Bioequivalence and Bioavailability Forum

Let GLM do the work [General Sta­tis­tics]

posted by d_labes  – Berlin, Germany, 2012-09-11 10:34 (5027 d 13:03 ago) – Posting: # 9175
Views: 7,422

Hi John!

❝ ... Shouldn't the CI be computed based on the following formula?

❝

❝ lower and upper CI = 100*exp(difference ± t 0.05,&df × SE difference).

This is done by Proc GLM for you (if you specify /CLPARM alpha=0.1 on the model statement). Without bothering to remember the correct formula for obtaining the confidence limits .

BTW: If your code was similar to mine, what was the problem?

Complete thread:

• 4-way crossover vs 2-way crossover jag009 2012-09-06 17:23 
  • Principle SAS code for more than 2 tmts d_labes 2012-09-07 12:05
    • Principle SAS code for more than 2 tmts jag009 2012-09-10 22:23
      • Let GLM do the workd_labes 2012-09-11 08:34
        • Let GLM do the work jag009 2012-09-11 16:46
          • Don’t worry about lenghty posts Helmut 2012-09-11 16:53
            • Don’t worry about lenghty posts jag009 2012-09-11 21:07
              • off-topic Helmut 2012-09-12 14:49
                • off-topic jag009 2012-09-13 18:23
        • Let GLM do the work jag009 2012-09-13 19:49