BBQHQThe world of cooking over fire

Sources

University of Georgia College of Family and Consumer Sciences

Calibrate food thermometer before Xmas cooking (UGA CAES Field Report)

What this source is

Authority
Academic
Territories
US
Primacy
secondary
Commercial interest
none
Retrieved
2026-08-22
State
Read and citable

Original document

By April Reese. The instrument half of every endpoint on this site, and until now BBQHQ argued for a thermometer without ever saying how to find out whether yours is telling the truth. The opening line is the argument: `Be sure to calibrate your thermometer before checking temperatures, or they don't mean anything.` The method is the ice slush rather than boiling water and the source says why — `The ice method is a much easier process to manage than boiling water and very accurate` — which also sidesteps the altitude correction boiling requires. Note the wait: 'This may take 5 to 10 minutes', far longer than people give it. And the honest ending for an instrument that cannot be adjusted: 'Replace the thermometer if it's more than a few degrees off.'

What it may be cited for

Safety, Technical

  • Explicitly not culinary practice: It describes how to check an instrument. It says nothing about how long anything takes or how it should taste.

What we recorded from it

  • Be sure to calibrate your thermometer before checking temperatures, or they don't mean anything.

    The University of Georgia states that a thermometer must be calibrated before it is used to check temperatures, or its readings do not mean anything.

  • The ice method is a much easier process to manage than boiling water and very accurate. [...] Make a 50/50 ice and water slush — at least half ice...Immerse the thermometer stem into the slush...Wait until the temperature reading stops changing. This may take 5 to 10 minutes. [...] The temperature should read 32 degrees F. If it doesn't, turn the calibrating nut under the dial.

    The University of Georgia gives the ice-slush method — a half-ice, half-water slush, the stem immersed, waiting five to ten minutes for the reading to settle, which should read 32 °F — and prefers it to boiling water.

    32 °F is the source's unit. It is 0 °C, and that is a definition rather than a conversion of a measured figure, so stating both is not the rounding hazard the catalogue usually guards against.
    'Turn the calibrating nut under the dial' describes a dial thermometer. Many instant-read digital instruments have no nut, which is why the replacement sentence below matters more than it looks.
    Three fragments from the same article, joined with bracketed ellipses.
  • Replace the thermometer if it's more than a few degrees off.

    The University of Georgia instructs that a thermometer more than a few degrees out should be replaced.

    'A few degrees' is not quantified by the source and is not quantified here. The catalogue does not invent a tolerance figure to make the sentence tidier.

Claims resting on it

  • A food thermometer should be checked against a known reference before it is trusted, using a half-ice half-water slush which should read 32 °F (0 °C) after five to ten minutes; a dial instrument is adjusted at the nut beneath the dial, and any instrument more than a few degrees out should be replaced.
  • A food thermometer can be checked in an ice-water slush — at least half ice — with the stem immersed and touching neither side nor bottom; it should read 32 °F (0 °C). Two extension services differ on how long to wait: Clemson gives at least thirty seconds, and the University of Georgia says the reading may take five to ten minutes to settle. Wait until it stops changing. A dial instrument is adjusted at the nut beneath the dial; one that cannot be adjusted, or that is more than a few degrees out, should be replaced. The University of Georgia's reason for checking at all is the blunt one: readings from an uncalibrated thermometer do not mean anything.

If this source changes, these are the claims that need rechecking. That is the point of recording provenance per claim rather than per page.

We’d like to measure which pages help, using anonymous analytics with no personal data and no search terms. Declining changes nothing about the site.