CENTER FOR NUTRITION POliCY AND PROMOTION
Gons"tr uction and Evaluation of a Diet Status Index
P. Peter .Basiotls, joanne R Guthrie, Sbanthy A. Bowman. and Susan 0. Welsh
Characteristics of Rural Residents and Vulnerability to
Alcohol Problem Behaviors
Elizabeth B. Robertson
Housing Trends
Nancy E. Schwenk
Research Summaries
Gender-Related Shifts in the Distribution of Wages
Changing Eating Patterns: Grains, Vegetables,
Fruit, and Sugars
38 The Development and Growth of Employer-Provided
Health Insurance
40 Are Women Leaving the Labor Force?
Regular Items
42 Charts From Federal Data Sources
44 Recent Legislation Affecting Families
45 Data Sources .
46 Journal Abstracts and Book Summary
47 Cost of Food at Home
48 Consumer Prices
49 Guidelines for Authors
To Our Readers:
This is the frrst issue published by the Center for Nutrition Policy
and Promotion under our new name, Family Economics and
Nutrition Review. Our intent is to include articles on nutrition
policy and research as well as family economics.
We invite the submission of manuscripts in each of these areas.
Please refer to "Guidelines for Authors," which appears on
page 49. A distinguished and diverse Editorial Board reflecting
both disciplines will help select manuscripts for publication.
Joan C. Courtless, Editor
The United States Department of Agriculture (USDA) prohibits discrimination in its
programs on the basis of race, color, national origin, sex, religion, age, disability, political
beliefs, and marital or familial 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 the USDA Office of
Communications at (202) 720-5881 (voice) or (202) 720-7808 (IDD).
To file a complaint, write the Secretary of Agriculture, U.S. Department of Agriculture,
Washington, DC 20250, or call (202) 720-7327 (voice) or (202) 720-1127 (TDD). USDA
is an equal employment opportunity employer.
Editor
Joan C. Courtless
Editorial Assistant
Jane W. Fleming
Family Economics and Nutrition
Review is written and published each
quarter by the Center for Nutrition
Policy 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,
Ageline, Economic Literature Index,
ERIC, Family Resources, PAIS, and
Sociological Abstracts.
Family Economics and Nutrition
Review is for sale by the Superintendent
of Documents. Subscription price
is $7.50 per year ($9.40 for foreign
addresses). Send subscription orders
and change of address to Superintendent
of Documents, P.O. Box 371954,
Pittsburgh, PA 15250-7954. (See
subscription form on p. 51.)
Suggestions or comments concerning
this publication should be addressed
to: Joan C. Courtless, Editor, Family
Economics and Nutrition Review, Center
for Nutrition Policy and Promotion,
USDA, 1120 20th St., NW, Suite 200
North Lobby, Washington, DC 20036.
Phone (202) 606-4816.
Center for Nutrition Policy and Promotion
Feature Articles
2
14
24
Construction and Evaluation of a Diet Status Index
P. Peter Basiotis, Joanne F. Guthrie, Shanthy A. Bowman, and
Susan 0. Welsh
Characteristics of Rural Residents and Vulnerability
to Alcohol Problem Behaviors
Elizabeth B. Robertson
Housing Trends
Nancy E. Schwenk
Research Summaries
34 Gender-Related Shifts in the Distribution of Wages
36 Changing Eating Patterns: Grains, Vegetables,
Fruit, and Sugars
38 The Development and Growth of Employer-Provided
Health Insurance
40 Are Women Leaving the Labor Force?
Regular Items
42 Charts From Federal Data Sources
44
45
46
47
48
Recent Legislation Affecting Families
Data Sources
Journal Abstracts and Book Summary
Cost of Food at Home
Consumer Prices
49 Guidelines for Authors
Volume 8, Number 2
1995
2
Feature Articles
Construction and Evaluation
of a Diet Status Index
By P. Peter Basi otis
Economist
Center for Nutrition Policy and Promotion
Joanne F. Guthrie
Nutritionist
Center for Nutrition Policy and Promotion
Shanthy A. Bowman
Nutritionist
Center for Nutrition Policy and Promotion
Susan 0. Welsh
Nutritionist
Cooperative State Research, Education, and Extension Service
Because there is a well-established link between diet and health, it is important
to be able to evaluate individuals' or groups' diets and to obtain information
on their diets' determinants. However, because diets are complex, it is
difficult to devise a summary measure of the overall diet to be used as an
outcome-or dependent-variable in univariate or multivariate studies of
diets' determinants. Current dietary guidance emphasizes both eating enough
food to meet the body's needs and avoiding excess intakes of certain food
components that have been linked to chronic diseases. Using USDA's
1989-91 Continuing Survey of Food Intakes by Individuals, this study
incorporates these two aspects of current dietary guidance to explore the
feasibility of constructing a Diet Status Index, a summary measure of the
overall diet. This exploratory index consists of two subindices: a Dietary
Adequacy Score and a Dietary Moderation Score. Conclusions from univariate
analyses by several respondent characteristics using the three measures
were similar with those from previous studies of diets' determinants. Based
on these results, it appears that the index and subindices have promise for
use as summary measures by researchers studying diet quality.
eople's diets are complex.
Every day in the United
States, individuals chooseand
consume-from a staggering
array of foods available to them.
The USDA Survey Nutrient Data Bank
currently contains food composition
data on almost 7,400 different foods.
For each food there is information on
about 30 nutrients and food components.
At the same time, the importance of diet
in maintaining good health is quite clear
(15). A good, or healthy, diet can help
people live longer and healthier lives,
with enhanced well-being. It also means
better economic productivity and lower
health care costs. Thus, the Federal
Government has a strong incentive to
monitor the population's diets (typically
through national food consumption
surveys) and, when necessary, to help
Family Economics and Nutrition Review
improve dietary status through nutrition
education and other efforts.
This raises two interesting questions
for nutrition educators and others concerned
with assessing diets and their
determinants. First, what exactly is a
good or healthy diet? And second, how
can a person's or a group's diet be
evaluated?
To answer the first question, since 1980
the U.S. Department of Agriculture
(USDA) and the U.S. Department of
Health and Human Services (DHHS)
have issued principles of a healthful
diet called the "Dietary Guidelines for
Americans" (14). These Guidelines
focus on obtaining a diet both sufficient
in nutrients and without excesses, since
excess intakes of certain food components
have been linked to chronic
diseases {15). The current Guidelines
are:
• Eat a variety of foods
• Maintain healthy weight
• Choose a diet low in fat, saturated
fat, and cholesterol
• Choose a diet with plenty of
vegetables, fruits, and grain
products
• Use sugars only in moderation
• Use salt and sodium only in
moderation
• If you drink alcoholic beverages,
do so in moderation
The Dietary Guidelines do not give
specific and detailed recommendations
on which foods to eat every day and
how much. This is done by the USDA/
DHHS Food Guide Pyramid (FGP) (16),
shown in figure 1. To come up with
these specific recommendations, USDA
scientists considered, among other
things, the number of servings per day
from major food groups and subgroups
that would ensure three dietary goals
would be met {17). These goals were
variety, proportionality, and moderation.
1995 Vol. 8 No.2
Figure 1. The Food Guide Pyramid: A guide to daily choices
Fats, Oils, & Sweets
USE SPARINGLY
Milk, Yogurt,
&Cheese
Group
2·3 SERVINGS
Vegetable
Group
3·5 SERVINGS
Meat, Poultry, Fish,
Dry Beans, Eggs,
& Nuts Group
2·3 SERVINGS
Fru~
Group
2-4 SERVINGS
Bread, Cereal,
Rice, & Pasta
Group
6·11
SERVINGS
Source: U.S. Department of Agriculture and U.S. Department of Health and Human SeNices.
Variety means eating a selection of
foods of various types that together
meet nutritional needs. Proportionality
means eating appropriate amounts of
various types of foods to meet nutritional
needs. And moderation means
avoiding too much of food components
in the total diet that have been linked
to diseases {17). For practical purposes,
variety and proportionality can be
combined to reflect dietary adequacy
(fig. 2, p. 4). In this context, adequacy
means eating the quantity and quality
of food that will satisfy the biological
needs of healthy people, so that the
recommended amounts of food energy
(calories), vitamins, minerals, and other
food components are consumed. The
FGP shows these ideas by suggesting
that people eat from all the food groups
but proportionately more of the food
groups at its base than at its top.
The answer to the second questionhow
can a person's or group's diet be
evaluated?-had to await the answer
to the first-what is a good or healthy
diet? Because a diet comprises many
components, it is difficult to judge one
overall. Still, those wanting to relate
overall diets to factors influencing
those diets need a summary measure,
or index, of the overall diet. Otherwise,
researchers are required to consider
multiple aspects of the diet, one at a time,
greatly complicating inferences on
determinants of the overall diet quality.
In this study, we propose a summary
measure for assessing the overall quality
of diets that can be used for identification
of factors influencing overall diets
and for prediction of dietary status of
groups of individuals sharing common
characteristics. This measure, the Diet
Status Index (DSI), builds and improves
upon previous measures of the overall
diet in two important ways: (1) it incorporates
aspects of both dietary adequacy
and dietary moderation, and (2) it is
a relatively simple measure, easy to
calculate from data readily available
in dietary survey data sets.
3
Figure 2. Conceptual framework for construction of diet status Index
Food Guide Pyramid
I PROTEIN 1-
I VITAMINA 1- l THIAMIN 1- I ADEQUACY l I MODERATION
I RIBOFLAVIN t-
~ H BREAD, CEREAL, I NIACIN t- RICE, PASTA GROUP 1 1
I VITAMINB8 f- I I I SATURATED H ~ TOTAL FAT
I VITAMIN B12 1- FAT VEGETABLE
GROUP I FOLATE f-
I VITAMINC t-I VITAMIN E 1-H ~ FRUIT GROUP
I IRON t-I ZINC 1-H ~ MILK, YOGURT
GROUP
I CALCIUM t-
I PHOSPHORUS 1-H MEAT. POULTRY, ~ FISH,
ALTERNATES
I MAGNESIUM f- GROUP
Previous Studies
A number of dietary summary measures
have been proposed in the past. Two of
the most popular ones have been the
Mean Adequacy Ratio (MAR) (1), and
the Index of Nutritional Quality (INQ)
(13). Recently the Diet Quality Index
(DQI) (1 1) was proposed.
The MAR is calculated by adding the
intakes of a number of nutrients of
interest expressed as a percentage of
the Recommended Dietary Allowances
(RDA)1 (9) for an individual, but truncated
at 100 percent, and then dividing
by the number of nutrients to get the
value of the index. A clear limitation
of the MAR is that it only addresses
the adequacy aspect of the diet.
1The Recommended Dietary Allowances are
average daily intakes which, if met, will provide
for individual variations in the nutritional needs of
practically all healthy persons living in the United
States under normal environmental stresses.
4
The INQ is calculated by dividing the
diet's caloric content by the food energy
intake recommended for the person's
age and gender. Then, the intakes of
a number of nutrients expressed as a
percentage of the RDA are divided by
this ratio and an MAR-type of index
is constructed based on these imputed
intakes. Because it is based on the nutrient
density of the diet, the INQ evaluates
the nutritional quality, not quantity,
of the diet. In a sense, the INQ does not
capture completely either the adequacy
or the moderation aspect of the overall
diet. Because people seem to underreport
their food intakes during dietary
survey interviews (7), the INQ may,
however, be used to attempt to capture
dietary adequacy while adjusting for
possible underreporting.
The recently proposed DQI captures
aspects of both adequacy and moderation
in the diet. It is based on eight
National Academy of Sciences (NAS)
recommendations (10). The NAS
recommendations are similar but not
l
1 1
CHOLESTEROL ll SODIUM I
identical to the Dietary Guidelines. In
addition, because of its relatively few
components, the DQI may not reflect
the quality of the total diet as well as
possible, given the additional information
available in the data sets (8). For
practical purposes, the DQI is hard to
calculate because it requires foods
consumed to be grouped appropriately,
a time-consuming task requiring considerable
decision-making by nutrition
experts. Variations in the food-grouping
processes may reduce comparability of
results across studies.
Sources of Data
Data used in this study are from USDA's
Continuing Survey of Food Intakes by
Individuals (CSFII), 1989-91. The sample
used for construction and evaluation of
the DSI consisted of adult (age 20 years
or over) males and females who were
not bedridden and females who were
not pregnant or breastfeeding.
Family Economics and Nutrition Review
The CSFII 1989-91 yielded 3 years of
data, with each yearly survey consisting
of approximately 1,500 all-income
households and a low-income sample
of approximately 750 households. Data
from the samples were combined using
the appropriate survey weights that were
supplied by the USDA. In the 1989-91
CSFII, USDA collected information
from individual household members
about what they eat and how much
(through one 24-hour recall followed
by a 2-day record), how they prepare
food, the time and name of each eating
occasion, and the source of the food.
The amounts of nutrients and other food
components from these food intakes,
calculated using the USDA Survey
Nutrient Data Bank, were provided by
the USDA. These amounts exclude
vitamin, mineral, and other supplements.
In addition, economic and sociodemographic
data and self-reported data on
the respondent's diet and health were
collected.
To ensure comparability across past
and future surveys, only intake data
obtained during the first 24-hour recall
were used for this study. All samples
were weighted to be representative of
the respective age-sex groups of the
U.S. population.
Methodology
Construction of the Diet Status Index
was based on the conceptual framework
shown in figure 2. This framework
reflects the advice of the FGP in the
context of the available data. Because
data on consumption of FGP food groups
are not readily available, nutrient and
food component data already available
in the data sets were used. Appropriate
levels of consumption of these nutrients
and food components result by following
the dietary recommendations of the
FGP (fig. 2) (17). Note, however, that
the converse is not necessarily true.
Adequacy is reflected in the daily consumption
of 15 nutrients for which the
RDA have been established and are also
1995 Vol. 8 No.2
available in the USDA Nutrient Data
Bank. These are: Protein, Vitamin A (in
retinol equivalents), Thiamin, Riboflavin,
Niacin, Vitamin B-6, Vitamin B-12,
Folate, Vitamin C, Vitamin E, Iron, Zinc,
Calcium, Phosphorus, and Magnesium.
Nutrients and dietary components such
as selenium and fiber were not included
because either data were not available,
or no established standards for consumption
exist. Since the RDA take into
account differences in requirements
due to age and sex, use of the RDA
allows interpersonal and interstudy
comparisons. Although there may be
some redundancy in using alliS RDA
nutrients available, statistical reliability
of the index is increased ( 4).
To construct the Dietary Adequacy
Score (DAS) component of the DSI,
the following steps were followed:
1. For each of 15 nutrient intakes, the
individual was assigned a 1 if the
intake was equal to or exceeded the
individual's RDA for the nutrient,
or assigned a 0, otherwise.
2. The intake scores were added up for
a possible minimum score of 0 and
a maximum score of 15.
3. The sum was then multiplied by
6-2/3 to adjust it to a scale from 0 to
100. This was the individual's DAS.
To construct the moderation component,
the following food components
were used (fig. 2):
1. Percent of food energy from total fat
(limit to 30 percent or less)
2. Percent of food energy from saturated
fat (limit to 10 percent or less)
3. Amount of dietary cholesterol (limit
to 300 mg or less)
4. Amount of sodium (limit to 2,400 mg
or less)
A good, or healthy,
diet can help people
live longer and
healthier lives, with
enhanced well-being.
It also means better
economic productivity
and lower health care
costs. Thus, the
Federal Government
has a strong incentive
to monitor the
population's diets ...
5
To construct the Dietary Moderation
Score (DMS) component of the DSI,
the following steps were followed:
I. For each of the four moderation food
components above, the individual
was assigned a I if the intake was
less than the recommended amount
for that component (30 percent for
food energy from total fat; I 0 percent
of food energy from saturated fat;
300 mg for dietary cholesterol; and
2,400 mg for sodium, and if the
person said that he or she rarely or
never adds salt to the food at the
table), or assigned a 0, otherwise.
2. The scores were added up for a
possible minimum score of 0 and a
maximum score of 4.
3. The sum was then multiplied by 25
to adjust it to a scale from 0 to IOO.
This was the individual's DMS.
The individual's DSI was then constructed
by adding the DAS and the DMS and
dividing by 2.
Because of the possibility of underreporting
of food intakes during survey
interviews (7), energy-adjusted scores
were also calculated. This was done by
dividing each nutrient or food component
above by the ratio of the individual's
food energy intake to his or her recommended
energy allowance (REA) and
then calculating the index scores as
above. However, this approach is valid
only if there is no systematic, or selective,
underreporting of specific foods.
Statistical tests were conducted using
the statistical program SUDAAN which
is designed for use with large complex
surveys (12). Statistical significance
was at the (p~. 05) level. T tests were
used for comparing means of two
groups, and multiple contrasts were
used for simultaneously comparing
means of three groups (12).
6
Selected Limitations
An obvious limitation of the DSI is that
the status of obesity is not reflected in
the three summary measures, although
it is a major diet-related health concern
in the United States. Including a measure
of obesity in the index construction
would increase complexity at this initial,
exploratory stage, and would complicate
the measures' evaluation. Use of IOO
percent of the RDA or the recommended
limit specified in constructing the DAS
and DMS is especially problematic,
since it assigns consumption of, say, 99
percent of the RDA for some nutrient to
the "inadequate" category. However,
even though 67 percent of the RDA has
been used as a cutoff point in the literature
(11), it was decided to use IOO percent
of the RDA because such intakes
would clearly meet the adequacy criteria.
Table 1. Estimated mean intakes of dietary components, U.S. adults, by
sex: 1 1989-91
Food component2 All Males Females
Sample size 10,088 4,169 5,919
Food energy3 78.0 82.2 74.2
Adequacy components3
Protein 136.6 I46.3 I27.6
Vitamin A 113.0 II1.8 114.2
Vitamin E 90.7 94.4 87.3
Vitamin C 159.4 173.0 146.8
Thiamin 122.8 127.0 119.0
Riboflavin 124.7 131.0 119.0
Niacin 135.9 I45.9 126.6
Vitamin B-6 92.9 97.7 88.5
Folate 132.4 143.6 I22.1
Vitamin B-12 250.3 309.7 195.6
Calcium 87.8 101.2 75.5
Phosphorus 143.1 167.8 120.4
Magnesium 82.4 85.3 79.7
Iron 127.8 164.7 93.9
Zinc 81.8 89.1 75.0
Moderation components
% Calories from fat 34.5 35.1 34.0
% Calories from saturated fat 11.9 12.2 11.7
Cholesterol (mg) 285.5 345.1 230.6
Sodium (mg) 3,163 3,903 2,482
1Estimates are based on USDA's 1989-91 Continuing Survey of Food Intakes by Individuals and are
weighted to represent the U.S. adult population living in households.
2Estimates exclude vitamin, mineral, and other supplements.
3Percentages of Recommended Energy Allowances (REA) and Recommended Dietary Allowances
(RDA) for protein through zinc.
Family Economics and Nutrition Review
The moderation food component cutoffs
were decided through similar reasoning.
Assigning equal weights, and thus
importance, to each nutrient or food
component in the index scores is also
problematic. However, in the absence
of comprehensive scientific guidance on
the subject, weighting each nutrient and
food component equally was deemed
the least problematic alternative. Since
the DSL DAS, and DMS were constructed
for use as outcome--or dependentvariables
in multivariate analyses or
for use with groups of individuals in
univariate work, the statistically undesirable
effects of large day-to-<lay variability
in nutrient consumption should be alleviated
by: The use of relatively many
nutrients for the DAS (4); the use of
cutoff points for the nutrients and food
components (5); and the large number
of individuals in the survey (3).
Table 2. Estimated mean index scores: Dietary status, adequacy, and
moderation, U.S. adults, by sex and age:1 1989-91
Dietary Dietary Dietary
Sample status adequacy moderation
Sex/Age size index2 score2 score2
Males (years)
20 to 50 2,618 42.9A 31.3A
48.7A 22.7A
Over 50 1,551 46.7 38.1
51.4 31.6
All 4,169 44.1s 33.5s
49.65 25.68
Females (years)
20 to 50 3,466
Over 50 2,453
All 5,919
All
20 to 50 6,084
Over 50 4,004
1.2
All 10,088 49.3 41.2
70.8 31.9
1Estimates are based on USDA's 1989-91 Continuing Survey of Food Intakes by Individuals and are
weighted to represent the U.S. adult population. See text for definitions of Dietary Status, Adequacy,
and Moderation.~--
"A" -Significantly different from the other age category at the ~.05 level.
"S" -Significantly different from the other sex category at the ~.05 level.
1995 Vol. 8 No.2
Evaluation
Evaluation of the DSI and its component
subindices relies on criterion validity
(4). That is, selected results obtained
by use of the DSI, DAS, and DMS are
compared with those of other studies that
draw inferences on factors influencing
the total diet.
Average intakes of food energy, the 15
adequacy nutrients, and the 4 moderation
food components are shown in table 1.
Average food energy intakes were substantially
below the Recommended
Energy Allowances (REA) for both
men and women. Average intakes for
11 nutrients for men and 9 for women
were above the RDA, whereas intakes
of 4 nutrients for men and 6 for women
were below. Of the four moderation
food components, only average cholesterol
intake by women met the moderation
criterion.
Sex and Age
The DSI, DAS, and DMS for the period
1989-91 are shown in table 2. In each
cell, the top number represents the
actual score, and the bottom number
represents the energy-adjusted score.
For all adult persons the DSI was 45.2.
The DAS, at 49.3, was higher than the
DMS (41.2). The energy-adjusted scores
were 51.3, 70.8, and 31.9, respectively.
Predictably, adjusting for energy intake
tends to increase the DAS and decrease
the DMS. For the sake of simplicity,
subsequent discussion will refer only
to the actual scores, unless there is
disagreement (in terms of statistical
significance) between the actual score
and the energy-adjusted score. Adult
women had a higher level of the DSI
than adult men, 46.3 and 44.1, respectively.
However, men had a much
higher mean DAS (54.7) than women
(44.3). This situation was reversed for
the DMS, where women had the higher
score, 48.3 compared with 33.5 for
men. These differences in the dietary
scores may reflect lower average food
intakes by women compared with those
by men (table 1).
7
8
... the "Dietary
Guidelines for
Americans" ... focus
on obtaining a diet
both sufficient in
nutrients and
without excesses,
since excess
intakes of certain
food components
have been linked
to chronic diseases.
Table 3. Estimated mean index scores: Dietary status, adequacy, and
moderation, U.S. adults, by sex and household income: 1 1989-91
Sex/
Income as percent of
poverty threshold
Males
At or below 130%
Over 130%
Females
At or below 130%
Over 130%
All
At or below 130%
Over 130%
Sample
size
1,374
2,795
2,565
3,354
3,939
6,149
Dietary
status
index2
44.5
46.8
Dietary
adequacy
score2
55.7
45.5
Dietary
moderation
score2
33.4
48.1
1Estimates are based on USDA's 1989-91 Continuing Survey of Food Intakes by Individuals and are
weighted to represent the U.S. adult population. See text for definitions of Dietary Status, Adequacy,
and Moderation.
~ ......
"I" -Significantly different from the other income category at the ~.05 level.
With the exception of the DAS for
males, older people (those over 50 years
of age) had substantially higher scores
for all three dietary measures (table 2).
The differences were more pronounced
for the DMS for both sexes, where
people over 50 years old had a DMS
of about eight points higher than those
20 to 50 years of age.
Income and Education
Averages of the three index scores by
income category are shown in table 3.
As expected (6), people from households
with incomes over 130 percent of the
Federal poverty threshold had higher
average scores for the DSI and the DAS
component than those from households
with incomes at or below 130 percent of
the poverty threshold. The DMS is very
similar within sex groups across income
levels. The statistically higher DMS for
all adults at or below 130 percent of
poverty may be an artifact of the significantly
greater percentage of females in
this low income group. Based on previous
studies, the evidence on the relationship
between income and diet quality is
conflicting (6).
Higher education levels were associated
with higher average levels for all three
dietary status measures (table 4). This
was true for both sexes. People with a
more-than-high-school education
averaged about four points higher on
Family Economics and Nutrition Review
Table 4. Estimated mean index scores: Dietary status, adequacy, and
moderation, U.S. adults, by sex and education level:1 1989-91
Males
Sex/
Education
Sample
size
Up to high school 2,662
More than high school 1,452
Females
Up to high school 4,088
More than high school 1,758
All
Up to high school 6,750
More than high school 3,210
Dietary
status
index2
48.7
Dietary
adequacy
score2
47.0
Dietary
moderation
score2
50.3
1Estimates are based on USDA's 1989-91 Continuing Survey of Food Intakes by Individuals and are
weighted to represent the U.S. adult population. See text for definitions of Dietary Status, Adequacy,
and Moderation ..
2Baltomaumller it~
"E" -Significantly different from the other education category at the pS.05 level.
the DSI than people with an up-to-highschool
education. The largest differences
were observed for the DAS. These
results are in general agreement with
those of previous studies (6).
Race
In general, people in the "Other" race
category of table 5, p. 10, had the best
scores for all dietary measures. The
Other category includes Asian/Pacific
Islander, Aleut, Eskimo, and American
Indian, and all other race classifications.
For almost all dietary measures, Whites
had higher scores than Blacks, a finding
in agreement with previous studies (6).
This result seems to reflect statistical
differences between the scores of
White and Black females.
1995 Vol. 8 No.2
Self Report and Self-Perception
of Diet Status
Persons who reported being on a special
diet (such as weight loss, low sodium,
diabetic, etc.) had substantially higher
averages for the three dietary measures
(table 6, p. 11). The exception was for
the males' actual DAS. The differences
were most pronounced for the DMS
where those on a special diet had higher
DMS by 9-12 points. Interestingly,
those on a special diet also tended to
have higher DAS than those not on a
special diet.
As might be expected from previous
research (2), there appeared to be a
direct association between self-reported
healthfulness of diet and dietary quality
... we propose a
summary measure
for assessing the
overall quality of
diets .... This measure,
the Diet Status Index
(DSI) .. .incorporates
aspects of both dietary
adequacy and dietary
moderation and is
a relatively simple
measure, easy to
calculate from data
readily available in
dietary survey data
sets.
9
as reflected by the three dietary measures,
especially for males (table 7, p. 12). For
females, those reporting that their diet
was excellent or very good had higher
DSI, DAS, and DMS averages than
those who rated their diet as good.
They, in turn, had better scores than
those who rated their diets fair or poor.
The pattern was similar for males, with
the exception of the DMS.
Summary and Conclusion
This study incorporated two aspects of
current dietary guidance to construct
the Diet Status Index, an exploratory
summary measure of the overall diet.
This index consists of two component
subindices, a Dietary Adequacy Score
and a Dietary Moderation Score. The
DAS is based on intakes of 15 nutrients,
expressed as percentages of the Recommended
Dietary Allowances. Because
of standards and data availability, the
DMS is based on intakes of only four
dietary components that should be
limited. These were: Percentage of
calories from fat, percentage of calories
from saturated fat, cholesterol, and
sodium. In an attempt to examine
effects of possible underreporting of
food during dietary surveys, the energyadjusted
DSI, DAS, and DMS were
also calculated and presented.
Several limitations were noted. Absence
of an obesity measure was a major limitation.
Future studies could incorporate
an obesity measure such as the Body
Mass Index (BMI) in the construction
of a DSI. Weighting all nutrients and
food components equally in the construction
of the DAS and DMS was
another major limitation. Refinement
of a nutrient-based DSI will require
scientific effort to determine nutrients'
relative importance in the diet, once the
minimum nutrient requirements have
been met. In this vein, scientific research
to determine the distribution of these
requirements in the population is also
needed. Although day-to-day variability
in food intake should not be a problem
10
Table 5. Estimated mean index scores: Dietary status, adequacy, and
moderation, U.S. adults, by sex and race: 1 1989-91
Males
White
Black
Other
Females
White
Black
Other
All
White
Black
Other
Sex/Race
Sample
size
3,570
400
199
4,738
869
312
8,308
1,269
511
Dietary
status
index2
47.2
Dietary
adequacy
score2
44.9
Dietary
moderation
score2
49.5
1Estimates are based on USDA's 1989-91 Continuing Survey of Food Intakes by Individuals and are
weighted to represent the U.S. adult population. See text for defmitions of Dietary Status, Adequacy,
and Moderation.
"0" -Significantly different from the "Other" category at the ~.05 level.
"B"- Significantly different from the "Black" category of the ~.OS level.
for summary measures of groups' diets
like the DSI, DAS, and DMS, it limits
use of such measures for assessing
individuals' diets. Research is needed
to improve on estimation accuracy of
individuals' usual intakes. Energyadjusting
nutrient and food component
intakes might compensate for possible
food underreporting in surveys, if underreporting
is random with respect to
individual foods. If, however, there is
selective underreporting, energy-adjusting
may lead to erroneous conclusions.
More research is clearly needed
to adequately assess the entire issue
of underreporting in dietary surveys.
The results show that, in general, those
who had better DSI scores were: Over
50 years of age, had household incomes
above 130 percent of the Federal poverty
threshold, and had more than a high
school education. Those who reported
being on a special diet and those who
Family Economics and Nutrition Review
Table 6. Estimated mean index scores: Dietary status, adequacy, and
moderation, U.S. adults, by sex and self-reported diet situation: 1 1989-91
Males
Sex/
Self-reported
diet situation
Sample
size
On special diet 441
Not on special diet 3,694
Females
On special diet 1,046
Not on special diet 4,831
AU ·
On special diet 1,487
Not on special diet 8,525
Dietary
status
index2
Dietary
adequacy
score2
Dietary
moderation
score2
1Estimates are based on USDA's 1989-91 Continuing Survey of Food Intakes by Individuals and are
weighted to represent the U.S. adult population. See text for definitions of Dietary Status, Adequacy,
and Moderation.
Bottom number is enen!V.adl:
"D" - Significantly different from those not on a special diet at the p$.05 level.
described their diets as healthful also
tended to have higher DSI scores.
Energy-adjusting the index scores had
only a minor effect on the results. For
most groups examined, the subgroup
with better DAS also had better DMS.
This was true, for example, for groups
compared by age, education, race,
special diet status, and self-rated diet
quality. It was not true, however, for
groups compared by sex. Women had
higher DMS, while men had higher
DAS, resulting in overall DSI scores
that differed far less than their components'
scores. This was probably due, at
least in part, to the large difference in
caloric intakes between the two groups
and because the upper limits for two of
1995 Vol. 8 No.2
the four moderation factors-cholesterol
and sodium-do not vary with caloric
requirements. It appears then, that when
comparing two groups of widely differing
caloric intakes, it may be advisable
to use the two component scores, rather
than the combined score. Alternatively,
researchers may wish to use energyadjusted
scores to examine diet quality
of such groups, either in place of or in
addition to absolute scores.
Based on these results, the DSI, DAS,
and DMS all appear to have promise for
use as summary measures by researchers
studying diet quality. Future studies
examining the properties of these
measures, perhaps in comparison with
other existing summary measures, such
... those who had better
DSI scores were:
Over 50 years of age,
had household incomes
above 130 percent of
the Federal poverty
threshold, and had
more than a high
school education.
Those who reported
being on a special
diet and those who
described their diets
as healthful also
tended to have
higher DSI scores.
11
12
... the value of the DSI
could be enhanced
by incorporating
consumption of
foods defined in the
FGP in its construction
and including the
DSI in publicly
released data sets.
Table 7. Estimated mean index scores: Dietary status, adequacy, and
moderation, U.S. adults, by sex and self-perception of healthfulness of
diet: 1 1989-91
Sex/
Self perception of
healthfulness of diet
Males
Excellent of very good
Good
Fair or poor
Females
Excellent or very good
Good
Fair or poor
All
Excellent or very good
Good
Fair or poor
Sample
size
1,778
1,653
705
2,268
2,413
1,203
4,046
4,066
1,908
Dietary
status
index2
43.0
41.2
49.6G,F
41.2
47.9G,F
41.2
Dietary
adequacy
score2
57.4G,F
53.9
49.2
47.8G,F
39.4
52.6G,F
43.9
Dietary
moderation
score2
34.9
32.0
33.2
43.0
43.1
38.5
1 Estimates are based on USDA's 1989-91 Continuing Survey of Food Intakes by Individuals and are
weighted to represent the U.S. adult population. See text for definitions of Dietary Status, Adequacy,
and Moderation.
"G" - Significantly different from the "Good" category at the ~.OS level.
"F" -Significantly different from the "Fair" category at the ~.OS level.
as the MAR, the INQ, and the DQI,
would further our understanding of their
strengths and limitations. 2 Should results
2 A promising future direction would utilize the
rules of the newly developed fuzzy set theory
("fuzzy logic") to reconstruct the DSI and its
component scores. This approach seems to address
the uncertainty, or fuzziness, inherent in dietary
assessment. For an informative, nontechnical
overview of fuzzy logic see, for example, the
article by Kosko and Isaka in the July 1993 issue
of Scientific American.
of these methodological studies prove
encouraging, future studies could utilize
the DSI to expand on the literature of
determinants of diet quality, by using it
in multivariate analyses, for example. In
addition, the value of the DSI could be
enhanced by incorporating consumption
of foods defined in the FGP in its construction
and including the DSI in publicly
released data sets. This should make the
DSI even more useful for relating overall
diet behavior to its determinants.
FamUy Economks and Nutrition Review
References
1. Abde1-Ghany, M. 1978. Evaluation of household diets by the index of nutritional
quality. Journal of Nutrition Education 1 0(2):79-81 .
2. Basiotis, P.P., Guthrie, J.F., Keane, T., and Welsh, S.O. 1992. Profiles of adults with
varying self-perceptions of dietary quality. Journal of the American Dietetic Association
92(9)(Suppl.):A-45.
3. Basiotis, P.P., Welsh, S.O., Cronin, F.J., Kelsay, J.L., and Mertz, W. 1987. Number of
days of food intake records required to estimate individual and group nutrient intakes
with defined confidence. The Journal of Nutrition 117 (9): 163 8-1641.
4. Carmines, E.G. and Zeller, R.A. 1979. Reliability and Validity Assessment. Sage
University Paper series on Quantitative Applications in the Social Sciences, Series No.
07-017. Sage Publications, Newbury, CA.
5. Greene, W.H. 1993. Econometric Analysis (2nd ed.). The MacMillan Publishing
Company, New York.
6. Morgan, K.J. 1986. Socioeconomic factors affecting dietary status: An appraisal.
American Journal of Agricultural Economics 68(5): 1240-1246.
7. Mertz, W., et al. 1991. What are people really eating? The relation between energy
intake derived from estimated diet records and intake determined to maintain body
weight. American Journal of Clinical Nutrition 54:291-295.
8. Murphy, S.P., Rose, D., and Lane, S. 1994. What is the proper use of a diet quality
index? Journal of the American Dietetic Association 94(9):968.
9. National Academy of Sciences, National Research Council, Food and Nutrition
Board. 1989. Recommended Dietary Allowances (lOth ed.). National Academy Press,
Washington, DC.
10. National Academy of Sciences, National Research Council, Food and Nutrition
Board. 1989. Diet and Health: Implications for Reducing Chronic Disease Risk.
National Academy Press, Washington, DC.
11. Patterson, R.E., Haines, P.S., and Popkin, B.M. 1994. Diet quality index: Capturing
a multidimensional behavior. Journal of the American Dietetic Association 94(1):57-64.
12. Shah, B.V., et al. 1991. SUDAAN User's Manual: Professional Software for Survey
Data Analysis for Multi-Stage Sample Designs. Research Triangle Institute, Research
Triangle Park, NC.
13. Sorenson, A.W., Wyse, B.W., Wittwer, A.J., and Hansen, R.G. 1976. An index of
nutritional quality for a balanced diet. Journal of the American Dietetic Association
68:236-242.
14. U.S. Department of Agriculture and U.S. Department of Health and Human Services.
1990. Nutrition and Your Health: Dietary Guidelines for Americans. Home and Garden
Bulletin No. 232.
15. U.S. Department of Health and Human Services, Public Health Service. 1988. The
Surgeon General's Report on Nutrition and Health. Summary and Recommendations.
DHHS (PHS) Publication No. 88-50211.
16. U.S. Department of Agriculture, Human Nutrition Information Service. 1992. The
Food Guide Pyramid. Home and Garden Bulletin No. 252.
17. Welsh, S., Davis, C., and Shaw, A. 1992. Development of the Food Guide Pyramid.
Nutrition Today November/December, pp. 12-23.
1995 Vol. 8 No.2 13
14
Characteristics of Rural Residents
and Vulnerability to Alcohol Problem
Behaviors
By Elizabeth B. Robertson
Social Scientist
Agricultural Research Service
Alcohol is reported to be the most widely abused substance in rural areas.
However, little has been done to identify alcohol consumption patterns and
problems among rural adults. In this paper, data from the 1990 and 1991
National Household Surveys on Drug Abuse were combined to produce a
sample of U.S. adults residing in areas outside urban settlements with fewer
than 2,500 residents. Relationships between seven personal, family, and
economic characteristics and the frequency of alcohol use, the quantity of
alcohol consumed, and markers of alcohol-related problems were examined.
Results indicate that some demographic characteristics are consistently
related to alcohol problem behavior among rural adults. These characteristics
include being male, being younger than the median age of 33, having
attained a high school or greater education, being employed or unemployed
(as opposed to being retired, disabled, a homemaker, or a student), and
being unmarried. Findings may be helpful to cooperative extension specialists,
other educators, and substance abuse prevention and treatment personnel.
T here is evidence that alcohol
is used at the same rate in
rural as in nonrural areas of
the United States (12). In
most cases, the use of alcohol is not
abusive. The majority of adults who
consume alcohol are referred to as
social drinkers because they rarely
experience negative effects from alcohol
use. The remaining drinkers fall into
two categories: Alcohol abusers, those
who are not dependent on alcohol but
who abuse or misuse it; and alcoholics,
those who are physically and psychologically
dependent upon alcohol (12).
The American Psychiatric Association
reports that about 13 percent of the U.S.
adult population experiences alcohol
abuse or dependence at some time in
their lives (J). The National Institute
on Alcoholism and Alcohol Abuse
(NIAAA) reports that about 10 percent
of adult Americans fall into the alcohol
abuse (4 percent) and dependent (6 percent)
categories (12).
Individuals who fall into any of the
three alcohol consumption categories
may experience serious alcohol-related
problems. For example, the impaired
judgement of the social drinker who
drives may result in a life-threatening
accident; the poor nutritional intake of
the abusive drinker may lead to longterm
health consequences; and the lack
of control of the alcoholic may lead to
high-risk behavior.
Alcohol is reported to be the most
widely abused substance in rural areas
(1 3). However,.little research has been
done to identify alcohol consumption
Family Economics and Nutrition Review
patterns and problems among rural
adults. Sociodemographic characteristics
may be helpful in identifying rural
adults at risk for problems associated
with alcohol consumption. In this
paper, the relationships between seven
personal, family, and economic characteristics
of rural adults and the frequency
of alcohol use, the quantity of alcohol
consumed, and markers of alcoholrelated
problems are examined.
Data and Sample
Combined data from the 1990 and 1991
National Household Surveys on Drug
Abuse (NHSDA) were used in these
analyses. The NHSDA utilizes a multistage
area probability sample design. In
1990 and 1991, the sample sizes were
9,259 and 32,594, respectively. The
1990 target population was the household
population of the 48 contiguous
States. In 1991, this definition was
broadened to include the civilian, noninstitutional
population of the entire
United States. This change introduces
some minor inconsistencies between
the 1990 and 1991 samples. However,
its impact is considered to be generally
inconsequential (10). For a more
complete discussion of the limitations
of the NHSDA data sets, see Robertson
(8).
Only those respondents who were age
21 years and older and who resided
in nonmetropolitan rural areas were
eligible for inclusion in this study. The
NHSDA uses the U.S. Census definition
of nonmetropolitan rural; that is,
the U.S. population residing in areas
(such as counties) outside urban settlements
including wilderness areas,
sparsely settled areas, farmland, and
small places with fewer than 2,500
residents (2). In all, 2,149 subjects fit
the criteria for inclusion. Population
weights of the NHSDA were corrected
to accurately reflect the 1992 census of
rural population by region.
1995 Vol. 8 No.2
Measures
Seven demographic characteristics were
considered in this paper: Gender, age,
race, education, work status, income,
and marital status. Age was split at the
rural subsample median of 33 years to
make two categories: Those below
the median and those at or above the
median. Race was also divided into two
categories: White and non-White. Small
numbers of respondents with Native
American, Alaskan Native, Asian, and
Pacific Islander heritage prevented the
construction of more descriptive race
groups.
After analysis to confirm that there
were no extreme outlier groups with
regard to the percentage who reported
any alcohol consumed in the past year,
all non-White respondents were grouped
into one category. Three levels of education
were distinguished: Less than high
school, high school, and more than high
school. Work status also has three categories:
Employed, unemployed, and
other, with the other category including
the retired, disabled, full-time students,
and homemakers. Per capita income
was split at the median of $8,000. Per
capita income was calculated by dividing
the household income by the number
of household members. Those at and
below the median were grouped together.
Three marital status groups were
considered: Currently married, divorced
or separated, and never married. A
small number of respondents were
widowed, and 72.9 percent of them had
consumed no alcohol in the past year.
This pattern of alcohol use was so
different from those of the other marital
status categories that they were eliminated
from the analyses of marital
status.
Frequency of alcohol consumption was
assessed with a question that read "On
the average, how often in the past 12
months have you had any alcoholic
beverage, that is, beer, wine, or liquor?"
... results demonstrate
that demographic
characteristics are
consistently related
to alcohol problem
behaviors among
rural adults.
15
Frequency of use was coded as daily,
weekly, monthly, less than monthly,
and no alcohol used in the past year.
For the majority of the analyses, the
daily and weekly categories were
collapsed because cell sizes for daily
use were too small to produce reliable
population estimates. Three measures
were used to assess quantity of alcohol
consumed: Having three or more drinks
in one sitting at least once in the past
month, having five or more drinks in
one sitting at least once in the past
month, and being drunk three or more
times in the past year. Twenty-one
problems related to alcohol consumption
in the past year were examined
individually (see table 1) and were used
to create three dichotomous alcohol
problem-behavior measures.
Analysis
Population weights of the NHSDA were
corrected to accurately reflect the 1992
census of rural population by region.
Following the derivation of population
estimates, population data were reverse
weighted and Chi Square analyses run.
All comparisons reported in this paper
were significant at p~.05 .
Results
In all, 55.6 percent of rural adults, age
21 years and older, reported consuming
alcohol in the past year. The figure
compares rural drinkers and nondrinkers
with regard to seven demographic
characteristics. The profile that emerges
of the typical rural alcohol user is of an
older, White, married male who has a
high school education, a job, and an
above-median income. In contrast, the
typical rural nondrinker is an older,
White, married female, with less than
a high school education, who either
works or is not in the labor force by
choice, and who has a below-median
income. Thus, while the profiles of the
drinking and nondrinking segments of
the rural adult population are the same
with regard to age, race, and marital
16
Table 1. Frequency of selected problems related to alcohol consumption
among rural adult drinkers
In past year Percent
1. Felt aggressive or cross while drinking
2. Got into a heated argument while drinking
13.0
6.6
1.1
6.4
5.8
3.4
8.3
6.2
2.4
9.1
2.4
3.6
3. Nearly lost job due to drinking
4. Told by spouse/date to cut down drinking
5. Told by relative to cut down drinking
6. Told by friend to cut down drinking
7. Tossed down drinks fast to get quick effect
8. Was afraid might be/become an alcoholic
9. Stayed drunk for more than 1 day
10. Could not remember things done while drinking
11. Had a quick drink when no one was looking
12. Hands shook after drinking day before
13. Sometimes got high/drunk while alone
14. Stayed out of work due to hangover
15. Was high or little drunk on the job
16. Could not stop drinking until completely intoxicated
11.2
2.8
3.3
5.0
1.2
5.2
17. Often had a drink first thing in morning
18. Kept drinking after promised self to quit
19. Needed more alcohol to get the same effect 20.7
10.5
6.3
20. Felt dependent on alcohol
21. Felt sick when tried to cut down on alcohol
status, they differ with regard to gender,
education, work status, and income.
Use of alcohol is not necessarily an
indication of alcohol problem behavior.
Thus, the profile of the typical problem
drinker in rural areas may not match
that of the typical user. Moreover, the
definition for problem drinking may
vary depending on the orientation of the
definer (1, 3, 5). In general, the following
factors are considered when defining
problem alcohol behavior: Frequency of
use, quantity consumed, and the consequences
of consumption. In this paper,
each of these three factors is considered
separately.
Frequency of Alcohol Consumption
Assessing frequency of alcohol consumption
may help to define problem
behavior through identifying those
individuals who drink regularly, such
as on a daily, weekly, or periodic (for
example, weekends) basis. Those who
drink with greater regularity may be
more likely to overindulge during
periods of heightened stress or celebration.
Hence, they may be more likely
to experience consequences associated
with errors in judgement or lack of
impulse control such as accidents and
aggressive encounters (12).
Family Economics and Nutrition Review
Percentage distribution by demographic factors
Rural drinkers (55.6%) Rural nondrinkers (44.4%)
54.1~
45.9
Gend.er
Male
Female 44.1...._._ 55.9
35.1 c==::::J___
64.9-
Ag.e
Below median 1
Above median
17.1
82.9
93.8 L
6.2 ··--------------'
Bace/ethnicity
White
Non-White
91.7L
8.3 -.---------------'
43.3---
56.7
ns~ 10.7
11.5
Education
<High school
High school
>High school
Work status
Employed
Unemployed
Other2
J.oomm
Below median
Above median
Marital status3
Married
Divorced/separated
Never married
40.0
38.0
22.1
43.8 81~
48.1
875i 7.1
5.4
1Median scores based on entire rural sample data, whereas percentages are for rural adults age 21 years and over.
2Retired, disabled, homemaker, student.
\vldowed were excluded.
Who are the most frequent alcohol
consumers? Sample size restrictions
make generalizations about daily use
problematic, thus drinking weekly or
more is used as the marker of most
frequent use. Table 2, p. 18, presents
frequency of use for the total rural
adult population by demographic
characteristics. Presumably, those in
the groups that report using alcohol
weekly or more frequently would be
more likely to experience alcoholrelated
problems than those who consume
alcohol less frequently. However,
binge drinkers may participate in less
1995 Vol. 8 No.2
frequent but more intense bouts of
alcohol consumption that can lead to
serious social and health consequences.
The weighted population estimates
reported at the top of table 2 are the
self-reported frequency of alcohol use
by rural adults. The largest group is that
which reported no use in the past year.
Only 3.9 percent reported daily use and
12.6 percent reported weekly use; these
numbers were collapsed to create the
weekly or more frequent category.
The demographic categories with the
highest weekly or more frequent use
rates are male, below-median age, high
school or greater educational attainment,
employed and unemployed, abovemedian
income, and not currently
married. A comparison of data from
the figure and table 2 reveals that the
profile of the frequent (weekly or more)
alcohol consumer differs somewhat from
that of the typical alcohol consumer.
Like the typical consumer, the typical
frequent consumer is male and has an
above-median income. The two classifications
differ but overlap with regard to
education and employment status. Rates
for the frequent consumer group were
17
Table 2. Frequency of alcohol consumption by rural adults (past year)
Weekly
Characteristics or more
All 16.5
Gender
Male 24.2
Female 8.9
Age
Below median 21.4
Above median 14.9
Race/ethnicity
White 16.4
Non-White 18.3
Education
Less than high school 13.2
High school 17.7
More than high school 18.2
Work status
Employed 19.8
Unemployed 19.6
Other1 10.7
Income
Below median 12.4
Above median 20.1
Marital status2
Married 14.7
Di vorcedlseparated 23.4
Never married 35.9
1Retired, disabled, homemaker, student.
2Widowed were excluded.
more inclusive than the groups defining
the typical drinker in that weekly or more
drinkers can be characterized as having
attained high school or greater education
and being either employed or seeking
employment, whereas the typical drinker
is employed and has a high school education.
Finally, the typical user and the
typical frequent drinker differ with
regard to age, race, and marital status.
The frequent user can be typified as
younger rather than older than the
18
Less than
Monthly monthly No use
Percent
14.7 24.4 44.4
15.9 20.5 39.5
13.5 28.3 49.3
23.1 28.4 27.2
11.9 23.1 50.2
14.6 25.1 43.9
14.8 15.3 51.6
9.2 18.6 59.0
16.0 25.9 40.4
18.4 28.5 34.8
16.8 26.6 36.8
24.8 25.2 30.4
9.7 20.9 58.8
13.4 24.5 49.7
16.0 24.3 39.7
14.9 25.6 44.8
19.5 24.8 32.2
17.5 21.2 25.4
median age and not married as opposed
to married. Moreover, among frequent
users, similar rates of weekly alcohol
use were evident for the two race groups,
whereas the typical drinker is White.
Thus, making distinctions between all
rural adult alcohol consumers and
frequent consumers does yield a somewhat
different profile. These distinctions
offer some hints about groups that are
at risk for problem alcohol behavior.
However, consuming alcohol daily or
weekly is not a problem unless consumption
is at a level that interferes
with social, occupational, psychological,
or physical functioning (1). Clearly,
without evidence of quantity consumed,
no conclusions about risk of problem
behavior can be drawn.
Quantity of Alcohol Consumed
The effect of alcohol varies with an
individual's weight, body chemistry,
and tolerance for alcohol. For example,
women are more susceptible than men
to the effects of alcohol because of
differences in metabolic enzyme activity
and body composition (9,14). Thus,
creating a generic "level of alcohol
consumption" definition or measure is
somewhat problematic. Several of the
many existing definitions of moderate
and heavier drinking follow:
• The Human Nutrition Information
Service (HNIS) defines moderate
drinking as no more than one drink
per day for women and two drinks
per day for men (9).
• The National Institute on Alcoholism
and Alcohol Abuse (NIAAA)
defines heavier drinking as drinking
two or more drinks per day or an
average daily intake of 1 or more
ounces of ethanol (7).
• The National Institute on Drug Abuse
(NIDA) defines heavy drinking as
five or more drinks in a row, at
least once in the past 2 weeks (4).
• The Substance Abuse and Mental
Health Services Administration
(SAMHSA) defines heavy drinking
as drinking five or more drinks per
day on each of 5 or more days in
the past 30 days (JJ).
In this study, three definitions of quantity
consumed are used. The first, three
or more drinks during at least one
sitting in the past month, is used as a
threshold measure between moderate
and heavier drinking. The second and
third measures, five or more drinks
during at least one sitting in the past
month and being drunk at least three
Family Economics and Nutrition Review
times in the past year, are viewed as
indicators of heavier use. These categories
are not mutually exclusive.
Table 3 shows that a substantial percentage
of individuals report alcohol use at
or above the level indicated by the three
measures of consumption. With few
exceptions, the percentage reporting
consumption by category for the demographic
variables is over 10 percent. In
addition to these generally high rates of
moderate to heavier alcohol consumption,
there are many strong categorical
differences. For example, more men
than women, more of those below the
median age than above the median age,
and more people who are not currently
married than who are, report moderate
to heavier consumption practices.
The profile of the drinker who consumed
three or more drinks in one sitting in the
past month matches that of the frequent
drinker. Thus, the profiles of weekly
consumers and the threshold to heavier
drinking consumers are very similar.
Also, the typical pattern for those who
reported drinking five or more drinks
at one sitting in the past month matches
that for those who report three or more
drinks except that the prevalence rate is
lower for those in the higher consumption
category.
The final measure of quantity consumed,
being drunk three or more times in the
past year, results in a somewhat different
pattern of demographic categories
than the other two quantity measures.
Patterns for gender and age are similar
with more younger people and males
reporting this consumption behavior.
However, several demographic categories
emerge that were not prominent before.
First, non-Whites report a higher prevalence
of this behavior than do Whites.
Further, those who were unemployed
have higher prevalence rates than those
in the other two categories, and those
who had never married have higher
prevalence rates than those who had
ever married. Finally, the high school
1995 Vol. 8 No.2
Table 3. Percentage of rural adults reporting moderate to heavier
drinking
3+ drinks1 5+ drinks Drunk3+
Characteristics per sitting in per sitting in times in
past month past month past year
All
Gender
Male
Female
Age
Below median
Above median
Race/ethnicity
White
Non-White
Education
Less than high school
High school
More than high school
Work status
Employed
Unemployed
Other2
Income
Below median
Above median
Marital status3
Married
Divorced/separated
Never married
22.2
31.6
13.9
37.9
17.7
22.5
25.0
16.6
26.3
23.8
28.9
29.9
11.7
19.6
25.5
20.3
40.0
42.7
12.7 18.3
18.7 23.9
7.1 14.9
24.8 38.7
8.8 12.8
12.8 18.9
12.9 25.0
9.7 12.8
14.8 24.4
13.3 18.9
17.6 23.6
18.4 34.2
4.4 10.2
11.6 19.5
14.0 19.2
10.6 17.7
26.6 29.4
30.3 37.6
1one drink is equivalent to 12 fluid ounces of beer, 5 fluid ounces of wine, or 1-1/2 fluid ounces of
distilled liquor.
2Retired, disabled, homemaker, student.
3widowed were excluded.
category has a higher percentage of
people who got drunk three or more
times in the last year than the other two
education categories. Thus, the typical
pattern that emerges with regard to the
third measure-got drunk three or more
times in the past year~an be described
as including younger, unemployed,
never married, minority, males who
have a high school education.
Problems Related to Alcohol
Consumption
The third means of defining problem
alcohol behavior is self-reported problems
that result from alcohol consumption.
The presence or absence of 21
individual problems was assessed.
Then the problems were grouped in
three ways. First, a count of all measures
was used to identify those individuals
who had experienced one or more
19
20
The consistent pattern
of associations
between being
younger in age and
alcohol problem
behaviors ... points
to this group as
a target for
interventions.
alcohol-related problems in the past year.
Second, a count of two measures related
to aggressive feelings and behavior was
made to identify those individuals who
had experienced one or both of those
problems in the past year. Finally, a
count of eight items that measure symptoms
of alcohol dependency as defined
by the Diagnostic and Statistical Manual
of Mental Disorders (DSM-Ill-R)1 {1)
was made to identify those individuals
who had experienced one or more
symptoms of alcohol dependence in
the past year.
Table 1 presents rates of problems
associated with alcohol consumption
for rural adult drinkers. Items are grouped
by content. The first two items refer
to aggressiveness while drinking. Relatively
high rates for both items, but
especially the items that refer to aggressive
feelings (13.0 percent), were .
reported. Items 3 through 13 include
social and physical consequences of
alcohol consumption. The content of
these items and the rates of experiencing
the problems they represent were
varied. Some problems had relatively
high rates: Getting high or drunk alone
( 11.2 percent); not remembering things
done while drinking (9.1 percent); and
tossing down drinks (8.3 percent).
Others had low rates: Nearly lost job
due to drinking (1.1 percent); stayed
drunk for more than one day (2.4 percent);
and had a quick drink when no
one was looking (2.4 percent).
The last eight items measure aspects of
eight of the nine symptoms of alcohol
dependence as defined by the DSMIll-
R {1). To be diagnosed as alcohol
dependent an individual would have to
exhibit three or more of these symptoms.
For those who reported drinking in the
past year, five of these eight measures
1Within the United States, this manual is the most
widely used among psychiatric clinicians for the
classification of mental disorders.
have prevalence rates of 5 percent or
greater. Most startling is the percentage
of rural drinkers who report needing
more alcohol to get the same effect
(20.7 percent) and feeling dependent
on alcohol (10.5 percent).
When the 21 measures reported in
table 1 were grouped, the percentages
of rural adults experiencing one or more
problems in the past year were high,
ranging from 9.2 percent to 34.1 percent
(table 4). Those who experienced one
or more problems can be characterized
as younger, single, male, with a high
school or greater educational attainment,
and either employed or unemployed
(compared with those who are retired,
disabled, homemakers, or students).
With two exceptions, those who reported
one or more problems related
to alcohol consumption are demographically
similar to those who were frequent
drinkers, and to those who reported
having three or more and five or more
drinks in a sitting in the past month. The
exceptions are income, for which there
is little difference across the two categories,
and employment status, for which
the unemployed have the highest rate
followed by the employed.
Aggression can be one of the most
serious problems related to alcohol
consumption because it is linked to
other problems including assaultive
behaviors {12). The aggressiveness
related to drinking measure presents
a profile similar to that of the one or
more problems associated with alcohol
measure. However, the magnitude of
differences in rates for some categories
makes conclusions about the typical
profile somewhat different. The most
notable difference is the lack of the
major disparity in rates between males
and females that was evident for all
other frequency, quantity, and consequence
measures. For the education
categories, those with high school or
more than high school education
generally are similar with regard to
alcohol problem behaviors. However,
Family Economics and Nutrition Review
Table 4. Percentage of rural adults experiencing alcohol-related
problems
One or more
problems Aggressiveness One or more
associated with related to symptoms of
Characteristics drinking drinking dependency
All 17.9
Gender
Male 23.9
Female 12.1
Age
Below median 32.8
Above median 13.0
Race/ethnicity
White 17.9
Non-White 18.6
Education
Less than high school 13.8
High school 19.7
More than high school 19.8
Work status
Employed 21.8
Unemployed 34.1
Other1 9.2
Income
Below median 18.4
Above median 17.6
Marital status2
Married 16.3
Divorced/separated 30.7
Never married 32.6
1Retired, disabled, homemaker, student.
2Widowed were excluded.
for aggressiveness while drinking, the
high-school graduate group reported
the highest rate of aggressiveness, but
its rate was only 3.3 percentage points
higher than that of the more-than-highschool
education group and 3.9 percentage
points higher than the less-than-highschool
education group.
1995 Vol. 8 No.2
7.7 9.2
9.2 13.6
6.2 4.9
16.3 15.6
4.9 7.1
7.7 8.9
8.5 13.3
5.9 7.9
9.8 10.1
6.5 9.4
10.0 11.3
19.9 16.5
2.2 4.7
8.5 9.9
7.0 8.5
6.4 7.5
15.3 20.9
18.0 19.4
The fmal measure of consequences or
problems related to alcohol use is: one
or more symptoms of alcohol dependency
based on the DSM-ill-R (1).
Interestingly, the pattern of demographic
characteristics that emerges for this
measure is very similar to that for the
drunk three or more times in the past
year measure. Like the profile for that
measure, the typical demographic
proftle for the dependency measure is
younger, unemployed, single, minority
male. The one area of difference between
the profiles for the dependency
and the drunk measures was education.
The magnitude of the difference in rates
for the three educational categories was
very small. This pattern is similar to that
seen for the aggressiveness measure,
though in the case of the dependence
measure, the differences are so small
that they are meaningless.
Conclusion
Seven measures of frequency of alcohol
consumption, quantity of alcohol consumed,
and problems associated with
alcohol consumption were examined
in conjunction with seven demographic
variables. The combined results demonstrate
that demographic characteristics
are consistently related to alcohol problem
behaviors among rural adults. With few
exceptions, these demographic characteristics
include being male, being
younger than the median age of 33,
having attained a high school or greater
education, being in or seeking to be in
the labor force, and being unmarried.
What were the exceptions to this proftle?
First, the male-dominated pattern
of alcohol problem behavior was less
evident for the aggressiveness measure
than for the other alcohol problem
behaviors. That is, the percentage of
women reporting one or both aggressive
feelings or actions was lower than that
of men, but the magnitude of the difference
between the two genders was not
as great for this measure as it was for
the other alcohol problem-behavior
measures. Many studies have noted the
connection between alcohol consumption
and aggression (12). Although
aggressive behavior is more often
reported by males than females,
aggressive feelings are likely to be
experienced by both genders under
the influence of alcohol.
21
Second, education was important in defining
those who drink more frequently,
drink moderate and heavier quantities,
and who experience one or more problems
related to alcohol consumption.
Specifically, higher percentages of those
with high-school or greater educational
attainment fit this profile. However,
education was not as important in
characterizing those who experienced
aggressiveness when drinking and
symptoms of alcohol dependency. That
is, there was little difference in rates
for these problem behaviors among
members of the three education categories.
Finally, the pattern of rates for the
employment categories was consistent
across all seven measures of alcohol
problem behaviors. In all cases, the
employed and unemployed categories
had higher rates than the "other" (those
who are retired, disabled, homemakers,
or students) category. However, there
also were differences in the rates of the
employed and unemployed groups. For
the number of drinks measures, three or
more and five or more drinks per sitting
in the past month, rates were similar.
For the remaining alcohol problembehavior
measures, members of the
unemployed category had appreciably
higher rates than members of the employed
category. Thus, in most cases,
the highest rates of problem behaviors
were for the unemployed, followed by
the employed, and then relatively low
rates for the "other" category. It is not
surprising, given the stress of being
unemployed, that this category has
the highest rates of alcohol problem
behaviors. What is surprising is the
relatively high rates for the employed
group.
Categories for two of the demographic
variables, race and income, were not
as consistently related to the alcohol
consumption and problem behaviors
as other demographic factors. However,
some patterns were evident. For race,
the only differences between Whites
and non-Whites were for being drunk
three or more times in the past year
and experiencing one or more symptoms
of alcohol dependence. Prevalence of
22
drunkenness and dependence were
higher for the non-White group. Being
drunk is, to an extent, a subjective state.
That is, at what point is someone drunk
and does the definition of drunkenness
differ across cultural contexts? If definitions
do differ, the results of this analysis
may be masked by those cultural
differences. On the other hand, if subjective
appraisals of drunkenness do
not differ across subcultures, then more
precise measures of race/ethnicity
should be employed to identify race/
ethnic groups where drunkenness occurs
more frequently and the personal and
cultural reasons for that behavior.
Those in the non-White category also
reported higher rates of one or more
alcohol dependence behaviors. This
fmding appears to support the previous
finding; however, the diverse membership
of the non-White group offers no
insights into which particular group(s)
account for these findings. Several
race/ethnicity minority groups have
substantial rural populations. Moreover,
some of these groups live in relative
physical and social isolation and experience
problems associated with poverty
and limited educational and occupational
opportunities which may exacerbate
dependency problems (6). Careful
study of patterns of consumption and
problems associated with alcohol consumption
among rural minority groups
is called for.
Income was interesting in that those
above the median income were more
likely to drink, to drink more frequently,
and to drink moderate-to-heavier amounts.
However, differences in rates for the
two income groups were not evident for
the heavier quantity measures and the
alcohol problem measures. Thus, while
it appears that in rural settings having a
higher income is related to drinking, it
is not related to the experience of alcohol
problem behaviors. Perhaps having
more expendable income makes it more
likely that one will drink at all, but does
not influence the likelihood of drinking
to excess or experiencing the consequences
associated with excessive
drinking.
Implications
Low population and distance between
settlements in rural areas influence the
availability of health care services,
including alcohol prevention and intervention
programming for adults (15).
These factors also make it especially
important that rural populations at high
risk for alcohol problem behavior be
clearly identified and that programs be
designed with specific groups in mind.
Results of this study give some insight
into the demographic characteristics of
those who would benefit from alcohol
prevention and intervention programs.
However, multivariate studies are
needed to determine combinations of
factors and processes that contribute
to high percentages of specific demographic
groups experiencing alcohol
problem behaviors. For example, what
factors account for rural males experiencing
higher rates of alcohol problem
behaviors than rural females? Are males
who earn their living in high-risk occupations
such as mining, forestry, firefighting,
and farming more likely to
experience alcohol problems than other
men? Is the seasonality of many rural
occupations related to patterns of
alcohol consumption?
The consistent pattern of associations
between being younger in age and
alcohol problem behaviors also points
to this group as a target for interventions.
However, much remains unclear from
these findings. For example, when did
these problem behaviors begin, what
aspects of the rural environment support
and encourage these problem behaviors,
and will these problem behaviors naturally
decrease with age or is the current
group of rural young adults likely to
experience alcohol problems across the
life course?
Alternately, one could examine characteristics
of the demographic categories
with lower prevalence of alcohol problem
behaviors for insights to what factors
and processes influence the absence and
presence of alcohol problem behavior.
Family Economics and Nutrition Review
For example, why do those with less
than a high school education drink less
and have fewer problems associated
with drinking than those with high
school and more? Does the educational
experience in the higher grades and in
college support the initiation and perpetuation
of alcohol problem behaviors?
If this is the case, rural alcohol prevention
programs need to continue to target
schools. Further, what aspects of the
marital relationship contribute to the
absence of these alcohol problem
behaviors? Perhaps in rural areas where
distance, poor transportation, and lack
of recreational facilities limit the sphere
of positive social interactions, supportive
family relations are especially important.
Other sources of meaningful interaction
and social support such as extended
family, church, and civic group memberships
should be examined to see if they
are also related to lower rates of alcohol
problem behaviors. If so, an important
aspect of prevention and intervention
programming in rural areas would be
the inclusion of social activity.
Finally, although all the problem
behaviors examined in this paper are
of concern, the findings regarding
aggressiveness while drinking may be
of particular importance. The limited
difference between males and females
with regard to this problem behavior
suggests that marital conflict among
couples where one or both partners drink
may be particularly volatile and may
spill over into parent-child relationships.
Aggressive conflict in isolated rural
settings is of particular concern as there
are fewer social controls to prevent
conflict from getting out of hand. Moreover,
distance in rural settings may
become a factor if aggressive feelings
are more likely to be played out behind
the wheel of a moving vehicle.
1995 Vol. 8 No.2
References
1. American Psychiatric Association. 1987. Diagnostic and Statistical Manual of Mental
Disorders. Third Edition, Revised.
2. Dahmann, D.C. and Dacquel, L.T. 1992. Residents of Farms and Rural Areas: 1990.
Current Population Reports, Population Characteristics. Series P-20, No. 457. U.S.
Department of Agriculture, Economic Research Service and U.S. Department of
Commerce, Bureau of the Census.
3. Hilton, M.E. 1988. Trends in U.S. drinking patterns: Further evidence from the past
20 years. British Journal of Addiction 83:269-278.
4. Johnston, L.D., O'Malley, P.M., and Bachman, J.G. 1991. Drug Use Among American
High School Seniors, College Students and Young Adults, 1975-1990. Volume 1. High
School Seniors. The University of Michigan Institute for Social Research and U.S.
Department of Health and Human Services, National Institute on Drug Abuse. DHHS
Publication No. (ADM)91-1813.
5. Knupfer, G. 1987. New directions for survey research in the study of alcoholic
beverage consumption. British Journal of Addiction 82:583-585.
6. O'Hare, W.P. and Curry-White, B. 1992. The Rural Underclass: Examination of
Multiple-Problem Populations in Urban and Rural Settings. Population Reference
Bureau.
7. Piani, A.L. and Schoenborn, C.A. 1993. Health Promotion and Disease Prevention:
United States, 1990. U.S. Department of Health and Human Services, Public Health
Service. DHHS Publication No. (PHS)93-1513, Series 10, No. 185.
8. Robertson, E.B. 1994. Trends in drug use: A comparison of metropolitan and nonmetropolitan
areas of the United States from 1975-1991. Family Econamics Review
1(4):2-10.
9. U.S. Department of Agriculture, Human Nutrition Information Service. 1993./fYou
Drink Alcoholic Beverages, Do So In Moderation. Dietary Guidelines for Americans.
Home and Garden Bulletin Number 253-8.
10. U.S. Department of Health and Human Services, Public Health Service. 1993.
National Household Survey on Drug Abuse: Main Findings 1991. DHHS Publication
No. (SMA)93-1980.
11. U.S. Department of Health and Human Services, Public Health Service. 1993.
Preliminary Estimates From the 1992 National Household Survey on Drug Abuse.
Advance Report Number 3.
12. U.S. Department of Health and Human Services, Public Health Services. 1990.
Alcohol and Health. DHHS Publication No. (ADM)90-1656.
13. U.S. General Accounting Office, Program Evaluation and Methodology Division.
1990. Rural Drug Abuse: Prevalence, Relation to Crime, and Programs. GAO/PEMD-
90-24.
14. Wisconsin Clearinghouse. 1991. Mind-Altering Drugs Series: Alcohol. Wisconsin
Department of Health and Social Services, Office of Alcohol and Other Drug Abuse.
15. Witherspoon, J.P., Johnstone, S.M., and Wasem, C.J. 1993. Rural TeleHealth:
Telemedicine, Distance Education and Informatics for Rural Health Care. U.S.
Department of Health and Human Services, Public Health Service.
23
24
Housing Trends
By Nancy E. Schwenk
Consumer Economist
Center for Nutrition Policy and Promotion
Recent trends in U.S. residential housing are reported using data from
the Federal Government and trade associations. Between 1992 and 1993,
prices for housing rose 2.7 percent, less than the overall inflation rate of
3.0 percent. Home sales in 1993 were at near-record levels; low mortgage
interest rates made entry into the housing market easier for first-time buyers.
Home prices have fluctuated over the past few years in all regions of the
country. Home ownership rates remain highest in the Midwest and the
South. The South accounts for 52 percent of the Nation's mobile homes,
compared with 35 percent of all housing units. The Department of Housing
and Urban Development is promoting the concept of universal design in
housing, implying that the housing components may be used by everyone,
including the elderly and disabled.
he impact of the housing
industry on the U.S. economy
is significant. New housing
construction stimulates the
economy through the creation of jobs,
wages, and tax revenues, and the demand
for goods and services created by new
construction is felt throughout the
economy. In addition, housing is the
biggest expenditure category of American
households, accounting for nearly one
of every three dollars spent (13). The
American dream of owning a home has
been realized by about two-thirds of
households (1). However, many firsttime
buyers were able to enter the housing
market in 1993 because of lower mortgage
interest rates-a key determinant
of home sales and starts (2). This article
presents fmdings from various Federal
Government and trade association
publications that survey trends in U.S.
residential housing. Topics include
prices, expenditures, demographic
characteristics of homeowners and
home buyers, and housing characteristics.
Housing Prices
Consumer Price Index
The shelter component of the Consumer
Price Index (CPI) consists of renters'
costs, homeowners' costs (including
household insurance), and maintenance
and repairs. Since 1960, the prices for
shelter have risen at about the same rate
as prices for all items, with three exceptions:
1969-70, 1978-80, and 1986
(ftg. 1) (12). During these periods, the
rise in shelter prices outpaced the rise
in prices for all items by more than
2 percentage points.
In addition to shelter, the housing
category of the CPI includes utilities,
furnishings, and housekeeping supplies
and services. The 2.7-percent rise in the
CPI for housing between 1992 and 1993
was less than the increase for all items,
3.0 percent (table 1, p. 26). The housing
component that showed the greatest
increase was household insurance, up
3.3 percent. The prices of piped gas,
electricity, and other utilities and public
Family Economics and Nutrition Review
Figure 1. Changes In consumer prices for shelter and all Items
1960-93 '
% change from
previous year
20
All Items
18
16 Shelter
14
12
10
8
6
4
2
0
1960 70 80 90 93
Socwe: U.S. DeplllfmfHit al Ubor, 8/tNU of Leor S.tllltb, CPI Detllled RllpOtt.
services were also up more than the
overall inflation rate. Increases of less
than I percent were observed for furnishings
and housekeeping supplies.
The price of fuel oil and other household
fuel commodities decreased 0.4
percent from 1992 (12).
Between 1983 and 1993, the CPI for all
items was up 45 percent, whereas the
CPI for housing rose 42 percent. During
the 1 0-year period, the price of household
insurance increased 42 percent;
fuels and other utilities, 21 percent;
and homefurnishings, 9 percent (12).
Sale Prices
Despite soft economies on the east and
west coasts, 5.1 million homes were
sold in the United States in 1993,just 5
percent below the all-time record high
of 19781 (2). Selling prices for homes
have fluctuated slightly over the past
few years (fig. 2, p. 27). In 1993, the
median sale price of a new single-family
1995 Vol. 8 No.2
home sold in the United States was
$126,500, up$6,500from 1989, whereas
the median sale price of an existing
single-family home was $106,800,
up $13,700 from 1989 (4).
Home prices in 1993 were above the
national median in the Northeast and
the West and below the national median
in the Midwest and the South (2). The
Northeast had the highest priced new
homes and the West had the highest
priced existing homes, whereas the
South had the lowest priced new homes
and the Midwest had the lowest priced
existing homes (5,9). According to the
National Association of Realtors' forecast,
home prices in 1995 should rise
at an annual rate of over 4 percent and
outpace the rise in consumer prices (5).
1The Atlanta metropolitan area led the Nation in
1993 in the number of residential building permits
issued, up 27 percent over 1992 (4). The Chicago
area was second in building permits issued, and
the Washington, DC, area was third.
••. 5.1 million ·homes
were sold In the
United States in
1993, just 5 percent
below the all-time
record high •••
25
In 1993, mortgage
interest rates hit
their lowest level
in 25 years with
long-term fixed
rates as low as
6.5 percent.
26
Table 1. Consumer Price Index for All Urban Consumers (CPI-U):
U.S. city average [1982-84 = 100]
Percent change from previous year
Group 1991 1992 1993
All items 4.2 3.0 3.0
Housing 4.0 2.9 2.7
Shelter 4.5 3.3 3.0
Renters' costs 1 6.1 3.4 2.5
Homeowners' costs 1 3.9 3.4 3.2
Household insurance 1 2.3 2.7 3.3
Maintenance and repairs 3.4 1.8 1.6
Maintenance and repair services 3.1 2.1 1.4
Maintenance and repair commodities 3.8 1.2 1.8
Fuel and other utilities 3.3 2.2 3.0
Fuels 2.1 1.3 2.9
Fuel oil and other household -4.7 -4.1 -.4
fuel commodities
Gas (piped) and electricity 3.0 2.0 3.2
(energy services)
Other utilities and public services 4.7 3.3 3.2
Household furnishings and operation 2.4 1.7 1.1
Housefurnishings .7 1.4 .5
Housekeeping supplies 3.0 .5 .8
Housekeeping services 6.2 3.6 2.8
1Indexes on a December 1982 = 100 base.
Source: U.S. Department of Labor, Bureau of Labor Statistics, CPI Detailed Report.
Mortgages
Mortgage interest rates affect both the
number of homes sold and the type of
loan home buyers choose. Mortgage interest
rates peaked in 1982 at an annual
average of 14.8 percent and then declined
rapidly (fig. 3, p. 28). Adjustablerate
mortgages (ARM's) reached their
peak in popularity in 1984 when 64
percent of mortgage loans for existing
homes (fig. 4, p. 29) and 59 percent of
mortgage loans for new homes were
adjustable (8). For the past several years,
ARM's have accounted for between
one-fifth and one-fourth of all loans.
ARM's are more popular in the West
than in the other regions of the country
(3).
Fixed-rate mortgages are more desirable
during periods of declining rates, such
as during the past several years. In
1993, mortgage interest rates hit their
lowest level in 25 years with long-term
fixed rates as low as 6.5 percent (3).
Between 1989 and 1993, the percentage
of home buyers financing their home
with a conventional loan, provided by
a private mortgage insurance company,
increased from 62 to 72 percent, whereas
the percentage using Federal Housing
Administration (FHA) loans, insured
by the Federal Government, dropped
from 14 to 9 percent. The percentage of
home buyers paying cash also dropped
during this period from 18 to 13 percent
(9).
Family Economics and Nutrition Review
Figure 2.
Median sale price of new single-family houses sold, by region,
1989-93
$thousands
170
160
150
140
130
120 : 110
100
90
80
70
1989 90 91 92 93
Median sale price of existing single-family houses sold, by region,
1989-93
$thousands
170
160 West
150 :-- ------=:::::: ' :
= 140 ·-- .;
130 Northeast
120
110 U.S. 100 : • : : South
90 : •
80
Midwest
70
1989 90 91 92 93
Source: u.s. Department of Commerce, Bureau of the Censu•. 1QQ4, CMracterlstlcs of New
Housing: 1993, Cu"ent Construction Reports, C25/93-A and NltiOnll AHOC/atlon of Res/tors,
1974, Real Estate Outlook: Market Trends and lns/Qhts, Vol. 1, No. 4.
1995 Vol. 8 No. 2
Expenditures
Consumer Expenditure Survey
According to the 1992 Consumer
Expenditure Survey, housing expenditures
accounted for 31 percent of
household expenditures-the largest
expenditure share. By comparison,
transportation accounted for 17 percent
and food, 14 percent of expenditures.
Households headed by people age 35
to 44 and households living in the West
had the highest housing expenditures,
whereas those under age 25 and Black
households had the lowest (table 2, p. 30).
The percentage of total expenditures
allocated to housing was highest for
households headed by people age 25 to
34, households living in the Northeast,
renters, and Black households (13).
Similar housing expenditure patterns
were found in 1982. However, compared
with 1992, the proportion of expenditures
allocated to housing in 1982 was
lower for nearly every demographic
group (13,14).
"Who's Buying Homes in America"
Survey
Data from this annual nationwide
survey (2) show that in 1993, flrst-time
home buyers saved for an average of
2.8 years for the downpayment and
looked at 12.9 houses before making a
purchase. They had an average mortgage
payment of $950, down from $1,046
in 1991. Repeat buyers looked at 15.6
houses before making a purchase and
had an average mortgage payment of
$1,076, down from $1,230 in 1991.
Maintenance and Repairs
Home maintenance includes painting,
replacing broken windows, and repairs
to air conditioning, walls, plumbing,
and so on. Maintenance does not include
landscaping and gardening costs.
Median home maintenance expenditure
in 1991 was $315. Maintenance expenditures
were reported by 63 percent of
households. Residents of the Northeast
27
28
Old homes were
not associated with
significantly larger
maintenance costs.
Figure 3. Contract Interest rates, conventional first mortgage
loans for purchase of existing single-family homes, 1965-92
annual averages
Percent
15
10
5
0 75 80 85 90
Source: U.S. Department of Commerce, Bureau of the Census, 1993, Statistical Abstract of the
United States, 1993 [113th ed.]
spent the most ($400), whereas residents
of the Midwest spent the least ($253).
Maintenance expenditures increased as
home size increased. Old homes were
not associated with significantly larger
maintenance costs. Households living
in homes built before 1950 spent $329
per year, compared with $310 spent by
households living in newer homes (6).
Home improvements, repairs, and alterations
include adding a room, replacing
a basic system such as a water heater or
central air-conditioning, and replacing
a roof. According to the American
Housing Survey, median expenditure
for home improvements, repairs, and
alterations over a 2-year period, 1990-
1991 , was $2,101. These expenditures
were reported by 49 percent of households.
Residents of the Northeast spent
the most ($2,409), whereas residents of
the South spent the least ($1,874) (6).
Demographic Characteristics
Homeowners
Nationwide, home ownership rates have
varied by only 2 or 3 percentage points
since 1960 (table 3, p. 30). By age of
householder, home ownership in 1993
ranged from 15.0 percent for those
under age 25, to 80.9 percent for those
age 60 to 64 years (1).
In 1993, the Midwest and the South had
higher home ownership rates than the
U.S. average of 64.5 percent, whereas
the Northeast and the West had lower
rates. The West was the only region in
which the home ownership rate had
decreased since 1960. The Northeast
had the greatest gain in home ownership
between 1960 and 1993-12 percent.
The States with the highest home ownership
rates in 1993 were Delaware (74.4
percent), West Virginia (73.3 percent),
and Michigan (72.6 percent); whereas
Hawaii (53.2 percent), New York (53.5
percent), and Alaska (56.0 percent) had
the lowest rates (1).
Family Economics and Nutrition Review
Figure 4. Percentage of mortgage loans for existing homes with adjustable rates, 1982-93
64
1982 83 84 85 86 87 88 89 90 91 92 93
Source: U.S. Department of Commerce, Bureau of the Census, 1994, Statistical Abstract of the United States, 1994 {114th ed.] and U.S.
Department of Commerce, Bureau of the Census, 1993, Statistical Abstract of the United States, 1993 {113th ed.]
The home ownership rate in 1993 was
79.1 percent for married-couple families,
up from 78.3 percent in 1983. For male
householders with no wife present, the
home ownership rate was 54.6 percent,
down from 59.2 percent in 1983; and
for female householders with no husband
present, the rate was 44.5 percent, down
from 47.0 percent. Among one-person
households, the 1993 rate was 43.2 percent
for males living alone, up from
38.3 percent in 1983; and 54.8 percent
for females living alone, up from 52.0
percent (1).
Home Buyers
During the 1980's, first-time buyers
accounted for less than 40 percent of
home buyers, but in 1993 they made
up 46 percent of the market (3). The
percentage of first-time buyers ranged
1995 Vol. 8 No. 2
from 41 percent in the South to 55 percent
in the Northeast (2). Never-married
single buyers accounted for about onethird
of frrst -time buyers in 1993, the
highest percentage in 18 years (3). Firsttime
buyers' average age was 31.6 years
in 1993, compared with 41.0 years for
repeat buyers. Of first-time buyers, 20
percent bought new homes and 82 percent
bought single-family homes. Of
repeat buyers, 24 percent bought new
homes and 87 percent bought singlefamily
homes (2).
Housing Characteristics
New Housing
Single-family houses being built today
are bigger than those built in years past.
In 1993, new single-family homes were
an average of 2,095 square feet and a
median of 1,945 square feet (table 4,
p. 3 2), 40 percent bigger than those
built in 1970 (4,8). Only 34 percent of
new single-family homes nationwide
had central air-conditioning in 1970, but
this proportion jumped to 63 percent in
1980 and to 78 percent in 1993 (4,8).
There is also a trend toward more homes
with 2-or-more-car garages, 4-or-more
bedrooms, and gas heat (4).
New housing characteristics vary among
regions of the country. The largest homes
were built in the Northeast and the
smallest, in the West. Installed central
air-conditioning was most prevalent in
new homes in the South (97 percent)
and least prevalent in the West (50 percent)
(9). On exterior walls, the use of
brick is increasing in the South, but decreasing
in the Northeast and Midwest,
whereas the use of wood is increasing
only in the West. Vinyl siding is being
used increasingly in all regions (9).
29
Table 2. Average annual expenditures for housing, by demographic
characteristics, 1992 and 1982
Percent of total
Mean dollars annual expenditures
Characteristic 1992 1982 1992 1982
AU households $9.477 $5,582 32 31
Region
Northeast 10,701 5,378 34 32
Midwest 8,504 5,595 30 31
South 8,422 5,323 30 30
West 11,150 6,251 33 32
Housing tenure
Homeowner 10,855 6,480 31 30
Renter 7,252 4,218 34 32
Head of household
Age (years)
<25 5,135 3,462 30 30
25-34 10,018 6,283 34 33
35-44 12,120 7,080 33 30
45-54 11,036 6,631 29 28
55-64 9.436 5,280 30 29
65 and over 6,733 3,851 33 34
Race
White and other 9,833 5,765 32 31
Black 6,718 4,158 34 31
Source: U.S. Department of Labor, Bureau of Labor Statistics, Consumer Expenditure Surveys: 1992
Interview Survey Documentation; and Interview Survey, 1982-83, Bulletin 2246.
Table 3. Home ownership rates, by region, selected years, 1960-93
1960 1970 1980 1990 1993
u.s. 62.1 64.2 65.6 63.9 64.5
Northeast 55.5 58.1 60.8 62.6 62.4
Midwest 66.4 69.5 69.8 67.5 67.4
South 63.4 66.0 68.7 65.7 66.1
West 62.2 60.0 60.0 58.0 60.4
Source: U.S. Department of Commerce, Bureau of the Census, 1994, Housing Vacancies and
Homeownership, Annual Statistics: 1993, Table 18, H1 1 1193-A.
30
Future Trends
Universal Design (11)
Many people have physical or mental
disabilities that threaten their independent
living. With the increase in the number
of older and disabled Americans, there
is a growing market for housing that
can be adapted for use by more peoplechildren,
older adults, and physically
handicapped-at little or no cost. This
concept has been called "universal
design." A universal feature is any
component of a house that can be used
by everyone. The following are examples
of universal design modifications:
• Replacing the traditional front
door knob with a level handle that
requires no gripping or twisting to
operate.
• Providing wider passageways and
doors between rooms, providing
easier access for wheelchair users.
• Placing clothes closet rods at
adjustable heights to accommodate
people of all heights and those in
wheelchairs.
• Locating the bathtub faucet close to
the outside rim of the tub, making it
easy for everyone to reach.
• Reinforcing the wall above the
bathtub with blocking so that the
wall can accept grab bars.
• Installing side-by-side refrigerator/
freezers, placing both the refrigerator
and freezer components within
reach of everyone.
• Installing stoves with front or side
mounted controls.
• Placing standard electrical receptacles
higher than usual above the floor.
All too often, elderly or disabled people
cannot remain in their homes because
their housing no longer meets their needs.
Housing with universal features such as
those listed above allows people to age
in place if they wish. The families and
friends of disabled and elderly people
can also benefit from universal design.
Family Economics and Nutrition Review
MoblleHomes
In 1991, 14.3 million Americans, about 6 percent of the total population, lived in mobile
homes1 (10). In 1993, the number of mobile homes shipped from factories to dealers was
254,000. However, the National Association of Realtors projects that the number of mobile
homes shipped will decrease in 1995 and again in 1996 (5). Median sale price of a new
mobile home in 1993 was $27,700, up $2,200 from 1992 (9).
Those who live in mobile homes are more likely to be White, nonelderly, and own their
mobile borne. In 1991, 80 percent of mobile homes were owner-occupied, compared
with 64 percent of all housing units. The majority of mobile homes (79 percent) are located
outside urbanized areas, with very few (5 percent) located in central cities. Nationwid~.
the percentage of housing units that were mobile homes ranged from 3 percent in
the Northeast to 10 percent in the South. Although 35 percent of all U.S. housing units
were in the South, 52 percent of the mobile homes were in the South (see figure below)
(10).
With a median of lf295 square feet in 1993, new mobile homes are considerably smaller
than the median size for all new housing units (9). There is a trend toward more mobile
homes with 3-or-more bedrooms, installed central air-conditioning, and placement outside
of mobile-home communities (9).
According to the 1991 American Housing Survey, households living in mobile homes
in 1991 had monthly housing expenses (mortgage payments or rent, real estate taxes,
property insurance, homeowners' fees, fuels and utilities, and trash collection) of $297,
compared with $459 for all households (10). Many of these households bought their
homes years ago and had very small mortgage payments or had paid off their mortgages.
Median maintenance expenditure in 1991 for households living in mobile homes was
$191. Median expenditure for home improvements, repairs, and alterations among those
living in mobile homes was $837 for the 2-year period, 1990-91 (6).
1The Census Bureau defines a mobile home as "a movable dwelling 8 feet or more wide and 40 feet
Percent distribution of U.S. housing units, by region, 1991
All housing units Mobile homes
Soutoe: u.s. Department of Commerce, Bureau of the Census and U.S. Depa11ment of Housing and Urban
Development, Office of Policy Development and Research, 1993, American Housing Survey for the
United States In 1991, Current Housing Repons, H150191.
1995 Vol. 8 No.2
Those who live in
mobile homes are
more likely to be
White, nonelderly,
and own their
mobile home.
31
Table 4. Characteristics of new, privately owned, single-family houses Future Home (15)
completed, selected years, 1970-93 In June 1994, Future Horne, the first
facility in the country designed to
Characteristic 1970 1980 1990 1993 demonstrate horne automation for the
disabled, opened in Baltimore County,
Total completed
Maryland. Developed by Volunteers
for Medical Engineering at a cost of
(in thousands) 793 957 966 1,039
over $500,000 in public and private
Floor area funds, electronic and automated features
Average (sq. ft.) 1,500 1,740 2,080 2,095 include: a remote control device that
Median (sq. ft.) 1,385 1,595 1,905 1,945 signals a master computer to open doors,
shut windows, tum on the shower, make
Percent distribution telephone calls, or check the temperature.
Number of stories Also, buttons, knobs, and switches are
1 73 61 47 48 marked in Braille, and telephones have
2 or more 17 31 49 48
enlarged numbers for the visually-
Split level 10 8 4 4
impaired and voice amplification for
the hearing-impaired. To accommodate
Bedrooms wheelchairs, doorways are extra wide
2 or less 13 17 15 12 and floors are flat with no raised
3 63 63 57 58 thresholds. The bathroom features a
4 or more 24 20 29 30 wheelchair-accessible shower and push-
Bathrooms button faucets that control both water
1-112 or less 20 10 13 12 flow and temperature. Wheelchair users
2 or more 80 90 87 88 can also electronically lower shelves
Heating fuel and counters.
Gas 62 41 59 66
The technology developed for Future
Electricity 28 50 33 29 Horne is also helpful to the elderly,
Oil 8 3 5 3 who are the largest potential group of
Other or none 5 3 2 users. Features designed with the elderly
Parking facility in mind include television-based
Garage 58 69 82 84 reminders to take medication and an
Carport 17 7 2 2 "anti-wandering" alarm that summons
No garage or carport 25 24 16 14 neighbors or family by telephone when
Central air conditioning 34 63 76 78
a resident leaves and fails to return.
Source: U.S. Department of Commerce, Bureau of the Census and U.S. Department of Housing and
Urban Development, 1994, Characteristics of New Housing: 1993, Current Construction Reports
C25/93-A and U.S. Department of Commerce, Bureau of the Census, 1993, Statistical Abstract of the
United States, 1993 [113th ed.].
32 Family Economics and Nutrition Review
References
1. Callis, R.R. 1994. Housing Vacancies and Homeownership, Annual Statistics:
1993. Current Housing Reports, Hl11~3-A. U.S. Department of Commerce, Bureau
of the Census.
2. Chicago Title and Trust Family of Title Insurers. 1994. Who's Buying Homes in
America. Chicago, IL.
3. Chicago Title and Trust Family of Title Insurers. 1994. Low Interest Rates Drove
1993 Home Buying to Near Record Levels. News Release. Chicago, IL.
4. National Association of Home Builders, Public Affairs Division. 1994. Housing
Backgrounder.
5. National Association of Realtors. 1994. Real Estate Outlook: Market Trends and
Insights, Vol. I, No.4.
6. Norry, L.J. and Williams, B.T. 1994. Homeowners, Home Maintenance, and
Home Improvements: 1991. Current Housing Reports, Series H121/93-4. U.S.
Department of Commerce, Bureau of the Census.
7. U.S. Department of Commerce, Bureau of the Census. 1994. Statistical Abstract
of the United States, 1994. [I 14th ed.)
8. U.S. Department of Commerce, Bureau of the Census. 1993. Statistical Abstract
of the United States, 1993. [113th ed.)
9. U.S. Department of Commerce, Bureau of the Census. 1994. Characteristics of
New Housing: 1993. Current Construction Reports, C25/93-A.
10. U.S. Department of Commerce, Bureau of the Census, and U.S. Department of
Housing and Urban Development, Office of Policy Development and Research.
1993. American Housing Survey for the United States in 1991. Current Housing
Reports, Hl50/91.
11. U.S. Department of Housing and Urban Development, Office of Public Affairs
and Office of Intergovernmental Relations. 1988. Universal Design: Housing for
the Lifespan of All People. HUD-1156-PA.
12. U.S. Department of Labor, Bureau of Labor Statistics. 1984 and 1994. CPI
Detailed Report. January issue.
13. U.S. Department of Labor, Bureau of Labor Statistics. Conswner Expenditure Surveys:
1992 Interview Survey Documentation.
14. U.S. Department of Labor, Bureau of Labor Statistics. 1986. Consumer Expenditure
Surveys: Interview Survey, 1982-83. Bulletin 2246.
15. Valentine, P.W. 1994. The future at his fingertip. The Washington Post, June 28
issue.
1995 Vol. 8 No.2 33
34
Research Summaries
Gender-Related
Shifts in the
Distribution of
Wages
The U.S. wage distribution grew more
dispersed and unequal in the 1980's.
The middle of the distribution thinned
out, with more wage earners located at
the upper and lower ends. In addition,
earnings of women grew faster, on
average, than those of men during this
period. This analysis focuses on the .
gender-related shifts that took place m
the Nation's wage distribution in the
1980's. Income and work experience
data collected in the March supplement
to the Current Population Survey are
used to approximate the wage distribution.
Annual wage and salary earnings
received from all jobs by people 15
years of age and older who usually
worked 35 or more hours a week for
50 or more weeks in the years 1979,
1989, and 1992 were included. All
earnings are reported in 1992 dollars.
Annual employment categories used in
this analysis were: low-wage employment
(earnings less than $12,000),lowto-
middle-wage employment ($12,000
to $23,999), middle-wage employment
($24,000 to $47,999), middle-to-highwage
employment ($48,000 to $59,999),
and high-wage employment ($60,000
or more). Mean earnings of year-round,
full-time workers rose from $30,485
in 1979 to $31,728 in 1989, whereas
median earnings stayed about the same
(see table). The proportion of workers
in the middle and in the low-to-middle
wage categories declined, while that of
workers earning lower and higher
wages each increased.
Shifts in the earnings distribution for
men were more pronounced than for
workers overall. The thinning of the
middle was more severe for men than
for all workers-the middle proportion
dropped from 53 percent of all men in
1979 to 45 percent in 1989. The proportion
with earnings below $24,000 increased
from 29 percent to 35 percent,
while the proportion with earnings
above $48,000 rose from 18 percent
to 20 percent.
The earnings situation for women was
quite different. Among full-time,
year-round workers, the proportion of
women with earnings between $24,000
and $48,000 increased from 27 percent
in 1979 to 35 percent in 1989. The
proportion of women earning between
$12,000 and $24,000 declined from 57
percent to 45 percent, and the proportion
earning under $12,000 increased
slightly from 14 to 15 percent.
During the 1980's, the median wage
and salary earnings of men fell from
$32,231 to $30,549 (in 1992 dollars)
or 5.2 percent. At the same time, wage
and salary earnings of women increased
from $18,960 to $20,932 or 10.4 percent.
The proportion of women working
full time, year round rose from 43 percent
to 51 percent, partly because of the
growing proportion of college-educated
women.
Of the nearly 15 million full-time, yearround
wage and salary workers added to
the labor force between 1979 and 1989,
45 percent were men and 55 percent
were women. Nearly 4.6 million
women entered the middle-earnings
ranks between 1979 and 1989, and
nearly 1.0 million women moved into
the upper ranks (earnings over $48,000).
During this period, the number of
men in the middle-earnings category
declined by 90,000, while those earning
in excess of $48,000 increased by 2.4
million.
For young adults age 20 to 29 years
with a high school education or less, the
proportions with low-wage employment
increased from 10 to 17 percent for men
and from 19 to 28 percent for women.
For college-educated women in this
Family Economics and Nutrition Review
Distribution of wage and salary earnings of full-time, year-round workers, by gender, 1979 and 1989
(in 1992 dollars)
Total Men
Intervals 1979 1989 1979 1989
Total (in thousands) 57,209 72,120 36,277 42,987
Percent
Less than $12,000 8.4 10.5 4.9 7.2
$12,000 to $23,999 36.2 34.6 24.0 27.4
$24,000 to $35,999 26.2 25.0 29.1 25.3
$36,000 to $47,999 17.4 15.9 24.3 19.6
$48,000 to $59,999 5.9 7.1 8.7 9.9
$60,000 to $71,999 2.4 2.6 3.7 3.9
$72,000 to $83,999 1.4 1.4 2.1 2.1
$84,000 and over 2.1 2.9 3.2 4.5
Mean $30,485 $31,728 $36,065 $37,051
Median $26,543 $26,023 $32,231 $30,549
Source: Ryscavage, P., 1994, Gender-related shifts in the distribution of wages, Monthly Labor Review 117(7):3-14.
age group, there was a large relative
increase into middle-wage employment
from 38 to 53 percent. Among young
college-educated men, there was a
modest decline in middle-wage employment
from 64 to 60 percent.
Of college-educated men age 30 to 54,
there was a slight increase in the percentage
of low-to-middle-wage employment
and a rise from 43 to 48 percent in
the proportion in middle-to-high-wage
and high-wage employment. Among
college-educated women in this age
group, most remained concentrated
in middle-wage employment, but the
proportions in middle-to-high- and
high-wage employment rose from 7
to 15 percent.
The shift in employment from industries
involved in the production of goods
to industries that provide services is
considered a primary contributor to
growing wage inequality. For men,
1995 Vol. 8 No. 2
43 percent of the employment gain
between 1979 and 1989 was in lowpaying
service-producing industries.
For women, 29 percent of the employment
gain was in low-paying serviceproducing
industries, whereas 61 percent
occurred in high-paying serviceproducing
industries.
The onset of the recession in 1990
changed the labor market situation.
The impact of the recession was felt
more strongly by men than by women,
with men's employment level declining
by 900,000 between 1989 and 1992.
During this period, employment for
women increased by 1.9 million.
Source: Ryscavage, P., 1994, Gender-related
shifts in the distribution of wages, Monthly Labor
Review 117(7):3-14.
Women
1979 1989
20,932 29,133
14.4 15.3
57.2 45.3
21.2 24.6
5.4 10.3
1.0 2.8
.3 .8
.2 .3
.3 .6
$20,816 $23,874
$18,960 $20,932
35
Changing Eating
Patterns: Grains,
Vegetables, Fruit,
and Sugars
Although diets are changing, a substantial
gap remains between public health
recommendations and consumers'
practices. Americans have shifted'to a
lower fat, higher carbohydrate diet in
the past decade but are still not eating
the amounts of high-fiber foods that
are recommended in the latest dietary
guidance. Furthermore, Americans are
eating more foods that contain large
amounts of refined sugar. This report,
using data from the U.S. Food Supply
Series, focuses on the change in American
eating patterns with regard to grain
products, legumes, vegetables, fruit,
caloric sweeteners, and beverages.
Whole-grain products, legumes, vegetables,
and whole fruit are high in fiber,
vitamins, and minerals-and contain
little added sugar. These foods are
consumed in relatively low amounts,
compared with more processed foods
that are stripped of fiber.
According to USDA surveys, the average
intake of carbohydrates increased from
43 to 49 percent of caloric intake
between 1977-78 and 1989-91. The
American Cancer Society and the
American Heart Association, among
others, recommend that the carbohydrate
content of the diet be increased
to 55-60 percent of the caloric intake.
The Daily Reference Value (daily values
on food labels) for total carbohydrates
is calculated as 60 percent of calories,
or 300 grams in a 2,000-calorie diet.
The Daily Reference Value for fiber is
based on a recommended intake of 25
grams in a 2,000-calorie diet.
36
Nationwide surveys demonstrate that
consumers are using nutrition labels in
making food selections; yet, only one
in four people consistently consider the
information about carbohydrates or
fiber. Concern about fiber in the diet
has changed little since 1985, never
exceeding 5 percent of the population.
The USDNDHHS Food Guide Pyramid
suggests that daily intake include 2-4
servings of fruit, 3-5 servings of vegetables,
and 6-11 servings of grain productsincluding
several servings of whole-grain
products. Also, frequent use of legumes
as meat alternates or as starchy vegetables
is recommended.
Grain Products
Americans have increased their consumption
of grain products in recent
years. From an annual average of 135
pounds in 1970-74, per capita use of
flour and cereal products increased to
146 pounds in 1980-83 and 187 pounds
in 1992.
Wheat is the major grain eaten in the
United States; wheat flour and other
products represented 74 percent of total
grain consumption in 1992. With increased
consumption of rice, com, and
oat products since 1980-83, however,
wheat's share of total grain consumption
declined 6 percentage points.
Despite the 28-percent increase in per
capita consumption from 1980-83 to
1992, average grain consumption is
still below recommended levels. Many
people continue to think that starchy
foods, such as bread and potatoes, are
fattening, but starches provide only
about 4 calories per gram, whereas fat
provides about 9 calories per gram.
Most calories in fact come from the
foods eaten with starchy foods, such as
butter or margarine, sour cream, gravy,
and jam or jelly. Six servings from the
bread and cereal food group represent
just over 20 percent of the day's total
on a 2,000-calorie diet.
In a study sponsored by the Food Marketing
Institute (FMI) and Prevention
magazine, only 14 percent of shoppers
reported eating more fiber in 1993 than
in 1990; in another study conducted for
the American Dietetic Association, only
15 percent had increased their consumption
of grains, cereal, or fiber to achieve
a more healthful diet.
Fruit and Vegetables
Consumption of fruit and vegetables
increased about 1 0 percent in the past
decade; vegetables accounted for most
of the increase, on a farm-weight basis.
The variety and seasonal availability
of fresh fruit and vegetables have expanded
because improved refrigeration
and transportation have created global
markets.
Prepared salads and salad bars have
become more popular over the past
decade. Restaurants introduced salad
bars in the late 1970's, and supermarket
chain stores added salad bars during
1982-84 and, more recently, a wide
array of prepared salads, as well. Fastfood
chains now offer prepackaged
salads.
Consumption of fresh fruit rose 15
percent above the 1980-83 annual
average. Americans' favorite fresh
fruit is bananas, followed by apples,
watermelons, oranges, cantaloupes, and
grapes. The per capita consumption of
apple juice has increased 47 percent,
accounting for 41 percent of total apple
.consumption-on a farm weight basisin
1992.
Beans and other legumes have returned
to the American culinary mainstream.
Dry bean, pea, and lentil use increased
26 percent, from 6.4 pounds per person
a year during 1980-83 to 7.9 pounds by
1992.
Family Economics and Nutrition Review
Although fruit and vegetables have
become more popular and available in
the past decade, consumption remains
well below the levels recommended
by government and health authorities.
More than a fourth of the population ate
no fruit and drank no fruit juice during
the 3 consecutive days of recordkeeping
in a 1989-90 USDA food intake survey.
Several factors influence the consumption
of fruit and vegetables. Between
1980 and 1992, retail prices more than
doubled (up 109 percent) for fresh produce,
while costs of other food items
rose at a much slower rate. Moreover,
consumers tend not to be aware of the
importance of eating recommended
amounts. Also, consumers' desire for
greater convenience has created a trend
toward drive-thru, carryout, and homedelivered
meals that has diminished the
popularity of salad bars at many fastfood
places.
Industry has responded by adding
convenience to the produce department.
New products and services have been
introduced, such as pre-cut fruit and
vegetables-often prepared for stirfrying
or microwaving with preparation
instructions provided. Packaged, freshcut
salad mixes have experienced a
93-percent increase in sales in 1 year.
Caloric and Low-Calorie
Sweeteners
Between 1980-83 and 1992, total per
capita use of caloric sweeteners rose
16 percent. Each American consumer
averages more than one-third pound of
added sugars a day.
The consumption of specific sugars has
changed over the past decade. Sucrose's
share in total caloric sweetener consumption
dropped from 62 percent in
1980-83 to 45 percent in 1992, while
com sweetener's share increased from
37 percent to 54 percent during the
same period. Most of the increase in
com sweeteners reflects an increased
1995 Vol. 8 No.2
use of high-fructose com syrup (HFCS),
which is significantly less expensive
than sucrose. Use of HFCS rose from
18 pounds per person in 1980 to 52
pounds in 1992. The low-calorie sweetener
(mainly aspartame and saccharin)
market grew from less than 6 percent of
the total sweetener market in 1980 to 15
percent in 1992; per capita use has more
than tripled since 1980.
Nutritionists are concerned about the
rise in consumption of fructose. Each
American now consumes significantly
more added fructose than in 1980.
Sucrose is half fructose; HFCS is 42
to 55 percent fructose. Evidence implicates
¢.ets high in fructose with increased
blood lipid levels. Some researchers
suggest that although there is no conclusive
evidence that a high sugar intake
is a risk factor for heart disease in the
general population, a small number of
"carbohydrate-sensitive" individuals
may be particularly sensitive to sugar
(especially fructose) and respond with
raised cholesterol and triglyceride
levels. Research in this area is ongoing.
Food processors are introducing new
products with "no added sugar" and
"reduced sugar." New sweeteners will
likely enter the market in the next
decade. In addition, the new food label,
which lists the amount of sugars in a
serving of food, can help those who are
trying to moderate their sugar intake.
Beverages
U.S. per capita soft drink consumption
increased 29 percent between 1980 and
1992. Supermarket customers spend
more money on carbonated soft drinks
than any other product scanned at the
checkout counter, excluding meat and
poultry. Soft drink makers have created
huge marketing and promotional campaigns
to encourage this growth. Sales
of fountain drinks are increasing because
they are promoted with "combomeals"
at fast-food places and because
the fast-food drink sizes have increased.
Per capita consumption of alcoholic
beverages declined steadily between
1981 and 1992. Annual average beer
consumption declined 11 percent, wine
consumption declined 18 percent, and
distilled spirits declined 50 percent.
These trends for decreased alcohol
consumption may reflect the smaller
percentage of the population at peak
drinking age, since the proportion of
the population over age 60 is increasing
and less likely to indulge. Sharply
higher Federal excise taxes on alcoholic
beverages beginning in 1991 may have
decreased alcoholic beverage consumption.
Furthermore, retail prices (as
measured in the Consumer Price Index
(CPI)) for packaged alcoholic beverages
increased 15.6 percent between
1990 and 1992, compared with a 3.4-
percent increase for the CPI for food
at home.
Putnam, J.J., 1994, American eating habits
changing: Part 2, Grains, vegetables, fruit, and
sugars, FoodReview 17(2):36-48.
Dietary Guidelines for
Alnerieau
• Bat a vQiy of foods
• Maintain healthy weight
• Choose e diet low in fat,
saturated fat, and cholesterol
• Cboole a diet With plenty of
vegemb1es, fruits, and grain
products
• Use sugars only .in moderation
• Use salt and sodium only in
moderation
• If you drink. alcobolic beverages,
do so in moderatiol)
37
The Development
and Growth of
Employer-Provided
Health Insurance
According to the Current Population
Survey conducted by the Bureau of the
Census for the Bureau of Labor Statistics
(BLS), about 35.7 million people under
the age of 65 were not covered by health
insurance in 1990. This is an increase of
2 million persons since 1988. Because
of this increase, as well as rising costs
for health care services, interest has
intensified in reforming the health care
system. In recent years, the Congress
has introduced many bills that attempted
to improve access, reduce the cost, and
modify the tax treatment of health care
benefits. In 1994, efforts focused on
changing the national health care system,
which relies heavily on health insurance
provided by employers.
Early Coverage
The earliest coverage for health services
began in 1798, when the Congress established
the United States Marine Hospital
Services for seamen. Compulsory
deductions for hospital services were
made from the salaries of seamen.
At first, most insurance policies protected
against lost income due to accidents,
rather than covering health services. In
1850, the Franklin Health Assurance
Co. of Massachusetts provided the first
accident policy. Travelers Insurance Co.
entered the field in 1863 and developed
accident insurance that provided a death
benefit and a weekly disability benefit.
During the 1870's and 1880's, companies
in the mining, lumber, and railroad
industries developed plans that covered
medical services. Group industrial clinics
were established to provide medical
care to employees for industrial accidents
and common illnesses. Many historians
credit the growth in health insurance to
38
the growing industrialization of America.
Employers and labor unions realized
that workers needed economic protection
against unforeseeable losses, which
can result from premature death and
disability.
In 1899, the Aetna Life Insurance Co.
and Travelers Insurance Co. offered a
new type of health plan that provided
coverage against loss due to temporary
total disability resulting from certain
diseases. This coverage was issued to
select, preferred-risk residents of towns
with a population of 5,000 or more. By
1908, most of the restrictions on these
plans were eliminated: most diseases
were no longer excluded, the premium
rate was abandoned, the 7-day waiting
period was no longer in effect, and a
medical examination was no longer
required for insurance.
In 1910, Montgomery Ward and Co.
sought a plan for its employees to
protect them from financial loss due to
illness or injury. The plan is regarded as
the Nation's first group health insurance
policy. The policy provided weekly
benefits equal to one-half of the employee's
weekly salary, with a minimum
benefit of $5 and a maximum of $28.85
per week, if the employee was unable
to work due to illness or injury. These
benefits were paid directly to the employee;
the company did not reimburse
for medical services.
Early in the 20th century, many people
advocated compulsory health insurance.
Proponents of compulsory insurance
sought to achieve two goals. First, they
wanted to "relieve poverty caused by
sickness by distributing individual
wage losses and medical costs through
insurance." Second, they wanted to
"reduce the social costs of illness by
providing effective medical care and
creating monetary incentives for disease
prevention." However, unions, physicians,
and insurance companies mounted
strong opposition to compulsory
insurance and all resolutions brought
before the U.S. House of Representatives
were defeated.
Blue Cross and Other Plans
Significant development in health insurance
took place in the 1930's because
of the Depression. Few people could
pay for hospital care, so most hospitals
were in serious fmancial straits. In the
first years of the Depression, more than
100 hospitals nationwide had failed,
and those that remained in business had
only about a SO-percent occupancy rate.
In 1929, a group of teachers at Baylor
University and the University Hospital
in Dallas, Texas, made arrangements to
provide coverage for room and board
and for specified ancillary services for
21 days at an annual premium of $6 per
teacher. This development is considered
a forerunner of what is now known as
Blue Cross.
Later came citywide plans with more
than one hospital. Individuals contributed
a small amount monthly to a
central fund that was redistributed to
participating hospitals. This fund
allowed hospitals to remain solvent
with payment of hospital bills guaranteed,
although coverage for dependents
was excluded. By 1935, 19 plans had
been created in 13 States.
Prepayment plans to cover physicians'
services (Blue Shield) paralleled the
development of Blue Cross plans. The
first Blue Shield plan-the California
Physicians Service-was founded in
California in 1939 and provided
physician services to employee group
members for $1.70 per month for each
member of the group. However, the
plan was limited to employees earning
less than $3,000 annually.
On the West Coast, at the same time
Blue Cross and Blue Shield plans were
being formed, Health Maintenance
Organizations (HMO's) were developing.
HMO's provide a wide range of
comprehensive health care services to
subscribers for a predetermined rate.
The largest and most widely known
HMO that was formed during the
1930's was Kaiser Permanente.
Family Economics and Nutrition Review
Employer-Provided Health
Insurance
By 1940, the population of the United
States was 132 million, but only 12
million people were covered by health
insurance. Blue Cross/Blue Shield
dominated the market with 50 percent
of those individuals with coverage,
followed by commercial insurance with
31 percent, and other plans including
HMO's, with 19 percent.
In 1942, the Congress enacted the
Stabilization Act, which limited the
amount of employer wage increases but
permitted the adoption of employee
insurance plans. This stimulated the
growth of plans through collective
bargaining agreements.
In 1949, the Liberty Mutual Insurance
Company introduced major medical
insurance to supplement basic medical
care expenses. Basic plans usually
cover facility and physician care in the
hospital. Major medical plans protect
individuals against extended illnesses
or injuries by providing coverage for
services not included in a basic plan
plus supplemental benefits after coverage
under the basic plan has expired.
Comprehensive major medical plansrather
than offering "additional" coverage
to a basic plan-cover a wide range
of medical services in one package.
Managed Care Plans
Between December 1971 and Decembe