I am a biologist looking for guidance regarding which statistical test to use when comparing variances in data.

I am comparing two theoretical distributions to the experimental data. The data in the distributions may not be normally distributed and there may not be an equal number data points. An example of my data is shown below:

enter image description here

So what I have done so for is to normalise the data on similar y-axis scales. I did this by taking the mean of each of the three distributions and plotting their difference from the mean. The distribution of the data is shown below:

enter image description here

The y-axis is the difference from the mean and the x-axis the type of data.

Using the normalised distribution I have been using F-Tests in R (with command var.test()) to statistically analyse the variance in the data. But the F-Tests assumes a normal distribution.

Is the F-Test the correct test to use in this case? If not can anyone suggest a more appropriate test?

  • $\begingroup$ How are these values from the 'theoretical' distributions obtained? $\endgroup$ – Glen_b Jan 18 '14 at 5:26
  • $\begingroup$ Obtained by computer software which tries to mimic the experimental. The software is a bit of a black box so I cant change any of the parameters. $\endgroup$ – Harpal Jan 19 '14 at 14:58
  • $\begingroup$ Is this for some coursework? $\endgroup$ – Glen_b Jan 19 '14 at 15:33
  • $\begingroup$ nope, research. $\endgroup$ – Harpal Jan 20 '14 at 0:10
  • 1
    $\begingroup$ Then it seems odd that the experimental values are equally spaced. This suggests that your values are not observations from a common distribution, in which case the F-test doesn't seem to be relevant. $\endgroup$ – Glen_b Jan 20 '14 at 0:44

Your Answer

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

Browse other questions tagged or ask your own question.