Hello Dr. Humaid Khan,
Thank you for your reply.
Assuming that our measuring devices have been certificated as "good" and it has a 1 year calibration validation, but then during in-house calibration there is a deviation found, by how much degree/size of deviation that we can tell the devices is out of tolerance already and should not be used. If the size of deviation is big e.g. 6 degrees Celcius but we apply the correction factor, is it still be acceptable?
This is Metrology which is a science unto itself. The comments below are probably simplifications.
The desired performance of a given measuring instrument relates to its usage, ie what you are measuring and to what required accuracy.
The typical "accuracy" delivered by measuring intruments is defined by their specifications (and often by cost). Often expressed by a maximum percentage error as a function of measurement value over defined ranges. Statistics can provide more elaborate answers.
The capability of a measuring instrument to meet its specification can be verified/validated by Calibration. And Statistics.
Nonetheless, there are typical "rules-of-thumb" for various instruments/functions which are commonly stated in calibration SOPs. These may be usable if the Procedure is validated and matches yr own usage requirements
As per post2, a comprehensive thermometer calibration certificate allows you to estimate the accuracy of its indicated readings for the test points used and predict the accuracy within the range employed.
Regarding post 3, if the unit referred is a thermcouple combination, being calibrated in ice-water and with no reset option i would first repeat with another, known accurate, probe. If same result send combination to the repair shop. If second result OK, discard the first probe and send it to the repair shop.