Thermometer Accuracy: Standard Limits by Type

This page describes products and general use; it is not medical advice.

Thermometer accuracy is how closely a displayed temperature matches a reference temperature, within a stated error limit. For clinical electronic thermometers that limit is set by the standards named in U.S. regulation, and it is a laboratory figure: it does not by itself tell you how well a reading on a person will agree with their true body temperature. Clinic staff, buyers and students use this page to read an accuracy specification correctly and to know which use conditions change a reading.

What “accuracy” means for a thermometer

In the standards below, accuracy is a maximum permissible error: the largest difference allowed between the thermometer’s display and a reference, across a stated temperature band. A smaller limit is a tighter specification. This is a different idea from precision (how closely repeated readings agree with each other); see accuracy vs precision in measurement.

In U.S. regulation, a clinical electronic thermometer is a device that measures the body temperature of a patient using a transducer, electronic signal amplification, conditioning and a display (21 CFR 880.2910). It is Class II, and is exempt from premarket notification provided testing shows it meets its specifications. The regulation names ISO 80601-2-56, ASTM E1965, ASTM E1112 and ASTM E1104 as suitable standards for validating specifications and performance.

Accuracy limits in the standards

The table gives the maximum permissible errors as summarized in Table 8 of Wang et al. (2022). The editions are the ones that paper lists, which may predate the current editions, so check the edition cited by a specific product’s labeling.

Standard (edition listed)	Thermometer type	Temperature band	Maximum permissible error
ASTM E1112 (2000, reapproved 2011)	Electronic thermometer	Below 35.8 °C (rated range 35.5 to 41.0 °C)	±0.3 °C
ASTM E1112	Electronic thermometer	35.8 to 37.0 °C	±0.2 °C
ASTM E1112	Electronic thermometer	37.0 to 39.0 °C	±0.1 °C
ASTM E1112	Electronic thermometer	39.0 to 41.0 °C	±0.2 °C
ASTM E1112	Electronic thermometer	Above 41.0 °C	±0.3 °C
ISO 80601-2-56 (2017)	Clinical thermometer, 34.0 to 43.0 °C	Within rated output range	±0.3 °C
ISO 80601-2-56 (2017)	Clinical thermometer	Within rated extended range	±0.4 °C
ASTM E1965 (1998, reapproved 2016)	Infrared, ear canal, 34.4 to 42.2 °C	Below 36.0 °C	±0.3 °C
ASTM E1965	Infrared, ear canal	36.0 to 39.0 °C	±0.2 °C
ASTM E1965	Infrared, ear canal	Above 39.0 °C	±0.3 °C
ASTM E1965	Infrared, skin, 22.0 to 40.0 °C	Over the display range	±0.3 °C

Who uses it: clinic and ward staff who take temperatures, buyers comparing manufacturer specifications, and students learning how device standards are written.

Laboratory accuracy versus clinical accuracy

Wang et al. note that none of these standards includes clinical accuracy requirements for infrared thermometers or thermometers: their limits describe agreement with a reference, not agreement with a person’s true temperature at a given body site. A device can therefore meet its specification and still read differently on a patient because of the measurement site and the conditions of use.

The same paper reports that the forehead site used by non-contact infrared thermometers tends to give poorer accuracy than full-face or inner-canthi approaches in infrared thermography research. That finding concerns research systems; it is not a ranking of consumer or clinical products.

What affects a reading

The FDA states that improper use of non-contact infrared thermometers may lead to inaccurate measurements, and that how and where the device is used may affect the result (page current as of 6 March 2026). Its guidance on use conditions includes:

• The device should be used in the environment the manufacturer specifies; the FDA notes that the environment can affect the measurement.
• The forehead should be clean, dry and uncovered. Headbands, excess clothing and facial cleansing products can change the reading.
• Hold the sensing area perpendicular to the forehead, with the person still, at the distance the manufacturer specifies.
• Do not touch the sensing area, and keep it clean and dry.
• Direct sunlight on the forehead is incorrect use.

For how these devices are used step by step, see how to use an infrared thermometer; for the broader product class, see what a medical thermometer is and infant digital thermometers.

Accuracy specification versus calibration

An accuracy specification states what error a device is allowed to have. Calibration is the comparison of a device against a reference to check or adjust that error; see what calibration is and how calibration intervals are set. Follow the manufacturer’s instructions for use for any calibration or verification of a specific thermometer.

Common mistakes

• Reading a “±0.2 °C” figure as a guarantee of clinical accuracy on every patient and site, when the standards test against a reference.
• Comparing specifications without checking the temperature band they apply to; the limits in the table change by band.
• Using an infrared thermometer straight from a different room, or on a forehead with a headband, sweat, cosmetics or sunlight on it.
• Comparing two products’ specifications that cite different standards or different editions.

FAQ

How accurate should a clinical thermometer be?

Under ASTM E1112 as summarized by Wang et al., an electronic thermometer’s allowed error is ±0.1 °C between 37.0 and 39.0 °C and up to ±0.3 °C at the ends of its range; ISO 80601-2-56 (2017) lists ±0.3 °C within the rated output range.

Does a thermometer that meets a standard read correctly on a person?

Not necessarily. The standards set laboratory limits against a reference, and measurement site and use conditions also affect the reading.

What can make an infrared forehead thermometer read wrongly?

The FDA lists headwear, the environment, facial cleansing products, direct sunlight, a wet or touched sensing area and the wrong distance or angle.

Is accuracy the same as calibration?

No. Accuracy is the specified error limit; calibration is checking a device against a reference.

Related supplies are listed in the lac.us category for diagnostic, monitoring and imaging supplies.

Sources

• 21 CFR 880.2910, Clinical electronic thermometer, Cornell Legal Information Institute, checked 2026-10-30: law.cornell.edu/cfr/text/21/880.2910
• U.S. FDA, Non-Contact Infrared Thermometers (content current as of 6 March 2026), checked 2026-10-30: fda.gov/medical-devices/general-hospital-devices-and-supplies/non-contact-infrared-thermometers
• Wang Q, Zhou Y, Ghassemi P, McBride D, Casamento JP, Pfefer TJ. Infrared Thermography for Measuring Elevated Body Temperature: Clinical Accuracy, Calibration, and Evaluation. Sensors 2022;22(1):215, Table 8, checked 2026-10-30: PMC8749858