search

spgarbet / tangram

Table Grammar package for R
66 stars 3 forks source link

multiple comparison correction support #71

Closed conge closed 1 year ago

conge commented 3 years ago

Hi Shawn,

I find this package very handy to create tables for reports, especially it provides statistical tests and reports p-values back.

But my question is, are the reported p-values corrected for the multiple comparison problem? if it is, what method was used?

Thanks,

Yang

spgarbet commented 3 years ago

Assuming you question refers to the default rmisc transform the answer is no correction is applied.

The rmisc transform is our departments recommended exploratory summary of data. The p-values are simply to look for something 'worthy of further investigation'.

Also a transform has no knowledge of the broader number of methods being applied. There is support in the hmisc module for specifying ones own statistical model to generate p-values. You're free to do whatever method you desire.

spgarbet commented 1 year ago

I spoke with Frank Harrell and these p-values have caused more confusion and aren't worthy of reporting in general. He's now recommending that they be removed entirely.