What Causes School Plate Waste? What Audit Data Links
Lunch start time, lunch length and milk format all show up in school audit data. Here is what the studies measured, who was in them, and why none of it proves cause.
What causes plate waste in schools?
Audit data cannot say. It can show what travels together with plate waste, and three things show up in the audit datasets we reviewed: when lunch starts, how long elementary students get to eat, and how milk is served. Each one is an association. Schools that differ on that factor also differ in how much they waste, which is not proof that changing the factor would change the waste.
This post draws on two sets of school audits, a 2024 PLOS ONE analysis and the World Wildlife Fund (WWF) Food Waste Warriors report from 2019, with one USDA national study for contrast. The schools in the audit datasets chose to take part, so the results describe those schools and not schools in general. For the wider numbers on school plate waste, see our school food waste statistics guide.
Who is in the audit data, and how far can you trust it?
The audit schools volunteered to weigh their own waste, so the data is useful but self-selected. If your district looks different from them, treat the numbers as leads to test, not benchmarks.
The PLOS ONE paper, by Adjapong, Bender, Schaefer and Roe, analyzes audits entered by staff and volunteers at 134 schools in 24 states between August 2018 and May 2023, about 490 audits in its lunch-timing figure. Some of the schools were in WWF’s Food Waste Warriors program and some were not. The authors call the sample "a self-selected convenience sample" and write that the data "should not be mistaken for a nationally representative sample." Schools were measured in pounds of waste per student served per meal, and the average was 0.29. The meals served were not weighed, so the paper cannot say what share of the food was wasted.
The WWF report covers 46 schools in nine cities across eight states, audited in spring 2019 with student-led audits run by local partner organizations. WWF writes that its "modest sample size is not representative" of the schools in the National School Lunch Program. It reports waste per student per year, which is a different unit from the PLOS paper’s per meal.
The USDA School Nutrition and Meal Cost Study, which observed 6,253 lunch trays in 165 schools, is the contrast case. The PLOS authors note that USDA’s data collection follows a strict sampling protocol and yields nationally representative estimates. It measured the share of calories left on trays, again a different unit. Keep the three sets of units apart when you compare them.
Is a later lunch period linked to less plate waste?
In these datasets, yes: lunches that start later go with less waste. The PLOS analysis found total waste about 25 percent lower during lunch periods scheduled between noon and 1 p.m. than during those starting before 11 a.m. (an effect on pounds per student, computed from a regression coefficient, so a percent change and not percentage points). The authors also report that waste in the 11 a.m. hour was significantly higher than in the noon hour. For schools with several lunch periods, the start time recorded is the earliest one.
The pattern is not uniform. In the PLOS paper the difference was not significant among older students, and not for produce waste on its own.
USDA’s national study points the same way in different units. The mean percentage of calories wasted was lower in lunch periods that started at 12:00 p.m. or later than in those that started before 11:30 a.m., 18 percent versus 20 percent. That is a small gap in a share of calories, and not comparable to the PLOS percent change in pounds. The PLOS authors also note that earlier work suggested very early or very late lunches may increase waste, and their data confirm the early side.
Is a longer lunch period linked to less plate waste?
Only for elementary students, only in one result, and with a data gap the authors flag. Across all schools, the PLOS paper found that lunch period length "had no significant net association with total waste or any subcomponent." Among elementary students, lunch periods of 30 minutes or more went with about 17 percent less total waste than shorter periods. The authors add that lunch duration was not reported for most audits, which gives them "less confidence in interpreting these results." Among older students the paper found no such link.
Combine the two timing factors and the picture is clearer, if still descriptive. The paper’s figure combining the two suggests that lunch periods that are short (under 30 minutes) and early in the day (before 11 a.m.) had the largest waste per student in the sample.
WWF’s 2019 report reached for the same explanation from the cafeteria floor. An Oregon team wrote that "students don’t have enough time to eat, some students were even eating while walking through the waste/recycling line." WWF put that in its list of possible drivers and added that "More research is needed to examine these possible correlations." Its recommendation was for lunches to be longer, held slightly later in the day, and/or with recess before lunch, "which could increase appetite and decrease wasted food." That is a recommendation, not a measured result. The PLOS paper had no information on whether recess came before or after lunch, so it could not test the recess idea at all.
Is how milk is served linked to milk waste?
Yes, by a wide margin, in both datasets, though neither was designed to prove that the format caused it. In the PLOS paper, schools where students could choose how much milk to take, through a commercial dispenser or by self-pouring from jugs, reported 76 percent less milk waste than schools relying on individual cartons. The paper does not say whether the milk was flavored or unflavored, so that question stays open.
Milk is a big enough share to matter. In the PLOS sample, the 0.29 pounds of waste per student per meal broke down into roughly 0.10 produce, 0.11 other solid food and 0.09 milk, with a small amount of other liquid.
WWF found the same direction. Schools with carton service averaged slightly more than 32 cartons of milk waste per student per year, and schools with dispensers averaged 4.5 cartons. The all-school average in that report was 28.7 cartons. The PLOS data and the WWF report both come out of WWF’s school audit work and may include some of the same schools, so the second should not be read as independent confirmation of the first.
The PLOS authors are careful about what this shows. Without pre- and post-change measurements, they write, "we can observe correlations that may be indicative of possible efficacy and warrant further evaluation in more rigorous data settings." Schools that install dispensers may differ in other ways: more staff time, more organization, a reason they volunteered to audit. A fair reading is that milk format is the strongest lead here, and your own before-and-after weighing is how you would find out whether it holds for your cafeterias.
Why can you not read these associations as causes?
Because the schools were not assigned to different lunch times or milk formats, and the datasets leave out things that matter. The PLOS authors list what is missing: competitive foods such as vending machines or nearby fast food, whether students may leave campus, the order of recess and lunch, and food quality. They also note that the schools "were not randomly selected, but rather represent a group of schools with the motivation and resources to conduct an audit."
Grade level adds a second tangle. Elementary schools had 52 percent more total waste than other schools in the paper’s descriptive results, and WWF also found waste falling as students get older, with a caution that high school data may be incomplete because students often eat outside the cafeteria. Lunch schedules and milk service can differ by grade, so a timing effect and a grade effect can be hard to pull apart.
The offer versus serve result is a useful warning. The PLOS paper found 38 percent more waste in schools using offer, the opposite of what many would expect, driven by middle and high schools. The authors give possible explanations other than offer making students waste more, such as students taking more food overall or schools that waste more being more likely to volunteer. Our offer versus serve guide goes through why the studies disagree. The lesson carries over here: a number from a self-selected audit is a reason to measure, not a reason to change the schedule.
How can a district test lunch timing and milk format in its own schools?
Change one thing at a time, measure before and after on a comparable menu, and write down the conditions. That turns an association from someone else’s schools into evidence about yours.
Record the lunch start time and the actual seat time, not only the period length on the bell schedule. Record the grade levels in each period and, if you can, whether recess comes before or after. Record the milk format, carton, jug or dispenser, and weigh milk separately from solid food so it does not swamp the rest, as in our plate-waste study guide. Use the same menu and the same school before and after a change, because a Tuesday with a popular entrée and a Thursday with a disliked vegetable will move the numbers more than a new schedule.
Treat a single audit week as a snapshot. A schedule change or a dispenser rollout is something to track over several weeks. A tray-return system such as CPIsight can record waste on every service day, which is useful for that kind of comparison, but a careful manual audit before and after works too. Start with the change you can actually make: lunch scheduling and bell times are often set at the school or district level, while milk format depends on equipment, staff and space.
If you are weighing these questions for your own district, our K-12 page describes how we work with school nutrition teams, and our guide to USDA school meal standards and plate waste covers upcoming federal changes you may want a baseline for.
Sources & References
Adjapong, E.S., Bender, K.E., Schaefer, S. and Roe, B.E., "School and meal characteristics associated with plate waste in K-12 cafeterias in the United States." PLOS ONE 19(12): e0299043, December 27, 2024. https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0299043
World Wildlife Fund, "Food Waste Warriors: A deep dive into food waste in US schools." WWF, 2019. https://files.worldwildlife.org/wwfcmsprod/files/Publication/file/21242e75x9_FoodWasteWarriorR_CS_121819.pdf
Fox, M.K. and Gearan, E., "School Nutrition and Meal Cost Study: Summary of Findings." Mathematica Policy Research for USDA Food and Nutrition Service, April 2019. https://fns-prod.azureedge.us/sites/default/files/resource-files/SNMCS_Summary-Findings.pdf
Stay in the loop
Get case studies, pilot updates, and food waste insights.
Keep reading
How K-12 Nutrition Directors Use Waste Data to Justify Budget Requests
School nutrition programs face constant budget pressure. Plate waste data gives directors the evidence they need to make compelling cases for investment.
Read moreK-12Does Offer vs. Serve Reduce Plate Waste? The Evidence
USDA says Offer vs. Serve cuts plate waste. A large audit dataset found more waste at offer middle and high schools. The rules, the studies, what to measure.
Read moreK-12USDA School Meal Standards 2027: Will Students Eat It?
What the USDA rule requires for sodium and added sugars by July 1, 2027, what may change, and how to test reformulated menus for plate waste first.
Read moreTest it on your own trays
CPIsight records plate waste at the tray return, so you can compare the weeks before and after a schedule or milk change on your own menu.