School Food Waste Statistics: What the Data Shows

These are the figures on US school plate waste that are worth quoting, each with what it actually measured. The sections below explain the methods and the caveats.

About 21 percent of the calories in school lunches went uneaten, according to USDA’s School Nutrition and Meal Cost Study, which observed 6,253 lunch trays in 165 schools during school year 2014–15.

In that same USDA study, vegetables were the most-wasted lunch component (31 percent of the vegetables on observed trays), followed by milk (29 percent), fruits and 100% fruit juice (26 percent), and side grains and breads (23 percent).

In World Wildlife Fund’s 2019 Food Waste Warriors audits of 46 schools, cafeteria plate waste averaged about 39.2 pounds of food per student per year, with milk tracked separately at about 28.7 carton-equivalents per student per year.

Scaling up from those 46 schools, WWF estimated that food waste in US schools, excluding milk, could be as much as 530,000 tons a year, while noting that its sample is not statistically representative of schools nationally.

A 2024 PLOS ONE analysis of audits from 134 schools in 24 states, submitted from 2018 to 2023 through a WWF-run audit database that includes Food Waste Warriors schools, found an average of 0.29 pounds of plate and beverage waste per student per meal.

The previous best national estimate, a 2002 USDA Economic Research Service review, put uneaten lunch calories at about 12 percent, based on data collected in 1991–92.

Short answer: the best national data, from school year 2014–15, found that roughly one-fifth of the calories served in school lunches went uneaten. There is no current national figure in pounds or tons; the pound figures come from audit programs that are useful but not representative.

The national number comes from the USDA School Nutrition and Meal Cost Study (SNMCS), published in April 2019. Trained field staff recorded what students took and how much of it they left, across 6,253 lunch trays in 165 schools. USDA describes it as the first national study in more than two decades to examine plate waste in school meals. Beyond calories, students left one-quarter or more of the vitamin A, vitamin C, vitamin D, calcium and potassium available in their lunches, which is the number to bring to a conversation about nutrition rather than cost. The data are from school year 2014–15; these are the most recent national plate-waste observations we found.

The pound figures come from World Wildlife Fund’s Food Waste Warriors report. Students and staff at 46 schools in nine cities across eight states ran cafeteria audits from January to June of the 2018–19 school year, 222 audit days in all, and the schools averaged about 39.2 pounds of food waste per student per year. From that sample WWF extrapolated the national estimate of up to 530,000 tons a year, and estimated that school food waste could be costing as much as $1.7B every school year. Both are projections from 46 schools, and WWF says plainly that its cohort is not statistically representative of the roughly 100,000 schools in the National School Lunch Program. Quote them as estimates, not measurements.

The PLOS ONE study by Adjapong and colleagues analyzed nearly 500 audit observations from 134 schools that uploaded audit data to a WWF-organized database between August 2018 and May 2023 (some were Food Waste Warriors schools, others were not), and found an average of 0.29 pounds of plate and beverage waste per student per meal.

Vegetables and milk. In the USDA study, the share of each component left uneaten was highest for vegetables (31 percent) and milk (29 percent), then fruits and 100% fruit juice (26 percent), side grains and breads (23 percent), and desserts and other items (20 percent). Entrées and meats or meat alternates were wasted least, at 16 and 14 percent. USDA notes that these levels are generally comparable to studies done before the updated nutrition standards took effect.

The WWF audits sorted waste by weight rather than by share of each item. Across all audits, organic waste totaled 19,074 pounds, made up of 32% fruit and vegetable waste, 26% milk waste and 42% other food waste. Weight shares favor heavy items, so they are not directly comparable to USDA’s percentages.

Milk is worth its own line on your report. WWF found an average of about 28.7 carton-equivalents of milk wasted per student per year across its 46 schools, with elementary schools higher (37.6) and middle schools lower (19.4). WWF converted milk to carton-equivalents for comparison regardless of how it was served. In the PLOS ONE data, the 0.29 pounds per meal split roughly evenly among produce (0.10 pounds), other food such as entrées and sides (0.11 pounds) and milk (0.09 pounds), with a small amount of other liquids.

Grade level, lunch timing and milk service show up most consistently, but every finding here is an association from observational data, not proof of cause. Elementary schools waste more: USDA found a higher share of each food wasted in elementary schools than in middle or high schools, and PLOS ONE found 52% more total waste in elementary schools than in other schools. Later lunches go with less waste: USDA found 18 percent of calories wasted in lunch periods starting at noon or later, versus 20 percent before 11:30 a.m. How milk is served (cartons versus dispensers) also shows up in the audit data. WWF found no strong statistical correlation between food waste and district income or free and reduced-price lunch rates.

Offer versus serve is where the studies disagree. USDA found that, among elementary schools, using offer versus serve was associated with significantly lower plate waste. PLOS ONE found 38% more waste in “offer” schools overall, but the difference came from middle and high schools; elementary schools had nearly identical waste either way. WWF’s 2019 audits, which overlap with the PLOS ONE data, pointed the same way, but only 6 of its schools used “serve,” and the difference was not statistically significant. We cover the evidence in more detail in our offer versus serve guide. For how upcoming changes to federal meal standards could show up on the tray, see USDA school meal standards and plate waste.

The three main sources measure different things, so their numbers cannot be stacked side by side. USDA’s study observed individual trays and reports the share of each item left uneaten. The WWF and PLOS ONE audits weigh the plate and beverage waste collected in the cafeteria during an audit and divide by the number of students served; they do not weigh the meals served, so they cannot say what percentage of the food was wasted. The 2002 ERS figure is a share of calories from a 1991–92 national study, and ERS itself warned that it may not reflect current conditions in schools. Do not compare 39.2 pounds a year with 0.29 pounds a meal or with 21 percent of calories as if they were the same measurement.

Know the sample before you quote it. USDA’s plate-waste findings represent public, non-charter schools that serve meals in a cafeteria setting and serve at least a minimum number of lunches a day. The PLOS ONE authors describe their data as a self-selected convenience sample of schools that chose to audit, covering only about half of US states, and say it should not be mistaken for a nationally representative sample. Lunch period length was not reported for the majority of their observations, and the authors caution against drawing causal inferences.

The WWF 2019 report and the PLOS ONE paper are not two independent samples. Both come from WWF-organized cafeteria audits, the PLOS ONE data include Food Waste Warriors schools, and the collection periods overlap, so treat them as one body of audit evidence, not two studies that confirm each other. For the cross-sector picture beyond schools, see our institutional dining food waste statistics.

Measure your own trays. National averages tell you the problem is real, but they cannot tell you which item on Tuesday’s menu is coming back to the dish room, and your district can sit well above or below any of the figures above.

Pick the metric to match the decision. If you are adjusting menus, you need waste by component, ideally as a share of what was served, like the USDA study. If you are tracking a trend or reporting progress, pounds per student served is enough, as long as you count students the same way every time. Either way, separate milk, record the number of students served, note when each lunch period starts, and measure across several days rather than one.

A one-time audit gives you a baseline; ongoing tracking tells you whether a change worked. We compare manual weighing, scales and camera-based systems in food waste tracking methods compared. For an example of a district measuring plate waste at its tray return, read the Seneca Valley School District case study, and see how we work with school nutrition programs on our K-12 page.