My problem is always "how accurate is accurate enough?" In Blood Banking, the standards read in integers, so we'd like to verify with something to the next decimal point (+ 0.1 C). However, when you start looking into LIG, infrared and electronic thermometers for bench use, you find that the accuracy does not approach 0.1 C, esp with most LIG thermometers that come with "certificates of accuracy". And we haven't even discussed other measurement variances yet ... It almost seems like most of what we do is a low-leveraged activiy of borderline utility, done only to show an inspector that we're doing something!