Factors influencing school and student participation in the school breakfast program 1977-78

              ,~ United States
r~U.j}) Department of
~ Agriculture
Food and Nutrition
Service
Office of Policy,
Planning and
Evaluation
Factors Influencing
School and Student
Participation in the
School Breakfast Program
Depository
I L J I ./
FACTORS INFLUENCING SCHOOL AND STUDENT PARTICIPATION
IN THE SCHOOL BREAKFAST PROGRAM 1977-78
AprU 1980
Office of Policy, Planning and Evaluation
Food and Nutrition Service
U. S. Department of Agriculture
ACKNOWLEDGEMENTS
Many people contributed to the successful completion of this study.
The principals and school foodservice managers
of the 515 schools in the sample responded to
the survey and the follow-up contacts of FNS.
The clerical staff members of the School Programs
Division prepared the questionnaires for mailing
and monitored the receipt of completed
questionnaires.
Suzanne Badini and Elaine Sinclair of the School
Programs Division assisted in telephone follow-up
and questionnaire editing.
Eleanora Battle and Carol Dubay of the Automated
Data Processing Division organized the keypunching
and computer support for the project.
Toni Walls and Jacqueline Eaton typed and retyped
the final report a number of times.
Ann Kilroy prepared the tables for the report.
Beatrice Rogers and Frances Zorn analyzed the
data and authored the report.
Frances Zorn was the project director of the study.
TABLE OF CONTENTS
ACKNOWLEDGEMENTS .
TABLE or COllTENTS
LIST OF TABLES . .
EXECUTIVE SUMMARY.
CHAPTER I
CHAPTER II
CHAPTER III
INTRODUCTION.
OBJECTIVES ..
SAMPLING AND METHODOLOGY.
STATISTICAL TREATMENT
DESCRIPTION OF SURVEY RESULTS
·.
Page
i
iii
v
vii
1
1
3
3
5
CHARACTERISTICS OF BREAKFAST AND NON-BREAKFAST SCHOOLS. 5
Grade Level. . . . 5
Economic Need. . . 5
Public Vs. Private Schools 5
School Enrollment. 5
Attendance 8
Facilities . . . . 8
Logistics 10
Competitive Foodservice. 10
School Breakfast and Other School Feeding Programs 11
Combined Effects of School Participation in the SBP. 15
STUDENT PARTICIPATION IN THE BREAKFAST AND LUNCH PROGRAMS 17
Grade Level. . 17
Economic Need. . 19
Meal Price . . . 19
School Enrollment. 19
Facilities . . . . 19
Competitive Foodservice. 22
Logistics. . . . . . . . 25
Combined Effects of Student Participation in the SBP and the NSLP 25
Breakfast Participation and Special SBP Program Characteristics 30
NON-BREAKFAST SCHOOLS . . . . . . . 30
Prior Participation. . . . . . 30
Reasons for .Non-Participation. 30
APPENDIXES ....
SURVEY INSTRUMENT
SELECTED BIBLIOGRAPHY ON THE SCHOOL BREAKFAST PROGRAM
35
35
53
iii
LISTS OF TABLES
Table 1
Table 2
Table 3
Table 4
Table 5
Table 6
Table 7
Table 8
Table 9
Table 10
Table 11
Table 12
Table 13
Table 14
Table 15
Table 16
Table 17
Table 18
Table 19
Table 20
Proportion of Meals Served Free and at a Reduced
Price in the School Breakfast Program and the
National School Lunch Program by Fiscal Year •
Percent of Breakfast and Non-Breakfast Schools
by Grade Level • • • • • •
Percent of Needy Students in Breakfast and
Non-Breakfast Schools. . • .
Average Enrollment of Breakfast and
Non-Breakfast Schools ..••.
Average Enrollment by Economic Need in Breakfast
and Non-Breakfast Schools •••
Average Daily Attendance in Breakfast and
Non-Breakfast Schools. • • • • • • • • • •
Percent of Students Using Different Methods of
Transportation to Breakfast and Non-Breakfast Schools.
Percent of Schools at Each Grade Level Which Have
Competing Food Service • .
Percent of Schools Having Open Campus by Grade Level
Breakdown of Food Programs Offered by Schools.
Percent of Schools Serving the National ~choc:!. !,u~,cJ-.
Pr~gram cy Grade Level . • • • • • • .
Average Price of Lunch In Breakfast and Non-Breakfast
Schools by Grade Level
Parameters of the Regression Model Testing Factors Influencing
School Participation in the School Breakfast Program • • • •
Average Daily Student Participation in the School Breakfast
and National School Lunch Programs • • • • • • •
Average Daily NSLP ·Participation in Breakfast and
Non-Breakfast Schools • • • . • •
Average Daily School Breakfast Program Participation in Needy and
Non-Needy Schools. • • ••••••
Average Daily National School Lunch Program Participation
by Level of Need for Breakfast and Non-Breakfast Schools •
Average Daily Participation in School Breakfast Program by
Location of Meal Consumption • • • • • • • • •
Average Daily Participation in the National School Lunch Program
by Location of Meal Consumption • • • • • • • •
Average Daily Participation in the School Breakfast F=ogram
by Open or Closed Campus • • • • • • • • • • • • • • • •
2
6
6
7
7
9
9
12
12
13
14
16
16
18
18
20
20
21
23
23
v
Table 21
Table 22
Table 23
Table 24
Table 25
vi
Average Daily School Breakfast Program
Participation with and without Competing Foodservice
Average Daily National School Lunch Program
Participation in Schools with and without Competing
Foodserv:l ce . . . . • . . • . . .
Parameters of Regression Model Showing Effect of School
and Program Characteristics on Average Daily Student
Breakfast Participation . . . . .
Parameters of Regression ~lode! Showing Effect of
School and Program Characteristics on Lunch Participation
Primary Reason for Non-Participation in the School
Breakfast Program . • . . . . . . . . . . . . . .
24
24
27
29
31
EXECUTIVE SUMMARY
SCHOOL PARTICIPATION
IN THE SCHOOL
BREAKFAST PROGRAM
Economic Need
Grade Level
Enrollment Size
Snack Bars
In 1978, the Food and Nutrition Service (FNS) conducted a survey of a
nationally representative sample of 625 schools to determine the factors
influencing school and student participation in the School Breakfast
Program (SBP). Principals and school foodservice managers for the schools
in the sample were mailed a questionnaire and were asked to provide data
for the month of October 1977. Five hundred and fifteen schools responded
to the questionnaire.
The data were analyzed to quantitatively measure factors which effect:
1. A school's participation in the SBP;
2. A student's participation in the SBP; and
3. A student's participation in the National School Lunch
Program (NSLP).
In addition, opinion data that were also collected as part of the survey were
analyzed to discover the major perceived barriers to the SBP, particularly
when a lunch program was already in operation.
Four factors distinguished breakfast from non-breakfast schools.
Whether or not a school is needy is the most important feature that
differentiates breakfast from non-breakfast! schools. Significantly more
economically needy schools offered the SBP. Schools that had 40 or more
percent of their enrollment approved to receive free or reduced price meals
were defined as economically needy. Forty-six percent of breakfast schools
and 16 percent of non-breakfast schools showed at least this level of poverty.
Elementary schools were significantly more likely to offer the SBP. Fifty­five
percent of SBP schools were elementary schools. Only 9.3 percent were
high schools and 24.8 percent were junior high schools. The remainder of
breakfast schools were special un~raded schools and consolidated schools that
contained grades from kindergarten through high school.
Breakfast schools were also significantly more likely to be larger than
non-breakfast schools. Breakfast schools had an average enrollment of 585;
non-breakfast schools averaged 484 students. ·
Schools that had a snack bar on the premises were more likely to offer the
breakfast program. There is no easy explanation for this finding, and no
data from the survey were collected which would allow further analysis of this
finding. It may be, however, that snack bars are the facility through which
breakfast is served in some schools.
Several factors that were hypothesized as characteristics which might dis­tinguish
breakfast from non-breakfast schools proved not to influence a
school's decision to participate in the SBP.
1 Non-breakfast schools include schools which offer no meal programs.
vii
Transportation
Facilities
Supervisory Time
FACTORS INFLUENCING
STUDENT PARTICIPATION
IN THE SCHOOL BREAKFAST
PROGRAM
Economic Need
Grade Level
Price
Length of the
Breakfast Period
Open Campus Policy
viii
No differences between breakfast and non-breakfast schools were apparent in
the modes of transportation students used to get to school. Most students
in non-breakfast schools were dri.vC'n to school, but the percentages of
students bused to school were not significantly different. This was true
for students riding both short and long periods on a bus.
Breakfast and non-breakfast schools were not distinguished by the availability
of kitchen facilities. Ten percent of breakfast schools and 13 percent of
non-breakfast schools did not have kitchens. The availability of a cafeteria
was also examined in the analysis. More non-breakfast schools than breakfast
schools did not have a cafeteria, but lack of a special eating room did not
prohibit a school from offering the SBP.
The amount of time between the arrival of teachers and administrators and
the start of classes was used as an indicator of the amount of time
available for supervision before class. There were no significant differences
on this amount of time in breakfast and non-breakfast schools.
Five factors were found to influence a student's participation in the School
Breakfast Program.
The chief factor influencing student participation in the SBP was the economic
need of students. Economically needy schools were defined as schools with at
least 40 percent of the students enrolled eligible for free and reduced price
meals. Students in these schools had significantly higher participation in
the SBP than non-needy schools. Needy schools had average daily participa­tion
rates of 45.1 percent of the school's attendance, compared to 22.1
percent participation by students in non-needy schools.
On a daily basis, more than twice as many elementary and junior high school
students participated in the SBP. As a percent of average daily attendance,
35.9 and 33.3 percent of the elementary and junior hi~h school students,
respectively, participated, compared to 15 percent of high school students.
As might be expected, non-needy student participation in the breakfast program
is price responsive; that is, participation is higher when the price is lower
for students who pay full price for their meals.
Two variables which may affect how much time a child may have to eat, the
degree of busing of students, and the length of time between the start of
the breakfast serving period and the beginning of classes, influenced student
participation in the breakfast program. Schools with 9hildren who ride buses
to school ate breakfast more often than students who used other modes of
transportation. Longer periods of time elapsing between the beginning of
breakfast and the start of classes were also positively associated with
SBP participation.
An open campus policy, where students are permitted to leave the school
grounds during the day, significantly reduces student breakfast participation,
even when the grade level and economic need of schools are taken into account.
The average participation of schools with an open campus policy is 28.5
percent compared to 36.5 percent of schools with a closed campus policy.
Meal Preparation Site
Availability of
Vending Machines
and Snack Bars
FACTORS INFLUENCING
STUDENT PARTICIPATION
IN THE NATIONAL SCHOOL
LUNCH PROGRAM
Meal Preparation Site
Availability of
Snack Bars
Length of the
School Day
The following factors did not influence student participation in the School
Breakfast Program:
Participation was not affected by whether the breakfast meal was prepared
on the school premises or prepared elsewhere and delivered to the school.
This is not surprising since breakfast is a relatively simple meal that can
be served with components that require little or no cooking. Generally,
85 percent of the schools prepared the breakfast meal on-site and 15 percent
prepared them at places other than the school.
The presence of vending machines or snack bars does not affect student
breakfast participation. When vending machines were available, 34.8 percent
of the students participated in breakfast compared to 33.7 percent of students
where the machines were not accessible. When snack bars were available, 27.4
percent of students participated compared to 34.6 percent when they were not
available. This difference, however, was not statistically significant. It
is not known whether the snack bars were open during the meal period.
The four factors that influenced a student's propensity to participate in
the SBP also influenced student participation in the NSLP. These variables
are economic need, grade level, the price charged, and an open campus policy.
The type of associations were the same: for example, students in economically
needy schools participated to a greater extent than students in non-needy
schools; lower grade students participated at a higher rate than higher grade
students. The strength of these associations, however, was usually weaker
for the lunch program. This was because student participation in the lunch
program is much higher than participation in the breakfast program. For
example, student participation was higher in both the lunch and the breakfast
program when there was a high percentage of needy children enrolled. However,
non-needy children were more likely to participate in the lunch program, and
therefore, the relationship of need to high participation because less
important.
There were, however, two interesting trends that were apparent in schools that
offered both the NSLP and the SBP. First, they had higher levels of partici­patiop
than schools that offered only the NStP. Second, lunch prices to
students who pay full price for their meals were lower, even after the effects
of economic need and grade level were considered.
Several other additional factors affected NSLP participation.
Participation in lunch, unlike breakfast, is significantly higher if the
meals are prepared at the school, 67 percent, compared with 59 percent in
schools that prepare lunch in locations away from the school.
Lunch service participation is negatively affected by the availability of a
snack bar. Only 49 percent of the students in attendance participated in the
NSLP when snack bars were available, compared to 61.3 percent when they were
not available.
Schools with the longest time between the start and end of classes are more
likely to have higher lunch participation.
ix
The Availability of
Vending Machines
Meal Consumption Site
Length of the
Lunch Period
PERCEIVED BARRIERS TO
SCHOOL PARTICIPATION
IN THE SBP
X
The following factors did not influence student participation in the NSLP:
The availability of vending machines did not significantly affect student
participation in the lunch ' program. When vending machines were available,
57.2 percent of those attending ate lunch, compared to 60.8 percent when
they were not available.
Lunch participation was examined separately for breakfast and non-breakfast
schools concerning meal consumption sites. In non-breakfast schools, the
location where meals are eaten is not a factor in the rate of participation.
In breakfast schools, high lunch participation rates were found if students
ate in a cafeteria, 75.4 percent, a multi-purpose room, 68.5 percent, or a
classroom, 65.8, rather than in a gymnasium, on the school grounds or some
other location.
The length of the school lunch period, the amount of time each student has
to be served and to eat, had no significant effect on the rate of student
participation.
The non-breakfast schools responding to the survey were asked the major
reason for non-participation in the SBP. The most frequently reported
answer was that the program was not needed, that parents in the community
were able to provide breakfast for their children, or that parents had
not requested the program. This response was mentioned three times as
often as the next most common response for a total of 36.4 percent of the
responses. Other responses included busing problems, lack of interest in
the program, lack of potential participation, and inadequate facilities.
CHAPTER 1: INTRODUCTION
OBJECTIVES
The USDA School Breakfast Program (SBP) was originally authorized as a pilot
program under Section 4 of the Child Nutrition Act of 1966. First priority
for participation was extended to schools in poor economic areas, and to
t hose schools in which a substantial proportion of students enrol led travell­ed
long distances daily. Federal reimbursement covered food costs for morning
meals in schools, and the Secretary of Agriculture established a payment rate
of 15 cents per meal for all meals regardless of whether they were served
free, at reduced price or at full price .to children. In schools determined
to be in "severe need" by State agencies, Federal financial assistance,
according to the 1966 law, was made available for up to 80 percent of the
operating cost of the program (food, l abor and other costs).
Between 1968 and 1975, legislation continued to authorize the SBP as a pilot
project, but extended its scope first to schools in which there was a need to
improve the nutrition and dietary practices of the children 'of working
mothers, and children from low income families, and later, to all schools that
applied for the program.
The .funding structure of the program grew to parallel that of the National
School Lunch Program (NSLP) . All meals received a basic reimbursement and
each free and reduced price meal received an additional Federal payment.
Reimbursement levels to States and localities ~ere determined by the number of
meals served. In addition, in schools designated to be in "severe need,"
increased reimbursement was made available for up to 100 percent of SBP
operating costs.
In 1975 Congress permanently authorized the SBP, extended the program to
public and private residential child care institutions, and directed the
Secretary to carry out planned outreach activities, in cooperation with State
educational agencies, that would make the program available in all schools
where it w.::s needed to provide adequate nutrition for children in attendance.
Although the School Breakfast Program has been available to all schools since
September 1972, only 25,000 s chools participated in October 1978, compared to
92,865 schools in the National School Lunch Program (NSLP). In these schools
the proportion of breakfasts served free and at a reduced price is substan­tially
higher than the comparable proportion of lunches in the NSLP. The
data presented in Table 1 illustrate this point.
USDA placed emphasis on the need to increase participation i ~ the SBP. How­ever,
the Department lacked systematic quantitative data to assist in the
development o.f outreach policies regarding the types of schools and children
participating in the program or to assess the nutritional contributions of
the SBP meal pattern. The information available to FNS related only to the
number of schools and children participating and the amount of Federal funds
expended. No data were available concerning program operations at the local
level. nor were data available concerning characteristics of schools or
students that delineate SBP participation.
This study of the School Breakfast Program focuses on three major objectives:
1. To develop a profile of schools and students that do and do not
participate in the SBP.
2. To identify factors that distinguish schools with a breakfast pro­gram
from those without a bn~ak fast program.
3. To identify barriers to participation in the SBP experienced by
schools and students.
1
Fiscal
Year
1967
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
TABLE 1
PROPORTION OF MEALS SERVED FREE AND AT A REDUCED PRICE
IN THE SCHOOL BREAKFAST PROGRAM AND THE NATIONAL SCHOOL LUNCH PROGRAM
BY FISCAL YEAR
Number of Schools Average Daily Average Daily Meals
Participating Attendance in Schools Meals Served Free and
sBpa/ NSLPb/ SBP NSLP SBP NSLP SBP
Millions Millions
752 72,944 .348 32.6 .077 18.5 76.1
3,325 74,861 1.382 36.0 .310 20.1 71.0
4,270 75,593 1.927 37.1 .485 20.9 71.6
6,609 79,924 3.058 39.0 .867 22.3 76.3
7,865 83,333 3.885 40.5 1.048 22.9 78.5
9,706 86,381 4.589 40.5 1.190 23.2 83.4
11,900 87,579 4.993 40.7 1.385 23.0 82.8
14,134 88,666 5.910 41.3 1. 774 23.2 82.1
16,835 89,718 7.688 41.40 1.992 23.5 83.4
23,729 92,973 8.856 41.11 2.74 26.8 85.1
28,200 92,865 10.410 41.24 3.13 26.2 82.8
~/ School Breakfast Program
~/ National School Lunch Program
Source: "Monthly Report on Child Nutrition Operations," Form FNS-10, From 1967-1978.
2
Served
Reduced Price
NSLP
Percent
12.2
15.1
20.7
26.1
32.4
35.0
37 .1
40.28
42.36
44.61
43.26
SAMPLING AND
METHODOLOGY
STATISTICAL TREA~NT
In addition to meeting these objectives, the study was undertaken to update
information obtained in previous surveys of school foodservice operations and
to assess recent changes in these operations. Prior to this study, the most
recent comprehensive studies of child nutrition programs were the "1972
National School Lunch Program Survey" and the 1975 "Special Milk Program
Evaluation and National School Lunch Program Survey." Both of these studies
were conducted by the Food and Nutrition Service. Other related literature
is cited in the bibliography.
The study was designed to survey a total of 650 schools through a mailed
questionnaire sent to a group of non-breakfast schools and a second group of
breakfast schools. The number of schools in each group was approximately
equal.
To choose the schools in the survey the following procedure was used:
1. The data base was the School Universe Director~ compiled and main­tained
by Curriculum Information Center, Inc. of Denver, Colorado, which
consisted in December 1975 of 100,974 public, parochial, and private schools.
2. Through a random selection procedure, 2,869 schools were drawn from
the directory and screened for the presence of the SBP by FNS RegionalOffices.
This screening resulted in the identificaticn of 379 &BP schools, and 2,L,R.5
non-SBP schoo:!.s (5 schools \vcre closed).
3. Each of the two groups was randomly ordered and through a second
random selection procedure 328 non-SBP and 324 SBP schools were selected for
the final survey sample.
4. Because of delays in clearance of the procedures to be used in
collecting the data, the survey was not mailed to the schools until December
1977. In several instances State agencies assisted the Department in gather­ing
the data by distributing or collecting the forms. However, FNS personnel
and consultants completed all editing and analysis·. A copy of the survey
instrument appears in the appendix of this report. Five hundred and fifteen
responses were received, 247 from schools with the breakfast program and 268
from schools without the program. These responses were edited for complete­ness
and accuracy of data. Where possible, incomplete responses were followed
up and completed by telephone.
At the time of the survey there were about six times as many non-breakfast
as breakfast schools in the country, but the sample used in this study con­tained
an approximately equal number of breakfast and non-breakfast schools.
Non-breakfast schools include schools that offer no Federal programs at all.
This occurred because in the second random selection procedure a smaller
proportion of non-breakfast schools than of breakfast schools was selected
from those chosen in the first random selection procedure. This dispropor­tionality
was necessary to insure that the sample contained an adequate
number of breakfast schools so that the analysis planned could provide an
accurate description of School Breakfast Program characteristics. Where the
two groups are analyzed separately, the fact that the groups were sampled
differently does not matter. But where the whole sample of schools is
discussed, the non-breakfast schools must be weighted more heavily than the
breakfast schools, so that the results will be representative of the entire
population of schools. In the sample, the two groups are equa~ to represent
the population, each breakfast school must count one sixth as much as a non­breakfast
school. Once procedures to weight the sample were implemented, the
reproportioned sample yielded 447 non-breakfast and 68 breakfast schools, a
total of 515 schools. The tables included in this report indicate whether
or not the number of schools listed as respondents was weighted for the
particular analysis.
3
•
4
Some of the information pre·sented in this study is descriptive. It presents
the similarities and differences between breakfast and non-breakfast schools.
An effort is made, however, to use some of these differences to explain why
the schools do or do not offer the SBP. To find out whether a variable had
explanatory power, statistical tests were used to determine whether the
differences between the two groups are significant or not. The level of
significance indicates the likelihood that an observed difference between
groups could have occurred by chance. For example, a significance level of
.05 indicates that there is a five percent probability of observing such a
difference in the population represented by the same. Significance is
usually reported as alpha (a).
There -is no set level at which a variable can be said to be statistically
significant. Researchers must make their own decision about how rigorous to
be in determining statistical significance. Two commonly used values are
a= .05 and a= .01. This means that at a= .05 and a= .01, respectively,
that 95 and 99 times out of 100 an observed difference between groups did
not occur by chance. These values are arbitrary, but have been adopted as
a convention by most statistical analysts. When possible, we have reported
the exact level of significance so that the reader may make his or her own
interpretation. In the test, we assume that levels higher than a= .05
are not statistically significant, but we discussed some interesting
findings with a significance level up to a = .10. Levels of significance are
reported in the text and tables of this report by both the confidence level
percentage and the applicable a level.
There is a wide variety of statistical tests to measure the significance of
a difference between two groups on one or several dimensions. In this study,
we have used the chi-square test for nominal and ordinal-level variables.
This test looks at the observed distribution of cases in the two groups, and
compares it with th~ distribution that would be expected if both groups were
really the same. We have used the student's t-test to determine the signi­ficance
of differences between means. Obviously, means can only be calcu­lated
for variables measured at the interval level or above. Analysis of
variance has been used to test the influence of nominal independent variables
(such as categories) on variance in a dependent variable measured at the in­terval
level or better. We have used multivariate ordinary least squares
regression analysis to determine the degree of association of each of a
number of independent variables on a given dependent variable with all the
other independent variables controlled.
The mathematical basis and assumptions underlying these these are too
technical for a blief presentation nere, but may be found in any standard
statistics text.!
!/ For example, Forcese and Richer, Social Research Methods (Prentice-Hall
1973), or Blalock, Social Statistics, (McGraw-Hill, second edition 1972).
CHAPTER II: DESCRIPTION OF SURVEY.RESULTS
CHARACTERISTICS OF
BREAKFAST AND
NON-BREAKFAST SCHOOLS
Grade Level
Economic Need
Public vs. Private
Schools
School Enrollment
The characteristic that most strongly distinguishes breakfast from non­breakfast
schools is grade level. The SBP is most commonly found in elemen­tary
and junior high schools, while senior high and consolidated schools are
less likely to have the program.!/ This information is summarized in
Table 2. All other results of this study must be interpreted in light of
the differences between elementary and secondary schools.
It should be mentioned here that, in schools that do offer the School Break­fast
Program, the rate of student participation is higher in lower grade­level
schools. Apparently younger students are likely to participate in the
breakfast program. This may be a reason why the administrators in secondary
schools do not initiate the program. This also might explain why relatively
few consolidated schools offer the program: they may see a low overall rate
of participation due to the older students in the school. It may also be
that administrators see the need for breakfast more clearly, with younger
children, while older children are believed to be able to take care of them­selves.
Schools having a higher percentage of needy students, those certified eligi­ble
for free or reduced-price meals, are much more likely to have a school
breakfast program than those that do not. The average percentage of needy
students in breakfast schools is 46 percent, while in non-breakfast schools
it is 16 percent, a difference significant at the 99 percent level;
ac.OOOOl. This difference remains significant when the sample is broken
down by grade level in Table 3, even though the percentage of needy students
in both groups of schools is consistently lower at higher grades. Schools
with a higher percentage of needy students apparently are more highly moti­vated
to offer the SBP. Of course, this may be because the program was
initiated exclusively in needy schools, and it is only eight years since the
program was made available to all schools compared to 34 years that the NSLP
has been available to all schools. School participation may also be
encouraged by the rate of student participation in the program, which is
higher in the needier schools. (Factors influencing student participation
in the breakfast program will be discussed later in this chapter).
It is possibly this factor of economic need that distinguishes public from
private schools in terms of their offering the SBP. About nine percent of
the unweighted study sample was made up of private schools. Of these, only
20 percent offered the program, compared with 51 percent of a sample as a
whole.
The average enrollment of schools offering School Breakfast Programs is larg­er
than that of non-breakfast schools. As summarized in Table 4, the
average enrollment of breakfast schools is 585; that of non-breakfast schools
is 494. This difference is significant at the .99 percent level; a= .01.
This finding is unexpected, since elementary schools, those most likely to
have the SBP,tend to be smaller than secondary schools. Nonetheless, at
every level, breakfast schools are larger than non-breakfast schools.
This difference is even more striking in view of the fact that less needy
schools tend to be larger, while those with a higher percentage of needy
students tend to have lower enrollment. Table 5 shows that schools with
40 percent or more needy students have an average enrollment of 466 students;
those with less than 40 percent needy students average 512. Interestin~ly,
ll Elementary schools are defined as those with any combination of kinder­garten
through grade six. Junior highs are those having seventh and
eighth or seventh through ninth grades. Senior highs in this study are
those with seventh through t\•Telfth or n:l.nth or tenth through twelfth
grades. Consolidated schools are those offering elementary through
senior high grades, and special schools are ungraded.
5
TABLE 2
PERCENT OF BREAKFASr AND NON-BR~AKFAST SCHOOLS
BY GRADE LEVEL ~
Grade Breakfast Non-Breakfast
Level Percent Number Percent Number
Elementary 55.7 149 41.6 102
Jr. High School 24.8 . 66 30.0 74
Jr.-Sr. High School 9.3 25 17.6 43
Consolidated 6.5 17 7.9 20
Special 3.7 10 3.0 7
TOTAL 100 267 100 246
~/ The data in this table are significant at 99 percent level; a = .009.
All
(weighted)
Percent Number
43.4 223
29.3 150
16.5 85
7.7 39
3.1 16
100 513
Source: Compiled from data collected during the School Breakfast Program Study, 1977-78.
TABLE 3
• PERCENT OF .NEEDY STUDENTS IN BREAKFAST AND
NON-BREAKFAST SCHOOLS ~/
Grade Level Breakfast Non-Breakfast All
(wei hted)
Elementary 47.0 18.4 27.7
Jr. High School 49.3 15.1 22.6
Jr.-Sr. High School 26.9 11.9 15.1
Consolidated 41.6 20.4 26.9
Special 54.7 22.3 30.5
All 46.0 16.0
Number of
Sample Schools 267 246 513
~/ The data in this table are significant at the 99.999 percent level: a = .00001.
Source: Compiled from data collected during the School Breakfast Study, 1977-78.
6
Grade Level
Elementary
Jr. High School
Jr.-Sr. High School
Conslidated
Special
All GrRrle
Levels
TABLE 4
MEAN ENROLLMENT OF BREAKFAST
AND NON-BREAKFAST SCHOOLS ~/
Breakfast
Schools
505
584
1,240
470
337
585
Mean Enrollment
Non-Breakfast
Schools
386
443
861
446
477
494
Number of
Schools
(weighted)
406
459
890
449
455
504
~/ The data in this table are significant at the 99 percent level; ~ = .01
Source: Compiled from data collected during School Breakfast Study, 1977-78.
Economic Need
Schools With More
Than 40 Percent of
Enrollment Needy
Schools With Less
Than 40 Percent of
Enrollment Needy
All Schools
Number of Schools
in Unweighted Sample
SBP
Schools
493
685
585
TABLE 5
MEAN ENROLLMENT BY ECONOMIC NEED IN SBP
AND NON-SBP SCHOOLS ~/
Number of Schoo~s
(weighted)
35
33
68
247
Non-SBP
Schools
444
498
493
Mean Enrollment
Number of Schools
(weighted)
45
402
447
268
~/ The data in this table are significant at the 99 percent level; ~ = .01.
Source: Compiled from data collected during School Breakfast Study, 1977-78.
All Nu~ber of Schools
Schools (weighted)
466 80
512 435
504 515
515
7
Attendance
Facilities
8
the size difference is much greater among breakfast schools. The size
difference may be more significant in the less needy schools because
among the needier schools, such a high percentage participate due to need
alone that the influence , of size as a motivation for offering the program
is masked.
There is a slight difference in the rate of attendance between breakfast and
non-breakfast schools. Schoqls with the breakfast program have a slightly
lower attendance rate, measured as a percent of total enrollment. This
difference is not explained by relative need. Attendance rates are slightly
lower among needy schools and the difference is significant at the 99 percent
level; ~ • .001 by t-test, but among needy and non-needy schools considered
separately (those with more or less than 40 percent needy students), attend­ance
at breakfast schools is still slightly lower. Attendance shows a small
negative correlation with school size, so that possibly the larger size of
breakfast schools accounts for some of the difference in attendance . (Table 6).
It is grade level, however, that apparently explains most of this difference
in average attendance rates between breakfast and non-breakfast schools . . In
an analysis that tested the effect of grade level and the presence of the
breakfast program on attendance rates, grade level was found to be highly
significant. The presence of a breakfast program, when the effect of
grade level is simultaneously considered, did not reach statistical
significance in influencing attendance rates.
The introduction of the School Breakfast Program appears to have no impact at
all on the attendance rate of a school. Attendance rates were compared at
schools one year prior to the introduction of the Breakfast Program, and in
the first and third years of the program. There was no discernible
difference in attendance, and a t-test comparing attendance at these
three points showed no significant differences.
It was thought that orle of the factors influencing the decision of a school
to participate in the School Breakfast Program might be the availability
of certain facilities at the school.
In this regard, it is interesting to note that the study found no significant
difference between breakfast and non-breakfast schools in the availability of
a kitchen. Of non-breakfast schools, 87 percent had a kitchen, compared with
90 percent of the breakfast schools, a statistically insignificant difference.
Similarly, children in 87 percent of non-breakfast schools were most likely
to take their meals in either a cafeteria or a "multi-purpose room," compared
with 91.5 percent of breakfast schools. Even with the two sample groups brok­en
down by grade level, no differences in these two variables were found.
There was a significant difference between breakfast and non-breakfast schools
in the availability of a cafeteria. Eighty percent of breakfast schools, and
only 64 percent of non-breakfast schools had one (a • .0001). This should not
affect a school's ability to serve breakfast, however. Breakfast is much
easier than lunch to serve, and may easily be offered in a classroom or other
non-specialized room.
While kitchens were equally available to both groups of schools, breakfast
schools were significantly more likely to have on-site preparation of lunch
than were non-breakfast schools. Among breakfast schools, 82 percent had
on-site lunch preparation, compared with only 63 percent of non-breakfast
schools, a • .00001. This may indicate that the types of kitchens that are
available vary between the two groups.
Grade Level
Elementary
Jr. High School
Jr.-Sr. High School
Consolidated
Special
Number of Sample
Schools:
weighted
unweightPrl
TABLE 6
AVERAGE DAILY ATTENDANCE IN BREAKFAST
AND NON-BREAKFAST SCHOOLS ~/
Percent of Total Enrollment
Breakfast Non-Breakfast
93.9 94.6
92.6 95.0
93.4 93.0.
92.7 94.3
92.7 94.3
68 445
2.47 266
All
(weighted)
94.8
94.9
93.0
94.1
93.3
513
513
~/ The data in this table are significant at 99 percent level; a = .001
Source : Compiled from data collected during School Breakfast Study, 1977-78.
TABLE 7
PERCENT OF STUDENTS USING DIFFERENT METHODS
OF TRANSPORTATION TO BREAKFAST AND NON-BREAKFAST SCHOOLS
Percent of Students
Number of Non-
Transportation Schools Breakfast Breakfast All
Bus, less than
30 minutes 497 .~/ 34.2 29.8 30.4
Bus, 30 minutes
to one hour 513 15.0 16.6 16.4
Bus, more than
1 hour 515 2.5 1.4 1.6
Walk or Ride
Bike 510 31.7 29.9 30.0
Driven 510 ll.8 16.3 15.7
Other 513 0.3 0.6 0.6
t-test
0. .09
N. s .E/
a .09
N.S.
a = .005
N.S.
~/ The number of schools may be less than 515 due to missing data points from non-response by
some schools.
E._/ N. S . . = Not statistically significant.
Source: Compiled from data collected during School Breakfast Study, 1977-7·8.
9
Logistics
Competitive Foodservice
10
It should be noted, though, that having a kitchen per se is not a regulatory
or practical prerequisite to offering the school breakfast. Ten percent of
breakfast schools had no kitchen available. This is important since schools
without the breakfast progr~m sometimes use the lack of kitchen facilities
as a reason for not participating in the SBP.
One objection raised to offering the School Breakfast Program is that it is
logistically difficult. Problems cited include a high percentage of children
who are bused to school, an early starting time for classes,and inadequate
supervisory time available from teachers or administrators.
However, as summarized in Table 7, no differences were found between breakfast
and non-breakfast schools in terms of the percentage of children riding the
bus, the amount of time between the arrival of teachers or administrators and
the start ·of classes, or the hour at which classes start (or at which the
first session starts for split session schools). The only difference in
student transportation was that more students in non-breakfast schools were
driven to school. With regard to busing, it io if anything>the schools that
do have the SBP which have a higher percentage of students riding the bus.
This shows that having students who are bused to school does not make it
difficult to offer the program.
It will be seen later in Table 13 that, even though there is no difference
between breakfast and non-breakfast schools in percentage of students riding
the bus, this percentage does influence the rate of participation of students
in schools that do have breakfast programs.
An important consideration in planning school feeding programs is the role
which is, or ought to be, played by competitive foodservice. Two aspects
of this issue were examined in this survey: the availability of competitive
foodservice outside and inside the school grounds. There was no direct
measure in this survey to ass~ss the availability of competitive foods out­side
the school, so the 'presence of an open campus policy was assumed to
allow students the liberty to find alternate, or competing, sources of food
at meal times. It was hypothesized that schools that have this policy might
be less likely to offer the SBP. The study found that breakfast schools were
more likely to operate on a closed campus system, where students are not
permitted to leave the schools grounds, than were non-bteakfast schools.
Forty-three percent of all schools have an open campus, but only 32 percent
of breakfast school~ compared with 44 percent of non-breakfast, had open
campuses, a statistically significant difference at the 99 percent level;
~ = 003.
Competing foodservice within schools was examined through questions con­cerning
vending machines and snack bars. Direct availability of non­federally
funded food might also decrease the likelihood of a school offering
the SBP. Non-breakfast schools were significantly more likely to have vending
machines in the cafeteria, although the numbers were small in both cases;
7.5 percent of non-breakfast and 2 percent of breakfast schools (6.8 percent
of the sample as a whole) had these machines, a= .007. This represents a
total of only 35 schools.
Schools that had vending machines outside the cafeteria were somewhat more
likely to serve school breakfasts. Of schools that had the machines (9.2
percent of all schools), 57 percent were in the SBP, compared with 47 percent
of those that had no machines outside the cafeteria. This difference was
significant at the 96 percent level; a = .04, based on a chi-square test.
When categorized by grade level, the difference was significant only for
junior high schools, and marginally significant, at the 92 percent level,
a = .08 for consolidated schools. It should be noted, though, that only 47
schools in the sample had these machines.
School Breakfast
and Other School
Feeding Programs
Schools with a snack bar are also more likely to participate in the School
Breakfast Program. Breakfast is offered at 57.4 percent of schools that have
a snack bar (only 8.2 percent of all schools in the sample), and at 47.5
percent of schools without this facility. This is true even though snack
bars are more commonly available to higher grade level schools. The differ­ence
is significant at the 99 percent level; a= .0001. No clear explanation
of this occurrence is apparent although it may be that schools use snack bars
as the serving facility to offer the brea~fast program. Schools with a high
percentage of needy students are less likely to have open campuses than
relatively less needy schools. Of needy schools, 29 percent were open-campus,
compared with 43 percent of the less needy schools (needy schools were defined
as those with more than 40 percent of students enrolled eligible for free and
reduced-price lunches). As we have mentioned, the age range served may
influence this difference, since there are more needy schools in the lower
grades.
Competitive foods of all kinds are more likely to be available at schools
serving older students as summarized in Table 8. The difference between
breakfast and non-breakfast schools in frequency of open campus, however, is
consistent in elementary, junior, and senior high schools, although it is
statistically significant only in elementary schools as indicated in Table 9.
When categorized by school type, breakfast schools show a consistently lower
availability of vending machines in the cafeteria, but statistical signifi­cance
is not reached for most groups because of the small numbers involved.
The significant difference in availability of vending machines outside the
cafeteria also disappears when the groups are broken down by grade level.
When the sample is categorized by grade level, the availability of a snack
bar in school is still positively associated with breakfast being served in
elementary schools, but no significant difference is found for other grades.
Of the nine elementary schools with a snack bar, seven serve some kind of
breakfast, six serve the school breakfast. This represents 75 percent of
snack bar schools, compared with 54 percent of the eler.:cntar:r schools ui::hout
a snac!· bar · that serve the SB:?. This difference is significant at the 92
percent level, a = .08. However, as 1-!ith vending machines, the r,_umber of
schools with the snack bars is quite low, so these differe~ces are probably
not very meaningful.
Table 10 shows the weighted number of schools that have different kinds of
meal and milk service available. The great majority of schools fall in the
class having federal lunch and milk programs, and no breakfast program. Only
slightly more than ten percent of schools do not offer some kind of lun·ch
similarly, only slightly over ten percent fail to offer milk. Less than five
percent of schools in the sample offer no foodservicc at all (and some of
these may have vending machines available). Schools which serve a school
breakfast are significantly more likely to serve a school lunch than are
non-breakfast schools.
Among the 108 "very needy" schools, defined as schools in which more than
50 percent of students are eligible for free or reduced-price meals, 90
percent had the SBP, while only 37 percent of the non-needy schools had it.
Since almost all these needy schools (better than 99 percent) have the NSLP,
no relationship with school breakfast could be identified. Among the non­needy
schools, however, those that have breakfast are significantly more
likely to serve lunch as well (a~ .00001).
When school grade level is considered, the positive relationship between the
breakfast and lunch programs does not reach statistical significance for
individual grade levels. For the sample as a whole, the relationship is
significant at a = .02, as shown in Table 11. Of non-breakfast schools, only
11
12
Grade Level
Elementary
Jr. High School
Jr.-Sr. High School
Consolidated
Special
All Grade Levels
x2 Significance
Level
Number of Schools
in Unweighted Sample
Open
Campus
45.7
30.5
54.1
62.7
14.0
43.0
.0001
515
TABLE 8
PERCENT OF SCHOOLS AT EACH GRADE LEVEL WHICH HAVE
COMPETIN~ FODDSERVICE
Percent of Schools
Vending Machines
In In Snackbar
Cafeteria School Total Available
3.1 3.7 4.5 4.0
4.6 5.7 7.9 8.3
18.1 24.3 36.1 15.7
9.9 24.6 30.3 14.1
10.5 0.0 10.5 10.5
6.8 9.2 12.9 8.2
.0001 .00001 .00001 .01
513 515
Source: Compiled from data collected during School Breakfast Study, 1977-78.
Grade Level
Elementary
Jr. High School
Jr.-Sr. High School
Consolidated
Special
Number of Schools
(weighted)
l
TABlcE 9
PERCENT OF SCHOOLS HAVING OPEN CAMPUS BY GRADE LEVEL
Number of
Schools in
Unweighted
Sample
223
149
85
39
16
Percent of Schools
Non-Breakfast Breakfast
48.6 31.4
31.6 21.3
55.3 39.1
61.9 68.8
12.5 22.2
443 67
~/ N.S. = Not statistically significant.
All
45.7
30.5
54.1
62.7
14.0
510
Source: Compiled from aata collected during School Breakfast Study, 1977-78.
Number of
Schools
(Weighted)
223
150
85
39
16
513
Significance
Level of
Difference
of Means
a = .008
a =.24 N.S.~1
a = .30 N.S.
a = .93 N.S.
a .55 N.S.
TABLE 10
BREAKDOWN OF FOOD PROGRAMS OFFERED BY SCHOOLS
(Weighted)
USDA Other No
Breakfast Lunch Milk Milk Milk Subtotal
NSLP 47 9 4 60
Non-Federal
Plate 1 0 0 1
School A la Carte 0 0 0 0
Breakfast
NSLP and
A la Carte 5 1 0 6
NSLP and
Plate 0 0 0 0
No Lunch 0 0 0 0 67
NSLP 7 2 0 9
Non-Federal
Plate 0 3 2 5
Non-Federal A la Carte 0 0 0 0
Breakfast
NSLP and
A la Carte 5 0 0 5
NSLP and
Plate 0 0 0 0
No Lunch 0 0 0 0 19
NSLP 1 0 0 1
Non-Federal
Plate 0 0 0 0
SBP and Non- A la Carte 0 0 0 0
Federal
Breakfast NSLP and
A la Carte 4 0 0 4
NSLP and
Plate 2 0 0 2
No Lunch 0 0 0 0 7
Continued On Next Page
13
14
TABLE 10--Continued
BREAKDOWN OF FOOD PROGRAMS OFFERED BY SCHOOLS
(Weighted)
USDA Other No
Breakfast Lunch Milk Milk Milk Subtotal
NSLP 255 28 22 305
Non-Federal
Plate 0 3 0 3
A la Carte 2 5 2 9
No Breakfast
NSLP and
A la Carte 43 7 2 52
NSLP and
Plate 0 0 0 0
No Lunch 27 7 22 56
TOTAL 399 65 54 518 ~/
~/ Total is high due to rounding of weighted observation.
Source: Compiled from data collected during School Breakfast Study, 1977-78.
TABLE 11
PERCENT OF SCHOOLS OFFERING THE NATIONAL SCHOOL LUNCH PROGRAM
BY GRADE LEVEL
Percent of Schools
Significance
Level of
Difference by
Grade Level Breakfast Non-Breakfast All Chi-Square Test
Elementary 93.3 90.1 91.6 N.S.~/
Jr. High School 100.0 81.2 83.4 N.S.
Jr.-Sr. High
School 95.7 74.5 76.1 .07
Consolidated 100.0 100.0 100.0
Special 88.9 87.5 87.7 N.S.
All Grade Levels 98.4 85.4 87.2 .02
~/ N.S. = Not statistically significant
Source: Compiled from data collected during School Breakfast Study, 1977-78.
425
518
Number of
Schools in
Weighted
Sample
204
150
85
39
16
515
Combined Effects of
School Participation
in the SBP
85 percent serve the Federal School Lunch, compared with 98.4 percent of
breakfast schools, and more than 12 percent serve no lunch at all, compared
with 0.4 percent, or only one of the breakfast schools. These differences
are highly significant.
It is also interesting to note that the price of the school lunch is signi­ficantly
higher in non-breakfast schools; 50 cents, compared with 46.9 cents
in breakfast schools. The price is also higher in schools serving older
children, but the difference between breakfast and non-breakfast schools is
consistent for elementary, junior, and senior highs as shown in Table 12.
This may be due to the fact that in needy schools, those more likely to
have the breakfast program, participation is more strongly affected by price
than in schools with less needy students, so that the administration tries to
keep prices low. Another possible explanation is that having two meal
programs makes possible economies of scale which lower meal costs. There was
no significant difference in the price charged for reduced price lunches. No
difference was found between breakfast and non-breakfast schools in whether
or not they offered milk under the Special Milk Program. About 78 percent
of the schools in both groups offer the SMP, and 12 to 13 percent offer milk,
which is not federally funded.
Up to this point the analysis of this report has only examined the relation­ships
that may exist between sets of two categories, or variables. However,
a relationship between two variables may be masked or altered by the effect
of a third variable not examined in such a two-way analysis. In order to
examine the relative independent effects of the various factors influencing
a school's decision whether or not to offer the School Breakfast Program, a
regression model was estimated. Regression analysis is a statistical
technique for controlling the influence of a set of related factors which
affect an outcome, or dependent variable, so that the effect of each factor
may be measured independently of the others. By looking at many factors at
once, regression analysis can clarify relationships suggested by two-way
analysis.
The regression model estimated here uses ordinary lea~t squares (OLS) analy­sis
and has, as a dependent variable,a dummy variable!/ set at 0 for non­breakfast
schools and at 1 for schools in the SBP. Other variables in the
model were:
ENROLL
NEED
LEVEL
KITCHEN
c~
Total school enrollment
Percent of students eligible for free
and reduced price meals
A dummy variable set at 0 for elementary
schools and at 1 for junior and senior
high school~/
A dummy variable set at 1 if schools had
a kitchen, 0 if they had no kitchen
A dummy variable set at 1 if schools had
either a cafeteria or a multi-purpose room
used for eating, and 0 otherwise
ll A dummy variable is used when the categories measured cannot be
expressed as continuous numerical data. In this instance values of zero
and one are assigned in lieu of this to measure the presence or absence of
a particular element in the model.
lf Special and consolidated schools were excluded from the analysis.
15
Grade Level
Elementary
Jr. High School
Jr.-Sr. High School
Consolidated
Special
All Grade Level
TABLE 12
AVERAGE PRICE OF LUNCH IN BREAKFAST
AND NON-BREAKFAST SCHOOLS BY GRADE LEVEL
Breakfast Non-Breakfast
Schools Schools
45.2 42.8
49.1 51.1
51.6 56.5
46.4 42.3
46.4 . 47.5
46.9 50.0
!!_I Significant at the 99 percent level; a • .0001.
Number of
Schools in
All Weighted
Schools Sample
47.7 201
50.8 106
56.0 68
42.9 26
47.3 12
49.5 !!_/ 412
Source: Compiled from data collected during School Breakfast Study, 1977-78.
TABLE 13
PARAMETERS OF THE REGRESSION MODEL TESTING FACTORS INFLUENCING SCHOOL PARTICIPATION
IN THE SCHOOL BREAKFAST PROGRAM
Dependent Variabl.e
Schools that do or do not offer the
School Breakfast Program
Parameters
ENROLL
NEED
LEVEL E./
KITCHEN
CAF ~/
SUPER i/
BUS
.869
.643
.382
. 202
.371
.465
.262
Number of Cases
431
F - test
5.071 !!_1
96.056 'pj
1.519
.202
.007
.196
• 175
a = .01
a = .OOld/
a = N.S.­a
= N.S •
a • N.S.
a N.S .
a = N.S.
--------- -----
16
a/ Significant at 99 percent level; a • .01.
b/ Significant at 99 percent level; a - .001. c/ Specialand consolidated schools were excluded from the analysis.
d/ N.S. - Not significant
~/ CAF = A dummy variable set at one if schools had either a cafeteria or multi-purpose room used
for eating, and 0 otherwise.
i/ SUPER = Supervisory time.
Source: Compiled from data collected during School Breakfast Study, 1977-78.
STUDENT PARTICIPATION
IN THE BREAKFAST AND
LUNCH PROGRAMS
Grade Level
SUPER
BUS
Supervisory time, i.~.', time (in minutes)
between the arrival of administrative or
teaching staff and the start of classes
Percentage of students who ride the bus
more than 30 minutes
The results of regression model are shown in Table 13. By far the most
important variable affecting school participation in the SBP is the percentage
of needy, students: schools with a higher proportion of needy children are
more likely to offer breakfast. As we have already seen, larger enrollment
is also significantly and positively associated with offering the breakfast
program. In this regression, grade level (between elementary and secondary)
loses its significance as a discriminator between breakfast and non-breakfast
schools, although we have already seen that lower grade scho~ls are still
shown as more likely to offer the program. In the regression, junior and
senior high schools were grouped together to make elementary and secondary
groups more equal. Since more junior than senior high schools offer the SBP,
this grouping of schools reduced the significance of grade level in the equa­tion.
The other variables tested in the model (availability of kitchen,
cafeteria, supervisory time, and percent of students bused more than 30
minutes) show no relationship at all to breakfast program participation.
The two way analysis discussed earlier in this report presented similar
conclusions, but it is worth noting since the lack of facilities and the
shortage of supervisors' time are commonly given as reasons for not offering
the breakfast program. Busing of students apparently does not influence the
decision of schools to offer the SBP, ·although we shall see that it does
influence participation by students if breakfast is offered. The results are
virtually unchanged if a step-wise regression is performed which considers
the effect of level first and school size and level of need last. This shows
that the high level of significance of school size and level of need is not
an artifact of co-variation with some other variable, and that these variables
do have independent explanatory power.
Student participation in school feeding programs is higher at lower grade
levels. Participation by students is also consistently higher for lunch than
for breakfast,as Tables 14 and 15 indicate. As has been suggested by the
Children's Foundation report on barriers to the School Breakfast Program!/
it may be that school administrators have too high expectations for
participation in the breakfast program, and so count the program a failure if
it does not match the NSLP in participation. Relevant to administrators'
concern for maintaining high participation rates, it is noteworthy that
participation in the School Lunch Program is consistently higher in schools
that also have the School Breakfast Program. This difference is highly
significant at all grade levels. The explanation for this is that lunch
participation is positively correlated with a number of variables that are
also associated with offering the School Breakfast Program, especially
economic need. However, participation is negatively correlated with size,
another factor positively associated with the presence of the School Breakfast
Program. In a regression equation that controlled for the effects of size,
economic need and grade level ; the presence' of the School Breakfast Program
1/ The Children's Foundation, Barriers to School Breakfast, Washington,
D.C.,-November 1978, p. 30.
17
18
•
TABLE 14
AVERAGE DAILY STUDENT PARTICIPATION IN THE SCHOOL BREAKFAST
AND NATIONAL SCHOOL LUNCH PROGRAMS
Percent of Average Daily Attendance
Number of Number of
Grade Level Breakfast Schools Lunch Schools
Elementary 35.9 136 62.3
Jr. High School 33.3 60 65.5
Jr.-Sr. High School 15.0 23 40.2
Consolidated 30.9 15 66.2
Special 59.5 9 75.9
All Grade Levels 33.5 243 60.5
!/ The remaining 88 schools offered neither the f.BP nor the SLP.
Source: Complied from data collected during School Breakfast Study, 1977-78 .
• TABLE 15
199
114
65
36
14
4271/
AVERAGE DAILY NSLP PARTICIPATION IN BREAKFAST AND NON-BREAKFAST SCHOOLS
Percent of Average Daily Attendance
Number of Number of
Breakfast Schools Non-Breakfast Schools All
Grade Level Schools (weighted) Schools (weighted) Schools
Elementary 77.5 36 58.9 163 62.3
Jr. High School 69.0 15 64.9 98 65.5
Jr.-Sr. High
School 47.1 6 39.4 58 40.2
Consolidated 71.9 4 65.4 32 66.2
Special 87.0 2 73.8 12 75.9
All Grade Levels 72.4 64 58.5 363
Source: Compiled from data collected during School Breakfast Study, 1977~78.
Number of
Schools
(weighted)
199
114
65
36
14
427
Economic Need
Meal Price
School Enrollment
Facilities
was still found to be significantly positively associated with student
participation in lunch. This regression analysis is discussed in greater
detail later in this chapter. These findings indicate that the joint
availability of the School Breakfast Program and the National School Lunch
Program can foster high participation rates by students in the lunch program.
At the very least, the availability of breakfast in schools will not reduce
participation in the lunch program.
Student participation in both breakfast and lunch is significantly higher in
schools with a higher percentage of needy students. Both lunch and breakfast
participation are highly correlated with the percent of needy students
enrolled in a school. These results are statistically significant at the
99 percent level; a= .001. Although the level of need of schools is higher
in lower grade schools, the relationship between breakfast participation and
need is consistent at all grade levels. Table 16 shows that among the needy
schools, participation in school breakfast does not drop off sharply at the
secondary level as it does in the less needy schools. Table 17 shows that
even when level of need is accounted for, lunch participation is higher in
schools where breakfast is offered, as well as being higher in needier
schools. An important finding of this study is the fact that, for both break­fast
and lunch, the rate of participation in school feeding among the needy
students in any school is much higher than among those who pay full price for
the meals . In breakfast schools, 46 percent of students are eligible for
free meals, but 57 percent of the lunches served were free. This means that
a higher proportion of non-needy students chooses not to eat the school lunc~
so a disproportionate number of needy children is served by the program.
Note that these figures do not indicate that free lunches are being served to
ineligible children. Rather, they show that the population of students who
eat school lunches is lower-income than the population of the school as a
whole. In non-breakfast schools, 16 percent of the enrolled students quali­fied
for free lunches, but 25 percent of the lunches served were free. The
di~ference in the percentage of lunches served to students eligible for free
and reduced-price lunches in non-breakfast schools was identical to
breakfast schools, but was about four percentage points higher. Overall, 30
percent of students are eligible for free lunches, while 46 percent of all
lunches served were free. The same is even ·more strongly the case for
breakfast: in breakfast schools, 46 percent of students are eligible for
free meals, but 79 percent of the breakfasts served during the study period
(October 197"7) were free. These differences are significant at the 99 per­cent
level, a = .00001.
Participation in both the SBP and the NSLP is price responsive, that is,
participation is higher when the price is lower. This is true for partici­pation
in both breakfast and lunch by students who pay full price for their
meals. These price and participation relationships are significant at the
99 percent level, a = .001. Of course, the higher participation rate by
needy students is another indication of price responsiveness, since these
students face a lower or a zero price for their meals.
Schools with a higher percentage of needy students tend to be smaller in
total enrollment. This relationship is significant at the 96 percent level;
a = .04, and is probably the reason that participation in both the SBP and
the NSLP also is higher at lower enrollment levels. One reason for this is
that elementary schools and junior high schools tend to have more needy
students than do high schools, and enrollment also tends to be smaller in
the lower grade schools.
Breakfast participation is slightly higher in schools where the breakfast is
prepared on-site: 35 percent, compared with 29 percent for off-site prepara­tion.
This difference is not significant, however; a = .15. School brf'c>.kfast
is a relatively simple meal that can consist entirely nf cold foods, so that
off-site preparation need not affect the acceptability of the meal.
19
TABLE 16
AVERAGE DAILY SCHOOL BREAKFAST PROGRAM PARJICIPATION
IN NEEDY AND NON~NEEDY SCHOOLS ~
Percent of Average Daily Attendance
Needy Non-Needy All
Grade Level Schools Schools Schools
Elementary 46.4 24.9 35.9
Jr. High School 39.6 22.4 33.3
Jr.-Sr. High School 41.0 9.6 15.0
Consolidated 38.1 26.2 30.9
Special 75.4 27.2 59.5
All Grade Levels 45.1 22.1 33.9
Number of
Schools
(Unweighted)
136
60
23
15
9
243
~/ Needy Schools have 40 percent or more needy students, significant at the 99 percent
level; ~ = .00001.
Source: Compiled from data collected during School Breakfast Study, 1977-78.
Need
Needy
Non-Needy
TABLE 17
AVERAGE DAILY NATIONAL SCHOO~ LUNCH PROGRAM PARTICIPATION /
BY LEVEL OF NEED FOR BREAKFAST AND NON-BREAKFAST SCHOOLS ~
Percent of Average Daily Attendance
Breakfast
Number of
Schools Non-Breakfast
Number of
Schools
81.1
64.1
32
32
66.5
57.2
42
323
~/ Needy Schools have 40 percent or more needy students, significant by test F
Source: Compiled from data collected during School Breakfast Study, 1977-78.
All
72.9
57.8
73; a
Number of
Schools
74
356
.00001.
TABLE 18.
AVERAGE DAILY PARTICIPATION IN THE SCHOOL BREAKFAST PROGRAM
BY LOCATION OF MEAL CONSUMPTION
Percent of Average Daily Attendance
Number of
Location Participation Sample Schools
Cafeteria 34.5 187
Multi-Purpose Room 31.6 37
Gymnasium 34.1 7
Classroom 39.3 6
School Grounds 26.8 5
Other 17.6 2
All Locations 33.9 244
Source: Compiled from data collected during School Breakfast
Study, 1977-78.
21
Competitive Foodservice
22
Participation in lunch is significantly higher (a = .01) if the meals are
prepared on-site: 67 percent compared with 59 percent for lunches prepared
away from the school. Other studies have found that hot lunches have a
higher acceptability than cold ones, and these are likely to be more
successful if prepared and served at one location.
This difference in lunch· participation is particularly noteworthy in light of
the fact that needier schools are more likely to have off-site preparation
(44 percent of schools with more than 40 percent needy students have off-site
lunch preparation, compared with 33 percent of less needy schools), and it has
already been discussed that participation is higher in the needy schools.
Among breakfast schools, the site of preparation of breakfast shows no relation
to whether or not the school is needy.
Breakfast participation was slightly higher for students who took their meals
in the classroom, multi-purpose room, or cafeteria; and slightly lower if meals
were taken in the gym or elsewhere on the school grounds. The difference was
neither great nor statistically significant, however. Table 18 shows that the
large majority, 77 percent, of breakfast schools serve meals in the cafeteria;
91 percent have meals in either a cafeteria or a multi-purpose room.
Lunch participation as shown in Table 19 in non-breakfast schools was also not
significantly different according to the location where meals were eaten. But
in breakfast schools, once again, the highest lunch participation rates were
found if students ate either in the cafeteria or multi-purpose room or the
classroom, and here the differences were significant. One possible explana­tion
for these differences is that access to other sources of food or to
other activities is greater if the students go to a gym or are allowed to go
out on school grounds. Supervision may be greater, and alternative activities
fewer, in the other locations. It . is also possible that some areas provide
a more pleasant atmosphere for eating.
Participation in the breakfast program is significantly reduced if the school
has an open campus. 'his is the case for every grade level as summarized in
Table 20. It is npt intuitively obvious that this should be so, since all
students are free to be outside the school grounds before the start of class,
and so are not compelled to be in school at breakfast time. Possibly more
food shops are available, and children may be more familiar with them if the
campus is open, so the incentive to eat breakfast off campus is greater for
these schools. Closed schools also tend to be lower grade schools, and
needier schools are also more likely to be closed; both of these character­istics
are associated with higher breakfast participation. Nonetheless,
breakfast participation is lower in open campus schools even with these
variables controlled.
Participation in breakfast is not significantly affected by the availability
of vending machines either in or outside the cafeteria, nor by the availabili­ty
of a snack bar in the school. This is true at all grade levels as
illustrated by Table 21.
Open campus policy is the only competing food system that significantly reduc­es
participation in breakfast. Lunch participation is reduced by having a
snack bar available, but open campus also has a highly significant effect.
These data are displayed in Table 22. It is therefore quite important to note
that open campus is a far more common policy than the competitive food
systems considered here. One third of breakfast schools and 41 percent of all
schools have an open campus, while only approximately ten percent of schools
have vending machines or snack bars.