The Calorie Number on the Menu Board Is Legally Allowed to Be Wrong

Since May 2018, the FDA has required chains with 20 or more locations to post calorie counts on menus and menu boards. The rule is one of the most visible pieces of food regulation in America. It is also more approximate than almost anyone realizes.

The tolerance is 20%

The FDA allows up to 20% error on nutrition labeling. That is not a loophole someone found: it is the compliance standard, and it exists because food is a biological product assembled by humans. Two chicken breasts are not identical. A sauce ladle is not a pipette.

The practical consequence: a posted 600-calorie entrée can legally be a 720-calorie entrée. On a 2,000-calorie day, that single item's allowance covers a sixth of your daily intake before anything has gone wrong.

What measurement found

Researchers analyzed over 250 items from more than 40 restaurants against laboratory measurement:

• 19% of individual samples differed from lab values by more than 100 calories.
• 15% of items exceeded their declared count by more than 20%, outside the tolerance, not just at the edge of it.
• A JAMA analysis found restaurant counts understate actual content by about 18% on average.

The error is not evenly distributed. Fast-food chains land closest, typically 7-15% off, because the whole operation is built around identical assembly: portioned patties, timed fryers, pumped sauces. Sit-down restaurants run 15-30% off, and smaller independent kitchens with hand-portioned components can be further still.

The cruel part: "light" items are the worst offenders

The most useful finding in this literature is also the most annoying. Items marketed as lighter options showed higher caloric variance than their indulgent counterparts.

The mechanism is not conspiracy, it is arithmetic. A 1,200-calorie burger is mostly fixed mass, patty, bun, cheese, all portioned. A 400-calorie salad is mostly variable mass: dressing poured by hand, cheese sprinkled by feel, croutons by the scoop, oil in the pan. Every one of those is a chef-variable component, and every chef-variable component is where the error lives. A heavy dressing pour can add 200 calories to a dish whose entire posted value is 400.

So the item you order when you are being careful is precisely the one whose number you should trust least.

How to use the numbers anyway

Posted calories are still worth reading: they are just a comparison tool, not a measurement.

1. Compare within a chain, not across chains. Two items on the same menu were counted by the same methodology, so the ranking between them is reliable even if both absolute numbers drift.
2. Trust fixed-assembly items more. Burgers, nuggets, and packaged sides are closer to their posted numbers than anything involving a ladle.
3. Assume the sauce is uncounted-ish. If a dish's calories are dominated by a poured or drizzled component, treat the posted figure as a floor.
4. Round up on the light stuff. Adding 20% to a salad's posted count gets you closer to reality than the menu board does.

If you want the ranking rather than the absolute number, the per-chain calorie breakdowns are the fastest way to see how items on the same menu stack against each other, which is the only comparison the data actually supports.