EVALUATING BEHAVIORAL NUDGES TO REDUCE HOUSEHOLD FOOD WASTE BY DEVELOPING A SMARTPHONE APP

• Sponsoring Institution: National Institute of Food and Agriculture
• Project Status: COMPLETE
• Funding Source: AFRI COMPETITIVE GRANT
• Reporting Frequency: Annual
• Accession No.: 1012390
• Grant No.: 2017-67023-26268
• Cumulative Award Amt.: $499,384.00
• Proposal No.: 2016-09934
• Project Start Date: Apr 1, 2017
• Project End Date: Mar 31, 2021
• Grant Year: 2017
• Program Code: [A1641]- Agriculture Economics and Rural Communities: Markets and Trade

Recipient Organization
OHIO STATE UNIVERSITY
1680 MADISON AVENUE
WOOSTER,OH 44691

Performing Department
Agric, Evrnml & Dev Econ

Non Technical Summary
About one-third of U.S. food is wasted with two-thirds of this waste attributed to consumers. The United States proposes to cut food waste in half by 2030 and, while many policies are available, those that change consumer waste behavior will be integral to achieving U.S. goals. However, we lack quality data about U.S. household food waste because popular data collection methods (diaries and waste stream analyses) suffer major shortcomings. Hence, rigorous evaluation of consumer interventions proves elusive.We propose to improve the accuracy and convenience of household food waste measurement so that consumer food waste program evaluations yield more power, less bias, less measurement error, and greater representation of targeted populations. We will improve measurement via the continued development of the FoodImage smartphone app, a technology leveraging the investigators' expertise in creating and deploying apps to measure food intake and deliver nutritional interventions in free-living household conditions. We will validate the app against current best practices with consumers in controlled laboratory settings and then deploy the app in free-living conditions as part of a pilot randomized controlled trial to evaluate if behavioral nudges in the form of personalized feedback and goals can reduce household food waste.The study will yield an app ready for future study deployment while analyses of the data will inform rapidly evolving policy discussions concerning optimal approaches to reduce food waste. These outcomes align with program priorities to understand the economics of food waste and to wield behavioral economics to address sustainable consumption behavior.

Animal Health Component

50%

Research Effort Categories

Basic

0%

Applied

50%

Developmental

50%

Classification

Knowledge Area (KA)	Subject of Investigation (SOI)	Field of Science (FOS)	Percent
607	5010	3010	50%
403	6010	3100	50%

Knowledge Area
403 - Waste Disposal, Recycling, and Reuse; 607 - Consumer Economics;

Subject Of Investigation
6010 - Individuals; 5010 - Food;

Field Of Science
3010 - Economics; 3100 - Management;

Keywords

behavioral econ

food waste

nudge

smartphone app

randomized controlled trials

social comparison

Goals / Objectives
Continued development of FoodImage, a smartphone app that measures household food waste and food-waste-related behaviors,Assess differences in the accuracy, time burden and perceived convenience of measurements taken with the FoodImage app versus the current state-of-the-art household food waste measurement approach (pen-and-paper diary), andUse the FoodImage app in a pilot randomized controlled trial designed to evaluate the effectiveness of reducing household food waste through behavioral nudges in the form of personalized feedback about food waste levels and household-specific reduction goals.

Project Methods
App DevelopmentThe FoodImage app will send participants' images and accompanying data (e.g., barcodes, PLU numbers, receipts, food descriptions, Participant ID number,timestamp) to a secure PBRC server where a team will analyze the images to estimate food waste. The program identifies a match for each food from the Food and Nutrient Database for Dietary Studies 5.0 and other sources to calculate energy and nutrient intake. Experienced raters use the program to estimate portion size via an existing and validated methodology. This process results in estimates of food waste in grams, which is then converted to energy and macronutrient content for each item.Data to translate food waste measurements into dollar values will be taken from scanned receipts where possible and from alternative sources (e.g., BLS food price data) as a backup. Greenhouse gas potential and resource burden for measured food waste will be taken from the literature.The unique coverage and capabilities of SmartIntake®Comprehensive measurement of food waste. The FoodImage app will quantify food waste at all stages where consumer food waste occurs (e.g., left-overs after a meal, cabinet or refrigerator/freezer purges, food preparation, etc).Coverage of all household members. This will require an articulation of household members during set up and initialization to account for the plate waste of each household member at each eating occasion and for the food preparation and purging activities of other members.Articulation of source, reasons and destinations for photographed food waste.Scripts will be developed where participants log the food waste's source, reason and destination.A drop-down menu featuring the most likely responses will allow for rapid entry while an 'other' option will allow for open ended written or voice recording responses to be analyzed by researchers later. Source options include: home-prepared meal, take-out meal, dine-out meal, leftovers, and other.Reason options include: after label date, moldy/spoiled, doesn't taste/look good, improper preparation, left out of refrigerator too long, plate waste, amount not enough to save, inedible parts, don't want to eat as leftovers, and other.Destination options include: garbage, disposal/sink, on-site compost, fed to pet, and other.Receipt scanning capability. An exploratory module will be developed in which receipts from food purchases are scanned.Subject RecruitmentAsample of 44 adults will be recruited from the Baton Rouge, Louisiana area to participate in Phase I (n=24) and Phase II (all 44 participants). Inclusion criteria 1) age 18-72 years, 2) self-reported body mass index 18.5 - 50 kg/m2, 3) conducts at least 75% of household shopping and food preparation. Exclusion criteria 1) not able to use an iPhone, 2) refusal or unable to use a smartphone app or kitchen food waste pen-and-paper diary to collect data in the lab during Phase I, 3) refusal or unable to use the smartphone app to collect data for up to a 14 day period in free-living conditions, and 4) households that purchase groceries less than 1 time per week.Participants will be compensated $30 for successful completion of Phase I and $175 for successful completion of Phase II (maximum compensation = $205).Phase I: Testing the FoodImage App with Consumers in a Laboratory Kitchen and Dining Setting Conduct a lab-based consumer study to test the measurement features of the FoodImage app under controlled conditions and to obtain user satisfaction data (via a questionnaire) and feedback (via small focus groups and cognitive interviews) about the app. Accuracy will also be evaluated for both the FoodImage app and the current best practice (a pen-and-paper food diary with food scale measurement) by comparing estimated food waste from these two methods to directly weighed food waste. Food waste will be simulated in the laboratory for this phase of the research by the researchers.The following conditions will be simulated in the PBRC's Ingestive Behavioral Lab's (IBL) Prep Kitchen, and participants will be asked to use each method to perform the following tasks:Measuring food waste created during a simulated food preparation setting with foods that include edible and non-edible food waste.Measuring food waste created during simulated eating conditions with different amounts of plate waste. The plate waste will differ in its ability to be saved as left-overs or disposed of via composting, garbage disposal, fed to pets, or garbage/landfill.Measuring food waste created during simulated cabinet and refrigerator clean-outs (purges) due to the discarding of spoiled foods, foods that exceeded the expiration date, foods that were edible but will not be eaten, left-overs that were not eaten, etc.Measurements from activities a, b, and c will be collected by participants with both the FoodImage app and via the pen-and-paper diary (with scale), while lab personnel directly weigh foods to provide the criterion value.Logging simulated food purchases from grocery stores and restaurants, including images of foods purchased and scans of receipts, barcodes, price look-up (PLU) numbers, etc.These data will be collected to determine the extent to which receipt scans are a viable method to quantify and value food purchases to calculate the dollar value of discarded foods.Participants then complete a survey to report satisfaction with the FoodImage app and the pen-and-paper diary. This is followed by either a small focus group or a cognitive interview as appropriatePhase II: Randomized Controlled Trial of Behavioral Nudges to Reduce Household food wastePhase II will occur in participants' natural environment (free-living conditions) and will involve the 24 participants from Phase I and an additional 20 more participants. Following confirmation of eligibility and informed consent, 44 participants will be trained to use the FoodImage app (for the 24 Phase I participants, this will occur at the end of Phase I) and asked to complete a survey of key household, demographic and attitudinal information. Participants will use the FoodImage app to capture data on food purchases, food waste that occurs during food preparation, food waste that is present after eating, and food waste from stored food purges. During collection, subjects will record the source, reason and destination of all food waste on the app.In Phase II all participants will capture baseline food waste levels for 7 days and then will be randomly assigned (1:1 ratio) to either:Control, who continue to use the app to record food waste for 7 additional days orTreatment, who are providedFeedback on the amount of food waste their household created during the first week,A goal to reduce the next week's food waste by 20% or more, andTips on how to reduce household food waste adapted from current consumer campaigns. A PBRC staff member will deliver tips to each respondent in this group via a remote informational session on Day 8 with printed copies of tips made available thereafter.A brief survey to assess comprehension of the tips and information provided.Randomization will occur via an optimal multivariate matching before randomization technique. A matching algorithm (e.g., Mahalanobis) creates pairs balanced on hypothesized drivers of food waste and then one member of each pair is randomly assigned to treatment. Key matching variables will include household size, demographics, and participation in Phase I.Both groups will expect to communicate with researchers on day 8. The control and treatment groups will have any questions about the app answered. The treatment group will also receive the additional information detailed above. Both groups complete a survey after day 14 that assesses satisfaction with the app and allows for open-ended feedback and suggestions about the app and, for the treatment group, about the perceived efficacy of the food waste reduction tips.

Progress 04/01/17 to 03/31/21

Outputs
Target Audience:The target audiences reached during the grant includes academics, policy makers, and practitioners in the food and waste management sectors. Academics and policy makers were reached via continued downloads of: an article published in PLOS ONE in 2018 that was viewed or downloaded >4000times and cited at least 33times; an article published in PLOS ONE in 2020 that was viewed or downloaded >1000timesand cited at least 3times; an article published in Resources, Conservation and Recycling that was published in 2020 and cited 3 times; and an article published in Applied Economics Perspectives and Policy in 2020 that has been cited 11 times. Academics and practitioners were also reached through presentations at multiple in person and online events annually. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?The project provided training for one Ph.D. student over one year in the intersection of economics research, MHealth interventions and wasted food. How have the results been disseminated to communities of interest?The results have been shared via both professional presentations to groups of academics, practitioners and policy makers and via peer-reviewed journal articles that have been documented as part of the annual reporting process. What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

Impacts
What was accomplished under these goals? 1.Continueddevelopment of FoodImage, a smartphone appthat measures household food waste and food-waste-related behaviors: The FoodImage smartphone app was created and is now available from Apple's App Store. Trained subjects will use app to capture photos that document food purchasing, food preparation, food intake and food discard in a manner that permits researchers to generate data about food acquisition and the amount and disposition of food wasted during preparation, consumption, and periodic storage clean outs. Researchers now have the ability to deploy this highly granular data collection system which, to our knowledge, is unique in its abilityto provide such adata collection approach that is also integrated with information delivery. 2. Assess differences in theaccuracy, time burden and perceived convenienceof measurements taken with the FoodImage app versus the current state-of-the-art household food waste measurement approach (pen-and-paper diary): The research team published an article in 2020 that documents that the FoodImage app provides food waste measures that are as or more accurate than previously state-of-the-art household food waste measurement approaches (pen-and-paper diaryies); required less time of research participants; and was perceived as more convenient by research participants. 3.Use the FoodImage app in a pilotrandomized controlled trialdesigned to evaluate the effectiveness of reducing household food waste throughbehavioral nudgesin the form of personalized feedback about food waste levels and household-specific reduction goal: The research team has a manuscript that summarizes the results of arandomized controlled trial that demonstrates the efficacy ofbehavioral nudgesin the form of personalized feedback about food waste levels in reducing the amount of food wasted during dining situations as measured using the FoodImage app. 4. The research pioneers the use of the MHealth technique of personalized coaching to address the issue of wasted food in households and provides evidence that this can be an effective approach for reducing the amount of food wasted in dining situations.

Publications

• Type: Journal Articles Status: Under Review Year Published: 2021 Citation: Roe, BE, D Qi, RA Beyl, KE Neubig, JW Apolzan, CK Martin. A Randomized Controlled Trial to Address Consumer Food Waste with a Technology-aided Tailored Sustainability Intervention
• Type: Conference Papers and Presentations Status: Published Year Published: 2020 Citation: Roe, Brian E., D. Qi, R.A. Beyl, K.E. Neubig, C.K., Martin, J.W. Apolzan. 2020. The FoodImageTM Smart Phone App: A Tool for Assessing Food Acquisition, Consumption & Waste, Webinar on Household Food Waste Measurement: Alternatives to Diaries and Digs, October 16.

Progress 04/01/19 to 03/31/20

Outputs
Target Audience:The target audiences reached during the year ending 31 March 2020 includes academics, policy makers, and practitioners in the food and waste management sectors. Academics and policy makers were reached via continued downloads of an article published in PLOS ONE during the previous reporting period that was viewed or downloaded 2179 times during the current reporting period and cited at least 7 times during the 2019 calendar year. Academics were also reached through a presentation at the Agricultural and Applied Economics Department of the University of Georgia. ? Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided? Nothing Reported How have the results been disseminated to communities of interest?Communities of interest reached during the year ending 31 March 2020 includes academics, policy makers, and practitioners in the food and waste management sectors. Academics and policy makers were reached via continued downloads of an article published in PLOS ONE during the previous reporting period that was viewed or downloaded 2179 times during the current reporting period and cited at least 7 times during the 2019 calendar year. Academics were also reached through a presentation at the Agricultural and Applied Economics Department of the University of Georgia. What do you plan to do during the next reporting period to accomplish the goals?An article documenting the Randomized Controlled Trial from Phase II will be completed and submitted for publication and presentations of the results will be shared at appropriate conferences and meetings.

Impacts
What was accomplished under these goals? The FoodImage smartphone app was deployed during Phase I and Phase II of the study. Phase I assesses the accuracy, time burden and perceived convenience of measurements taken with the FoodImage app versus the current state-of-the-art household food waste measurement approach (pen-and-paper diary conducted via visual estimation or visual estimation aided with scale). These data were collected in a laboratory setting from subjects tasked with assessing food purchases from a shopping trip and food wasted in standardized food preparation, food consumption and food storage disposal settings. One publication has been accepted that documents the performance of FoodImage compared to alternative approaches. Phase II of the studied featured deployment of the FoodImage app in a pilot randomized controlled trial that evaluates the effectiveness of reducing household food waste through behavioral nudges in the form of personalized feedback about food waste levels and household-specific reduction goals provided to a randomly selected group of participants who tracked their food purchases, food intake and food waste using the app in their daily lives. The data is continuing to be cleaned and analyzed with outlines for a publication being developed.

Publications

• Type: Conference Papers and Presentations Status: Published Year Published: 2019 Citation: Roe, Brian E. 2019. Emerging Issues in the Economics of Food Waste, invited presentation, University of Georgia, April 24.
• Type: Journal Articles Status: Submitted Year Published: 2020 Citation: Roe, Brian E., Danyi Qi, John W. Apolzan, and Corby K. Martin. Selection, Intake, and Plate Waste Patterns of Leftover Food Items From a Pilot Study with U.S. Consumers.
• Type: Journal Articles Status: Awaiting Publication Year Published: 2020 Citation: Roe, Brian E., Danyi Qi, Robbie A. Beyl, Karissa E. Neubig, Corby K. Martin & John W. Apolzan. (2020). The Validity, Time Burden, and User Satisfaction of the FoodImageTM Smartphone App for Food Waste Measurement Versus Diaries: A Randomized Crossover Trial. Resources, Conservation & Recycling, 160(Sep): 104858. DOI: 10.1016/j.resconrec.2020.104858

Progress 04/01/18 to 03/31/19

Outputs
Target Audience:The target audiences reached during the year ending 31 March 2019 includes academics, policy makers, and practitioners in the food and waste management sectors. Academics were reached via continued downloads of an article published in PLOS ONE during the previous reporting period that was viewed 1157 times and downloaded 706 times during the current reporting period. Academics were also reached through a paper presentations at the Annual Meetings of the Agricultural and Applied Economics Association in August, 2018 and at the Annual Meetings of the Southern Agricultural Economics Association in February, 2019. Policy makers and practitioners were reached as part of presentations at the WasteExpo, North America's largest solid waste, recycling and organics tradeshow serving the private and public sectors, and presentations to regional practitioners groups in the agribusiness and nutrition sectors. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided? Nothing Reported How have the results been disseminated to communities of interest?Communities of interest have been reached via the presentations detailed earlier in the report and through downloads of the February 2018 PLOS ONE article detailed in the 'Target Audience' section of this report. What do you plan to do during the next reporting period to accomplish the goals?Data organization and cleaning for Phase I data were underway at the close of this reporting period with analysis and manuscript preparation slated for the next reporting period. Phase II data, whichhas also been collected, will be organized, cleaned and analyzed in the next reporting period along with manuscript preparation.

Impacts
What was accomplished under these goals? The FoodImage smartphone app was deployed during Phase I and Phase II of the study. Phase I assesses the accuracy, time burden and perceived convenience of measurements taken with the FoodImage app versus the current state-of-the-art household food waste measurement approach (pen-and-paper diary conducted via visual estimation or visual estimation aided with scale). These data were collected in a laboratory setting from subjects tasked with assessing food purchases froma shopping trip and food wasted in standardized food preparation, food consumption and food storage disposal settings. Data organization and cleaning for Phase I data were underway at the close of this reporting period with analysis and manuscript preparation slated for the next reporting period. Phase II of the studied featured deployment of the FoodImage app in a pilot randomized controlled trial that evaluates the effectiveness of reducing household food waste through behavioral nudges in the form of personalized feedback about food waste levels and household-specific reduction goals provided to a randomly selected group of participants who tracked their food purchases, food intake and food waste using the app in their daily lives. The data has been collected and will be organized, cleaned and analyzed in the next reporting period along with manuscript preparation .

Publications

• Type: Conference Papers and Presentations Status: Published Year Published: 2018 Citation: Qi, Danyi, Roe, Brian E., Apolzan, John, Allen, Ray, and Martin, Corby. 2018 The Role of Incidental Learning on Reducing Household Food Waste in Free-Living Condition, AAEA Annual Meetings, Washington, DC, August 6.
• Type: Other Status: Published Year Published: 2018 Citation: Roe, Brian E. 2018. Learning Not to Waste Food: The Power of Attentiveness & Avoiding Mixed Messages, Invited Presentation, WasteExpo, Las Vegas, NV, April 24.
• Type: Other Status: Published Year Published: 2018 Citation: Roe, Brian E. 2018. Food Waste: An Introduction and Overview, Invited Presentation, Northwest Ohio Ag-Business Breakfast Forum, Bowling Green, OH, August 16.
• Type: Other Status: Published Year Published: 2018 Citation: Roe, Brian E. 2018. Learning Not to Waste Food: Lessons from Research & Campus Collaborations, Invited Presentation, Ohio Academy of Nutrition and Dietetics, Columbus Member Meeting, Columbus, OH, October 8.
• Type: Conference Papers and Presentations Status: Published Year Published: 2019 Citation: Roe, Brian E., Qi, Danyi, Apolzan, John, and Martin, Corby. 2019. Selection, Intake, and Plate Waste Patterns of Leftover Food Items from a Pilot Study with U.S. Consumers, Selected Paper, Annual Meetings, Southern Agricultural Economics Association, Birmingham, AL, February 5.

Progress 04/01/17 to 03/31/18

Outputs
Target Audience:The target audiences reached during the year ending 31 March 2018 include academics, policy makers, practitionersin the food and waste sectors, and the general population. Academics were reached via the publication of an article in PLOS ONE on 14 February 2018 that was viewed 445 times by 31 March 2018 either at the PLOS ONE website or via Pub Med. Academics also were reached througha selected presentation at the Annual Meetings of the Agricultural and Applied Economics Association in July, 2017. Policy makers were reached as part of an invitedpresentation at the Convening Event for Food Waste for theFoundation for Food and Agricultural Research (FFAR) in November, 2017. Practitioners in the food, waste and sustainability sectors were reached through presentations by PI Roe at the College & University Recycling Coalition's National Campus Recycling Workshopin October, 2017 and at the Midwest Food Recovery Summit in September, 2017. The general public was reached through news coverage of the PLOS ONE article that appeared in outlets such as Science Daily (www.sciencedaily.com). Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?One graduate student at Ohio State has participated in the development of the research/data collection protocol development and provided feedback on many intermediate versions of the FoodImage app that helped lead to the final app currently posted on Apple's App Store. Further this student presented a paper at a national professional society meeting that uses techniques that will be applicable to the data that will be collected in the upcoming reporting year. How have the results been disseminated to communities of interest?Several presentations have been made throughout the course of the reporting period while one manuscript was published using pilot data that used in the application for this grant. A press release was circulated and interviews with popular press reporters were given after the publication that appeared in PLOS ONE to facilitate sharing with the public. What do you plan to do during the next reporting period to accomplish the goals?The data collection under goals 2 and 3 will be starting during May 2018 and completed during the next reporting period. This will allow for data analysis and writing to commence before the end of the next reporting period.

Impacts
What was accomplished under these goals? 1. The FoodImage app was completed and uploaded to the Apple App Store and is now available for download. 2 and 3. The research and data collection protocols were finalized and submitted to the Pennington Biomedical Research Center's IRB for review during this reporting period, while planning commencedduring the reporting period tobegin scheduling facilities for subject recruitment and data collectionduring the next reporting year.

Publications

• Type: Journal Articles Status: Published Year Published: 2018 Citation: Roe, Brian E., John W. Apolzan, Danyi Qi, H. Raymond Allen, and Corby K. Martin. (2018) Plate waste of adults in the United States measured in free-living conditions, PLOS ONE. 13(2):e0191813.
• Type: Conference Papers and Presentations Status: Published Year Published: 2017 Citation: Qi, Danyi, Martin, Corby K., Apolzan, John W. and Roe, Brian E. Exploring Individual Food Waste Patterns from In-Home Observation of Plate Waste, track session paper, Annual Meetings of the Agricultural and Applied Economics Association, Chicago, IL, August 1, 2017.
• Type: Conference Papers and Presentations Status: Published Year Published: 2017 Citation: Roe, Brian E. Learning Not to Waste: The Power of Attentiveness & Avoiding Mixed Messages, Invited Presentation, Midwest Food Waste Recovery Summit, Des Moines, IA, September 6, 2017.
• Type: Conference Papers and Presentations Status: Published Year Published: 2017 Citation: Roe, Brian E. 2017. Learning Not to Waste Food: Lessons from Research and Campus Collaborations, Keynote presentation, College and University Recycling Coalition Workshop, San Antonio, TX, October 15.
• Type: Conference Papers and Presentations Status: Published Year Published: 2017 Citation: Roe, Brian E. 2017. Identifying Research Gaps Across the Food System: Challenges & Opportunities, Invited Presentation, From Food Waste to Food Security and Beyond: Identifying Research Gaps Across the Food System Convening Event, Foundation for Food and Agricultural Research, Arlington, VA, November 16.