.0058
(0.000660)b (0.006) (0.003)
HISQ -0.000013 0.000008
(0.000010) (0.000004)
LNHI 0.1854
(0.194)
FS 0.1154 0.1070 0.0132
(0.097) (0.093) (0.037)
LNFS -0.6169
(0.223)
B -0.0056 0.0026
(0.002) (0.005)
R2 0.1898 -0.0871 0.0358 0.0891
(0.166) (0.203) (0.134) (0.095)
E2 0.1178 -0.4194 0.2322 0.0486
(0.166) (0.256) (0.181) (0.136)

68








TABLE A-6.-continued

Program status

FS non- non-FS non-FS
FS EFNEP EFNEP EFNEP non-EFNEP
Variable n = 49 n = 43 n = 73 n =71

S2 0.1442 0.2680 0.1251 0.1319
(0.154) (0.193) (0.138) (0.111)
A 0.0420 0.1320 0.0958 0.0522
(0.'081) (0.087) (0.061) (0.448)
W -0.1966 0.4680 -0.0514 -0.4260
(0.153) (0.211) (0.160) (0.189)
Y 0.0097 -0.0199 0.0008 0.0114
"(0.007) (0.018) (0.006) (0.006)
F 0.1134 0.1158 0.0333 -0.2670
(0.120) (0.290) (0.133) (0.134)
H -0.0169 0.0530 0.3256 0.2184
(0.206) (0.328) (0.246) (0.173)
LC3 -0.0121 -0.5605 0.1646 -0.3751
LC4 -0.3390 -0.7925 -0.0374 -0.1232
LC5 -0.5634 -0.7005 0.3095 -0.2395
LC6 -0.1632 0.0808 0.2868 -0.3478
Group F Stat. (1.290) (0.515) (0.774) (1.627)
FE 0.000625 -0.0317 0.0033 0.0025
(0.002702) (0.015) (0.002) (0.002)
FESQ 0.000068
(0.000034)
MP 0.0299 -0.0627
(0.044) (0.032)
MPSQ -0.0014
(0.001)
LNMP 0.8763
(0.461)
LNE2 -0.1369 0.7499
(0.119) (0.463)
CME 0.0394 0.3031 0.0138 0.0803
(0.043) (0.687) (0.026) (0.354)
Intercept 0.5172 3.176 -1.824 0.8162
(0.520) (1.40) (1.40) (0.531)
---------------------------------------
R2 .6261 .4844 .5508 .4091
F 2.344 1.034 3.250 2.038

aA complete definition of variables appears on pp. 22-25.
bStandard errors appear in parentheses.


69








TABLE A-7.-STATISTICAL SUMMARY EQUATIONS [4.7, 4.8, 4.9, 4.10],
OLS COEFFICIENTS, DEPENDENT VARIABLE, NUTRIENT
ADEQUACY RATIO IRON, NAR3, POLK COUNTY,
FLORIDA, 1976,

Program status

FS non- non-FS non-FS
FS EFNEP EFNEP EFNEP non-EFNEP
Variable n = 49 n = 43 n = 73 n = 71

HI 0.0034 0.0037
(0.001)b (0.001)
HISQ -0.0000017 0.000005
(0.0000010) (0.000002)
LNHI 0.4502 0.1760
(0.151) (0.090)
FS 0.0421 -0.0025 0.0518
(0.051) (0.032) (0.191)
LNFS -0.2696
(0.301)
B 0.0036
(0.004)
BHI -0.000010 -0.000016
(0.000005) (0.000006)
BFS 0.000970
(0.000444)
R2 -0.1167 -0.1400 0.0516 0.0483
(0.090) (0.072) (0.062) (0.049)
E2 -0.0962 -0.1779 -0.1812 0.0301
(0.092) (0.099) (0.107) (0.070)
S2 -0.0763 0.1694 -0.0596 0.0066
(0.071) (0.073) (0.066) (0.057)
A -0.0777 0.6010 0.0354 -0.0666
(0.050) (0.031) (0.031) (0.023)
W 0.0943 0.0980 -0.0605 0.1475
(0.085) (0.082) (0.071) (0.098)
Y 0.0022 -0.0148 -0.00003 0.0066
(0.003) (0.007) (0.002) (0.003)
F 0.0194 -0.1120 0.0354 -0.0907
(0.070) (0.113) (0.062) (0.069)
H -0.2020 -0.0530 -0.0017 0.2159
(0.108) (0.115) (0.115) (0.089)
LC3 -0.1360 0.0744 -0.1638 0.0613
LC4 -0.2630 0.1442 -0.2492 -0.0185


70








TABLE A-7.--continued

Program status

FS non- non-FS non-FS
FS EFNEP EFNEP EFNEP non-EFNEP
Variable n = 49 n = 43 n = 73 n = 71

LC5 -0.4398 0.2979 -0.2428 0.1220
LC6 -0.2591 0.3965 -0.0416 -0.3700
Group F Stat. (1.959) (0.569) (1.950) (2.581)
FE -0.0073 -0.0114 -0.000127 -0.000369
(0.004) (0.008) (0.001223) (0.000806)
FESQ 0.000016 0.000019
(0.000007) (0.000018)
MP 0.0053
(0.004)
LNMP 0.0723
(0.057)
LNEI 0.0378 0.0130
(0.072) (0.040)
LNE4 0.1566 0.0566
(0.066) (0.078)
HILNE4 -0.000358 -0.000233
(0.00018) (0.000222)
CME -0.0188 -0.2845 0.0063 0.1526
(0.026) (0.267) (0.012) (0.183)
Intercept 0.5742 0.1762 0.4508 0.8198
(0.346) (0.716) (0.245) (0.274)

R2 .6414 .7458 .3699 .3817
F 1.789 2.934 1.361 1.818

aA complete definition of variables appears on pp. 22-25.
bStandard errors appear in parentheses.














71








TABLE A-8.-STATISTICAL SUMMARY EQUATIONS [4.7, 4.8, 4.9, 4.10],
OLS COEFFICIENTS, DEPENDENT VARIABLE, NUTRIENT
ADEQUACY RATIO VITAMIN A, NAR4, POLK COUNTY,
FLORIDA, 1976.

Program status

FS non- non-FS non-FS
FS EFNEP EFNEP EFNEP non-EFNEP
Variablea n = 49 n = 43 n = 73 n = 71

HI 0.0151 0.0029
(0.011)b (0.0016)
HISQ -0.000036
(0.000019)
LNHI -0.0023 -1.4818
(0.114) (0.764)
FS 0.3746 -0.1516 -0.3541
(0.164) (0.522) (0.208)
LNFS -0.4895
(0.345)
B -0.0175 0.0254
(0.005) (0.018)
BHI 0.000018 -0.000041
(0.000008) (0.000047)
R2 -0.1974 -0.1534 0.1196 -0.0134
(0.240) (0.351) (0.215) (0.195)
E2 -0.1333 -0.6058 0.1067 0.0729
(0.232) (0.462) (0.313) (0.273)
S2 0.3343 0.7328 0.1841 -0.1365
(0.209) (0.345) (0.218) (0.236)
A -0.3420 0.1533 0.0522 0.1031
(0.121) (0.149) (0.095) (0.089)
W 0.2357 0.4546 0.1576 -0.5627
(0.231) (0.373) (0.247) (0.378)
Y -0.0040 -0.0453 0.0020 0.0017
(0.010) (0.032) (0.009) (0.012)
F -0.1386 -0.0490 -0.2792 0.0928
(0.206) (0.545) (0.207) (0.270)
H 0.1582 0.2631 0.1938 0.2967
(0.291) (0.569) (0.390) (0.345)
LC3 0.5109 0.1879 0.1802 0.2428
LC4 0.5831 -0.1249 -0.0333 0.2667
LC5 -0.8984 -0.1902 0.5836 0.1073
LC6 0.9529 0.5623 0.0403 -0.5561
Group F Stat. (4.78) (0.69) (0.87) (0.54)

72








TABLE A-8.-continued

Program status

FS non- non-FS non-FS
FS EFNEP EFNEP EFNEP non-EFNEP
Variable n = 49 n = 43 n = 73 n = 71

FE 0.0090 -0.0749 -0.0022 0.0274
(0.004) (0.035) (0.004) (0.139)
FESQ 0.000178 -0.000082
(0.000082) (0.000042)
MP -0.0740
(0.050)
LNMP 0.4606 1.0254
(0.310) (0.727)
LNE3 0.4591 0.3131
(0.151) (0.164)
HILNE4 -0.0009 -0.0011
(0.0004) (0.0005)
CME -0.0350 0.5429 -0.0266 1.5781
(0.068) (1.91) (0.041) (0.819)
Intercept -0.9743 6.2400 -0.8405 6.7180
(0.897) (2.510) (1.470) (3.740)

R2 .6474 .5162 .3039 .2619
F 2.170 .881 1.157 0.971

aA complete definition of variables is found on pp. 22-25.
bStandard errors appear in parentheses.




















73








TABLE A-9.--STATISTICAL SUMMARY EQUATIONS [4.7, 4.8, 4.9, 4.10],
OLS COEFFICIENTS, DEPENDENT VARIABLE, NUTRIENT
ADEQUACY RATIO VITAMIN C, NAR5, POLK COUNTY,
FLORIDA, 1976.

Program status

FS non- non-FS non-FS
FS EFNEP EFNEP EFNEP non-EFNEP
Variable n = 49 n = 43 n = 73 n = 71

HI 0.0004 0.0092 -0.0040 -0.0147
(0.001)h (0.008) (0.002) (0.005)
HISQ -0.000028 0.000031
(0.000015) (0.000010)
FS -0.4241 -0.8703 -0.5265 0.6114
(0.231) (0.374) (0.178) (0.280)
HIFS 0.0013 0.0013 -0.0015
(0.001) (0.0004) (0.0007)
B -0.0094 -0.0131
(0.008) (0.010)
BFS 0.0022 0.0040
(0.001) (0.002)
R2 0.2257 -0.6449 0.0654 -0.0895
(0.240) (0.341) (0.212) (0.191)
E2 0.2060 0.1578 -0.1149 -0.2833
(0.236) (0.441) (0.309) (0.274)
S2 -0.0865 0.0898 0.3787 0.0457
(0.208) (0.321) (0.208) (0.225)
A -0.1677 0.2387 0.1151 0.0832
(0.124) (0.134) (0.105) (0.091)
W -0.2473 0.5162 -0.2128 -0.1508
(0.236) (0.351) (0.239) (0.383)
Y -0.0138 -0.0129 0.0037 0.0221
(0.010) (0.284) (0.008) (0.012)
F 0.0242 -0.3377 0.0493 0.0129
(0.200) (0.511) (0.203) (0.283)
H -0.3533 0.4087 0.5768 0.6599
(0.294) (0.519) (0.383) (0.346)
LC3 0.9688 1.2535 -0.1888 -0.3893
LC4 1.0767 0.9488 -0.6325 -0.5604
LC5 0.6622 0.6534 -0.3847 -0.4644
LC6 0.9354 -0.4088 0.1785 -0.6696
Group F Stat. (1.580) (0.800) (1.016) (1.044)



74








TABLE A-9.-continued

Program status

FS non- non-FS non-FS
FS EFNEP EFNEP EFNEP non-EFNEP
Variable n = 49 n = 43 n = 73 n = 71

FE 0.0027 (0.008) 0.0067 0.0024
(0.003) -0.0078 (0.004) (0.003)
LNMP 0.3761 0.3063
(0.177) (0.171)
LNE3 0.0697 -0.0540
(0.095) (0.057)
CME 0.0550 1.0625 0.0340 1.0001
(0.062) (1.140) (0.040) (0.721)
Intercept 0.0073 2.9610 1.4930 1.0840
(0.994) (2.070) (0.971) (1.110)

R2 .5041 .5719 .3822 .3458
F 1.307 1.366 1.758 1.447

"aA complete definition of variables appears on pp. 22-25.
bStandard errors appear in parentheses.




























75



















REFERENCES


[1]. Adrian, J. 1974. "Specification of the Effects of Income and Other
Socioeconomic Factors on the Consumption of Food Nutrients."
Unpublished Ph.D. dissertation. Knoxville: University of Tennes-
see.
[2]. Bayless, T. M., and D. M. Paige. 1971. "Disaccharide Intolerance in
Feeding Problems," Proceedings of the Western Hemisphere Nu-
trition Congress III, Miami, Florida. September.
[3]. Becker, G. S. 1971. Economic Theory. New York: Alfred A. Knopf,
Inc.
[4]. Blackorby, C., G. Lady, D. Nissen, and R. R. Russel. 1970. "Homo-
thetic Separability and Consumer Budgeting," Econometrica. 38(4)
486-472.
[5]. Council on Foods and Nutrition. 1970. "Malnutrition and Hunger
in the United States," Journal of the American Medical Associa-
tion. 213 (2):1-5.
[6]. Davidson, D. S., J. Livermore, P. Anderson, and S. Kaufman. 1962.
"The Nutrition of a Group of Apparently Healthy Aging Persons,"
American Journal of Clinical Nutrition. 10(March) :181-199.
[7]. Duval, E. M. 1967. Family Development. New York: J. B. Lippin-
cott Co.
[8]. Engles, J. A. 1969. "Diets of Low Income Families," presented at
the 46th Annual Agricultural Outlook Conference, Washington,
D.C., February 19.
[9]. Expanded Food and Nutrition Education Program, Polk County, Flor-
ida. Annual County Report, December 1975.
[10]. Feaster, J. G., and G. B. Perkins. 1972. Families in the Expanded
Food and Nutrition Education Program: Comparison of Food
Stamp and Food Distribution Program Participants and Nonpartici-
pants. USDA-ERS Report No. 246. Washington, D.C.: Govern-
ment Printing Office.
[11]. Federal Register. 1976. Vol. 41(90).
[12]. Federal Register. 1977. Vol. 42(85).
[13]. Gorman, W. M. 1959. "Separable Utility and Aggregation," Econo-
metrica. 27(3):469-481.
[14]. Houthakker, H. S. 1960. "Additive Preferences," Econometrica.
28(1):62-87.
[15]. Johnston, J. 1972. Econometric Methods. Second edition New York:
McGraw Hill Book Company.
[16]. Lancaster, K. L. 1966. "A New Approach to Consumer Theory,"
Journal of Political Economy. 74:132-157.

77








[17]. Lane, S. 1975. Food Distribution and Food Stamp Program Effects
on Nutritional Achievement of Low Income Households in Kern
County, California. Unpublished paper.
[181. Leidenfrost, N. 1975. EFNEP: Accomplishments and Future Needs.
USDA Extension Service HE-89.
[19]. Lewin, K. 1942. "Forces Behind Food Habits and Methods of
Change," in The Problem of Changing Food Habits. National Re-
search Council Bulletin 108:35-65.
[20]. Madden, P., S. J. Goodman, and H. A. Gunthrie. 1975. "Validity of
the 24-Hour Dietary Recall with Elderly Subjects," Journal Series
No. 4882. The Pennsylvania State University Agriculture Experi-
ment Station.
[21]. Madden, P., and M. Yoder. 1972. Program Evaluation: Food Stamps
and Commodity Distribution in the Rural Areas of Central Penn-
sylvania. The Pennsylvania State University Agricultural Experi-
ment Station Bulletin 780.
[22]. Maslow, A. H. 1943. "Theory of Human Motivation," Psychological
Review. 50:105-127.
[23]. National Center for Health Statistics. 1972. Health Characteristics
of Low Income Persons. Vital and Health Statistics Series 10, No.
74, DHEW No. (HMS) 73-1500. Washington: U.S. Government
Printing Office.
[24]. National Research Council. 1973. Recommended Dietary Allow-
ances. Food and Nutrition Board, National Academy of Sciences.
Washington.
[25]. Neenan, P. H. 1977. "Impact of the Food Stamp and Expanded
Food and Nutrition Education Programs on Nutrient Demand of
Low Income Households: Polk County, Florida." Unpublished
M.S. Thesis. Gainesville: University of Florida.
[26]. Paarlberg, D. 1976. "Food and Economics," presented at American
Dietetic Association symposium, Food: The Realities of Economics,
Politics and Population. Boston. October 12.
[27]. Southworth, H. M. 1945. "The Economics of Public Measures to
Subsidize Food Consumption," Journal of Farm Economics. 27(1)
38-66.
[28]. Strotz, R. H. 1959. "The Utility Tree-A Correction and Further
Appraisal," Econometrica. 27(3):482-488.
[29]. Swanberg, K. G., and E. Shipley. 1975. "The Nutritional Status of
the Rural Family in East Cundinamarca, Colombia," Food Re-
search Institute Studies. 14(2) :111-125.
[30]. U.S. Bureau of the Census. 1972. Census Population: 1970. Detailed
Characteristics Florida. Final Report PC (1) D11. Washington.
U.S. Government Printing Office.
[31]. U.S. Bureau of the Census. 1975. Current Population Reports. Series
P-60, No. 98, "Characteristics of the Low Income Population:
1973," Washington: U.S. Government Printing Office.
[32]. U.S. Food and Nutrition Service. 1975. Food Stamp Program. Wash-
ington: U.S. Government Printing Office.
[33]. West, D., and D. Price. 1976. "The Effects of Income, Assets, Food
Programs and Household Size on Food Consumption," American
Journal of Agricultural Economics. 58(4):725-730.


78








[34]. Youland, D., and A. Engle. 1976. "Practices and Problems in
HANES," Journal of the American Dietetic Association. 68(1):
22-25.
[35]. Young, C. M., G. C. Hagan, R. E. Tucker, and W. E. Foster. 1952.
"A Comparison of Dietary Study Methods-Dietary History vs.
Seven Day Record vs. 24-Hour Recall," Journal of the American
Dietetic Association. 28:218-221.
















































79





















Institute l d d Agricultural SEioca


TEACHING IFA S
RESEARCH
EXTENSION













This public document was promulgated at an annual cost of
$5,046.75 or a cost of $1.68 per copy to provide information on
the effects the Food Stamp Program and the Expanded Food and
Nutrition Education Program on food expenditure behavior and
nutritional status of low income rural households in Florida.





University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs