Family Economics and Nutrition Review [Volume 11, Number 4]

              17. d4'J: \ \ ..
.t--=- LJ ') ') ::_) CENTER FOR NUTRITION POLICY AND PROMOTION SERIALS DEP.U.RTMENT
Research Articles
3 Do Third Graders Eat Healthful Breakfasts?
Johanna T Dwyer et a!.
19 Comparison of Economic Status of Elderly Households:
Nonmetropolitan Versus Metropolitan Residence
Cara Janette Miller and Catherine P. Montalto
31 How Marketers Reach Young Consumers: Implications for Nutrition
Education and Health Promotion Campaigns
Vivica Kraak and David L. Pelletier
Research Briefs
42 Changes in Consumers' Knowledge of Food Guide Recommendations.
1990-91 Versus 1994-95
Joanne F. Guthrie and Brenda Derby
49 Dietary Guidance on Sodium: Should We Take It With A Grain of Salt?
Etta Saltos and Shanthy Bowman
52 Could There Be Hunger in America?
Bruce W. Klein
Research Summaries
55 Women's Labor Force Participation
57 USDA's Healthy Eating Index and Nutrition Information
60 Eating Breakfast: Effects of the School Breakfast Program
Regular Items
63 Research and Evaluation Activities in USDA
66 Federal Statistics: USDA Food and Nutrition Programs
68 USDA Food Plans: Cost of Food at Home
69 Consumer Prices
70 Index of Authors in 1998 Issues
72 Index of Articles in 1998 Issues
73 Reviewers for 1998
UNITED STATES DEPARTMENT OF AGRICULTURE
Volume 11 , Number 4 ~~~~~~lfll~llii~~i~
1998
3 0510 1800240 0
Dan Glickman, Secretary
U.S. Department of Agriculture
Shirley R. Watkins, Under Secretary
Food. Nutrition, and Consumer Services
Rajen Anand, Executive Director
Center for Nutrition Policy and Promotion
CarolS. Kramer-LeBlanc, Deputy Executive Director
Center for Nutrition Policy and Promotion
P. Peter Basiotis, Director
Nutrition Policy and Analysis Staff
The U.S. Department of Agriculture (USDA) prohibits discrimination in all its programs
and activities on the basis of race, color, national origin, gender, religion, age, disability,
political beliefs, sexual orientation, or marital or family status. {Not all prohibited bases apply to
all programs.) Persons with disabilities who require alternative means for communication of
program information (Braille, large print, audiotape, etc.) should contact USDA's TARGET
Center at (202) 720-2600 (voice and IDD).
To file a complaint of discrimination, write USDA, Din:ctor, Office of Civil Rights, Room
326-W, Whitten Building, 14th and Indq>endence Avenue, SW, W asbington, DC 20250-9410 or
call (202) 720-5964 (voice and IDD). USDA is an equal opportunity provider and employer.
Editor-in-Chief
CarolS. Kramer-LeBlanc
Editor
Julia M. Dinkins
Features Editor
Mark Uno
Managing Editor
Jane W. Fleming
Contributors
Bruce W. Klein
Nancy E. Schwenk
Lisa Bente
Family Economics and Nutrition Review Is
written and published each quarter by the
Center for Nutrition PoUcy and Promotion,
U.S. Department of Agriculture, Washington, DC.
The Secretary of Agriculture has determined
that publication of this periodical Is necessary
In the transaction of the public business
required by law of the Department.
This publication Is not copyrighted. Contents
may be reprinted without permission, but credit
to Family Economics and Nutrition Review
would be appreciated. Use of commercial or
trade names does not Imply approval or consti­tute
endorsement by USDA. Family Economics
and Nutrition Review Is Indexed In the following
databases: AGRICOLA, Agellne, Economic
Literature Index, ERIC, Family Studies,
PAIS, and Sociological Abstracts.
Family Economics and Nutrition Review Is
for sale by the Superintendent of Documents.
Subscription price Is $12.00 per year ($15.00
for foreign addresses). Send subscription
orders and change of address to Superin­tendent
of Documents, P.O. Box 371954,
Pittsburgh, PA 15250-7954. (See subscription
lorm on p. 75.)
Original manuscripts are accepted for publi­cation.
(See "guidelines for authors• on back
Inside cover.) Suggestions or comments con­cerning
this publication should be addressed
to Julia M. Dinkins, Editor, Family Economics
and Nutrition Review, Center for Nutrition
Polley and Promotion, USDA,1120 20th St.
NW, Suite 200 North Lobby, Washington, DC
20036. Phone (202)606-4876.
The Family Economics and Nutrltlon Review
Is now available on (http:llwww.usda.gov/cnpp)
the Internet (see p. 74).
Center for Nutrition Policy and Promotion
PfloP£rrry
,.,EieAfl~r rifF
J• .
tJ,v 2 l
{Jf}jl " 1999.____...-...-.<..--'Y Of
R h Arti I at "" Nort.J... esearc c es I,Jfee,, '' Caro!J·
3 Do Third Graders Eat Healthful Breakfasts?
s. vr0 lfla
• Johanna T. Dwyer et al.
19 Comparison of Economic Status of Elderly Households:
31
Nonmetropolitan Versus Metropolitan Residence
Cara Janette Miller and Catherine P. Montalto
How Marketers Reach Young Consumers: Implications
for Nutrition Education and Health Promotion Campaigns
Vivica Kraak and David L Pelletier
Research Briefs
42 Changes in Consumers' Knowledge of Food Guide
Recommendations, 1990-91 Versus 1994-95
Joanne F. Guthrie and Brenda Derby
49 Dietary Guidance on Sodium: Should We Take It With
A Grain of Salt?
Etta Saltos and Shanthy Bowman
52 Could There Be Hunger in America?
Bruce W. Klein
Research Summaries
55
57
60
Women's Labor Force Participation
USDA's Healthy Eating Index and Nutrition Information
Eating Breakfast: Effects of the School Breakfast Program
Regular Items
63 Research and Evaluation Activities in USDA
66 Federal Statistics: USDA Food and Nutrition Programs
68 USDA Food Plans: Cost of Food at Home
69 Consumer Prices
70 Index of Authors in 1998 Issues
72 Index of Articles in 1998 Issues
73 Reviewers for 1998
Volume 11, Number 4
1998
Would you like to publish in Family Economics and Nutrition Review?
Family Economics and Nutrition Review will consider for publication articles concerning economic and
nutritional issues related to the health and well-being of families. We are especially interested in studies
about U.S. population groups at risk-from either an economic or nutritional perspective. Research
may be based on primary or secondary data as long as it is national or regional in scope or of national
policy interest. Articles may use descriptive or econometric techniques.
Family Economics and Nutrition Review has a new feature: Research Briefs. We defme Research Briefs
as short research articles. Our guidelines are found on the back inside cover of each issue.
We invite submission of Research Briefs; manuscripts may contain findings previously presented at
poster sessions if not published in proceedings (except for abstract).
Manuscripts may be mailed to Julia M. Dinkins, Editor, Family Economics and Nutrition Review,
Center for Nutrition Policy and Promotion. See guidelines on back inside cover for complete address.
Johanna T. Dwyer
Tufts University
Mary Kay Ebzery
Tufts University
Theresa A. Nicklas
Tulane University
Henry A. Feldman
New England Research Institutes
Marguerite A. Evans
National Heart, Lung, and Blood
Institute
Michelle M. Zive
University of California, San Diego
Leslie A. Lytle
University of Minnesota
Deanna H. Montgomery
University of Texas
Ann L. Clesi
Tulane University
Anne Garceau
Tufts University
Milton Z. Nichaman
University of Texas
1998 Vol. 11 No.4
Research Articles
Do Third Graders Eat
Healthful Breakfasts?
Breakfast nutrient consumption patterns of third graders were examined
using data from the Child and Adolescent Trial for Cardiovascular Health
(CATCH). Twenty-four-hour recalls, assisted with a food record, were
collected in 96 public schools in four States. Ninety-four percent of the
children reported eating breakfast on the day of the survey: 80 percent ate
at home, 13 percent at school, 3 percent at both locations, and 4 percent
elsewhere. Breakfast eaters had lower total daily intakes of fat as a percentage
of calories (33 vs. 35 percent) but had higher intakes of calories, vitamins,
and minerals than did nonbreakfast eaters. Breakfast contributed 18 percent
of total daily caloric intakes; 19 to 34 percent of vitamin and minerals; 13 to
14 percent of total fat, saturated fat, and cholesterol; and 17 percent of sodium
intakes. Hispanic and African American children had higher fat and saturated
fat breakfast intakes than did Caucasian children. Interventions are needed
to encourage primary school students to consume healthful breakfasts.
[!] he diets of primary school
children are high in food
energy, total fat, saturated
fat, and sodium (8,18,22,35).
These children are particularly vulnerable
to high intakes of saturated fat ( 18,35 ),
and their intakes of calcium, iron, zinc
(1 ,4), vitamins A, B6, and C (6) are
sometimes low. Compared with earlier
generations, today's primary school
children are increasingly overweight
· (9,60). Poor diets and less-than-optimal
nutritional status may influence later
risks for cardiovascular disease and
other chronic degenerative diseases
(31 ,32,33). Attention to the quality of
children's diets is, therefore, warranted.
Breakfast contributes substantially to
the nutritional quality of children's
diets ( 15,26,28,36,38,40,43,54). Eating
breakfast is related positively to children's
cognitive function and school perform­ance
(6,23,48,49,50,62), especially for
low-income (30) and undernourished
children (48). Children's breakfast
consumption is also related inversely
to two risk factors for cardiovascular
disease that persist into adulthood (31 ):
body weight and total blood cholesterol
levels (51). Between 5 and 31 percent
of school-age children skip breakfast­a
particularly common practice among
African American girls (27,37,38). Both
skipping breakfast and consuming an
inadequate breakfast increase the likeli­hood
of dietary inadequacies that are
not compensated for by other meals or
snacks ( 17,27,36,54 ).
The Child and Adolescent Trial for
Cardiovascular Health (CATCH) was
a multicenter field trial designed to test
the effects of school- and family-based
interventions designed to reduce risk
3
factors for cardiovascular disease ( 47).1
CATCH provided a unique opportunity
to examine the dietary intakes of a large
ethnically and geographically diverse
group of children ( 19,61 ).
This article describes breakfast con­sumption
patterns and nutrient contribu­tion
of breakfast meals, measured prior
to intervention, when the CATCH sample
was in third grade. We compare findings
with national goals and results from
similar studies. The results may be use­ful
in designing and evaluating health
promotion strategies for improving the
diets of children.
Methods
Subjects
The total CATCH sample consisted
of 5,106 elementary school students
from 96 public schools in California,
Louisiana, Minnesota, and Texas.
Twenty-four schools were in each State:
14 treatment and 10 control. Fifty-nine
of these schools (61 percent) had a
School Breakfast Program. Before
implementing the CATCH intervention,
we randomly selected a subsample of
3,486 students from the four States to
provide representative 24-hour dietary
recalls. Of this subsample, those students
who gave their consent and for whom a
blood cholesterol level was available
were interviewed for baseline measure­ment
(fall 1991) when they were third
graders (n=1,920). To evaluate CATCH
intervention effects, we also measured
students' intakes using a 24-hour dietary
recall at follow-up in spring 1994 when
they were fifth graders.2 The final sub­sample
(n= 1 ,920) was representative
1The main results of the trial are reported else­where
(21).
2Results are presented in detail elsewhere (27).
4
of the entire CATCH sample on factors
such as age, race/ethnicity, and other
demographic characteristics. The mean
age was 8.7 years (range of 7.6 to 11.2
years) for the third graders who partici­pated.
Dietary Assessment
The interview was a 24-hour recall,
assisted with a food record, a method
that had been validated for use with
third graders (20 ). CATCH staff asked
students to record-briefly-all food
and beverages consumed "from the
time they woke up until the time they
went to bed." The amounts were omitted.
The next day, CATCH staff asked each
student, during a 24-hour dietary recall
interview, to recall everything consumed
the previous day . The students' food
records were used as a memory prompt.
Using three-dimensional food models,
two-dimensional shapes, and measuring
utensils, children estimated portion
sizes. Then they provided the names
(breakfast, lunch, snack, and supper),
time, and source of each meal (e.g.,
home, school, restaurants).
CATCH staff collected school breakfast
menus and detailed information on recipes,
prepared food products, and preparation
methods to coincide with the 24-hour
dietary recall. Thus, we were able to
describe precisely the nutrient intakes
from school breakfast meals.3 Informa­tion
was not collected on the use of
vitamin and mineral supplements or salt
added at the table, so results reflect only
food intake.
Trained and certified interviewers used
a standard protocol to collect 24-hour
recalls from each child. We used the
3Details of theCA TCH school meal assessment
and quality assurance procedures are published
elsewhere (12 ).
Minnesota Nutrition Data System, version
2.2 (food database 4A and nutrient data
19, 1990) to calculate breakfast and total
daily nutrient intakes. This data system
is designed to allow users to link the
24-hour recall with relevant nutrient data
on school breakfast. We coded foods
and beverages that children consumed
at breakfast as breakfast items, and for
each child, we summed the nutrients for
all foods that had a breakfast code. To
ensure data quality, we excluded recalls
from the analysis if the amount reported
could not be verified with documentation
about the intake's unusual size (collected
at the study site by nutritionists) and
if it also exceeded the 99th percentile
values for portions commonly eaten by
children (45).
Statistical Analysis
Of the l ,920 children in the sample,
46 were excluded for quality assurance
reasons and 2 because meal codes were
not specified. The sample for nutrient
analysis (n=1,872) was representative of
the CATCH group by gender (50 percent
each), race/ethnicity (69 percent Cauca­sian,
12 percent African American, 15
percent Hispanic, and 4 percent Other),
and site. (Data are not shown.) Among
the analysis sample, 1,765 reported eating
breakfast either at home, at school, or
both places. Seventy-seven additional
students reported eating breakfast at
"Other" places (e.g., day care, day camp,
a friend's house, a store, or in transit)
and were not included in the nutrient
analyses. Total daily and breakfast intakes
for sodium (p<0.04) for students eating
breakfast at "Other" places were signifi­cantly
higher than sodium intakes for
students eating breakfast at home or at
school. Also, breakfast intakes of students
eating breakfast at "Other" places were
significantly higher for cholesterol
(p<0.007), protein (p<0.02), and vitamin A
(p<0.05) and lower for carbohydrate
Family Economics and Nutrition Review
(p<0.01), compared with the intakes of
students eating breakfast at home or at
school. (Data are not shown.)
To eliminate small cell sizes based on
ethnicity and source of breakfast, we
excluded 5 students from the analysis.
Thus, for the primary analysis of break­fasts
eaten at home, at school, or in both
locations, 1,683 students were in the
sample.
To analyze nutrient intakes at breakfast
meals and the percentage of contribution
of breakfast to the total daily intake, we
used a mixed linear model. We analyzed
the dependent variables both in absolute
units and relative to the total energy
content of breakfast. Site, gender, race/
ethnicity, and source of meal were in­cluded
as fixed independent effects.
We assessed interaction terms for gender
with race/ethnicity and source of the
meal with gender, site, and race/ethnicity.
A random effect accounted for between­school
variation among sites. Means
were adjusted for all factors in the model.
Means and standard errors were trans­formed
back to the original units for
presentation when log or square root
transformations were used to reduce
skewness. We used version 6.11 of
Statistical Analysis System (SAS)
for all computations (29,52).
Results
Breakfast Patterns of Third Graders
Overall, 94 percent of the students
reported eating breakfast (table 1). No
Asian American students and only 4
percent of Caucasian students reported
skipping breakfast, compared with 11
percent of Hispanic and 8 percent of
African American students (p<0.001).
(Data are not shown.) Two percent of
1998 Vol. 11 No. 4
the third graders in Minnesota skipped
breakfast, compared with 5 percent in
California, 6 percent in Louisiana, and
10 percent in Texas (p<0.001). Less
than one-sixth of the CATCH schools
in Minnesota and California provided a
School Breakfast Program ( 14 and 13
percent, respectively), compared with
all of the CATCH schools in Louisiana
and Texas. (Data are not shown.)
Where Third Graders
Ate Breakfast
Most of the students who ate breakfast,
did so at home: 84 percent. Only 13 per­cent
ate breakfast at school, and 3 percent
ate it both at home and at school. Vari­ations
in breakfast consumption patterns
among sites were striking. Ninety-eight
percent of the students in Minnesota
reported eating breakfast, followed by
95 percent of those in California, 94
percent in Louisiana, and 90 percent
in Texas. More students in Texas and
Louisiana ate breakfast at school (29 and
22 percent, respectively), compared with
students in California and Minnesota (2
and 1 percent, respectively).
Differences were not evident in the
number of children eating breakfast at
home versus at school when only those
schools with a School Breakfast Program
were examined. (Data are not shown.)
Texas and Louisiana ( 28 percent, each)
still had a higher participation rate for
school breakfast, compared with California
and Minnesota (14 and 4 percent, respec­tively).
(Data are not shown.) That is,
simply offering the School Breakfast
Program alone did not explain differ­ences
among sites. It is difficult to know
which factor was associated with this
variation, because site and ethnicity are
confounded.
Most of the students
who ate breakfast, did
so at home: 84 percent.
Only 13 percent ate
breakfast at school,
and 3 percent ate it
both at home and at
school.
5
Table 1. Breakfast eating and sources of breakfast of the CATCH sample at baseline in the third grade
Nutrient analysis sample1
Reported eating Source of breakfast
Characteristic breakfast Total Home School Both
N %2 N N %3 N %3 N %3
Total 1,765 94 1,683 1,409 84 218 13 56 3
N %4 N N %5 N %5 N %5
Gender
Boys 878 94 835 697 83 112 13 26 3
Girls 887 94 848 712 84 106 12 30 4
Race/ethnicity
Caucasian 1,240 96 1,180 1,082 92 76 6 20 2
African American 207 92 198 115 58 59 30 24 12
American Hispanic 253 89 248 153 62 83 33 12 5
Other 65 97 57 57 100 0 0 0 0
Site(% with School Breakfast
Program)
California (13%) 431 95 423 412 97 7 2 4 I
Louisiana (I 00%) 416 94 397 275 69 87 22 35 9
Minnesota (14%) 484 98 440 434 99 3 1 3 1
Texas ( 100%) 434 90 423 288 68 121 29 14 3
1Native Americans and Asian Americans were combined with "Other" race/ethnicity for nutrient analysis. Eighty-two students were excluded: 77 who ate
~reakfast in places other than at home or school and 5 students who ate breakfast at school.
Percentage of total substudy sample ( 1,872).
3Percentage of nutrient analysis sample.
4Percentage of site, gender, or race/ethnicity stratum.
5Percentage of site, gender, or race/ethnicity stratum in nutrient analysis sample.
Contribution of Breakfast to
Third Graders' Total Daily
Nutrient Intake
Students who ate breakfast consumed,
on average, significantly more calories
daily than those who did not eat break­fast:
1,952 versus 1,544 calories (table
2). Breakfast eaters also had higher
intakes of protein, percentage of energy
from carbohydrates, sodium, cholesterol,
and most vitamins and minerals. Means
for both breakfast eaters and non breakfast
eaters met at least two-thirds of the
Recommended Dietary Allowances
(RDA's) for energy, protein, vitamin A,
ascorbic acid, iron, and zinc (34).
6
Non breakfast eaters' mean intakes fell
short of the RDA' s for vitamin A and
calcium. Compared with nonbreakfast
eaters, breakfast eaters consumed a sig­nificantly
higher percentage of calories
from carbohydrates (54 vs. 52 percent).
Total fat intake for both groups exceeded
the recommendation of 30 percent of
calories from fat (58): 33 percent for
breakfast eaters and 35 percent for non­breakfast
eaters. Daily sodium (2,891 mg)
and cholesterol (204 mg) intakes among
breakfast eaters were higher than those
of non breakfast eaters (2,259 and 142
mg, respectively). Although cholesterol
intakes of breakfast eaters and their
counterparts met recommended guide­lines
of no more than 300 mg per day
(33), sodium intakes of breakfast eaters
exceeded the guideline.
Breakfast contributed about 18 percent
of the third graders' mean energy intakes,
17 percent of total protein, 22 percent of
carbohydrate, and 13 percent of total fat
consumed (table 3). Fourteen percent of
total daily amounts of both saturated fat
and cholesterol, 17 percent of sodium,
and 19 to 34 percent of daily vitamin
and mineral intakes came from break­fast.
(Data are not shown.) Compared
with girls' breakfasts, those for boys
Family Economics and Nutrition Review
Table 2. Total daily nutrient intakes of children eating breakfast,
compared with those not eating breakfast, 1 CATCH
Non breakfast Breakfast eaters
Breakfast eaters eaters vs.
Variable Goal N=l,765 N= 107 non breakfast
eaters
p2
Energy (calories) >1,34cf 1,952 (17) 1,544 (48) <0.001
Protein (% calories) NA4 14.7 (0.1) 15.1 (0.1) 0.30
Carbohydrate(% calories) NA4 54.0 (0.2) 51.5 (0.8) 0.003
Total fat (% calories) <303 32.5 (0.2) 34.6 (0.7) 0.002
Saturated fat (% calories) <103 12.7 (0.11) 13.0 (0.3) 0.28
Sodium (mg) <2,400S 2,891 (33) 2,259 (97) <0.001
Cholesterol (mg) <3005 204.4 (3.5) 141.6 (I 0.1) <0.001
Protein (g) >193 71.9 (0.7) 58.0 (2.2) <0.001
Vitamin A (RE) >4673 908 (15) 455 (39) <0.001
Ascorbic acid (mg) >303 89.8 (2.1) 52.0 (5.4) <0.001
Iron (mg) >73 13.4 (0.2) 9.1 (0.5) <0.001
Calcium (mg) >8716 1,043 (15) 745 (37) <0.001
Zinc (mg) >73 9.59 (0.11) 7.18 (0.28) <0.001
1 Adjusted mean (standard error): model-adjusted by site, race/ethnicity, and gender.
2-resting hypothesis of equal mean between breakfast eaters and nonbreakfast eaters.
3Goal based on the Dietary Guidelines for Americans (58), National School Lunch Program and School
Breakfast Program: School Meals Initiative for Healthy Children (57).
4Does not apply.
5V alues are two-thirds of the 1989 Recommended Dietary Allowances ( 34) for 7- to I 0-year-old children.
6Value is two-thirds of the 1998 Dietary Reference Intake (55).
supplied 1 to 4 percent more of their
daily intakes on 11 of the 12 nutrients
analyzed (p<0.05). Boys' and girls'
intake of ascorbic acid was not signifi­cantly
different.
The contribution of breakfast to total
daily intakes of fat, saturated fat, and
cholesterol differed by site (all p<O.OOl).
(Data are not shown.) Breakfast at all
sites provided 20 percent or more of
daily intakes of ascorbic acid and iron.
Differences in other nutrients were also
evident (p<0.05).
1998 Vol. 11 No. 4
Nutrient Content of Third Graders'
Breakfast Meals
Table 4 presents mean breakfast intakes
of food energy and selected nutrients
among third graders overall, by gender,
and by race/ethnicity. All interaction
terms in table 4, as well as tables 5
and 6, were statistically nonsignificant
(p>O.l 0); thus, results are tabulated for
main effects only (e.g., site, gender,
race/ethnicity, and source of meal). The
tables also provide one-quarter of the
RDA goals (34), the Dietary Guidelines'
goals (58) recommended by the U.S.
Department of Agriculture's (USDA)
-· Students who ate
breakfast consumed ...
more calories daily
than those who did
not eat breakfast: ...
7
Table 3. Percent contributions of breakfast to daily nutrient intakes,1 CATCH
Overall Gender Site
Variable Goal2 N = 1,683 Boys Girls p3 California Louisiana Minnesota Texas p3
Energy (calories) 500 18.4 (0.3) 19.0 (0.4) 17.8 (0.4) 0.004 17.3(0.6) 18.8 (0.6) 17.8 (0.8) 19.7 (0.6) 0.06
Protein (g) 16.5 (0.3) 17.0 (0.4) 16.0 (0.4) 0.03 16.2 (0.7) 17.0 (0.6) 15.6 (0.8) 17.3 (0.6) 0.36
Carbohydrate (g) NA4 21.6 (0.4) 22.2 (0.5) 21.0 (0.4) 0.02 21.2 (0.8) 20.6 (0.6) 21.9 (1.0) 22.7 (0.7) 0.19
Total fat (g) 12.7 (0.4) 13.3 (0.5) 12.2 (0.4) 0.04 11.2 (0.7) 15.1 (0.7) 10.6 (0.9) 14.4 (0.7) 0.0001
Saturated fat (g) 14.3 (0.4) 14.9 (0.5) 13.7 (0.5) 0.03 13.0(0.8) 17.1 (0.7) 11.7(1.0) 15.8 (0.8) 0.0001
Sodium (mg) <600 16.6 (0.4) 17.4 (0.5) 15.7 (0.4) 0.0007 16.6 (0.8) 16.6 (0.6) 15.7 (0.9) 17.4(0.7) 0.55
Cholesterol (mg) <75 14.1 (0.5) 14.9 (0.7) 13.3 (0.6) 0.03 13.3 (1.1) 16.8 ( 1.0) 10.8 (1.2) 15.8 (1.0) 0.0015
Vitamin A (RE) 175 34.4 (0.8) 36.5 (1.0) 32.3 (0.9) 0.0003 33.9 (1.6) 32.7 (1.2) 35.8 (2.2) 35.0 (1.4) 0.52
Ascorbic acid (mg) 11 23.3 (0.9) 23.4 (1.1) 23.1 (1.1) 0.81 19.5(1.7) 22.6 (1.5) 24.1 (2.4) 27.2 (1.8) 0.02
Iron (mg) 3 26.9 (0.5) 28.0 (0.7) 25.9 (0.6) 0.005 26.0 (1.1) 23.8 (0.8) 29.3 (1.5) 28.8 (1.0) 0.0004
Calcium (mg) 325 26.4 (0.5) 27.3 (0.7) 25.6 (0.6) 0.01 27.0 (1.1) 27.7 (0.9) 24.3 (1.3) 26.9 (1.0) 0.22
Zinc (mg) 3 19.4 (0.4) 20.2 (0.5) 18.7 (0.5) 0.01 19.4 (0.8) 19.3 (0.6) 20.0 (1.1) 19.1 (0.7) 0.93
1 Adjusted mean (standard error); N=l ,683 children.
2Goals based on National School Lunch Program and School Breakfast Program: School Meals Initiative for Healthy Children (57), 1989 Recommended
Dietary Allowances (34), and National Academy of Sciences, Diet and Health: Implications for Reducing Chronic Disease Risk (33).
3Testing hypothesis of equal means across gender or site.
4NA- not applicable.
School Meal Initiative for Healthy
Children (57), and the Diet and Health
Report of the National Academy of
Sciences ( 3 3).
Overall, the adjusted mean energy intake
at breakfast was 337 calories, with about
14 percent of energy from protein, 65
percent from carbohydrate, 23 percent
from total fat, and 10 percent from satu­rated
fat (table 4 ). Mean sodium and
dietary cholesterol intakes from break­fast
were 459 and 32 mg, respectively.
The average energy intake at breakfast
was significantly lower among girls
than boys (317 vs. 358 calories). Similar
results were noted for protein intake
expressed in grams. Compared with
girls, boys consumed significantly more
8
sodium, dietary cholesterol, vitamin A,
iron, calcium, and zinc at breakfast. But
gender differences disappeared after
adjustment for differences in food
energy intakes. (Data are not shown).
Compared with other students, African
American and Hispanic students con­sumed
higher percentages of energy in
their breakfasts from total fat (23 and 26
percent, respectively) and saturated fat
(11 and 12 percent, respectively) (table 4).
Compared with other children, Hispanic
children consumed less energy from
carbohydrates (61 percent vs. 65 to 68
percent). The students' intakes of energy,
calcium, and zinc at breakfast did not
meet the dietary goals for any of the
race/ethnic groups.
The nutrient profiles of breakfasts
differed among sites, with Minnesota
breakfasts having the most healthful
nutrient profiles (table 5). Compared
with other breakfasts, those in Minnesota
had the lowest percentage of calories
from fat (19 percent), saturated fat (8
percent), and dietary cholesterol (21
mg). Also, breakfasts in Minnesota had
the highest percentage of calories from
carbohydrate (70 percent), vitamin A
(363 RE), and iron (4.3 mg). Compared
with breakfasts at other sites, those in
Texas and Louisiana had more total
fat, saturated fat, and dietary cholesterol;
exceeded the goal for saturated fat and
sodium; but did not contain more food
energy. Breakfasts in Louisiana were
also lower in vitamin A, ascorbic acid,
Family Economics and Nutrition Review
Table 4. Energy and selected nutrients for breakfast meals, by gender and race/ethnicity ,I CATCH
Gender Race/ethnicity
African
Variable Goal2 Overall Boys Girls p3 Caucasian American Hispanic Asian Other p3
Energy (calories) 500 337 (7) 358 (9) 317 (8) <0.001 333 (78) 347 (16) 342(15) 314(32) 396 (42) 0.30
Protein(% calories) NA4 13.6 (0.2) 13.6 (0.2) 13 .6 (0.2) 0.84 13.7 (0.2) 13.0 (0.4) 13.7 (0.4) 14.3 (0.9) 10.8 (1.0) 0.013
Carbohydrate(% calories) NA4 65.0 (0.5) 64.9 (0.6) 65.2 (0.6) 0.68 65.6 (0.6) 65.9 (1.2) 61.3 (1.2) 65.0 (2.9) 68.3 (3 .0) 0.014
Total fat (% calories) <30 23.1 (0.4) 23.4 (0.5) 22.9 (0.5) 0.49 22.5 (0.5) 23.4 (1.1) 26.3 (1.0) 21.8 (2.4) 21 .7 (2.5) 0.011
Saturated fat(% calories) <10 10.4 (0.2) 10.5 (0.3) 10.2 (0.3) 0.30 9.9 (0.2) 10.7 (0.5) 12.2 (0.5) 9.4 (1.2) 9.8 (1.2) <0.001
Sodium (mg) <600 459 (12) 491 (14) 428 (13) <0.001 456 (13) 483 (27) 447 (24) 487 (60) 449 (59) 0.76
Cholesterol (mg) <75 32.0 (1.7) 34.9 (2.2) 29.1 (2.0) O.Ql5 29.2 (1.8) 38.7 (4.6) 39.1 (4.4) 33 .0 (9.6) 37.3 (10.6) 0.06
Protein (g) 7.0 11.7 (0.2) 12.5 (0.3) 11.0 (0.3) <0.001 11.7 (0.3) 12.0 (0.6) 11.9 (0.6) 11.6 (1.3) 10.7 (1.3) 0.95
Vitamin A (RE) 175 309 (9) 335 (12) 284 (11) <0.001 314(11) 332 (23) 269 (20) 379 (57) 241 (47) 0.067
Ascorbic acid (mg) 11 21.2 (1.2) 21.6 (1.4) 20.8 (1.4) 0.56 20.1 (1.3) 25.8 (3.0) 20.8 (2.6) 24.4 (6.5) 36.0 (8.1) 0.064
Iron (mg) 3 3.8(0.1) 4.1 (0.1) 3.5(0.1) <0.001 3.8 (0.1) 4.3 (0.3) 3.4 (0.2) 3.8 (0.6) 3.1 (0.6) 0.11
Calcium (mg) 3255 273 (6) 293 (8) 255 (7) <0.001 278 (7) 272 (15) 266 (14) 248 (32) 205 (30) 0.38
Zinc (mg) 3 1.70(0.05) 1.85 (0.06) 1.57 (0.05) <0.001 1.67 (0.05) 1.84 (0.12) 1.70 (0.11) 2.26 (0.35) 1.47 (0.23) 0.13
1 Adjusted mean (standard error); N= 1,683 children.
2Goals based on National School Lunch Program and School Breakfast Program: School Meals Initiative for Healthy Children (57).
3Testing hypothesis of equal means across gender or race/ethnicity.
4NA - not applicable.
5Value is one-quarter of the 1998 Dietary Reference Intake (55).
and iron, compared with other sites. At
all sites, the breakfasts eaten by children
did not meet intake goals for energy,
calcium, and zinc.
Most breakfast intakes were similar,
whether eaten at home or at school
(table 6). Children who reported eating
breakfasts both at home and at school,
however, had significantly (p<0.05)
higher breakfast intakes of food energy,
protein, and of most other nutrients.
Breakfast intakes for percentage of food
energy from saturated fat and sodium
exceeded goals for children eating
breakfast both at home and at school.
Their breakfast intakes were 705 Kcal,
compared with 326 Kcal for those eating
1998 Vol. 11 No. 4
breakfast at home only and 334 Kcal
for those eating breakfast at school only
(p<0.05). Similarly, total daily energy
intakes were 2,397 Kcal for children
who consumed breakfasts both at home
and at school, compared with 1,928 Kcal
for children who ate breakfast at home
only and 1,976 Kcal for those who ate
breakfast at school only. (Data are not
shown.) No differences were apparent
in body mass indices by gender or by
race/ethnicity for the children who ate
breakfast at both places on the same day
versus those who ate breakfast once: at
home or at school. (Data are not shown.)
Most (63 percent) of those eating break­fast
at both home and school were from
Louisiana.
Mean food energy and most selected
nutrient intakes from breakfast were
not significant by source of the meal
(i.e., whether eaten at home or school or
both) (table 6). The exception was iron.
Compared with home breakfasts, school
breakfasts, on average, contributed sig­nificantly
lower amounts of iron (2.3 vs.
3.8 mg) and contributed less than the 3-
mg dietary goal. This finding persisted
across sites, gender, and the three race/
ethnic groups (p>0.20 for interaction;
data are not shown). Whether consumed
at home or at school, both breakfasts
exceeded goals for percentage intake
from saturated fat (10 and 11 percent,
respectively); both were low in energy,
calcium, and zinc. In Louisiana and
9
Table 5. Energy and selected nutrients for breakfast meals by site,1 CATCH
Site
Variable Goal2 Overall California Louisiana Minnesota Texas p3
Energy (calories) 500 337 (7) 312(13) 336 (12) 342 (17) 361 (13) 0.08
Protein(% calories) NA4 13.6 (0.2) 14.4 (0.3) 13.4 (0.3) 13.3 (0.4) 13.3 (0.3) 0.07
Carbohydrate(% calories) NA4 65.0 (0.5) 67.0 (1.0) 60.1 (0.8) 69.9 (1.3) 63.2 (0.9) <0.001
Total fat(% calories) <30 23.1 (0.4) 21.0 (0.9) 27.5 (0.7) 18.8 (1.1) 25.2 (0.8) <0.001
Saturated fat (% calories) <10 10.4 (0.2) 9.6 (0.4) 12.3 (0.4) 8.3 (0.6) 11.2 (0.4) <0.001
Sodium (mg) <600 459 (12) 419 (22) 463 (20) 446 (28) 510 (22) 0.04
Cholesterol (mg) <75 32.0 (1.7) 29.7 (3.3) 40.0 (3.2) 20.6 (3.6) 39.8 (3.4) <0.001
Protein (g) 7.0 11.7 (0.2) 11.4 (0.5) 11.4 (0.4) 11 .5 (0.6) 12.6 (0.5) 0.24
Vitamin A (RE) 175 309 (9) 307 (19) 240 (14) 363 (26) 333 (17) <0.001
Ascorbic acid (mg) 11 21.2 (1.2) 18.2 (2.2) 17.2 (1.8) 24.0 (3.2) 26.1 (2.4) O.Ql5
Iron (mg) 3 3.8 (0.1) 3.6 (0.2) 3.0 (0.2) 4.3 (0.3) 4.2 (0.2) <0.001
Calcium (mg) 3255 273 (6) 274 (12) 256 (10) 283 (16) 281 (11) 0.32
Zinc (mg) 3 1.70 (0.05) 1.71 (0.09) 1.57 (0.07) 1.86 (0.13) 1.68 (0.08) 0.23
1 Adjusted mean (standard error); N=1 ,683 children.
2Goals based on National School Lunch Program and School Breakfast Program: School Meals Initiative for Healthy Children (57).
3Testing hypothesis of equal means across site.
4NA - not applicable.
5Value is one-quarter of the 1998 Dietary Reference Intake (55).
Texas, breakfasts consumed at school
were higher (p<0.02) in the mean per­centage
of energy from total fat and
saturated fat and lower (p<0.03) in
energy from carbohydrate than were
breakfasts consumed at home. The rela­tive
contribution of breakfast to total
daily intakes did not vary by source of
breakfast (e.g., home or school). (Data
are not shown.)
Discussion
We found that only 6 percent of the
third grade students in the Child and
Adolescent Trial for Cardiovascular
Health (CATCH) skipped breakfast.
This is the same predicted rate for 6- to
1 0-year-olds included in the USDA's
10
School Nutrition Dietary Assessment
study (SNDA) ( 15 ). Other large studies
of primary school children, however,
reported higher percentages of children
who skipped breakfast ( 14, 15,44 ). In
the SNDA study, but not in the CATCH
study, the percentage of students who
ate breakfast were constant across regions
of the country, whether or not the child's
school offered a School Breakfast Program.
But where children who ate breakfast
did differ among sites, more CATCH
third graders than SNDA 6- to 18-year­olds
consumed breakfast at home (84
vs. 69 percent). Comparisons are diffi­cult,
however, because older children
skip breakfast more often than younger
children do ( 15). Sixteen percent of
CATCH students ate a School Breakfast
Program meal, compared with the 25-
percent prediction for 6- to 1 0-year-olds
in the SNDA study. Three-fifths of
CATCH schools provided a School
Breakfast Program; about two-fifths
of schools in the SNDA study did so
(61 vs. 45 percent, respectively).
SNDA concluded that the availability
of a School Breakfast Program did not
influence whether a student ate breakfast.
The Bogalusa Heart Study, however,
reached the opposite conclusion. In the
Bogalusa longitudinal study, prior to
widespread availability of the School
Breakfast Program, the percentage of
children who skipped breakfast was
high, ranging from 9 percent in 1973 to
30 percent in 1979. When the School
Family Economics and Nutrition Review
Table 6. Energy and selected nutrients for breakfast meals, by source
of meal, 1 CATCH
Home School Home and school p2
Variable N = 1,409 N=218 N=56 (Home v. school)
Energy (calories) 326 (6) 334 (31) 705 (82) 0.76
Protein (% calories) 13.7(0.1 ) 13 .4 (0.8) 12.6 (1.1) 0.70
Carbohydrate(% calories) 65.2 (0.4) 64.6 (2.6) 63 .0 (3.3) 0.87
Total fat (% calories) 22.9 (0.4) 23 .7 (2.2) 26.5 (2.8) 0.74
Saturated fat (% calories) 10.2 (0.2) 11.3 (1.0) 11.7 (1.3) 0.30
Sodium (mg) 448 (10) 427(51) 838 (90) 0.70
Cholesterol (mg) 31.2 ( 1.4) 27.3 (8 .0) 71.2 (16.3) 0.66
Protein (g) 11 .4 (0.2) 11.2 (1.2) 22.4 (2.1 ) 0.85
Vitamin A (RE) 307 (8) 260 (43) 546 (79) 0.30
Ascorbic acid (mg) 21.4 (1.0) 15.8 (4.8) 37.3 (9.2) 0.31
Iron (mg) 3.8 (0.1) 2.3 (0.4) 9.0 (1.1) <0.01
Calcium (mg) 264 (5) 276 (30) 518 (53) 0.73
Zinc (mg) 1.65 (0.04) 1.67 (0.23) 3.70 (0.65) 0.89
1 Adjusted mean (standard error); N= I ,683 children.
2-resting hypothesis of equal means between breakfast eaters by source of meal.
Breakfast Program became widely
available, the percentage of students
skipping breakfast declined dramatically
( 42 ). In CATCH, the availability of the
School Breakfast Program did not affect
the percentage of students who skipped
breakfast. Compared with students in
Minnesota and California (84 and 79
percent, respectively), lower percentages
of students in Texas and Louisiana (63
and 70 percent, respectively) ate break­fast
at home, and slightly higher percent­ages
skipped breakfast, even after we
controlled for the availability of the
School Breakfast Program. Although
household income data were unavailable
for individual CATCH children, we
suspect that Texas and Louisiana
schools had more children from poor
and minority families (as determined
by ethnic distribution and number of
1998 Vol. 11 No. 4
children eligible for free or reduced­price
school meals at each site).
The contribution of breakfasts eaten
at home or at school as a percentage of
total daily intakes was similar for most
nutrients. However, for the small number
of children who consumed breakfast
both at home and at school, daily food
energy intakes were higher, mostly
accounted for by the extra food energy
at breakfast. Children who ate two
breakfasts, however, did not weigh
more than other children weighed.
Because most of those eating breakfast
twice came from sites where more schools
were considered low income, it is possible
the children were from poor families
with limited access to food at other
meals and snacks, and the children
relied on the School Breakfast Program
to supplement their intakes. Alterna­tively
, the children may have been
especially hungry, because they were
growing rapidly.
In a related study by our group ( 11 ), we
found the amount of calories provided
by 5 consecutive days of CATCH
school breakfast menus at baseline was
similar to the data reported here. In the
SNDA study, breakfasts consumed at
home provided only 18 percent of the
RDA for food energy for students over­all,
and only 10 percent of the students
who participated in the School Breakfast
Program met or exceeded the target of
25 percent of the RDA for food energy
at breakfast (7). Food energy provided
in the School Breakfast Program in
CATCH conformed to the program's
regulations at the time of the baseline
study.
Regulations adopted after theCA TCH
program started require that school
breakfasts provide 25 percent of the
RDA of 2,025 Kcal per day for children
6- to 11-years-old or about 500 Kcal
and an equivalent proportion of other
nutrients (57). If schools provide only
25 percent of the RDA, on average, it
is unlikely that 25 percent will be con­sumed,
because children rarely eat all
of their food. In other analyses, however,
we found that CATCH third graders'
intakes of both total daily energy and
macronutrient intakes were adequate
(19). Snacks and other meals consumed
throughout the day may have compen­sated
for reduced intakes at breakfast in
this study. Because total dietary intakes
of students nationwide exceeded the
RDA for energy (8), perhaps 25 percent
of the RDA is not as critical for food
energy consumption at breakfast as it
is for vitamins and minerals.
II
12
Breakfast eaters also
had higher intakes of
protein, percentage
of energy from
carbohydrates,
sodium, cholesterol,
and most vitamins
and minerals.
When the SNDA students' daily dietary
intakes were examined, researchers
found that students participating in the
School Breakfast Program consumed
more than the 25-percent target of the
dietary goals for fat, saturated fat, and
cholesterol that is specified by the
National Cholesterol Education program
( 31 ). Those eating breakfasts at home
consumed less of these nutrients and
food energy (7). In contrast, students'
breakfast intakes, regardless of whether
they were at home, at school, or at both
home and school, exceeded the 25-percent
target of the RDA's for most nutrients
(except zinc). This result underscores
the contributions of breakfast to nutri­tional
quality (7,34 ).
Many aspects of SNDA's data collection
and methods of analysis were similar to
those used by CATCH. SNDA, however,
did not incorporate analysis of actual
school recipes and vendor foods into
the 24-hour recalls of students who ate
school meals: This may have required
greater use of generic recipes and food
entries (defaults) than were used in
CATCH analysis. Using defaults can
result in higher nutrient estimates over­all
and may explain some of the differ­ences
in food energy contributions of
the School Breakfast Program between
the two studies (2).
CATCH third graders consumed break­fasts
that were consistent with national
nutrition goals for dietary intakes of
total fat (no more than 30 percent of
energy), saturated fat (10 percent or less
of energy), sodium (600 mg or less), and
cholesterol (75 mg or less) (31,58). For
CATCH third graders, overall, breakfasts
contributed only 13 percent of their
daily total fat, 14 percent of their satu­rated
fat, and 16 percent of their sodium
intakes. Hence, consumption at other
meals or snacks must be responsible for
the excessive 24-hour intakes of these
nutrients ( 19 ). Overall, school breakfast
intakes did not meet the goal of less
than 10 percent of energy from saturated
fat among Hispanics ( 12 percent of
calories) or among children in Louisiana
(12 percent) and Texas (11 percent).
Variation by sites suggests regional
differences in food preparation methods,
and types of foods consumed may also
influence the nutrients consumed at
breakfast (37). To meet fat intake goals
for Healthy People 2000 (59), we need
intervention efforts that focus on school
meals and breakfasts among children
in these race/ethnic groups; in different
regions; and for lunches, snacks, and
dinners.
Mean intakes of protein (g), vitamin A,
ascorbic acid, and iron at breakfast
contributed at least 30 percent of the
RDA's for these nutrients for all gender,
regional, and race/ethnic groups among
CATCH third graders. The exception
was among the small number of girls of
"Other" race/ethnicity ( 34 ). These find­ings
confirm the importance of school
breakfasts in enhancing the quality of
children's nutrient intakes ( 41 ). Based
on the new calcium DRI's (55), intakes
of calcium at breakfast were below the
25-percent goal of the RDA's for all
groups. Average daily calcium intakes,
however, met about 80 percent of the
AI (adequate intake).
Among CATCH third graders (and also
among participants in other studies such
as SNDA), mean zinc intakes at break­fast
were less than one-fourth of the
RDA. But on a daily basis, the children's
intakes reached recommended levels;
therefore, there was little cause for
concern (8). One way to improve the
zinc content of school breakfasts, while
meeting the dietary goals for fat intake,
is to include fortified, ready-to-eat cereals.
Family Economics and Nutrition Review
For example, a recent study shows that
children who consumed ready-to-eat
cereal at any time in a 24-hour period
had significantly higher total daily intakes
of zinc, compared with those who did
not consume ready-to-eat cereals ( 42).
When we analyzed the 24-hourrecalls,
we found that iron in the meals of the
School Breakfast Program in CATCH
schools was about one-third of the RDA
(31 to 34 percent) ( 11 ). Among third
graders eating breakfast at school, iron
intake at breakfast was slightly lower
(23 percent of the RDA) than the desired
percentage of the RDA. Among those
eating breakfast at home, iron intake
was higher (38 percent of the RDA)
than the desired percentage. We attribute
this finding to children not eating all
their breakfast and sampling variability.
The SNDA study, in contrast, found
iron intakes at breakfast were adequate
(40 to 43 percent of the RDA), regard­less
of the source of the meal (7).
The study reported here has several
limitations. Socioeconomic status could
not be assessed for each child, thus relevant
adjustments could not be made for factors
that could have produced different findings
for the subgroups. Use of only a single
24-hour recall on each child is another
limitation. Thus, usual intakes could not
be assessed. Also, evidence shows that
24-hour recalls systematically under­estimate
food intakes by 10 to 20 percent;
therefore, actual intakes may have been
higher than those reported. But no reason
exists to suspect that breakfast intakes
were underreported differentially ( 1 7).
Hence, it is likely that among CATCH
third graders, mean total calorie intakes
may have been higher than the 18 percent
of the RDA reported here.
Moreover, our data consist of weekday
food intake; it is likely that breakfasts
1998 Vol. 11 No. 4
vary between weekdays and weekends
(39). Some children may have reported
snacks as part of the breakfast meal, and
others may have reported foods eaten at
breakfast as snacks. This type of report­ing
introduces error into the analysis.
Because it was not feasible to collect
quantitative data on discretionary salt
used by this population, our estimates
of total dietary sodium are incomplete.
Also, we did not measure intakes from
vitamin and mineral supplements.
Conclusion
Our most striking finding confirms the
adage that children who eat breakfast
tend to have more healthful daily intakes
than those who do not eat breakfast.
Also, eating breakfast-at home or at
school-increased children's daily in­takes
of several vitamins and minerals
and decreased the percentage of calories
from fat. Although breakfast is a valuable
meal for children, it is less and less
likely to be consumed by adults ( 16 ).
If the availability of breakfast at home
decreases because parents are not eating
it, the availability of school breakfast
becomes more important for enhancing
the chances that children will eat health­ful
breakfasts.
There are, however, economic and other
barriers to implementing breakfasts in
many schools. Thus, encouraging
breakfast consumption-at home or at
school-should be a priority in health
promotion programs for children. This
is particularly important among African
American and Hispanic students who
skip breakfast more often and in regions
of the country where skipping breakfast
is more prevalent. It is important among
adolescents because breakfast consump­tion
tends to decline during the second
decade of life. Skipping breakfast is
more prevalent among children from
low-income than higher income families,
but low-income children are also more
likely to participate in the School Break­fast
Program when it is available than
are higher income children.
Information on changes in the food
supply (13) and in children's eating
patterns ( 1, 44) must be considered if
health promotion programs about
children's meals are to be effective.
Therefore, it is important to monitor
children's eating behaviors and dietary
intakes (3,46,53,56). It is also important
that intervention programs and new
initiatives for healthy children provide
strategies for decreasing fat, saturated
fat, and sodium in breakfasts. These
programs also need to include recom­mendations
on how to incorporate foods
that are energy-dense and rich in vitamins
and minerals.
The U.S. Department of Agriculture
and others have joined in a campaign on
child nutrition and health that has made
child nutrition an immediate priority
(25 ). Children must be guided to make
healthful decisions. We nutritionists,
policymakers, and information multipliers
must direct new efforts to better under­stand
children's eating behaviors and
psychosocial factors that influence their
food-related decisions.
Acknowledgments
This research was supported by funds
from the National Heart, Lung, and
Blood Institute (HL-398880, HL-39906,
HL-39852, HL-399927, HL-39870).
We thank Marion Bernstein for her help
in preparing the manuscript.
13
14
References
1. Albertson, A., Tobelmann, R., Engstrom, A., and Asp, E. 1992. Nutrient intakes
of 2 to 20 year old American children: 10 year trends. Journal of the American
Dietetic Association 92( 12): 1492-1496.
2. Arneson, P. 1990. The Minnesota Nutrition Data System. InK. Price and M.
Stevens (Eds.), Version 2.2 User's Manual (p. 51). University of Minnesota,
Minneapolis, MN.
3. Berenson, G., Arbeit, M., Hunter, S., Johnson, C., and Nicklas, T. 1991. Cardio­vascular
health promotion for elementary school children. Hyperlipidemia in childhood
and the development of atherosclerosis. Annals of the New York Academy of Sciences
623:299-313.
4. Berenson, G., Strong, W., Williams, C., Haley, N., Mancini, M., Nicklas, T., et al.
1989. Coronary artery disease prevention: Cholesterol, a pediatric perspective.
Preventive Medicine 18( 3 ):323-409.
5. Bingham, S.A. 1987. The dietary assessment of individuals, methods, accuracy,
new techniques and recommendations. Nutrition Abstracts and Reviews (Series A)
57:705-742.
6. Connors, C. and Blouin, A. 198211983. Nutrition effects on behavior of children.
Journal of Psychiatry Research 17(2): 193-201.
7. Devaney, B., Gordon, A., and Burghardt, J. 1993. The School Nutrition Dietary
Assessment Study: Dietary intakes of program participants and nonparticipants. In
L. Berenson (Ed.), Mathematica Policy Research, Inc. Contract No. 53-3198-0-16.
Food and Nutrition Service, U.S. Department of Agriculture, Princeton, NJ.
8. Devaney, B., Gordon, A., and Burghardt, J. 1995. Dietary intakes of students.
The American Journal of Clinical Nutrition 61 (Suppl 1 ):205S-212S.
9. Division of Health Examination Statistics, National Center for Health Statistics,
Division of Nutrition and Physical Activity, National Center for Chronic Disease
Prevention and Health Promotion, CDC. 1997. Update: Prevalence of Overweight
Among Children, Adolescents and Adults-United States 1988-94. Morbidity and
Mortality Weekly Report 46:199-201.
10. Dwyer, J. 1995. The School Nutrition Dietary Assessment Study. The American
Journal of Clinical Nutrition 6l(Suppl): 173S-177S.
11. Dwyer, J.T., Hewes, L.V., Mitchell, P.D., Nicklas, T.A., Montgomery, D.H.,
Lytle, L.A. , Snyder, M.P., Zive, M.M., Bachman, K.J., Rice, R. , and Parcel, G.S.
1996. Improving school breakfasts: Effects of the CATCH Eat Smart program on
the nutrient content of school breakfasts. Preventive Medicine 24:413-422.
12. Ebzery, M., Montgomery, D., Evans, M., Hewes, L., Zive, M., Reed, D., Rice,
R. , Hann, B., and Dwyer, J.T. 1996. School meal data collection and documentation
methods in a multisite study . School Food Service Research Review 20:69-77.
Family Economics and Nutrition Review
13. Friend, B., Page, L., and Martson, R. 1979. Food consumption patterns in the
United States: 1909-13 to 1976. In R. Levy, B. Rifkind, B. Dennis, and N. Ernst
(Eds.), Nutrition, Lipids and Heart Disease. Raven Press, New York, NY.
14. Gleason, P. 1995. Participation in the National School Lunch Program and the
School Breakfast Program. The American Journal of Clinical Nutrition 61(Suppl1 ):
213S-220S.
15. Gordon, A., Devaney, B., and Burghardt, J. 1995. Dietary effects of the National
School Lunch Program and the School Breakfast Program. The American Journal of
Clinical Nutrition 61 (Suppl 1 ):221 S-231 S.
16. Haines, P.S., Guilkey, D.K., and Popkin, B.M. 1996. Trends in breakfast
consumption of US adults between 1965 and 1991. Journal of the American Dietetic
Association 96:464-470.
17. Hanes, S., Vermeersch, J., and Gale, S. 1984. The National Evaluation of School
Nutrition Programs: Program impact on dietary intake. The American Journal of
Clinical Nutrition 40(8):390-413.
18. Johnson, R., Guthrie, H., Smickilas-Wright, H., and Wang, M. 1994. Charac­terizing
nutrient intakes of children by sociodemographic factors. Public Health
Reports 109(3):414-420.
19. Lytle, L., Ebzery, M., Nicklas, T., Montgomery, D., Zive, M., Evans, M., et al.
1996. Nutrient intakes of third graders: Results from the Child and Adolescent Trial
for Cardiovascular Health (CATCH) baseline survey. Journal of Nutrition Education
28:338-347.
20. Lytle, L., Nichaman, M., Obarzanek, E., Glovsky, E., Montgomery, D., Nicklas, T.,
et al. 1993. Validation of 24-hour recalls assisted by food records in third grade
children. Journal of the American Dietetic Association 93( 12): 1431-1436.
21. Luepker, R., Perry, C., McKinlay, S., Nader, P., Parcel, G., Stone, E., et al.
1996. Outcomes of a field trial to improve children's dietary patterns and physical
activity: The Child and Adolescent Trial for Cardiovascular Health (CATCH).
Journal of the American Medical Association 27 5( 10 ):768-77 6.
22. McDowell, M., Briefel, R., Alaimo, K., Bischof, A., Caughman, C. , Carroll, M.,
eta!. 1994 (October). Energy and macronutrient intakes of persons ages 2 months
and over in the United States: Third National Health and Nutrition Examination
Survey, Phase 1, 1988-91. Advance Data from Vital and Health Statistics; No. 255.
U.S. Department of Health and Human Services, National Center for Health Statistics,
Publication No. DHHS-PHS-95-1250, Hyattsville, MD.
23. Michaud, C., Musse, N., Nicholas, J., and Mejean, L. 1991. Effects of breakfast
size on short term memory, concentration, mood and blood glucose. Journal of
Adolescent Health 12( 1 ):53-57.
1998 Vol. 11 No. 4 15
16
24. Moffitt, R. 1995. Commentary on who participates in the National School Lunch
Program and the School Breakfast program. The American Journal of Clinical Nutrition
61 (Suppl):250S-251S.
25. Monsen, E.R. 1995. Our children are our future. Journal of the American
Dietetic Association 95:1082.
26. Morgan, K., Zabik, M., and Leveille, G. 1981. The role of breakfast in the nutrient
intake of 5- to 12-year-old children. The American Journal of Clinical Nutrition
34(7): 1418-1427.
27. Morgan, K., Zabik, M., and Stampley, G. 1986. Breakfast consumption patterns
of U.S. children and adolescents. Nutrition Research 6:635-646.
28. Morris, P., Bellinger, M., and Hass, E. 1991 (September). Heading for a health
crisis: Eating patterns of America's school children, p. 42. Public Voice for Food
and Health Policy, Washington, D.C.
29. Murray, D. and Wolfinger, R. 1994. Analysis issues in the evaluation of community
trials: Progress toward solutions in SAS/STAT MIXED. Journal of Community
Psychology 140-154.
30. Myers, A., Sampson, A., Weitzman, M., Rogers, B., and Kayne, H. 1989. School
breakfast program and school performance. American Journal of Diseases of Children
143:1234-1239.
31. National Cholesterol Education Program. 1991 (September). Report of the expert
panel on blood cholesterol levels in children and adolescents. U.S. Department of
Health and Human Services, Public Health Service, National Institutes of Health,
National Heart, Lung, and Blood Institute. Publication No.: NIH-NHLBI-91-2732,
Bethesda, MD.
32. National Heart, Lung, and Blood Institute's Task Force on Blood Pressure in
Children, Report of the Second Task Force on Blood Pressure Control in Children.
1987 (January). U.S. Department of Health and Human Services, Public Health
Service, National Institutes of Health, Bethesda, MD.
33. National Research Council, Committee on Diet and Health, Food and Nutrition
Board. 1989. Diet and Health: Implications for Reducing Chronic Disease Risk,
p. 749. National Academy Press, Washington, DC.
34. National Research Council, Subcommittee on the Tenth Edition of the RDAs,
Food and Nutrition Board. 1989. Recommended Dietary Allowances. National Academy
Press, Washington, DC.
35. Nicklas, T. 1995. Dietary studies of children: The Bogalusa Heart Study Experience.
Journal of the American Dietetic Association 95( I 0): 1127-1133.
36. Nicklas, T., Bao, W., and Berenson, G. 1993. Nutrient contribution of the breakfast
meal classified by source in 10 year old children: Home versus school. School Food
Service Research Review 17(2): 125-131.
Family Economics and Nutrition Review
37. Nicklas, T., Bao, W., Webber, L., and Berenson, G. 1993. Breakfast consumption
affects adequacy of total daily intake in children. Journal of the American Dietetic
Association 93(8):886-891.
38. Nicklas, T., Farris, R., Bao, W., and Berenson, G. 1995. Temporal trends in
breakfast consumption patterns of 10 year old children: The Bogalusa Heart Study.
School Food Service Research Review 19(2):72-80.
39. Nicklas, T.A., Farris, R.P., Bao, W., Webber, L.S., and Berenson, G.S. 1997.
Differences in reported dietary intake of 10 year old children on weekdays compared to
Sunday: The Bogalusa Heart Study. Nutrition 17:31-40.
40. Nicklas, T., Farris, R., Myers, L., and Berenson, G. 1995. Dietary intakes of
electrolytes, minerals and trace metals in children's diets: The Bogalusa Heart
Study. Journal of Advances in Medicine 8(4):241-258.
41. Nicklas, T., Montgomery, D., and Mitchell, P. Micronutrient intake of third
grade children. Unpublished manuscript.
42. Nicklas, T., Myers, L., and Berenson, G. 1994. Impact of ready to eat cereal
consumption on total dietary intake of children: The Bogalusa Heart Study. Journal
of the American Dietetic Association 94(3):316-318.
43. Nicklas, T.A., O'Neil, C.E., and Berenson, G.S. Breakfast: Nutrient contribution,
secular trends and the role of ready to eat cereals: A review of the Bogalusa Heart
Study. American Journal of Clinical Nutrition. (in press).
44. Nicklas, T., Webber, L., Srinivasan, S., and Berenson, G. 1993. Secular trends in
dietary intakes and cardiovascular risk factors of 10 year old children: The Bogalusa
Heart Study (1973-1988). American Journal of Clinical Nutrition 57:930-937.
45. Pao, E., Fleming, K., Guenther, P., and Mickle, S. 1982. Foods commonly eaten
by individuals: Amounts per day and per eating occasion. USDA Home Economics
Research Report, No. 44. Human Nutrition Information Service, Washington, DC.
46. Perry, C., Klepp, K., and Sillers, C. 1989. Community wide strategies for cardio­vascular
health: The Minnesota Heart Health Program: Youth Program Health
Education Research. Theory and Practice 4( 1 ):87-101.
47. Perry, C., Stone, E., Parcel, G., Ellison, R., Nader, P., Webber, L., and Luepker, R.
1990. The Child and Adolescent Trial for Cardiovascular Health (CATCH). Journal
of School Health 60( 8):406-413.
48. Pollitt, E. 1995. Does breakfast make a difference in school? Journal of the
American Dietetic Association 95( 10 ): 1134-1139.
49. Pollitt, E., Leibel, R., and Greenfield, D. 1981. Brief fasting, stress, and cognition in
children. The American Journal of Clinical Nutrition 34(8): 1526-1533.
1998 Vol. 11 No. 4 17
50. Pollitt, E., Lewis, N., Garza, C., and Shulman, R. 1982/1983. Fasting and cognitive
function. Journal of Psychiatric Research 17 (2 ): 169-17 4.
51 . Resnicow, K. 1991. The relationship between breakfast habits and plasma
cholesterol levels in school children. Journal of School Health 61(2 ): 81-85.
52. SAS/ST AT Software: Changes and enhancements, release 6.07. 1992. SAS
technical report P-229 (289-368). SAS Institute, Inc., Cary, NC.
53. Simons-Morton, B., Parcel, G., Baranowski, T., Forthofer, R., and O'Hara, N.
1991. Promoting physical activity and a healthful diet among children: Results of a
school based intervention study. American Journal of Public Health 81(8):986-991.
54. Skinner, J., Salvetti, N., Ezell, J., Penfield, M., and Costello, C. 1985. Appalachian
adolescents' eating patterns and nutrient intakes. The Journal of the American
Dietetic Association 85(9 ): 1093-1099.
55. Standing Committee on the Scientific Evaluation of Dietary Reference Intakes,
Food and Nutrition Board, Institute of Medicine. 1999. Dietary Reference Intakes
for Calcium, Phosphorus, Magnesium, Vitamin D, and Fluoride. National Academy
Press, Washington, DC.
56. Stone, E. 1990. ACCESS: Keystones for school health promotion. Journal of
School Health 60(7):298-300.
57. U.S. Department of Agriculture, Food and Consumer Service, 7 CFR Pts. 210-220.
1995 (June 13). National School Lunch Program and School Breakfast Program:
School Meals Initiative for Healthy Children: Final rule. Federal Register
60( 113 ):31188-31222.
58. U.S. Department of Agriculture and U.S. Department of Health and Human
Services. 1995. Nutrition and Your Health: Dietary Guidelines for Americans.
(4th ed.). U.S. Department of Agriculture. Home and Garden Bulletin No. 232.
59. U.S. Department of Health and Human Services. 1990. Healthy People 2000:
National Health Promotion and Disease Prevention Objectives. U.S. Public Health
Service, Washington, DC. Publication No. DHHS-PHS-92-50213.
60. Webber, L., Harsha, D., Nicklas, T., and Berenson, G. 1994. Secular trends in
obesity in children. In L. Filer, R. Lauer, and R. Luepker (Eds.), Prevention of
atherosclerosis and hypertension beginning in youth. Lea and Febiger, Philadelphia, PA.
61. Webber, L., Osganian, S., Luepker, R., Feldman, H., Stone, E., Elder, J., et al.
1995. Cardiovascular risk factors among third grade children in four regions of the
United States: The CATCH Study. American Journal of Epidemiology 141:428-439.
62. Wyon, D. and Wyon, I. 1992. The effects of breakfast on school work. ILSI-NS
Nutrition and Central Nervous System Function Conference, February 26.
18 Family Economics and Nutrition Review
Cara Janette Miller
The Ohio State University
Catherine P. Montalto
The Ohio State University
1998 Vol. 11 No. 4
Comparison of Economic
Status of Elderly Households:
Nonmetropolitan Versus
Metropolitan Residence
Elderly households in nonmetropolitan areas have lower economic status
than do their metropolitan counterparts, as determined by several measures:
Income, expenditures, and financial assets. Data from the 1990-94 Consumer
Expenditure Survey indicate that nonmetropolitan elderly households have
80 to 83 percent as much income and 79 to 82 percent as much expenditures
as metropolitan elderly households. We find that after controlling for age,
education, gender, marital status, race, home ownership, and presence of
at least one earner in the household, nonmetropolitan and metropolitan
differences persist, but as expected, are somewhat smaller. The multivariate
models that control for demographic characteristics indicate that nonmetro­politan
elderly households have 83 to 88 percent as much income and 81 to
85 percent as much expenditures as metropolitan elderly households. We
discuss the public policy implications of these persistent nonmetropolitan and
metropolitan differences in economic status.
s the number and percentage
of elderly people in the
United States continue to
increase, there is much
concern over the financial well-being
and economic status of this growing
segment of the population. For 50 years,
the elderly population has benefitted
from the creation and expansion of public
programs and, as a whole, has experi­enced
increases in income and wealth
and declines in poverty rates ( 14). These
improvements in economic status, how­ever,
conceal high risks of poverty still
faced by some subgroups of the elderly
population.
Previous research has linked economic
well-being of the elderly population to
age, living arrangements, gender, marital
status, and race (2,3,7,8,1 1,13). How­ever,
in research examining risks of
poverty and low economic status among
the elderly, geographic location has
received less attention. The limited
research that has compared nonmetro­politan
and metropolitan elders confirms
the relative economic disadvantage of
nonmetropolitan elders.1 For example,
1The U.S. Bureau of the Census defines a metro­politan
area as a county or counties containing a
place or urbanized area of 50,000 people or more
with a total population of I 00,000, including
adjacent communities that have a high degree of
economic and social integration with the central
city. A nonmetropolitan area refers to counties
outside a metropolitan area. The metropolitan
and nonmetropolitan focus is used in this research
because work and residence patterns are likely to
be tied more closely to metropolitan and nonmetro­politan
residence than to urban/rural residence.
19
nonmetropolitan elderly households are
more likely to be poor and to have
lower incomes, compared with their
metropolitan counterparts ( 6,8, 10,15 ).
Compared with elders in urban and
metropolitan areas, elders living in rural
and nonmetropolitan areas are more
sparsely located and receive less media
attention (5).
Research analyzing differences by
geographic location of residence is
important because the elderly are over­represented
in rural and nonmetropolitan
areas. About three times as many elders
live in metropolitan areas as in nonmetro­politan
areas. Elderly people, however,
make up higher percentages of non­metropolitan
populations, compared
with metropolitan populations ( 19).
Income is the most commonly used
indicator of economic status. Income
captures one resource of elderly house­holds
but ignores the use of savings or
accumulated financial assets that elders
can use to meet current economic needs.
For example, income flow generally
decreases dramatically when people retire,
but retired people often use savings
and other assets to purchase goods and
services. If these resources are ignored,
the economic status of the elderly will
be underestimated. Thus, measures of
household expenditure or financial
assets may be important indicators of
economic status, particularly for elderly
households.
Measures of economic status should be
adjusted for household need to represent
more accurately a household's economic
status ( 4 ). Measures of total household
income and total household expenditure
ignore differences in need across house­holds
of different sizes. If household
size is ignored, the relative economic
status of larger households will be over-
20
estimated. Per capita and equivalent
measures are frequently used to adjust
for household need. Per capita estimates
are obtained by dividing household re­sources
by the number of persons living
in the household. This measure implies
that household need (and therefore cost)
increases proportionately as household
size increases ( 1 ). Equivalent estimates
are obtained by dividing household
resources by a household equivalence
factor, allowing for economies of scale
that vary with size of the household and
characteristics of household members
(12).
Differences in economic status of non­metropolitan
and metropolitan elderly
households may be partially explained
by nonmetropolitan and metropolitan
differences in demographic charac­teristics
that are related to economic
status. Research has established that
being relatively young, more educated,
married, and White are associated posi­tively
with the economic status of the
elderly population (2,3,7,8,1 1,13).
Also, nonmetropolitan and metropolitan
differences in economic status may be
partially explained by differences in
"opportunity structures" in nonmetro­politan
versus metropolitan areas.
"Opportunity structures" refers to
potential residential and employment
opportunities in a geographic area. It also
refers to socioeconomic characteristics
of the area that influence the availability
and quality of employment and the like­lihood
different groups of people have
for obtaining employment ( 16). People
living in nonmetropolitan areas face
different economic and labor market
opportunities than do those living in
metropolitan areas ( 17). People in non­metropolitan
areas often have more
limited choices; they are less likely
than their metropolitan counterparts to
pursue postsecondary education and are
more likely to have low-paying, unstable
jobs (5). These disadvantages persist
through people's years in the labor market
and influence the level of resources that
are available to them to pay for goods
and services during retirement.
This study contributes to the research
on differences in the economic status of
elderly households in nonmetropolitan
versus metropolitan areas. It examines
the magnitude of differences in economic
status by using multiple measures of
economic status. Further, multivariate
analysis is used to examine whether
nonmetropolitan and metropolitan
differences in economic status remain
when other demographic correlates of
economic status are controlled. Thus,
the persistence of a nonmetropolitan
and metropolitan difference in a multi­variate
framework would support the
theory that residential and employment
opportunities in specific geographic
areas influence differences in economic
status.
Methods
Data and Sample
The data for this research are from the
interview component of the 1990-94
Consumer Expenditure Survey (CE)
conducted by the U.S. Bureau of the
Census for the U.S. Bureau of Labor
Statistics (BLS) (20). The CE's data on
income, expenditure, and total liquid
financial assets were used to construct
indicators of economic status for each
household. Household is used to refer to
a BLS consumer unit. The BLS defines
a consumer unit as (I) all members of a
particular housing unit who are related
by blood, marriage, adoption, or other
legal arrangements; (2) two or more
people living together who pool their
incomes to make joint expenditure
Family Economics and Nutrition Review
decisions; or (3) a person living alone
or sharing a household with others or
living as a roomer in a private home or
lodging house or in permanent living
quarters in a hotel or motel, but who is
financially independent (20). For this
study, only households that were inter­viewed
in four consecutive quarters
(excluding the initial bounding interview)
between the first quarter of 1990 and
the fourth quarter of 1994 were included.
Expenditures over the four consecutive
quarters were summed to obtain actual
annual household expenditure for each
household. All dollar values were
adjusted to 1994 dollars.
To be included in the analysis, households
had to be complete income reporters.
Ninety percent of nonmetropolitan
elderly households and 89 percent of
metropolitan elderly households in the
sample were classified by BLS as
complete income reporters. A house­hold
is classified as a complete income
reporter if the respondent provides values
for major sources of income, such as
wages and salaries, self-employment,
and Social Security. Also, to be included,
the householder had to be 65 years old
or older, and the household could not
contain children less than 18 years old.
Households with dependent children
have different needs and available
resources. Thus, they were expected to
differ systematically from households
without dependent children.
The final sample consisted of 3,334
elderly households: 751 nonmetropolitan
and 2,583 metropolitan. About 25 percent
of elderly persons live in non metropolitan
areas ( 19 ). The unit of analysis for this
research is households with a house­holder
65 years old or older. About 23
percent of the elderly households in the
sample were located in nonmetropolitan
areas.
1998 Vol. 11 No. 4
Measures of Economic Status
Multiple measures of economic status
were used to compare the economic status
of nonmetropolitan and metropolitan
elderly households because there is no
agreement on the best measure to use.
By using several measures, we were able
to determine whether the results differed
based on the empirical measure used.
The measures differed both in the spe­cific
economic resource measured (i.e.,
income, expenditure, and financial assets)
and in the method used to adjust for
household needs (i.e., per capita and
equivalent measures). The specific
measures consisted of total, per capita,
and equivalent annual household income
and expenditure and the value of house­hold
financial assets: the sum of money
in savings, checking, and brokerage
accounts, U.S. savings bonds, stocks,
bonds, mutual funds, and securities.
(See box.)
Per capita household income (expendi­ture)
was calculated by dividing total
household income (expenditure) by
the number of persons living in the
household. Equivalent household income
(expenditure) was calculated by dividing
total household income (expenditure)
by the household's equivalence factor
implicit in the U.S. poverty thresholds.
Definitions for Each Measure of Economic Status
Ratio: Ratio of mean value for nonmetropolitan households to mean value for
metropolitan households.
Total annual household income: Reported household before-tax income excluding
the value of food stamps.
Per capita annual household income: Total household income divided by
household size.
Equivalent annual household income: Total household income divided by the
household equivalence factor.
Total annual household expenditure: Sum of four quarters of reported total house­hold
expenditure.
Per capita annual household expenditure: Total household expenditure divided by
household size.
Equivalent annual household expenditure: Total household expenditure divided
by the household equivalence factor.
Total financial assets (total sample): Sum of money in savings, checking and
brokerage accounts, U.S. savings bonds, stocks, bonds, mutual funds, and securities
for the total sample of 751 nonmetropolitan and 2,583 metropolitan households.
Total financial assets (subsample): Total financial assets for the subsample of
526 nonmetropolitan and 1,879 metropolitan households with some positive amount
of financial assets.
21
The poverty thresholds are the most
widely recognized absolute standard
of need in the United States and are
commonly used in studies of relative
economic status. The 1994 poverty
thresholds used in this research are for
households with a householder 65 years
old or older and containing no related
children under age 18. The equivalence
factor was calculated by dividing the
poverty threshold for a given household
size by the poverty threshold for a one­person
household. For example, the
poverty threshold for a two-person
household ($8,958) was divided by
the poverty threshold for a one-person
household ($7,108) to yield an equiva­lence
factor of 1.26 for a two-person
household.
According to this scale, an elderly couple
needs 26 percent more income than a
single elderly person needs to achieve
the same level of well-being. This implies
large returns-to-scale in consumption.
In contrast, budget share-based scales
typically have smaller returns-to-scale.
The relative economic status of non­metropolitan
and metropolitan elderly
households does not change substantively
when a budget-share scale is used instead
of the implicit scale in the poverty
threshold (4).
The assumptions regarding economies
of scale underlying the various measures
are different: Total household income or
expenditure assumes infinite economies
of scale, per capita income or expendi­ture
assumes no economies of scale;
and equivalent income or expenditure
assumes finite economies of scale and
thus is between the two extremes.
The value of household financial assets
was used as a separate indicator of eco­nomic
status because these assets are
very liquid and are commonly used by
22
elderly households to purchase goods
and services. Home equity represents
a less liquid asset than do financial
assets, and the appropriate treatment of
home equity in the analysis of relative
economic status is much more contro­versial.
Home equity is the most impor­tant
component of wealth for elders.
The same is true for other age groups
in the United States.
However, elders' ability to use this wealth
to purchase other goods and services
requires them to sell their house or use
market mechanisms such as second
mortgages, home equity loans, and
reverse mortgages to convert home
equity to a more liquid form. In reality,
most retired elderly people do not sell
their homes or use reverse mortgages to
finance their consumption ( 1 8). There­fore,
in this research, we excluded home
equity from the measures of economic
status. A dichotomous variable equal to
one if the reference person was a home­owner,
zero otherwise, was included
as an independent variable in the multi­variate
analysis. This controlled for any
correlation between home ownership and
income, expenditure, and financial assets.
Excluding home equity has two poten­tially
opposing effects. Most elderly
own their homes, but home ownership
varies by nonmetropolitan and metro­politan
residence. Nonmetropolitan
elderly households are more likely than
their metropolitan counterparts to own
their homes and to do so without a
mortgage. (In the sample, 83 percent
of nonmetropolitan and 77 percent of
metropolitan elders were homeowners.
Seventy-three percent of nonrnetropolitan
and 62 percent of metropolitan elders
owned their home without a mortgage.)
However, home equity also varies by
nonmetropolitan and metropolitan
residence. The median value of homes
is higher in metropolitan areas than in
nonmetropolitan areas, a reflection, in
part, of the higher land values in metro­politan
areas (21 ).
Relative to the economic status of
metropolitan elderly households, home
ownership rates for nonmetropolitan
elderly households would improve their
economic status, and lower home values
would lower it. Thus, it is difficult to
determine the net effect of excluding
home equity on our results regarding
the relative economic status of non­metropolitan
and metropolitan elderly
households. However, it is likely that
ignoring home equity as an economic
resource is more critical in intergenera­tional
comparisons of economic status
than in nonmetropolitan and metropolitan
comparisons among elderly households.
The influence of home ownership, home
equity, and housing choice on the relative
economic status of nonmetropolitan and
metropolitan elderly households is an
important topic for further research.
Empirical Analysis
First, we compared the measures of eco­nomic
status between nonmetropolitan
and metropolitan elderly households.
We used two sample t-tests to identify
statistically significant differences in the
mean value of the measures of economic
status between nonmetropolitan and
metropolitan elderly households. Then,
we calculated nonmetropolitan to metro­politan
ratios for each measure of eco­nomic
status to determine the magnitude
of differences between the groups. A
ratio of one indicates equivalent eco­nomic
status at the mean values; a ratio
less than one indicates lower economic
status of nonmetropolitan elderly house­holds
relative to metropolitan elderly
households.
Family Economics and Nutrition Review
Second, differences in economic status
of nonmetropolitan and metropolitan
elderly households may be partially
explained by differences in demographic
characteristics that are related to eco­nomic
status. Hence we summarized
demographic characteristics and used
appropriate statistical tests to identify
characteristics that were significantly
different between nonmetropolitan and
metropolitan elderly households.
Third, we used multivariate regressions
to examine determinants of economic
status and to ascertain whether nonmetro­politan
and metropolitan differences
remained when demographic charac­teristics
were controlled. Regression
equations were estimated on the total
sample of elderly households, and a
dichotomous variable for nonmetro­politan
residence was included as an
explanatory variable. Separate equations
were estimated for each measure of
economic status.
Results
Comparisons of Economic Status
of Nonmetropolitan and Metro­politan
Elderly Households
The eight measures of economic status
produced consistent results (table 1 ).
In general, adjusting the measures for
household need reduced the magnitude
of the nonmetropolitan and metropolitan
differences between elderly households,
and the differences were larger based
on expenditure measures, compared
with income measures. However, the
magnitude of these differences was
never greater than 3 percentage points.
What was the economic status of non­metropolitan
elderly households­compared
with their metropolitan
counterparts? Results showed that the
1998 Vol. Jl No. 4
mean values of measures of economic
status for nonmetropolitan elderly
households were lower than those for
metropolitan elderly households. This
was true for all measures analyzed in
this research. Ratios showed that non­metropolitan
elderly households had
80 to 83 percent as much income and
spent 79 to 82 percent as much as their
metropolitan counterparts. The equivalent
and per capita measures of income and
expenditure produced nonmetropolitan
and metropolitan ratios that were slightly
larger (indicating smaller differences)
than the ratios based on total income
and total expenditure. Differences in
ratios for financial assets were more
pronounced between the two groups.
For the total sample, the value of financial
assets for nonmetropolitan elderly
households was 72 percent as much as
that of their metropolitan counterparts.
Among those households with some
positive amount of financial assets,
the ratio for financial assets increased
to 75 percent.
Demographic Characteristics
of Nonmetropolitan and
Metropolitan Elderly Households
The demographic characteristics of non­metropo)
jtan and metropolitan elderly
households were significantly different
(table 2). Compared with metropolitan
elderly households, higher percentages
of reference persons in nonmetroplitan
elderly households were male, White,
and married. The percentage of reference
persons with at least a high school
diploma was higher for metropolitan
households, compared with nonmetropo)jtan
households. A higher percentage of non­metropolitan
elders owned their homes
and reported that there were no earners
in the household. The age of the refer­ence
person in nonmetropolitan versus
metropolitan households did not differ
significantly .
For the total sample, the
value of financial assets
for nonmetropolitan
elderly households was
72 percent as much as
that of their metropolitan
counterparts.
23
Table 1. Mean value of measures of economic status of nonmetropolitan
and metropolitan elderly households
Nonmetropolitan Metropolitan
Ratio1 Measure of economic status (N=751) (N=2,583)
Total annual household income 18,157 22,715 0.80
(15,716) (19,730)
Per capita annual household income 11,635 13,970 0.83
(9,605) (10,567)
Equivalent annual household income 15,282 18,615 0.82
(12,620) (14,690)
Total annual household expenditure 16,247 20,449 0.79
(10,619) (14,843)
Per capita annual household expenditure 10,608 12,934 0.82
(6,434) (8,626)
Equivalent annual household expenditure 13,774 16,956 0.81
(8,419) (11 ,389)
Total financial assets (total sample/ 18,763 26,079 0.72
(36, 174) (47,022)
Total financial assets (subsample)3 26,788 35,850 0.75
(40,669) (51 ,860)
Note: Standard deviations are in parentheses. There are statistically significant nonmetropolitanand
metropolitan differences at the mean value of all measures of economic status at the 99-~rcent confidence
level. The two sample t-test was used. The test statistic was constructed as (XI.X2)/(s12/n1 + s22/n2)
where X;, s?, and n; are the mean, estimate of variance, and number of observations for the ith sample.
The test statistic has a !-distribution.
1Ratio of mean value for nonmetropolitanhouseholds to mean value for metropolitan households.
2Sum of money in savings, checking and brokerage accounts, U.S. savings bonds, stocks, bonds, mutual
funds, and securities for the total sample of751 nonmetropolitanand 2,583 metropolitan households.
3-rotal financial assets for the subsample of 526 nonmetropolitan and 1,879 metropolitan households
with some positive amount of financial assets.
Nonmetropolitan and metropolitan
differences in gender, race, and marital
status would suggest higher economic
status for nonmetropolitan households
relative to metropolitan households;
differences in education and presence
of at least one earner in the household
would suggest higher economic status
for metropolitan households, compared
with nonmetropolitan households.
Previous research documents the correla­tion
of age, education, gender, race, and
marital status with economic status of
24
elderly persons (2,3,7,8,1 1,13). Differ­ences
in the composition of elderly
households in nonmetropolitan and
metropolitan areas suggest that economic
status should be higher in nonmetropolitan
areas (the exceptions: education and
presence of at least one earner in the
household). However, across all measures
of economic status that we analyzed,
economic status is lower among non­metropolitan
elderly households. To
separate the contribution of demographic
composition and nonmetropolitan resi­dence,
we used multivariate analysis to
examine the magnitude of nonmetropolitan
and metropolitan differences in economic
status, controlling for differences in
demographic characteristics.
Determinants of the Economic
Status of Elderly Households
We used multivariate regression analysis
to determine whether the nonmetropolitan
and metropolitan difference in economic
status remained-once the independent
effects of demographic characteristics
were controlled. Multivariate regression
results show the effect of each inde­pendent
variable while simultaneously
controlling for the effects of all other
independent variables.
Each measure of economic status was
used as an independent variable in
separate equations. The independent
variables included measures of age,
education, gender, marital status, and
race ofthe reference person; home owner­ship;
earners; and nonmetropolitan and
metropolitan residence. We measured
each as follows:
• Age and education-with categorical,
dichotomous variables to allow for
nonlinear effects on economic
status.
• Age of the reference person-with
three categorical dichotomous
variables: 65 to 74 years of age (the
reference category), 75 to 84 years
of age, and 85 years and over.
• Educational attainment of the
reference person-with five cate­gorical,
dichotomous variables:
Elementary school or Jess including
no formal schooling (the reference
category), at least some high school,
high school graduation, at least some
college, and college graduation or
more.
Family Economics and Nutrition Review
• Gender of the reference person­with
a dichotomous variable equal
to one if the reference person was
male.
• Marital status-with a dichotomous
variable equal to one if the reference
person was manied. Thus, reference
persons who were widowed,
divorced, separated, or never
married were all coded as zero.
• Race-with a dichotomous variable
equal to one if the reference person
was White.
• Home ownership-with a dichoto­mous
variable equal to one if the
reference person was a homeowner.
• Earners-with a dichotomous
variable equal to one if there were
no earners in the household.
• Residence-with a dichotomous
variable equal to one if the reference
person lived in a nonmetropolitan
area, zero otherwise.
The effects of the independent variables
on economic status are similar across
the measures of economic status, with
most of the independent variables having
statistically significant effects. Table 3
presents statistically significant results.2
2The R2 statistic is a commonly used index of 2 how well an estimator fits the sample data. The R
statistic indicates the percentage of the variation in
the dependent variable that is explained linearly
by the variation in the set of independent variables.
The R 2 statistic adjusted to account for degrees of
freedom is called the "adjusted-R2
" R2 statistics
are sensitive to the range of variation of the de­pendent
variable; in general, measures of R2 are
inversely related to the amount of varia!ion in .
the dependent variable. The adjusted-R statJsncs
for the eight regression models estimated in this
research vary in a manner consistent with our
expectations. The amount of variation in the finan­cial
asset variables is large relative to the amount
of variation in the income and expenditure measures,
resulting in lower measures ofR2 in ~he models
for financial assets. In general, the R measure IS
largest for total income (expenditure), slightly
smaller for equivalent income (expenditure), and
declines further for per capita income (expenditure).
1998 Vol. II No.4
Table 2. Characteristics of elderly households by nonmetropolitan and
metropolitan residence1
Total Nonmetropolitan Metropolitan
Characteristic2 (N=3,334) (N=751) (N=2,583)
Percent
Reference person
Age (in years)
65-74 57 55 58
75-84 35 36 34
85 and over 8 9 8
Education***
Elementary school or less 25 36 22
Some high school 18 18 18
High school graduate 29 25 30
Some college 15 12 16
College graduate or more 13 9 14
Male** 54 58 53
Married*** 46 49 45
White*** 89 95 88
Household
Homeowner*** 78 83 77
No earners** 70 73 69
1Data are column percentages.
~he test statistic for the categorical and dichotomous variables was constructed as :E(O; - E;)2/E; where
0; and E; refer to the observed and expected frequency, respectively, for a given cell. The test statistic
has a chi-squaredistribution.
**Characteristics between norunetropolitan and metropolitan elderly households are significantly different at
p$.01.
***Characteristics between nonmetropolitanand metropolitan elderly households are significantly
different at p$.001.
Across all measures of economic status,
households with a reference person who
was more highly educated, male, White,
and a homeowner had higher economic
status, compared with counterparts. All
other things equal, metropolitan elderly
households, on average, had higher eco­nomic
status than did nonmetropolitan
elderly households.
The effects of age, being married, and
having no earners in the household varied
with the specific measure of economic
status. Age did not have a statistically
significant effect on income measures of
economic status for elderly households
when the other variables were controlled.
However, age was negatively associated
with expenditure measures of economic
status and positively associated with
financial asset measures of economic
status.
Being married was positively associated
with all but the per capita measures of
economic status. This result is reasonable,
25
Table 3. Multivariate regression: Measures of economic status1
·
2
Total Per capita Equivalent Total Per capita
annual annual annual annual annual
household household household household household
income income income expenditure expenditure
Intercept 14721.00 10175.00 11961.00 10521.00 7899.97
Coefficients
Age of reference person (omitted: 65-74 years)
75-84 years -1257.92 -760.93
85+ years
Education of reference person (omitted: elementary school or less)
Some high school 2145.07 1132.21 1615.10
High school graduate 5221.65 3651.99 4606.46 2716.71 2410.33
Some college 8683.65 5514.70 7289.64 6659.47 4584.19
College graduate+ 15364.00 10323.00 13256.00 12116.00 8214.30
Male 3282.39 2481.22 2890.71 1819.36 1199.98
Married 8286.53 -2642.17 3341.77 7287.01 -2237.65
White 2359.44 2165.56 3529.52 2967.45
Homeowner 3528.59 1640.94 2607.92 2231.84 780.45
No earners -10450.00 -3271.27 -6359.07 -6077.04 -1154.47
Nonmetropolitan -3711.23 -1659.44 -2577.81 -3762.53 -1848.67
Adjusted R2 0.3063 0.1643 0.2469 0.3267 0.1569
F value 123.640 55.612 92.049 135.772 52.702
N 3334 3334 3334 3334 3334
1 Statistically significant coefficients only, p s .05.
2Detailed tables are available from the second author.
because the multivariate analysis revealed
the effect of being married, while holding
income constant. Because being married
was positively correlated with household
size, it would be negatively correlated
with a per capita measure.
Having no earners in the household was
negatively associated with the income
and expenditure measures of economic
status but did not have a statistically
significant effect on financial asset
measures of economic status.
The multivariate analysis confirmed this:
the nonmetropolitan and metropolitan
differences in economic status persisted
even after controlling for age, education,
gender, marital status, race, home owner­ship,
and presence of at least one earner
in the household. Multivariate results
showed that nonmetropolitan elderly
households had 83 to 88 percent of the
income, and spent 81 to 85 percent as
much as metropolitan elderly households
spent (table 4). Similarly to the bivariate
results presented in table I, the eight
Equivalent Total Total
annual financial financial
household assets assets
expenditure (total sample) (subsample)
8820.03 -11729.00 -13794.00
-1022.86 5980.51
-1263.44 10146.00 16018.00
2688.87 13998.00 16331.00
5808.73 17902.00 17871 .00
10517.00 33531 .00 36685.00
1520.04 7953.43 9422.05
3018.43 4289.77 7726.08
3377.94 11517.00 13605.00
1504.85 9174.80 11602.00
-3195.37
-2737.83 -8241.85
0.2457 0.1059 0.1082
91.463 33.901 25.316
3334 3334 2405
measures of economic status produced
consistent results regarding the relative
economic status of nonmetropolitan
and metropolitan elderly households.
In general, adjusting the measures for
household need reduced the magnitude
of the nonmetropolitan and metropolitan
differences, and the differences were
larger based on expenditure measures,
compared with income measures. How­ever,
the magnitude of these differences
was never greater than 5 percentage
points.
26 Family Economics and Nutrition Review
Table 4. Measures of economic status of nonmetropolitan and metropolitan
elderly households based on multivariate results
Coefficient on
nonmetropolitan Sample
Measure of economic status variable1 mean value2 Ratio3
Total annual household income -3,711 21,706 0.83
(19,010)
Per capita annual household income -1,659 13,444 0.88
(10,402)
Equivalent annual household income -2,578 17,864 0.86
(14,316)
Total annual household expenditure -3,763 19,502 0.81
(14,112)
Per capita annual household expenditure -1,849 12,410 0.85
(8,240)
Equivalent annual household expenditure -2,738 16,239 0.83
(10,872)
Total financial assets (total sample)4 -6,692 24,431 0.73
(44,908)
Total financial assets (subsample)5 -8,242 33,868 0.76
(49,762)
Note: Standard deviations are in parentheses.
1 Estimated coefficient on the nonmetropolitandichotomous variable in the regression equation for each
measure of economic status.
2Mean value of the measure of economic status for the total sample (N=3,334).
3Ratio of mean value· for nonmetropolitanhouseholds to the mean value for metropolitan households
implied by the multivariate results. The actual ratio was calculated as I + (estimated coefficient/sample
mean value).
4Sum of money in savings, checking and brokerage accounts, U.S. savings bonds, stocks, bonds, mutual
funds, and securities for the total sample of751 nonmetropolitanand 2,583 metropolitan households.
5Total financial assets for the subsample of 526 nonmetropolitanand I ,879 metropolitan households
with some positive amount of financial assets.
What about the ratios for assets? The
nonmetropolitan and metropolitan ratio
of total financial assets was 73 percent
for the total sample and 76 percent for
the subsample when differences in
demographic characteristics were
controlled. These nonmetropolitan and
metropolitan ratios were larger (indicat­ing
smaller differences) than the ratios
that did not control for differences in
demographic characteristics (table 1).
1998 Vol. 11 No.4
These results suggest that some portion
of the nonmetropolitan and metropolitan
differences in economic status is due to
differences in demographic characteristics
of the nonmetropolitan and metropolitan
elderly households. However, the result
that the measures of economic status of
nonmetropolitan elderly households are
never greater than 88 percent of the
comparable measures for metropolitan
elderly households confirms the persist­ence
of relatively lower economic status
of nonmetropolitan elderly households.
After geographic
differences in population
composition are controlled,
nonmetropolitan elderly
households still have
lower relative economic
status, but the magnitude
of the nonmetropolitan
and metropolitan differ-ences
becomes slightly
smaller.
27
Summary
Nonmetropolitan elderly households
have lower economic status, on average,
than metropolitan elderly households
have-across measures based on income,
expenditure, and financial assets. The
magnitude of the non metropolitan and
metropolitan difference in economic
status varies slightly with the specific
measure used. The bivariate results
indicate that the economic status of
nonmetropolitan elderly households is
17 to 21 percent lower than the economic
status of metropolitan elderly households,
depending on the income or expenditure
measure used.
After geographic differences in popula­tion
composition are controlled, non­metropolitan
elderly households still
have lower relative economic status, but
the magnitude of the nonmetropolitan
and metropolitan differences becomes
slightly smaller. However, the actual
magnitude of the difference is still fairly
large. Based on the multivariate results,
the economic status of nonmetropolitan
elderly households is 12 to 19 percent
lower than the economic status of
metropolitan elderly households, that is,
depending on the income or expenditure
measure used.
Implications
The explanation for the lower economic
status of nonmetropolitan elderly house­holds
does not lie completely in variation
in population composition. One plausible
explanation is that the lower economic
status of nonmetropolitan elderly house­holds
results from the more limited
"opportunity structure" in nonmetropolitan
areas. Persons living in nonmetropolitan
28
areas have poorer employment experi­ences,
resulting from both lower educa­tional
attainment and poorer employment
opportunities available in nonmetropolitan
areas (5). The lower lifetime earnings
result in lower economic status in later
life.
Economic resources are only one factor
contributing to overall well-being or
quality of life. Quality of life is influ­enced
by access to goods and services
through the marketplace and through
nonmarket production (objective factors),
as well as by subjective factors: including
emotional well-being, life satisfaction,
and support networks.
Price levels, which influence access to
goods and services through the market­place,
and nonmarket production are
likely to differ between nonmetropolitan
and metropolitan areas. If prices in non­metropolitan
areas are systematically
lower than prices are in metropolitan
areas and if nonmarket production is
greater in nonmetropolitan areas than
in metropolitan areas,4 then actual non­metropolitan
and metropolitan differences
in levels of well-being will be much
smaller than indicated by this research.
It is possible that nonmetropolitan elderly
households actually enjoy higher levels
of well-being than their metropolitan
counterparts, when differences in price
levels and nonmarket production are
considered.
Subjective factors are more difficult to
measure than income or expenditure but
should be considered for a more compre­hensive
assessment of the overall well-
4Nonmarket production is likely higher in non­metropolitan
areas than in metropolitan areas.
Why? Because nonmetropolitanelders are more
likely than metropolitan counterparts to have
extended family structures and more highly
developed community networks for support.
being of the elderly. Previous research
documents conflicting evidence regarding
the correlation between objective and
subjective dimensions of well-being.
(For an overview of research on subjective
dimensions of well-being, see Lee and
Lassey (9)). The notion that metropolitan
elderly fare better than nonmetropolitan
elderly in objective terms and therefore
should also fare better on measures of
subjective well-being is not confirmed
in empirical research. In a study of rural
and urban elderly, the rural elderly scored
as well or better than urban elderly
scored on measures of subjective well­being
(9). Further research should explore
the causal processes of subjective well­being
and the contribution of subjective
factors to overall well-being and quality
oflife.
The overrepresentation of the elderly in
rural and nonmetropolitan areas may
suggest that elderly people perceive the
quality of life to be higher in nonmetro­politan
areas and prefer living in these
areas. People in metropolitan areas
who prefer nonmetropolitan living may
relocate to nonmetropolitan areas later
in life. However, less than 10 percent
of those aged 65 and over move to a
new house. And of those elderly people
who move, less than 10 percent leave a
metropolitan area and move to a non­metropolitan
area ( 19).
Overall well-being and quality of life
are influenced by both objective and
subjective factors. Therefore, low relative
economic status associated with non­metropolitan
residence should not be
ignored. Because of nonmetropolitan and
metropolitan differences in residential
and employment opportunities, a blanket
approach to improving economic status
will not be effective. Different problems
and needs demand different solutions.
Family Economics and Nutrition Review
Most policy aimed at improving eco­nomic
status focuses on human capital
strategies. Public policy designed to
increase human capital through increased
and better education and employment
opportunities should be effective in
improving economic status of young
people throughout their lives including
their later years. Further, improving
employment prospects of working-age
persons through job training and retrain­ing
should effectively raise the economic
status of prime-age Americans. However,
strategies to improve the economic status
of elderly Americans, and specifically
elderly Americans living in nonmetro­politan
areas, cannot rely on efforts to
increase human capital. Strategies to
improve the economic status of the
elderly today must focus on improving
the level of income transfers to persons
with low lifetime earnings and interrupted
labor force participation. Forward­looking
strategies for improving the
economic status of future groups of
elders need to focus on availability
and access to good jobs that help
individuals acquire adequate financial
resources for retirement.
1998 Vol. II No. 4
References
I. Blaylock, J.R. and Blisard, W.N. 1990. Economic Well-Being and Household
Size: Alternative Ways of Analyzing Demographic Information on Households.
Agricultural Economic Report No. 640. U.S. Department of Agriculture, Economic
Research Service.
2. Bound, J., Duncan, G.J., Laren, D.S., and Oleinick, L. 1991. Poverty dynamics in
widowhood. Journal of Gerontology 46( 3 ):S 115-S 124.
3. Choi, N.G. 1996. The never-married and divorced elderly: Comparison of economic
and health status, social support, and living arrangement. Journal of Gerontological
Social Work 26( I &2):3-25.
4. Crystal, S. and Shea, D. 1990. The economic well-being of the elderly. Review of
Income and Wealth 36(3):227-247. .
5. Duncan, C.M. and Sweet, S. 1992. Introduction: Poverty in rural America. In C.
Duncan (Ed.), Rural Poverty in America (pp. ix-xxvii). Auburn House, New York.
6. Glasgow, N. 1988. The Nonmetro Elderly: Economic and Demographic Status.
Rural Development Research Report No. 70. U.S. Department of Agriculture,
Economic Research Service.
7. Hardy, M.A. and Hazelrigg, L.E. 1993. The gender of poverty in an aging population.
Research on Aging 15( 3 ):243-278.
8. Kivett, V.R. and Schwenk, F.N. 1994. The consumer expenditures of elderly
women: Racial, marital, and ruraVurban impacts. Journal of Family and Economic
Issues 15(3):261-277.
9. Lee, G.R. and Lassey, M.L. 1980. Rural-urban differences among the elderly:
Economic, social and subjective factors. Journal of Social Issues 36(2):62-74.
10. McLaughlin, D.K. and Jensen, L. 1993. Poverty among older Americans: The
plight of nonmetropolitan elders. Journal of Gerontology 48(2):S44-S54.
29
11. Rendall, M.S. and Speare, Jr., A. 1993. Comparing economic well-being among
elderly Americans. Review of Income and Wealth 39( I): 1-21.
12. Ringen, S. 1996. Households, goods, and well-being. Review of Income and
Wealth 42(4):421-431.
13. Schwenk, F.N. 1991. Women 65 years or older: A comparison of economic
well-being by living arrangement. Family Economics Review 4(3):2-8.
14. Schwenk, F.N. 1993. Changes in the economic status of America's elderly
population during the last 50 years. F amity Economics Review 6( 1 ): 18-27.
15. Schwenk, F.N. 1994. Income and consumer expenditures of rural elders. Family
Economics Review 7(3):20-27.
16. Tickamyer, A.R. 1992. The working poor in rural labor markets: The example
of the Southeastern United States. In C. Duncan (Ed.), Rural Poverty in America
(pp. 41-61). Auburn House, New York.
17. Tickamyer, A.R. and Bokemeier, J. 1993. Alternative strategies for labor market
analyses: Multi-level models of labor market inequality. In J. Sin gel mann and F. A.
Deseran (Eds.), Inequality in Labor Market Areas (pp. 49-79). Westview Press,
Boulder, CO.
18. Torrey, B.B. and Taeuber, C.M. 1986. The importance of asset income among
the elderly. The Review of Income and Wealth 36:443-449.
19. U.S. Bureau of the Census. 1996. 65+ in the United States. Current Population
Reports, Special Studies, P23-190.
20. U.S. Bureau of Labor Statistics. 1996. Consumer Expenditure Survey: 1994
Interview Survey CD ROM/Public Use Tape Documentation.
21. Ziebarth, A.C. and Meeks, C.B. 1998. Public policy issues and financing for
rural housing. Advancing the Consumer Interest 10( I):11-19.
30 Family Economics and Nutrition Review
Vivica Kraak
Cornell University
David L. Pelletier
Cornell University
1998 Vol. ll No.4
How Marketers Reach Young
Consumers: Implications for
Nutrition Education and
H~alth Promotion Campaigns
The advertising industry aggressively seeks to understand, anticipate, and
influence the perceived needs and desires of young consumers. Because
marketers have taken an increasingly disciplined approach to market
research, they have gained a wealth of information about children and
teenagers. This paper reviews the research methods marketers use to
gain information about young consumers to design targeted marketing
campaigns. The paper provides an overview of the advertising techniques,
styles, and channels marketers use to reach children and teenage youth. It
discusses how current market research can be used in a social marketing
framework to design more effective nutrition education and health promotion
campaigns for young consumers.
ommercialism permeates
the lives of children and
teenage youth. It is generally
defined as the means of
communication that creates consumer
awareness and induces the desire for
products, thus increasing consumer
demand and commercial profit (24). The
Center for the Study of Commercialism
describes commercialism as "ubiquitous
product marketing that leads to a pre­occupation
with individual consumption
to the detriment of society" ( 16). One
top executive of an advertising firm
said, "It isn't enough to just advertise
on television ... you've got to reach kids
through the day-in school, as they're
shopping in the mall... or at the movies.
You've got to become part of the fabric
of their lives" (6).
Much research exists that assesses the
specific influence of television advertising
on children's food- and nutrition-related
decisions and behavior over the past
two decades (26). Few studies or
reviews, however, have attempted to
examine the presence of commercialism
in promotional mediums such as school
lesson plans, movies, magazines,
games, and kid's clubs.
This paper describes the research methods
and type of information gathered by
marketers for advertising campaigns
targeted to children and teenage youth.
The paper also describes the advertising
techniques, styles, and channels marketers
use to reach young consumers. Then the
paper discusses how current knowledge
of market research methods, marketing
31
strategies, and techniques can be used­within
a social marketing framework­to
design more effective health promotion
and nutrition education campaigns that
encourage healthful eating habits among
children and teenage youth.
How Marketers Reach
Children and Teenage Youth
Marketers are extremely interested in
children as consumers because children
themselves spend billions of dollars
annually, influence household purchases,
and are future adult consumers ( 33 ). A
lifetime customer may be worth $100,000
to a retailer (23 ). Hence, the advertising
industry aggressively pursues efforts to
understand and anticipate the needs and
desires of young consumers (23 ). With
more sophisticated market research tech­niques,
marketers have gained a wealth
of information about children and teen­agers.
A review of the research methods
marketers use provides insight into the
type of information they seek: informa­tion
that allows them to design marketing
strategies for young consumers.
To obtain opinions, feedback, and insights
from children and teenage youth, market
researchers use various research methods.
Some are focus groups, written or tele­phone
surveys, individual or group inter­views,
picture drawing, story-telling,
secret ballot, and observational field
studies. Manufacturers and retailers
will often contract with independent
market research firms that have exten­sive
experience working with children
and teenagers. These manufacturers and
retailers design engaging advertising
campaigns to sell products or services
to this lucrative market with the goal of
increasing their market share. A 1990
market survey, based on the responses of
49 corporations and advertising agencies
32
that market children's products, revealed
that $16.1 million was spent on children's
research. This research was related to
product, concept, commercial tests,
audience segmentation, programming,
packaging, promotions, print advertise­ments,
brand name identification, and
pricing (12).
According to the marketing literature,
four essential elements help marketers
reach children: First, marketers keep
their efforts child-focused. Second, they
ask children the right questions and
select appropriate outcome measures
(e.g., product recognition, attention level
or in-store behavior, likability rating,
verbal recall, and conventional indicators
of product preference). Third, marketers
keep corporate attention focused on
children's needs (using seminars, quali­tative
interviews, and periodic testing of
products and communication strategies).
Fourth, marketers complement intuition
with theory when designing their research
(15).
Market researchers caution against using
standard research methods that are used
with adults when children are studied.
Adult marketers may understand adult
consumers intuitively, but they tend to
read adult meanings into what children
say ( 1 5). Using conventional focus groups
with children, for instance, can lead to
"follow-the-leader" group dynamics.
The result: Inadequate data, misleading
interpretations, unhappy clients, and
dissatisfied customers (30).
Experienced focus group moderators
believe that overcoming the effects of
peer pressure is a challenge. One way
to reduce the influence of peer pressure
is to ensure that the children in the
group are unacquainted with each
other. Moderators suggest keeping
focus group members within a 2-year
age span, because younger children
may be intimidated by older youth.
Moderators also suggest separating
boys and girls: girls tend to answer
more frequently when genders are
mixed ( 31 ). It is also recommended
that an adult moderator be replaced
with a trained youth peer to obtain
more reliable information. Another
recommendation: collect information
in familiar surroundings such as in
schools,.at summer camps, or at sporting
events (30).
Market researchers believe that surveys
must be engaging and user-friendly.
For example, the 1991 Simmons Kids
Study, the first syndicated multimedia
survey of children, researched the direct
purchase and purchase influence habits
of children ages 6 to 14 (see box). It
used a "through-the-book" magazine
method, a television diary, and a product
questionnaire ( 4 ). Marketers also use
written and visual scales, the latter
designed especially for children with
limited verbal skills. Smile and star
scales are the most common types of
visual cues market researchers use.
However, market researchers also use
card sorts and cartoon figures to determine
product appeal, purchase influence and
purchase interest, and product appropri­ateness
based on children's age and
gender ( 12).
By having children draw pictures, market
researchers have learned a great deal
about how children perceive the shopping
experience. This technique, like story­telling
and secret ballot, is especially
useful for children who may not express
themselves well verbally (22). Observa­tional
field studies are particularly in­strumental
in helping market researchers
study parent-child interactions in stores.
Market researchers operate from the
premise that the purchasing process
Family Economics and Nutrition Review
Definitions of Key Marketing Terms
Direct purchase habits: those habits related to the purchase of goods and services
that (:hildren or teenagers make for themselves.
Purchase influence habits: the array of habits related to a child's or teenager's
influence on family purchases. This includes toys and clothes; housing items,
televisions, and stereo equipment; and family items, vacations, and food.
Through-the-book magazine method: a research strategy that uses a magazine fom1at
to obtain personal information from children regarding product identification and
preferences.
Secret ballot: a research method that asks children to make a choice and then
write it down or whisper it to the researcher to keep it confidential.
Advertorial: a technique used by marketers to encourage children to read a magazine
advertisement. Marketers make the advertisement look like a game, puzzle, advice
column, or comic strip.
Product placement: the placement of brand name products in movies to deliver
promotional messages to viewers.
tends to be more impulsive than planned.
They have found that observational field
studies give a more accurate picture of
what influences children's consumer
behavior than will verbal interviews
with children, parents, or both in the
marketplace. Market researchers have
used observational field studies to deter­mine
which factors influence in-store
decisions. Their intent was to develop
new marketing strategies targeted to
families with children of various ages
(28).
Advertising Styles,
Techniques, and Channels
Successful marketing is based on cor­rectly
representing customer lifestyles
and making products relevant to their
lives. A range of advertising styles,
techniques, and channels are used to
reach children and youth to foster brand
1998 Vol. 11 No.4
loyalty and encourage product use. Some
approaches are market segmentation;
television advertising; sales promotions
at schools, stores, and sporting events;
multimedia exposure; celebrity endorse­ment;
kid's clubs; product placement;
and advertorials. Also, retailers, manu­facturers,
wholesalers, the media, schools,
and corporate donors are creating mutu­ally
beneficial partnerships to gain access
to, and capture the attention of, young
consumers. One of their goals is to
develop a market for tomorrow's adult
consumers.
Market Segmentation
The basic premise of market segmenta­tion
is that different groups of consumers
have diverse attitudes, interests, and
behaviors. And, by acknowledging these
differences, marketers believe they can
increase their chances of influencing
consumers' behaviors. Segmentation
involves describing the potential market's
physical, behavioral, demographic,
psychographic, and geographic charac­teristics
(25). Gender, age, socioeconomic
status, and ethnicity are four ways that
advertisers segment the youth market.
Although marketers usually segment
young consumers into three age categories
(2- to 5-year-olds, 6- to 11-year-olds,
and 12- to 17-year-olds), there is agree­ment
on two points-large gaps exist in
understanding young consumers' behavior
and the existing age categories may be
initially helpful but are arbitrary (32).
Marketers often segment age with several
other factors, such as gender and socio­economic
status. Only recently have
marketers acknowledged the importance
of ethnic minority subcultures. Marketers
tend to assume that the preferences and
consumer habits of various ethnic groups
are not significantly different among
young children, but these preferences
and habits become significant during
older childhood and adolescence when
ethnic and cultural identities are formed
( 32). The ability to understand and
depict cultural nuances and the use of
appropriate language are the two greatest
challenges faced by marketers and
educators in effectively reaching ethnic
minority groups that are distinct and
heterogeneous.
Television Advertising
Television has been identified as the
medium that provides the widest and
most frequent reach for younger children.
Children ages 2 to 11 watch an average
of 26 hours of television each week
(6,26). In a 3-hour setting, a child may
watch about 30 minutes of advertising,
totaling 20-40 advertisements each hour
depending on their length (26) and
may be exposed to as many as 22,000-
25,000 commercials each year ( 13 ).
Television commercials use attention-
33
34
Although in-school
multimedia can be viewed
as a useful way to educate
children and teenage
youth, it has been
increasingly criticized
as a form of
"commercialization of
the classroom" when
provided by corporations
in exchange for advertising
promotions and test
marketing within
educational
environments.
getting techniques such as attractive
models and familiar songs and jingles;
they provide easily stored and recalled
images from memory; they motivate
children to retain information by high­lighting
the relevant, desired behavior;
and they are highly repetitious (29).
Advertisers are now looking beyond
children's programs to reach the larger
audience of children who are watching
prime-time television or listening to the
radio with their parents because it is an
opportunity to reinforce the connection
between children's independent purchases
and their influence on family purchases.
Marketers who want to focus on children's
personal spending choose media that
deliver messages to a large number of
children in their desired target group.
Marketers who want to take advantage
of young people's power to influence
family purchases choose commercials
or television programs that reach children
or teenage youth together with their
parents (32).
Sales Promotion
Sales promotion is a commonly used
method for reaching young consumers
in places where they are often found. The
objectives and strategies marketers use
need to be well-defined to capture the
attention and interest of the desired
target audience. For example, sales
promotions occur at rock concerts,
beaches, malls, and sports events; in
stores; and even at school. They are used
to motivate children and teens to make
purchases at places they or their parents
regularly shop, such as cosmetic counters,
convenience stores, supermarkets, and
fast-food restaurants. Premiums and
sweepstakes prizes are often distributed
to appeal to children's and teens' tastes
and desires (27).
Manufacturers, wholesalers, retailers,
the media, and corporate donors fre­quently
engage in cross-selling, the
practice of combining promotional
efforts to sell a concept, product, or
service. Disney, for instance, has launched
cross-selling campaigns worth millions
of dollars to promote its films and char­acters
in exchange for the sale or place­ment
of other companies' products into
Disney films (7). Disney has marketing
agreements with several companies,
including Coca-Cola, Proctor and
Gamble, Kraft, and McDonald's.
Nationally, McDonald's produces and
delivers more than 200 different advertise­ments
annually. This fast-food chain
spends about $740 million in advertising,
has earned an internationally recognized
name that is synonymous with fast-food,
and has built a reputation as "the children's
marketer" ( 17). The company uses a
multifaceted sales promotion approach
to reach ethnic youth by using radio and
cable television to deliver messages to
African Americans and Hispanics, and it
uses network television to air commercials
to the general population. McDonald's
strives to make parents feel good about
taking the family to the restaurant chain:
both mothers and children surveyed put
McDonald's at the top of their list for
likability. McMoms, a program that
targets bilingual mothers of children
ages 2 to 7, inserts bilingual response
cards into its "Happy Meal" boxes.
In return, mothers receive Spanish
language newsletters and promotions.
Sports, youth, and community angles
are used in McMoms' promotional
advertising, which also includes scratch
card contests, games on the place mats,
and toy car give-a-ways ( 17).
Family Economics and Nutrition Review
Multimedia Exposure
Using television commercials to reach
children and youth is rapidly becoming
more expensive and less efficient. Chil­dren
are increasingly being exposed to
different types of mass media, including
radio, magazines and newspaper sections
written especially for them, and inter­active
computer technology (21 ).
Although in-school multimedia can
be viewed as a useful way to educate
children and teenage youth, it has been
increasingly criticized as a form of
"commercialization of the classroom"
when provided by corporations in
exchange for advertising promotions
and test marketing within educational
environments. Because of chronic
funding shortages, school districts have
welcomed advertisers to underwrite the
cost of educational materials, equipment,
and services. Thus, school districts have
been viewed as silent partners in adver­tising
to children (7,26).
Celebrity Endorsement
Heroes, heroines, and role models can
motivate children and teenage youth to
buy products and services. The celebrities
most admired by children are entertainers
or athletes. McDonald's and Pepsi, for
example, have used Michael Jordon and
Michael Jackson, respectively, to endorse
food and beverage products targeted to
children and teenage youth (27). Celeb­rity
endorsements encourage children
to buy products for their status appeal.
The status products being marketed are
costly, and celebrity commercials are
becoming increasingly slick. Today's
children are contending not only with
the celebrity appeal in television and
magazine advertisements (7,26) but also
with peer pressure from friends who
see the same commercials. Children
must also face the financial realities of
wanting products that they do not need
and/or their parents cannot afford (7).
1998 Vol. 11 No.4
Kid's Clubs
Some corporations (Nickelodeon, Fox,
Burger King, and Disney) have created
kid's clubs. A kid's club establishes an
ongoing relationship with its members
by providing membership cards and
participatory activities that are depend­ent
on spending money. Research has
suggested that kid's clubs promote con­sumerism,
reinforce commercial interests
by building brand loyalty, and provide a
convenient vehicle to deliver commercial
messages and perpetuate ongoing
advertising to children. Many of these
clubs use their enrollment databases to
distribute coupons for club merchandise
(7;26).
Product Placement and
Advertorials
Advertisers have paid between $10,000
and $1 million to display brand name
products in movies, with the price in­creasing
if an actor uses a product rather
than if the product is only shown. Some­times,
movie studios and producers
accept merchandise or promotional
support in exchange for placing a product
(7,26). For instance, Burger King was
depicted in Teenage Mutant Ninja Turtles
in exchange for promoting the movie
before its release. Products can also be
placed in prime-time television programs,
comic strips, and video games.
The opportunity to reach children and
teenage youth with print media has
expanded over the past decade. Over
160 magazines are targeted to children,
many of which carry hidden advertise­ments-
in editorials, comics, games,
and puzzles. The resulting advertorials
or hidden advertisements have been
described as "subliminal inducements"
that can mislead and deceive children
(7,26).
Other Advertising Styles
and Techniques
Marketers specializing in advertising to
children and teenage youth have learned
which advertising styles and techniques
work well with specific segmented groups
and have provided the rationale for why
they believe these styles are effective.
An executive of one marketing firm
offers these 10 tips to make children
notice messages:
1. Be aware of age differences in
the market;
2. Make sure the product or service
has a point of view and a unique
selling proposition;
3. Use child-appropriate language to
reinforce a feeling of peer-group
belonging and bonding;
4. Pay close attention to the l