Body fat percentage: scans, scales, calipers and the Navy method

Body fat percentage is the share of your weight that is fat, and every common way to measure it is an estimate. For one person, smart scales, calipers, the Navy tape method and photo apps are often several points off a lab reference.^{[6, 12, 16, 21]} The most useful number is the change on one method, measured the same way each time.^[6]

How each method works, and how far off it tends to be

Body fat percentage

Fat mass as a share of total body weight. Everything else, including muscle, bone, organs and water, is fat-free mass.

Many studies judge methods against a four-compartment model, which one review calls the gold standard.^[2] The table shows how far off each method tends to be for a single person, from the studies cited.

DEXA and other lab methods

A dual-energy X-ray absorptiometry (DEXA) scan sends two X-ray beams through you as you lie on a table.^[1] One review calls it the preferred reference method today.^[1] The radiation dose of a whole-body scan is under 10 microsieverts, about one day of natural background radiation at sea level.^[3]

DEXA is not perfect. Most studies comparing it with the four-compartment model found significant differences.^[2] It cannot measure body water, so it assumes a fixed water level and counts extra water as lean tissue.^[1] When one clinic replaced its scanner with a new one of the same model, body fat readings in volunteers differed by 1.1 points on average, which is why new scanners are cross-calibrated.^[4]

Underwater weighing was the gold standard for decades.^[1] It means breathing out as far as you can, then going fully underwater, several times, which many people find very uncomfortable.^[1] The Bod Pod measures body volume by air displacement.^[5] On average it agrees with underwater weighing within 1 point.^[5] The few studies against multi-part lab models suggest both read about 2 to 3 points low.^[5]

Smart scales and other bioimpedance devices

A bioimpedance scale sends a small current through your body. Water carries it easily and fat resists it, and lean tissue is about 73% water, so the device estimates fat from the resistance.^[1] That makes the reading very sensitive to how much water you are carrying.^[1]

In 73 healthy adults, 15 devices were checked against a four-compartment model.^[6] They repeated themselves well, but their average errors ranged from 3.5 points too low to 11.7 points too high.^[6] In hospital patients, three smart scales got weight right but missed DEXA fat mass by 2.2 to 4.4 kg, and the authors said they are not accurate for body composition.^[7]

• Water. Drinking 2 liters of water raised measured body fat by 1.3 to 2.6 points on bioimpedance devices.^[8]
• Meals and time of day. In 18 adults measured through one day, body fat readings varied by 8.8% from highest to lowest in women and 9.9% in men.^[9]
• Your body type. European clinical nutrition guidance says bioimpedance works well in healthy people with normal water balance, but not for routine use at very high or very low BMI.^[10]

So a jump from one day to the next may be water, not fat. For change over 12 to 16 weeks, average errors ranged from 0.4 points under to 1.3 points over.^[6] Even so, about half the devices tracked change poorly, and the authors advise caution.^[6]

Skinfold calipers

Calipers pinch the skin at set sites, and an equation turns the thicknesses into an estimate of body density, then body fat.^{[11, 12]} The Jackson-Pollock equations for women had standard errors of 3.6 to 4.0 points against underwater weighing, across the original and cross-check samples.^[12] In 43 adults, calipers read 4.53 points lower on average than a Bod Pod.^[13]

Skill matters. Missing a site by as little as 1 cm can change the result, and there are more than 100 equations for turning folds into body fat.^[11] In one athlete, the same folds gave 4 to 8% body fat depending on the equation.^[11] One upside: skinfolds appear to be the least affected by day-to-day changes.^[11]

The US Navy tape method

The Navy method, adopted by all US military services in 2002, estimates body fat from a few tape measurements and height, in inches.^{[14, 17]} Men measure the neck and the abdomen at the navel. Women measure the neck, the waist at its narrowest and the hips.^[14]

The 2002 US Defense Department instruction said the method had an error of plus or minus 1 percent.^[14] Later studies disagree. In 481 women and 926 men starting Army basic training, the tape read 6.0 points below DEXA on average.^[15] Over 8 weeks of basic training, DEXA found body fat fell 4.0 points in women, while the tape found no change.^[15]

In 2023 the Army announced a one-site version that uses only the waist at the navel and body weight, and it became the only tape method in 2024.^[18] When researchers applied it to waist sizes taken from 3D phone scans, it was close to DEXA on average but read 1.3 points high against a four-compartment model, and was off by about 4 points for a typical person.^[16]

The Army has since dropped the tape test. A 2026 directive rescinded the 2023 one and made the waist-to-height ratio its only body composition measure, with a standard below 0.55.^[31]

AI estimates from photos

Photo apps estimate body fat from your shape in phone pictures.^[19] In a study written mostly by Amazon staff, 134 adults had a photo estimate and a DEXA scan.^[19] The photo estimate was off by 2.16 points on average, and 95% of results fell between 5.5 points under and 4.7 points over.^[19] The model was trained on people in form-fitting clothes, and loose clothing may give inaccurate results.^[19]

A study of 929 adults, with authors from another company, found errors of about 2.8 to 2.9 points against DEXA.^[20] A separate test of a different app against a four-compartment model found errors of about 5 points, and the app read too low in people with more body fat.^[21] Results also depend on which DEXA machine is treated as the truth.^[22]

Tracking change is less studied. In 50 cadets over 12 weeks, changes from 3D optical scans agreed with DEXA changes better than most other methods, though the authors note a single group and small changes.^[23] The authors of the 929-person study call for more research on tracking change over time.^[20]

What is a healthy body fat percentage?

When researchers proposed ranges in 2000, they noted that no accepted published ranges existed.^[24] They worked out the body fat that lines up with the usual BMI lines, using lab data from 1,626 adults, and called the result provisional.^[24]

The same study gave higher figures for Asian adults, for example 25% to 35% for women aged 20 to 39.^[25] The authors said one universal set of ranges would be hard to build.^[26] These ranges come from lab methods, so a tape or scale reading is not directly comparable with them.

Why it can matter more than BMI: across 25 studies, standard BMI cut-offs missed about half of people with excess body fat.^[27] A 2025 expert commission on obesity says excess body fat should be confirmed by a direct body fat measurement or a measure such as waist size, not BMI alone.^[28] Our guide to waist-to-height ratio covers the tape-measure option.

How to track it so the number means something

1. 1Use one device. Researchers advise using the same device to follow change over time.^[6]
2. 2Same conditions. Measure in the morning after an overnight fast, after using the bathroom, in minimal clothing.^[6]
3. 3Average several days. Track a multi-day average rather than single readings.^[6]
4. 4Do not compare methods. A DEXA result and a scale reading differ by method as well as by any real change. Even DEXA scanners need cross-calibration before they can be compared over time.^{[4, 26]}

When a body fat number falls, it is worth knowing whether muscle went with it. Read losing weight without losing muscle, or start with how to lose weight and keep it off.

When to talk to a clinician