Body Fat Measurement Methods Compared (And Their Real Error)

In this guide

1. There is no true reading
2. The methods, with their error
3. The US Navy tape method
4. Why bioimpedance scales swing so much
5. Consistency beats accuracy
6. What the numbers mean
7. Common questions

There is no true reading

Worth stating plainly at the start: nobody has ever directly measured your body fat percentage. Direct measurement requires cadaver dissection. Everything available to a living person is an inference — from density, from electrical resistance, from X-ray attenuation, from the thickness of a skin fold — and every inference rests on assumptions that are truer for some bodies than others.

This is not a reason for despair. It is a reason to change what you ask of the measurement. A number that is consistently three percentage points high is useless as a truth and excellent as a tracker, because the error cancels when you compare two readings.

The methods, with their error

Two things stand out. Even DEXA — the method everyone treats as truth — carries one to three percentage points of error, and readings differ between machines and software versions. And the free tape method is not dramatically worse than calipers in untrained hands. The expensive methods buy you a better single reading; they do not buy you better tracking.

The US Navy tape method

Developed by Hodgdon and Beckett in 1984 for military body composition screening, this method uses circumference measurements — neck and waist for men, with hips added for women — plus height, in a logarithmic regression.

Its logic is sound: it is essentially measuring where fat is deposited, and circumference at the waist relative to the neck is a reasonable proxy for adiposity in most bodies. Its failure modes are equally logical. Someone with an unusually thick neck from heavy training gets read as leaner than they are. Someone with a naturally narrow waist and high total fat gets read as leaner too. And because the equation is regression-derived from a military population, it drifts at the extremes of body composition.

The practical requirements matter more than the equation. Measure in the morning before eating, tape snug but not compressing the skin, level all the way round, waist at the navel, neck below the larynx, and take each measurement three times and average them. Our body fat calculator runs the Navy equation alongside a BMI-based estimate so you can see how far apart two methods land on the same body.

Why bioimpedance scales swing so much

Bioimpedance analysis passes a small current through your body and infers composition from the resistance it meets. Fat conducts poorly, lean tissue conducts well — largely because lean tissue holds far more water.

Which is precisely the problem: the measurement is dominated by hydration status. A foot-to-foot scale reads mainly your legs and extrapolates the rest, so it is sensitive to everything that moves water around:

• Time of day — morning readings differ substantially from evening.
• Hydration — a litre of water changes the reading.
• Recent exercise — inflammation and fluid shifts distort it for hours.
• Sodium and carbohydrate intake the day before.
• Skin temperature and how damp your feet are.

This is why a scale can show a four-point change in two days, which is physiologically impossible for actual fat. Four to eight points of error is the honest range. Use it if you like, but only as a trend line, measured at the same time of day under the same conditions — and never take a single reading seriously.

Consistency beats accuracy

Here is the reframe that makes this whole topic manageable. Suppose you use the tape method and it reads three points low. Your readings across four months are 22, 21, 19.5, 18 percent — all three points off, so your real values were 25, 24, 22.5, 21. The absolute numbers were wrong throughout. The trend, the rate of change, and every decision you would make from them were all correct.

So the rules are simple:

1. Pick one method and never switch mid-project. Comparing a DEXA scan in January with a bathroom scale in March tells you nothing.
2. Standardise the conditions. Same time of day, same hydration state, before eating, before training.
3. Measure monthly, not daily. Real composition change is slower than any method’s noise floor.
4. Track the direction, not the digit. “Down two points in eight weeks” is actionable. “19.4 percent” is false precision.
5. Cross-check with a tape and a camera. Waist circumference and monthly photographs are the most honest feedback available, and both are free.

What the numbers mean

Women carry more essential fat than men for reasons of reproductive and hormonal function — this is biology, not a fitness gap, and comparing across the two columns is meaningless. Below the essential range, health suffers in both sexes; sustained very low body fat is associated with hormonal disruption, menstrual dysfunction, poor sleep and impaired immune function.

Two closing uses for whichever number you land on. It converts your weight into lean body mass, which lets you set protein from lean mass rather than total weight — see the protein target table. And it tells you whether to cut, bulk or recomp, which is the one decision where being three points off genuinely does not change the answer.