Body Fat Percentage — DEXA, Smart Scales, and What the Numbers Actually Mean

A reader writes in frustrated. Their bathroom smart scale shows their body fat percentage drifting up by 2 points across a single week — despite no obvious change in habits. They bought a higher-end scale. Same result. They went and got a DEXA scan to confirm. The DEXA result was 6 points lower than either scale showed.

The methods aren't broken. They are measuring related but different things, and the only error is treating the numbers as interchangeable.

What Each Method Actually Measures

Smart scales (bioelectrical impedance). A weak electrical current passes through the body. Fat resists current; muscle and water conduct. The scale estimates body composition from the current's path. The accuracy depends heavily on hydration — a glass of water before stepping on changes the reading by 1–3 points. Time of day matters. So does the surface skin temperature. Smart scales are useful for tracking trends, useless for absolute values.

Skinfold calipers. A trained measurer pinches fat at specific anatomical sites — typically 3, 4, or 7 sites depending on the protocol — and feeds the millimeter measurements into a regression equation. The error band is roughly ±3% for an experienced measurer, much higher for a beginner. Calipers measure subcutaneous fat specifically and ignore visceral fat, which is the fat that actually matters for cardiometabolic risk.

DEXA scans (dual-energy X-ray absorptiometry). Two X-ray beams of different energies pass through the body. Bone, lean tissue, and fat attenuate the beams differently, allowing the scan to map each tissue separately. The error band is roughly ±1–2% in controlled conditions. DEXA is the closest thing to a gold standard available without dissection — and it reports regional breakdowns (trunk fat vs. limb fat) that no other consumer-accessible method delivers.

BodPod (air displacement plethysmography). A sealed chamber measures the volume of air the body displaces, then derives density and from that, fat percentage. The accuracy is similar to DEXA in controlled settings, but BodPods are increasingly rare and the result depends on accurate lung-volume corrections.

Why the Numbers Diverge

The methods are estimating the same physical quantity through wildly different physical signals. They all assume normal distributions of bone density, normal hydration, and average proportions. When you don't fit those assumptions, the methods disagree in different directions.

A muscular person with low hydration. Smart scale reads high (low conductivity), DEXA reads accurate, calipers read accurate. Result: smart scale says 22%, DEXA says 17%.

An older adult with low bone density. DEXA's formulas assume average bone — low bone density inflates the apparent fat percentage. The DEXA number for a person with osteoporosis can read 2–4 points higher than the actual value.

An endurance athlete with high water content. Smart scales drastically underestimate fat percentage because their formulas assume water content of around 60%. Endurance athletes can run 65–70%, making the scale read several points low.

Someone immediately post-meal or post-workout. Every method shifts. Eat before measuring and you get a different reading than measuring fasted. Workout shifts blood flow and water distribution for hours.

What the Numbers Tell You (And What They Don't)

Body fat percentage is one signal in a constellation. It does not tell you anything about strength, cardiometabolic risk, or aesthetic body composition on its own.

For cardiometabolic risk, waist-to-height ratio beats body fat percentage. Visceral fat is what damages metabolic health, and visceral fat is reflected far better in waist measurement than in any total body fat number. A waist measurement under half your height is the simple threshold that correlates with most risk markers.

For strength training progress, fat percentage is too noisy to use weekly. Use monthly averages, ideally measured under identical conditions (same time of day, same hydration state, same protocol). Track the trend, not the individual reading.

For weight loss, the DEXA scan is most useful for distinguishing fat loss from muscle loss. A scale reading shows total weight change. A DEXA breaks it into lean mass and fat mass. Most weight loss protocols target 70–80% fat / 20–30% lean. If the DEXA shows you've lost equal lean and fat, the protocol needs adjustment.

What to Do Practically

For most people, three habits resolve most of the confusion.

Pick one method and stick to it. Switching methods week to week guarantees noise. Use the same smart scale at the same time of day after the same morning routine, or schedule DEXA scans on the same date each quarter. Comparing this week's measurement to last week's only makes sense if both used identical conditions.

Track averages, not single readings. A seven-day moving average filters most hydration and timing noise out. A single reading is almost meaningless. The average is where the signal lives.

Anchor on one absolute reading per year. A single DEXA scan annually gives you a calibrated reference. Smart scale numbers between DEXAs are useful as relative changes from that anchor — not as standalone truths.

The numbers from each method are real numbers. They are also each measuring something slightly different, and pretending otherwise turns a useful tool into a source of confusion. The body fat percentage that matters most is the one you can measure consistently over time. The exact value is almost beside the point.