Are bioimpedance scales accurate? What the studies show

By PounceFit Team · Updated on 2026-10-06

Technical content for health and fitness professionals. It reproduces formulas and protocols published in the scientific literature, with the sources at the end of the page. It is not a recommendation and does not replace the judgment of the professional who runs the assessment.

A bioimpedance scale is accurate at repeating its own reading, but not at giving one person's exact body fat percentage. In a study of 15 devices, the error between repeated readings was at most 0.49%, while the mean error against the reference method ranged from 3.5 points under to 11.7 points over, depending on the device (Siedler et al., 2023). That makes it better suited to tracking a trend, under standardized conditions, than to pinning down a number.

What the studies show

Repeatability. Siedler et al. (2023) tested 73 healthy adults aged 19 to 50 on 15 bioimpedance devices, 14 of them consumer-grade, in three formats: hand-to-hand, foot-to-foot and hand-and-foot (eight electrodes). The comparison was against the four-compartment model, a laboratory method. All devices repeated their readings well: precision error of 0 to 0.49%.

Error in a single reading. In the same study, the mean error against the reference model ranged from 3.5 points under to 11.7 points over, and the standard error of the estimate ranged from 3.1 to 7.5 points, depending on the device (Siedler et al., 2023).

Error in change. In a subgroup of 37 people reassessed 12 to 16 weeks later, the body fat change measured by the devices had standard errors of 1.7 to 2.6 points. The authors concluded that some models, particularly eight-electrode devices and certain foot-to-foot models, may be useful for tracking body composition over time (Siedler et al., 2023).

Systematic review. Oliver et al. (2026) pooled 12 studies comparing bioimpedance devices with the four-compartment model in healthy adults. The mean body fat error was small, from 3.5 points under to 4.4 points over, but the limits of agreement typically spanned 15 to 20 percentage points. For fat-free mass, the limits often exceeded 6 kg either way. The conclusion was that, at the individual level, bioimpedance is not equivalent to the reference method.

In practice: on a group average, the scale comes close; for one person, it can be far off, and the error depends on the device.

Why the scale gets it wrong

The scale does not measure fat. It measures the body's impedance and uses an equation to estimate fat-free mass; fat is what is left of body weight (Kyle et al., 2004, part I). Three things weigh on that path:

• The equation. It was built in a specific population. ESPEN calls for an equation validated for the age, sex and ethnicity of the person assessed (Kyle et al., 2004, part II).
• The current path. The research standard places electrodes on the hand and foot. Some devices use other positions, such as foot-to-foot (Kyle et al., 2004, part I). A foot scale therefore measures a different current path, and the equation needs to have been built or validated for that kind of device. See bioimpedance: what it is and how it works.
• The state of the body. Food, drink, exercise and hydration change impedance (Kyle et al., 2004, part II).

What changes the reading

According to ESPEN (Kyle et al., 2004, part II):

• Eating or drinking can lower impedance by 4 to 15 ohms over the following 2 to 4 hours. Being fasted or fed is what affects the result most.
• Right after exercise, resistance drops by about 3% and reactance by about 8%; values return to normal within an hour.
• At extreme BMI or with abnormal hydration, routine use is not recommended until further validation.

How to standardize the measurement

The ESPEN protocol (Kyle et al., 2004, part II) includes:

• Height and weight measured on the spot; self-reported values are not valid.
• Fasting and no alcohol, ideally for 48 hours; shorter periods may be acceptable in clinical practice.
• Empty bladder before the measurement.
• No exercise in the previous 48 hours.
• For repeated measurements, always at the same time of day, noting the menstrual cycle phase.
• Clean skin where it touches the electrodes.

Under these conditions, ESPEN considers body composition follow-up possible at a BMI between 16 and 34 with normal hydration, read with caution. Changes smaller than 1.5 to 2 kg fall within the precision limit of the method.

The scale in PounceFit

PounceFit, the body assessment software, reads the scale over the computer's Bluetooth, stores the raw impedance in ohms and calculates body composition with the client data already on file. The professional picks the formula in Settings: the scale manufacturer's (the default) or published scientific equations, which the app flags as built for a different kind of device and not validated for scales.

Scale assessments keep their own history and never mix with skinfolds or with another scale. The scales the app recognizes are listed under compatible scales, and the bioimpedance scale for body assessment page explains what the app measures. To compare with the caliper, see bioimpedance vs skinfolds.