Buy the wrong thermometer and you won’t necessarily know it — you’ll just get a number that quietly disagrees with the truth. For an older or disabled household tracking a temperature after a fall, an infection, or a course of new medication, that half-degree of error is not academic: it’s the difference between watching and waiting versus picking up the phone. This guide sets out, in plain terms, how the three thermometer types actually compare, where each one goes wrong in practice, and what NHS guidance says about using them at home.
Three technologies, three different jobs
Every home thermometer on the market falls into one of three families, and each measures a genuinely different thing.
- Digital oral/underarm thermometers use a probe placed under the tongue or in the armpit. Oral readings are the most direct measurement of core temperature you can take without medical equipment, but the technique has to be followed precisely — the tip needs to sit under the tongue, towards the back, with the lips closed around it. Underarm (axillary) readings from the same device typically come out around 0.2–0.3°C lower than an oral reading, so the two aren’t directly interchangeable for tracking a trend.
- Ear (tympanic) thermometers use an infrared sensor to read heat from the eardrum, which sits close to the body’s core temperature-regulating centre. Correctly aimed, they can agree closely with clinical-grade thermometers and deliver a result in a couple of seconds — useful for someone who can’t comfortably hold a probe under their tongue. But "correctly aimed" is doing a lot of work in that sentence: a build-up of ear wax, an ear canal that’s slightly the wrong shape for the probe, or even a hearing aid still in place can all pull the reading off.
- Forehead (temporal artery / no-contact infrared) thermometers scan the skin over the forehead without touching it, which makes them appealing for someone who is asleep, distressed by contact, or has sensitive skin. The trade-off is that they’re measuring skin surface temperature, not core temperature, and that surface reading is easily thrown off by sweat, a warm room, a draught, direct sunlight, or simply holding the device at the wrong distance. Evidence on this type is mixed: some models track reasonably well against a baseline reading but respond more slowly to a genuine change in temperature, and contactless thermometers as a group have tended to read lower than oral devices, with weaker sensitivity for actually catching a fever.
Accuracy at a glance
| Type | What it measures | Typical strength | Where it goes wrong |
|---|---|---|---|
| Digital oral | Core temperature (mouth) | Most direct reading; low cost | Needs correct technique; skewed by recent food or drink |
| Digital underarm | Core temperature (approximate) | Easiest for someone who can’t hold a probe steady | Reads 0.2–0.3°C lower than oral; needs the arm held still |
| Ear (tympanic) | Near-core (eardrum) | Fast, good agreement when aimed correctly | Wax, poor aim, hearing aids in place |
| Forehead (infrared) | Skin surface | No contact; good for sleeping or touch-sensitive users | Room temperature, sweat, sunlight, distance from skin |
The mistake that matters more than the model you buy
Most of the "my thermometer is wrong" complaints we’d expect to hear aren’t really about the device — they’re about technique, and that’s true across all three types. An oral reading taken straight after a hot drink will run high. An ear reading taken without pulling the ear up and back first (the manufacturer instructions on every tympanic thermometer we’re aware of say to do this) will sit against the ear canal wall rather than the eardrum. A forehead reading taken in a cold hallway straight after someone’s come in from outside will read low. None of this is a fault with the device; it’s a fault with the five seconds before the reading was taken. If you’re caring for someone else and using the same thermometer over days or weeks to track a trend, the single most useful thing you can do is use the same type, on the same site, the same way every time — consistency matters more than which family of thermometer you picked.
What counts as a fever, and what the NHS says to avoid
NHS guidance defines a high temperature in adults as 38°C or above, against a normal baseline that sits around 37°C and varies a little between individuals and through the day. Guidance is also clear that a thermometer isn’t always necessary — feeling unusually hot to the touch on the chest or back, shivering, sweating, or flushed skin can indicate a high temperature even before you reach for a device, and a reading just under 38°C doesn’t rule out a genuine fever if the person still feels hot, cold or shivery.
On device choice specifically, NHS guidance steers households away from glass thermometers (a breakage and mercury-exposure risk that has no place in a modern medicine cabinet) and away from forehead adhesive strips, which are not considered accurate enough to rely on. Some infrared thermometers are also flagged as potentially falling short of UK regulatory standards, which is a reminder to check for a CE or UKCA mark and a listed manufacturer before buying rather than picking the cheapest unbranded option on a marketplace listing.
A watchdog’s buying checklist
Before you buy a home thermometer for yourself or for someone you care for, it’s worth checking:
- Can the user physically use it correctly? An oral thermometer is only as accurate as the person’s ability to hold it still under the tongue for the full reading time. If that’s not realistic, an ear or forehead device removes that dependency.
- Is it CE or UKCA marked, from a named manufacturer? This is a basic floor, not a guarantee of accuracy, but its absence is a genuine warning sign.
- Does the packaging state a clinical accuracy range (most reputable digital and ear thermometers state something in the region of ±0.2°C under test conditions)? If a listing makes no such claim at all, treat that as missing information, not a neutral fact.
- Will you use it consistently? A slightly less "accurate" device used the same way every time will give you a more useful trend than switching between types.
Choosing on evidence, not the loudest packaging claim
None of the three thermometer families is universally "the most accurate" — each is a trade-off between measurement precision, ease of use, and how forgiving it is of imperfect technique. For most independent adults, a well-made digital oral thermometer, used correctly, remains the simplest way to get a reliable reading. For someone who can’t safely or comfortably hold a probe in place, a correctly aimed ear thermometer is the better compromise. Forehead devices earn their place for quick, non-invasive checks on someone who’s asleep or touch-sensitive, provided you understand that a normal-looking forehead reading in a genuinely unwell person shouldn’t be treated as the final word. The choice that matters isn’t which one wins a lab test — it’s which one you and the people you’re monitoring will actually use properly, every time.