0
$\begingroup$

I have completed a series of Mann Whitney U tests in R, and am looking for the degrees freedom.

I'm comparing two datasets of n1=29 and n2=5, with StandardDev1 0.4525 and StandardDev2 0.3652.

What can I use from these to calculate DF? Is there a way to do it? I've found this website:

http://web.utk.edu/~cwiek/TwoSampleDoF

But I'm not sure of its methods or applicability.

Many thanks, Chris

Improve this question
$\endgroup$

1 Answer 1

Reset to default
1
$\begingroup$

As there are no parameter values being estimated from the data in the non-parametric Mann-Whitney test, it doesn't really involve degrees of freedom in the same way as t-tests and other parametric tests do.

What you should do is report the number of cases in each dataset, some central measure (typically the median) for each dataset, the value of the U statistic, and the p-value. See the Wikipedia page for an example of how to proceed. With so few data points, a dotplot directly comparing the values in the 2 datasets could be quite helpful to the reader. With only 5 cases in one dataset, a skeptical reader might otherwise think that your results are being affected by outliers.

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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