Guides
Mifflin-St Jeor vs. Harris-Benedict: Which Equation Should You Trust?
Two calculators, two different calorie targets — because two different equations. Where each formula comes from, why they disagree, and which to use.
By Daniel Brandt · · 10 min read
Run your numbers through two calorie calculators and you'll often get two answers, sometimes 100–200 calories apart. That's rarely a bug. It's usually two different equations under the hood — most often Mifflin-St Jeor versus some version of Harris-Benedict — and knowing why they disagree tells you which number to trust.
Short version: use Mifflin-St Jeor unless you have a measured body-fat percentage, in which case Katch-McArdle deserves a look. The long version, with the actual formulas and the evidence, follows.
Harris-Benedict: the 1919 original
James Harris and Francis Benedict published their equations in 1919, from calorimetry on 239 subjects. For their era they were remarkable, and they anchored clinical nutrition for most of a century. But the sample was small, skewed young and lean, and later validation kept finding the same bias: overestimation of resting metabolism, typically by around 5%, worst in heavier and older people.
A 1984 revision by Roza and Shizgal re-fit the coefficients on more data (that's the 'revised Harris-Benedict' our calculator offers), improving things without fully fixing the tendency to run high. When a calculator quietly hands you a generous calorie budget, an original-flavor Harris-Benedict implementation is a common reason.
Mifflin-St Jeor: the 1990 correction
In 1990, Mifflin and St Jeor derived new coefficients from about 500 adults spanning normal weight to obese — a sample that looked far more like the modern population. The structure is elegantly simple: BMR = 10 × weight(kg) + 6.25 × height(cm) − 5 × age, plus 5 for men or minus 161 for women.
The decisive endorsement came from the American Dietetic Association's systematic comparison, which found Mifflin-St Jeor predicted measured resting metabolic rate within 10% for more adults than any other equation — the closest thing metabolism math has to an official verdict. It has been the default in serious calculators (including ours) ever since.
Katch-McArdle: for when you know your body fat
Both equations above infer your metabolically active tissue from weight, height, age, and sex. Katch-McArdle skips the inference: BMR = 370 + 21.6 × lean mass(kg). Feed it an accurate body-fat percentage and it handles the physiques that break weight-based formulas — the 95 kg lifter at 12% fat (underestimated by Mifflin) and the very lean or very high-body-fat cases in both directions.
The catch is input quality. Consumer body-fat measurements (home scales, handheld devices) can be off by 5+ percentage points, and garbage in beats any equation's precision. Use Katch-McArdle with a DEXA scan or a consistent skinfold protocol; otherwise Mifflin-St Jeor's assumptions are the safer bet.
The verdict, and why it matters less than you think
Our recommendation mirrors the evidence: Mifflin-St Jeor by default; Katch-McArdle if you genuinely know your lean mass; Harris-Benedict for continuity with older plans or clinical charts. The typical disagreement between them — 50 to 150 kcal — sounds decisive but is smaller than two other error sources everyone carries: the activity multiplier (picking 'moderate' over 'light' swings ~300 kcal) and food-logging error (routinely 20% and up).
So pick the best equation, then spend your perfectionism where it pays: an honest activity level, and two weeks of scale data to calibrate the estimate into your number. Every equation is a starting point; your logbook is the finish line.
Quick answers
Mifflin-St Jeor, for most healthy adults. The American Dietetic Association's systematic review found it predicted measured resting metabolic rate within 10% more often than any competitor, while Harris-Benedict tends to overestimate, especially in heavier and older individuals.
Three reasons, in rising order of impact: different BMR equations (±50–150 kcal), different activity-multiplier definitions (±200–400 kcal), and different goal adjustments. Any two well-built calculators land in the same neighborhood; the exact house number always comes from testing your target against real scale data.