Statistical tests and data type

[miaoj5]

I have data in log2 format. I just came across some article which says that logged data should be avoided for statistical tests like T-tests, F-tests or Mann-Whitney tests.

I am doing these tests in excel so wondering if anybody who performs such tests regularly in excel can flash some light??

[Jerry W. Lewis]

What was the context of that statement? If the data were lognormally
distributed, taking logs before doing these tests would be mandatory.

Jerry

miaoj5 wrote:

> I have data in log2 format. I just came across some article which says
> that logged data should be avoided for statistical tests like T-tests,
> F-tests or Mann-Whitney tests.
>
> I am doing these tests in excel so wondering if anybody who performs
> such tests regularly in excel can flash some light??

[miaoj5]

Actually, the data that I am working with is a microarray data and I have log2 transformed as well as natural data. Log 2 tranformation was used for better normalisation in the data series.
Is this what you wanted to know?

The article that I talked about was explaning these tests in general