I am doing my thesis, and I need your help in interpreting the result.

I calculated the median, IQR and range of percentages because the data are not normally distributed. But now I need to compare my result with other results interpreted as (mean) and (CI). How can I do that?

  • 7
    $\begingroup$ I'm sorry but if you expect us to help you more information would be necessary. For instance: What is the research question you are trying to answer? What Kind of data are you using? What variables are you trying to compare and how are they measured? You need to provide further Information. Otherwise we cant help you. Do I understand you correctly that you dont want to use a simple t-test for equality of means because your data is not normally distributed? With kind regards, Alex $\endgroup$ – JustMe Nov 1 '17 at 12:12

In many cases, you can use a t-test to compare means of non-normally distributed data sets, as the mean of a large number of non-normally distributed data is approximately normally distributed:

Why is a sum of skewed left distribution normal distributed according to the central limit theorem

If you do not want to go that route, you would need to make assumptions about the distribution of the data in the other studies. Just because they publish mean and CI does not mean, they found normally distributed values. If, for example, there was reason to assume, those other data were poisson-distributed or any other reasonable assumption, you could simulate a lot of datasets from suitable distributions and compare that with your data. Do you have any helpfull knowledge in that respect?

| 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.