6

I would like to integrate T-statistics values from multiple, independent tests.

How can this be done?

Since my degrees of freedom are many ($n>1000$) I was thinking to use the Stouffer's method, which is really for Z-statistics score. However, with a high $df$ value, the T distribution and the Gaussian distribution should be identical.

Can I use Stouffer's method to integrate T values at $df>1000$?

Chill2Macht
  • 5,639
  • 4
  • 25
  • 51
Federico Giorgi
  • 189
  • 1
  • 11
  • 2
    See [this answer](http://stats.stackexchange.com/questions/136415/is-there-a-t-test-equivalent-to-stouffers-z-test?rq=1) for the formula – mdewey Aug 23 '16 at 08:50

0 Answers0