Family Economics and Nutrition Review [Volume 9, Number 1]

              \,[NTEJrfOR NUTRITION POLICY AND PROMOTION
Feature Articles
2 Income and Spending of Poor Households With Children
MarkLino
14 Demographic and Economic Determinants of Household
Income Polarization Among the States in America
Mohamed Abdel-Ghany
Factors Influencing Rural Southern Elders' Life Satisfaction
Julia M. Dinkins and Retia Scott Walker
Research Summaries
40 Cholesterol Measurement
42
47 Measuring Years of Healthy Life
49 Optimal Calcium Intake
Regular Items
52 Charts From Federal Data Sources
54 Recent Legislation Affecting Families
55 Research and Evaluation Activities in USDA
58 Data Sources
59 Journal Abstracts
60 Cost of Food at Home
61 Consumer Prices
62 Guidelines for Authors
Dan Glickman, Secretary
U.S. Department of Agriculture
Ellen Haas, Under Secretary
Food, Nutrition, and Consumer Services
Eileen Kennedy, Executive Director
Center for Nutrition Policy and Promotion
Jay Hirschman, Director
Nutrition Policy and Analysis Staff
Editorial Board
Mohamed Abdel-Ghany
University of Alabama
Rhona Applebaum
National Food Processors Association
Johanna Dwyer
New England Medical Center
Jean Mayer USDA Human Nutrition Research Center
on Aging at Tufts University
Helen Jensen
Iowa State University
Janet C. King
Western Human Nutrition Research Center
U.S. Department of Agriculture
C.J. Lee
Kentucky State University
Rebecca Mullis
Georgia State University
Suzanne Murphy
University of California-Berkeley
Donald Rose
Economic Research Service
U.S. Department of Agriculture
Ben Senauer
University of Minnesota
Laura Sims
University of Maryland
Retia Walker
University of Kentucky
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 $8.00 per year ($1 0.00
for foreign addresses). Send subscription
orders and change of address to Superinten­dent
of Documents, P.O. Box 371954,
Pittsburgh, PA 15250-7954. (See subscrip­tion
form on p. 63.)
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)~16.
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) 720-2791 . To file a complaint, write
the Secretary of Agricu~ure, U.S. Department
of Agricu~ure, Washington, DC 20250, or call
(202) 720-7327 (voice) or (202) 720-1127
(TDD). USDA is an equal employment
opportunity employer.
Center for Nutrition Policy and Promotion
Feature Articles
2
14
24
Income and Spending of Poor Households With Children
MarkLino
Demographic and Economic Determinants of Household
Income Polarization Among the States in America
Mohamed Abdel-Ghany
Factors Influencing Rural Southern Elders' life Satisfaction
Julia M. Dinkins and Retia Scott Walker
Research Summaries
40 Cholesterol Measurement
42 The Effects of Health Insurance on Consumer Spending
47 Measuring Years of Healthy Life
49 Optimal Calcium Intake
Regular Items
52
54
55
58
59
60
61
62
Charts From Federal Data Sources
Recent Legislation Affecting Families
Research and Evaluation Activities in USDA
Data Sources
Journal Abstracts
Cost of Food at Home
Consumer Prices
Guidelines for Authors
Volume 9, Number 1
1996
2
Feature Articles
Income and Spending of Poor
Households With Children
By Mark Lino
Economist
Center for Nutrition Policy and Promotion
This study examines the income and spending of poor households with
children using data from the 1990-92 Consumer Expenditure Survey. Poor
households were defined as those whose income and total expenditures fell
below the poverty threshold. The majority of poor households were headed
by a single parent, and the majority of the heads of poor households did
not have a high school diploma. Food stamps was the most often received
income source of these households and made up 21 percent of their before­tax
income. Housing, food, and transportation accounted for approximately
78 percent of the total expenditures of poor households. Although these
budgetary components accounted for a high proportion of total expenditures,
83 percent of these households did not own a home, and 45 percent did not
own a vehicle. Implications of the results of this study for policy and program
purposes are discussed.
oor households with children
are one of the most vulner­able
groups in the U.S.
population. Their reduced
economic state affects not only their
current situation but also the future
prospects of their children. Past research
has tended to focus on the income of
these families. Little attention has been
devoted to their allocation of resources.
In order to provide a more complete
picture of the economic situation of
these households, this study examines
the expenditures of these households
as well as their income. In doing so, it
addresses a gap in the economics litera­ture
on poor households and should be
of use to policymakers and professionals
concerned with these families.
Data Source
Data used in this study are from the
interview component of the 1990-92
Consumer Expenditure Survey (CE),
conducted by the Bureau of the Census
for the Bureau of Labor Statistics. The
CE is an ongoing survey that collects
data on expenditures, income, and
major sociodemographic characteristics
of consumer units (for this study, the
term consumer unit will be used inter­changeably
with household). A national
sample of consumer units, representing
the civilian noninstitutionalized popula­tion,
is interviewed over the course of
a year. The 1990-92 survey contains
information from approximately 60,000
interviews.
Family Economics and Nutrition Review
There is a rotating sample design: each
quarter, a portion of the sample consists
of new consumer units introduced to
replace consumer units that complete
their participation in the survey. Each
quarter is deemed an independent sample
and is treated as such to incorporate the
weights. Data from each quarter were
therefore aggregated and expenditures
annualized.
Households with at least one child under
age 18 in the home and that were com­plete
income reporters were selected for
analysis. Complete income reporters are
households that had provided values for
major sources of income, such as wages
and salary, food stamp benefits, and
Social Security; however, even com­plete
income reporters may not have
provided a full accounting of all income
from all sources. Approximately 86
percent of households surveyed in the
1990-92 CE were complete income
reporters. The unweighted sample of
complete income reporters consisted
of 18,327 households with children; of
these, 1,625 were deemed to be poor.
Data were weighted to represent the
population of interest.
To place poor households with children
in perspective, nonpoor households with
children were also analyzed. Tests of
statistical significance (Chi-square and
t-tests) were performed between the
two groups using unweighted data and
reported at the .01level. The .01 level
of statistical significance was selected
rather than the more traditional .05 level
to compensate for any possible cluster­ing
effect present in the data. Almost
all differences in characteristics, income,
and expenditures between the two groups
were statistically significant at the .01
level; hence, all comparisons are signifi­cant
unless noted.
1996 Vol. 9 No.1
Defining Poor Households
To study poor households, the first step
is to defme "poor." Typically, having
an income below the U.S. poverty
threshold (the weighted average thresh­old
differs by household size) has been
used as the definition. This definition
poses problems, especially with the CE,
because of nonreporting and under­reporting
of various sources of income
(and because no income imputation is
made for nonresponses in the CE).
As the average income of CE families
in the lowest income quintile is below
that found in Census reports and their
total expenditures are twice their income
( 10,12), it is likely that poor families
in the CE either do not report certain
sources of their income or they under­report
them. Although part of the
expenditure-income disparity may
reflect purchasing on credit, it seems
unreasonable that such a large amount
of credit could be obtained. Using
solely an income measure with the CE
would likely result in many households
being classified as poor, when in fact
they are not. Some other definition for
poor households is needed. Two other
definitions that have been used by
researchers involve total expenditures
and receipt of various forms of public
assistance.
The use of total expenditures as a proxy
for income to gauge households that fall
below the poverty threshold has some
support in the economics literature. The
permanent income hypothesis suggests
that people smooth out their consump­tion
over their lifetime based on their
estimated lifetime income ( 3 ). Whereas
annual income is subject to transitory
shocks, such as temporary unemploy­ment,
annual consumption or total
expenditures are not likely to vary as
much and therefore may be viewed as a
measure of estimated lifetime income.
A study by McGregor and Borooah (6)
found that a total expenditure-based
measure, as opposed to an income­based
measure, was a better indicator of
poor households based on criteria such
as ownership of consumer durables.
This measure, however, failed to ac­count
for families with children in the
CE who had low expenditures and a
high savings rate; some families were
putting money aside for future retire­ment,
a new home, and/or children's
education. For these families, their
expenditures may have fallen below
the poverty threshold, even though their
income did not-so they were not what
is usually regarded as poor households.
Receipt of public assistance is another
possible way to identify poor households.
To receive various forms of public
assistance, such as Aid for Families
with Dependent Children (AFDC) or
food stamps, a household must meet
some set low-income criteria. Receipt
of money or in-kind benefits from one
or more of these welfare programs there­fore
would seem to be a reasonable way
to identify poor households. However,
many poor households that are eligible
to receive various forms of public assis­tance
do not apply for them (2). They
may be unaware of their eligibility or
if they are aware, they choose not to
apply. Analysis of the CE data con­firmed
this. Some households with both
low income and low total expenditures
did not receive any forms of public
assistance.
3
4
The majority
(52 percent) of
heads of poor
households did
not have a high
school diploma;
only 2 percent had
a college degree.
Given the problems with income under­reporting
in the CE and with various
measures of low income, this study
used a measure based on both before­tax
income and total expenditures to
define "poor" households. Their income
substantially exceeded their expenditures.
Specifically, households were defined
as poor if their before-tax income and
total expenditures fell below the poverty
threshold. The use of both income and
expenditures alleviates the problems
associated with using either individually.
Households that underreport their in­come
such that it fell below the poverty
threshold would not be categorized as
poor if their expenditures were above
the poverty threshold. Similarly, house­holds
with low expenditures and an
income above the poverty threshold
would not be categorized as poor.
It should be noted that the definition of
poor used in this study is rather strict.
Of the households with children in the
sample, 9 percent were classified as poor.
By comparison, during the 1990-92
period, 16 to 18 percent of families with
children were classified as being in
poverty according to a Census report
( 11 ). In addition, this definition of poor
households (before-tax income and total
expenditures below the poverty thresh­old)
may include some nonreporters or
underreporters of income with low
expenditures.
Characteristics
The characteristics of poor households
in this study are similar to those ob­tained
in Census reports ( 11) and there­fore
will only be briefly discussed and
compared with nonpoor households.
Average age of the household head 1 for
poor households with children was 34
and for nonpoor households, 37 (table 1).
Average household size was 4.4 for poor
households. The average size of non­poor
households was 3.9. The majority
of poor households with children (52
percent) were composed of a single
parent (of whom 97 percent were mothers)
and their children onli The actual
proportion of single-parent households
in the poor population is likely higher
since single parents residing with ex­tended
family members are included
in the "other" category. In contrast,
74 percent of nonpoor households with
children were composed of a married
couple and their children only.
The majority (52 percent) of heads of
poor households did not have a high
school diploma; only 2 percent had a
college degree. For nonpoor households
with children, 15 percent of heads did
not have a high school diploma and
27 percent had a college degree. Fifty­seven
percent of poor households with
children were White and 43 percent
were non-White; 21 percent were
Hispanic (and could be of any race).
A higher proportion of nonpoor house­holds
with children were White (86
percent) and a lower proportion were
Hispanic (10 percent). A higher percentage
of poor households with children resided
in the urban Midwest (31 percent) than
in other areas.2 In the CE, urban areas
may be identified by region, but rural
areas are for the overall United States.
1The household head is defmed as the person
who owns or rents the home; in cases where there
is joint ownership or renting status, the head is
arbitrarily decided so is actually a co-head.
2Urban areas are defined as Metropolitan
Statistical Areas (MSA's) and places outside
an MSA of 2,500 or more people; rural areas are
places of fewer than 2,500 people outside an MSA.
Family Economics and Nutrition Review
Table 1. Characteristics of poor and nonpoor households with children,*
1990-92
Characteristic
Average age of head 1
Average household size
Household type
Husband-wife
Single parent (divorced/separated)
Single parent (never married)
Single parent (widowed)
Othe?
Education of head
No high school diploma
High school diploma
Some college
College degree
Race
White
Black
Other
Etbnicity*
Hispanic
Non-Hispanic
Region3
Urban
Northeast
South
Midwest
West
Rural
Poor
34
4.4
30
27
24
1
18
52
33
13
2
57
39
4
21
79
15
22
31
21
11
Percent
Non poor
37
3.9
74
12
3
1
10
15
32
26
27
86
11
3
10
90
17
21
27
21
14
1 The household head is defmed as the person who owns or rents the home; in cases where there is joint
~wnership or renting status, the head is arbitrarily decided.
Includes husband-wife or single-parent households residing with others, and grandparents or others
~roviding primary care for children.
Urban areas are defined as Metropolitan Statistical Areas (MSA's) and places outside an MSA of
2,500 or more people; rural areas are places of fewer than 2,500 people outside an MSA.
*All differences in characteristics between poor and nonpoor households were statistically significant at
p ~ .0 I based on unweighted data.
1996 Vol. 9 No.1
Sources of Income
Poor households with children reported
income from a variety of sources (table
2, p. 6). Food stamps was the most often
received income source with 69 percent
of poor households reporting income
from this source. Given the income of
these households was below the poverty
threshold and eligibility for food stamps
is set at 130 percent of this threshold,
one would expect an even higher pro­portion
to have received food stamps.
Food stamps, however, also has an
asset qualification. 3 In addition, as
previously discussed, many families
eligible for public assistance, such as
food stamps, do not participate in these
programs. Food stamp benefits were
received by 6 percent of non poor house­holds
with children. As food stamp
eligibility is set at 130 percent of the
poverty threshold for families with
children, near-poor households would
be eligible.
Wages or salary and public assistance
were the next two most often received
income sources of poor households;
54 percent of poor households received
each of these sources. Although the
majority of poor households received
income from wages or salary, many of
the household heads worked part time
(fig. 1, p. 7). For nonpoor households
with children, wages or salary was the
most often received income source,
received by 94 percent of these
households.
3 Assets of these families were not analyzed
because the CE does not contain detailed asset
data.
5
Income from alimony, child support,
or regular contributions4 was received
by 14 percent of poor households with
children. Since more than half of these
households were single-parent house­holds
and child support is included in
this source, this proportion may seem
low. Many single parents with children,
however, do not have child support
awards,s and even when they do, the
full amount due is often not paid ( 4 ).
Twenty-four percent of poor households
received income from other sources,
which includes income from pensions,
Supplemental Security income, unem­ployment
compensation, or owned
businesses. Eight percent received
Social Security income (which includes
disability insurance payments), but only
2 percent received interest or dividend
income. By comparison, 30 percent of
nonpoor households had interest or
dividend income.
Average Income
Before-tax income of poor families with
children averaged $8,633 and per capita
income averaged $1,962 (table 3). After­tax
total and per capita income were
slightly higher than before-tax income,
probably because of the Earned Income
Tax Credit that provides a direct grant to
households whose credit exceeds their
tax liability. The after-tax per capita
income of nonpoor households was
'7hese three income sources are combined in the
CE public use tape; "regular contributions" are
periodic payments from a nongovernment, non­household
source, such as extended family.
5The reasons for single mothers not having a child
support award are, in order of prevalence: Did not
want award, did not pursue award, other reasons,
father unable to pay, father could not be located,
other settlement/father in household, and final
agreement pending ( 4 ).
6
Table 2. Percentage of poor and nonpoor households with children
with income source,* 1990-92
Income source Poor Nonpoor
Wages or salary 54 94
Public assistance 54 4
Food stamps 69 6
Alimony, child support, or regular contributions 1 14 11
Interest or dividends 2 30
Social Security 8 4
Other2 24 30
1Regular contributions are periodic payments from a nongovernment, nonhousehold source.
2Includes income from pensions, Supplemental Security Income, unemployment compensation, or
owned businesses.
*All differences in income sources between poor and non poor households were statistically significant
at p ~ .01 based on unweighted data.
Table 3. Income of poor and nonpoor households with children, 1990-92
Income source
Before-tax income*
Per capita*
After-tax income*
Per capita*
Wages and salary*
Public assistance*
Food stamps*
Alimony, child support, and regular
contributions* 1
Interest and dividends*
Social Security
Other*2
Poor Nonpoor
$8,633 $41,670
1,962 10,685
8,688 37,873
1,975 9,711
Percentage of before-tax income
35.1 86.9
26.7 0.4
21.2 0.3
2.7 1.1
0.1 1.1
4.8 0.8
9.4 9.4
1Regular contributions are periodic payments from a nongovernment, nonhousehold source.
2Includes income from pensions, Supplemental Security Income, unemployment compensation, and
owned businesses.
*Differences in dollar amounts between poor and nonpoor households were statistically significant at
p ~ .01 based on unweighted data.
Family Economics and Nutrition Review
Figure 1. Employment status1 of heads of poor and nonpoor households2 with children,* 1990-92
Poor Non poor
71%
15%
4%
Employed full time, full year • Employed part time, full year • Employed full time, part year
• Employed part time, part year • Not working
1 Full-time, full-year employment is defined as working 35 or more hours per week, 50 or more weeks per year, including any time off with pay.
Part-time, full-year employment is working less than 35 hours per week for 50 or more weeks per year, including any time off with pay. Full-time,
part-year employment is working 35 or more hours per week for less than 50 weeks per year, including any time off with pay. Part-time, part-year
employment is working less than 35 hours per week for less than 50 weeks per year, including any time off with pay.
2The household head is defined as the person who owns or rents the home; in cases where there is joint ownership or renting status, the head is
arbitrarily decided.
*Difference in employment status between poor and' non poor households was statistically significant at p 5. .01 based on unweighted data.
approximately five times that of poor
households. The dollar amounts
received from each source of income,
except Social Security, were signifi­cantly
different for poor than nonpoor
households.
Figure 1 shows the employment status
of poor household heads: 15 percent
were employed full time,6 33 percent
were employed part time (28 percent
were considered part time because they
worked part of the year),? and 52 percent
were not employed. When employment
status of poor household heads and re­ceipt
of wages or salary by households
6pull-time employment is defined as working
35 or more hours per week, 50 or more weeks
per year, including any time off with pay.
1996 Vol. 9 No.1
are compared, a higher percentage of
poor households received wages or
salary than had an employed household
head. This difference probably indicates
another person(s) in these households,
such as a spouse or older children, was
employed-and not the household head.
Of the household heads not employed,
most (see table at right) reported not
working because they were taking care
of their family; only a small percentage
reported they could not fmd work.
7Part-time employment includes working: (I) part
time for the full year (working less than 35 hours
per week for 50 or more weeks per year, including
any time off with pay), (2) full time for part of the
year (working 35 or more hours per week for less
than 50 weeks per year, including any time off
with pay), and (3) part time for part of the year
(working less than 35 hours per week for less than
50 weeks per year, including any time off with
pay).
Reason head of poor
households with
children not employed
Taking care of family
Illness
Could not find work
Other (includes retired
and going to school)
Percent
65
18
8
9
Most heads in nonpoor households
worked full time (71 percent) or part
time (22 percent). Again, for these
households, the discrepancy between
the percentage of heads who were
7
employed and the percentage of house­holds
reporting wage or salary income
is likely because a spouse-and/or older
child-was employed.
Wages and salary accounted for the
largest share (35 percent)8 of before-tax
income for poor households. Public
assistance and food stamps made up the
next largest shares (27 and 21 percent,
respectively). Alimony, child support,
and regular contributions provided only
3 percent of income; in dollar terms this
amounted to about $230. For those re­ceiving
this income source, the average
amount received by families with two
children was $1,670. This amount was
low compared with estimates of the
"cost of raising a child" --expenditures
on two children in the average single­parent
household ranged from $7,430
to $11,080 in 1991 (5). The bulk of
income (87 percent) for non poor house­holds
was derived from wages and
salary.
The incomes of the household groups
examined do not include the value of
some noncash benefits, such as medi­caid
and public housing. These benefits
would raise the effective income of
poor households. A study by the Census
Bureau found that the poverty rate in
1990 declined when various noncash
benefits were taken into account (9).
However, even with these benefits,
the income of poor households
remained low.
8Households with and without income from a
particular source were used to calculate percent
shares from that source.
8
Table 4. Percentage of poor and non poor households with children by
expenditures incurred, 1990-92
Expenditures Poor Non poor
Housing 100 100
Food 100 100
At home* 99 100
Away from home* 50 91
Transportation* 80 99
Clothing* 85 96
Health care* 32 84
Entertainment* 69 95
Personal care* 41 81
Education or reading* 39 82
Child care* 7 31
Home furnishings or equipment* 50 80
Alcoholortobacco* 51 69
Retirement or pensions* 56 97
Miscellaneous* 1 34 81
1Includes life insurance, cash contributions, finance charges excluding mortgages and vehicles, and
occupational expenses.
*Differences in expenditures incurred between poor and nonpoor households were statistically
significant at p ~ .01 based on unweighted data.
Expenditures
All households with children, regardless
of income, reported housing and food
expenditures9 (table 4). For a description
of these and other expenses, see box.
Half of poor households with children
reported food-away-from-home expenses,
compared with 91 percent of nonpoor
households. Eating out is probably a
luxury for many poor households.
9rt should be noted that in the CE data larger
expenditures are more likely to be remembered
than smaller expenditures; therefore, these larger
expenses are likely to be reported with more reli­ability
than smaller ones. For example, a house­hold
is lilcely to remember their monthly rent, but
may forget some of the food items they purchased
in a given month.
Yang and Basi otis ( 13) found income
to be positively related to food-away­from-
home expenditures.
A much smaller proportion of poor than
nonpoor households reported out-of­pocket
health care expenses (32 vs. 84
percent). If poor households have access
to employer-provided insurance, they
may not incur health care expenses out­of-
pocket. However, approximately half
of heads of poor households were unem­ployed.
Some may receive free medical
care through medicaid; a Census Bureau
study found that in 1987-89, one-third of
people with incomes below the poverty
threshold were covered by medicaid
Family Economics and Nutrition Review
Description of Expenditures
1. Housing: Shelter (mortgage interest, property taxes, or rent;
maintenance and repairs; and home insurance) and utilities
(gas, electricity, fuel, telephone, and water). It should be noted
that for homeowners, housing expenses do not include mortgage
principal payments.
2. Food: Food and nonalcoholic beverages purchased at grocery
stores, convenience stores, and specialty stores including purchases
with food stamps; dining out at restaurants; and household expendi­tures
on school meals.
3. Transportation: The net outlay on purchase of new and used
vehicles, vehicle finance charges, gasoline and motor oil, vehicle
maintenance and repairs, vehicle insurance, and public
transportation.
4. Clothing: Apparel items; footwear; and clothing upkeep services
such as dry cleaning, alteration and repair, and storage.
5. Health care: Medical and dental services not covered by insurance,
prescription drugs and medical supplies not covered by insurance,
and health insurance premiums not paid by employer or other
organization.
6. Entertainment: Fees and admissions, televisions, radios and
sound equipment, and services.
7. Personal care: Appliances for personal care use, such as
electric shavers; haircuts; and cosmetics.
8. Education and reading: Tuition, books, supplies, and other
fees for elementary school, high school, and college, as well as
newspapers and magazines.
9. Child care: Day care outside the home and baby-sitting or home
care for children.
10. Home furnishings and equipment: Furniture, floor coverings,
major appliances, and small appliances.
I 1. Alcohol and tobacco: Alcoholic beverages purchased at stores
and restaurants, and cigarettes and other tobacco products.
12. Retirement and pension: Deductions for Social Security, pri­vate
pensions, and self-employment retirement plans.
13. Miscellaneous: Life insurance, cash contributions, finance
charges excluding mortgages and vehicles, and occupational
expenses.
throughout this period (8). Also, some
households may. go without medical
care.
Child-care expenses were incurred by
7 percent of poor households and
99 percent). A smaller proportion of
poor than nonpoor households incurred
entertainment, personal care, and educa­tion
or reading expenses. Poor house­holds
may consider these expenses as
luxuries given their economic status.
Average Expenditures
Total expenditures averaged $9,986
for poor households with children,
compared with $35,815 for nonpoor
households with children (table 5, p. 10).
For poor households, total expenditures
exceeded their after-tax income by 13
percent. The difference may be caused
by underreporting of income, incurring
debt or drawing on savings to cover
expenses, or rnisreporting expenses
paid by others.
31 percent of non poor households.
The relatively small percentage of poor
households with child-care expenses,
compared with the percentage having
employed heads (48 percent), may seem
surprising. However, much child care is
provided by a spouse or other relatives,
such as grandparents (7), who are likely
not paid. In addition, many employed
heads of poor households worked part
time so they may be able to be home
when their children return from school.
Children may also be latchkey children.
Fifty-six percent of poor households
reported retirement or pension expenses,
which include Social Security deduc­tions
(Social Security deductions are
considered an expense in the CE and are
not subtracted from after-tax income).
By comparison, 97 percent of nonpoor
households reported retirement or
pension expenses. Having these expenses
is related to the employment status of
adult household members with implica­tions
for their retirement years. Without
Social Security or pensions, they will
likely remain disadvantaged and on
public assistance.
Housing accounted for the largest sharelO
of total expenses for poor households
with children (37 percent, fig. 2, p. 11).
For homeowners, the shelter component
of housing includes payments of mort­gage
interest but not mortgage principal;
A smaller proportion of poor households
reported transportation expenses com­pared
with nonpoor households (80 vs.
1996 Vol. 9 No.1
10Households with and without expenses on a
particular budgetary component were used to
calculate percent shares on that component.
9
10
Food made up
the second largest
share of total
expenses for poor
households at
32 percent-double
the percentage
share of nonpoor
households.
Table 5. Expenditures of poor and nonpoor households with children,*
1990-92
Expenditures
Total expenditures
Per capita
Housing
Food
At home
Away from home
Transportation
Clothing
Health care
Entertainment
Personal care
Education and reading
Child care
Home furnishings and equipment
Alcohol and tobacco
Retirement and pensions
Miscellaneous 1
Poor
$9,986
2,270
Nonpoor
$35,815
9,183
Percentage of total expenditures
37.0 25.8
31.7 15.8
30.0 12.1
1.7 3.7
9.0 18.9
6.4 5.3
1.5 4.2
3.1 5.4
1.0 0.8
0.5 2.0
0.4 1.8
2.1 4.2
3.1 1.6
2.7 10.8
1.5 3.4
1Includes life insurance, cash contributions, finance charges excluding mortgages and vehicles, and
occupational expenses.
*All differences in dollar amounts between poor and nonpoor households were statistically significant at
p :5 .01 based on unweighted data.
mortgage principal payments are consid­ered
a reduction of liabilities in the CE
and not an expense. The effective housing
expenses of homeowners would, there­fore,
be higher than reported here. Most
poor households (83 percent) rented
their homes (fig. 3). A very small per­centage
of households stated they occu­pied
a dwelling without payment; these
people were classified as renters. Eight
percent owned with a mortgage and
9 percent owned without a mortgage.
Many of those owning without a mort-gage
resided in mobile homes, which
are much less costly than other forms of
housing. By comparison, 68 percent of
nonpoor households owned their homes.
Food made up the second largest share
of total expenses for poor households at
32 percent-double the percentage share
of nonpoor households. However, the
annual food expense of poor households
was approximately $2,500 less, even
though poor households had a larger
average household size. It should be
Family Economics and Nutrition Review
Figure 2. Expenditure shares: Poor and nonpoor households
with children, 1990-92
Poor Non poor
/~~
37%
C Housing • Transportation • Clothing . Al other
Figure 3. Housing tenure of poor and nonpoor households with
children,* 1990-92
Percent
8
Poor 9
83
60
Nonpoor 8
32
LJ Own {with mortgage) • Own {without mortgage) . Rent
*Difference in housing tenure between poor and nonpoor households was statistically significant at
P .$ .01 based on unweighted data.
1996 Vol. 9 No. I
noted that although food expenses
include the value of food stamps used,
the value of other food program benefits,
such as WIC (Special Supplemental
Nutrition Program for Women, Infants,
and Children) and free meals at school,
are not included in food expenditures.
For households that receive these
benefits, effective food expenses are
likely higher than reported here.
Transportation expenses accounted for
9 percent of the total expenditures of
poor households, compared with 19
percent for nonpoor households. This
difference may be attributable to differ­ences
in vehicle ownership between the
two groups--45 percent of poor house­holds
did not own a vehicle, whereas
only 6 percent of nonpoor households
did not own one (fig. 4, p. 12).
Other budgetary components each
made up less than 10 percent of total
expenditures for poor households.
Health care accounted for 2 percent of
total expenses, compared with 4 percent
for nonpoor households. Child care ac­counted
for less than 1 percent of total
expenses for poor households. Alcohol
and tobacco were a higher proportion
of total expenses, but a smaller dollar
amount, for poor households than for
nonpoor households. Retirement and
pensions made up 3 percent of total
expenses for poor households and 11
percent for nonpoor households. The
higher share for nonpoor households
reflects the presence of an employed
head with Social Security and pension
deductions.
11
Figure 4. Vehicle ownership of poor and nonpoor households
with children,* 1990-92
Poor Non poor
70%
• Own no vehicle • Own one vehicle L Own two or more vehicles
*Difference in vehicle ownership between poor and nonpoor households was statistically
significant at p 5 .01 based on unweighted data.
Summary and Discussion
This study examined the income and
expenditures of poor households with
children using the 1990-92 CE, thereby
filling a gap in the economics literature
on the expenditures of these families.
Because of limitations with the income
data in the CE, it was necessary to
define poor households using other
variables in addition to income. The
definition that was developed was
based on income and total expenditures
of households in relation to the poverty
threshold. This definition can be used
in future research. Comparisons of this
measure of poor households to other
measures (e.g., receipt of public assis­tance,
income in the lowest quintile)
would determine the extent of any
difference in such measures.
12
Poor households with children were
more likely to receive food stamps than
any other source of income. Food stamps
also provided about one-fifth of their
income, indicating the importance of
this Federal program to the economic
security of poor households. Probably
more than any other program, food
stamps provides a safety net for poor
households.
The total expenditures of poor house­holds
with children exceeded their after­tax
income. If households are assuming
debt to cover expenses, this debt is add­ing
to their precarious economic status.
Housing and food accounted for nearly
70 percent of total expenditures of poor
households with children, compared
with 42 percent for nonpoor households.
Although $3 out of every $8 spent went
to housing, most poor households were
renters. Therefore, they are not building
up equity in a home and are vulnerable
to rises in rental prices.
Food expenditures of poor households
were about $2,500 less than those of
their nonpoor counterparts. Poor house­holds
also had a higher average house­hold
size. There is some evidence that
lower food spending puts people at
nutritional risk ( 1 ). Research needs to
examine more closely the food situation
of poor households to see how their
lower food spending affects their diet.
Many poor households do not own a
vehicle. This limits their job opportu­nities.
When designing policies and
programs aimed at moving poor people
into the labor force, their dependence
on public transportation must be
considered.
A sizeable proportion of poor house­holds
did not have health care expenses,
including insurance premiums. Future
research should more closely examine
the health care situation of poor house­holds.
Employer-provided insurance
and medicaid may help many of these
households; others may be going with­out
medical care.
Family Economics and Nutrition Review
References
1. Davis, C.G. 1982. Linkages between socioeconomic characteristics, food expenditure
patterns, and nutritional status of low income households: A critical review. American Journal of
Agricultural Economics 64(5): 10 17-1025.
2. Duncan, G.J. 1984. Years of Poverty, Years of Plenty. Institute for Social Research,
University of Michigan, Ann Arbor, MI.
3. Friedman, M. 1957. A Theory of the Consumption Function. Princeton University Press,
Princeton.
4. Lester, G.H. 1991. Child Support and Alimony: I989. Current Population Reports, Consumer
Income. Series P-60, No. 173. U.S. Department of Commerce, Bureau of the Census.
5. Lino, M. 1993. Expenditures on a Child by Families, I992. U.S. Department of Agriculture,
Agricultural Research Service
6. McGregor, P.P.L. and Borooah, V .K. 1992. Is low spending or low income a better indicator of
whether or not a household is poor: Some results from the 1985 Family Expenditure Survey.
Journal of Social Policy 21 (1):53-69.
7. O'Connell, M. and Bachu, A. 1992. Who's Minding the Kids? Child Care Arrangements:
Falli988. Current Population Reports, Series P-70, No. 30. U.S. Department of Commerce,
Bureau of the Census.
8. Short, K. 1992. Health Insurance Coverage: I987-I990. Current Population Reports,
Household Economic Studies. Series P-70, No. 29. U.S. Department of Commerce, Bureau
of the Census.
9. U.S. Department of Commerce, Bureau of the Census. 1991. Measuring the Effect of
Benefits and Taxes on Income and Poverty: 1990. Current Population Reports, Consumer
Income. Series P-60, No. 176-RD.
10. U.S. Department of Commerce, Bureau of the Census. 1993. Money Income of Households,
Families, and Persons in the United States: I992. Current Population Reports, Consumer
Income. Series P60-184.
11. U.S. Department of Commerce, Bureau of the Census. 1993. Poverty in the United States:
I992. Current Population Reports, Consumer Income. Series P60-185.
12. U.S. Department of Labor, Bureau of Labor Statistics. 1993. Consumer Expenditures in
I992. Report 861.
13. Yang, H.W. and Basiotis, P.P. 1988. Expenditures on food away from home of low-income
households-analysis using USDA 1985 and 1986 Continuing Survey of Food Intakes by
Individuals (CSFII) Data. American Journal of Agricultural Economics 70(5): 1209-1210.
1996 Vol. 9 No.1 13
14
Demographic and Economic
Determinants of Household Income
Polarization Among the States in
America
By Mohamed Abdei-Ghany
Professor
The University of Alabama
Using data from the 1990 Census, this paper examines the effects of
household characteristics and factors related to the structure of the
economy on income polarization among the States. Results indicate that
women's labor force participation rate, the unemployment rate, education,
and the percentage of manufacturing workers to service workers contribute
to the determination of income polarization. Implications for public policy
are discussed.
[!] he results of recent studies
(3, 5, 11, 14, 16, 21, 22, 23,
28, 31, 33) conclusively
demonstrate that income
distribution in America has become less
equal. However, the causes that have
led to this change are still debatable.
Some researchers argue that it has been
the result of changes in the demographic
characteristics of the population, i.e.,
supply-s.ide factors such as the increase
in female-headed households, the shift
in age distribution caused by the maturity
of the baby-boom generation, and the
rise in women's labor force participation.
Others point to changes in the structure of
the national economy, i.e., demand-side
factors such as changes in occupational
and industrial structure and technology.
This study examines differences in
the distribution of household income
among the States in America in 1989.
It relates these differences to variations
in supply-side and demand-side factors.
The analysis thus focuses on demo­graphic
factors as well as economic
conditions affecting household income
distribution.
Background and
Related Literature
The distribution of income among
families reflects not only the economic
structure of the society but also the oppor­tunities,
situations, and proprieties of
family life (35). Understanding the
factors and conditions precipitating the
increases in family income inequality
and what this situation means for the
family is paramount for devising social
policies ( 16).
Family Economics and Nutrition Review
One of the major worries that is associ­ated
with rising levels of income
inequalities is the increasing bipolariza­tion
of income. Bradbury ( 5) noticed
that the shrinking of the middle class
would not be a reason for concern if
families were generally getting Iicher.
However, her data showed that median
family incomes adjusted for the rate of
inflation fell, and the percentages of
families with higher and lower income
increased.
Changes in the distribution of income
may be a result of responses to changes
in the characteristics of families. For
example, an increasing proportion of
families headed by females can lead to
an increase in the number of families
with low income (27). On the other
hand, changes in the economic structure
of the society, such as unemployment
rate or changes in the occupational and
industrial mixture of jobs, may alter the
distribution of income. Actual changes
in the income distribution of American
families are determined by the com­bined
effect of several factors.
Kuznets (19) and Paglin (26) argued
that the shift in the demographic compo­sition
of the population in the postwar
era towards younger, older, and female­headed
units fostered greater inequality
within the various family types.
Women's labor force participation has
been debated in the literature in terms
of how it affects income distribution.
In the 1960's and early 1970's, a major
percentage of the wives who joined the
labor force were from families where
husbands had lower than average earnings
(18); this participation reduced the
income inequalities among families.
Sweet (32) and Mincer (24) both using
data from the 1960 census, Smith (30)
for the period 1960-70, Danziger (7) for
1996 Vol. 9 No.1
the period 1967-74, Harris and Hedderson
( 12) for the period 1967-76, and Bartlett
and Poulton-Callahan ( 1) for the period
1951-76, showed that rising labor force
participation by women has, actually,
reduced income inequality.
In the late 1970's and 1980's, more
wives from families where husbands
had above-average incomes entered
the labor force. Consequently, this situ­ation
led to speculation that a further
increase in female's labor force partici­pation
could result in an increase in
income inequality ( 15 ).
However, studies by Horvath (15) using
data for the year 1977, Beston and van
Der Gaag (2) covering the period 1968-
80, and Grubb and Wilson ( 11) for the
1967-88 period indicated that increasing
labor force participation by wives actu­ally
continued to serve as an equalizing
factor regarding household income
inequality.
Compositional changes in the age
structure of the population could affect
income distribution. An increase in the
number of household heads under age
25 or over age 65 (whose households
have relatively low incomes) would
increase income inequality. Lawrence
(20) suggested that the entry of the
baby-boom generation into the labor
force and the resulting changes in the
age distribution of the work force pro­vide
a powerful explanation of income
inequality.
Among the macroeconomic factors
that affect income distribution is the
unemployment rate. Horowitz ( 13 ),
studying the 1954-71 period, concluded
that unemployment increased income
inequality within and among members
of various races.
In this study, the
income polarization
ratio, defined as
bottom-to-top
quintile income
ratio, is used to
measure income
inequality.
15
Some researchers have argued that
increased bipolarization of income in
America has been caused by shifts in
the occupational and industrial mix of
jobs in the economy. They attribute the
shifts to declining employment in
manufacturing industries and growth
of high technology industries, service­producing
industries, and low-paying
occupations (4, 10, 21). Kosters and
Ross ( 17) pointed out that wages for
service workers are about 83 percent
of manufacturing wages.
Rosenthal (27), however, examining the
period from 1973 to 1982, concluded
that the changes in the occupational
structure alone do not support the claims
of bipolarization. Also, the results of a
study by Davidson and Reich (9) indi­cated
that during the 1970-85 period,
employment loss in manufacturing was
at the tails rather than at the middle of
the industry wage distribution and as
a result, employment shifts out of
manufacturing had an equalizing effect.
According to the authors, the increase in
inequality can be accounted for mainly
by increasing wage differentials among
industries.
Changes in the provision of transfer
income may also affect income in­equality.
Studies showed that public
transfers have equalizing effects on
income distribution (8, 34).
Education, as measured by the percent
of the population completing high
school, was found to be inversely
related to income inequality (6, 25, 29).
To sum up, explanations of the increase
in income inequality in America
include the following: (1) increased
labor force participation by women
from families with higher-than-average
incomes, (2) growing numbers of youth
16
and elderly who command lower
incomes than other age groups, (3) an
increase in unemployment rate has a
differential effect on inter-industry
wages leading to greater income in­equality,
(4) a decline in manufacturing
employment may cause a reduction in
the share of employment near the center
of wage distribution, (5) a reduction in
public transfers would increase the
number of low-income units, and (6) an
increase in the percent of the population
completing high school would reduce
income inequality.
Methodology
Data
The source of data for this study was
the 1990 census (36). Tabular data
from published reports were used in
the analysis.
Measures of Inequality
There are several measures of income
inequality. They include, but are not
limited to, the Gini Coefficient, Theil's
Index of inequality, coefficient of
variation, incidence of poverty, stand­ard
variation, standard variation of the
logarithm of income, the normalized
interquartile range, and the income
polarization ratio. Each of the measures
has different properties and is sensitive
to different dimensions of the
distribution.
In this study, the income polarization
ratio, defmed as bottom-to-top quintile
income ratio (22), is used to measure
income inequality. In calculating quin­tiles
of income, midpoints were used for
the closed income classes and a Pareto
curve was fitted to the open-ended class
of income to approximate the mean
measure of income (for more detailed
methods of computation, consult
Maxwell (22), pp. 142-145).
Variables
Pretax incomes earned by households in
1989 were used for the calculations of
income polarization ratios. A household
consists of all the persons who occupy
a housing unit. The incomes of house­holds
rather than families or individuals
are used in this study. A household is an
income-pooling unit, whereas families
do not include households made up of
individuals. In this study, the calculated
income polarization ratios refer to in­equality
of pretax money income, and
the ratios constitute the dependent
variable in the statistical model.
The independent variables measuring
demographic characteristics of house­holds
and economic structure of the
State are explained as follows:
(1) Female's labor force participation:
labor force participation rate for
females 16 and older.
(2) Dependency ratio: summation of
number of individuals under 18
and over 64 divided by number
between 18 and 64 years old,
represented as a percentage.
(3) Industry: ratio of manufacturing
workers to service workers.
(4) Unemployment rate: unemployment
rate for persons 16 years and older.
(5) Government assistance: average
annual public assistance income.
( 6) Education: percentage of population
completing high school.
Previous analysis indicated a correlation
(.614) between the variable "Female­headed
households" and "Female's
labor force participation." Therefore,
"Female-headed households" was
omitted from the regression model.
Family Economics and Nutrition Review
Model and Statistical Procedure Table 1. Percentage distribution of household income by State by quintiles, 1989
Ordinary least squares regression was
used to regress the independent vari- Quintiles
ables on the income polarization ratios. State Lowest Second Third Fourth Highest
The following model was estimated:
U.S. 3.6 9.5 15.7 24.1 47.1
Gi =a+ b1Xli + ... + b6X6i + ei
Alabama 3.1 9.3 15.3 24.1 48.1
Alaska 4.6 10.4 16.5 24.9 43.6
Ariwna 3.9 9.9 15.2 23.4 47.6
where G refers to the income polariza- Arkansas 3.5 9.3 15.0 24.1 48.0
tion ratio; a is a constant term; Xli ... X6i California 3.9 9.8 15.8 23.8 46.7
denote the independent variables; Colorado 4.1 9.9 16.0 24.1 45.9
Connecticut 4.1 9.8 15.6 23.3 47.2
b1 ... b6 are parameters to be estimated; Delaware 4.4 10.8 17.0 24.4 43.3
ei is random disturbance term; and i is a District of Columbia 2.7 8.2 14.2 22.7 52.1
subscript corresponding to the 50 States Florida 3.9 9.8 14.8 22.6 48.9
and the District of Columbia. Georgia 3.4 9.6 15.4 24.2 47.5
Hawaii 4.7 10.4 16.4 24.4 44.2
Idaho 4.4 10.7 15.7 23.4 45.7
An appropriate specification for the Illinois 3.6 9.7 16.4 24.0 46.4
model is a logistics regression. How- Indiana 4.2 10.4 16.0 24.0 45.3
ever, ordinary least squares regression Iowa 4.3 10.7 16.1 23.8 45.1
Kansas 4.0 10.2 15.5 23.2 47.0
yielded the same qualitative results as Kentucky 3.2 9.1 15.3 24.2 48.2
a logistics regression, so the ordinary Louisiana 2.9 8.3 14.8 23.9 50.0
least squares model is reported for Maine 4.3 10.5 16.1 23.8 45.3
simplicity. Maryland 4.4 10.5 16.6 24.3 44.1
Massachusetts 3.5 10.1 16.5 24.6 45.2
Michigan 3.7 9.5 16.7 24.8 45.3
The model was tested for heteroscedas- Minnesota 4.1 9.9 16.7 23.9 45.3
ticity using the White Test and the Mississippi 3.1 8.0 14.9 24.6 49.4
Breusch-Pagan Test. Results showed Missouri 3.7 9.9 15.3 23.4 47.7
that the null hypothesis indicating no
Montana 3.9 10.5 15.7 24.2 45.7
Nebraska 4.3 10.7 15.8 23.6 45.5
heteroscedasticity was accepted using Nevada 4.3 10.3 16.3 23.4 45.7
both measures. New Hampshire 3.3 10.4 16.5 23.0 46.7
New Jersey 4.0 9.9 16.0 24.0 46.1
Empirical Results and Discussion New Mexico 3.4 10.0 15.3 23.7 47.5
New York 3.2 8.9 15.6 23.8 48.5
North Carolina 3.8 10.2 15.7 23.6 46.7
Differences in Inequality North Dakota 4.0 10.7 15.8 24.2 45.2
Table 1 shows quintile share distribution Ohio 3.8 10.0 15.7 24.1 46.4
for all of the States and the District of Oklahoma 3.5 9.8 15.2 23.6 47.9
Columbia. The poorest fifth of house-
Oregon 4.2 10.3 15.6 23.3 46.6
Pennsylvania 3.9 9.8 15.4 23.9 47.0
holds earned 4. 7 percent of total income Rhode Island 3.9 9.8 16.8 24.3 45.1
in the State of Hawaii, compared with South Carolina 3.5 10.3 15.9 24.1 46.2
only 2.7 percent of total income in the South Dakota 3.9 10.7 15.6 23.8 46.0
District of Columbia.
Tennessee 3.2 9.8 15.2 23.4 48.4
Texas 3.3 9.5 14.8 23.5 48.9
Utah 4.6 10.8 16.3 23.7 44.6
The richest fifth of households obtained Vermont 4.6 10.5 16.5 23.9 44.5
43.3 percent of all income in the State Virginia 3.8 10.3 16.5 24.4 45.0
of Delaware, compared with 52.1 per-
Washington 4.2 10.2 16.7 24.0 44.9
West Virginia 3.5 8.9 15.1 47.7
cent of all income in the District of Wisconsin 4.5 16.2 45.1
Columbia. It should also be noted Wyoming 4.3 16.3
that the middle quintile income share
(middle 20 percent of the population) Percentages in quintiles may not add up to 100 because of rounding.
1996 Vol. 9 No.1 17
Table 2. Income inequality within the United States, 1989 received 17.0 percent of all income in
the State of Delaware, compared with
Inequality Index of Income only 14.2 percent in the District of
State rank inequality polarization ratio Columbia. The percentages for the 50
States and the District of Columbia are
u.s. 100.00 13.08 shown in table 1.
District of Columbia 1 147.55 19.30
Louisiana 2 131.80 17.24
Mississippi 3 121.79 15.93 The income polarization ratios, defmed
Alabama 4 118.65 15.52 as the top-to-bottom share ratios, are
New York 5 115.90 15.16 calculated and presented in table 2.
Tennessee 6 115.60 15.12
Kentucky 7 115.14 15.06
As the ratio is about 13 for the United
Texas 8 113.30 14.82 States, the top quintile of households in
New Hampshire 9 108.18 14.15 1989 received $13 of income for every
New Mexico 10 106.80 13.97 $1 received by the bottom quintile.
Georgia 11 106.80 13.97
Arkansas 12 104.82 13.71
Oklahoma 13 104.66 13.69 In table 2, the income polarization ratio
West Virginia 14 104.20 13.63 for each of the States is expressed as a
South Carolina 15 100.92 13.20 percentage of the income polarization
Massachusetts 16 98.70 12.91 ratio of the United States. The income
Missouri 17 98.55 12.89
lllinois 18 98.55 12.89 inequality in each State is expressed as
Florida 19 95.87 12.54 a percentage of the inequality that exists
North Carolina 20 93.96 12.29 in the United States. For example, the
Michigan 21 93.58 12.24 District of Columbia has an index of
Ohio 22 93.35 12.21 inequality of 147.55. This means that Arizona 23 93.27 12.20
Pennsylvania 24 92.12 12.05 the District of Columbia's income
California 25 91.51 11.97 polarization ratio is 47.55 percent
Virginia 26 90.52 11.84 greater than the income polarization
South Dakota 27 90.14 11.79 ratio for the Nation, i.e., incomes are
Kansas 28 89.83 11.75
Montana 29 89.60 11.72 47.55 percent more unequally distrib-
Rhode Island 30 88.38 11.56 uted in the District of Columbia than in
New Jersey 31 88.07 11.52 the entire Nation. Hawaii, on the other
Connecticut 32 88.00 11.51 hand, has an index of inequality of
North Dakota 33 86.39 11.30
Colorado 34 85.55 1l.l9 71.86, indicating that incomes in
Oregon 35 84.79 11.09 Hawaii are 28.14 percent more equally
Minnesota 36 84.48 11.05 distributed than in the country as a
indiana 37 82.49 10.79 whole.
Washington 38 81.73 10.69
Nevada 39 81.27 10.63
Nebraska 40 80.89 10.58 It is clear from the figure that all Western
Maine 41 80.50 10.53 States with the exception of New Mexico
Iowa 42 80.20 10.49 have indices of inequality less than 100,
Idaho 43 79.43 10.39
Wyoming 44 79.13 10.35
revealing lesser inequality in income
Maryland 45 76.63 10.02 distribution than in the Nation as a whole.
Wisconsin 46 76.62 10.02 On the other hand, most of the Southern
Delaware 47 75.23 9.84 States have indices of inequality greater
Utah 48 74.16 9.70 than 100, showing greater inequality
Vermont 49 73.93 9.67
Alaska 50 72.48 9.48 in income distribution than the Nation.
Hawaii 51 71.86 9.40
18 Family Economics and Nutrition Review
Indices of income inequality, 1989
= <80 80-89
The States are ranked in order of inequality
in table 2. The District of Columbia has
the most unequal distribution whereas
Hawaii has the most equal.
Determinants of Inequality
Table 3, p. 20, provides descriptive
information regarding the independent
variables used in the analysis. The rate
offemale's labor force participation
ranges from 41.7 percent in the State
of Alaska to 50.8 percent in the District
of Columbia.
The dependency ratio signifying the
percentage of those under 18 and over
64 years to those between 18 and 64
years was the highest in the State of
Utah at 82.2 percent and the lowest in
the District of Columbia at 47.3 percent.
1996 Vol. 9 No. 1
• 90-99 • >100
The variable education, which refers
to the percentage of population com­pleting
high school, ranges from 23.9
percent in Nevada to 82.8 percent in
Pennsylvania. The variable industry,
representing the ratio of maimfacturing
workers to service workers, was the
highest in North Carolina at 95 .6 per­cent
and the lowest in the District of
Columbia at 9.3 percent.
Unemployment rate was the highest at
9.6 percent in Louisiana and the lowest
at 3.5 percent in Hawaii. The govern­ment
assistance variable, representing
the average annual income provided by
the government, ranges from $2,800 per
household in Mississippi to $5,972 per
household in California.
... most of the
Southern States
have indices of
inequality greater
than 100, showing
greater inequality
in income distribu­tion
than the Nation.
19
Table 3. Household characteristics and macroeconomic factors in the United States, 1989
Female's labor Dependency Unemployment Government
State force participation ratio Education Industry rate assistance
Alabama 45.5 64.3 76.7 77.9 6.9 2,985
Alaska 41.7 54.8 35.6 17.7 8.8 4,934
Arizona 44.6 66.2 37.0 37.2 7.2 3,711
Arkansas 45.8 70.4 67.8 77.6 6.8 2,901
California 43.4 56.7 59.2 50.6 6.6 5,972
Colorado 45.4 56.6 45.3 36.3 5.7 3,638
Connecticut 46.2 57.2 62.3 62.6 5.4 4,864
Delaware 46.8 57.8 51.9 60.2 4.0 4,012
District of Columbia 50.8 47.3 43.6 9.3 7.2 3,927
Florida 45.8 68.0 34.9 31.0 5.8 3,803
Georgia 46.2 58.1 66.3 64.3 5.7 3,210
Hawaii 44.3 57.6 65.8 17.3 3.5 5,272
Idaho 43.5 74.5 52.1 48.2 6.1 3.321
Illinois 45.4 62.2 75.4 61.0 6.6 3,925
Indiana 45.6 63.4 72.3 85.6 5.7 3,613
Iowa 46.0 70.2 78.8 54.7 4.5 3,784
Kansas 45.0 68.1 62.9 50.5 4.7 3,740
Kentucky 44.4 62.7 78.1 65.1 7.4 3,282
Louisiana 44.8 67.2 80.6 36.5 9.6 3,114
Maine 45.7 62.5 70.6 60.7 6.6 3,557
Maryland 46.9 54.1 53.3 29.2 4.3 3,915
Massachusetts 47.0 56.5 76.0 49.6 6.7 4,711
Michigan 45.4 62.3 77.8 77.4 8.2 4,369
Minnesota 46.3 64.4 75.6 54.2 5.1 4,426
Mississ1ppi 46.4 71.0 77.9 79.4 8.4 2,800
Missouri 46.1 65.9 70.8 58.8 6.2 3,314
Montana 44.8 69.8 60.0 22.0 7.0 3,620
ebraska 45.9 70.4 71.4 39.6 3.7 3,729
Nevada 44.0 54.6 23.9 13.3 6.2 3,908
ew Hampshire 46.3 57.2 45.8 72.7 6.2 3,722
New Jersey 45.8 51.8 62.6 50.9 5.7 4,298
New Mexico 44.2 67.4 54.6 22.9 8.0 3,325
New York 46.3 58.3 80.2 39.0 6.9 4,469
North Carolina 46.2 57.1 71.7 95.6 4.8 3,143
North Dakota 44.5 71.5 74.3 17.8 5.3 3,688
Ohio 45.5 63.3 75.9 73.8 6.6 3,736
Oklahoma 44.7 66.9 64.8 43.6 6.9 3,279
Oregon 44.8 64.6 49.0 55.1 6.2 3,798
Pennsylvania 45.3 63.7 82.8 61.3 6.0 4,041
Rhode Island 46.7 60.0 70.0 69.6 6.6 4,503
South Carolina 46.3 60.7 69.4 91.1 5.6 3,111
South Dakota 45.5 76.1 71.0 32.6 4.2 3,261
Tennessee 45.9 60.3 70.0 79.3 6.4 3,035
Texas 44.0 62.8 71.1 44.3 7.1 3,011
Utah 44.0 82.2 69.6 45.9 5.3 3,733
Vermont 46.5 59.2 59.0 44.3 5.9 3,966
Virginia 45.4 54.0 57.1 46.5 4.5 3,394
Washington 44.2 60.6 51.6 54.9 5.7 4,489
West Virginia 42.6 56.9 78.0 46.0 9.6 3,545
Wisconsin 46.1 65.7 78.4 81.9 4,356
Wyoming 43.8 67.4 43.4 18.5 3,410
20 Family Economics and Nutrition Review
The analysis that follows quantifies the
effects of the differences in the above
discussed independent variables on the
distribution of household income.
Table 4 presents parameter estimates
for the regression model of income
polarization for the United States in
1990. The results indicate that women's
labor force participation rate, educa­tion,
and unemployment rate are statisti­cally
significant determinants of income
inequality among the States. The ratio
of manufacturing workers to service
workers (industry) is a marginally
significant determinant of income in­equality
among the States. The results
of this research also suggest that the
States' distribution of income has been
impervious to the effects of dependency
ratio and government assistance.
The rate of women's labor force partici­pation
is positively related to income
inequality. This finding is in support of
the contention that a further increase in
women's labor force participation would
lead to an increase in income inequality
(15).
There is an inverse relationship between
education and income inequality. This
fmding is in support of previous studies
(6, 25, 29). The results also show that as
the ratio of manufacturing workers to
service workers increases, income
inequality decreases. This finding is
consistent with past studies (4, 10, 21).
Other results show that as the rate of
unemployment increases, the top
quintile of households gains income
shares at the expense of the lowest
quintile.
Table 4. Regression estimates of income polarization ratios for the
United States, 1989
Variable
Female's labor force participation
Dependency ratio
Education
Industry
Unemployment rate
Government assistance
Intercept
Adjusted R2
F
*p <.I.
***p< .001.
1996 Vol. 9 No.1
Regression coefficient
(standard error)
.619***
(.133)
-.008
(.028)
-.175***
(.039)
-.020*
(.008)
.729***
(.134)
-.00043
(.0003)
-3.791
.74
24.8***
Conclusions and Implications
The increase in income inequality in the
last decade in America has been viewed
with anxiety and with concern that the
country is drifting in the direction of
"haves" and "have-nots." Understanding
why the income distribution has become
more bipolarized is an issue that is both
of inherent interest for family economists
and of relevance for public policy.
The ranking of States according to the
measure of income inequality presented
in this paper should help State policy­makers
to be aware of the extent of
income inequality in their State and be
cognizant of the social and economic
policies that impact upon income
inequality.
Among household demographic differ­ences,
the level of education has an
impact on income distribution in the
States. Also, among the independent
variables reflecting the national economic
structure, the unemployment rate and
women's labor force participation are
statistically significant in impacting on
income distribution.
From a public policy viewpoint, several
approaches must be considered simulta­neously.
Reduction of the unemployment
rate through the creation of more jobs,
either directly through government sub­sidized
employment or indirectly through
government stimulation of the economy,
would influence the distribution of in­come.
Various avenues for providing
additional education and training for
the disadvantaged segments of our
population need to be explored and
publicized. Efforts to promote the long­range
economic advantages of "remain­ing
in school" should be revitalized at
all levels of government. Such policies
would help to close the gap between the
"haves" and the "have-nots."
21
22
References
1. Bartlett, R.L. and Poulton-Callahan, C. 1982. Changing family structures and the
distribution offamily income: 1951 to 1976. Social Science Quarterly 63( 1):28-37.
2. Betson, D. and van Der Gaag, J. 1984. Working married women and the distribution of
income. The Journal of Human Resources 19(4):532-543.
3. Blackburn, M.L. and Bloom, D.E. 1985. What's happening to the middle class?
American Demographics 7(1):19-25.
4. Bluestone, B. and Harrison, B. 1982. The Deindustrialization of America. Basic Books,
New York.
5. Bradbury, K.L. 1986. The shrinking middle-class. New England Economic Review
September/October, pp. 41-55.
6. Braun, D. 1991. The Rich Get Richer: The Rise of Income Inequality in the United States
and the World. Nelson-Hall Publishers, Chicago, IL.
7. Danziger, S. 1980. Do working women increase family income inequality? Journal of
Human Resources 15(3):444-451.
8. Danziger, S. and Plotnick, R. 1977. Demographic change, government transfers, and
income distribution. Monthly Labor Review 102(4):7-11.
9. Davidson, C. and Reich, M. 1988. Income inequality: An inter-industry analysis.
Industrial Relations 27(3):263-286.
10. Grubb, W.N. and Wilson, R.H. 1989. Sources of increasing inequality in wages and
salaries, 1960-80. Monthly Labor Review 112(6):3-13.
11. Grubb, W.N. and Wilson, R.H. 1992. Trends in wage and salary inequality, 1967-88.
Monthly Labor Review 115(6):23-39.
12. Harris, R.J. and Hedderson, J.J. 1981. Effects of wife's income on family income
inequality. Sociological Methods and Review 10(2):211-232.
13. Horowitz, A.R. 1974. Trends in the distribution of family income within and
between racial groups. In G.M. von Furstenberg, B. Harrison, and A.R. Horowitz, eds.,
Patterns of Racial Discrimination, Vol. 1/. Lexington Books, Lexington, MA.
14. Horrigan, M.W. and Haugen, S.E. 1988. The declining middle-class thesis: A
sensitivity analysis. Monthly Labor Review 111(5):3-13.
15. Horvath, F. 1980. Working wives reduce inequality in distribution of family earnings.
Monthly Labor Review 103(7):51-53.
16. Kosters, M.H. 1992. The rise in income inequality. The American Enterprise 3(6):28-37.
17. Kosters, M.H. and Ross, M.N. 1988. A shrinking middle class? The Public Interest
90(1 ):3-27.
18. Kreps, J. 1971. Sex in the Market Place: American Women at Work. Johns Hopkins,
Baltimore, MD.
Family Economics and Nutrition Review
19. Kuznets, S. 1974. Demographic aspects of the distribution of income among families:
Recent trends in the United States. In W. Sellekaerts, ed., Econometrics and Economic
Theory: Essays in Honor of Jan Tinbergen (pp. 223-245). International Arts and Sciences
Press, New York.
20. Lawrence, R.Z. 1984. Sectoral shifts and the size of the middle class. The Brookings
Review 3( 1 ):3-11.
21. Levy, F. 1987. Dollars and Dreams: The Changing American Income Distribution.
Russell Sage Foundation, New York.
22. Maxwell, N.L. 1990. Income Inequality in the United States, 1947-1985. Greenwood
Press, London.
23. McMahan, P.J. and Tschetter, J.H. 1986. The declining middle class: A further analysis.
Monthly Labor Review 109(9):22-27.
24. Mincer, J. 1962. Labor force participation of married women: A study of labor supply.
In National Bureau of Economics, Committee for Economic Research, Aspects of Labor
Economics (pp. 63-105). Princeton University Press, Princeton, NJ.
25. Nord, S. 1984. An economic analysis of changes in the relative shape of the interstate
size distribution of family income during the 1960's. The American Economist 28(2): 18-25.
26. Paglin, M. 1975. The measurement and trend of inequality: A basic revision. American
Economic Review 65(4):598-609.
27. Rosenthal, N. 1985. The shrinking middle class: Myth or reality? Monthly Labor Review
108(3):3-10.
28. Ryscavage, P. and Henle, P. 1990. Earnings inequality accelerates in the 1980s. Monthly
Labor Review 113(12):3-16.
29. Sale, T.S. 1974. Interstate analysis of the size distribution of family income, 1950-1970.
Southern Economic Journal40(3):434-441.
30. Smith, J. 1979. The distribution of family earnings. Journal of Political Economy 81(5,
Part 2):Sl63-S192.
31. Strobel, F.R. 1993. Upward Dreams, Downward Mobility: The Economic Decline of the
American Middle Class. Rowman and Littlefield Publishers, Inc., Lanham, MD.
32. Sweet, J. 1971. The employment of wives and inequality offamily income. Proceedings
of the American Statistical Association, pp. 1-5.
33. Thurow, L. 1987. A surge in inequality. Scientific American 256(5):30-31.
34. Treas, J. 1983. Trickle down or transfers? Postwar determinants of family income
inequality. American Sociological Review 48(4):546-559.
35. Treas, J. and Walther, R.J. 1978. Family structures and the distribution of family income.
Social Forces 56(3):866-880.
36. U.S. Department of Commerce, Bureau of the Census. 1993. 1990 Census of Population:
General, Social, and Economic Characteristics.
1996 Vol. 9 No. I 23
24
Factors Influencing Rural Southern
Elders' Life Satisfaction
By Julia M. Dinkins
Consumer Economist
Center for Nutrition Policy and Promotion
Retia Scott Walker
Dean
College of Human Environmental Sciences
University of Kentucky
Using 1987-88 data from a regional project involving 11 States, this study
focused on four dimensions of well-being as measured by rural Southern
elders' (n = 2,951) satisfaction with their economic status, independent living,
social interactions, and psychological status. Findings show that, overall,
rural Southern elders' satisfaction with their status is significantly affected by
some perceived and actual housing, nutrition, and clothing status variables
as well as socioeconomic and demographic characteristics, mobility, and
concerns about loneliness and the location of their home. With all other
variables controlled, actual housing and actual nutrition (measured by
nutrition-related illnesses), race, physical disabilities, an9 a concern about
loneliness were significantly related to all dimensions of well-being. Results
are useful to policymakers who address health and health care, long-term
care, social and community-based services, housing, financial security, and
community involvement issues for the elderly.
hen planning the agenda
for the 1995 White House
Conference on Aging, a
panel of expert policy
researchers suggested the following
characteristics be used to determine
how conference recommendations
should address the needs of special
groups in the elderly population: Race
and ethnicity; gender; urban, rural, and
suburban residence; elders 85 years
and older; the poor and near-poor; and
veterans (25 ). Meeting the needs of the
elderly requires consideration of the
heterogeneity of this population.
This research focuses on one of those
special groups-rural Southern elders.
In 1990, 13 percent of all persons in
the Southern region 1 were 65 years and
older. Thirty-one percent of Southerners
65 years and older lived in rural areas
(32). The South has a higher share of
the Nation's poor as indicated by poverty
rates and income (a determinant of
poverty status). The Southern region
had a poverty rate of 16 percent in 1991,
1 Alabama, Arkansas, Delaware, District of
Columbia, Florida, Georgia, Kentucky, Louisiana,
Maryland, Mississippi, North Carolina, Oklahoma,
South Carolina, Tennessee, Texas, Virginia, and
West Virginia.
Family Economics and Nutrition Review
compared with 12 to 14 percent for the
other regions (33). In 1991, the median
household income in the South was
$27,000, compared with $33,000 in the
Northeast, $32,000 in the West, and
$30,000 in the Midwest (31). Also, the
South has a higher percentage of adults
with multiple disadvantages2 (21 ). Older
and rural householders have less income
than younger and urban householders.
In 1991, householders 65 years and
older had a median household income
of$17,000, compared with younger
householders whose median household
income was $35,000 (31). In 1989,
mean income of rural elder households
was $15,400 compared with $20,400
for urban elder households (28).
The U.S. population continues to age.
The median age was 32.8 years in 1990;
it is expected to increase to 35.5 years
by the tum of the century and peak at
39.1 years in 2035. Although current
estimates indicate that one in eight
Americans are 65 years and older, by
2020 one in six and by 2030 one in five
Americans are expected to be elderly
(7). In 1990, the median age for persons
in all urban and rural areas (32.5 and
34.1 years) and Southern urban and
rural areas (33.4 to 34.7 years) was
similar ( 32 ). Because of these demo­graphic
trends and the characteristics of
the region where they live, studying the
well-being of the Southern elderly will
help in determining how best to meet the
needs of a graying U.S. population.
2Includes disadvantages such as higher rates
of poverty, high school dropouts, and public
assistance.
1996 Vol. 9 No. 1
Previous Studies
Housing
Adequate housing is an important com­ponent
of life satisfaction. The degree
to which families are satisfied with
their housing is influenced by their
age, values, ability to function within
the home, and repair needs. A study
of rural elders in two Southern States
found they tend to be more satisfied
with their housing compared with
younger cohorts ( 10 ). Also, elders'
housing decisions are more likely than
those made by younger cohorts to be
influenced by economic and personal
values.3
A study of the rural South found that
when elders compared their housing
situation with that of other elders they
know, they believed their own housing
situation was worse (9 ). Data from the
U.S. Departments of Commerce and
Housing and Urban Development (35)
show that 8 percent of U.S. elderly
householders lived in homes with
plumbing, heating, upkeep, and electrical
problems in 1991. Among those with
these housing problems, 38 percent said
the problems were severe. For those
who described their problems as severe,
41 percent lived in the South and 35
percent lived in rural areas. Other findings
indicate that those 65 years and older
spent less on home maintenance than
those 25 to 64 years old ( 34 ).
3"Economy-place[s) emphasis on the economic
uses of goods and services. They [individuals)
base choices on selling price and what they con­sider
sound business judgement. They are conser­vative
and take only calculated risks .... Personal­view[
s) the physical and social environment from
a personal perspective. The group is individualistic
and desires independence and self-expression"
(10).
Other factors that may influence hous­ing
satisfaction for the elderly include
the size of the home and costs associ­ated
with adapting the home to meet
changing needs. A study of older women
in a Southern State found that married
women were most dissatisfied with the
size of their house (too small), followed
by maintenance and yard work problems
(4). Of pre-retirees (40 years and older)
in some Western and Midwestern States,
22 percent believed the cost of modifying
their home to accommodate a wheelchair
would be prohibitive ( 18 ).
Nutrition
Another important determinant of elders'
well-being is their diet. In Healthy
People 2000 Review 1993, five leading
causes of death--<:oronary heart disease,
some cancers, stroke, noninsulin de­pendent
diabetes mellitus, and coronary
artery disease-are attributed, in part,
to Americans' diets (36). "Diets high in
calories, fat, saturated fat, cholesterol,
and salt, and low in such fiber-containing
foods as fruit, vegetables, and whole­grain
products, are associated with risks
of those diseases" (11, p. iii). Poor diets
also influence other conditions (e.g.,
overweight and osteoporosis) that affect
well-being.
Some segments of the population are
still more likely than others to be under­nourished
(20, 36). Among the elderly,
being undernourished is related to in­appropriate
food intake, poverty, social
isolation, living arrangements, disability,
diseases, and chronic use of medications
(5, 12, 22). Elderly women who con­sumed
low amounts of protein (1.47
glkg body cell mass) were more likely
to experience functional losses (in lean
tissue, muscle functioning, and immune
response) than those who received
adequate amounts of protein (2.94 glkg)
(6).
25
Although the elderly need to be con­cerned
about excessive energy intakes,
maintaining diets that ensure adequate
energy intake to meet the RDAs is
also important among this population.
Murphy et al. (20) found that among
people 65 to 84 years old, higher energy
intake (kcal)4 was positively associated
with the amount spent on food, number
of meals consumed, percentage of kcal
from snacks, and good or excellent
self-described health status when other
variables were held constant. Also, for
elderly men, weight was positively
associated with energy intake. A factor
that negatively influenced elderly men's
energy intake was the percentage of
kcal from cereals. Women were more
likely than men to have diet and medical
problems that were negatively related to
higher energy intakes. Women and men
65 to 84 years old who had poor diets
were likely to be dieting to lose weight
and did not like breakfast.
Living alone may influence dietary status
of the elderly. Compared with recently
widowed elders, those who were married
rated mealtime as an enjoyable time
more often (26). Murphy et al. (20)
found that women 65 to 84 years old
who lived with their spouse had higher
reported energy intakes than those who
lived alone or with others. Among elders
who lived alone, those with higher in­come
were more likely than those with
lower income to believe that health and
nutrition were related ( 3 ).
4'The energy requirement of an individual is the
level of energy intake from food that will balance
energy expenditure when the individual has a
body size and composition, and level of physical
activity, consistent with long-term good health;
and that will allow for the maintenance of
economically necessary and socially desirable
physical activity" (38).
26
Clothing
Reports on elders' well-being generally
do not focus on their clothing needs.
However, the psychosocial benefits as
well as the protective role of clothing
are important to perceptions of well­being
across the life cycle ( 14 ). Rural
Southern elders' concerns for clothing
are more likely to be influenced by
costs, style, and fit than by socio­demographic
characteristics (9 ). In
1992, people 65 years and older had
an average before-tax family income of
$20,890. This was the lowest average
income of any age group, except for
those less than 25 years old. Elders
spent 4 percent of their total expenditure
for apparel, compared with 33 percent
for housing, 16 percent for food, 12 per­cent
for health, 16 percent for transporta­tion,
and 19 percent for other goods and
services ( 37 ).
A study on garments worn to maintain
thermal comfort showed that the elderly
place a higher priority on comfort (92
percent) and washability (73 percent)
than fashion (21 percent) when staying
indoors. However, when going out,
fashion becomes more of a priority
(50 percent)5 (16).
Other Selected Factors
Living independently and degree of
homeboundness reflect elders' physical
disabilities and the type of assistance or
support received (15,23). A U.S. Depart­ment
of Commerce report showed that
in 1991-92 among the 48.9 million
5Percentages do not equal 100 because of multiple
responses or nonresponses. For example, elders
were asked if they believed their clothing was
fashionable when they stayed at home/when they
went out (yes or no); if the clothing was comfort­able
when they stayed at home/went out (yes or
no).
disabled6 people, 34 percent were 65
years and older ( 19 ). Disabled elderly
women were more likely than disabled
elderly men to use personal and/or tech­nical
assistance; use of assistance and
devices (such as canes, wheelchairs,
grab bars, and walkers) has a negative
impact on subjective7 perceptions of
well-being among the elderly (24).
Another factor that may affect elders'
perceptions of well-being is living
arrangement. In 1993, 24 percent of
Americans 65 to 74 years old and 40
percent of those 74 years old and older
lived alone (27). Compared with elders
who lived with others, those who lived
alone--especially rural women-were
more likely to be economically vulnerable
(1, 29). Elders who lived alone and had
more severe physical problems were
more likely than those with less severe
physical limitations to experience finan­cial
strain. Also, elders who lived alone
were likely to experience biophysical,
psychological, financial, and social
isolators8 ( 13 ).
&rhe author used data from the Survey of Income
and Program Participation (SlPP). The definition
of disability is broader than the one used in other
Bureau of Census reports. A person was disabled
if any of the following criteria were met: "(a) used
a wheelchair; (b) had used a cane or similar aid
for 6 months or longer; (c) had difficulty with a
functional activity; (d) had difficulty with an ADL
[activity of daily living]; (e) had difficulty with
an IADL [instrumental activity of daily Living];
or (f) was identified as having a developmental
disability or a mental or emotional disability" ( 19,
p. A-1). Also, reported figures exclude persons
living in nursing homes or other institutions.
7Subjective perceptions of well-being was defined
as "satisfaction with health, finances, family rela­tions,
friendships, housing, recreational activity,
religion, self-esteem, and transportation" (24,
p. S205).
8Biophysical isolators include limitations in mobility
and hearing loss; psychological isolators include
changes in self-esteem and roles; financial isolators
include ability to purchase needed goods and
services; and social isolators include limited
contact with family and friends.
Family Economics and Nutrition Review
Compared with rural Southern elders
who lived with others, those who lived
alone were less concerned about food9
and more concerned about housing10
issues (9). A study by the American
Association of Retired Persons (2)
found that elders' concerns for utilities,
property taxes, homeowners' or renters'
insurance, mortgage or rent payments,
and upkeep and maintenance were influ­enced
by different socioeconomic and
demographic characteristics (including
health limitations, race, annual income,
gender, age, and marital status). Another
study concluded that older, female, and
black elders who lived alone were more
likely than their respective counterparts
who lived with others to have fmancial
difficulties because of their lower in­come
and greater likelihood of having
physical limitations (17).
Expenditure patterns are indicators of
economic status. Although rural elders
spend a higher percentage of their after­tax
income than do urban elders (99 per­cent
vs. 95 percent), rural elders spend
less than urban elders on most goods
and services. Exceptions are home
furniture and equipment, gas and oil
for transportation, and health care
expenditures (28).
Compared with the youngest cohort
of Southern elders (age 65 to 74), the
other cohorts (age 75 to 84 and 85+)
9Elders who believed their food situation was
better than other elders they knew were signifi­cantly
less concerned about their food situation
than those who said their food situation was about
the same or worse than that of other elders they
knew. Also, elders who said food cost was not an
issue were significantly less likely than those who
said food cost was a serious issue to believe food
was a concern for them.
10Elders differed significantly on their concerns
regarding their housing situation compared with
others they knew, the repairs needed, repair costs,
difficulty meeting housing costs, and the amount
spent on maintenance and upkeep.
1996 Vol. 9 No.1
had less positive perceptions of overall
well-being (satisfaction with economic
status, independent living, social interac­tions,
and psychological status) and
well-being related to independent living
and social interactions. However, as
age increased, elders' satisfaction with
their economic situation increased (8).
Age, household income, household
net worth, perceived locus of control,
and perceived income adequacy were
related to satisfaction with financial
status among rural households in the
West and Midwest (30).
Previous studies suggest the complexity
and interdependence of factors that
influence well-being of the elderly
population. This study considers that
complexity and the multidimensional
nature of well-being as measured by
satisfaction with different aspects of
life, specifically by examining rural
Southern elders.
Models
The conceptual model for this study
suggests that the following factors may
affect elders' satisfaction: Actual and
perceived status for nutrition, housing,
and clothing; selected socioeconomic
and demographic characteristics;
concerns; and degree of mobility (fig. 1 ).
For this study, elders' satisfaction with
their economic status, independent
living, social interactions, and psycho­logical
status are examined.
Linear models for satisfaction were
estimated with four ordinary least squares
regressions. Satisfaction dimensions
(economic status, independent living,
social interactions, and psychological
status); perceived and actual housing,
nutrition, and clothing status; and age
were continuous variables. Other vari­ables
were treated as dummy variables.
... actual housing
and actual nutrition
(measured by
nutrition-related
illnesses), race,
physical disabilities,
and a concern about
loneliness were
significantly related
to all satisfaction
dimensions.
27
Figure 1. Proposed model of rural Southern elders' satisfaction
28
Independent Variables
Actual and Perceived Status
Housing
Beliefs about home repair needs
Access to and condition of appliances
and rooms
Nutrition
Dietary beliefs and practices
Nutrition-related illnesses
Clothing
Beliefs about selection and purchase
Physical conditions affecting clothing
Socioeconomic/Demographic
Characteristics
Race
Age
Gender
Locality
Education
Monthly income
Living arrangements
Housing tenure
(~ _________ c_ o _n_c_e_rns _ _______~ )
Location of home
Loneliness
(~ __ M_ o_bil_ity __ ~)
Physical disabilities
Dependent Variables
Satisfaction With:
(~ ___ E_c_o_n_o_m_i_c_s_ta _ tu_s ___ ~)
(~ ___In _depe_n_d_e_n_t_L_i_v_in_g_ _~ )
(~ ___ s oc_i_alin_t_e_rac_t_io_n_s __ ~)
(~ ___P_s y_c_h_o_lo_g_ica_l_S_t_atu_s_ _~ )
Family Economics and Nutrition Review
Each estimation model used the standard specification:
(1) Yi
where
y
bo
= bo + biXii + b2Xi2 + ... + bkXik
the predicted value of the dependent variable
the value of the dependent variable when the independent
variables equal 0
the change in the dependent variable associated with one unit
change in each independent variable when other independent
variables are held constant
The empirical model for each satisfaction dimension (SATD) was:
(2) SATD
where
AHOUS
PHOUS
ANCILL
PNDIET
A CLOTH
PCLOTH
AGE
WHITE
TOWN
HSCH
COLL
$400-$699
$700
ONEHHLD
OWNER
DISABLE
LOCALSER
LONELY
FEMALE
bo + b1AHOUS + b2PHOUS + b3ANC1LL + b4PNDIET +
bsACLOTH + b6PCLOTH + b7AGE + bgWHlTE + bgTOWN +
b10HSCH + b11COLL + bi2$400-$699 + biJ$700 +
b140NEHHLD + b1sOWNER + b16DISABLE + b11LOCALSER +
b1sLONELY + b19FEMALE
actual housing status
perceived housing status
actual nutrition status
perceived nutrition status
actual clothing status
perceived clothing status
age of elder
race of elder (the omitted category being "black")
town in rural county with 50,000 people or less (the omitted category
being "rural fann/nonfarm areas")
l if education of elder was high school, 0 otherwise
I if education of elder was college, 0 otherwise (the omitted category
being "less than high school")
I if elder's income was $400 to $699
1 if elder's income was $700 and above (the omitted category being
"less than $400")
I if household size was 1, 0 otherwise (the omitted category being
"multiperson household")
1 if elder owned home, 0 otherwise (the omitted category being "renter'')
I if elder was disabled, 0 otherwise (the omitted category being
"not disabled")
I if location was a serious concern, 0 otherwise (the omitted category
being "not a serious issue")
I if loneliness was a concern, 0 otherwise (the omitted category being
"not a concern")
Gender of elder (the omitted category being "male")
1996 Vol. 9 No. 1
To determine if multicollinearity
(correlation_2: .70) existed, coefficients
were examined. Because marital status
and number of people in the household
appeared to be highly correlated, marital
status was not included in the models.
Data and Sample
This study uses data from the "Quality
of Well-Being of the Rural Southern
Elderly: Food, Clothing, Shelter"
regional research project. The study was
funded by The Council of Administrators
of Family and Consumer Sciences, the
Association of Research Directors, and
the U.S. Department of Agriculture's
Cooperative State Research, Education,
and Extension Service (CSREES). The
data set contains information on elders'
socioeconomic and demographic charac­teristics,
concerns, health problems,
housing statu , dietary practices and
nutritional status, clothing acquisition
and preferences, and life satisfaction.
Cooperating States were Alabama,
Arkansas, Georgia, Kentucky, Maryland,
Mississippi, Missouri, South Carolina,
Tennessee, Texas, and Virginia (fig. 2).
To obtain a representative sample of
elderly people living in rural counties
of the South, the 1980 U.S. Census
population tapes were used to determine
the total population, the elderly popula­tion,
and median income by county.
Systematic random procedures based
on the proportion of elderly population
of each county were used to choose six
counties (three in South Carolina) from
a list of rural counties, that is, those
with no more than 30 percent urban
population.
Each participating State had 60 sam­pling
units with five elderly households
per sampling unit. Using a list of the
cumulative number of elderly people
29
30
... women were more
likely than men to
be satisfied with
their ability to live
independently ...
Figure 2. States participating in Quality of Well-Being of the
Rural Southern Elderly: Food, Clothing, Shelter regional research
project
I
I
T I
I
~ 1
in each enumeration district, the 60
sampling units were allocated to enu­meration
districts using sampling
intervals of l/60th of the total elderly
population in each of the six counties.
Equal probability of selection methods
were used to determine sample cluster
or sample unit starting points within the
six rural counties.
Face-to-face interviews were used to
collect data from June 1987 through
November 1988. The initial sample
consisted of 3,284 people 65 years old
and older who were noninstitutional­ized,
ambulatory, and who lived in rural
counties of the South. The sample for
this study consisted of2,951 elderly­those
who answered 24 or 25 (the high­est
possible for this study) items on the
life satisfaction scale. 11
Definition and Treatment of
Dependent Variables
The dependent variables were satisfac­tion
related to (1) economic status, (2)
independent living, (3) social interac­tions,
and (4) psychological status (see
box, p. 32). The satisfaction constructs
were introduced accordingly: "I [the
interviewer] would like to now focus on
how satisfied you are with your life at
the present time .... tell me if you are"
very satisfied (VS=4), satisfied (S=3),
dissatisfied (DS=2), or very dissatisfied
(VD=l). Each scale was summated.
11 Eighty-seven percent answered 25 of the life
satisfaction items, and 3 percent answered 24
items.
Family Economics and Nutrition Review
Table 1. Rural Southern elders: Descriptive statistics for continuous variables
Variable Mean
Standard
deviation
Number of
components to
score
Maximum
score
Mean as
percentage of
potential
maximum score
Dependent variables-
Satisfaction with:
Economic status
Independent living
Social interactions
Psychological status
16.58
21.72
16.27
21.52
3.37
3.52
2.31
2.67
6
7
5
7
Independent variables
Actual housing1
Perceived housing2
Actual nutrition3
Perceived nutrition4
Actual clothing5
Perceived clothing6
Age
49.34 9.69 24
2.21 0.94
5.00 1.00 4
20.76 3.60 5
5.62 0.99 5
7.87 1.79 6
73.84 7.34 NA
1 Access to and condition of selected durable goods and rooms scale.
2Home repair needs scale.
3Nutrition-related chronic illnesses cale.
4Dietary beliefs and practices scale.
5Physical conditions affecting clothing selection scale.
6Ciothing selection and purchases scale.
The economic status scale described
how satisfied elders were with their
present income; life savings; the amount
of money available for food, housing,
and clothing; and their ability to meet
personal and household expenses. The
mean score was 16.58 (table 1), and
Cronbach's alpha was .91. That is, 91
percent of the variance in the scores
on the economic status scale was
accounted for by true differences.
1996 Vol. 9 No. 1
The independent living scale focused
on ability to perform some household
chores, solve problems, and make
decisions. The mean score was 21.72
(Cronbach's alpha= .89). The mean
score for social interactions was 16.27.
This dimension measured satisfaction
with involvement in religious activities
and contact with others. Psychological
status measured satisfaction with time
spent alone, life accomplishments,
home safety, living arrangements, and
24
28
20
28
72
4
8
25
10
18
NA
69
78
81
77
69
55
63
83
56
44
NA
adjustment to retirement and retirement
age. The mean score was 21.52. Eighty­three
to 87 percent of the variance in
the scores on the psychological status
and social interactions dimensions,
respectively, was accounted for by
differences in elders' perceptions.
Respondents' mean scores, as a percent­age
of the maximum score that could
be obtained, ranged from 83 percent
(perceived nutrition) to 44 percent
(perceived clothing).
31
Satisfaction Dimensions
Economic Status
How satisfied are you with your present income?
How satisfied are you with your life savings?
Are you satisfied with your ability to meet personal and household expenses?
How satisfied are you with the amount of money you have to spend for
(a) clothing?
(b) housing?
(c) food?
Independent Living
Are you satisfied with your ability to take care of your household chores?
How satisfied are you with your ability to get around without help from others?
How satisfied are you with your ability to solve your own problems?
Are you satisfied with your ability to make ypur own decisions?
How satisfied are you with the [sic] ability to
(a) prepare your own meals?
(b) travel?
(c) take care of personal hygiene needs?
Social Interactions
How satisfied are you with the contact you have with
(a) family?
(b) friends?
(c) neighbors?
(d) young people?
How satisfied are you about the extent to which you are involved in religious
activities?
Psychological Status
How satisfied are you about spending time alone?
How satisfied are you with your activities since retirement?
How satisfied are you with your life accomplishments?
How satisfied are you with the safety of your home?
How satisfied are you with your living arrangements?
How satisfied are you with adjustments you have made since retirement?
How satisfied are you about reaching retirement age?
32
Definition and Treatment
of Independent Variables
Independent variables were actual and
perceived status of housing, nutrition,
and clothing, socioeconomic and demo­graphic
characteristics, mobility, and
elders' concerns about loneliness and
location of their home in relation to
neighbors and services.
Housing Status
Actual housing consisted of one
summated scale: Presence of selected
durabJe goods, condition of durable
goods, accessibility of rooms in the
home, and condition of rooms. The
higher the score, the more likely elders
were to have the selected durable goods
in working order and accessible rooms
in good condition. Perceived housing
had one summated scale consisting of
home repair needs. Elders were asked:
" .... How would you rate the condition
of your present home?" 12 The more
repairs believed necessary, the higher
the score on this dimension.
Nutrition Status
Nutrition-related illnesses was the
actual nutrition status measure. Elders
were asked if they had diabetes, heart
problems, high blood pressure, or
atherosclerosis. The higher the score,
the more likely elders were to have
nutrition-related health problems.
Dietary beliefs about nutritional prac­tices
was the perceived nutrition status
summated scale. Elders stated if they
never (1), seldom (2), sometimes (3),
almost always (4), or always (5):
Believed they ate nutritious meals,
12Response choices: No repairs needed, only a
few repairs, many minor repairs, or many major
repairs needed.
Family Economics and Nutrition Review
thought what they ate affected how they
felt, believed they made an effort to eat
the right amount of food, thought they
tried to choose the right kinds of foods
to eat, and believed what they ate
would affect their health. Higher scores
indicated elders' beliefs about their
nutritional status were positive.
Clothing Status
To describe actual clothing status,
physical conditions that affect clothing
selection were used in a summated
scale. Elders were asked if arthritis/
rheumatism, humpback, swayback,
enlarged waist or abdomen affected
the type of clothing selected. The more
conditions reported, the higher the
score. The perceived clothing status
scale measured elders' beliefs about
clothing purchases and selection.
They indicated if they purchased age­appropriate
and easy-on/easy-off
clothing, if their budget was adequate
for purchasing needed clothes, and if
they were able to fmd styles that were
suitable for their figure type. The higher
the score, the more positive elders felt
about clothing selections.
Results
Characteristics of Elders
Most were White, female, had less than
a high school education, and were not
physically disabled. Also, most owned
their home and believed the location of
their home was not a serious issue. A
majority of the elders lived in rural
farm/nonfarm areas, lived with their
spouse or others, had a monthly income
over $400, and thought loneliness was
a serious concern (table 2).
1996 Vol. 9 No. 1
Table 2. Rural Southern elders: Descriptive statistics for categorical
variables
Gender
Female
Male
Race
White
Black
Variables
Rural county residence
Farm/nonfarm
Town
Education
Less than high school
High school or technical/trade
College
Respondent's monthly income
<$400
$400-$699
$700+
Household size
One
Two or more
Housing tenure
Owner
Renter
Home's location is serious concern
Yes
No
Loneliness is a serious issue
Yes .
No
Physically disabled
Yes
No
n Percent
2,323 79
622 21
2,339 79
608 21
1,729 59
1,214 41
1,987 67
573 20
378 13
1,379 48
897 32
581 20
1,327 45
1,594 55
2,471 84
474 16
825 28
2,101 72
1,599 55
1,317 45
469 16
2,482 84
33
Table 3. Rural Southern elders' satisfaction: Ordinary least squares
regression results
Satisfaction with
Economic Independent
Variables status living
Actual housing1 .154** .153**
Perceived housing2 -.186** - .043*
Actual nutrition3 -.086** -.151 **
Perceived nutrition 4 .010 .096**
Actual clothing5 - .039* - .078**
Perceived clothing6 -.174** -.053**
White (Black) .145** -.070**
Age .119** -.111 **
One-person household -.008 - .095**
(Multiperson household)
Female (male) .025 .035*
Town (farm/nonfarm) -.046** -.006
High school7 .041* .036*
(Less than high school)
College7 .048** .022
$400-$699 ( <$400)8 .005 -.005
$7008 .120** .012
Owner (Renter) .074** .027
Location is serious concern -.017 .005
(Not serious)
Loneliness is a concern -.070** -.144**
(Not a concern)
Disabled (Not disabled) -.036* - .259**
iF .29 .27
F ratio 62.71 ** 56.71 **
1 Access to and condition of selected durable goods and rooms scale.
2Home repair needs scale.
3Nutrition-related chronic illnesses scale.
'*Dietary beliefs and practices scale.
Social
interactions
.174**
.005
- .084**
.082**
- .002
- .022
-.099**
-.006
-.076**
.039*
-.081 **
.070**
.021
.013
.006
-.010
-.050*
-.099**
-.118**
.11
18.88**
5Physical conditions affecting clothing selection scale.
6C!othing selection and purchases scale.
7High school is 12th grade or technical/trade school. College is I or more years.
8Respondent's monthly income.
* P:5 .05.
** p :5 .0 1.
34
Psychological
status
.185**
-.070**
-.069**
.078**
.077**
- .101**
-.072**
.033
- .004
.041*
- .009
.026
.040*
.010
.066**
.020
-.018
-.173**
- .083**
.19
36.58**
Satisfaction Related to Actual
and Perceived Status
Results reveal that when other variables
were controlled, rural Southern elders
who were pleased with their actual
housing status were significantly more
likely than those who were not pleased
to be satisfied across all dimensions­economic
status, independent living,
social interactions, and psychological
status (table 3). Elders whose homes
needed repair were significantly less
satisfied with their economic status, abil­ity
to live independently, and psycho­logical
status. Additional analysis
shows that 41 percent of Black elders
and 31 percent of White elders said
they had major or many minor home
repair needs (fig. 3). The presence of
nutrient-related illnesses was negatively
related to all satisfaction dimensions.
Elders who believed there was a connec­tion
between food-related behavior and
health were significantly more likely
than those who believed otherwise to
be satisfied with their ability to live
independently, their social interactions,
and their psychological status. Figure 4
shows that between 84 and 73 percent
of the rural Southern elders had positive
beliefs and practices related to nutrition.
However, compared with other beliefs
and practices, a larger percentage of
elders never or seldom believed what
they consumed affected their health
(14 percent).
The more physical conditions elders
had that affected actual clothing status,
the more likely rural Southern elders
were to indicate significant dissatisfaction
with their economic status and inde­pendent
living and significant satisfac­tion
with their psychological status.
Elders who were more positive about
their perceived clothing status were
significantly less satisfied with their
Family Economics and Nutrition Review
Figure 3. Actual housing status: Home repair needs of rural
Southern elders by race, 1986*
100
80
60
40
20
0
Blacks Whites
0 No repairs 0 Few repairs • Many minor repairs • Major repairs
*Significantly different at p .s .01 .
perceived economic, independent living,
and psychological status compared with
elders who were less positive. Results
related to perceived clothing status
appear counter-intuitive.
Satisfaction Related to
Socioeconomic and Demographic
Characteristics
Race was significantly related to all
satisfaction dimensions, when other
factors were controlled. Compared with
Black elders, White elders were signifi­cantly
more likely to be satisfied with
their economic status and less likely
to be satisfied with their ability to live
independently, social interactions, and
psychological status.
1996 Vol. 9 No.1
As elders aged, they were significantly
more likely to be satisfied with their
economic status and less satisfied with
their ability to live independently. As
people age, the likelihood of living
alone increases. Census data show that
in 1993, 14 percent of people 55 to 64
years old, 24 percent of those 65 to 74
years old, and 40 percent of those 75
years old and older lived alone (27).
Living alone significantly influences
some areas of satisfaction among the
rural Southern elderly. One-person
households were less likely than other
households to be satisfied with their
ability to live independently as well
as less satisfied with their social
interactions.
Elders whose homes
needed repair were
significantly less
satisfied with their
economic status,
ability to live
independently,
and psychological
status.
35
Figure 4. Perceived nutritional status: Elders' beliefs, 1986
100
80
60
40
20
0
"I believe that I eat
nutritious meals. •
"I believe what I eat
makes a difference
in how I feel."
"I make an effort
to eat the right
amount of food. •
"I try to select the
right kinds of foods. •
"I think what I eat
will affect my health."
Almost always--always • Sometimes • Never--seldom
Among those 65 years and older, women
are more likely than men to live in one­person
households. Census data indicate
that 32 percent of women 65 to 74 years
old live alone compared with 13 percent
of men in this age group. Among women
75 years and older, 52 percent lived
alone versus 20 percent of men. Rural
Southern elderly women were more
likely than the men to be satisfied with
their ability to live independently, with
their social interactions, and with their
psychological status.
Elders in rural towns were significantly
less likely than elders in rural fann/
nonfarm areas to be satisfied with their
economic status and social interactions.
36
Compared with elders with less than a
high school education, those with a high
school or technical education were
significantly more satisfied with their
economic situation, ability to live inde­pendently,
and social interactions.
Those with a college education were
significantly more likely to be pleased
with their economic and psychological
status but no more likely to be pleased
with other aspects of their life, compared
with elders with less than a high school
education.
For the rural Southern elderly, income
was not a strong predictor of satisfaction.
Two significant relationships existed,
when all other variables were controlled.
Compared with those whose personal
income was less than $400 per month,
those whose personal income was $700
or more per month were more satisfied
with their economic and psychological
status.
Home ownership for the elderly often
means that the housing unit is not mort­gaged.
Elderly homeowners can allo­cate
more of their fixed income to
nonhousing expenditure categories
than can elderly renters. Compared with
renters, homeowners were significantly
more likely to be satisfied with their
economic status.
Family Economics and Nutrition Review
Satisfaction Related to Mobility
and Concerns
Rural Southern elders who were physi­cally
disabled and those who were
lonely were significantly less likely
than others to be satisfied with their
economic status, ability to live inde­pendent!
y, social interactions, and
psychological status. Also, elders who
believed the location of their home was
a serious concern were significantly less
satisfied with social interactions than
were those who believed location was
not a concern.
Conclusions and Implications
Results support the need to address most
of the 1995 White House Conference on
Aging proposed agenda items: Health
and health care, long-term care, social
and community-based services, housing,
financial security, and community
involvement (25). Other studies from
the regional project "Quality of Well­Being
of the Rural Southern Elderly:
Food, Clothing, Shelter" focused on
actual or perceived status or provided
descriptive information. This study
provides a more comprehensive frame­work
on elders' well-being, as measured
by their satisfaction with different areas
of life. Follow-up studies may concern
the effectiveness of intervention strate­gies
that could influence rural Southern
elders' satisfaction with life.
This study focuses on rural Southern
elders' well-being as measured by their
satisfaction across several dimensions
and includes objective and subjective
evaluations of their housing, nutrition,
and clothing status. Findings show
that, overall, rural Southern elders'
satisfaction with their economic status,
independent living status, social interac­tions,
and psychological status is a
1996 Vol. 9 No. 1
multidimensional construct that is sig­nificantly
affected by some perceived
and actual housing, nutrition, and clothing
status variables as well as socio­economic
and demographic charac­teristics,
mobility, and concerns. With
all other variables controlled, actual
housing and actual nutrition (measured
by nutrition-related illnesses), race,
physical disabilities, and a concern
about loneliness were significantly
related to all satisfaction dimensions.
Income, concern about location, and
housing tenure were less likely than
other variables to predict satisfaction.
For the elderly, housing has various
connotations beyond shelter and eco­nomics.
These include independence­being
able to take care of household
chores and personal needs without
assistance from others, living arrange­ments,
and safety.
Nutrition-related health problems
influence life satisfaction of the rural
Southern elderly. Having a diet-related
chronic illness affects all areas of life
satisfaction. These findings suggest
dietary intake, health status, and life
satisfaction may be included in a frame­work
for examining food-related
behavioral changes in this population.
Also, although the elderly spend a smaller
percentage of their total expenditure
on clothing compared with younger
cohorts, proper fit, costs, and styles
remain important factors. Additional
research needs to be done to explore
why elders who were satisfied with
clothing selections were less satisfied
with their economic and psychological
status and ability to live independently.
Race, age, and household type can help
determine the types of services needed
by rural Southern elders. Policies that
address elders' well-being need to focus
on issues related to living in one-person
households, adapting the environment
to accommodate elders' changing
physical conditions, and extending
elders' contact with others. Age should
be considered when professionals are
determining the adequacy of elders' in­come
and issues related to independent
living. Elders tend to become more
satisfied with their economic status as
they age, more concerned about their
ability to live independently, and they
are likely to live in one-person house­holds
at some point.
Living independently requires pre­retirees
to make decisions about retire­ment
earlier rather than later in life, and
elders need to consider interventions,
such as the development of a strong
social network, at an earlier age. Also,
the varied needs of the physically handi­capped
elder should be addressed to
foster independent living and social
interactions.
Policies that focus on meeting the needs
of rural Southern elders must be multi­faceted.
To address elders' economic
status without considering other aspects
of living leaves them vulnerable to
other factors that may reduce their
perceptions of well-being.
37
38
References
I. Aging America: Trends and Projections. 1991 Edition. Prepared by the U.S. Senate Special
Committee on Aging, the American Association of Retired Persons, the Federal Council on Aging,
and the U.S. Administration on Aging.
2. American Association of Retired Persons. 1990. Understanding Senior Housing for the 1990s.
Washington, DC.
3. Axelson, M.L. and Penfield, M.P. 1983. Food- and nutrition-related attitudes of elderly persons
living alone. Journal of Nutrition Education 15( 1):23-27.
4. Baillie, S.T. and Peart, V. 1992. Determinants of housing satisfaction for older married and
unmarried women in Florida. Housing and Society 19(2): 101-116.
5. Brech, D.M. 1994. The elderly: At risk for malnutrition. Journal of Home Economics 86(2):47-49.
6. Castandea, C., Charnely, J.M., Evans, W.J., and Crim, M.C. 1995. Elderly women accommodate
to a low-protein diet with losses of body cell mass, muscle function, and immune response. American
Journal of Clinical Nutrition 62( 1 ):30-39.
7. Day, J.C. 1993. Population Projections of the United States, by Age, Sex, Race, and Hispanic
Origin: 1993 to 2050. Current Population Reports. P25-1104. U.S. Department of Commerce,
Bureau of the Census.
8. Dinkins, J.M. 1992. Perceptions of well-being among three age cohorts of rural southern elders.
In Annual Agricultural Outlook Conference Proceedings
9. Dinkins, J.M. 1993. Meeting basic needs of rural southern elders. Journal of Home Economics
85(1 ): 18-24.
10. Earhart, C.C., Weber, M.J., and McCray, J.W. 1994. Life cycle differences in housing perspectives
of rural households. Home Economics Research Journal22(3):309-323.
II. Frazao, E. 1995. The American Diet: Health and Economic Consequences. Agricultural
Information Bulletin No. 711. U.S. Department of Agriculture, Economic Research Service.
12. Gerrior, S.A., Guthrie, J.F., Fox, J.J., Lutz, S.M., Keane, T.P., and Basi otis, P.P. 1995. Differences
in the dietary quality of adults living in single versus multi person households. Journal of Nutrition
Education 27(3):113-119.
13. Hansen-Gandy, S. and Pestle, R. 1993. Addressing elder isolation: Intervention strategies. Journal
of Home Economics 85(3):31-35.
14. Hoffman, A.M. (ed.). 1976. The Daily Needs and Interests of Older People. Charles C. Thomas
Publisher, Springfield, IL.
15. Hughes, S.L., Edelman, P.L., Singer, R.H., and Chang, R.W. 1993. Joint impairment and self­reported
disability in elderly persons. Journal of Gerontology 48(2):S84-S92.
16. Khan, S., Roper, L., and Rogers, M. 1993. Older adults: Clothing preferences for thermal
comfort in cold weather. Journal of Consumer Studies and Home Economics 17:187-195.
17. Lee, Hee-Sook. 1994. Factors influencing financial strain on elderly people who live alone in
the U.S.A. Journal of Consumer Studies and Home Economics 18:265-278.
18. McFadden, J.R., Brandt, J.A., and Tripple, P.A. 1993. Housing for disabled persons: To what
extent will today's homes accommodate persons with physical limitations? Home Economics
Research Journal22( 1):58-82.
Family Economics and Nutrition Review
19. McNeil, J.M. 1993. Americans With Disabilities: I991-92. Current Population Reports,
Household Economic Studies. P70-33. U.S. Department of Commerce, Bureau of the Census.
20. Murphy, S.P., Davis, M.A., Neuhaus,J.M., and Lein, D. 1990. Factors influencing the dietary
adequacy and energy intake of older Americans. Journal of Nutrition Education 22(6):284-291.
21. O'Hare, W. 1994. People with multiple disadvantages live in rural areas, too. Rural Development
Perspectives 9(2):2-6.
22. Payette, H., Gray-Donald, K., Cyr, R., and Boulier, V. 1995. Predictors of dietary intake in a
functionally dependent elderly population in the community. American Journal of Public Health
85(5):677-683.
23. Pearlman, D.N. and Crown, W.H. 1992. Alternative sources of social support and their impacts
on institutional risk. The Gerontologist32(4):527-535.
24. Penning, M.J. and Strain, L.A. 1994. Gender differences in disability, assistance, and subjective
well-being in later life. Journal of Gerontology 49(4):S202-S208.
25. Pillemer, K., Moen, P., Krout, J. and Robison, J. 1995. Setting the White House Conference on
Aging Agenda: Recommendations from an expert panel. The Gerontologist 35(2):258-261.
26. Rosenbloom, C.A. and Whittington, F.J. 1993. The effects of bereavement on eating behaviors
and nutrient intakes in elderly widowed persons. Journal of Gerontology 48(4):S223-S229.
27. Saluter, A.F. 1994. Marital Status and Living Arrangements: March 1993. Current Population
Reports, Population Characteristics. P20-478. U.S. Department of Commerce, Bureau of the Census.
28. Schwenk, F.N. 1992. Economic status of rural older Americans. In Annual Agricultural
Outlook Conference Proceedings.
29. Schwenk, F.N. 1994. Income and consumer expenditures of rural elders. Family Economics
Review 7(3):20-27.
30. Sumarwan, U. and Hira, T.K. 1993. The effects of perceived locus of control and perceived
income adequacy on satisfaction with financial status of rural households. Journal of Family and
Economic Issues 14(4):343-364.
31. U.S. Department of Commerce, Bureau of the Census. 1992. Money Income of Households,
Families, and Persons in the United States: 1991. Current Population Reports, Consumer Income.
Series P-60, No. 180.
32. U.S. Department of Commerce, Bureau of the Census. 1992. 1990 Census of Population,
General Population Characteristics, United States. 1990 CP-1-1.
33. U.S. Department of Commerce, Bureau of the Census. 1992. Poverty in the United States:
1991. Cunent Population Reports, Consumer Income. Series P-60, No. 181.
34. U.S. Department of Commerce, Bureau of the Census. 1994. House Beautiful-Patterns of
Home Maintenance. SB/94-7.
35. 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, H150/9J.
36. U.S. Department of Healtl1 and Human Services, National Center for Health Statistics. 1994.
Healthy People 2000 Review, 1993. DHHS Publication No. (PHS)94-1232-l.
37. U.S. Department of Labor, Bureau of Labor Statistics. 1992 Consumer Expenditure Survey.
Unpublished data.
38. Young, V.R. 1992. Energy requirements in the elderly. Nutrition Review 50(4):95-101.
L996 Vol. 9 No. 1 39
40
Research Summaries
Cholesterol
Measurement
Elevated levels of serum blood choles­terol
have been shown to be positively
correlated with increased rates of
coronary heart disease, a leading cause
of death for both men and women in
the United States. There were 478,530
deaths attributed to coronary heart dis­ease
in 1991, according to the American
Heart Association. In addition, 1.5
million Americans were expected to
· suffer a heart attack in 1994. Total costs
associated with coronary heart disease
are estimated to be $56.3 billion per
year-$37 .2 billion spent on hospital
and nursing home services, $8.7 billion
on physicians and nurses services,
$2.4 billion on drugs, and $8 billion
in lost output.
Because of the large sums being spent
on treatment of coronary heart disease,
prevention has been emphasized. In
1985, the National Institutes of Health
(NIH) founded the National Cholesterol
Education Program (NCEP) to encourage
Americans to have their cholesterol
measured and to modify their diets.
Related to the NCEP, the U.S. General
Accounting Office was asked to review
and evaluate: How cholesterol is
measured, the accuracy and precision
of cholesterol measurement techniques,
what factors influence cholesterol levels,
and the potential effect of uncertain
measurement.
The NCEP defines an adult's risk status
according to serum cholesterol levels,
including total, high-density lipoprotein
(HDL), and low-density lipoprotein
(LDL), in conjunction with other coro­nary
heart disease risk factors. Choles­terol
levels are classified as: Desirable
(below 200 mg/dL), borderline high
(200-239 mg/dL), and high (240 mg/dL
or above). Accurate cholesterol test
results are needed to provide clinical
guidelines for identifying and treating
people who are particularly at high risk
of heart disease. Positive risk factors
are:
• Hypertension (140/90 rnm Hg or
higher, or on antihypertensive
medication)
• Current cigarette smoker
• Diabetes
• Family history of myocardial
infarction or sudden death before
age 55 in father or male sibling,
before age 65 in mother or female
sibling
• Age: male 45 years or over or
female 55 years or over or post­menopausal
and not on estrogen
replacement therapy
• Low HDL cholesterol (less than
35 mg/dL)
The treatment goal is to reduce LDL
cholesterol, first with diet and then with
cholesterol-lowering drugs if diet is not
successful. The average total serum
cholesterol for adults is about 205
mg/dL, which is slightly above NCEP's
borderline-high category. Of the adult
population, 52 million people (29 per­cent)
are candidates for dietary therapy.
Of this group, 12.7 million (7 percent of
the adult population) are candidates for
drug therapy, often for life.
An NCEP panel of experts in 1988
found considerable inaccuracy in
cholesterol testing in the United States.
They and a subsequent panel in 1990
made recommendations about how
cholesterol measurement could be
standardized and improved. They
recommended that two separate
Family Economics and Nutrition Review
cholesterol measurements be averaged
together, with further testing if the first
two varied substantially. The panels
also established the goal that by 1992
a single total cholesterol measurement
should be accurate within +/- 8.9 percent.
The Health Care Financing Administra­tion
(HCF A) also established testing
requirements for total cholesterol
(+/- 10 percent) and HDL cholesterol
( +1- 30 percent).
Nearly two-thirds of American adults
have had a cholesterol test in the past
5 years and thus know their cholesterol
number. However, cholesterol levels
should be viewed as a range rather than
as an absolute fixed number. Individuals
and physicians should be aware of
cholesterol measurement variability;
decisions to classify patients and begin
treatment need to be based on the
average of multiple measurements and
the assessment of other risk factors.
Under controlled conditions, particu­lar~
y research, clinical, and hospital
laboratories, cholesterol measurement
is reasonably accurate and precise.
Less is known about the performance
of cholesterol measurement in other
settings, such as physician's offices,
commercial laboratories, and mass
public health screenings.
Over 40 manufacturers have about 160
devices on the market that use different
technologies and chemical formulations
to conduct cholesterol tests, making it
difficult to standardize measurement.
Under the Clinical Laboratory Improve­ment
Amendments of 1988, HCFA is
conducting laboratory inspections to
assess quality control procedures and
test results on all medical equipment,
including cholesterol testing. Studies of
desk-top analyzers have found accuracy
1996 Vol. 9 No. 1
problems for total and HDL measure­ments,
with misclassification rates for
some devices ranging from 17 to nearly
50 percent.
Biological and behavioral factors such
as diet, exercise, and illness cause an
individual's cholesterol level to vary,
accounting for up to 65 percent of total
variation. The average biological vari­ation
of total cholesterol is 6.1 percent;
HDL cholesterol, 7.4 percent; and LDL
cholesterol, 9.5 percent. Biological
variation is caused by behavioral factors
such as diet, exercise, and alcohol con­sumption,
and clinical factors such as
illness, medications, and pregnancy.
Changes in the consumption of satu­rated
fats and cholesterol raise or lower
serum cholesterol levels, although indi­viduals
tend to respond quite differently
to changes in diet.
Recent studies have found that differ­ences
in the way blood samples are
collected and handled can have differ­ent
results. Capillary (fmger-stick)
samples were found to be more variable
than venous samples-an important
·finding since capillary samples are
taken in screening ~ettings and are used
in recently FDA-approved and marketed
home test kits.
The total error in cholesterol testing
measurement associated with analytical
and biological variability can have im­portant
consequences. If the total error
is assumed to be 16 percent (equivalent
to the sum of the NCEP goal for analytical
variability plus the average biological
variability derived from a synthesis of
existing studies), then a single measure­ment
of total cholesterol known to be
240 mg/dL could be expected to range
from 201 to 279 mg/dL, and a single
measurement of HDL cholesterol
known to be 35 mg/dL could range
from 24 to 46 mg/dL.
Important consequences can be
associated with measurement error.
In a worst-case scenario, two types of
diagnostic errors could occur: false­positive
or false-negative results. A
false-positive screen could result in
treating someone who in fact has a de­sirable
total, HDL, or LDL cholesterol
level. A false-negative would i