Download:
pdf |
pdfAttachment L
Economic Research Service Report “Analysis of the Current
Population Survey Food Security Supplement Split-Panel Test”
Economic
Research
Service
Technical
Bulletin
1963
September 2023
Analysis of the Current
Population Survey Food Security
Supplement Split-Panel Test
Alisha Coleman-Jensen and Matthew P. Rabbitt
RecommendedResearch
citation formatService
for this publication:
Economic
Coleman-Jensen, A., & Rabbitt, M.P. (2023). Analysis of the current population survey food security
www.ers.usda.gov
supplement split-panel test (Report No. TB-1963), U.S. Department of Agriculture, Economic
Research Service.
Cover image assets sourced from Getty Images.
Use of commercial and trade names does not imply approval or constitute endorsement by USDA.
To ensure the quality of its research reports and satisfy governmentwide standards, ERS requires that all research reports with
substantively new material be reviewed by qualified technical research peers. This technical peer review process, coordinated
by ERS' Peer Review Coordinating Council, allows experts who possess the technical background, perspective, and expertise
to provide an objective and meaningful assessment of the output’s substantive content and clarity of communication during
the publication’s review.
In accordance with Federal civil rights law and U.S. Department of Agriculture (USDA) civil rights regulations and policies, the
USDA, its Agencies, offices, and employees, and institutions participating in or administering USDA programs are prohibited
from discriminating based on race, color, national origin, religion, sex, gender identity (including gender expression), sexual
orientation, disability, age, marital status, family/parental status, income derived from a public assistance program, political
beliefs, or reprisal or retaliation for prior civil rights activity, in any program or activity conducted or funded by USDA (not all
bases apply to all programs). Remedies and complaint filing deadlines vary by program or incident.
Persons with disabilities who require alternative means of communication for program information (e.g., Braille, large print,
audiotape, American Sign Language, etc.) should contact the responsible Agency or USDA's TARGET Center at (202)
720-2600 (voice and TTY) or contact USDA through the Federal Relay Service at (800) 877-8339. Additionally, program information may be made available in languages other than English.
To file a program discrimination complaint, complete the USDA Program Discrimination Complaint Form, AD-3027, found
online at How to File a Program Discrimination Complaint and at any USDA office or write a letter addressed to USDA and
provide in the letter all of the information requested in the form. To request a copy of the complaint form, call (866) 632-9992.
Submit your completed form or letter to USDA by: (1) mail: U.S. Department of Agriculture, Office of the Assistant Secretary
for Civil Rights, 1400 Independence Avenue, SW, Washington, D.C. 20250-9410; (2) fax: (202) 690-7442; or (3) email: program.
[email protected].
USDA is an equal opportunity provider, employer, and lender.
Economic
Research
Service
Technical
Bulletin
1963
September
2023
Analysis of the Current Population
Survey Food Security Supplement
Split-Panel Test
Alisha Coleman-Jensen and Matthew P. Rabbitt
Abstract
The U.S. Department of Agriculture (USDA) has monitored the extent and severity of food insecurity
in U.S. households for more than 25 years. Data on food security is collected annually as part of the
Current Population Survey Food Security Supplement (CPS-FSS). USDA, Economic Research Service
(ERS) undertook a process to update the survey instrument used for CPS-FSS data collection to ensure
that questions accurately reflect respondent engagement with the current food environment and to
decrease nonresponse on select questions and lower respondent burden. The updated instrument was
created through a process of expert content review and cognitive testing and implemented by the U.S.
Department of Commerce, Bureau of the Census in a split-panel data collection in September 2020. A
portion of the survey sample received the revised survey instrument, and the other portion received the
standard survey instrument, which was implemented in 1995 and has been consistent since 2008. This
process ensured that any unforeseen outcomes of the revisions did not affect the annual food security
estimates that USDA, ERS produced. The split panel forms two samples that were weighted to represent the U.S. population. Estimates for food spending, food security, and participation in Federal and
community nutrition assistance were compared. Based on the analysis of the split-panel test, the test
instrument performed well. The differences observed in the test and the standard instrument for food
spending and community nutrition assistance were expected, given the changes in the survey questions.
The food security measure is based on the Rasch measurement model. Rasch analysis was used to assess
the comparability of the food security questions across the split-panel samples. The analysis confirmed
that the minor changes to the food security section are unlikely to affect the measurement of food insecurity or affect the comparability of year-to-year estimates.
Keywords:
Food security, food insecurity, food spending, food pantry, soup kitchen, emergency kitchen, free
groceries, free meals, material well-being, material hardship, food security measurement, food security
data, split-panel survey, Current Population Survey Food Security Supplement, CPS-FSS
About the authors
Alisha Coleman-Jensen is the deputy director for data development and Matthew P. Rabbitt is a
research economist in the USDA, Economic Research Service (ERS) Food Economics Division.
Acknowledgments
The authors thank Anne Byrne, Michele Ver Ploeg, Christian Gregory, and Jay Variyam of USDA,
ERS for their reviews. The authors also thank Elaine Waxman from the Urban Institute, Emily
Engelhard from Feeding America, Rebecca Nemec from USDA, Office of the Chief Economist,
and reviewers from USDA, Food and Nutrition Service for their insightful feedback. Thanks also to
USDA, ERS editors Jeff Chaltas, Elaine Symanski, and Grant Wall, and Xan Holt of USDA, ERS for
layout and design.
Contents
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1
Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3
Modifications to the CPS-FSS Test Instrument. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4
Methods and Results: Food Spending . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5
Methods and Results: Food Security Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8
Questions Used To Assess the Food Security of Households in the Current Population Survey Food
Security Supplement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13
Methods and Results: Federal Nutrition Assistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .18
Methods and Results: Community Nutrition Assistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .21
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .24
Appendix A: Test Questionnaire–September 2020 Food Security Supplement . . . . . . . . . . . . . .27
ii
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
Summary
September 2023
A report summary from the Economic Research Service
Analysis of the Current Population Survey Food
Security Supplement Split-Panel Test
Alisha Coleman-Jensen and Matthew P. Rabbitt
What Is the Issue?
The U.S. Department of Agriculture (USDA) has monitored the extent and
severity of household food insecurity in the United States for more than 25
years. USDA defines food-insecure households as those that had difficulty at
some time during the year providing enough food for all their members because
of a lack of resources. USDA, Economic Research Service (ERS) sponsors the
annual, nationally representative Food Security Supplement (FSS) to the Current
Population Survey (CPS). The FSS data are collected by the U.S. Department
of Commerce, Bureau of the Census and then analyzed by ERS to produce the
annual report series titled “Household Food Security in the United States.”
Economic
Research
Service
Technical
Bulletin
1963
September 2023
Analysis of the Current
Population Survey Food Security
Supplement Split-Panel Test
Alisha Coleman-Jensen and Matthew Rabbitt
As ERS approached 25 years of food security data collection, researchers worked
with the Census Bureau and USDA’s Food and Nutrition Service to review the
FSS survey content, make revisions, and conduct cognitive interviews to ensure the data collected continued to be
relevant, current, and useful. The Census Bureau collected split-panel test data as a supplement to the September
2020 CPS to assess differences in data collected from the standard instrument and test instrument across food
spending, food security, and nutrition assistance. This report presents findings in the order they are asked about
in the modified survey. Modifications included changing the ordering of sections, updating language about food
spending, and revising questions about community nutrition assistance use.
What Did the Study Find?
ERS updated food spending questions to reflect changes in terminology, the retail environment, and technology:
• Most differences in usual food spending between the split-panel samples were not statistically significant. For
both the standard and test instrument medians, usual food spending per person per week was about $60.
• The results suggest that the modified questions functioned at least as well as the standard instrument and
show possible improvement, with slightly lower nonresponse.
ERS is a primary source of economic research and analysis from the U.S. Department of Agriculture, providing timely
information on economic and policy issues related to agriculture, food, the environment, and rural America.
www.ers.usda.gov
The food security section was moved earlier in the survey, the language used in two questions about child food
security was standardized, and a lead-in to one screening variable was modified:
• Psychometric analyses of the food security measure based on the Rasch measurement model showed that all
the food security items that comprise the adult and child food security scales captured similar levels of the
severity of food hardship in the standard and test instruments, suggesting the instruments would produce
comparable measures of food security.
• Based on the results of the psychometric analyses, modest differences in prevalence were not due to differences in how respondents interpreted the food security module. Based on the September 2020 standard
instrument, 9.7 percent of U.S. households were food insecure, and based on the September 2020 test instrument, 10.7 percent of U.S. households were food insecure. The December 2020 prevalence of food insecurity
previously published in USDA’s annual food security report was 10.5 percent.
• Differences in prevalence between the test and standard instrument may be related to differences in sample
characteristics and nonresponse bias related to the Coronavirus (COVID-19) pandemic.
• Given the findings from this study and from past studies regarding the stability of the food security scale
to similar minor modifications, the authors expect that the proposed updates to the FSS instrument will
continue to produce comparable estimates to previous years.
Questions used to assess participation in community nutrition assistance were revised significantly to ask about
receipt of free groceries and free meals. These items are not comparable because of changes in wording:
• For both questions about free groceries and free meals in the test supplement, the reported receipt was higher
than for the questions about food pantries and soup kitchens in the standard instrument. In the standard
instrument, 5.9 percent of households reported receiving food from food pantries; in the test instrument, 6.9
percent of households reported receiving free groceries. In the standard instrument, 0.5 percent of households reported receiving meals from soup kitchens; in the test instrument, 2.1 percent of households reported
receiving free meals.
• Based on expert review, cognitive testing, and the results of the test instrument, the revised questions on
community nutrition assistance appear to be an improvement and function well.
How Was the Study Conducted?
USDA, ERS developed a revised survey instrument for the FSS in collaboration with the U.S. Census Bureau
and USDA’s Food and Nutrition Service. A split-panel test of the standard and revised FSS survey instrument was
implemented in the September 2020 CPS. In the split-panel test, half of respondents received the standard survey
instrument and half received the test instrument. The data were analyzed to assess performance of the revised
instrument. The split panels form two samples that were weighted to represent the U.S. population. Estimates for
food spending, food security, and participation in Federal and community nutrition assistance were compared
across the two samples. Psychometric analyses using conditional maximum likelihood (CML) Rasch measurement
models were performed to assess how respondents interpreted the food security module.
www.ers.usda.gov
Analysis of the Current Population
Survey Food Security Supplement
Split-Panel Test
Introduction
The U.S. Department of Agriculture (USDA) has sponsored annual data collection of the Food Security
Supplement (FSS) since 1995. The supplement was first sponsored by the USDA, Food and Nutrition Service
(FNS); the USDA, Economic Research Service (ERS) assumed sponsorship beginning in 1998. The FSS has
been added to the monthly Current Population Survey (CPS) each December since 2001 to form the Current
Population Survey Food Security Supplement (CPS-FSS). Prior to that, the FSS was not in the same month
each year. The survey collects information on household food spending, food security, and the use of Federal
and community nutrition assistance programs. The CPS-FSS is used for annual monitoring of household
food security in the United States (Coleman-Jensen et al., 2022). Food insecurity is a key indicator of wellbeing for the U.S. population. Two examples of the focus on food insecurity at the Federal level are Healthy
People 2030, an initiative of the U.S. Department of Health and Human Services’ (HHS) Office of Disease
Prevention and Health Promotion, which identifies U.S. public health priorities, and the Federal Interagency
Forum on Child and Family Statistics, a collection of 23 Federal agencies involved in research and activities
related to children and families. The CPS-FSS has also been used by USDA, ERS researchers and other stakeholders to research correlates and determinants of food insecurity. For a full history of the development of the
food security measure and data, see Coleman-Jensen (2015).
The CPS-FSS survey questions have been largely unchanged since collection began in 1995. Minor modifications were made in the early years of implementation; however, the survey content and wording have been
consistent since 2008. As USDA, ERS was approaching 25 years of data collection with the CPS-FSS, agency
researchers believed it was important to critically review the survey instrument. To ensure the continued
relevance of the survey questions, ERS contracted with the U.S. Department of Commerce, Bureau of the
Census to review the CPS-FSS instrument as it was implemented at the time, develop changes as needed,
and conduct cognitive testing of any possible revisions to the survey instrument. Cognitive testing is used
to examine how well survey respondents understand survey questions—if the questions make sense to the
respondents and if they interpret the questions as intended. Cognitive testing or cognitive interviewing
involves asking respondents about how they interpreted or understood questions as they complete a survey
(for more information, see Kephart et al., 2021).
To begin, USDA, ERS researchers reviewed the survey items in collaboration with research staff at USDA,
FNS and survey methods staff at the Census Bureau’s Center for Behavioral Science Methods in the Research
and Methodology Directorate. USDA, ERS also considered previous comments about the survey received
from academic food security experts, researchers, experts in the Federal Government, and practitioners in
other organizations. The review perceived some wording in the survey items to be outdated. For example,
questions about food spending referred to produce stands and meat markets but did not ask about spending
at farmers’ markets or online food purchases. Thus, some survey items were updated. Based on the review,
the authors included minor modifications to some of the 18 items in the food security measure to standardize
consistent language and reduce burden. The burden on respondents can be reduced in a few ways, such as by
reducing the time it takes to complete a survey or making the survey easier to complete. One way to make the
1
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
survey easier to complete is to make the wording of questions more consistent. An example is to use “wasn’t
enough money for food” across items in place of “we just couldn’t afford more food” used in only a few items.
Questions regarding nutrition assistance were added to reflect new and changing Federal and community
nutrition assistance programs. For example, a new question asks about the receipt of a free or reduced-price
meal or snack at an after-school program.
In addition, the reordering of questions was proposed. The food security items were moved earlier in the
survey interview with a possible effect of reducing nonresponse to those items from respondents who would
stop part-way through and not complete the survey. A new questionnaire was developed for testing that
incorporated revisions to some survey items and a reordering of items. The revised questionnaire went
through three rounds of cognitive testing led by the Census Bureau in 2019. The findings and recommendations from the cognitive test are included in a published Census Bureau report (Kephart et al., 2021).
In 2020, USDA, ERS sought and received U.S. Office of Management and Budget (OMB) approval for a
one-time split-panel test to assess the revised CPS-FSS survey instrument that resulted from the review and
cognitive testing process. This split-panel test was a separate data collection from the regular annual CPS-FSS
data collection used to monitor the Nation’s food security. USDA, ERS researchers proposed adding a splitpanel test because they did not want the test instrument to affect annual food security estimates to cause an
unnecessary break in the long-time series. The data from this one-time split-panel test are being used to assess
the functioning of the updated questionnaire and the effect of modifying some survey items on key outcomes
of interest, like the measure of food security.
The timing of the data collection was not ideal because the split sample data were collected in September
2020 during the Coronavirus (COVID-19) pandemic. The timing may have affected response rates, nonresponse bias, and the use of nutrition assistance programs like school meal programs and community nutrition
assistance. However, the advantage of a split panel is that the sample receiving the standard instrument and
the sample receiving the test instrument were subject to the same pandemic conditions.
This report presents analyses comparing the two instruments implemented in the September 2020 CPS
supplement—the FSS standard instrument and the FSS test instrument. Based on expert review and recommendations, as well as cognitive testing, the authors believe the modifications made in the test instrument
were valid and important updates. This report documents how the modifications and updates may affect the
estimates produced from the FSS data. The authors examined the survey items that form the three sections of
USDA’s annual food security report series titled Household Food Security in the United States: food spending,
household food security, and participation in Federal nutrition assistance programs. Also examined are questions on community nutrition assistance that are presented annually in the report titled Statistical Supplement
to Household Food Security in the United States. Responses to the individual food security items are also
presented in the supplement each year (Coleman-Jensen et al., 2022b).
A preliminary version of this report with the analyses included was previously submitted to OMB’s Office
of Information and Regulatory Affairs (OIRA) as part of the Information Collection Request (ICR) process
under the Paperwork Reduction Act to obtain approval for Federal data collection. The test instrument examined in this report was submitted as part of the ICR package to obtain approval to use the revised survey
instrument beginning with the regular December 2022 CPS-FSS data collection. Approval was obtained in
2022, and the revised survey instrument was implemented in December 2022. All documentation for the
ICR package is available online at the OMB’s Office of Information and Regulatory Affairs website under
control number 0536-0043, the number assigned to the CPS-FSS.
2
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
Data
The general methodology for food security measurement is described in USDA, ERS’s annual food security
report (Coleman-Jensen et al., 2022) and other publications (Bickel et al., 2000; National Research Council,
2006).1
In a split-panel test, a portion of the survey sample receives one questionnaire, and the other portion receives
an alternate questionnaire. The sample for the split panel was intentionally chosen so that it would not
affect the regular CPS-FSS data collection. The CPS sample contains eight months-in-sample groups. The
survey is designed so that households are interviewed for 4 consecutive months. After the four monthly
interviews, respondents are out of the sample for 8 months; they are interviewed again in the same 4 months
the following year. These become their months-in-sample 5 through 8. For example, in each month of the
CPS, “month-in-sample group 1” includes respondents in the survey for the first month; that group becomes
“month-in-sample group 2” in the next monthly interview. Each month of the CPS contains respondents
from all 8 months-in-sample groups. Three-quarters of the regular monthly CPS sample members were
eligible for the September 2020 split-panel test. Only three-quarters were eligible because the authors did
not want to interview households in the test that may also be interviewed in the regular December 2020
CPS-FSS. CPS months-in-sample groups 2, 3, 4, 6, 7, and 8 were eligible to be interviewed in the September
2020 split panel. CPS months-in-sample groups 1 and 5 (one-quarter of the total sample) were not included
because those groups would be eligible for interview in the regular December 2020 CPS-FSS data collection. The eight months-in-sample groups are representative subsamples of the CPS sample. Therefore, the
September 2020 split panel was representative because it included six complete month-in-sample groups
that comprised six representative subsamples. For the split-panel test in September 2020, half of the sample
interviewed received the standard food security instrument, and half received the revised test instrument that
resulted from the expert review and cognitive interviewing.
The September 2020 CPS consisted of 46,300 households. Of these, 11,216 households were not eligible
for the food security supplement because they were in months-in-sample groups 1 or 5. The remaining
35,084 households were eligible for the split panel. For the split panel, 17,496 households were assigned to
the test instrument, and 17,588 were assigned to the original instrument. A total of 26,292 interviews were
completed. For the standard instrument, 13,160 interviews were completed (74.8 percent response). For the
test instrument, 13,132 interviews were completed (75.1 percent response).
Each split panel was weighted separately to represent the U.S. population. All statistics in this report were
calculated by applying the food security supplement weight for the appropriate panel (test instrument or
standard instrument). Unweighted estimates were specifically noted as unweighted. Unless otherwise noted,
statistical differences described in this report are significant at the 90-percent confidence level. Standard
errors were calculated using balanced repeated replication (BRR) methods based on replicate weights
computed for the September 2020 CPS-FSS by the Census Bureau. Replicate weights were computed separately for each panel.
The CPS-FSS includes questions on food spending, food security, and the use of nutrition assistance
programs. Each section was modified, although most were minor updates to the wording of questions.
Questions were also reordered, with the food security questions moved earlier in the module, immediately
after the food spending questions. As a result, the questions on Federal nutrition assistance were moved after
the food security questions. This change was made for two reasons: first, to possibly reduce missing responses
1 See also a listing of food security measurement and technical research on the USDA, ERS website under Food Security in the U.S., Survey Tools,
and Technical Reports and Food Security Measurement Research.
3
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
on the food security questions from respondents who started the survey but dropped off part way through;
and second, to reduce any possibility that asking about nutrition assistance prior to the food security items
could influence responses to the food security questions. Modifications to the instrument are shown in
appendix A, which includes the test instrument with revisions highlighted in yellow. For a summary of modifications and updates, see box, “Modifications to the CPS-FSS Test Instrument.”
Modifications to the CPS-FSS Test Instrument
Overview
The Current Population Survey Food Security Supplement (CPS-FSS) was modified and evaluated
in the split-panel test in September 2020. Modifications were made to the instrument across food
spending, food security, and nutrition assistance. Changes included reordering the sections, updating
language on food spending, and revising questions on community nutrition assistance use.
Screeners
Screener questions are used to reduce respondent burden by screening households out of those sections
if they are unlikely to experience food insecurity or participate in nutrition assistance. One change was
made to the location of the screening questions because sections of the instrument were reordered. The
screening question about running short of money and trying to make food or food money go further
(variable name “HES9”) and the food sufficiency screening question asking households to describe
whether they have enough of the kinds of food they want to eat (variable name “HESS1”) are now
placed immediately after one another. The lead-in for HES9 was removed, and the former lead-in for
HESS1 became the lead-in for HES9 in the test instrument. Because HESS1 is located earlier in the
test instrument, this question was also used to screen households into the Federal nutrition assistance
sections. In the standard instrument, only income and HES9 were used to screen households in the
Federal nutrition assistance sections. No changes to screening were made based on income. Screening
procedures in the food security section and community nutrition assistance section were unchanged
from the standard to the test instrument. (Note: HES9 and HESS1 are the variable names in the
public-use data file. In the questionnaire in the appendix, these questions are S9 and SS1.)
Food Spending
Questions in the section on food spending were modified to reflect changes in terminology, the retail
environment, and technology. In both instruments, the section first asks respondents to recall all
the places they purchased food and how much they spent, and then asks them to report usual food
spending. Questions that ask respondents to think about all the places they spent money on food
purchases and how much they spent last week were modified. For example, the new wording asks
about online food purchases and purchases at farmers’ markets that were not included when the survey
was developed in the 1990s.
Food Security
The section on food security was moved earlier in the survey instrument. A few questions about child
food insecurity were modified to standardize the resource constraint to “there wasn’t enough money
for food.” Also, the lead-in to the screening variable (HES9) was dropped due to moving sections in
the questionnaire.
4
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
Federal Nutrition Assistance
The section on Federal nutrition assistance was moved after the food security section. The lead-in to
the first question (HESP1) changed due to moving the section. Questions on school meals were modified slightly to refer to “reduced-price” meals instead of “reduced-cost” meals. A new question was
added on receipt of free or reduced-price afterschool meals and snacks.
Community Food and Nutrition Assistance
New questions that ask about the receipt of free groceries and free meals were added to the section on
community food and nutrition assistance. These questions were tested and recommended by experts in
the charitable feeding sector. The section no longer asks separate questions about free meals received by
older adults but includes those programs in the more general questions about the receipt of free meals.
Methods and Results: Food Spending
In USDA’s annual food security reports (e.g., Coleman-Jensen et al., 2021), food spending is reported for
various demographic and household groups and by food security status. The survey questions about food
spending aim to help respondents reflect on all the places the household purchased food in the previous
week. USDA does not report food spending for the individual questions in the survey. Rather, respondents
are first asked to report the amount of money their household spent on food in the week before the interview,
including purchases made with Supplemental Nutrition Assistance Program (SNAP) benefits. Total spending
for food, based on responses to the series of questions about places where food was purchased and how much
was spent, was verified with the respondent. The respondent is then asked how much the household usually
spends on food per week.
The questions in this section were modified based on expert review and cognitive testing to better reflect the
places where households currently purchase food compared with when the instrument was first developed.
For example, the test instrument asks about spending at Walmart and Target in addition to supermarkets
and grocery stores. The test instrument also asks about food purchases at dollar stores, pharmacies, club
stores, farmers’ markets, and online—all places that were not previously included. Low-income households
continued to be prompted to include spending on food purchases made with SNAP benefits, as has been
implemented in previous FSS data collections. In one case, where the food spending question was modified to include convenience stores where SNAP benefits may be used, the prompt to include purchases made
with SNAP was included in the revised test question. The survey asks about the specific places to prompt the
respondent to recall all the places they purchased food. The total usual food spending is the focus of USDA’s
analysis of this section, not the specific spending at each type of retail outlet. The question that forms the
basis for usual food spending was not revised in the test instrument. However, since the questions before it
about all the places where the households purchased food in the last week were revised in the test instrument,
differences may exist in how respondents reported their usual food spending between the test and standard
instrument.
Table 1 is formatted similarly to table 7 in USDA’s annual food security reports. Usual food spending by split
panel is shown in the table. Usual food spending was adjusted for household size and composition in two
ways. First, researchers divided each household’s usual weekly food expenditure by the number of household
5
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
members, yielding the “usual weekly food spending per person” for that household. The second adjustment
more precisely accounts for the different food needs of households by comparing each household’s usual
food spending to the estimated cost of the Thrifty Food Plan (TFP) for that household in September 2020
(USDA, 2020).2 USDA’s TFP serves as a national standard for a nutritious, minimal-cost diet. Each household’s reported usual weekly food spending was divided by the household-specific cost of the TFP based on
the age- and gender-specific cost of the TFP for each household member and the number of persons in the
household.
The medians of each of the two food-spending measures (spending per person per week and total spending
relative to the cost of the September 2020 TFP) were estimated at the national level and for households
in various categories. Medians are reported rather than averages (means) because medians are not unduly
affected by some unexpectedly high values of food spending. Thus, the median better reflects what a typical
household spent.
Some interviewed households did not respond to the food-spending questions or reported zero usual food
spending and were excluded from the analysis. For the standard instrument, the exclusions represented 7.8
percent of all U.S. households (6.8 percent of unweighted respondents). For the test instrument, the exclusions represented 6.9 percent of all households (6.3 percent of unweighted respondents). Nonresponse to
the food-spending questions was somewhat lower with the test instrument, which could be due to random
variation or a better understanding of the question based on improvements made from the expert review
and cognitive interviews. In other words, the revised food-spending questions may have elicited more valid
responses.
The results for food spending for the standard instrument and test instrument are shown in table 1. The
left side of the table shows the standard instrument sample; the right side shows the test instrument sample.
According to the findings for the standard instrument, the reported usual weekly food spending per person
was $60, and the median weekly food spending relative to the cost of the TFP was 1.32. The TFP adjusts for
food price inflation and adjusts more precisely for the food needs of persons in different age or gender categories. A ratio of household food spending relative to the TFP that is above 1.0 indicates the household spends
more than the cost of the TFP; a ratio below 1.0 means the household spends less than the cost of the TFP.
The findings for the test instrument showed that the reported usual weekly food spending per person was also
$60, and the median weekly food spending relative to the cost of the TFP was 1.36.
Differences between the panel samples in food spending relative to the TFP were tested for statistical significance. Differences were statistically significant for only a few categories (denoted by an asterisk in table 1).
When differences were significant, reported food spending was higher for test instrument households. Higher
reported food spending suggests an improvement in the instrument because food spending in the CPS-FSS
may be underreported compared with other sources. For example, USDA, ERS’s reported normalized annual
food expenditures by households suggest a weekly average in food spending of $225 per household in 2020
and $98 per capita (Zeballos & Sinclair, 2023; see the USDA, ERS Food Expenditure Series on the USDA,
ERS website under Data Products; specifically under Current Food Expenditure Series, see the excel file
“Normalized food expenditures by all purchasers and household final users”). Some seasonal variation exists
in food spending, so the total annual expenditures from the food expenditure series used to approximate
average weekly spending are not precisely comparable with food spending reported for a single week. Medians
are reported in table 1. For those with positive reported food spending, the average or mean food spending
2 The cost of the Thrifty Food Plan (TFP) is revised each month to account for inflation in food prices and was revised significantly in 2021
(USDA, 2021). For this analysis, TFP costs are from prior to the 2021 revision. The TFP is used as the basis for setting maximum SNAP benefits.
6
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
per person per week in the standard instrument was $68; the average food spending per person per week in
the test instrument was $69.
Most differences in usual food spending between the split-panel samples are not statistically significant.
However, the results suggest that the modified questions functioned at least as well as the standard instrument and show possible improvement, with slightly lower nonresponse and slightly higher reported food
spending.
Table 1
Median weekly household food spending per person and relative to the household cost of the
Thrifty Food Plan (TFP), by instrument version, September 2020
Standard instrument
Median weekly food spending
Number of
households1 in
thousands
Category
All households
120,143
Test instrument
Median weekly food spending
Per peson
Relative to
household
cost of
September
2020 TFP
U.S. dollars
Ratio
60.00
1.32
121,280
Number of
households1
in thousands
Per peson
Relative to
household
cost of
September
2020 TFP
U.S. dollars
Ratio
60.00
1.36
Household composition:
With children <18 years
34,860
48.00
1.21
35,854
50.00
1.24
At least one child <6 years
14,617
44.00
1.28
14,818
43.33
1.25
Married-couple families
22,876
50.00
1.25
23,479
46.67
1.22
Female head, no spouse
8,490
45.00
1.16
9,127
50.00
1.26 *
Male head, no spouse
3,039
42.00
1.09
2,790
50.00
1.18
NA
NA
NA
NA
NA
NA
85,283
65.00
1.42
85,426
65.00
1.44
50,494
57.50
1.29
50,934
60.00
1.36 *
Other household with child2
With no children <18 years
More than one adult
Women living alone
19,108
70.00
1.52
19,141
70.00
1.52
Men living alone
15,681
80.00
1.63
15,351
80.00
1.63
Adults aged 65 and older
36,555
Adults aged 65 and older living
13,374
alone
Race/ethnicity of household reference persons:
50.00
1.22
36,874
55.00
1.24
60.00
1.30
13,937
60.00
1.3
60.00
1.40
79,521
60.00
1.42
White, non-Hispanic
79,012
Black, non-Hispanic
15,339
50.00
1.14
15,073
50.00
1.19
Hispanic3
16,697
50.00
1.23
17,591
50.00
1.22
Other, non-Hispanic
9,095
60.00
1.38
9,095
56.00
1.29
Under 1.00
9,486
50.00
1.08
8,783
50.00
1.06
Under 1.30
14,616
50.00
1.08
14,598
50.00
1.08
Household income-to-poverty ratio:
Under 1.85
23,323
50.00
1.09
23,314
50.00
1.09
1.85 and over
66,939
65.00
1.47
67,051
63.33
1.47
Income unknown
29,882
50.00
1.22
30,915
55.00
1.30 *
Area of residence:4
Inside metropolitan area
103,172
60.00
1.38
103,910
60.00
1.40
In principal cities
35,232
60.00
1.43
35,338
60.00
1.42
Not in principal cities
52,485
60.00
1.40
52,593
60.00
1.42
Outside metropolitan area
16,971
50.00
1.13
17,370
50.00
1.14
Continues on next page >
7
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
Continued from previous page
Standard instrument
Median weekly food spending
Number of
households1 in
thousands
Category
Per peson
Relative to
household
cost of
September
2020 TFP
U.S. dollars
Ratio
60.00
1.47
Test instrument
Median weekly food spending
Per peson
Relative to
household
cost of
September
2020 TFP
U.S. dollars
Ratio
20,695
62.50
1.48
Number of
households1
in thousands
Census geographic region:
Northeast
20,462
Midwest
25,509
50.00
1.26
25,946
56.67
1.30
South
46,872
55.00
1.27
46,871
57.50
1.30
West
27,300
60.00
1.41
27,768
60.00
1.33
NA = not available because fewer than 100 interviewed households in the category.
* Indicates a statistically significant difference between the standard and test instrument in the ratio of food spending relative to the
household cost of the September TFP, with 90 percent confidence.
1 Totals exclude households that did not answer the questions about spending on food or reported zero usual food spending. For the
standard instrument, these exclusions represented 7.8 percent of all households (6.8 percent of unweighted respondents). For the test
instrument, these exclusions represented 6.9 percent of all households (6.3 percent of unweighted respondents).
2 Households with children in complex living arrangements.
3 Hispanics may be of any race.
4 Metropolitan refers to residence areas as defined in the 2013 Office of Management and Budget Delineation files.
Source: USDA, Economic Research Service using data from U.S. Department of Commerce, Bureau of the Census, September 2020
Current Population Survey, Food Security Supplement.
Methods and Results: Food Security Measures
The primary focus of the CPS-FSS is to collect data on food security in U.S. households. The data provide
information for annual monitoring of the Nation’s well-being and enable research to better understand the
determinants of and policy impacts on food security. It is important to understand how any changes to the
survey instrument may affect food insecurity prevalence estimates. The methods underlying the food security scale have been described in other publications (Bickel et al., 2000; Coleman-Jensen et al., 2022). The
food security questions and rules for assigning each household a food security status are shown in the box
“Questions Used To Assess the Food Security of Households in the CPS Food Security Supplement.” The
measurement methods are unchanged between the standard and test instruments. In the test instrument,
some sections were reordered, which resulted in a change to the lead-ins to the items used to screen households into the food security module (see box, “Modifications to the CPS-FSS Test Instrument” and in-depth
on pages 4-5).
The characteristics of respondent households for each version of the instrument are shown in table 2. Because
food security status is closely related to respondent characteristics, it helps to identify any differences in the
sample characteristics. These estimates are weighted and include households that responded to the split panel
and had valid food security status. The total number of households across selected household characteristics,
as well as the percentage of households for each characteristic, are shown in the table. The characteristics of
respondent households with valid food security status are quite similar across the standard and test instruments. The largest difference in the percentage of households is for households with an income-to-poverty
8
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
ratio under 1.0. For the standard instrument, 8.0 percent of all respondent households were in this lowest
income category. For the test instrument, 7.2 percent of all respondent households were in the under-1.0
household income-to-poverty ratio category. However, a nearly equal percentage of households in each split
panel had incomes below 130 percent of the Federal poverty line (12.1 percent of households for the standard
instrument and 12.0 percent for the test instrument). The test instrument sample had a somewhat larger share
of households with children compared with the standard instrument sample—29.1 percent of test instrument
households included children, and 28.6 percent of standard instrument households included children. Note
that the characteristics of respondent households are not expected to be influenced by the different instruments because households were randomly assigned to either the test or standard instrument. Rather, sample
distributions are provided to understand how the two samples may differ and whether they may affect differences in items between the instruments.
Overall food insecurity prevalence rates between the September 2020 standard and test instrument panel
samples were compared. Table 3, panel A, shows statistically significant differences in both the prevalence
of food insecurity and very low food security between the standard instrument and the test instrument.
The prevalence of food insecurity in the standard instrument in September 2020 was 9.7 percent, significantly below the prevalence in the September test instrument (10.7 percent). The prevalence of very low food
security with the September standard instrument was 3.5 percent, significantly below the prevalence in the
September test instrument (4.0 percent).
Table 3, panel B shows the results for households with children. The prevalence of household food insecurity
for households with children was statistically significantly different between the September 2020 standard
instrument (12.8 percent) and the September test instrument (14.6 percent).
Differences in the prevalence of food insecurity between the test instrument and standard instrument may
be partly due to sampling error, which had a greater impact on the smaller samples in the split panel that was
potentially more problematic during the COVID-19 pandemic when response rates were lower. Sampling
variation is inherent in any survey that relies on a sample. However, sampling error during the pandemic may
have been greater, resulting in nonresponse bias. Sampling error and nonresponse bias may have resulted in
greater variation in the food security estimates than would be typical of other full CPS-FSS data collections.
Respondents were assigned to a given panel randomly, so no systematic differences between the panel samples
should exist. However, nonresponse bias or sampling error related to the pandemic challenges may have
resulted in unsystematic differences between the samples. Evidence from a Census Bureau analysis suggests
that nonresponse during the pandemic was more strongly associated with income than in prior years of the
CPS data collections (Rothbaum & Bee, 2021). This is an area for future research with the CPS-FSS.
The response rates were similar between the test and standard instruments. For all households in both instrument panel samples, 0.2 percent of all households were missing food security status because they had no
valid responses on those items (weighted; see footnote 1, table 3). For the standard instrument, 22 households (unweighted) were excluded, and for the test instrument, 26 households (unweighted) were excluded
due to having no valid responses to the food security questions. For households with children, 0.1 percent
of households (weighted) in the standard instrument panel and 0.3 percent of households (weighted) in the
test instrument panel were excluded due to nonresponse to the food security items (see footnote 2, table 3).
For the standard instrument, three households with children (unweighted) in the sample were excluded, and
for the test instrument, seven households with children (unweighted) in the sample were excluded based on
nonresponse to the food security items.
9
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
Table 4 shows the percentage of households affirming each of the individual food security items. A lower
percentage of households in the September standard instrument affirmed the food security items when
compared with households in the September test instrument, with statistically significant differences marked
with an asterisk. Note that the child items with changes in wording between the test and standard instrument were not statistically significantly different.
Given the modest differences observed in the prevalence of food insecurity and rates of affirmative responses
to the food security items used to construct food security measures between the standard and test supplements, additional psychometric analyses were conducted to better understand the differences and their
implications for future food security measurement activities. Psychometric analyses specifically assess how
households responded to the food security items and if household response patterns were as expected. These
analyses can also be used to assess whether food security measures across different surveys are comparable.
Two measurement subsamples were constructed for the psychometric analyses based on households administered the standard and test instruments in the September 2020 CPS-FSS. Households were included in the
standard and test measurement subsamples if their incomes were below 185 percent of the Federal poverty
line,3 had no incomplete or missing responses to the food security questions, and had no extreme raw scores
on the 10-item adult and 8-item child food security scales. Extreme raw scores are a measurement term
simply meaning the respondents have zero affirmative responses or all affirmative responses. These households were omitted from the subsamples because households with extreme raw scores are not identified under
the assumptions of the conditional maximum likelihood (CML) Rasch measurement model USDA uses
to develop and continually monitor the properties of adult and child food security scales. Households with
extreme raw scores do not provide information to the Rasch measurement model and are, therefore, not
useful in comparing the responses to the food security items across the instruments to determine the severity
of items. Most households with extreme scores had a zero, meaning no affirmative responses to the food security items. The resulting subsamples for the adult food security scale consisted of 1,192 and 1,241 households
administered the standard and test instruments, respectively. For the analysis of the child food security scales,
the subsamples consisted of 316 households administered the standard instrument, and 296 households
administered the test instrument. The subsamples consisted of households that affirmed at least one of, but
not all, the food security items.
The four measurement subsamples were used to separately estimate the parameters of the Rasch measurement model for the adult and child food security scales for households administered the standard and test
instruments. Researchers used the ERS Rasch program developed in SAS by USDA, ERS. Based on the
measurement theory underlying the Rasch measurement model, any differences in the estimates of severity
parameters for households administered the standard and test instruments may indicate respondents perceive
the food security questions differently or reflect differences in the way they experience food hardships. If this
were to occur, then the standard and test instruments would measure different levels of the severity of food
hardship, leading to biased comparisons of food security measures—such as the prevalence of food insecurity—based on these scales.
Estimates of the Rasch severity parameters for the items for the adult food security scale are shown separately
in figure 1 for households administered the standard and test instruments. The food security items that are
3 The authors use this income threshold for their analyses because it is the initial screening threshold for a household to be administered the food
security items in the CPS-FSS. While this results in the exclusion of a small proportion of high-income households from analyses, it mitigates bias
associated with income that could result from the screening procedure (Nord, 2012). Households are also initially screened into the food security items
based on reported food hardship in addition to income, but this is a much smaller group, and the respondents were less likely to answer more than the
first three questions in the food security scale. The advantage of screening based on income is to reduce respondent burden for higher income households unlikely to experience food insecurity. Early research suggested that only a very small proportion of higher income households that were screened
out of the food security items based on income would be classified as food insecure were they to be asked the full module (Bickel et al., 2000).
10
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
estimated to capture a greater severity of food hardship in the test instrument relative to the standard instrument are indicated by a red X located above the equal severity line. Those that capture a lesser severity of
food hardship in the test instrument are indicated by a red X below the equal severity line. All the items that
comprise the adult food security scale captured similar levels of the severity of food hardship in the standard
and test instruments, suggesting these items would produce comparable measures of adult food security.
Figure 1
Comparisons of the adult food security scale question severity parameters, 2020 September
CPS-FSS test versus standard instrument
Item severity, test instrument
9
whldayf
whlday
losewt
8
hungry
7
cutskipf
6
cutskip
eatless
5
4
balmeal
fnotlast
3
worried
2
2
3
4
5
6
7
8
9
Item severity, standard instrument
Test instrument
Equal severity
balmeal = Couldn’t afford to eat balanced meals; cutskip = Adult(s) cut size of meals or skipped meals; cutskipf = Adult(s) cut size or
skipped meals in 3 or more months; eatless = Respondent ate less than felt he/she should; fnotlast = Food bought didn’t last, and (I/
we) didn’t have money to get more; hungry = Respondent hungry but didn’t eat because couldn’t afford food; losewt = Respondent
lost weight; whlday = Adult(s) did not eat for a whole day; whldayf = Adult(s) did not eat for whole day in 3 or more months; worried
= Worried food would run out before (I/we) got money to buy more. (For complete wording of questions, see box “Questions Used To
Assess the Food Security of Households in the Current Population Survey Food Security Supplement.”)
Note: Question severity parameters estimated from separate conditional maximum likelihood (CML) Rasch measurement models for
households administered the standard and test instruments in the 2020 September Current Population Survey Food Security Supplement.
Source: USDA, Economic Research Service using data from U.S. Department of Commerce, Bureau of the Census, September 2020
Current Population Survey Food Security Supplement.
Figure 2 provides estimates of the Rasch severity parameters for the child food security scale items, estimated
separately for households administered the standard and test instruments. Like the findings for the adult
food security items, the findings of the child food security items were similar for households administered the
standard and test instruments. Therefore, food security measures based on these instruments captured similar
experiences and are comparable. In addition, the proposed wording changes for “The child in (my/our) house11
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
hold was/The children were) not eating enough because (I/we) just couldn't afford enough food” and “In the
last 12 months, (was the child/were the children) ever hungry but you just couldn't afford more food?” does
not appear to have affected the way households with children interpreted these items. See box “Questions
Used To Assess the Food Security of Households in the CPS Food Security Supplement” for a comparison of
wording between the standard and test questions.
Figure 2
Comparisons of the child food security scale question severity parameters, 2020 September
CPS-FSS test versus standard instrument
Severity parameter, test instrument
13
chwhlday
12
11
chskipf
chskip
10
chhungry
9
chcut
8
chenuf
7
6
chbal
5
4
chfewfd
3
2
2
3
4
5
6
7
8
9
10
11
12
13
Severity parameter, standard instrument
Test instrument
Equal severity
chbal = Couldn’t feed child(ren) balanced meals; chcut = Cut size of child(ren)’s meals; chenuf = Child(ren) were not eating enough;
chfewfd = Relied on few kinds of low-cost food to feed child(ren); chhungry = Child(ren) were hungry; chskip = Child(ren) skipped
meals; chskipf = Child(ren) skipped meals in 3 or more months; chwhlday = Child(ren) did not eat for a whole day. (For complete
wording of questions, see box “Questions Used To Assess the Food Security of Households in the Current Population Survey Food
Security Supplement.”)
Note: Question severity parameters estimated from separate conditional maximum likelihood (CML) Rasch measurement models for
households administered the standard and test instruments in the 2020 September Current Population Survey Food Security Supplement (CPS-FSS).
Source: USDA, Economic Research Service using data from U.S. Department of Commerce, Bureau of the Census, September 2020
CPS-FSS.
While the psychometric analyses provide evidence that standard and test instruments produced similar food
security measures, it is important to note that the test instrument screened in more households than the
standard supplement. Screeners are used to reduce respondent burden so that higher income households or
households showing no food hardship are not asked additional food security questions. A similar proportion of households administered the standard instrument (38.2 percent) and test instrument (38.1 percent)
12
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
were initially screened into the food security module (based on income and variables HES9 and HESS1).
However, households administered the test instrument were more likely than households administered the
standard instrument to screen past the first internal food security screener (18.2 percent versus 17.1 percent)
and second screener (8.3 percent versus 7.3 percent). Households administered the standard instrument (12.8
percent) and test instrument (12.9 percent) had similar rates of passing the child food insecurity screener.
Given the decline in response rates during the past few years to the monthly CPS and CPS-FSS, screening
additional households through the initial and internal food security screeners to the most severe food security
items ensures that USDA, ERS is able to reliably produce statistics for even the most severe forms of food
insecurity. The differences in the percentages of households screened through the food security module based
on the version of the instrument may be related to the placement of the food security items earlier in the test
instrument survey before the questions on nutrition assistance, but the authors could not definitively determine that to be the cause.
The Rasch analyses suggest that the test instrument measured food insecurity comparably to the standard
USDA instrument. When USDA instituted similar modifications in previous years, those modifications also
did not appreciably affect the food security measure. For example, in the 2007 CPS-FSS, food security items
were reordered in the questionnaire so that all household- and adult-referenced questions were administered
first, followed by the child-referenced items. Changes were also made in internal screener specifications to
accommodate the new order of items in 2007, but those changes resulted in only negligible differences in
item responses (Nord, 2008). The changes in item ordering did not result in a break in the food security
series, and food insecurity estimates were comparable to previous years (Nord et al., 2008). Also, in the 2006
CPS-FSS questionnaire, the resource constraint was standardized for the items addressed to adults, “Were
you ever hungry?” and “Did you lose weight?” to “Because there wasn’t enough money for food,” so that
the wording of the resource constraint was more consistent across questions (Nord et al., 2006; Nord et al.,
2007). Again, this minor change did not result in a break in the series or a change in the comparability of the
estimates from year to year. Given the current findings and past findings regarding the stability of the food
security scale to similar minor modifications, the authors expect that the proposed updates to the CPS-FSS
instrument will continue to produce estimates comparable to previous years.
Questions Used To Assess the Food Security of Households in the Current
Population Survey Food Security Supplement
1. “We worried whether our food would run out before we got money to buy more.” Was that often,
sometimes, or never true for you in the last 12 months?
2. “The food that we bought just didn’t last, and we didn’t have money to get more.” Was that often,
sometimes, or never true for you in the last 12 months?
3. “We couldn’t afford to eat balanced meals.” Was that often, sometimes, or never true for you in the
last 12 months?
4. In the last 12 months, did you or other adults in the household ever cut the size of your meals or skip
meals because there wasn’t enough money for food? (Yes/No)
5. (If yes to question 4) How often did this happen—almost every month, some months but not every
month, or in only 1 or 2 months?
6. In the last 12 months, did you ever eat less than you felt you should because there wasn’t enough
money for food? (Yes/No)
13
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
7. In the last 12 months, were you ever hungry, but didn’t eat, because there wasn’t enough money for
food? (Yes/No)
8. In the last 12 months, did you lose weight because there wasn’t enough money for food? (Yes/No)
9. In the last 12 months did you or other adults in your household ever not eat for a whole day because
there wasn’t enough money for food? (Yes/No)
10. (If yes to question 9) How often did this happen—almost every month, some months but not every
month, or in only 1 or 2 months?
Note: Questions 11–18 were asked only if the household included children age 0–17.
11. “We relied on only a few kinds of low-cost food to feed our children because we were running out
of money to buy food.” Was that often, sometimes, or never true for you in the last 12 months?
12. “We couldn’t feed our children a balanced meal because we couldn’t afford that.” Was that often,
sometimes, or never true for you in the last 12 months?
13. Standard: “The children were not eating enough because we just couldn’t afford enough food.”
Was that often, sometimes, or never true for you in the last 12 months?
Test: “The children were not eating enough because there wasn’t enough money for food.” Was that
often, sometimes, or never true for you in the last 12 months?
14. In the last 12 months, did you ever cut the size of any of the children’s meals because there wasn’t
enough money for food? (Yes/No)
15. Standard: In the last 12 months, were the children ever hungry but you just couldn’t afford more
food? (Yes/No)
Test: In the last 12 months, were the children ever hungry because there wasn’t enough money for
food? (Yes/No)
16. In the last 12 months, did any of the children ever skip a meal because there wasn’t enough money
for food? (Yes/No)
17. (If yes to question 16) How often did this happen—almost every month, some months but not
every month, or in only 1 or 2 months?
18. In the last 12 months, did any of the children ever not eat for a whole day because there wasn’t
enough money for food? (Yes/No)
Coding of Responses
Questions 1–3 and 11–13 are coded as affirmative (i.e., possibly indicating food insecurity) if the
response is “often” or “sometimes.” Questions 5, 10, and 17 are coded as affirmative if the response is
“almost every month” or “some months but not every month.” The remaining questions are coded as
affirmative if the response is “yes.”
14
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
Assessing Food Security Status in Households Without Children
Households without children are classified as food insecure if they report 3 or more indications of
food insecurity in response to the first 10 questions; households are classified as having very low food
security if they report 6 or more food-insecure conditions out of the first 10 questions.
Assessing Food Security Status in Households with Children Age 0–17
Households with children are classified as food insecure if they report 3 or more indications of food
insecurity in response to the entire set of 18 questions; households are classified as having very low food
security if they report 8 or more food-insecure conditions in response to the 18 questions.
The food security status of children in the household is assessed by responses to the child-referenced
questions (questions 11–18). Households reporting two or more of these conditions are classified as
having food insecurity among children. Households reporting five or more are classified as having very
low food security among children.
Table 2
Characteristics of respondent households with valid food security status, by instrument version,
September 2020
Standard instrument
Test instrument
Number of
households in
thousands
Percent of
total number of
households
Number of
households in
thousands
Percent of
total number of
households
130,124
100.0
129,976
100.0
37,275
28.6
37,864
29.1
Category
All households1
Household composition:
With children <18 years
With children <6 years
15,750
12.1
15,573
12.0
Married-couple families
24,101
18.5
24,679
19.0
Female head, no spouse
9,317
7.2
9,645
7.4
Male head, no spouse
3,362
2.6
3,050
2.3
Other household with child2
495
0.4
490
0.4
92,849
71.4
92,112
70.9
More than one adult
54,719
42.1
54,553
42.0
Women living alone
20,997
16.1
20,797
16.0
With no children <18 years
Men living alone
17,133
13.2
16,762
12.9
40,585
31.2
40,382
31.1
15,200
11.7
15,528
11.9
White, non-Hispanic
84,804
65.2
84,934
65.3
Black, non-Hispanic
16,940
13.0
16,492
12.7
Hispanic3
18,361
14.1
18,764
14.4
Other, non-Hispanic
10,018
7.7
9,786
7.5
With adults aged 65 and older
Adults aged 65 and older living
alone
Race/ethnicity of households:
Continues on next page >
15
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
Continued from previous page
Standard instrument
Test instrument
Number of
households in
thousands
Percent of
total number of
households
Number of
households in
thousands
Percent of
total number of
households
10,357
8.0
9,392
7.2
Under 1.30
15,765
12.1
15,538
12.0
Under 1.85
24,902
19.1
24,525
18.9
1.85 and over
69,962
53.8
69,869
53.8
Income unknown
35,260
27.1
35,582
27.4
111,982
86.1
111,484
85.8
Category
Household income-to-poverty ratio:
Under 1.00
Area of residence:⁴
Inside metropolitan area
In principal cities
38,518
29.6
38,340
29.5
Not in principal cities
56,628
43.5
55,862
43.0
Outside metropolitan area
18,142
13.9
18,492
14.2
Census geographic region:
Northeast
22,326
17.2
22,181
17.1
Midwest
27,885
21.4
27,859
21.4
South
50,631
38.9
50,238
38.7
West
29,282
22.5
29,698
22.8
1 Totals exclude households for which food security status is unknown because household respondents did not give a valid response
to any of the questions in the food security scale. For all respondent households, these exclusions represent for the September 2020
standard instrument: 218,000 households (0.2 percent of all households) and for the September 2020 test instrument: 315,000 households (0.2 percent of all households). Totals also exclude households that did not respond to the September Food Security Supplement. Survey weights are adjusted to account for supplement nonresponse so that weighted estimates from each split panel are
nationally representative. Percentages may not sum to 100 due to rounding.
2 Households with children in complex living arrangements.
3 Hispanics may be of any race.
4 Metropolitan refers to residence areas as defined in the 2013 Office of Management and Budget Delineation files.
Source: USDA, Economic Research Service using data from U.S. Department of Commerce, Bureau of the Census, September 2020
Current Population Survey, Food Security Supplement.
16
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
Table 3
Households by food security status and instrument version, September 2020
Panel A: All households
Food insecure
Total1
Category
Food secure
All
With low food
security
With very low
food security
Thousands
Thousands
Percent
Thousands
Percent
Thousands
Percent
Thousands
Percent
130,124
117,511
90.3
12,613
9.7*
8,025
6.2
4,588
3.5*
129,976
116,080
89.3
13,896
10.7
8,738
6.7
5,158
4.0
September standard instrument
All households
September test instrument
All households
Panel B: Households with children
Total2
Category
Food-secure
households
Food-insecure
households
Households
with food-insecure children
Households
with very low
food security
among children
Thousands
Thousands
Percent
Thousands
Percent
Thousands
Percent
Thousands
Percent
37,275
32,506
87.2
4,769
12.8*
2,320
6.2
200
0.5
37,864
32,331
85.4
5,533
14.6
2,707
7.1
311
0.8
September standard instrument
All households with children
September test instrument
All households with children
*Difference between the September standard instrument and September test instrument is statistically significant with 90 percent
confidence (t >1.645).
1 Totals exclude households for which food security status is unknown because household respondents did not give a valid response
to any of the questions in the food security scale. For all households (panel A), these exclusions represent:
• September 2020 standard instrument: 218,000 households (0.2 percent of all U.S. households).
• September 2020 test instrument: 315,000 households (0.2 percent of all U.S. households).
2 Totals exclude households for which food security status is unknown because household respondents did not give a valid response
to any of the questions in the food security scale. For households with children (panel B), these exclusions representations:
• September 2020 standard instrument: 28,000 households (0.1 percent of all U.S. households).
• September 2020 test instrument: 115,000 households (0.3 percent of all U.S. households).
Source: USDA, Economic Research Service using data from U.S. Department of Commerce, Bureau of the Census, September 2020
Current Population Survey, Food Security Supplement.
17
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
Table 4
Responses to items in the food security scale, standard instrument and test instrument, September
2020
Households affirming item
Scale item1
Standard
Test
Percent of U.S. households2
Household items:
Worried food would run out before (I/we) got money to buy more
14.0
14.7
Food bought didn't last and (I/we) didn't have money to get more
10.9
11.6
Couldn't afford to eat balanced meals
10.4
11.8*
Adult(s) cut size of meals or skipped meals
5.7
6.3*
Respondent ate less than felt he/she should
5.7
6.5*
Adult(s) cut size or skipped meals in 3 or more months
4.1
4.7*
Respondent hungry but didn't eat because couldn't afford
3.0
3.2
Respondent lost weight
2.0
2.2
Adult(s) did not eat for whole day
1.1
1.1
Adult(s) did not eat for whole day in 3 or more months
0.8
0.8
Relied on few kinds of low-cost food to feed child(ren)
11.6
11.8
Couldn't feed child(ren) balanced meals
6.6
7.4
Child(ren) were not eating enough (question revised in test instrument to “…
there wasn’t enough money for food?”)
3.0
2.9
Adult items:
Child items:
Cut size of child(ren)’s meals
Child(ren) were hungry (question revised in test instrument to “…there wasn’t
enough money for food?”)
Child(ren) skipped meals
1.5
1.6
0.6
1.0
0.4
0.6
Child(ren) skipped meals in 3 or more months
0.2
0.4
Child(ren) did not eat for whole day
0.1
0.1
*Difference between the September standard instrument and September test instrument is statistically significant with 90 percent
confidence (t > 1.645).
1
The full wording of each question includes explicit reference to resource limitation. See box “Questions Used To Assess the Food
Security of Households in the Current Population Survey Food Security Supplement” for item wording.
2
Households not responding to an item are omitted from the calculations of percentages for that item. Household without children
are omitted from the calculation of child-referenced items.
Source: USDA, Economic Research Service using data from U.S. Department of Commerce, Bureau of the Census, September 2020
Current Population Survey Food Security Supplement.
Methods and Results: Federal Nutrition Assistance
USDA’s annual food security report presents statistics on the use of Federal nutrition assistance by food security status. The CPS-FSS data on nutrition assistance are not intended to describe total participation in these
programs. Participation rates of eligible households and characteristics of participation are available from
USDA, FNS. The data in the CPS-FSS and information presented in USDA’s report are meant to inform
understanding of the relationship between nutrition assistance programs and food security.
18
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
In the CPS-FSS, all households with reported annual income below 185 percent of the Federal poverty
threshold were asked questions about their participation in Federal nutrition assistance. To minimize respondent burden in the standard instrument, households with annual incomes above that range were not asked
these questions unless they indicated some level of difficulty in meeting their food needs on the first of the
two preliminary screener questions asked of all households (HES9). In the test instrument, because of the
change in the location of the nutrition assistance questions, the second preliminary screener question on food
insufficiency was also used as a screener question for the nutrition assistance items.
In the split panel, a few differences existed between the standard and test instruments. In the test instrument,
the questions on Federal nutrition assistance were moved after the questions on food security. The lead-in to
one of the screener questions (HES9) was also changed due to the change in location of the question. The
lead-in preceding HES9 in the test instrument was formerly the lead-in for HESS1 (food sufficiency) in the
standard instrument. Thus, the change was an alteration in which item the lead-in preceded rather than an
entirely new lead-in. The previous lead-in to HES9 was removed from the test instrument. The standard and
test lead-ins for HES9 are:
Standard: People do different things when they are running out of money for food in order to make
their food or their food money go further.
In the last 12 months, since September of last year, did you ever run short of money and try to make
your food or your food money go further?
Test: The next questions are about the food eaten in your household in the last 12 months, since
September of last year, and whether you were able to afford the food you need.
In the last 12 months, since September of last year, did you ever run short of money and try to make
your food or your food money go further?
Also, a lead-in for the question about participation in SNAP was added (HESP1) in the test instrument due
to the new location in the instrument. There was no lead-in for the standard instrument because the question
about SNAP immediately followed question HES9. For both these items, the wording of the question did not
change. Questions about free and reduced-price lunch were reworded to use reduced-price language instead
of reduced-cost meals, which is more consistent with the Federal programs.
Standard: In the past 12 months, did (you/anyone in this household) get SNAP or food stamp benefits?
Test: Sometimes people need help getting food for their household. There are many programs that can
help.
In the past 12 months, did (you/anyone in this household) get SNAP or food stamp benefits?
A new question was added about the receipt of free or reduced-price meals or snacks at an afterschool
program. USDA, FNS staff recommended that the authors ask about these programs during the expert
review of the CPS-FSS instrument. Afterschool supper programs have become more common in recent years.
There was a slight error in the test instrument for this item, and only households that reported receipt of free
or reduced-price school lunches were asked the item. In the regular implementation of the revised instrument, all low-income or screened-in households with school-age children will be asked this item, regardless
of whether they receive free or reduced-price school meals. One potential concern with the item before the
test was that only a small number of households might affirm the item. Because the survey was administered
in September 2020 during the COVID-19 pandemic, when many school districts had implemented online
learning, and only households reporting receipt of free or reduced-price school lunch were asked the item, the
19
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
number of affirmative responses was likely suppressed. However, 163 households affirmed the item with the
test, and the total is likely to be higher when schools are open for in-person learning, the item is not restricted
to free or reduced-price school lunch recipients, and the item is administered in a full CPS-FSS sample.
The item will likely have sufficient affirmations in a regular implementation to be useable for research and
analysis.
Question included in September 2020 test instrument: During the past 30 days, did any children
in the household (between 5 and 18 years old) receive a free or reduced-price meal or snack at an after
school program?
Note: The item is only asked for households that have school-age children. For households with younger
children as well, the parenthetical statement is added to indicate that the question refers only to schoolage children.
Table 5 shows the prevalence of food insecurity by participation in Federal nutrition assistance programs.
While differences in screening between the test and standard instrument resulted in more households being
screened into these sections based on reported food insufficiency, this table is limited to respondents with
reported incomes within the range to make them eligible for the programs, and these households would all
automatically be screened in based on having a low income. Statistically significant differences exist between
the standard and test instruments in the prevalence of food insecurity for SNAP participants. The prevalence
of food insecurity among low-income households that received SNAP benefits in the previous 12 months
was 39.3 percent in the standard instrument, significantly lower than 46.7 percent in the test instrument.
In December 2020, 45.4 percent of low-income households that received SNAP benefits in the previous 12
months were food insecure (for the December 2020 estimates, see table 8 in Coleman-Jensen et al., 2021).
For both the test and standard instruments, a consistent pattern with previous years emerges, with higher
food insecurity rates among low-income SNAP participants than among low-income nonparticipants.
The results suggest that the modified nutrition assistance items functioned similarly to the standard instrument in the pattern of results between households that reported receiving assistance and income-eligible
respondents who did not report receiving benefits. No statistically significant differences existed in the
prevalence of food insecurity for those receiving free or reduced-price school lunches or Special Supplemental
Nutrition Program for Women, Infants, and Children (WIC) or income-eligible nonparticipants. A statistically significant difference was noted for the prevalence of food insecurity among households that received
SNAP benefits in the previous 12 months. However, this may be due to factors besides the change in location
of the SNAP questions in the survey and the different lead-in between the standard and test instruments. The
statistical difference may be related to differences in the sample since estimates from the test instrument show
greater food insecurity overall. Including the additional food insufficiency screening variable to screen households into the nutrition assistance section of the questionnaire and adding a lead-in to the SNAP variable
may have affected SNAP reporting, but any possible effect does not appear to be detrimental to the results.
Cell sizes are relatively small for these estimates, so sampling errors or misreporting may have an undue effect
on the estimates. The test instrument included 672 low-income households, and the standard instrument
included 706 households that reported receipt of SNAP benefits and had valid food security data. Previous
research showed that nutrition assistance participation is underreported by household survey respondents in
the CPS (see footnote 45 in Coleman-Jensen et al., 2021; Meyer & Goerge, 2011).
20
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
Table 5
Percentage of households by food security status and participation in selected Federal nutrition
assistance programs, by instrument version, September 2020
Standard instrument
Test instrument
Food insecure
Food
secure
All
With
low
food
security
With
very
low
food
security
Food insecure
Food
secure
Percent
All
With
low
food
security
With
very
low
food
security
Percent
Income less than 130 percent of poverty line:
Received SNAP benefits in previous 12 months
60.7
39.3
20.0
19.3
53.3
46.7*
26.4
20.3
Did not receive SNAP benefits in previous 12
months
78.3
21.7
12.4
9.3
75.1
24.9
14.5
10.4
Income less than 185 percent of poverty line; school-age children in household:
Received NSLP free or reduced-price school
lunch in previous 30 days
68.8
31.2
21.8
9.3
62.3
37.7
26.6
11.1
Did not receive NSLP free or reduced-price
school lunch in previous 30 days
71.5
28.5
17.0
11.5
71.0
29.0
19.1
10.0
Income less than 185 percent of poverty line; children under age 5 in household:
Received WIC in previous 30 days
70.4
29.6
14.5
15.1
62.4
37.6
30.8
6.9
Did not receive WIC in previous 30 days
76.0
24.0
16.6
7.4
70.0
30.0
18.6
11.4
SNAP = Supplemental Nutrition Assistance Program, formerly the Food Stamp Program. NSLP = National School Lunch Program.
WIC = Special Supplemental Nutrition Program for Women, Infants, and Children.
* Difference between the September standard instrument and that September test instrument is statistically significant with 90 percent confidence (t >1.645).
Source: USDA, Economic Research Service using data from U.S. Department of Commerce, Bureau of the Census, September 2020
Current Population Survey, Food Security Supplement.
Methods and Results: Community Nutrition Assistance
In USDA’s statistical supplement to the annual food security report, statistics on the use of community
nutrition assistance programs are presented by food security status (Coleman-Jensen et al., 2022b). The
combined use of community and Federal nutrition assistance is also reported, as is the use of food pantries
by household characteristics. Community nutrition assistance refers to free food available from food pantries,
food banks, churches, etc., as well as free meals received from community soup kitchens and other sources.
Anti-hunger advocates had previously commented to USDA, ERS staff that the statistics from the CPS-FSS
likely underestimate the use of community nutrition assistance (Weinfield et al., 2014), do not cover the full
range of how free food is made available, and are dated in the terminology used in the questions. As a result,
the authors revised the items based on recommended wording from Feeding America. Feeding America is
a national nonprofit organization that provides food to nearly every county in the United States through
a network of food banks and charitable programs. The organization has expertise in the charitable feeding
system and conducts research on food security and community nutrition assistance. Feeding America had
previously conducted cognitive testing of these survey items about free groceries and free meals in 2017, and
these questions were fielded in the Urban Institute’s Wellbeing and Basic Needs Survey (McDonald et al.,
2017). These items underwent cognitive testing with the Census Bureau and, based on the recommendations
of expert review and cognitive testing, these new items were included in the food security test instrument.
21
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
USDA, ERS has explicitly acknowledged in the annual statistical supplement to the household food security report that estimates for the use of emergency or soup kitchens understate actual use because sample
households are selected from an address-based list (Coleman-Jensen et al., 2022b). While the revisions to
the wording of the questions may improve estimates of reported use of community food providers, underreporting may remain because the sample does not include people who were homeless at the time of the survey.
Screening for the community nutrition assistance questions is the same as screening for the food security
questions, and the screening procedures did not change for the test beyond dropping the lead-in for question
HES9. Households are screened into this section based on reporting a low income or reporting some level
of food hardship (in response to questions HES9 and HESS1). While screening is important for reducing
respondent burden and reducing irritation or awkwardness of interviewers asking respondents questions that
do not apply to their household or appear inappropriate based on earlier responses, the limitation is that some
households using community nutrition assistance may not pass through these screeners, and their use of these
programs would not be reported.
One item in the standard instrument asked about emergency food from food pantries, and another asked
about meals from a soup kitchen. These items were revised significantly for the test instrument:
Standard instrument:
Food pantries: In the last 12 months, did you or other adults in your household ever get emergency food
from a church, a food pantry, or food bank?
Soup kitchens: In the last 12 months, did you or other adults in your household ever eat any meals at a
soup kitchen or shelter?
Test instrument:
Free groceries: In the last 12 months, did you or anyone in your household ever get free groceries from a
food pantry, food bank, church, or other place that helps with free food?
Free meals: In the last 12 months, have you or anyone in your household received a free meal from a
church, shelter, home-delivered meal service like Meals on Wheels, or other place that helps with free
meals?
In response to the questions about free groceries and free meals in the test supplement, the reported receipt
was higher than for the questions about food pantries and soup kitchens in the standard instrument. This is
consistent with the authors’ expectations for the test items. The language for the test items is more inclusive
in several ways, including asking if “anyone in your household” received assistance rather than just “adults,”
referring to “free groceries” rather than only “emergency food,” asking about “home-delivered meals,” and by
asking about any “other place that helps with free food/meals.”
Differences in statistical significance are not presented in table 6 because these items are not comparable
given the changes in wording. The new questions about the receipt of free groceries and free meals appear to
have functioned adequately. For both questions in the test panel, the receipt of free food was more common
for food-insecure households and most common for households with very low food security. This is to be
expected as the use of nutrition assistance, both Federal and community, is more common among food-insecure households. In addition, in the test panel, the reported receipt was more common among low-income
households. For households with incomes under 185 percent of the poverty level, 21.9 percent reported
receipt of free groceries, and 7.5 percent reported receipt of free meals. By contrast, in the test panel, among
households with incomes of 185 percent of the poverty level and over, 3 percent reported receipt of free
22
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
groceries, and 0.7 percent reported receipt of free meals (estimates not shown in table). These findings further
confirm that the relationship between the receipt of community nutrition assistance as obtained from the test
instrument and household income is as expected. Based on expert review, cognitive testing, and the results of
the test instrument, the revised questions on community nutrition assistance appear to be an improvement
and function well.
Table 6
Use of community nutrition assistance,1 by standard and test instrument, September 2020
Standard Instrument
Food pantries
Total
Thousands
All households2
129,564
Soup kitchens
Users
Total
Thousands
Percent
7,584
5.9
Thousands
Users
Thousands
129,574
597
Percent
0.5
Households by food security status:3
Food secure
117,108
3,618
3.1
117,121
224
0.2
Food insecure
12,454
3,966
31.8
12,437
374
3.0
Low food security
7,963
2,174
27.3
7,938
206
2.6
Very low food security
4,491
1,793
39.9
4,499
167
3.7
Test Instrument
Free groceries
Total
All households2
Free meals
Users
Total
Users
Thousands
Thousands
Percent
Thousands
Thousands
Percent
129,198
8,888
6.9
129,103
2,705
2.1
Households by food security status:3
Food secure
115,597
4,355
3.8
115,533
1,448
1.3
Food insecure
13,566
4,533
33.4
13,546
1,257
9.3
Low food security
8,530
2,598
30.5
8,529
579
6.8
Very low food security
5,036
1,934
38.4
5,018
678
13.5
1 Questions about the receipt of community nutrition assistance are not comparable between the standard and test instruments. Percentages between the two instruments are not tested for statistical significance because the wording of the items changed enough
that there is no expectation of comparability.
2 Totals for “All households” exclude households that did not answer the question about food pantries/free groceries or emergency
kitchens/free meals. For all households, these exclusions represent:
•
September 2020 standard instrument: 0.6 percent of all households had missing data on food pantries, and 0.6 percent of
all households had missing data on soup kitchens.
•
September 2020 test instrument: 0.8 percent of all households had missing data on free groceries, and 0.9 percent of all
households had missing data on free meals.
3 Totals also exclude households for which food security status is unknown because household respondents did not give a valid
response to any of the questions in the food security scale.
Source: USDA, Economic Research Service using data from U.S. Department of Commerce, Bureau of the Census, September 2020
Current Population Survey, Food Security Supplement.
23
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
Conclusions
Based on the analysis of the split-panel test in the September 2020 CPS-FSS, the authors conclude the test
instrument performed reasonably well. The observed differences in the test and standard instruments for food
spending and community nutrition assistance were as expected. The Rasch analysis confirms that the minor
changes to the food security section are unlikely to affect the measurement of food insecurity or the comparability of estimates from year to year. The differences in the prevalence of food insecurity between the test and
standard instruments may be related to sampling error, which was more of a concern during the early months
of the COVID-19 pandemic.
This analysis, along with the recommendations from the expert review and from cognitive testing, indicates
that updates to the CPS-FSS September 2020 test instrument should be implemented in the regular CPS-FSS
data collection moving forward. USDA, ERS sought and obtained approval from OMB to implement these
changes in the regular CPS-FSS data collection beginning in December 2022. These changes to the regular
CPS-FSS December instrument will be documented for researchers using the data for analyses and for stakeholders who read or cite the future annual reports in the series Household Food Security in the United States
and Statistical Supplement to Household Food Security in the United States.
24
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
References
Bickel, G., Nord, N., Price, C., Hamilton, W.T., & Cook, J.T. (2000). Guide to measuring household food security, (revised 2000). U.S. Department of Agriculture, Food and Nutrition Service.
Coleman-Jensen, A., (2015, October 5). Commemorating 20 years of U.S. food security measurement. Amber
Waves, U.S. Department of Agriculture, Economic Research Service.
Coleman-Jensen, A., Rabbitt, M.P., Gregory, C.A., & Singh, A. (2021). Household food security in the United
States in 2020 (Report No. ERR-298). U.S. Department of Agriculture, Economic Research Service.
Coleman-Jensen, A., Rabbitt, M.P., Gregory, C.A., & Singh, A. (2022), Household food security in the United
States in 2021 (Report No. ERR–309). U.S. Department of Agriculture, Economic Research Service.
Coleman-Jensen, A., Rabbitt, M.P., Gregory, C.A., & Singh, A. (2022b). Statistical supplement to household food Security in the United States in 2021 (Report No. AP–105). U.S. Department of Agriculture,
Economic Research Service.
Federal Interagency Forum on Child and Family Statistics. (2021). America’s children: Key national indicators
of well-being, 2021. U.S. Government Printing Office, Washington, DC.
Healthy People 2030. (n.d.). Healthy people 2030 leading health indicators (LHIs). U.S. Department of Health
and Human Services, Office of Disease Prevention and Health Promotion.
Kephart, K., Katz, J., Virgile, M., Terry, R., & Holzberg, J. (2021). Cognitive testing results for the current
population survey food security supplement. Research report series: Survey methodology. (Report No. 2021–
06). Center for Behavioral Science Methods, Research, and Methodology Directorate, U.S. Department
of Commerce, Bureau of the Census.
McDonald, E., Engelhard, E., & Hake, M. (2017). Recommendation of survey questions to measure receipt
of charitable food assistance on FoodAPS 2.0. Memo to Elizabeth Larimore, Economic Research Service,
USDA, September 29, from Feeding America.
Meyer, B.D., & Goerge, R. (2011). Errors in survey reporting and imputation and their effects on estimation of
food stamp participation. Working Paper, The Harris School, University of Chicago.
Nord, M. (2008). December 2007 microdata file: technical documentation (CPS Food Security Supplement).
U.S. Department of Agriculture, Economic Research Service.
Nord, M. (2012). Assessing potential technical enhancements to the U.S. household food security measures (Report
No. TB–1936). U.S. Department of Agriculture, Economic Research Service.
Nord, M., Andrews, M., & Carlson, S. (2006). Household food security in the United States, 2005 (Report No.
ERR-29). U.S. Department of Agriculture, Economic Research Service.
Nord, M., Andrews, M., & Carlson, S. (2007). Household food security in the United States, 2006 (Report No.
ERR-49). U.S. Department of Agriculture, Economic Research Service.
Nord, M., Andrews, M., & Carlson, S. (2008). Household food security in the United States, 2007 (Report No.
ERR-66). U.S. Department of Agriculture, Economic Research Service.
25
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
Rothbaum, J., & Bee, A. (2021, March 30). Coronavirus infects surveys, too: Survey nonresponse bias and the
coronavirus pandemic (Report No. WP2020-10). Social, Economic, and Housing Statistics Division
working papers, U.S. Department of Commerce, Bureau of the Census Bureau.
U.S. Department of Agriculture. (2020). Official USDA food plans: Cost of food at home at four levels, U.S.
average, September 2020. U.S. Department of Agriculture, Center for Nutrition Policy and Promotion.
U.S. Department of Agriculture. (2021). Thrifty Food Plan, 2021 (Report No. FNS-916). U.S. Department of
Agriculture, Food and Nutrition Service.
Weinfield, N.S., Mills, G., Borger, C., Gearing, M., Macaluso, T., Montaquila, J., & Zedlewski, S. (2014).
Hunger in America 2014, national report prepared for Feeding America by Westat and Urban Institute.
Zeballos, E., & Sinclair, W. (2023). USDA Economic Research Service’s food expenditure series data
product. U.S. Department of Agriculture, Economic Research Service.
26
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
Appendix A: Test Questionnaire–September 2020 Food
Security Supplement
Test Questionnaire to test recommended changes from Census Cognitive Testing
Draft test questionnaire January 3, 2020
NOTE: This test version starts with the 2019 CPS-FSS questionnaire. HRPOOR may need to be updated
before final production in September 2020. Changes from 2018 CPS-FSS to 2019 CPS-FSS questionnaire:
There are updates to the POORCK specification (HRPOOR variable). Three new household size and income
combinations will be screened into the module in 2019 with POOR now equal to 1 for these groups.
NOTE: Changes are highlighted in yellow below in POORCK specification and shown here:
NUMHOU=1 and FAMINC=8 set POOR=1
NUMHOU=4 and FAMINC=12 set POOR=1
NUMHOU=5 and FAMINC=13 set POOR=1
NOTE: Update year where appropriate to 2020. Update month where appropriate from December to September. Changes for test are highlighted in yellow.
SPECIFICATIONS
Set SUPTM SUPTM is a timer for the entire supplement.
------------------------------------------------------------NOTES:
and are State-specific names for the SNAP/Food Stamp
program.
------------------------------------------------------------POORCK
NUMHOU = 1 and FAMINC = 1, 2, 3, 4, 5, 6, 7, or 8 set POOR=1 else if
NUMHOU = 2 and FAMINC = 1, 2, 3, 4, 5, 6, 7, 8, or 9 set POOR=1 else if
NUMHOU = 3 and FAMINC = 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 set POOR=1 else if
NUMHOU = 4 and FAMINC = 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, or 12 set POOR=1 else if
NUMHOU = 5 and FAMINC = 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12 or 13 set POOR=1 else
if
NUMHOU = 6 and FAMINC = 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, or 13
set POOR=1
else if
NUMHOU = 7 and FAMINC = 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, or 14
set POOR=1
else if
NUMHOU = 8 and FAMINC = 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, or 14
set POOR=1
else if
NUMHOU= 9-16 and FAMINC = 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, or 15
set POOR=1
else if
Set POOR = 2.
-------------------------------------------------------------
27
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
[LEAD for one-person households] This month we are asking some questions about food
used in your household and the ways you are managing to meet your food needs.
[Lead for two or more person households] This month we are asking some questions about
food used in your household and the ways you are managing to meet your food needs. (The
best person to answer would be the adult most knowledgeable about food shopping and
meal preparation. Would that be you or someone else?)
[Change respondent if needed and available, otherwise continue with current respondent]
[If group quarters, go to S9]
I. FOOD EXPENDITURES
S1A
If more than one HHMEM=1 has an AGE equal to or greater than 10 fill with second option else fill with first option.
These first questions are about all the places at which you bought food LAST
WEEK. By LAST WEEK, I mean from Sunday through Saturday.
First, did (you/anyone in your household) shop for food at a supermarket, grocery
store, Walmart or Target LAST WEEK?
<1>
<2>
Yes
No
Blind or
SCOMPL
If onpath entry, D or R in S1A then set SCOMPL=1 otherwise set SCOMPL=0
S1B
If more than one HHMEM=1 has an AGE equal to or greater than 10 fill with second option else fill with first option.
Think about other places where people buy food, such as dollar stores, pharmacies,
club stores, farmers markets, or online. Did (you/anyone in your household) buy food
from any stores such as these LAST WEEK?
<1>
<2>
Yes
No
Blind or
S1C
If more than one HHMEM=1 has an AGE equal to or greater than 6 fill with second option
else fill with first option in first parenthetical.
If one or more HHMEM=1 with AGE between 6 and 18 then fill second parenthetical else
fill blank.
28
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
LAST WEEK, did (you/anyone in your household) buy food at a restaurant, fast food
place, cafeteria, deli, convenience store, or vending machine? (Include any children
who may have bought food at the school cafeteria).
<1>
<2>
Yes
No
Blind or
S1D
If more than one HHMEM=1 has an AGE equal to or greater than 10 fill with
second option else fill with first option.
Did (you/anyone in your household) buy food from any other kind of place LAST
WEEK?
<1>
<2>
Yes
No
Blind or
SCKA
If onpath entry of <2>, or in S1A, S1B, S1C and S1D then skip to S8
else go to SLEAD.
SLEAD
Now I'm going to ask you about the TOTAL amount you spent on food
LAST WEEK in all the places where you bought food. Then, since LAST
WEEK may have been unusual for you, I will ask about the amount you
USUALLY spend.
Proceed
SCKB
If onpath entry of <1> in S1A then ask S2 else skip to SCKC.
S2
If more than one HHMEM=1 has an AGE equal to or greater than 10 fill with second option else fill with first option.
If POOR=1 then fill second parenthetical else fill blank.
How much did (you/anyone in your household) spend in total at supermarkets,
grocery stores, Walmart or Target LAST WEEK (including any purchases made with
or food stamp benefits)?
ENTER <0> IF RESPONDENT CAN ONLY GIVE RANGE
$_ _ _.00
S2CK
If entry of <0> in S2 goto S2CK1A else store entry in S2O. If S2O is between $1.00 and
$450.00 goto S3A else if S2O is equal to D or R go to SCKC otherwise go to S2RC.
29
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
S2CK1A
**************DO NOT READ TO RESPONDENT *******************
Enter range reported by respondent
_ _ _.00
If entry is D or R, store in S2O
<1-999> GOTO S2CK1B
GOTO SCKC
S2CK1B
Enter range reported by respondent
_ _ _.00
If entry is D or R, store in S2O
<1-999> GOTO S2RG
GOTO SCKC
S2RG
Add the entries in S2CK1A and S2CK1B and divide by 2. Store the answer in S2O. If S2O
is between $1.00 and $450.00 go to S3A otherwise go to S2RC.
S2RC
*****************DO NOT ASK THE RESPONDENT*******************
Amount spent recorded as: (entry in S2O) Is this entry correct?
<1>
<2>
S2COR
YES (GO TO S3A)
NO (GO TO S2COR)
***************DO NOT ASK THE RESPONDENT****************
Incorrect entry was recorded as:
Correct entry is:
(entry in S2O)
$_ _ _.00 (store entry in S2O)
S3A
How much of the (fill with S2O) was for non-food items, such as pet food, paper
products, alcohol, detergents, or cleaning supplies?
Enter <1> for whole dollar amount (GOTO S3)
Enter <2> if respondent can only give range (GOTO S3CK2A)
$_ _ _.00
Blind or (GOTO SCKC)
30
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
S3
Enter whole dollar amount
$_ _ _.00
Store amount in S3O
<1-100> GOTO SCKC
<0, ge 101> GOTO S3RC
S3CK2A
******************DO NOT READ TO RESPONDENT*****************
Enter range reported by respondent
_ _ _.00
If entry is D or R, store in S3O
<1-999> GOTO S3CK2B
GOTO SCKC
S3CK2B
Enter range reported by respondent
_ _ _.00
If entry is D or R, store in S3O
<1-999> GOTO S3RG
GOTO SCKC
S3RG
Add the entries in S3CK2A and S3CK2B and divide by 2. Store the answer in S3O. Do not
allow entry in S3O to be greater than entry in S2O. If S3O is between $1.00 and $100.00
go to SCKC otherwise go to S3RC.
S3RC
*************DO NOT ASK THE RESPONDENT**********************
Amount spent recorded as: (entry in S3O)
Is this entry correct?
<1>
<2>
S3COR
YES (GO TO SCKC)
NO (GO TO S3COR)
***************DO NOT ASK THE RESPONDENT******************
Incorrect entry was recorded as:
Correct entry is:
(entry in S3O)
$_ _ _.00 (store entry in S3O)
Do not allow entry in S3O to be greater than entry in S2O.
31
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
SCKC
If onpath entry of <1> in S1B then ask S4 else skip to SCKD.
S4
If more than one HHMEM=1 has an AGE equal to or greater than 10 fill with second option else fill with first option.
If POOR=1 then fill second parenthetical with first option else fill with blank.
How much did (you/your household) spend at stores such as dollar stores, pharmacies, club stores, farmers markets, or online LAST WEEK (including any purchases
made with or food stamp benefits)?
Enter whole dollar amount
Enter <0> if respondent can only give range
$_ _ _.00
Blind or (GO TO SCKD)
S4CK
If entry of <0> in S4 go to S4CK1A else store entry in S4O. If S4O is between $1.00 and
$300.00 go to S5 else if S4O is D or R go to SCKD otherwise go to S4RC.
S4CK1A
**************DO NOT READ TO RESPONDENT*******************
Enter range reported by respondent
_ _ _.00
If entry is D or R, store in S4O
<1-999> GOTO S4CK1B
GOTO SCKD
S4CK1B
Enter range reported by respondent
_ _ _.00
If entry is D or R, store in S4O
<1-999> GOTO S4RG
GOTO SCKD
S4RG
Add the entries in S4CK1A and S4CK1B and divide by 2. Store the answer in S4O. If S4O
is between $1.00 and $300.00 go to S5A otherwise go to S4RC.
32
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
S4RC
*****************DO NOT READ TO RESPONDENT******************
Amount spent recorded as: (entry in S4O)
Is this entry correct?
<1> Yes (GO TO S5A)
<2> No (GO TO S4COR)
S4COR
***************DO NOT READ TO RESPONDENT********************
Incorrect entry was recorded as: (entry in S4O)
Correct entry is:
$_ _ _.00 (store entry in S4O)
S5A
How much of the $(fill with S4O) was for non-food items, such as pet food,
paper products, alcohol, detergents, or cleaning supplies?
Enter <1> for whole dollar amount (GOTO S5)
Enter <2> if respondent can only give range (GOTO S5CK1A)
$_ _ _.00
Blind or (GOTO SCKD)
S5
Enter whole dollar amount
$_ _ _.00
Store amount in S5O
<1-100> GOTO SCKD
GOTO S5RC
S5CK1A
******************DO NOT READ TO RESPONDENT*****************
Enter range reported by respondent
_ _ _.00
If entry is D or R, store in S5O
<1-999> GOTO S5CK1B
GOTO SCKD
33
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
S5CK1B
Enter range reported by respondent
_ _ _.00
If entry is D or R, store in S5O
<1-999> GOTO S5RG
GOTO SCKD
S5RG
Add the entries in S5CK1A and S5CK1B and divide by 2. Store the answer in S5O. Do not
allow entry in S5O to be greater than entry in S4O. If S5O is between $1.00 and 100.00 go
to SCKD else go to S5RC.
S5RC
***************DO NOT ASK THE RESPONDENT***************
Amount spent recorded as : (entry in S5O)
Is this entry correct?
<1>
<2>
S5COR
Yes (GO TO SCKD)
No (GO TO S5COR)
***************DO NOT ASK THE RESPONDENT*******************
Incorrect entry was recorded as:
Correct entry is:
(entry in S5O)
$_ _ _.00 (store entry in S5O)
Do not allow entry in S5O to be greater than entry in S4O.
SCKD
If entry of <1> in S1C then ask S6 else skip to SCKE
S6
If more than one HHMEM=1 has an AGE equal to or greater than 10 fill with second option else fill with first option.
How much did (you/your household) spend for food at restaurants, fast food places, cafeterias, delis, convenience stores, and vending machines LAST WEEK, not
including alcohol purchases (including any purchases made with (SNAP) or food
stamp benefits)?
Enter whole dollar amount
Enter <0> if respondent can only give range
$_ _ _.00
Blind or
If entry is 1-999, D, or R store in S6O
34
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
<0> GOTO S6CK1A
<1-200, D, R> GOTO SCKE
GOTO S6RC
S6CK1A
**************DO NOT READ TO RESPONDENT *******************
Enter range reported by respondent
_ _ _.00
If entry is D or R, store in S6O
<1-999> GOTO S6CK1B
GOTO SCKE
S6CK1B
Enter range reported by respondent
_ _ _.00
If entry is D or R, store in S6O
<1-999> GOTO S6RG
GOTO SCKE
S6RG
Add the entries in S6CK1A and S6CK1B and divide by 2. Store the answer in S6O. If S6O
is between $1.00 and $200.00 go to SCKE else go to S6RC.
S6RC
***************DO NOT ASK THE RESPONDENT*****************
Amount spent recorded as : (entry in S6O)
Is this entry correct?
<1>
<2>
S6COR
Yes (GO TO SCKE)
No (GO TO S6COR)
**************DO NOT ASK THE RESPONDENT********************
Incorrect entry was recorded as:
Correct entry is:
(entry in S6O)
$_ _ _.00 (store entry in S6O)
SCKE
If entry of <1> in S1D then ask S7 else skip to SCKF.
S7
If more than one HHMEM=1 has an AGE equal to or greater than 10 fill with second option else fill with first option.
35
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
How much did (you/your household) spend for food at any other kind of
place LAST WEEK?
Enter whole dollar amount
Enter <0> if respondent can only give range
$_ _ _.00
Blind or
If entry is 1-999, D, or R store in S7O
<0> GOTO S7CK1A
<1-150> GOTO SCKF
GOTO S7RC
GOTO SCKF
S7CK1A
**************DO NOT READ TO RESPONDENT *******************
Enter range reported by respondent
_ _ _.00
If entry is D or R, store in S7O
<1-999> GOTO S7CK1B
GOTO SCKF
S7CK1B
Enter range reported by respondent
_ _ _.00
If entry is D or R, store in S7O
<1-999> GOTO S7RG
GOTO SCKF
S7RG
Add the entries in S7CK1A and S7CK1B and divide by 2. Store the answer in S7O. If S7O
is between $1.00 and $150.00 go to SCKF otherwise go to S7RC.
S7RC
*************DO NOT ASK THE RESPONDENT**********************
Amount spent recorded as: (entry in S7O)
Is this entry correct?
<1>
<2>
YES (GO TO SCKF)
NO (GO TO S7COR)
36
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
S7COR
**************DO NOT ASK THE RESPONDENT*******************
Incorrect entry was recorded as:
Correct entry is:
(entry in S7O)
$_ _ _.00 (store entry in S7O)
SCKF
If any amounts 0 or over in S2O, S4O, S6O or S7O then add together and store in
SFDAMT. If any amounts 0 or over in S3O or S5O, then add these together and store in
SNFAMT. Subtract SNFAMT from SFDAMT and store the result in S8O.
S8A
If (entry of D or R in S2, S4, S6, and S7) or (S8O equals 0) or (onpath entry of <2>, ,
or in S1A, S1B, S1C and S1D) then fill first parenthetical with first option else fill with
second option.
If more than one HHMEM=1 has an AGE equal to or greater than 10 fill remaining parentheticals with second option else fill with first.
If POOR= 1 fill last parenthetical with first option else fill with blank.
(Let's see, it seems that (you/your household) did not buy any food LAST WEEK. /
Let's see, (you/your household) spent about (fill with S8O) on food LAST WEEK.)
Now think about how much (you/anyone in your household) USUALLY (spend/
spends). How much (do you/does your household) USUALLY spend on food at
all the different places we've been talking about IN A WEEK? (Please include any
purchases made with or food stamp benefits). Do not include nonfood items such as pet food, paper products, detergent or cleaning supplies.
Enter <1> for whole dollar amount
Enter <2> if respondent can only give range
Blind or
If entry is D or R store in S8OU
<1> GOTO S8
<2> GOTO S8CK1A
GOTO S9
S8
Enter whole dollar amount
$_ _ _.00
Store amount in S8OU
<0> GOTO S8ZA
<1-450> GOTO S8B
GOTO S8RC
GOTO S9
37
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
S8CK1A
**************DO NOT READ TO RESPONDENT *******************
Enter range reported by respondent
_ _ _.00
If entry is D or R, store in S8OU
<1-999> GOTO S8CK1B
GOTO S9
S8CK1B
Enter range reported by respondent
_ _ _.00
If entry is D or R, store in S8OU
<1-999> GOTO S8RG
GOTO S9
S8RG
S8ZA
Add the entries in S8CK1A and S8CK1B and divide by 2. Store the answer in S8OU. If
S8OU is between $1.00 and $450.00 go to S8B otherwise go to S8RC.
If that is because you shop for food only once in a while, how much would you say
you spend in total in a month?
NOTE: This question previously asked “If that is because you shop for food infrequently, how much would
the weekly average be over several weeks?” The previous version asked the respondent to determine the weekly
average. The new question asks the respondent how much they spend in a month. The survey software instrument will then calculate the weekly average for the respondent. Census will need to provide guidance on how
this change should be reflected in the interviewer instructions below. We still want to enter a weekly average in
the data even though the question now asks for a monthly average.
Enter <1> for whole dollar amount or if the amount really is zero
Enter <2> if respondent can only give range
Blind or
If entry is D or R store in S8OU
<1> GOTO S8Z
<2> GOTO S8ZCK1A
GOTO S9
S8Z
Enter whole dollar amount
$_ _ _.00
Store amount in S8OU NOTE: Divide by 4 before storing in S8OU
<0> GOTO S8RC
<1-450> GOTO S8B
GOTO S8RC
38
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
S8ZCK1A
**************DO NOT READ TO RESPONDENT *******************
Enter range reported by respondent
_ _ _.00
If entry is D or R, store in S8OU
<1-999> GOTO S8ZCK1B
GOTO S9
S8ZCK1B
Enter range reported by respondent
_ _ _.00
If entry is D or R, store in S8OU
<1-999> GOTO S8ZRG
GOTO S9
S8ZRG
Add the entries in S8ZCK1A and S8ZCK1B and divide by 2. Store the answer in S8OU.
NOTE: Divide by 4 before storing in S8OU If S8OU is between $1.00 and $450.00 go to S8B
otherwise go to S8RC.
S8RC
*************DO NOT ASK THE RESPONDENT********************
Amount spent recorded as: (entry in S8OU)
Is this entry correct?
<1>
<2>
S8COR
Yes (GO TO S8B)
No (GO TO S8COR)
*************DO NOT ASK THE RESPONDENT********************
Incorrect entry was recorded as: (entry in S8OU)
Correct entry is:
$ _ _ _ _.00
II. MINIMUM SPENDING NEED TO HAVE ENOUGH FOOD
S8B
If NUMHOU = 1 then fill parenthetical with first option else fill with second
option.
In order to buy just enough food to meet (your needs/the needs of your
household), would you need to spend more than you do now, or could you
spend less?
39
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
************************DO NOT READ************************
<1>
More (GO TO S8C)
<2>
Less (GO TO S8D)
<3>
Same (GO TO S9)
Blind or (GO TO S9)
S8C
About how much MORE would you need to spend each week to buy just
enough food to meet the needs of your household?
Enter whole dollar amount
Enter <0> if respondent can only give range
$_ _ _.00
Blind or
If entry is 1-999, D, or R store in S8CO
<0> GOTO S8CCKA
<1-999, D, R> GOTO S9
S8CCKA
**************DO NOT READ TO RESPONDENT *******************
Enter range reported by respondent
_ _ _.00
If entry is D or R, store in S8CO
<1-999> GOTO S8CCKB
GOTO S9
S8CCKB
Enter range reported by respondent
_ _ _.00
If entry is D or R, store in S8CO
<1-999> GOTO S8CRG
GOTO S9
S8CRG
Add the entries in S8CCKA and S8CCKB and divide by 2. Store the answer in S8CO.
GOTO S9.
40
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
S8D
About how much LESS could you spend each week and still buy enough food
to meet the needs of your household?
Enter whole dollar amount
Enter <0> if respondent can only give range
$_ _ _.00
Blind or
If entry is 1-999, D, or R store in S8DO
<0> GOTO S8DCKA
<1-999, D, R> GOTO S9
S8DCKA
**************DO NOT READ TO RESPONDENT *******************
Enter range reported by respondent
_ _ _.00
If entry is D or R, store in S8DO
<1-999> GOTO S8DCKB
GOTO S9
S8DCKB
Enter range reported by respondent
_ _ _.00
If entry is D or R, store in S8DO
<1-999> GOTO S8DRG
GOTO S9
S8DRG
Add the entries in S8DCKA and S8DCKB and divide by 2. Store the calculated value in
S8DO. GOTO S9.
III. FOOD SUFFICIENCY AND FOOD SECURITY
S9_Lead
The next questions are about the food eaten in your household in the last 12
months, since September of last year, and whether you were able to afford the food
you need.
41
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
S9
In the last 12 months, since September of last year, did you ever run short of money
and try to make your food or your food money go further?
<1> Yes (GO TO SS1)
<2 > No (GO TO SS1)
Blind or (GO TO SS1)
If NUMHOU = 1 then fill parenthetical with first option else fill with second
option.
SS1
Which of these statements best describes the food eaten in your household—enough
of the kinds of food (I/we) want to eat, enough but not always the kinds of food
(I/we) want to eat, sometimes not enough to eat, or often not enough to eat?
<1>
<2>
<3>
<4>
Enough of the kinds of food we want to eat
Enough but not always the kinds of food we want to eat
Sometimes not enough to eat
Often not enough to eat
Blind or
----------------------------------------------------SX1CK
If POOR=2 and (SS1= <1> or ) and (S9= <2> or ), then go to END OF SUPPLEMENT else ask SS2
-----------------------------------------------------NOTE:
ADULT_COUNT=number of household members with AGE>=18 or PURRP=1, 2, 3, 13,
or 14
CHILD_COUNT=number of household members with AGE<=17 and PURRP>=4 and
PURRP not 13 or 14
-----------------------------------------------------SS2
If ADULT_COUNT=1 for the household then fill first option in parenthetical else fill second option.
Now I'm going to read you several statements that people have made about their
food situation. For these statements, please tell me whether the statement was OFTEN true, SOMETIMES true, or NEVER true for (you/your household) in the last
12 months.
The first statement is "(I/We) worried whether (my/our) food would run out before (I/we) got money to buy more." Was that OFTEN true, SOMETIMES true, or
NEVER true for (you/your household) in the last 12 months?
<1> Often true
<2> Sometimes true
<3> Never true (GO TO SS3)
Blind or (GO TO SS3)
42
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
SSM2
Did this ever happen in the last 30 days?
<1> Yes
<2> No
Blind or
SS3
"The food that (I/we) bought just didn't last, and (I/we) didn't have money to get
more." Was that OFTEN, SOMETIMES or NEVER true for (you/ your household)
in the last 12 months?
<1> Often true
<2> Sometimes true
<3> Never true (GO TO SS4)
Blind or (GO TO SS4)
SSM3
Did this ever happen in the last 30 days?
<1> Yes
<2> No
Blind or
SS4
"(I/we) couldn't afford to eat balanced meals." Was that OFTEN, SOMETIMES or
NEVER true for (you/ your household) in the last 12 months?
<1> Often true (GO TO SSM4)
<2> Sometimes true (GO TO SSM4)
<3> Never true (GO TO SX2CK)
Blind or (GO TO SX2CK)
SSM4
Did this ever happen in the last 30 days?
<1> Yes
<2> No
Blind
-----------------------------------------------------------SX2CK If SS1 = <3> or <4> OR SS2 = <1> or <2> OR SS3 = <1> or <2> OR SS4 =
<1> or <2> then go to SH2
else go to SX4CK.
------------------------------------------------------------
43
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
SH2
If ADULT_COUNT=1 for the household then fill first option in parenthetical else fill second option.
In the last 12 months, did (you/you or other adults in your household) ever cut the
size of your meals or skip meals because there wasn't enough money for food?
<1> Yes
<2> No (GO TO SH3)
Blind or (GO TO SH3)
SHF2
How often did this happen—almost every month, some months but not every month,
or in only 1 or 2 months?
<1> Almost every month
<2> Some months but not every month
<3> Only 1 or 2 months
Blind or
SHM2
If ADULT_COUNT=1 for the household then fill first option in parenthetical else fill second option.
Now think about the last 30 days. During that time did (you/you or other adults in
your household) ever cut the size of your meals or skip meals because there wasn't
enough money for food?
<1> Yes
<2> No (GO TO SH3)
Blind or (GO TO SH3)
SHMF2
How many days did this happen in the last 30 days?
______number of days (GO TO SH3)
<1-30>
Blind or (GO TO SH3)
SH3
In the last 12 months, did you ever eat less than you felt you should because there
wasn't enough money for food?
<1> Yes
<2> No (GO TO SH4)
Blind or (GO TO SH4)
44
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
SHF3
How often did this happen—almost every month, some months but not every month,
or in only 1 or 2 months?
<1> Almost every month
<2> Some months but not every month
<3> Only 1 or 2 months
Blind or
SHM3
Did this happen in the last 30 days?
<1> Yes
<2> No (GO TO SH4)
Blind or (GO TO SH4)
SHMF3
In the last 30 days, how many days did you eat less than you felt you should
because there wasn't enough money for food?
______ number of days
<1-30>
Blind or
SH4
In the last 12 months, were you ever hungry but didn't eat because there wasn't
enough money for food?
<1> Yes
<2> No (GO TO SH5)
Blind or (GO TO SH5)
SHF4
How often did this happen—almost every month, some months but not every month,
or in only 1 or 2 months?
<1> Almost every month
<2> Some months but not every month
<3> Only 1 or 2 months
Blind or
SHM4
Did this happen in the last 30 days?
<1>Yes
<2> No (GO TO SH5)
Blind or (GO TO SH5)
45
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
SHMF4
In the last 30 days, how many days were you hungry but didn't eat because there
wasn't enough money for food?
_____ number of days
<1-30>
Blind or
SH5
In the last 12 months, did you lose weight because there wasn't enough money for
food?
<1> Yes
<2> No (GO TO SX3CK)
Blind or (GO TO SX3CK)
SHM5
Did this happen in the last 30 days?
<1> Yes
<2> No
Blind or
----------------------------------------------------------------SX3CK
If SH2=<1> OR SH3=<1> OR SH4=<1> OR SH5=<1> then continue to SSH1 else skip to
SX4CK
----------------------------------------------------------------SSH1
If ADULT_COUNT=1 for the household then fill first option in parenthetical else fill second option.
In the last 12 months, did (you/you or other adults in your household) ever not eat
for a whole day because there wasn't enough money for food?
<1> Yes
<2> No (GO TO SX4CK)
Blind or (GO TO SX4CK)
SSHF1
How often did this happen—almost every month, some months but not every month,
or in only 1 or 2 months?
<1> Almost every month
<2> Some months but not every month
<3> Only 1 or 2 months
Blind or
46
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
SSHM1
If ADULT_COUNT=1 for the household then fill first option in parenthetical else fill second option.
Now think about the last 30 days. During that time did (you/you or other adults in
your household) ever not eat for a whole day because there wasn't enough money for
food?
<1> Yes
<2> No (GO TO SX4CK)
Blind or (GO TO SX4CK)
SSHMF1
How many times did this happen in the last 30 days?
______ times
<1-30>
Blind or
-------------------------------------------------------------------SX4CK
If CHILD_COUNT is greater than or equal to 1 in household go to SS5 else skip to SP1.
------------------------------------------------------------SS5_LEAD
Now I'm going to read you several statements that people have made about the food
situation of their children. For these statements, please tell me whether the statement was OFTEN true, SOMETIMES true, or NEVER true in the last 12 months
for any child under 18 years old living in the household.
SS5
If ADULT_COUNT=1 for the household fill first, third, fourth and fifth parenthetical with
first option else fill with second option.
If CHILD_COUNT=1 then fill second parenthetical with first option else fill with second
option.
"(I/we) relied on only a few kinds of low-cost food to feed (the child in (my/our)
household/the children) because (I was/we were) running out of money to buy
food. Was that OFTEN, SOMETIMES or NEVER true for (you/ your household) in
the last 12 months?
SSM5
<1> Often true
<2> Sometimes true
<3> Never true (GO TO SS6)
Blind or (GO TO SS6)
Did this ever happen in the last 30 days?
<1> Yes
<2> No
Blind or
47
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
SS6
If ADULT_COUNT=1 fill first, third, fourth and fifth parenthetical with first option else fill
with second option.
If CHILD_COUNT=1 then fill second parenthetical with first option else fill with second
option.
"(I/we) couldn't feed (the child in (my/our) household/the children) a balanced
meal, because (I/we) couldn't afford that." Was that OFTEN, SOMETIMES or
NEVER true for (you/your household) in the last 12 months?
<1> Often true (GO TO SSM6)
<2> Sometimes true (GO TO SSM6)
<3> Never true (GO TO SH1)
Blind or (GO TO SH1)
SSM6
Did this ever happen in the last 30 days?
<1> Yes
<2> No
Blind or
SH1
If ADULT_COUNT=1 then fill second, third, and fourth parenthetical with first option else
fill with second option.
If CHILD_COUNT=1 then fill first parenthetical with first option else fill with second option.
"(The child in (my/our) household was/The children were) not eating enough
because there wasn’t enough money for food." Was that OFTEN, SOMETIMES or
NEVER true for (you/ your household) in the last 12 months?
<1> Often true
<2> Sometimes true
<3> Never true (GO TO SX5CK)
Blind or (GO TO SX5CK)
SHM1
Did this ever happen in the last 30 days?
<1> Yes
<2> No
Blind or
48
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
-----------------------------------------------------------SX5CK
If SS5 = <1> or <2> OR SS6 = <1> or <2> OR SH1 = <1> or <2> go to SSH2 else skip to
SP1.
-----------------------------------------------------------SSH2
If CHILD_COUNT=1 then fill with first option else fill with second option.
In the last 12 months, did you ever cut the size of (the child's/any of the children's)
meals because there wasn't enough money for food?
<1> Yes
<2> No (GO TO SSH3)
Blind or (GO TO SSH3)
SSHF2
How often did this happen—almost every month, some months but not every month,
or in only 1 or 2 months?
<1>
<2>
<3>
Almost every month
Some months but not every month
Only 1 or 2 months
Blind or
SSHM2
Did this happen in the last 30 days?
<1> Yes
<2> No (GO TO SSH3)
Blind or (GO TO SSH3)
SSHMF2
If CHILD_COUNT=1 then fill with first option else fill with second option.
In the last 30 days, how many days did you cut the size of (the child's/any of the
children's) meals because there wasn't enough money for food?
______days
<1-30>
Blind or
49
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
SSH3
If CHILD_COUNT=1 then fill with first option else fill with second option.
In the last 12 months, (was the child/were the children) ever hungry because there
wasn’t enough money for food?
<1> Yes
<2> No (GO TO SSH4)
Blind or (GO TO SSH4)
SSHF3
How often did this happen—almost every month, some months but not every month,
or in only 1 or 2 months?
<1> Almost every month
<2> Some months but not every month
<3> Only 1 or 2 months
Blind or
SSHM3
Did this happen in the last 30 days?
<1> Yes
<2> No (GO TO SSH4)
Blind or (GO TO SSH4)
SSHMF3
If CHILD_COUNT=1 then fill with first option else fill with second option.
In the last 30 days, how many days (was the child/were the children) hungry because
there wasn’t enough money for food?
______ number of days
<1-30>
Blind or
SSH4
If CHILD_COUNT=1 then fill with first option else fill with second option.
In the last 12 months, did (the child/any of the children) ever skip a meal because
there wasn't enough money for food?
<1> Yes
<2> No (GO TO SSH5)
Blind or (GO TO SSH5)
50
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
SSHF4
How often did this happen—almost every month, some months but not every month,
or in only 1 or 2 months?
<1> Almost every month
<2> Some months but not every month
<3> Only 1 or 2 months
Blind or
SSHM4
If CHILD_COUNT=1 then fill with first option else fill with second option.
Now think about the last 30 days. Did (the child/any of the children) ever skip a
meal during that time because there wasn't enough money for food?
<1> Yes
<2> No (GO TO SSH5)
Blind or (GO TO SSH5)
SSHMF4
How many days did this happen in the last 30 days?
______days
<1-30>
Blind or
SSH5
If CHILD_COUNT=1 then fill with first option else fill with second option.
In the last 12 months, did (the child/any of the children) ever not eat for a whole day
because there wasn't enough money for food?
<1> Yes
<2> No (GO TO SP1)
Blind or (GO TO SP1)
SSHM5
Did this happen in the last 30 days?
<1> Yes
<2> No
Blind or
All responses go to SP1
51
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
IV. FOOD PROGRAM PARTICIPATION
SP1_LEAD
Sometimes people need help getting food for their household. There are many programs that can help.
SP1
If hhnum=1 fill with first option else fill with second.
In the past 12 months, since September of last year, did (you/anyone in this household) get or food stamp benefits?
<1> Yes (GO TO SP2)
<2> No (GO TO SP6CK)
Blind or (GO TO SP6CK)
NOTE: Since the test interview will be in September, it will change how months of SNAP receipt and
receipt within 30 days of interview are asked about. SP2 and SP2DCK and SP2D below may need more
revision.
SP2
In which months since September of last year were or food stamp
benefits received?
DO NOT READ LIST. MARK ALL THAT APPLY
<1> January 2020
<2> February 2020
<3> March 2020
<4> April 2020
<5> May 2020
<6> June 2020
<7> July 2020
<8> August 2020
<9> September 2020
<10> October 2019
<11> November 2019
<12> December 2019
<13> All
---------------------------------------------------------------------------------SP2DCK
If SP2 = 8 AND ≠ 9 AND ≠ 13 go to SP2D else go to SP3
----------------------------------------------------------------------------------
52
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
SP2D
If hhnum = 1 fill with first option else fill with second.
On what date in August did (you/your household) receive or food
stamp benefits?
SP2D
SP3
Day______
<1-31>
Blind or
If hhnum=1 fill with first option else fill with second.
How much did (you/your household) receive the last time you got
or food stamp benefits?
$ _ _ _ .00
Blind or
------------------------------------------------------------SP3CK
Store entry in SP3O. If SP3O is between $1.00 and $700.00 go to SP6CK otherwise go to
SP3RC.
-------------------------------------------------------------
SP3RC
*************DO NOT ASK THE RESPONDENT*************
AMOUNT RECEIVED RECORDED AS: (entry in SP3O)
IS THIS ENTRY CORRECT?
<1> YES (GO TO SP6CK)
<2> NO (GO TO SP3COR)
SP3COR
**********DO NOT ASK THE RESPONDENT************
INCORRECT ENTRY WAS RECORDED AS: (entry in SP3O) CORRECT ENTRY IS:
$_ _ _.00 (store entry in SP3O)
Items SP3 through SP3COR go into making the out variable SP3O. This is the amount
received in food stamp benefits.
------------------------------------------------------------SP6CK
If HHMEM=1 and AGE is 5 THROUGH 18 for anyone in the household ask SP6 else skip
to SP7ACK.
-------------------------------------------------------------
53
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
SP6
If 1 or more children under 5 years old, fill parenthetical, else fill blank.
During the past 30 days, did any children in the household
(between 5 and 18 years old) receive free or reduced-price lunches at school?
<1> Yes
<2> No (GO TO SP7ACK)
Blind or (GO TO SP7ACK)
SP7
If 1 or more children under 5 years old, fill parenthetical, else fill blank.
During the past 30 days, did any children in the household (between 5 and 18 years
old) receive free or reduced-price breakfasts at school?
<1> Yes
<2> No
Blind or
NEW 1
If 1 or more children under 5 years old, fill parenthetical, else fill blank.
During the past 30 days, did any children in the household (between 5 and 18 years
old) receive a free or reduced-price meal or snack at an after school program?
<1> Yes
<2> No
Blind or
Note: Question “NEW 1” is a newly added question. It does not replace any previous item.
----------------------------------------------------------SP7ACK
If HHMEM=1 and AGE is 0-6 for anyone in the household ask SP7A
else skip to SP8CK.
----------------------------------------------------------SP7A
If only 1 child under age 6 fill with first option else fill with second option
During the past 30 days, did any children in the household (under 6 years old) receive
free or reduced-price food at a day-care or Head Start program?
<1> Yes
<2> No
Blind or
54
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
------------------------------------------------------------SP8CK
If [(SEX=2 and AGE = 15-45) OR (AGE 0-4)] and HHMEM=1 for anyone in the household then ask SP8 else skip to “NEW 2”
------------------------------------------------------------SP8
If [(SEX=2 and AGE=15-45) for any hh member and (AGE=0-4) for any hh member] then
fill second option else if
(SEX=2 and AGE=15-45) for any hh member and (no AGE=0-4) for any hh member then
fill first option else fill third option.
During the past 30 days, did any (women/women or children/children) in this
household get food through the WIC program?
<1> Yes
<2> No (GO TO “NEW 2”)
Blind or (GO TO “NEW 2”)
SP9
If [(SEX=2 and AGE=15-45) for any hh member and (AGE=0-4) for any hh member] then
fill second option else if
(SEX=2 and AGE=15-45) for any hh member and (no AGE=0-4) for any hh member then
fill first option else fill third option.
How many (women/women or children/children) in the household got WIC foods?
Number ______
Blind or Yes
<2> No
Blind or
55
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
SC2
If more than one person in household fill with second option, else fill with first
option.
During the past 30 days, did (you/anyone in this household) go to a
community program or senior center to eat prepared meals?
<1> Yes
<2> No
Blind or
NOTE: Items SC1 and SC2 will be deleted. These questions are incorporated into “NEW 3” below.
For items “NEW 2” and “NEW 3,” if only 1 HHMEM = 1 and (AGE>=18 or PURRP <=3)
in household then fill first parenthetical with first option else fill with second option.
NEW 2
(Replaces SC3) In the last 12 months, did (you/you or anyone in your household) ever
get free groceries from a food pantry, food bank, church, or other place that helps
with free food?
<1> Yes
<2> No (GO TO SC3A)
SCF3
Blind or (GO TO SC4)
How often did this happen—almost every month, some months but not every month,
or in only 1 or 2 months?
<1> Almost every month
<2> Some months but not every month
<3> Only 1 or 2 months
Blind or
SCM3
Did this happen in the last 30 days?
<1> Yes (GO TO SC4)
<2> No (GO TO SC4)
SC3A
Is there a food pantry, food bank, church, or other place in your community where
you could get free groceries if you needed it?
<1> Yes
<2> No
Blind or
56
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
NEW 3
(Replaces SC4) In the last 12 months, have (you/you or anyone in your household)
received a free meal from a church, shelter, home-delivered meal service like Meals
on Wheels, or other place that helps with free meals?
<1>Yes
<2> No (GO TO NEW 4)
Blind or (GO TO NEW 4)
SCF4
How often did this happen—almost every month, some months but not every month,
or in only 1 or 2 months?
<1> Almost every month
<2> Some months but not every month
<3> Only 1 or 2 months
Blind or
SCM4
Did this happen in the last 30 days?
<1> Yes
<2> No
Blind or
NEW 4
Is there a church, shelter, home-delivery meal service like Meals on Wheels, or other
place in your community where you could get free meals if you needed it?
<1> Yes
<2> No
Blind or
NOTE: “NEW 4” is a newly added item. It does not replace any previous items.
SSPCK1
(Interviewer query, do not read)
Did you use the Spanish language version for none, some, or all of the questions
about having adequate food?
<1> None
<2> Some
<3> All
*****END OF SUPPLEMENT*****
57
Analysis of the Current Population Survey Food Security Supplement Split-Panel Test, TB-1963
USDA, Economic Research Service
| File Type | application/pdf |
| Author | Coleman-Jensen, Alisha - REE-ERS, Washington, DC |
| File Modified | 2024-11-06 |
| File Created | 2024-11-06 |