Performing Department
(N/A)
Non Technical Summary
We propose to conduct and evaluate a consumer pilot campaign that supports the development and refinement of educational messages and materials that meaningfully reduce the amount of wasted food created in U.S. households. We will work with three U.S. cities to implement and evaluate a pilot household food waste reduction campaign that builds from materials that are likely to be used in future national consumer food waste education campaigns.The major goals of this proposal are to (1) identify and refine educational messages and materials appropriate for a household food waste reduction campaign that can meaningfully reduce the amount of wasted food created among U.S. households; (2) conduct and assess pilot household food waste reduction campaigns in three cities in the United States, (3) anticipate challenges a national household food waste reduction campaign would encounter during a national roll out, and (4) project the food waste reduction potential and cost-benefit ratio of different national strategies for a household food waste reduction campaign.The evaluation will use two complementary measurement approaches - route-level waste audits and household level suveys - to ensure a robust and valid evaluation of campaign effects. Measurement will occur in both treatment and a carefully selected control group within each city to ensure valid estimation of campaign effects on household food waste generation. The results of the evaluation will inform projections of food waste reductions from different nationwide campaign roll out strategies and cost-benefit analyses of different national campaign roll out strategies. These efforts will also identify likely roll out challenges that may face any regional or national campaigns and provide recommendations for additional testing that could be crucial to ensure success of future large-scale campaigns.
Animal Health Component
50%
Research Effort Categories
Basic
0%
Applied
50%
Developmental
50%
Goals / Objectives
The major goals of this proposal are to:Identify and refine educational messages and materials appropriate for a household food waste reduction campaign that can meaningfully reduce the amount of wasted food created among U.S. households.Conduct and assess pilot household food waste reduction campaigns in three cities in the United States.Anticipate challenges a national household food waste reduction campaign would encounter during a national roll out.Project the food waste reduction potential and cost-benefit ratio of different national strategies for a household food waste reduction campaign.
Project Methods
We propose the following:Task 1. Campaign Formulation. We will engage with experts from government, non-governmental organizations, and the private sector familiar with consumer food waste reduction campaigns. If materials have been identified by federal agencies for use in future national campaigns, the pilot campaigns will focus on fine tuning these materials. If such materials have not been identified, then we will discuss the relative merits of adopting existing campaigns (e.g., Oregon's 'Don't Let Good Food Go Bad!', Ohio's 'Save More Than Food') for use in the pilot with selection based on: evidence of campaign efficacy; the adaptability of the campaign to pilot cities; inclusiveness and accessibility of campaign materials; and expert assessment of potential impact. The exact materials to be shared and the communications channels will be determined in cooperation with officials from the selected pilot cities to ensure local suitability and to improve efficacy.Task 2. Pilot City Selection. We will develop a geographically dispersed list of possible pilot cities to ensure that the campaign reaches areas with substantial populations enrolled in USDA FNS programs. We will work with USDA and US EPA to select communities based on characteristics such as climactic conditions, socioeconomic profile, food system infrastructure, composting prevalence, past exposure to campaigns, among others. We will approach city officials to seek cooperation in conducting a pilot. Funds are budgeted to offset cities' personnel time spent supporting the pilot.Task 3. Baseline Assessment. The assessment unit is the municipal waste hauler's residential collection route. 84 residential waste collection routes suitable for randomization (half treatment, half control) will be identified in each city, permitting attrition of ~5% while still meeting the target of 80 routes to ensure the power needed to identify a 20% food waste reduction effect of a campaign.We assume routes have 500 residences, implying data collection representing ~42,000 residences per city. A contractor intercepts garbage trucks after route completion, collects a randomly selected sample of waste (~200 pounds), transports it to a designated sorting location, and sorts it into 10 comprehensive, mutually-exclusive categories: produce, pasta/bread/rice/cereal/beans, meat and fish, soups/sauces, potato, snacks, dairy and eggs, beverages, inedible (i.e., food never intended for consumption, like bones, pits and peels), and all other (non-food and non-beverage) items. The fraction of the sample by weight as assessed by electronic scale in each category provides baseline measurement.Because this form of assessment does not assess the absolute level of food waste generated on a per household basis does not measure wasted food that is fed to animals, discarded down the drain or composting, we will also recruit households from each route to complete surveys. We will use the OSU/RECIPES National Household Food Waste Tracking Survey along with correction factors being generated by Roe's FFAR/USDA-OCE project to generate estimates of the level of food wasted by route households. We recognize these data will be affected by self-selection bias and plan to use post-hoc econometric methods to adjust for such biases.This survey will be customized to permit assessment of household member awareness of food waste topics and reduction approaches prior to the circulation of the campaign materials. Depending on the city and its communications infrastructure, recruitment will occur via mailers to addresses on the selected routes or via emails and texts to household contacts on the control and treatment routes. In either case, the communications will invite participants to connect with a link to complete the survey and highlight participation incentives. Use of the Tracking Survey will also permit us to compare the level of self-reported waste in pilot cities against contemporaneous self-reported household food waste as captured by ongoing OSU/RECIPES National Household Food Waste Tracking survey, which uses the same instrument. This will permit assessment of whether the waste amounts and patterns among households in participating cities are distinct from a comparable sample drawn from elsewhere in the United States.Task 4. Campaign Implementation. We will create 42 route pairs from each city's 84 routes where each pair features similar characteristics as determined through the use of Census data and expert opinion of city personnel. One route in each pair is randomly assigned to treatment. Treatment residences receive campaign materials via mailings delivered via the U.S. postal service and, where possible, via email or text message. Delivery of campaign materials via mass media will be avoided to ensure integrity of the control group. Task 5. Follow Up Assessment.Follow up assessments will mirror the baseline measurements and be repeated twice: 1-2 months and 6-8 months post campaign. Questions will be added to the survey to assess residents' awareness of and favorability towards campaign materials. City officials will be canvased for feedback concerning campaign roll out and impacts.Task 6. Analysis. A difference in differenceanalysis will be conducted in which the differences between the pre- and post-intervention measures across the treatment routes will be contrasted to the pre/post differences among the control routes. An example regression equation would be:Yi= b0+ b1*Treatmenti+ b2*Posti+ b3*Treatmenti*Posti+ b4*Xi+ ei,where Yiis an outcome variable for uniti(e.g., fraction of routei's sample that is wasted food), Treatmenti equals one if unitiwas in the treatment group (e.g., got campaign materials) and zero otherwise, Postiequals one if the observation occurred after the campaign initiated and zero otherwise, Xidenotes other characteristics of the unit (e.g., route-level income and housing type), and eirepresents the remaining factors that influence the outcome but are not measured. Regression analysis yields estimates of coefficients (b0- b4) with b3the coefficient that indicates the average effect of the treatment on the treated.Outcome variables to be assessed include all-category wasted food (as a fraction of the total solid waste sample), food scraps (inedible food waste as a fraction of the total solid waste sample), wasted food by category (each category as a fraction of the total solid waste sample), surveyed food waste amounts (in levels for total wasted food and waste by food category), awareness of food waste as a topic, awareness of the specific food waste reduction campaign, and self-assessed food waste reduction effort.Analyses within particular groups (city, socio-economic strata, dominant housing form) will be conducted. The estimates will help predict reductions in food waste associated with different national campaign roll out strategies. Projections will be based upon any heterogeneity identified among the pilot cities using differences in demographics across regions as documented by Census. The approach provides total reduction in food waste possible from implementing campaigns across the entire United States. Optimization exercises will identify the greatest reduction in food waste for different levels of national campaign funding and the cost-benefit ratio achievable for different funding levels. The optimization approach will be sufficiently flexible to permit restrictions, e.g., minimum proportions of total campaign funds expended in different geographic regions or to prioritize waste reduction among targeted demographic segments.