Guide

How accurate are hygrometers?

A good capacitive sensor is ±2–3% relative humidity when new, a consumer one ±5%, and a hair hygrometer ±5–10% and worsening. In practice, drift and placement usually contribute more error than the specification does — and for a weather-app reading, distance to the nearest station dominates everything else.

The specifications

Typical accuracy when new and correctly used.
Type	Accuracy	Resulting dew point error at 20°C
Chilled mirror	±0.2°C dew point	±0.2°C
Professional capacitive	±2% RH	about ±0.3°C
Good consumer capacitive	±3% RH	about ±0.5°C
Basic consumer device	±5% RH	about ±0.8°C
Hair tension	±5–10% RH	±0.8–1.6°C
Drifted / uncalibrated	±10% or worse	1.6°C and up

Those are the numbers on the box. They are also, in practice, the least important part of the picture.

The three things that beat the specification

1. Drift

Sensing materials change. Polymer films are altered by contaminants, solvents and repeated condensation; hair loses elasticity and takes a set. The device keeps responding to humidity — just no longer along the curve its calibration assumes.

About 1% per year for a decent capacitive sensor, faster for everything else. Nothing warns you. The reading stays plausible and slowly stops being true, which is why the salt test is worth doing on anything you rely on.

2. Placement

A hygrometer reports the air at the sensor, and humidity varies sharply within a single room. Near a cold exterior wall the air is close to saturation; over a radiator it is far from it; in the middle of the room it is somewhere between.

None of those readings is wrong. They are measurements of different air, and the instrument cannot tell you which one you meant.

Outdoors the same problem is worse, and it is why official observations sit inside a radiation shield: a sensor in direct sun measures its own heated enclosure, and since relative humidity depends on temperature, a few degrees of solar heating produces a large humidity error.

3. Response time

An instrument that has not settled is not reading the air; it is reading where it has been. Walk a capacitive sensor from a cold hallway into a warm bathroom and it needs a little time. A hair hygrometer needs considerably more.

Where the error is for an app reading

Completely different profile, and worth understanding separately.

A weather-service humidity figure, by contribution to the error.
Source	Size
Distance to the nearest observation	Dominant. Humidity varies over hundreds of metres.
Model interpolation between stations	Significant. Grid spacing is kilometres.
The station's own sensor	±2–3%. Small, and well maintained.
The conversion arithmetic	Negligible. Well under 0.1%.

The ordering surprises people. The instrument is the best-controlled part of the chain; the geography is the weak link. A valley, a lake, a car park and a wood a few hundred metres apart can genuinely differ by several degrees of dew point at the same moment.

This is stated plainly on where the data comes from, because an app that presented an interpolation as a measurement would be overstating what it knows.

What accuracy is good enough

Depends entirely on the question:

Is it muggy? Will it dry? Will the windows fog? ±5% is ample. The thresholds are broad.
Is my humidor at 70%? ±2% after calibration, or the cigars suffer.
Should this room have a dehumidifier? Trends matter more than absolutes; ±5% is fine as long as it is consistent.
Regulatory or process control? Calibrated, traceable, and checked on a schedule.

For most people, most of the time, a cheap sensor that has been salt-tested once beats an expensive one that never has.

The most useful thing you can do

Test what you own. Twenty minutes, a tablespoon of salt and a sealed bag will tell you your instrument's actual error instead of the one printed on the packaging — and that number is the one that matters. The procedure is here.

Common questions

Why do my two hygrometers show different readings?

Three likely reasons, in order: they are measuring genuinely different air, since humidity varies over short distances; one or both have drifted since manufacture; and their specifications may permit several percent of difference even when both are within tolerance.

How much do hygrometers drift?

Roughly 1% of relative humidity per year for a decent capacitive sensor, faster for resistive types, and substantially faster for hair hygrometers. After five years an uncalibrated device can easily be outside anything its box claimed.

How do I know if my hygrometer is accurate?

Test it. The salt test uses saturated sodium chloride to create a known 75% environment, needs only salt, water and a sealed bag, and will tell you your instrument's actual error rather than its advertised one.

Are cheap hygrometers worth it?

For trends and rough levels, yes. For anything where a few percent matters — a humidor, an instrument case, a curing process — calibrate it first, because the error out of the box is frequently larger than the specification.

Is a humidity reading accurate at very low or very high humidity?

Usually worse at both ends. Most sensors are specified over a middle band and degrade outside it, and prolonged exposure to condensation can permanently shift a polymer film's response.

Sources

WMO Guide to Instruments and Methods of Observation (WMO-No. 8)

Published 12 August 2026