I have a set of temperature datasets all from different products. I wish to calculate the variation in the products. I have tried using coefficient of variation. But because the temperature data varies from say 20-40 deg C, the mean would be high and standard deviation low, the coefficient of variation is very low. I wish to know an alternative in this case to give a proper representation.

It would be great if you can help.

Ex. 27.2 30.4 30.4 31.0 31.9 35.2
SD=2.6; CV=8.4

  • $\begingroup$ Dear Keerthi, maybe you can explain why you think the fact that the coefficient of variation is low is a problem? Also, usually, the coefficient of variation is defined as $\dfrac{\sigma}{\mu}$, so in your example would be $\dfrac{2.6}{31} = 0.083$. For some reason, you seem to use that value multiplied by 100. Also, the wikipedia page on coefficient of variation cautions using them with interval data, the example use temperatures. $\endgroup$ – Antoine Vernet Aug 19 '16 at 8:09
  • $\begingroup$ Thank you Vernet! yes i am multiplying it by 100 to just give it as a percentage. I have used the same method to calculate the CV for precipitation as well.. and in my case the CV for precipitation should be lower than that for temperature. Which I am unable to achieve. $\endgroup$ – Keerthi Aug 20 '16 at 3:47

Please refer to CV wiki - https://en.wikipedia.org/wiki/Coefficient_of_variation#Alternative

First, CV may not be relevant for Celsius scale as it takes both positive & negative values. You can convert it to Kelvin & compute.

There are alternatives provided on the same page

| cite | improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.