This page contains documentation for USDA's National Household Food Acquisition and Purchase Survey (FoodAPS). The first phase of this data outreach provides the following information:
FoodAPS collected the data from a nationally representative, stratified sample of 4,826 households between April 2012 and January 2013. Stratification was based on participation in the Supplemental Nutrition Assistance Program (SNAP) and total household income; the four strata were:
- Households receiving SNAP benefits, with a target sample size of 1,500,
- Non-SNAP households with income at or less than the poverty level, with a target sample size of 800,
- Non-SNAP households with income at or above the poverty level but no more than 185 percent of that level, with a target sample size of 1,200, and
- Non-SNAP households with higher income, with a target sample size of 1,500.
The FoodAPS sample of households was selected using two sample frames and a multi-stage sample design covering the contiguous United States. The first stage of the sampling process selected a stratified sample of 50 Primary Sampling Units (PSUs, defined as counties or groups of contiguous counties). The selection was proportional to size (PPS), with the measure of size (MOS) for each PSU being a composite that reflected the overall sample targets and the estimated population in each PSU for each of the four strata.
Data collection procedures
FoodAPS collected data on all food acquisitions and purchases by all members of sampled households over a 7-day period. This section provides information on the following aspects of data collection:
The primary food shopper or meal planner in the household was identified as the Primary Respondent (PR), and was asked to complete 2 in-person interviews (see Initial interview 16x16 - PDF and Final interview 16x16 - PDF ) and 3 brief telephone interviews (see Telephone interview 16x16 - PDF ) over the course of 9 days. All household members were also asked to track and report all food acquisitions during a 1-week period; scan barcodes on food products; save their store and restaurant receipts; and write information in a food book.
After a household was recruited into the study, the field interviewer trained the primary respondent in the use of the food reporting tools (see Initial visit script 16x16 - PDF ). A video program was also used to train the respondents (contact John A. Kirlin for a copy of the training video).
A typical data collection week for the household included the following:
- Initial household visit (Day 0):
- PR provided consent and completed the Initial household interview.
- PR received training on the use of food books and scanner; the PR was responsible for training other household members, as needed.
- PR received an income worksheet to complete during the study week in preparation for the Final interview (see Final household visit).
- Information was collected on food acquisitions beginning on the day after the initial household visit (Day 1).
- For each day from Day 1 to Day 7, household members or the PR:
- Recorded all food acquisitions in their food books.
- Indicated all meals and snacks eaten on the Meals and snacks form (see Meals and snacks form).
- For Days 2, 5, and 7, the PR called Mathematica’s Survey Operations Center to report his or her household's food acquisitions. When no food acquisitions for a given day were reported, these instances were reviewed to verify that no food was actually acquired that day and to distinguish these reports from non-participation in the study.
- Final household visit (Day 8):
- PR completed the Final household interview.
- Interviewer reviewed the completeness of the food books, issued the base incentive check and gift cards, and collected all food books and scanner.
- Interviewer asked PR to fill out the Respondent feedback form.
Meals and snacks form
Each household used the Meals and snacks form 16x16 - PDF to report the meals and snacks consumed by each household member on each day of the study week.
Data collection instruments
A variety of data collection instruments were used in FoodAPS, and these instruments are available in pdf format. This section provides background information about some of the key survey instruments:
Household screening tools
To determine a household's eligibility to participate in FoodAPS, screening interviews were conducted to collect information about household size, income, and SNAP participation. A series of questions was used to determine household eligibility within the income subgroups. Household size was defined as the number of people "who live together and share food." Household income was identified in two ways: (1) respondents identified all types of income received by the household by referring to a list of sources presented by the interviewer on a hand-held card; and (2) respondents reported the range of their total household income (the sum from all sources) by identifying income ranges on another hand-held card.
The screening interviews also included questions to identify the primary meal planner and food shopper in the household. As eligible households were identified, the field interviewer asked to speak with the primary food shopper; explained the study and what would be expected of participants; explained that participation would be voluntary; and asked the food shopper to participate as the primary respondent for his or her household.
Two versions of the Screening Interview were used during field operations. Version 1 was used until mid-August 2012, at which time several changes were made to the instrument and screening protocol. The changes that were incorporated into Version 2 include:
- A shortened introduction that revised the text and eliminated a question about receipt of the study’s advance postcard;
- The timing of the offer of a $5 unconditional incentive was changed from being offered immediately to the respondent to being offered after an initial refusal (the $5 incentive was provided at the end of a Version 2 screener even if the household did not answer any screening questions);
- Four questions were added for households that refused to complete the full screener;
- The questions were changed for households that were eligible for the study but refused to participate; and
- Space was added to the form for interviewer observations about the gender, age, race/ethnicity, and language of the person contacted.
The new questions and observations were added to provide more information for the analysis of patterns of non-response to the study.
See the actual interview tools on the Overview page.
Household food books
When filling out their food books, participants were asked to distinguish between "food and drinks brought into the home" and "meals, snacks, and drinks you got outside the home." ERS refers to food and drinks brought into the home as food at home (FAH) and meals, snacks, and drinks obtained outside the home as food away from home (FAFH). Although this documentation uses FAH and FAFH, these terms were not used on the data collection instruments or during interactions with respondents.
Based on information provided to participants, the following descriptions for FAH and FAFH apply:
Food at home (FAH) = foods and drinks that are brought home and used to prepare meals for consumption at home or elsewhere (for example, food used to make a sandwich that you bring to work). Places to get FAH include grocery stores, food pantries, and gardens.
Food away from home (FAFH) = foods and drinks that are obtained and consumed away from home, and prepared foods that are brought home or delivered. Places to get FAFH include cafeterias, restaurants, takeout places, parties, schools, church dinners, and senior centers.
The survey used four complementary tools to track food acquisitions during the survey week:
- A hand-held scanner for scanning barcodes on FAH items;
Primary respondent book 16x16 - PDF (for the primary respondent), with written instructions for scanning foods brought home; pictures and barcodes to scan variable weight items that might not have attached barcodes; data collection forms for reporting all places where foods were acquired each day (daily lists); data collection forms for reporting FAH acquisitions (blue pages); and data collection forms for reporting FAFH acquisitions (red pages);
Adult food book 16x16 - PDF for adults age 19 and older (other than the primary respondent), with data collection forms for reporting all places where foods were acquired each day (daily lists) and FAFH acquisitions (red pages); and
Youth food book 16x16 - PDF for respondents age 11-18, with data collection forms for reporting FAFH acquisitions (red pages).
Adults were asked to use their own food books—either the Primary respondent book or Adult food book—to record foods acquired by children under age 11.
Final interview and income worksheet
During the Final interview 16x16 - PDF , the field interviewer collected detailed information about sources and amounts of income for all household members. To prepare for the income questions, the primary respondent received an income worksheet during the initial interview. The income worksheet was designed to improve data quality and reduce respondent burden (see Income worksheet 16x16 - PDF ). The worksheet assisted the primary respondent in collecting information from other household members, referencing documents, as needed, and recording this information at his or her convenience before the final interview. Nearly 57 percent of households completed all or portions of the income worksheet prior to the final interview.
Respondent feedback form
After completing the final household interview and distributing incentives, interviewers asked respondents to complete a self-administered Respondent feedback form 16x16 - PDF . This paper questionnaire contained four questions:
- How often did respondents complete the Meals and Snacks Form?
- Was it easy to get other household members to participate in the data collection efforts?
- Was it easy to track foods? and
- Did household members change food acquisition behavior during the week because of the survey?
97.6 percent of those who completed the final interview submitted a feedback form.
Use of incentives
FoodAPS offered households monetary incentives to complete the various components of the data collection activities. As seen in prior research and the field test for this survey, the use of incentives can boost response rates, especially among hard-to-reach populations. The incentives were designed both to maximize responses to the household screening interview and to encourage participation throughout the week-long survey.
FoodAPS offered a $5 token of appreciation to all households that were contacted for screening. Initially, this token of appreciation was provided unconditionally, aiming to prevent refusals at the first point of contact instead of attempting to convert refusals afterwards. Although the timing of the incentive offer changed midway through field operations, all contacted households continued to receive the $5 even if they did not complete the screening interview.
Households that were eligible to participate in the study were offered a multi-part incentive designed to encourage initial agreement to participate in the week-long survey and to motivate households to stay engaged throughout the data collection week. This multi-part incentive included a base incentive (a $100 check) for primary respondents; up to three $10 gift cards to encourage PRs to initiate the telephone call-ins for food reporting on Days 2, 5, and 7; one $10 gift card for each additional household member age 11-14 who tracked their food acquisitions; and one $20 gift card for each additional household member age 15 and older who tracked their food acquisitions.
Households received all incentives at the end of the data collection week during the final visit from the field interviewer.
In response to intense interest in the accessibility of affordable healthy food and the potential influence of food deserts on food acquisition, FoodAPS added a geographic component to investigate how the local food environment affects food spending patterns in the United States. The FoodAPS Geography component (FoodAPS-GC) builds upon the FoodAPS survey data with the collection of data on the local food environment in the 50 primary sampling units (PSUs) in the survey. Information collected through the FoodAPS-GC includes the location of different types of retailers, measures of access to these retailers, measures of food prices and prices of food categories by retailers, as well as information about socio-demographic and food-related policies.
Data for all 50 FoodAPS PSUs are provided at different geographic levels—census block groups and census tracts, school districts, and counties—depending upon the availability of data. The FoodAPS-GC data come from a number of sources; details about the data and data sources can be found in the FoodAPS-GC codebook. The geographic data will be linked to the survey respondents and made available on a restricted basis to approved data users through the National Opinion Research Center (NORC). FoodAPS-GC data will be available in two separate files: the first file will include data on retail location, food store access, socio-demographic data, and food policy data. The second file will contain data on food prices. Both files will be available in Summer 2014.
In the Fall of 2014, ERS plans to release five descriptive research reports based on FoodAPS data which will examine food demand relationships that previously could not be investigated in detail because the requisite data did not exist (see Research Topics). More in-depth examination of food policy issues will follow, both by ERS economists and outside researchers.
ERS plans to make restricted data files from FoodAPS available to external researchers at the same time as the descriptive reports are published (see Data Access).
If it is determined that publically available data would not create too high a risk that confidential information could be disclosed, a public use database will be made available at a later date. Not only do all identifiers need to be deleted from a public use file, but responses to a number of variables need to be modified or even deleted to ensure that sample sizes in table cells are large enough to prevent identification of a household within that cell.