Error Propagation

I did an experiment measuring the frequency of the note a ukulele string made when plucked, at different temperatures. I calculated the temperature by using dT=VIt/mc. Now to the question. I used that formula to calculate the percentage error of temperature, by adding the percentage errors of each of the components. One of the percentage uncertainties for temperature I got was 20%, when the temperature was 27.4°C. The problem is, I don't know if, to find the actual uncertainty value, I need to take 20% of 27.4°C (so around +-5.5°C), or if I should convert the temperature to Kelvin (300.4K) and take the 20% of this (+-60K). I don't really see how the Kelvin version makes sense, but this is an important paper, and to be professional, I know temperature should be in SI, so Kelvin. What should I do?
- Sarah (age 17)
Vienna, Austria

It's a little hard to read that equation in text. What it sounds like is that your equation gives how much T differs from some reference value as a function of how much the frequency differs from the frequency at that reference value. Then the percent errors would be in those temperature differences. However, I may not have read the equation right.

Mike W.

(published on 10/31/2013)