ERS LAFA_Generic clearance memo 1 17 23

To: Anthony Nerino, Desk Officer, Office of Management and Budget

From: Linda Kantor, U.S. Department of Agriculture, Economic Research Service

Via: Julie Parker, Clearance Officer, U.S. Department of Agriculture, Economic Research Service

Date: January 17, 2023

Subject: Request to Conduct Exploratory Semi-Structured Interviews on Retail Food Loss under Generic Clearance for Survey Research Studies (OMB Control #0536-0073)

The purpose of this memorandum is to obtain Office of Management and Budget (OMB) clearance for the first of two waves of semi-structured, exploratory interviews with food retailers to inform the design of a nationally representative study on retail-level food loss. The memorandum begins by providing information on the background and purpose for the proposed data collection, including details on prior communications with OMB, the U.S. Census Bureau, and other organizations. Next, we provide a description of the research plan for the semi-structured interviews. The memorandum concludes with the burden estimate.

Background and Purpose

Food loss is an issue of national and global importance with implications for nutrition security, environmental sustainability, and economic growth. In 2015, the U.S. Department of Agriculture (USDA) and the U.S. Environmental Protection Agency (EPA) jointly announced a goal to reduce U.S. national food loss and waste by 50% by 2030 from a 2010 baseline. USDA needs rigorously derived, nationally representative estimates of food loss and waste over time to document progress toward the 2030 goal. In a 2019 report, the U.S. Government Accountability Office identified limited data and information about food loss and waste as one of three key challenges to reducing food loss and waste in the United States.¹

The Economic Research Service’s (ERS’s) Loss-Adjusted Food Availability (LAFA) data series is one of two federal government sources of food loss data,² along with EPA.³ The LAFA data series provides the loss-adjusted per capita amount of food, calories, and Food Pattern Equivalents (aka “servings”) available for consumption in the United States for 215 commodities and estimates the amount of food loss at retail and consumer levels nationwide. Since LAFA’s creation in 1997, ERS has undertaken a series of initiatives to update the original underlying loss factors, many of which were from the 1970s or earlier. ERS estimated new retail-level loss shares for fresh fruits and vegetables and some meat and poultry products in 2009⁴ and 2016⁵ by calculating the difference in sales and shipment weights for the selected foods.

For the 2009 study, ERS obtained updated food loss coefficients (i.e., loss factors) at the retail level for individual fresh fruits, vegetables, and meat and poultry products and aggregate coefficients for all fish and all shellfish from the Nielsen Perishables Group, Inc., an independent consulting firm. The Perishables Group used 2005–2006 data from a convenience sample of six large national and regional supermarket retailers from their proprietary database. The proprietary data represented point-of-sale data collected from conventional U.S. grocery store chains that had more than $2 million in sales annually.

The data sample included information from over 600 stores from six large national and regional food retail chains located in all four U.S. regions—East, South, Central, and West. Data were collected for 2005 and 2006 full calendar years. The sample did not include convenience stores; megastores; club stores; and small, family-owned, independent “mom-and-pop”–type grocery stores. To be used in the analysis sample, data from each store under consideration had to have both weekly shipment data on a particular food commodity (e.g., actual purchases, measured in pounds, of fresh boneless chicken breast shipments sent from a supplier to the store) and point-of-sale data on consumer purchases (e.g., scanner data showing the pounds of fresh boneless chicken breasts sold). The Perishables Group used these two types of information to match shipment data to purchase data so that loss coefficients could be calculated as the residual for each fresh food commodity in the study.

The Perishables Group solicited additional qualitative loss and shipment data and insights for meat, poultry, and seafood (i.e., fish and shellfish, both farm-raised and wild caught) from more than 10 retailers’ supermarket executives. Retailers did not track meat, poultry, and seafood shipment data as effectively as they tracked produce shipment data so detailed shipment data were not available for these departments. The Perishables Group developed and distributed a supporting study among their retail contacts to obtain additional insight about loss rates for meat, poultry, and seafood. The new coefficients for both fresh fruits and vegetables and meat, poultry, and seafood were incorporated into the LAFA data series for the study commodities but had little impact on aggregate per capita availability because the new coefficients were, on average, close to the existing estimates.

In 2016, ERS commissioned the Perishables Group to update the 2009 coefficients with 2011–2012 data for fresh fruits, vegetables, meat, poultry, and seafood. The convenience sample included one large national and four regional supermarket retailers from their proprietary database, which provided data from roughly 2,900 stores in 45 states and the District of Columbia. The sample did not include convenience stores, supercenters, club stores, or mom-and-pop grocery stores. To identify a loss coefficient for each retailer, commodity, and study year (2011 and 2012), total supplier shipment data were paired with corresponding total point-of-sale data (aggregated across all stores for each retailer in the sample).

Average loss coefficients were then calculated for each commodity by equally weighting the coefficients by retailers providing estimates for that commodity. Because of the omission in the retailer calculations of random-weight meat, poultry, and seafood, the updated shrink estimates were not appropriate for updating food loss estimates at the retail level for these commodities.

ERS then used the new loss coefficients to analyze how the loss-adjusted per capita availability for the study commodities would change if the 2011–2012 coefficients were used. The updated estimates for fruits and vegetables, which included both UPC-coded and random-weight items, were found to be generally close to the current loss assumptions, but there would be greater impacts for fresh meat, poultry, and seafood if the new estimates were used. For fresh meat, poultry, and seafood, data were only available for case-ready items with a UPC, or roughly two-thirds of that market, but not for random-weight items (i.e., loose items sold by weight, roughly one-third of that market). Based on recommendations from an ERS-sponsored expert panel, ERS incorporated the updated 2011–2012 fresh fruit and vegetable loss factors from 2011 to the current year and used linear interpolation for the years 2007–2010.

In 2017, an ERS-sponsored expert panel, charged with identifying data gaps and technical weaknesses in the LAFA data series, recommended that ERS develop updated and nationally representative retail-level loss factors for all LAFA commodities, beyond just fruits and vegetables.⁶ In addition, the panel recommended that ERS collect data for the estimation from a wider range of grocery retailers, including supercenters and independent grocers.to improve generalizability.

In September 2018, ERS contracted with RTI International to develop updated, nationally representative, and fully documented estimates of retail food loss in the United States for more than 200 LAFA commodities and to answer the following questions:

• What is the amount of the available U.S. food supply at the retail level that goes uneaten in total and for each LAFA commodity or commodity group?

• What are the major drivers of food loss at the retail level and how do these drivers differ across retail commodities (e.g., fresh produce, dairy, meats and poultry, and frozen foods)?

Before filing a clearance package for the planned Field Test, which was to take place before conducting a full nationally representative data collection, ERS met with OMB in May 2019 to discuss the proposed study design. Under the original plan, RTI would develop loss coefficients for LAFA commodities as the difference between the total annual weight of product sold in stores and the total annual weight of product shipped to stores expressed as a percentage of the shipment weight for the most recent year available. For the Field Test, RTI would collect detailed product-level shipment and sales data from a sample of 20 retail stores (e.g., grocery stores), along with data on intra-store transfers of food (i.e., products used to prepare foods on site) and charitable food donations, if available. Product donations and intra-store transfers would be reflected in shipments into the store but would not be included in sales data because they are not sold directly to customers. These data would be needed to adjust the calculations so that they would not be included in the loss factor estimates. The plan was to collect store-level data from the corporate headquarters and provide an electronic template that could be used to provide data in a standardized format for each sampled store. Each sampled store would also be asked to complete a short survey on the drivers of food loss in the sampled store, that is, why food is unable to be sold (e.g., spoilage, theft, product recalls).

ERS’s OMB desk officer and other staff expressed concerns about the coverage of Nielsen TDLinx, a proprietary commercial database that was proposed as the sample frame for the Field Test and subsequent full data collection. For example, because TDLinx excludes grocery stores with less than $1 million in annual sales (e.g., small Mom & Pop/corner grocery stores) in its grocery trade classification, OMB was concerned that smaller stores that may have different food loss patterns might be excluded from the data collection. In addition, OMB expressed doubts about the willingness of food retailers to share their confidential data with a government agency, which would negatively affect response rates. They were also concerned about asking companies to use a standardized electronic template for data capture that may not be compatible with how some retailers store and maintain their data. To address these concerns, OMB asked ERS/RTI to consult with the U.S. Census Bureau to see if they might have a lower cost and lower burden mandatory survey vehicle that would accomplish the study goals. This approach would give RTI access to the Census Bureau’s larger and more complete sample frame and increase the likelihood of survey response by leveraging the Census Bureau’s preexisting relationships with corporate entities.

In July 2019, ERS and RTI began a series of meetings with staff from the Census Bureau’s Economic Reimbursable Surveys Division (ERSD) to discuss how ERS and Census might work together to collect the data that RTI would need to calculate loss factors for LAFA commodities, starting with a Field Test. Five in-person meetings or conference calls were held between July 2019 and January 2020. Our initial discussions focused on the suitability of different survey vehicles for collecting the study data, including the Commodity Flows Survey (CFS), the Annual Retail Trade Survey (ARTS), and the Annual Business Survey (ABS). The CFS collects data from wholesalers rather than retailers and was deemed unsuitable. The ARTS could have potentially been a good fit for the data collection but was unavailable as it was undergoing methodological redesign and testing. The ABS was, therefore, the only remaining survey vehicle available for the study.⁷

The ABS collects firm-level data from all nonfarm employer businesses filing Internal Revenue Service tax returns with receipts of $1,000 or more. About 300,000 total employer businesses across all industries were surveyed (at the establishment level) in 2018 and 2019. Estimates generated from the survey include number of employer firms; sales and receipts; annual payroll; and employment by sex, ethnicity, race, and veteran status and are available at the national, state, metropolitan statistical area, county, and economic place levels. Estimates are also available by North American Industry Classification System sector. Additionally, the ABS is designed to incorporate new content each survey year based on topics of relevance, thus making it potentially suitable to add new content for a nationally representative data collection on retail food loss. New questions are submitted to OMB each year for approval.

However, as we progressed through our meetings with Census/ERSD and they learned more about the detailed type of data that we planned to request from retailers, significant obstacles were identified. The ABS does not ask responding businesses to upload data, although technically they could be asked to do this. In addition, the ERSD is not familiar with collecting product-level data at the level of detail that ERS would need (e.g., UPC and random-weight level for LAFA commodities, like beef, cheese, and eggs). ERSD staff suggested using the ABS to pilot test a method for collecting store-level data on shipments and sales. For example, one test could ask respondents to input the requested data into a preformatted spreadsheet, while another could test a questionnaire format. Another test could assess respondents’ ability to attach a large dataset to the online survey. The earliest that Census could conduct such a test would be 2022 with a minimum of two rounds of testing anticipated. This would have taken us past the period of performance of the RTI contract, which ends in September 2022 (the contract has subsequently been extended by 1 year to September 2023). During our final meeting in January 2020, Census/ERSD, RTI, and ERS jointly decided to abandon the ABS approach and replace it with additional exploratory work that would inform the appropriate methodology for a Field Test and ultimately for a full nationally representative data collection.

ERS and RTI subsequently learned that the nonprofit group ReFED,⁸ which has a Memorandum of Understanding with USDA’s Office of the Chief Economist to collaborate on food loss and waste measurement issues (Appendix A), already had a pilot project underway to collect sales and shipment data (in dollar value) from companies that are members of the Pacific Coast Collaborative⁹ to estimate retail-level food loss. The pilot is a test of ReFED’s newly developed “Grocery Store Calculator,” part of their Insights Engine Food Waste Monitor¹⁰ that will calculate retail-level food loss amounts using sales and shipment data inputted by food retailers and prices and package weight data supplied by third-party firms. At OMB’s request, RTI and ERS met several times with ReFED in 2020 to learn more about their study and discuss concerns that respondents may be unduly burdened with similar data requests from two different organizations. After extensive discussions over several months, ERS and ReFED mutually concluded that our two studies were different enough in timing and substance; for example, ReFED’s loss estimates will be at the broad retail department level (e.g., dairy, meat, produce, and frozen food), versus the commodity level (e.g., milk, yogurt, chicken, beef and lettuce) needed by ERS for the LAFA data series, that it was appropriate to pursue our studies separately without any formal collaboration.

Research and Analysis Plan for the Semi-structured Interviews

To learn more about the nature of data that might be available to estimate food loss factors by LAFA commodity, RTI will conduct up to 25 semi-structured interviews with food retailers’ corporate staff. The interviews will be divided into two waves: 10 companies in Wave 1 and up to 15 companies in Wave 2.

The findings from these interviews will help us assess the feasibility of our original methodology for a nationally representative retail food loss study (see description on page 5), including retailer recruitment, data collection, and analysis.

Following Wave 1, RTI will prepare a memorandum for ERS that summarizes the findings of the interviews and recommendations for revisions to the interview guide for a second wave of interviews. If we learn during the interviews that the planned approach is not viable or that companies offer alternative approaches for measuring food loss given how they maintain data, we will adapt the interview guide and approach for the Wave 2 interviews to gather additional information on the best methodology for a subsequent Field Test. We will submit a separate OMB Generic Clearance package for the Wave 2 interviews that will reflect what we learned from Wave 1.

Selection and Recruitment Procedures

Soft Quota Ranges for Recruitment

The unit for the semi-structured interviews is the corporate headquarters of a food retail company (referred to as “company” or “companies” for brevity) that owns supermarkets, supercenters, or club stores.¹¹ In this context, companies own individual retail stores or chains of retail stores. For the proposed semi-structured interviews, companies will be interviewed, not individual stores, to learn about what data are maintained for estimating food loss and at what level of the organization the data are maintained.

The study is limited to supermarkets, supercenters, or club stores because they account for 85% of food and beverage retail sales in the United States.¹² Retail companies with different organizational structures (i.e., independently owned, regional chain, national chain) will be interviewed. As described in more detail below, RTI plans to work with the Food Industry Association (known as FMI) to recruit member companies to participate in some of the interviews, which will help reduce burden because RTI will not have to contact as many companies or make multiple contacts within a company to identify the target person for the interview. Companies that are not members of FMI will also be interviewed to be inclusive of all retail companies. Also, companies from different geographic locations will be interviewed to provide coverage across the four Census regions of the United States (the Northeast, Midwest, South, and West).

Soft quotas with ranges will be used for type, organizational structure, and recruiting source (i.e., FMI member or not FMI member) to ensure that a diverse mix of companies (Table 1) are interviewed (there are no quotas for region, a mix of regions will be purposively selected). Because a small number of companies own supercenters or club stores, only one or two of these company types will be interviewed. In addition, since only 10 interviews are being conducted, not all combinations of type and organizational structure will be included in the final set of interviews. The final number of companies within the soft quota ranges that is represented in each type and organizational structure category will likely be determined by company responsiveness. For example, for type, RTI might complete interviews with eight companies that own supermarkets, one company that owns a supercenter, and one company that owns a club store. The total number of companies interviewed will not exceed 10.

Table 1. Soft Quota Ranges for Recruitment for the Semi-Structured Interviews by Retailer Type, Organizational Structure, and Recruiting Source^a,b

Type

• Seven or eight companies that own supermarkets

• One or two companies that own supercenters

• One company that owns club stores

Organizational structure

• Three or four independent operators (companies that own 10 or fewer stores)

• Three or four regional chains

• Three or four national chains

Source

• Seven or eight companies that are members of FMI (the company and individual will be recruited by FMI, i.e., referral)

• Two or three companies that are not members of FMI

^a RTI will interview companies that have headquarters in different regions of the United States to provide coverage across the four U.S. Census regions (the Northeast, Midwest, South, and West); there are no soft quotas for region; a mix of regions will be purposively selected.

^b Some companies may own multiple banners, which is a set of stores defined by a common name and unified advertising programs. For these companies, RTI will purposively choose one banner that aligns with the soft quota ranges because procedures for maintaining data for estimating food loss may vary by banner.

Sample Size and Expected Response Rates

We assumed that up to 50 companies will need to be contacted to recruit 10 companies for the Wave 1 interviews. This estimate is based on RTI’s experience contacting food retailer companies earlier in this study. In an initial step under the original study design, in 2018, RTI had informal discussions with seven food retailer companies (ranging in size and type) as well as organizations of interest (ReFED, the Food Industry Association [referred to as FMI], and the National Grocers Association) to help inform the study design for the original project, and 28% of companies contacted agreed to talk with RTI by telephone (memorandum on informal discussions provided as Appendix B). For the proposed semi-structured interviews, we adopted a conservative approach to estimating participation and assumed 20% of contacted companies would participate in the interview after the initial contact. Additionally, based on RTI’s experience with the earlier informal discussions, we estimated that up to four contacts (with potentially different people) would be required to identify the correct person to participate in the interview.

Target Individual to Interview Within a Company

Target individuals for the interviews are corporate staff who are knowledgeable about how product data are maintained across their company’s individual stores (e.g., this may be someone in operations or supply chain management, a procurement specialist, or a retail/supermarket buyer). From our informal discussions with food retailers, it is our understanding that such staff will be best able to answer the interview questions. As described in more detail below, RTI research staff will use a script to identify the target individual within the company to interview for non-FMI members (we anticipate that FMI will identify the target individual for the companies that it recruits).

Selection of Companies for the Interviews and Recruitment Procedures

RTI will use Nielsen’s TDLinx database to select companies for the interviews. TDLinx is a proprietary commercial database, which includes the name, address, and corporate owner of individual food stores with at least $1 million in sales.¹³ It also provides information on store characteristics and type. Once selected, participants will be contacted by RTI staff using the procedures described below for FMI and non-FMI member companies..

Non-FMI Members

We plan to recruit two to three companies that are not FMI members. To encourage participation among companies that are not FMI members, RTI asked the National Grocers Association, which represents independent, privately owned grocery stores, about their willingness to communicate their support for the study to their member companies. In past studies of this type, RTI has found that such efforts help demonstrate the legitimacy and need for the study and may motivate companies to participate.¹⁴ The National Grocers Association has agreed to share information about the study with their membership through their newsletter the week before recruiting starts (once at the beginning of the week in their general newsletter and once in their government relations–specific newsletter at the end of the week) (Appendix C).

RTI will purposively select 20 companies from TDLinx to provide a mix of types, organizational structures, and regions (Table 1). These companies will primarily be companies classified as independent operators because FMI generally represents larger regional or national chains. RTI staff will contact companies from this list to recruit them for an interview. For any regional and national chains, (that are not FMI members) RTI will initially identify and contact someone within the company’s sustainability group (e.g., chief of sustainability, director of sustainability) to get their support for the interviews and then work with them to identify the individual(s) who maintain(s) data on product shipments and sales as the individual to take part in the interview. In discussions with ReFED, they indicated that they are using a similar approach. We understand smaller, independent operators might not have a sustainability staff/office; in these cases, RTI will identify and contact the corporate director of operations (or similar title) as a starting point and work with them to identify the appropriate person to interview. RTI will search online for publicly available information to identify the initial contact and an email address and/or phone number. Table 2 details the contact/recruitment procedures for non-FMI members.

Table 2. Contact/Recruitment Procedures for Non-FMI Members

Email address is not available

RTI will contact the company by telephone. Appendix D-1 provides the script for making the initial contact, with separate scripts for independent operators and regional/national chains. After the appropriate contact has been identified, RTI will call this person using the script provided in Appendix D-2 and, if the person agrees to participate, schedule a mutually agreeable date/time for an interview during the call or by sending a scheduling email (Appendix E-2). Appendix E-2 will include two attachments with additional information on the study; the informed consent form (Appendix F), and the list of potential interview topics (Appendix G).

Email address is available

RTI will send an email initially (Appendix E-1). If a response to the email is received, RTI will follow up by sending a recruitment/scheduling email (Appendix E-2) with the attachments, or if the email response suggests a phone contact would be preferred, then RTI will contact them by telephone using the script provided in Appendix D-2 and then send the scheduling email (Appendix E-2) with the two attachments noted above. If there is not a response to the email and a phone number is available, RTI will call this person using the script provided in Appendix D-2 and, if the person agrees to participate, schedule a mutually agreeable date/time for an interview during the call and then send the scheduling email (Appendix E-2) with the two attachments noted above. If the person prefers to schedule by email, then RTI will send a scheduling email (Appendix E-2) with the two attachments noted above.

FMI Member Companies

Following a July 2021 meeting with FMI leadership and members of their Sustainability Executive Committee, FMI agreed to recruit seven or eight member food retailers to participate in the interviews. To help FMI identify companies to recruit, RTI will provide FMI with a list of 20 purposively selected food retail companies from TDLinx to provide a mix of types, organizational structures, and regions (Table 1). An additional list will be provided, if needed. FMI will provide contact information (email and phone number) for the individual at each company that has indicated a willingness to be interviewed. Because FMI is recruiting these companies and the individual has already indicated a willingness to participate, RTI will “by-pass” the recruiting script that includes background information on the study and a script to identify the target respondent and go directly to scheduling an interview using the email in Appendix E-2. Appendix E-2 will include two attachments with additional information on the study; the informed consent form (Appendix F), and the list of potential interview topics (Appendix G). If the person identified by FMI indicates that someone else within the company should be interviewed, then RTI will use the contact procedures for non-FMI members as described in Table 2.

Follow-Up Attempts and Reminders

For non-FMI members, it may take up to four telephone call or email attempts to different people within the company to identify the correct person(s) to interview. This may also be the case if the initial contact suggested by FMI for an FMI member company refers RTI to a different person to schedule an interview. It may be possible that within such a contact attempt RTI gets transferred to multiple people within a company.

For both FMI and non-FMI members, if a response is not received from the initial call or email within 3 business days, RTI will call again or send a follow-up email. RTI will make up to four contact attempts to a company over a 2- to 3-week period. If both an email address and phone number are available, RTI will make two email attempts and two phone call attempts. If a response is not received after four attempts, the company will be coded as a nonresponse (Appendices H-1, H-2, and H-3 are scripts and emails for the follow-up attempts)

Companies that agree to an interview will get a reminder from RTI a few days before the event (Appendix I). The reminder email will include the following attachments: the informed consent form (Appendix F), and the list of potential interview topics (Appendix G).

Interview Procedures

The interviews are intended to provide a better understanding of the types of information that might be available to calculate retail food loss and expand on some of the topics that were discussed in the informal discussions. RTI interviewers will use an interview guide (Appendix J) to provide structure to the discussion. The guide contains structured probes so that all participants will initially be asked the same set of questions, but interviewers will probe spontaneously when further clarification is needed.

The interviews will be conducted via secure Zoom video conference by two RTI staff members—one person to lead the discussion and one person to take notes. The RTI interviewer will begin the interview by asking companies about their current procedures for tracking food loss. The interviewer will then solicit feedback on RTI’s proposed methodology (i.e., using data on product shipments and sales weights, adjusted for intra-store product transfers, and food donations) to calculate food loss coefficients for individual LAFA commodities or commodity groups. The interviewer will follow up with additional questions about how such data, if available, are maintained (e.g., by UPC or store department; daily, monthly, or annually) and about companies’ preferred method for providing available data (e.g., by responding to survey questions, filling out a formatted spreadsheet, or uploading data). The interviewer will ask about the anticipated burden for providing the data. The interviewer will also ask companies thoughts about giving participants in a full nationally-representative study with a benchmarking report that would compare their company’s food loss performance with an aggregated industry average.

Interviewer Training

RTI project staff responsible for developing the interview guide and procedures will conduct the interviews. Interviewers and note takers will train for the interviews by doing practice runs using the interview guide.

Incentive Payments

No incentive payments will be offered.

Informed Consent

Before participating in the interviews, RTI will ask the person(s) to be interviewed to electronically sign an informed consent using a FedRAMP-certified Qualtrics survey link (Appendix F). The informed consent will be included in the scheduling email (Appendix E-2) that will be sent to everyone who will be participating in the interview. Additional information about the procedures for obtaining informed consent is provided in the section below on confidentiality procedures.

Confidentiality Procedures

RTI’s Institutional Review Board (IRB) reviewed the Not Human Research Determination request for the study and determined that the proposed activity is not research involving human subjects as defined by Department of Health and Human Services regulations (Appendix K). Thus, RTI IRB oversight of this activity is not required.

RTI has assessed and secured its information systems in accordance with the Federal Information Security Management Act, which appears as Title III of the E-Government Act of 2002 and the National Institute of Standards and Technology (NIST). For information systems containing data protected under the Confidential Information Protection and Statistical Efficiency Act (CIPSEA), the security characterization reflects a moderate potential impact and requires that RTI store and work with any data in its Federal Information Processing Standards Moderate Enhanced Security Network.

The informed consent form (Appendix F) will provide assurances regarding data confidentiality and security as described below. If multiple people within a company participate in the interview, each one will be emailed a copy of the informed consent form prior to the interview. If a person who has not previously provided informed consent will be participating in the interview, RTI will email them the form and have them sign electronically before starting the interview (i.e., in real time). Statement on CIPSEA protections: “The information you provide will be used for statistical purposes only. Your response will be kept confidential and any person who willfully discloses ANY identifiable information about you or your company is subject to a jail term, a fine, or both. This study is conducted in accordance with the Confidential Information Protection and Statistical Efficiency Act of 2018, Title III of Pub. L. No. 115-435, codified in 44 U.S.C. Ch. 35 and other applicable Federal laws.”

• RTI’s Zoom platform is configured to meet NIST-moderate requirements.

• The interview notes (typed and saved in individual Word documents) will be saved on RTI’s Enhanced Security Network. Additionally, RTI will prepare an Excel file that summarizes responses to the interviews using a matrix format for organizing and summarizing the interview data to prepare a written report. The summary Excel file will be saved on RTI’s Enhanced Security Network. Any handwritten notes taken during the interviews will be shredded after the notes have been transcribed into the Word and Excel files.

• RTI will use the interview responses to prepare a summary report for ERS.

• At the end of the study, the interview responses in Word/Excel files and the Excel file used to organize and summarize the interview data will be transferred to ERS’s secure FISMA-moderate cloud-based network using the FISMA-moderate Kiteworks file sharing network. RTI will destroy the files on their Enhanced Security Network and provide documentation of the destruction to ERS.

Burden Estimate

• The total response burden for Wave 1 is estimated to be 29.56 hours.

• The estimated total response burden for the participants group is 16.24 hours, or about 96 minutes per company. The estimated total response burden for companies who go through the recruitment process but do not participate is 13.32 hours, or about 20 minutes per company. Table 3 provides a detailed breakdown of the response burden for the recruitment and interview process.

• RTI’s previous experience suggests that participating in a recruitment call (Appendices D-1 and/or D-2) or reading the recruitment email (Appendices E-1and/or E-2) will take an average of 5 minutes, and it may be necessary to talk to multiple people within the company (e.g., transferred to a different person during the call or asked to call/email a different person, which would count as an additional contact). We assumed that RTI will make a maximum of four attempts to identify the correct person(s) to interview within a selected company. For estimating burden, we assumed that all companies will require four contacts to reflect the potential maximum burden (Appendices H-1, H-2, and H-3 provide the follow-up recruitment scripts and email). Although companies and target respondents that are recruited by FMI are likely to be more responsive, we do not know that with certainty. Also, FMI may encounter challenges recruiting companies/target respondents, so RTI staff may have to contact FMI member companies and go through the recruiting process. For these reasons, we erred on the side of caution and assumed the same burden for FMI and non-FMI member companies so that the burden estimates reflect the potential maximum burden. We did not include time spent by FMI staff for recruiting companies and target respondents because FMI staff are not participating in the interviews themselves.

• For companies that agree to take part in the interview, we assumed that reading and completing the informed consent form (Appendix F) and reviewing the list of potential interview topics (Appendix G) will take an average of 15 minutes.

• We assumed that one company will go through the process of preparing for the interview, including submitting the consent form, and gathering materials, but will drop out before reading the reminder email and participating in the interview itself.

• For companies that agree to take part in the interview, we assumed that reading the reminder email (Appendix I) will take 1 minute.

• The interviews will take an average of 60 minutes following the semi-structured interview guide (Appendix J).

Table 3. Reporting Burden

Instrument	Appendix	Sample Size	Freq	Respondents				Nonrespondents				Total Burden Hours
				Count	Freq x Count	Minutes/Resp	Burden Hours	Count	Freq x Count	Minutes/Resp	Burden Hours
Recruitment call and/or email contact #1a	D-1, D-2, E-1, and/or E‑2	50	1	10	10	5	0.83	40	40	5	3.33	4.16
Recruitment call and/or email contact #2a	H-1, H-2, and/or H-3	50	1	10	10	5	0.83	40	40	5	3.33	4.16
Recruitment call and/or email contact #3a	H-1, H-2, and/or H-3	50	1	10	10	5	0.83	40	40	5	3.33	4.16
Recruitment call and/or email contact #4a	H-1, H-2, and/or H-3	50	1	10	10	5	0.83	40	40	5	3.33	4.16
Informed consent, list of interview topics	F, , and G	11	1	11b	11	15	2.75	39	39	0	0.00	2.75
Reminder email	I	10	1	10	10	1	0.17	40	40	0	0.00	0.17
Interviewsc	J	10	1	10	10	60	10.00	0	0	0.0	0.00	10.00
Total burden							16.24				13.32	29.56

^a The sample unit for this study is food retail companies, not individual people. We assumed that RTI may need to contact multiple people within each company to identify the correct person (s) to interview, and each contact will take an average of 5 minutes and that up to four contacts will need to be made. Therefore, the minutes per response column reflects the amount of time required for the contact and not the number of people contacted within the company.

^b We assumed that one company will go through the process of preparing for the interview, including reading and submitting the consent form, and reading the list of interview topics, but will drop out before participating in the interview itself.

^c Multiple people within a company may participate in the interview. Because the sample unit for this study is food retail companies, not individual people, the minutes for completing the interview does not reflect participation by multiple people.

• The estimated total number of contacts for the Wave 1 interviews is 50 companies. This number can be broken down into two groups: an estimated 40 companies that go through the recruitment process but decline to participate (i.e., nonrespondents) and 10 companies that participate (i.e., respondents).

The following statements will be displayed in the informed consent for the interview:

The data are being collected under the Confidential Information Protection and Statistical Efficiency Act, otherwise known as CIPSEA. CIPSEA provides protection for information collected for statistical purposes under a pledge of confidentiality. CIPSEA-protected information is not subject to Freedom of Information Act requests. RTI will combine your interview responses with the responses from other companies to prepare a summary report for USDA, ERS.

Assurance of Confidentiality: The information you provide will be used for statistical purposes only. Your response will be kept confidential and any person who willfully discloses ANY identifiable information about you or your company is subject to a jail term, a fine, or both. This study is conducted in accordance with the Confidential Information Protection and Statistical Efficiency Act of 2018, Title III of Pub. L. No. 115-435, codified in 44 U.S.C. Ch. 35 and other applicable Federal laws.

NOTIFICATION TO RESPONDENT OF ESTIMATED BURDEN: Public reporting burden for this collection of information is estimated to average 127 minutes per response, including the time for reviewing instructions and completing and reviewing the collection of information. An agency may not conduct or sponsor, and a person is not required to respond to, a collection of information unless it displays a currently valid OMB control number. The valid OMB control number for this information collection is #0536-0073. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing this burden, to Linda Kantor, U.S. Department of Agriculture, Economic Research Service ([email protected]).

Contact Information

The contact person for questions regarding this data collection is:

Linda Kantor

[email protected]

Appendices

• Appendix A: Memorandum of Understanding—USDA and ReFED

• Appendix B: Memorandum on Informal Discussions Conducted by RTI

• Appendix C: Text for National Grocers Association Newsletter

• Appendix D-1: Recruitment Script for Non-FMI Members (no contact identified)

• Appendix D-2: Recruitment Script for Non-FMI Members (contact identified)

• Appendix E-1: Initial Recruitment Email for Non-FMI Members

• Appendix E-2: Scheduling Email—Participant Identified for non-FMI members or FMI Referral

• Appendix F: Informed Consent for Interviews

• Appendix G: Informational List of Interview Questions and Topics

• Appendix H-1: Recruitment Follow-Up Script for Non-FMI Members (no contact identified)

• Appendix H-2: Recruitment Follow-Up Script for Non-FMI Members (contact identified)

• Appendix H-3: Recruitment Follow-Up Email for Non-FMI Members

• Appendix I: Reminder Email

• Appendix J: Interview Guide

• Appendix K: IRB Determination

1 U.S. Government Accountability Office. (2019, June). Food loss and waste: Building on existing federal efforts could help to achieve national reduction goal. GAO-19-391. https://www.gao.gov/assets/710/700118.pdf

2 U.S. Department of Agriculture, Economic Research Service. (2020, November 12). Loss-Adjusted Food Availability documentation. https://www.ers.usda.gov/data-products/food-availability-per-capita-data-system/loss-adjusted-food-availability-documentation/

3 EPA maintains data on the flows of excess food and food waste throughout the food system and generates annual estimates of generation and management of excess food and food waste by sector. See https://www.epa.gov/sites/production/files/2020-06/documents/food_measurement_methodology_scoping_memo-6-18-20.pdf for additional information.

4 Buzby, J. C., Wells, H. F., Axtman, B., & Mickey, J. (2009, March). Supermarket loss estimates for fresh fruit, vegetables, meat, poultry, and seafood and their use in the ERS Loss-Adjusted Food Availability data. https://www.ers.usda.gov/webdocs/publications/44306/10895_eib44.pdf?v=2533.6

5 Buzby, J. C., Bentley, J. T., Padera, B., Campuzano, J., & Ammon, C. (2016, June). Updated supermarket shrink estimates for fresh foods and their implications for ERS Loss-Adjusted Food Availability data. https://www.ers.usda.gov/webdocs/publications/44100/eib-155.pdf?v=4892.8

6 RTI International. (2018, May). Expert panel on technical questions and data gaps for the ERS Loss-Adjusted Food Availability (LAFA) data series (Contractor and Cooperator Report. No. CCR-70). U.S. Department of Agriculture. https://www.ers.usda.gov/webdocs/publications/92409/ccr-70.pdf?v=5391

7 See https://www.census.gov/programs-surveys/acs.html for additional information on the ABS. See abs_2021.pdf (census.gov) for the questionnaire for the 2021 ABS.

8 ReFED is a national nonprofit working to end food loss and waste across the U.S. food system (https://refed.com/).

9 Pacific Coast Collaborative. (n.d.). Food waste reduction: Benefits for the planet, people – and your business. https://pacificcoastcollaborative.org/food-waste/

10 ReFED. 2020. Insights Engine Food Waste Monitor: 2020 Methodology, https://insights.refed.org/uploads/documents/refed_insights_enginefood_waste_monitormethodology_vfinal_2021.04.12.pdf?_cchid=4f1d745d5f50363e83869ef5f3b2e962

11 A supermarket sells a wide variety of food, beverages, and household products. It has a wider selection than grocery stores but is smaller and more limited in the range of merchandise than a supercenter or club store. Supercenters are large stores that combine nonfood mass merchandise with supermarkets. Club stores are large outlet stores that sell food and beverages in bulk and require consumers to buy a membership.

12 Muth, M.K., Kinney, S., Gargano, M., Looby, C., & Siegel, P. (2021, April). User documentation: Store weights for InfoScan data, 2012–2018. RTI International.

13 Cho, C., McLaughlin, P.W., Zeballos, E., Kent, J., & Dicken, C. (2019). Capturing the complete food environment with commercial data: A comparison of TDLinx, ReCount and NETS databases (TB-1953). U.S. Department of Agriculture, Economic Research Service. https://www.ers.usda.gov/publications/pub-details/?pubid=92628

14 Cates, S. C., Viator, C. L., Karns, S. A., & Muth, M. K. (2008). Food safety practices of meat slaughter plants: Findings from a national survey. Food Protection Trends, 28(1), 26–36.