**■ ft c>; /-/ • /AA9 "'^ aA* ^.
ocnn-d-oi
USDA Measuring Food Security in the United States
United States
Department of
Agriculture
Food and
Nutrition
Service
Office of
Analysis,
Nutrition, and
Evaluation
Household Food Security in the
United States, 1995-1998
Advance Report
7/7 <jQ~Q3l/?3 &
AUG 0 5 mi
Measuring Food Security in the United States:
Household Food Security in the United States 1995-1998 (Advance Report)
Summary of Findings
This advance report introduces the second installment in the planned series ofUSDA reports
presenting estimates ofthe food security status of the American people. The estimates are based
on a new direct survey measure of the severity of food insecurity and hunger due to lack of
resources to obtain food, as experienced in U.S. households, utilizing specially designed data
collected for USDA by the Bureau of the Census in each year since 1995.
This report is the first to compare estimated levels of food insecurity and hunger in successive
years, and thus to provide information on changes occurring over time. The findings show:
• essentially stable levels and rates of food insecurity and hunger in 1995-1996;
• substantially declining levels and rates of food insecurity and hunger in 1997; and
• reversal of the declining trend in 1998, bringing the rate of food insecurity among all U.S.
households back essentially to the same level it held in 1995 (10.3% in 1995,10.2% in
1998).
• One positive note in the recent changes is that households with hunger—the more severe
range of food insecurity measured-increased much more slowly than food-insecure
households without hunger, leaving the 1998 rate of households with hunger lower than the
1995 rate (3.6 percent of all households in 1998, compared with 3.9 percent in 1995).
Other detailed summary figures are presented in the first two pages of the report.
The present report does not suggest any interpretations as to the possible causes for the changes
observed; that will be more feasible and meaningful after at least one more year of data
(collected in April 1999) also has been analyzed.
Rather, the present report focuses on the special methods required to obtain consistency of
estimates, for purposes of accurate comparison, across these particular four years of the food-security
data. The need for this special handling stems from changes that were made from year
to year in the screening procedures used by Census in collecting the data. A re-design of the
food-security questionnaire introduced with the August 1998 survey resolved this problem
successfully, it is believed, so that future year's data will be consistent and comparable without
need of special handling. Given this problem of inconsistent screening in the 1995-1997 data,
the present report also serves as a preliminary bridging report for the 1998-1999 period (and
beyond) in comparison with the earlier period.
TS
MEASURING FOOD SECURITY IN THE UNITED STATES
HOUSEHOLD FOOD SECURITY IN THE UNITED STATES 1995-1998
(ADVANCE REPORT)
Correction
Please note the following correction in Table 2C (page 7) of the enclosed report:
In the third column, all food insecure households as a percent of total households with a
household income-to-poverty ratio under 1.85 should read 21.7 (not 4.7); the number of such
households (6,849) is reported correctly.
Additional Information
Office of Analysis, Nutrition, and Evaluation
Food and Nutrition Service, USDA
3101 Park Center Drive
Alexandria VA 22302
(703)305-2134
http://www.fhs.usda.gov/fhs
£ July 1999
MEASURING FOOD SECURITY IN THE UNITED STATES
HOUSEHOLD FOOD SECURITY IN THE UNITED STATES
1995 - 1998 (ADVANCE REPORT)
Food and Nutrition Service United States Department of Agriculture July 1999
The long-running expansion of the U.S.
economy and the continuing strength of
the nation's nutrition safety net has helped a
large majority of American households
achieve or maintain food security.
Preliminary estimates indicate that during
the 12 months ending in August 1998, 93
million U.S. households, or 89.8 percent of
all households, were food secure. That is,
they had access at all times to enough food
for an active healthy life, with no need for
recourse to emergency food sources or other
extraordinary coping behaviors to meet their
basic food needs. More than 230 million
Americans lived in food-secure households
during this period.
At the same time, about 10.5 million U.S.
households (10.2 percent of all households)
were food insecure, meaning that they did
not have this same access to enough food to
fully meet basic needs at all times. About
36 million persons lived in these food-insecure
households, with children
accounting for nearly 40 percent of this
group.
Clearly, despite the great strength of the
U.S. economy and the nation's nutrition
safety net, many American families and
individuals still struggle to meet basic needs.
Among the estimated 10.5 million
households that experienced some degree of
food insecurity in 1998, 3.7 million reached
a level of severity great enough that one or
more household members were hungry at
least sometime during the period due to
inadequate resources for food. Altogether,
some 6.6 million adults and 3.4 million
children lived in such households in 1998
(Figures 1 and 2).
Table 1 shows estimated prevalence levels
based on comparable data and consistent
methods applied across the four years
between 1995 and 1998. The trend over the
period shows the food security of U.S.
households improving through 1997, then
declining again in the 12 months ending in
August 1998. Comparing the rates of U.S.
food security (insecurity) between 1995 and
1998 shows virtually no overall change for
the period as a whole. Food secure
(insecure) households were 89.7 (10.3)
percent of all households in 1995 and 89.8
(10.2) percent in 1998.
This report was prepared by Gary Bickel and Steven Carlson of the Food and Nutrition Service and Mark Nord of
the Economic Research Service. The data for 1995-1997 were prepared by James Ohls, Abhijay Prakash, Larry
Radbill, and Allen Schirm of Mathematica Policy Research, Inc.; the 1998 data were prepared by Mark Nord. The
information reported is based on data collected for USDA by the U.S. Bureau of the Census in the Food Security
Supplement to the Current Population Survey of April 1995. September 1996, April 1997, and August 1998.
FIGURE 1
Food Security Status of U.S. Households:
Preliminary Prevalence Estimates, 1995-1998
D
Food
Secure
Food
Insecure,
Total
Food
Insecure
Without
Hunger
Food
Insecure
With
Hunger
FIGURE 2
Food Insecurity and Hunger in U.S. Households:
Preliminary Prevalence Rate Estimates, 1995-1998
Food Insecure,
Total
Food Insecure
Without Hunger
Food Insecure
With Hunger
i
BLANK
PAGE
TABLE 1
Preliminary Food Security Prevalence Estimates, 1995-1998
1995 a/ 1995 b/ 19% b/ 1997 b/ 1998 b/
(000) (%) (000) (%) (000) (•/.) (000) (•/.) (000) (%)
All U.S. Households 100,210 (100.0) 100,445 (100.0) 101,508 (100.0) 102,373 (100.0) 103,480 (100.0)
Food Secure 88,266 88.1 90,097 89.7 90,964 89.6 93,459 91.3 92,972 89.8
Food Insecure: 11,943 11.9 10,348 10.3 10,544 10.4 8,914 8.7 10,509 10.2
Without Hunger 7,783 7.8 6,402 6.4 6,407 6.3 5,760 5.6 6,820 6.6
With Hunger 4,160 4.2 3,946 3.9 4,137 4.1 3,154 3.1 3,689 3.6
Adult Members (total) and by 190,550 (100.0) 191,063 (100.0) 193,608 (100.0) 195,180 (100.0) 196,972 (100.0)
Household Food Security Level:
Food Secure 169,590 89.0 172,862 90.5 175,003 90.4 179,420 91.9 174,761 88.7
Food Insecure: 20,997 11.0 18,200 9.5 18,606 9.6 15,761 8.1 22,210 11.3
Without Hunger 14,066 7.4 11,611 6.1 11,582 6.0 10,601 5.4 15,646 7.9
With Hunger 6,931 3.6 6,589 3.4 7,024 3.6 5,160 2.6 6,564 3.3
Child Members (total) and by 70,160 (100.0) 70,279 (100.0) 71,172 (100.0) 70,948 (100.0) 71,296 (100.0)
Household Food Security Level:
Food Secure 56,480 80.5 58,048 82.6 58,218 81.8 60,589 85.4 57,252 80.3
Food Insecure: 13,689 19.5 12,231 17.4 12,953 18.2 10,359 14.6 14,044 19.7
Without Hunger 9,427 13.4 8,131 11.6 8,537 12 7,444 10.5 10,653 14.9
With Hunger 4,262 6.1 4,100 5.8 4,416 6.2 2,915 4.1 3,391 4.8
Source:
Estimates based on data from the April 1995, September 1996, April 1997, and August 1998 Food Security Supplement to die
Current Population Survey, prepared by Mathematica Policy Research, Inc.and U.S. Department of Agriculture, Economic Research Service.
Notes:
a/ Estimates based on exact application of method reported in Hamilton et al., 1997a and 1997b.
b/ Consistent estimates across years, based on weighted data from samples that passed the least restrictive common screen in all years,
and with imputed values for all missing items (present in < 2 percent of sample households).
The more severe range of food insecurity
reported here—the category of food insecure
households with evidence of hunger among
household members—followed a similar
year-to-year trend, but over the period as a
whole registered a decline in the prevalence
of hunger among household members. In
1995, 3.9 percent of all households were
food insecure with hunger while the
comparable figure for 1998 was 3.6 percent.
The profile of household characteristics
shown in Tables 2A-2D reveals the variation
in rates of food insecurity and hunger in
various household types. For example, the
national rate of food insecurity in 1998 was
exceeded substantially in high-risk groups
such as single-woman headed households
with children (31.9%), Hispanic and Black
households generally (21.8 and 20.7 percent
respectively), and households with incomes
below the official poverty line (35.4%).
Overall, households with children
experienced food insecurity at more than
double the rate for households without
children (15.2 versus 7.2 percent). Among
households with children, only married-couple
families showed a lower rate of food
insecurity (9.6%) than the national average.
The prevalence of food-insecurity in central
cities (14.2%) and rural areas (10.6%)
substantially exceeded that of suburbs and
other metropolitan areas outside central
cities (7.6%). Regionally, the rate was
higher in the South and West (11.1 and
12.2% respectively) and lower in the
Midwest (7.7%).
This pattern of variation in prevalence of
food insecurity across household types and
characteristics is closely followed by the
variations observed in prevalence of hunger
across household types. The national
prevalence rate for hunger in 1998—present
in 3.6 percent of all households—was
exceeded by hunger prevalence among
families headed by a single woman (10.4%),
men living alone (5.6%), Black and
Hispanic households generally (8.2 and 6.7
percent respectively), households below the
poverty line (13.5%), central-city
households (5.6%), and those in the South
and West (4.0 and 4.2 percent respectively).
Below-average hunger prevalence occurred
in households with elderly (1.8%), married-couple
families with children (2.0%), White
non-Hispanic households generally (2.5%),
and households at or above 185% of the
poverty line (1.2%).
Background
This advance report of preliminary findings
for the period 1995-1998 introduces the
second installment in the annual series,
Measuring Food Security in the United
States. The series was inaugurated in
September 1997 with the three-volume U.S.
Department of Agriculture (USDA) report,
Household Food Security in the United
States in 1995 (Hamilton et al., 1997a and
1997b; Price et al., 1997). Both the earlier
and the present report utilize a new direct
survey measure of the severity and extent of
food insecurity and hunger among U.S.
households due to lack of resources to obtain
food. The measure was developed over
several years by the Food Security
Measurement Project, an ongoing
collaboration among Federal agencies,
academic researchers, and commercial and
non-profit private organizations (Carlson, et
al., 1999; Olson, 1999).
TABLE 2A
1995 - Prevalence of Food Security, Food Insecurity, and Hunger
by Selected Characteristics of Households
Food Secure
Food Insecure:
All Without Hunger With Hunger
Total (000) (%) (000) (%) (000) (%) (000) (%)
All Households 100,445 90,097 89.7 10,348 10.3 6,402 6.4 3,946 3.9
Adults in households 191,063 172,862 90.5 18,200 9.5 11,611 6.1 6,589 3.4
Children in households 70,279 58,048 82.6 12,231 17.4 8,131 11.6 4,100 5.8
Household Composition
With Children < 6 18,003 14,955 83.1 3,047 16.9 2,149 11.9 898 5.0
With Children < 18 37,520 31,728 84.6 5,791 15.4 3,940 10.5 1,851 4.9
Married Couple Families 26,810 24,159 90.1 2,650 9.9 1,943 7.2 707 2.6
Female Head, No Spouse 8,811 5,975 67.8 2,836 32.2 1,784 20.2 1,052 11.9
Male Head, No Spouse i,897 1,592 83.9 305 16.1 213 11.2 92 4.8
Other Households with Child * 663 544 82.1 120 18.1 52 7.8 68 10.3
With No Children < 18 62,925 58,369 92.8 4,556 7.2 2,461 3.9 2,095 3.3
More Than One Adult 36,777 34,864 94.8 1,914 5.2 1,137 3.1 777 2.1
Women Living Alone 15,161 13,728 90.5 1,433 9.5 798 5.3 635 4.2
Men Living Alone 10,325 9,214 89.4 1,092 10.6 475 4.6 617 6.0
Households With Elderly 23,776 22,598 95.0 1,178 5.0 731 3.1 447 1.9
Elderly Living Alone 10,069 9,505 94.4 565 5.6 319 3.2 246 2.4
Race and Hispanic Ethnicity
White Non-Hispanic 78,100 72,334 92.6 5,766 7.4 3,588 4.6 2,178 2.8
Black Non-Hispanic 11,906 9,342 78.5 2,565 21.5 1,522 12.8 1,043 8.8
Hispanic** 7,753 6,008 77.5 1,745 22.5 1,132 14.6 613 7.9
Other Non-Hispanic 2.687 2,414 89.8 273 10.2 161 6.0 112 4.2
Household Income-to-Poverty Ratio
Under 0.50 5,603 3,495 62.4 2,107 37.6 1,173 20.9 934 16.7
Under 1.00 15,924 10,791 67.8 5,133 32.2 3,064 19.2 2,069 13.0
Under 1.30 21,953 15,562 70.9 6,392 29.1 3,805 17.3 2,587 11.8
Under 1.85 32,182 24,130 75.0 8,051 25.0 4,910 15.3 3,141 9.8
1.85 and Over 58,400 56,634 97.0 1,766 3.0 1,169 2.0 597 1.0
Income Not Known 9,863 9,333 94.6 531 5.4 322 3.3 209 2.1
Area of Residence
Inside Metropolitan Area *** 76,907 69,072 89.8 7,837 10.2 4,788 6.2 3,059 4.0
In Central City 24,701 21,197 85.8 3,505 14.2 2,132 8.6 1,373 5.6
Not In Central City 37,059 34,138 92.1 2,921 7.9 1,748 4.7 1,173 3.2
Outside Metropolitan Area 23,066 20,579 89.2 2,487 10.8 1,606 7.0 881 3.8
Census Geographic Region
Northeast 19,491 17,804 91.3 1,687 8.7 1,049 5.4 638 3.3
Midwest 23,656 21,454 90.7 231 9.3 1,365 5.8 836 3.5
South 35,891 31,934 89.0 3,958 11.0 2,495 7.0 1,463 4.1
West 21,407 18,905 88.3 2,503 11.7 1,493 7.0 1,010 4.7
• •• «»• ~ See End Notes
TABLE 2B
1996 - Prevalence of Food Security, Food Insecurity, and Hunger
by Selected Characteristics of Households
Food Insecure:
Food Secure All Without Hunger With Hunger
Category Total (000) 89.6 (000) (%) (000) (%) (000) (%)
All Households ' 101,508 90,964 89.6 10,544 10.4 6,407 6.3 4,137 4.1
Adults in households 193,608 175,003 90.3 18,606 9.6 11,582 5.9 7,024 3.6
Children in households 71,172 58,218 81.8 12,953 18.2 8,537 11.9 4,416 6.2
Household Composition:
With Children < 6 17,560 14,526 82.7 3,034 17.2 2,105 11.9 929 5.2
With Children < 18 37,460 31,677 84.5 5,783 1S.4 3,868 10.3 1,915 S.l
Married Couple Families 26,738 24,034 89.8 1,894 7.0 1,904 7.1 800 2.9
Female Head, No Spouse 8,721 5,987 68.6 2,733 31.3 1,735 19.0 998 11.4
Male Head, No Spouse 2,001 1,656 82.7 346 17.2 229 11.4 117 5.8
Other Household with Child * 736 600 81.4 137 18.6 99 13.4 38 5.1
With No Children < 18 64,048 59,287 92.5 4,760 7.4 2,538 3.9 2,222 3.4
More Than One Adult 37,806 35,790 94.6 2,016 5.3 1,122 2.9 894 2.3
Women Living Alone 14,961 13,471 90.0 1,490 9.9 789 5.2 701 4.6
Men Living Alone 10,545 9,426 89.3 1,119 10.6 529 5.0 590 5.5
Households With Elderly 24,087 22,882 95.0 1,205 5.0 766 3.1 439 1.8
Elderly Living Alone 9,760 9,188 94.1 572 5.8 352 3.6 220 2.2
Race/Ethnicity of Households:
White Non-Hispanic 77,830 72,131 92.6 5,699 7.3 3,423 4.4 2,276 2.9
Black Non-Hispanic 12,083 9,386 77.6 2,697 22.3 1,670 13.8 1,027 8.4
Hispanic** 8,165 6,360 77.8 1,805 22.1 1,097 13.4 708 8.6
Other Non-Hispanic 3,430 3,088 90.0 342 9.9 217 6.3 125 3.6
Household Income-to-Poverty Ratio:
Under 0.50 6,442 3,962 61.5 2,450 38.2 1,339 20.7 1,111 17.3
Under 1.00 14,548 9,456 65 5,093 35.0 2,930 20.1 2,163 14.8
Under 1.30 20,252 13,942 68.7 6,340 31.2 3,705 18.2 2,635 12.9
Under 1.85 32,814 24,551 74.8 8,264 25.1 5,004 15.2 3,260 9.9
1.85 and Over 57,123 55,385 96.9 1,738 3.0 1,077 1.8 661 1.1
Income Not Known 11,571 11,028 95.3 542 4.6 326 2.8 216 1.8
Area of Residence:***
Inside Metropolitan Area 81,285 73,010 89.8 8,275 10.1 4,945 6.0 3,330 4.0
In Central City 26,262 22,664 86.3 3,597 13.6 2,123 8.0 1,474 5.6
Not In Central City 40,956 37,683 92 0 3,273 7.9 1,950 4.7 1,323 3.2
Outside Metropolitan Area 20,223 17,955 88.7 2,268 11.2 1,461 7.2 807 3.9
Census Geographic Region:
Northeast 19,740 18,204 92.2 1,536 7.7 925 4.6 611 3.0
Midwest 24,132 22,046 91.3 2,089 8.7 1,273 5.2 813 3.3
South 35,792 31,543 88.1 4,249 11.8 2,612 7.3 1,637 4.5
West 21,811 19,171 87.7 2,673 12.2 1,596 7.3 1,077 4.9
*.**,*•♦--See End Notes
TABLE 2C
1997 - Prevalence of Food Security, Food Insecurity, and Hunger
by Selected Characteristics of Households
Food Insecure:
Food Secure All Without Hunger With Hunger
Category Total (000) 89.6 (000) (%) (000) (%) (000) (%)
All Households 102,373 93,459 91.2 8,914 8.7 5,760 5.6 3,154 3.1
/vdults in households 195,180 179,420 91.3 15,761 8.1 10,601 5.4 5,160 2.6
Children in households 70,948 60,589 85.4 10,359 14.6 7,444 10.4 2,915 4.1
Household Composition:
With Children < 6 17,306 14,944 86.3 2,363 13.6 1,764 10.1 599 3.4
With Children < 18 37,497 32,681 84.1 4,816 12.8 3,437 9.1 1,379 3.6
Married Couple Families 26,323 24,333 92.4 1,989 7.5 1,503 5.7 486 1.8
Female Head, No Spouse 8,973 3,460 72.0 2,512 27.9 1,681 18.7 831 9.2
Male Head, No Spouse 2,201 1,887 85.7 314 14.2 252 11.4 62 2.8
Other Household with Child * 698 595 85.1 104 14.8 73 10.3 31 4.4
With No Children < 18 64,877 60,778 93.6 4,098 63 2,324 3.5 1,774 2.7
More Than One Adult 37,788 36,056 95.4 1,732 4.5 1,078 2.8 654 1.7
Women Living Alone 15,411 14,180 92.0 1,231 7.9 678 4.4 553 3.5
Men Living Alone 10,980 9,947 90.5 1,032 9.3 495 4.5 537 4.8
Households With Elderly 24,420 23,306 95.4 1,115 4.5 766 3.1 349 1.4
Elderly Living Alone 10,323 9,816 95.0 508 5.6 329 3.1 179 1.7
Race/Ethnicity of Households:
White Non-Hispanic 78,021 73,311 93.9 4,711 6.4 2,914 3.7 1,797 2.3
Black Non-Hispanic 12,289 10,006 81.4 2,283 18.5 1,490 12.1 793 6.4
Hispanic** 8,384 6,755 80.5 1,628 19.4 1,167 13.9 461 5.4
Other Non-Hispanic 3,680 3,386 92.0 293 7.9 189 5.1 104 2.8
Household Income-to-Poverty Ratio:
Under 0.S0 6,206 4,085 65.8 2,121 34.1 1,312 21.1 809 13.0
Under 1.00 14,157 9,871 69.7 4,287 30.2 2,656 18.7 1,631 11.5
Under 1.30 20,158 14,812 73.4 5,346 26.5 3,359 16.6 1,987 9.8
Under 1.85 31,560 24,712 78.3 6,849 4.7 4,434 14.0 2,415 7.6
1.85 and Over 58,980 57,492 97.4 1,488 2.5 964 16 524 0.8
Income Not Known 11,833 11,255 95.1 385 3.2 362 3.0 213 1.8
Area of Residence:***
Inside Metropolitan Area 82,005 75,001 91.4 6.704 8.1 4,436 5.4 2,568 3.1
In Central City 26,314 23,241 88.3 3,073 11.6 1,840 6.9 1,233 4.6
Not In Central City 41,354 38,591 93.3 2,764 6.6 1,832 4.4 932 2.2
Outside Metropolitan Area 20,369 18,458 90.6 1,911 9.3 1,325 6.5 586 2.8
Census Geographic Region:
Northeast 19,897 18,342 92.1 1,556 7.8 972 4.8 584 2.9
Midwest 24,555 22,699 92.4 1,854 7.5 1,136 4.6 718 2.9
South 35,844 32,452 90.5 3,392 9.4 2,262 6.3 1,130 3.1
West 22,077 19,966 90.4 2,111 9.5 1,391 6.3 720 3.2
• •• ••• -- See End Notes
TABLE 2D
1998 - Prevalence of Food Security, Food Insecurity, and Hunger
by Selected Characteristics of Households
Food Insecure:
Food Secure All Without Hunger With Hunger
Category Total (000) (•/.) (000) (%) (000) (%) (000) (%)
All Households 103,480 92,972 89.8 10,509 10.2 6,820 6.6 3,689 3.6
Adults in households 196,972 174,761 88.7 22,210 11.3 15,646 7.9 6,564 3.3
Children in households 71,296 57,252 80.3 14,044 19.7 10,653 14.9 3,391 4.8
Household Composition:
With Children < 6 17,176 14,381 83.7 2,796 16.3 2,132 12.4 664 3.9
With Children < 18 38,178 32,365 84.8 5,812 15.2 4,216 11.0 1,596 4.2
Married Couple Families 26,415 23,873 90.4 2,542 9.6 2,019 7.6 523 2.0
Female Head, No Spouse 8,826 6,013 68.1 2,813 31.9 1,898 21.5 915 10.4
Male Head, No Spouse 2,167 1,832 84.5 336 15.5 225 10.4 111 S.l
Other Household with Child * 769 647 84.2 122 15.9 75 9.7 47 6.1
With No Children < 18 65,302 60,607 92.8 4,695 7.2 2,603 4.0 2,092 3.2
More Than One Adult 38,691 36,634 94.7 2,057 5.3 1,219 3.2 838 2.2
Women Living Alone 15,525 14,091 90.8 1,434 9.2 807 5.2 627 4.0
Men Living Alone 11,086 9,882 89.1 1,203 10.9 577 5.2 626 5.6
Households With Elderly 24,478 23,131 94.5 1,346 5.5 913 3.7 433 18
Elderly Living Alone 10,129 9,577 94.6 552 5.4 349 3.4 203 2.0
Race/Ethnicity of Households:
White Non-Hispanic 78,294 72,700 92.9 5,594 7.1 3,650 4.7 1,944 2.5
Black Non-Hispanic 12,529 9,941 79.3 2,588 20.7 1,560 12.4 1,028 8.2
Hispanic** 8,721 6,823 78.2 1,898 21.8 1,313 15.1 585 6.7
Other Non-Hispanic 3,937 3,508 89.1 429 10.9 298 7.6 131 3.3
Household Income-to-Poverty Ratio:
Under 0.50 5,205 3,187 61.2 2,018 38.8 1,202 23.1 816 15.7
Under 1.00 12,35a 7,980 64.6 4,378 35.4 2,712 21.9 1,666 13.5
Under 1.30 18,018 12,261 68.0 5,757 32.0 3,630 20.1 2,127 11.8
Under 1.85 29,540 21,985 74.4 7,555 25.6 4,870 16.5 2,685 9.1
1.85 and Over 61,775 59,482 96.3 2,293 3.7 1,558 2.5 735 1.2
Income Not Known 12,165 11,505 94.6 660 5.4 391 3.2 269 2.2
Area of Residence:***
Inside Metropolitan Area 83,189 74,824 89.9 8,364 10.1 5,361 6.4 3,003 3.6
In Central City 26,682 22,903 85.8 3,778 14.2 2,286 8.6 1,492 5.6
Not In Central City 42,196 38,969 92.4 3,227 7.6 2,178 5.2 1,049 2.5
Outside Metropolitan Area 20,291 18,148 89.4 2,142 10.6 1,458 7.2 684 3.4
Census Geographic Region:
Northeast 19,635 17,852 90.9 1,784 9.1 1,161 5.9 623 3.2
Midwest 24,321 22,446 92.3 1,875 7.7 1,235 5.1 640 2.6
South 37,328 33,188 88.9 4,139 11.1 2,653 7.1 1,486 4.0
West 22,196 19,486 87.8 2,710 12.2 1,770 8.0 940 4.2
*, **, ••• - See End Notes
8
In the National Nutrition Monitoring and
Related Research Act of 1990, Congress
recognized the need to bolster the scientific
and data resources devoted to assessing
nutritional well-being in the U.S. population.
The Ten-Year Plan developed under the Act
specified the following task, undertaken by
the Food Security Measurement Project:
Recommend a standardized mechanism
and instrument(s) for defining and
obtaining data on the prevalence of "food
insecurity" or "food insufficiency" in the
U.S. and methodologies that can be used
across the NNMRR Program and at State
and local levels.'
Beginning in 1992, staff of the USDA Food
and Nutrition Service (FNS) examined the
existing research literature in the field,
giving particular attention to the conceptual
basis for measuring the severity of food-insecurity
and hunger and to the practical
problems of developing a survey instrument
feasible for use in sample surveys at national
as well as State and local levels.
In January 1994, FNS joined with the U.S.
Public Health Service, Centers for Disease
Control and Prevention National Center for
Health Statistics (NCHS) in sponsoring a
national Conference on Food Security
Measurement and Research. The conference
brought together leading academic experts
and other private researchers with key staff
of the concerned Federal agencies. The
conference identified the consensus existing
among researchers in the field as to the
most appropriate conceptual basis for a
national measure of food insecurity and
hunger and reached working agreement as to
the best operational form for implementing
1 See end notes.
such a measure in national surveys (USDA,
1995).
After extensive cognitive assessment, field
testing, and analysis by the U.S. Bureau of
the Census in collaboration with FNS and
the interagency working group, a food
security questionnaire was fielded by the
Bureau as a supplement to the Current
Population Survey (CPS) of April 1995.2
Essentially the same CPS supplement was
implemented again in September 1996,
April 1997, August 1998, and April 1999.
While maintaining the same content, a
revised questionnaire design was introduced
in August 1998 to reduce respondent burden
and improve data quality.
In September 1995, a competitive contract
was awarded to Abt Associates, Inc. to
analyze the 1995 CPS food security data in a
cooperative venture with FNS, the
interagency working group, and other key
researchers involved in developing the
questionnaire.
The Abt team applied state-of-the-art scaling
methods developed and used most widely in
the educational testing industry to the CPS
data. This work produced a measurement
scale for the severity of deprivation in basic
food needs as experienced in U.S.
households, based on a core set of food-security
indicators from the CPS data.3
Extensive testing was carried out to establish
the validity and reliability of the scale and
its applicability across various household
types in the broad national sample.
A second, categorical measure of food-security
also was created, corresponding to
four designated ranges of severity on the
underlying near-continuous scale. Abt
classified each survey household into one of
these four categories based on its scale
score, which in turn depends on the overall
pattern of response by the household to a
core set of 18 CPS food security questions.
Attaching population weights supplied by
the Census Bureau, Abt then estimated
household prevalence levels for each of the
four severity-range categories, designated
as: food secure, food insecure without
hunger, food insecure with moderate hunger,
and food insecure with severe hunger
(Hamilton et al., 1997a, 1997b and Price et
al., 1997).
Following collection of the September 1996
and April 1997 CPS food security data, a
second research contract was awarded to
Mathematica Policy Research, Inc. (MPR) to
carry out the following tasks:
• reproduce the earlier results obtained
from the 1995 CPS data;
• replicate the 1995 analysis with the 1996
and 1997 data sets;
• assess the stability and robustness of the
measurement model when applied
independently to the separate data sets.
Beginning with the August 1998 CPS, the
questionnaire format of the food security
supplement was redesigned for improved
screener efficiency and reduced respondent
burden. Coincident with this redesign, the
USDA Economic Research Service (ERS)
assumed sponsorship of the Census Bureau's
annual food-security data collection for
USDA. ERS has analyzed the 1998 data,
applying the procedures developed and
established for USDA in the Abt and MPR
research. The MPR findings, which will be
presented in full in a forthcoming final
report, establish the stability and robustness
of the food-security measure over the 1995-
1997 period. The ERS work, which will be
presented in a subsequent report, has found
continuing stability of the measure in 1998.
Method
The means by which the U.S. food-security
scale measures the severity of food
insecurity and hunger as experienced in
American households has been described in
several places (Hamilton et al., 1997a,
1997b; Andrews et al. 1998; Bickel et al.
1998; Carlson et al., 1999.) Briefly, in each
household surveyed, the person most
knowledgeable about the food eaten there is
asked a series of questions about conditions
and behaviors occurring for them over the
past 12 months that are known to
characterize households having difficulty
meeting basic food needs. The household's
overall pattern of responses to a validated set
of 18 such questions, reflecting a wide range
of severity of food insecurity, determines the
food security status of the household. Each
question specifies lack of money to buy food
as the reason for the condition or behavior
so that voluntary fasting or dieting to lose
weight are not included in the measure.
The wide range of severity covered by the
indicator questions is reflected in the widely
differing proportions of households that
affirm the various scale items (Table 3). For
example, 12.8 percent of households in 1998
reported that they worried whether their
food would run out before they got money to
buy more, while less than 1 percent reported
that adults went whole days without eating
in 3 or more months because there wasn't
enough money for food, and only 0.2
percent of households with children reported
that children went whole days without eating
sometime during the year. Comparable
scale item response figures for 1995-1997
will be included in the final report for those
10
years. (Full question wordings are presented
in Hamilton et al. 1997a, Price et al. 1997,
and Bickel et al. 1998.)
The preliminary results presented in this
advance report are based on the CPS Food
Security Supplement data for 1995-1998,
edited to create consistency among data files
for comparative analysis across the four
years. Two sets of results are shown in
Table 1 for 1995. The first column is
comparable to the results reported in
Hamilton et al. (1997a) and the second
reflects the technical adjustments applied to
establish consistency across the four years.
The main adjustment required for strict
comparability of estimates was to apply
common screening criteria to each year's
data. In collecting the CPS food-security
data, an initial screener is applied to higher-income
households to avoid undue burden
for households showing no preliminary
signs of any food adequacy problem. In the
1996 and 1997 surveys, a similar screen was
applied to lower-income households as well.
The households screened out were deemed
to be food-secure, based on negative
responses to the broad initial screening
questions. However, analysis of the 1995
data indicated that we should expect a small
but non-trivial proportion of the low-income
households screened out in 1996 and 1997
to be classified as food-insecure if they had
been asked the full set of food-security
indicator questions.
Thus, the 1996 and 1997 data under-represent
somewhat the prevalence of food
insecurity, while the 1995 and 1998 data,
which are nearly comparable as collected,
provide more complete information.
In the present report, in order to achieve
strict comparability across years, households
in the 1995 and 1998 data are, in effect,
subject to the 1996-1997 screening rules.4
Those households that would be scieened
out under the more restrictive rules simply
are deemed to be food secure. This method
achieves cross-year comparability, but at the
cost of understating the prevalence of food
insecurity. For example, the best estimate of
food-insecurity prevalence in 1998, using all
data available, was 11.9 percent, while the
estimate based on the more restrictive
screening (Table 2D) was 10.2 percent. The
estimated prevalence of hunger is less
affected by the screening change. The full-data
estimate for 1998 is 3.8 percent,
compared with estimated prevalence of 3.6
percent based on the more restrictive screen.
For future data collections, continued use of
the 1998/1999 screen is anticipated. Thus,
food-security and hunger prevalence
statistics based on the full data as collected
will be directly comparable from 1998
onward, while the present report provides
the bridging comparison from 1998 to the
earlier years.5
Other technical refinements which were
applied to each year's estimates presented in
this report, and which produce slight
variations from the original 1995 estimates,
include: fitting the model to population-weighted
rather than unweighted data; and
revised treatment of item non-response
(present in under 1% of sample households).
11
TABLE 3
Food Security Scale Item Responses, 1995 and 1998 *
Scale Items • Percent of
Households
Affirming Item -
Household Items:
1995 1998
Worried food would run out before (I/we) got money to buy more 13.0 12.8
Food bought didn't last and (I/we) didn't have money to get more 10.6 10.8
Couldn't afford to eat balanced meals 10.0 9.1
Adult Items:
Adult(s) cut size of meals or skipped meals 6.5 6.0
Respondent ate less than felt he/she should 6.3 5.7
Adult(s) cut size or skipped meals in 3 or more months 4.7 4.2
Respondent hungry but didn't eat because couldn't afford 3.1 2.6
Respondent lost weight 1.7 1.6
Adult(s) did not eat for whole day 1.5 1.3
Adult(s) did not eat for whole day in 3 or more months 1.1 0.9
Child Items:
Relied on few kinds of low-cost food to feed child(ren) 13.2 13.6
Couldn't feed child(ren) balanced meals 8.7 8.4
Child(ren) were not eating enough 4.8 4.4
Cut size of child(ren)'s meals 2.0 1.6
Child(ren) were hungry 1.7 1.1
Child(ren) skipped meals 0.8 0.8
Child(ren) skipped meals in 3 or more months 0.6 0.5
Child(ren) did not eat for whole day 0.2 0.2
Source: Current Population Survey Food Security Supplement, April 1995 and August 1998.
Notes:
a/ Item response frequencies weighted to population totals. All estimates are based on least restrictive common
screen, 1995-1998.
b/ The actual wording of each item includes explicit reference to resource limitation, e.g., "because (I was/we were)
running out of money to buy food," or "because there wasn't enough money for food."
c/ Households not responding to item are excluded from the denominator. Households without children are
excluded from the denominator of child-referenced items.
12
Caveats Notes:
The preliminary findings presented in
this advance report are based on
application of a consistent methodology
to the 1995, 1996, 1997, and 1998 data
sets. Analysis of all these data sets is
still in progress for presentation in
subsequent final reports. These ongoing
analyses may produce further
refinements to the prevalence estimates
for these years, or in their form of
presentation. Consequently, final
estimates may differ in minor detail from
those presented in the present report.
The measure of children in food-insecure
households with hunger is not, as such, a
valid estimate of the number of children
directly experiencing hunger, but a
rather wide upper-bound for this figure.
In most households, children present are
shielded from food deprivation until
the level of deprivation among adult
members is quite severe. Work is
currently under way to develop a more
accurate estimate of children's hunger
from the CPS data.
Other sources of possible estimation bias
in the prevalence estimates presented
include: the omission of homeless
persons from the CPS household-based
sample, probable underreporting bias of
unknown size, and a potential over-estimation
bias resulting from the highly
skewed distribution of households across
the range of severity measured by the
scale.
1 Task V-C-2.4, U.S. Department of Health and
Human Services and U.S. Department of Agriculture:
Ten-Year Comprehensive Plan for the National
Nutrition Monitoring and Related Research Program.
Federal Register 1993, 58:32 752-806.
2 The Current Population Survey is a representative
national sample of approximately 50,000 households
conducted monthly by the U.S. Bureau of the Census
for the U.S. Department of Labor, Bureau of Labor
Statistics. Its primary purpose is to monitor labor
force participation and employment in the nation and
each of the 50 states. The CPS also collects
specialized supplementary data each month
sponsored by various Federal agencies. USDA plans
to collect the Food Security Supplement on a regular
annual basis, alternating between the April and
September CPS.
3 The food security scale reported here is based on
the Rasch measurement model, an application of
maximum likelihood estimation in the family of Item
Respoi.se Theory models (Wright, 1977, 1983).
These statistical measurement models were
developed in educational testing, where test items
vary systematically in difficulty and the overall score
measures the level of difficulty that the tested
individual has mastered. In the present application,
the severity of food insecurity recently experienced
by household members is analogous to the level of
test difficulty that an individual has mastered.
4 Other minor differences in the screening criteria
used in different years also exist. As a result, the
common screen applied in the four years is actually
somewhat more restrictive than the screen applied in
any single year, representing the least restrictive
screening rules that could be applied uniformly
across all four years.
5 The bridging comparison between 1998 and earlier
years will be presented by detailed breakdown in a
future full report.
Notes to Tables 2A- 2D.
* Households with children in complex living
arrangements, e.g., children of other relatives or
unrelated roommate or border.
13
* * Hispanics may be of any race.
*** Subtotals do not add to metropolitan totals
because central city residence is not identified for
some areas. Households not identified as to area
were 0.88 percent of all households.
References:
Hamilton, W.L., J.T. Cook, W.W. Thompson, L.F.
Buron, E.A. Frongillo, Jr., CM. Olson, and C.A.
Wehler (1997a). Household Food Security in the
United States in 1995: Summary Report of the Food
Security Measurement Project. Report prepared for
the USDA Food and Consumer Service, Alexandria,
VA. (available: www.usda.gov/fcs/fcs.htm).
Hamilton, W.L., J.T. Cook, W.W. Thompson, L.F.
Buron, E.A. Frongillo, Jr., CM. Olson, and C.A.
Wehler (1997b). Household Food Security in the
United States in 1995: Technical Report. Report
prepared for the USDA Food and Consumer Service,
Alexandria, VA. (www.usda.gov/fcs/fcs.htm).
Price, C, W.L. Hamilton, and J.T. Cook (1997).
Household Food Security in the United States in
1995: Guide to Implementing the Core Food
Security Module. Report prepared for the USDA
Food and Consumer Service, Alexandria, VA.
USDA Food and Consumer Service, Office of
Analysis and Evaluation (1995). Food Security
Measurement and Research Conference: Papers and
Proceedings. Alexandria, VA.
Carlson, S.J., M.S. Andrews, and G.W. Bickel
(1999). Measuring Food Insecurity and Hunger in
the United States: Development of a National
Benchmark Measure and Prevalence Estimates. 77*?
Journal ofNutrition 129:51OS-516S.
Olson, CM. (Ed., 1999). Symposium: Advances in
Measuring Food Insecurity and Hunger in the U.S.
Sponsored by the American Society for Nutritional
Sciences as part of Experimental Biology 98, April
1998, San Francisco, CA. Published as supplement
to The Journal ofNutrition 129:504S-528S.
Andrews, M., G. Bickel, and S. Carlson (1998).
Household Food Security in the United States in
1995: Results from the Food Security Measurement
Project. Family Economics and Nutrition Review
11:(1&2): 17-28.
Bickel, G., M. Andrews, and S. Carlson (1998). The
Magnitude of Hunger: In a New National Measure
of Food Security. Topics in Clinical Nutrition
13(4): 15-30.
Wright, B. D. (1977). Solving Measurement
Problems with the Rasch Model. Mesa Psychometric
Laboratory, The University of Chicago, College of
Education. Chicago, IL.
Wright, B. D. (1983). Fundamental Measurement in
Social Science and Education. Mesa Psychometric
Laboratory, The University of Chicago, College of
Education. Chicago, IL.
14
The U.S. Department of Agriculture (USDA) prohibits discrimination in all its programs and activities on the basis
of race, color, national origin, gender, religion, age, disability, political beliefs, sexual orientation, or family status.
Persons with disabilities who require alternative means for communications of program information (Braille, large
point, audiotape, etc.) should contact USDA's TARGET Center at (202)720-2600 (voice and TTD).
To file a complaint of discrimination, write USDA Director, Office of Civil Rights, Room 326-W, Whitten
Building, 14* and Independence Avenue, S.W., Washington, DC 20250-9410 or call (202)720-5964 (voice and
TTD). USDA is an equal opportunity provider and employer.
15