How to Run a Plate-Waste Study in Your School Cafeteria

A plate-waste study measures what students leave on their trays, usually broken out by food item or meal component. A food waste audit is broader and can include kitchen prep waste, overproduction and packaging. This guide covers student trays only. For a kitchen-wide audit, see how to conduct a food waste audit.

Plate-waste studies come in two designs. In a tray-level design, you know what each sampled student took and measure what is left of each item, so you can report the share of each food that went uneaten. In an aggregate design, you collect leftovers from many students, sort them by food category, and divide by the number of trays. USDA’s Economic Research Service (ERS) describes both in a 2002 report to Congress. The free student audit guides from EPA and World Wildlife Fund use the sorted aggregate design: leftovers go into labeled buckets and volunteers weigh them, and the EPA guide also has volunteers count the trays. USDA’s national School Nutrition and Meal Cost Study was tray-level: field interviewers recorded the foods on each tagged tray and the amount left.

A note for Rhode Island districts: as of November 2026, the state’s amended school waste law (H7342 Substitute A, signed in June 2026) requires schools to report their waste audit results to the state Department of Education, which publishes them. Reporting is currently scheduled to start January 1, 2027 and to recur every three years. The law defines a waste audit as an analysis of a facility’s waste stream: what recyclables and waste it generates, how much of each category is recovered or discarded, and what can be composted. That is not a plate-waste study.

Neither the EPA guide nor the WWF guide sets a minimum number of trays or days. Plan for coverage instead: every lunch period and serving line, trays chosen at random, and more than one menu day.

Lunch periods matter. USDA’s study spread its plate-waste observations evenly over all meal periods and serving lines, and it found that lunch periods starting at noon or later had a lower share of calories wasted than those starting before 11:30 a.m.

Menu days matter too. One USDA-funded study summarized by ERS ran its school observations over five consecutive days, in 1991–1992. The EPA guide recommends scouting on a day with the same menu as the planned audit day, and running follow-up audits on days with the same or a similar menu so the same items can be compared. Pick the days you want answers about, such as the entrée you are thinking of dropping.

For tray-level work, sample at random. ERS describes weighing studies that use a randomly selected set of trays, and the University of Washington study described below picked cafeteria tables by random cluster sample. For a bucket audit of every tray, people are the constraint: the EPA guide recommends at least four volunteers plus a team leader, ideally four volunteers for every 100 students eating, and one audit station for every 200 students.

Sort and record waste by the components the meal is planned around. As of November 2026, the federal lunch rules require schools to offer five components daily: fruits, vegetables, grains, meats or meat alternates, and fluid milk. Under offer versus serve, students may decline two components at lunch but must take at least half a cup of fruit or vegetable. Measuring along those lines shows which part of the reimbursable meal is coming back.

Choose buckets before lunch. WWF’s guide uses five: fruit, vegetable, milk, entrée and other. That is quick, but there is no grains bucket, and a pizza slice counts as entrée rather than grain plus meat alternate. The EPA guide suggests a bucket for each entrée option, notes that a high school may need 12 or more buckets, and says more categories make the results more useful.

Weigh one serving of each item, and count how many went out. WWF’s guide has volunteers ask the cafeteria manager for permission to weigh one unpackaged serving of each food before lunch. For a bucket audit you also need the number of servings of each item that went out, from the production record or a tally, because under offer versus serve or with two entrée choices not every tray carries every item. Divide the bucket weight by servings out times serving weight to get the share wasted, a bucket-level version of the per-tray formula ERS describes. Without those two numbers, you only have pounds.

Handle milk separately. The EPA guide warns that a half-full bucket of liquid can weigh up to 25 pounds and says to weigh liquid buckets at half full. In the University of Washington study, milk cartons were hard to judge by eye, so assessors poured what was left into a measuring cup and recorded it to the nearest quarter-cup.

Set unopened items aside, as both guides do, so you can see what could go to a share table or donation.

Each method trades accuracy against time, space and people. Weighing is the reference standard. Trained visual estimation is more convenient and, in one school validation study, agreed closely with the scale. Photos let you code trays later.

Weighing. ERS describes weighing the edible food on a randomly selected set of trays before the meal (or using a typical serving weight), then weighing what is left of each item. Its advantage is detailed, accurate information. ERS lists the drawbacks as cost, time, space to hold trays until they are weighed, and impracticality for samples of more than 50 to 100 children, a limit it attributes to a 1980 study by Comstock and colleagues. Bucket weighing is simpler but gives only category totals.

Visual estimation. Trained observers rate how much of each item is left on a fixed scale. ERS lists the advantages as saving space and arguably time, being cleaner, and possibly needing fewer people; the main disadvantage is that ratings are not exact proportions. The common school version is the quarter-waste method: none, a quarter, half, three-quarters or all of the item left. A University of Washington team tested it against weighed waste on 748 trays in four middle schools and five high schools in one district in May 2014. For 87 percent of foods, agreement with the measured weight was substantial or almost perfect. Weaker results came from items that were hard to isolate or define, such as rice served with a sauced entrée, or salads where only dressing and croutons were left. Observers needed only two hours of training, and the authors call the method minimally disruptive to cafeteria flow. The limits: one district, middle and high schools only, one visit per school.

Photographs. Staff photograph each tray before and after the meal and code the photos later. A UK validation study in one primary and one secondary school found that estimates from the photos came close to weighed measurements, and its authors note that photo records can be examined by more than one independent coder. Someone still has to code every photo, and the study tested trained people reading photos, not any automated system.

A fourth option, asking students how much they ate, is less expensive than weighing, but ERS notes it relies on children’s recall and is subject to bias. Use student input for reasons, not amounts.

Use one sheet per lunch period. The EPA guide’s sample weight log records the empty bucket weight, the number of trays and a weight for each food type; its interview sheet records the location, date and lunch period, each unfinished item and the student’s reason. A fuller sheet has three parts.

Header: school, date, lunch period and start time, serving line, menu, method (bucket, tray-level visual or photo), number of trays in the sample, and who recorded it.

Reference numbers: each menu item, the component it counts toward, the weight of one unpackaged serving, the number of servings that went out, and the empty-bucket weight to subtract.

Results: for a bucket audit, one line per weighing with food type and gross weight, plus a count of unopened items. For a tray-level study, one row per tray with each component scored on your scale. If you talk to students, add a reasons column. The EPA guide asks for one main reason per item and says answers like “didn’t like it” are not specific enough, so ask a follow-up such as what about the taste they didn’t like.

Leave student names off. You don’t need them to answer a menu question.

Start with permission and transparency, and collect less than you think you need. This is not legal advice, and your district’s policies come first.

WWF’s guide makes getting permission from the school administration the first step, along with informing cafeteria staff, and includes a sample lunchroom announcement telling students the school is collecting information about the food in the cafeteria. The EPA guide has interviewers explain that they are taking a survey before asking students about their leftovers.

Research studies go further. The University of Washington study’s protocols were approved by the university’s institutional review board, and each student gave verbal consent. The UK photo study was approved by a university ethics board and obtained written parental consent. If you partner with a university or plan to publish, expect that kind of review.

For a district-run study: record trays, not names. If you take photos, frame the tray rather than the student, and check your district’s photo and media policy first. Tell families and staff what you are measuring and why, and decide in advance who will see the data and how long you will keep it.

Turn raw weights into a few comparable numbers, then look for the items that come back most and the reasons students give. Calculate the share wasted for each item if you have serving weights and servings counts, or pounds per tray if you don’t, and keep milk on its own line so it doesn’t swamp the solid food. Compare lunch periods and grade levels within your own schools before comparing with anyone else; national figures measure different things in different ways, and our school food waste statistics guide explains which ones can be compared.

Then act on the specific item. If one vegetable keeps coming back, look at recipe, presentation and portion before dropping it. If whole components are taken and thrown away, look at how offer versus serve is working on your line; our offer versus serve guide covers what the studies found by grade level. If waste is highest in the earliest lunch, look at the schedule. WWF’s guide lists share tables, offer versus serve, recess before lunch and bulk milk dispensers among its next steps. Whatever you change, measure again on the same menu once the change is fully in place, as the EPA guide recommends.

A one-off study is enough to set a baseline or answer one well-defined question, such as whether students eat the new entrée. It falls short when you need to know whether a change held, how waste moves across a menu cycle, or how schools compare, because a few days of data cannot separate a real change from an unusual day. The EPA guide encourages multiple audits in a school year, but manual studies take volunteers, space and time, so they are hard to run often. The alternative is continuous measurement at the tray return, by scale or camera, which trades setup cost for data from every service day. We compare the options in food waste tracking methods compared.

Run a manual study first so you know what you are looking for. If its questions need weekly or daily answers, look at continuous tracking. CPIsight is one such system, a camera and scale at the tray return. For an example, read the Seneca Valley School District case study, and see how we work with school nutrition teams on our K-12 page.