Bioequivalence and Bioavailability Forum

Not to round is the question [General Sta­tis­tics]

posted by d_labes  – Berlin, Germany, 2011-09-22 11:11 (5380 d 13:28 ago) – Posting: # 7385
Views: 19,714

My dear captain,

where is your problem ?

All serious statistical software I'm aware of have functions for the t-quantile which accepts non-integer degrees of freedom.

So why round/truncate the df's in any way? If you can get serious statistical software for free (R project)?

We know that the Welch-Satterthwaite procedure is an approximate solution to the problem of testing means with unequal variances. Do we need an approximation to the approximation?

BTW: Seems Martin's theorem "Never never never never use Excel. Not even for calculation of arithmetic means." may be extended to OpenOffice Calc ?

Complete thread:

• Welch/Satterthwaite in practice ElMaestro 2011-09-21 17:04 
  • Not to round is the questiond_labes 2011-09-22 09:11
    • my problem ElMaestro 2011-09-22 12:34
      • Ways out d_labes 2011-09-27 11:54
  • Welch/Satterthwaite in practice yicaoting 2011-09-29 16:29
    • Welch/Satterthwaite in practice ElMaestro 2011-09-29 17:39
    • Spreadshit addiction Helmut 2011-09-30 16:52
      • what a divine post! ElMaestro 2011-09-30 18:54
        • what a divine post! Helmut 2011-09-30 19:41
          • what a divine post! ElMaestro 2011-10-01 13:03
            • what a divine post! Helmut 2011-10-01 13:20
      • Spreadshit addiction yicaoting 2011-10-01 05:36
        • Spreadshit addiction ElMaestro 2011-10-01 11:13
          • GNU addiction Helmut 2011-10-02 03:20
        • Spreadshit addiction Helmut 2011-10-01 13:03
          • Spreadshit addiction yicaoting 2011-10-01 16:39
          • Spreadshit addiction yicaoting 2011-10-03 07:23
            • Spreadshit addiction Helmut 2011-10-03 15:59
              • Spreadshit addiction yicaoting 2011-10-03 17:35
                • Spreadshit addiction Helmut 2011-10-03 17:41