The most accurate body-fat test, ranked: every method compared
By The DexaExplained Team · Last verified
Disclosure: some links below are affiliate or referral links — if you book or buy through them we may earn a commission, at no cost to you. It never changes a price or a chart. How we make money.
Educational body-composition information, not medical advice. Nothing here diagnoses, treats, or prevents any condition, and this site does not interpret bone-density results. Talk to your clinician about anything on a scan report that concerns you.
Every method on this page produces a body-fat percentage. They do not produce the same body-fat percentage, and the spread between the best and worst is bigger than most of the marketing admits. Short version: the 4-compartment model is the research standard, DEXA is the best test you can actually book, capsule-and-current devices (Bod Pod, InBody) are decent estimates, and home smart scales are trend tools with real absolute error.
Every method, one table
| Method | What it actually measures | Typical cost | Error vs criterion | Regional data |
|---|---|---|---|---|
| 4-compartment model | Body volume + total body water + bone mineral, combined | Research labs only | The criterion itself | No |
| DEXA | X-ray attenuation of fat, lean, bone | $40–185 (local pricing) | Low; repeat precision under 1% CV for fat mass | Yes — arms, legs, trunk, visceral estimate |
| Hydrostatic weighing | Body density via water submersion | University labs | Moderate (2-compartment assumptions) | No |
| Bod Pod | Body density via air displacement | $40–80 at university labs | Underestimated body fat vs DEXA in direct comparison | No |
| InBody (multi-frequency BIA) | Electrical impedance → equation | Often free at gyms; $25–50 standalone | Underestimated body fat vs DEXA; BIA-class limits of agreement span 15–20 points | Segmental estimates |
| Smart scale (2-electrode BIA) | Foot-to-foot impedance → equation | One-time purchase | High for absolute values (same BIA limits) | No |
| Calipers | Skinfold thickness at set sites | Very cheap | Entirely skill-dependent | Subcutaneous sites only |
| Tape / Navy formula | Circumferences → equation | Effectively free | Rough estimate | No |
The criterion: the 4-compartment model
Research needs a referee, and it's the 4C model: measure body volume, total body water, and bone mineral separately, then solve for fat with minimal assumptions. The 2026 systematic review by Oliver et al. — the source for the BIA numbers below — uses it as the criterion method, which is the standard position in the validation literature. You cannot book a 4C assessment at a strip-mall clinic; it exists to keep every other method honest.
The tier you can book: DEXA, Bod Pod, hydrostatic
DEXA is the closest thing to 4C money can buy. It resolves fat, lean soft tissue, and bone mineral separately and reports them per region — which is why it's the reference device in comparison studies and the basis of the NHANES national dataset behind our percentile charts. Repeat-scan precision in weight-stable adults: 0.74% CV for total fat mass, same-day. The full error story lives in our DEXA accuracy breakdown.
Bod Pod and hydrostatic weighing both measure one thing — body density — and convert it to fat percentage with a two-compartment equation. Hydrostatic dunks you in a tank; the Bod Pod seals you in an air capsule. In the biggest three-way comparison (Antonio et al. 2019, 155 trained adults), the Bod Pod underestimated body fat relative to DEXA and was statistically indistinguishable from an InBody 770. Full comparison here.
DEXA is cheaper than its reputation — often $40–60 self-pay depending on your metro.
Check DEXA prices near you →The estimate tier: InBody and smart scales
All bioelectrical impedance (BIA) devices — the gym's InBody and the bathroom smart scale alike — run a current through you and estimate composition from impedance plus equations. Oliver et al.'s 2026 systematic review put BIA devices against the 4C criterion across 34 evaluations: mean bias ranged from −3.5 to +4.4 percentage points of body fat, which sounds fine until you see the individual limits of agreement — typically spanning 15 to 20 percentage points. Group averages behave; individuals get told 18% when they're 26%.
The honest frame for a home smart scale: the absolute number is unreliable, but a daily reading taken under the same conditions (morning, post-bathroom, pre-coffee) is a legitimately useful trend line between real measurements. That's the job it's good at — and it's the same job the InBody does for gym-goers, with more electrodes.
The manual tier: calipers and tape
Calipers measure subcutaneous fat at specific sites and feed the folds into a prediction equation. In experienced hands they're a defensible tracking tool; in inexperienced hands the site selection and pinch pressure vary more than your body does. The tape-measure (Navy) method converts a few circumferences into an estimate — free, fast, and rough. Both share a virtue: consistency costs nothing, so they work as trend tools if one person measures the same way every time.
The honest playbook
- DEXA once or twice a year for the real numbers and the regional data.
- A BIA device in between — smart scale or gym InBody — for the trend, never the absolute value.
- Never compare across methods. Every device gets its own lane; progress is judged within a lane.
Frequently asked questions
What is the most accurate body-fat test?
The 4-compartment model — which combines body volume, total body water, and bone mineral measurements — is the criterion method research uses to judge everything else. You can't book one commercially. Of the tests you can actually buy, DEXA comes closest: it measures fat, lean, and bone separately with same-day repeat precision under 1% CV for fat mass, and it's the method national reference data is built on.
Are smart scales accurate for body fat?
For the absolute number, no. A 2026 systematic review of bioelectrical impedance devices against the 4-compartment model found small average bias but individual limits of agreement spanning 15 to 20 percentage points of body fat — one person's reading can be off by a lot even when the group average looks fine. For a daily trend measured under consistent conditions, they're genuinely useful. Treat the trend as signal and the absolute number as decoration.
Is DEXA the gold standard for body fat?
It's the practical gold standard — the reference method in most comparison studies and the basis of the NHANES national dataset — but the research criterion is the 4-compartment model, which adds a total-body-water measurement DEXA doesn't make. For anything you can walk in and buy, DEXA is the benchmark.