Not Losing Weight in a Calorie Deficit? Check the Log First

By Eric Parker, founder of TrueCal. Published October 10, 2026. Sources checked October 10, 2026.

Written with AI assistance from the cited studies. Every claim was checked against its source. Editorial policy.

Key takeaways

When the scale doesn't follow a calorie deficit, the food record is the first thing worth checking, because studies that measured intake directly found records come in low. In Lichtman's 1992 study of people who weren't losing weight on diets they reported as low in calories, measured energy use was within 5 percent of predicted, and the gap was in the records.¹
This happens to trained people too. In Champagne's 2002 study, registered dietitians who kept 7 day weighed food records came in an average of 223 calories a day under their measured energy use, and women who weren't dietitians came in 429 under.⁴
Restaurant numbers vary item by item. Urban's 2011 lab tests of 269 restaurant foods found the stated calories right on average, yet 19 percent of items held at least 100 calories more than stated.⁵
USDA's national diet interview goes through each day more than once. After the first list of foods, it asks about forgotten foods, then the amount and additions for each food, then anything else. The same passes work on your own log.⁷⁶
Check the scale side too. Compare weekly averages over several weeks rather than one morning against another.

Why am I not losing weight in a calorie deficit?

The deficit you see is the difference between two estimates, what the log says you ate and what a calculator says you burn. Studies that measured both directly found the food side of that math tends to run low, so a deficit on paper can be smaller in real life.¹²

The best known test of this is Lichtman's 1992 study. The researchers looked at people who said their diets weren't working and asked whether their bodies burned less than expected or their records missed food. Total energy use and resting metabolic rate came out within 5 percent of predicted, and there was no meaningful difference from a comparison group in the energy used to digest food or exercise. The difference was in the records.¹

That finding is about food records, not about effort or honesty. Records came in low across the studies below, including records kept by people trained to keep them.²⁴³

How accurate is calorie counting?

The reference method is doubly labeled water, a lab test that measures how much energy a person used over a stretch of days. Schoeller's 1995 review of studies that checked food records against it found that people's individual underestimates of 50 percent were not uncommon in several groups. Even in groups where the average bias was small, the error for one person had a standard deviation approaching 20 percent.²

The OPEN study, published by Subar and colleagues in 2003, checked 484 adults the same way. On 24 hour recalls, where an interviewer walks through the previous day, men reported 12 to 14 percent less energy than they used and women 16 to 20 percent less. On food frequency questionnaires, which ask how often you eat things, the gap was 31 to 36 percent for men and 34 to 38 percent for women.³

So calorie counting is useful and it's imperfect in a known direction. In the OPEN study, a detailed walk through one day came much closer to measured intake than questions about usual eating.³²

Do food records run low even for dietitians?

Champagne and colleagues tested this in 2002 at the Pennington Biomedical Research Center. Ten registered dietitians and ten women of similar age and weight who weren't dietitians each kept a 7 day weighed food record while doubly labeled water measured their energy use over the same days.⁴

Everyone was told the goal was to record food as accurately as possible, and that the records would be compared with the lab measurement. The dietitians still came in 223 calories a day under on average, a gap small enough that it wasn't statistically different from what they burned. The other women came in 429 calories a day under, a significant gap.⁴

The authors read this as experience helping. They concluded that familiarity with and interest in keeping food records may lead to more reliable estimates, which suggests practice with a log counts for something.⁴

How accurate are restaurant calorie counts?

Food bought at restaurants supplies about 35 percent of daily energy intake for people in the US, according to Urban and colleagues. In 2011 they measured 269 items from 42 quick serve and sit down restaurants in a lab and compared the results with the stated calories.⁵

On average the stated numbers held up, off by 10 calories a portion. Single items were another story. Fifty of the 269 items, 19 percent, held at least 100 calories more than stated, and items with lower stated calories tended to hold more than stated.⁵

So a restaurant's stated number is right on average, but any single item can be well off. The trend across many days says more than any one meal out.⁵

Is the activity estimate part of the gap?

It can be. The people in Lichtman's study overreported their physical activity by 51 percent on average, so both sides of their math leaned the same way, intake looked lower and burn looked higher.¹

If you set your target with a calorie deficit calculator, check which activity level you picked. Choosing the level that matches a normal week, rather than your most active week, keeps the burn side of the estimate closer to real.

How to check if your food log is complete

USDA's Automated Multiple Pass Method is the interview behind What We Eat in America, the diet part of the national NHANES survey. It's built on the idea that one pass through a day misses things, so it goes through the day several times with a different question each time.⁷⁶

You can run the same passes on your own log at the end of the day, or the next morning.

Quick list. Write down every meal and drink you remember, without amounts yet.
Forgotten foods. Walk through the day by place and time, the drive, the desk, the kitchen while cooking, the evening, and add anything that wasn't a meal.
Detail. For each item, add the amount and what went on it or in it, the spread, the dressing, the oil in the pan, the milk in the coffee.
Final check. Ask once more whether anything else was eaten or drunk, and add it.

USDA step	What it asks for	In the example day
Quick List	Foods and drinks from the previous day	Oatmeal, turkey sandwich, pasta⁷
Forgotten Foods	Foods forgotten during the quick list	Latte, granola bar, two cookies⁷
Detail Cycle	Description, amount and additions for each food	Brown sugar, milk, mayonnaise, olive oil, parmesan⁷
Final Probe	Anything else consumed	Bites while cooking, glass of wine⁷

Is the scale telling the whole story?

The other half of the check is the weight data. Weigh under the same conditions each time, write every reading down, including the high ones, and compare weekly averages across several weeks. Day to day swings are covered in the water weight guide linked below.

If the log has been complete for a while and the trend still sits flat, the weight loss plateau page linked below covers how the body adjusts over a long diet. If your log is complete and your weight trend still isn't what you'd expect, a clinician is the person to ask about medical causes.

Worked example

Here's one made up day of a food record, run through the passes of USDA's interview to show what each pass can add. The quick list says oatmeal, a turkey sandwich and pasta with tomato sauce. The forgotten foods pass adds a latte on the way to work, a granola bar in the afternoon and two cookies from a meeting. The detail pass asks for amounts and additions, which brings in brown sugar and milk on the oatmeal, mayonnaise on the sandwich, olive oil in the pasta pan and parmesan on top. The final probe adds a few bites of pasta while cooking and a glass of wine with dinner. In this example day, the first list had three entries and the finished record has thirteen, with each pass adding a few.⁷

A closer look

The top results list medical and hormonal reasons, and none of them cite a study that measured what people actually ate. Lichtman's 1992 study in the New England Journal of Medicine did. It took 10 people with a history of not losing weight on diets they reported as low in calories, measured their energy use and their actual intake for 14 days, and found their energy use within 5 percent of predicted. Their records showed intake 47 percent below what they ate, on average, and they overreported their physical activity by 51 percent. The authors concluded the gap came from intake higher than reported and activity lower than estimated, not from an abnormal metabolism.¹

Questions and answers

Is it my metabolism?

In Lichtman's study of people who believed their bodies resisted dieting, measured energy use was within 5 percent of predicted and the gap was in intake and activity records. For changes over a long diet, see the plateau page linked below.¹

Does counting calories still work if records run low?

A record is still the closest view most people have of what they eat. The studies show where it drifts, so the useful move is to make it complete, with amounts and additions, rather than to drop it.³⁴

Are app and AI calorie estimates accurate?

They're estimates. TrueCal's own fact sheet says nutrition results can be wrong and depend on the meal details given, including portions and ingredients, so a description with amounts and additions gets a better number than a vague one.⁸

Do I need to weigh my food?

Not necessarily. The food scale guide linked below covers when weighing helps, and the portion control guide covers estimating by eye.

Add the small things in one line

Log calories by hand for free. With Pro, type what you ate in plain words, like a splash of milk in my coffee or a tablespoon of oil in the pan, and TrueCal turns it into calories and macros. You can edit a portion later if it's off.

Get TrueCal on the App Store Get TrueCal on Google Play

Sources

New England Journal of Medicine. Lichtman, Pisarska, Berman, Pestone, Dowling, Offenbacher, Weisel, Heshka, Matthews, Heymsfield. 1992.
Metabolism. Schoeller. 1995.
American Journal of Epidemiology. Subar, Kipnis, Troiano, Midthune, Schoeller, Bingham, Sharbaugh, Trabulsi, Runswick, Ballard-Barbash, Sunshine, Schatzkin. 2003.
Journal of the American Dietetic Association. Champagne, Bray, Kurtz, Monteiro, Tucker, Volaufova, Delany. 2002.
JAMA. Urban, McCrory, Dallal, Das, Saltzman, Weber, Roberts. 2011.
American Journal of Clinical Nutrition. Moshfegh, Rhodes, Baer, Murayi, Clemens, Rumpler, Paul, Sebastian, Kuczynski, Ingwersen, Staples, Cleveland. 2008.
USDA Agricultural Research Service, AMPM Features. U.S. Department of Agriculture, Agricultural Research Service. 2024.
TrueCal official facts. TrueCal. 2026.