2

I've the following data format

Entity Measure1 Measure2
A1     1.5      1.4
A1     1.9      1.7
A1     1.8      1.5
B1     2.7      2.5
B1     3.0      2.9
B1     3.1      2.7
B1     2.8      2.9
C1     1.7      1.5
C1     1.7      1.8
...

I wish to do a pairwise T-test for the Entity factor and at Measure1 and Measure2 levels. I can do this in R but is there any way to do in R Commander GUI? In R Commander the paired t-test considers all the values of Measure1 and Measure2 columns and doesn't perform the t-test groupwise. The 'independent sample t-test' option is grayed out too.

Ram
  • 3,092
  • 10
  • 40
  • 56
WoA
  • 173
  • 2
  • 2
  • 12

1 Answers1

3

If you abandoned the paired-t.test formalism you could do this with regression:

 lm( Measure2 ~ Measure1 + Entity, data=dfrm)

This will give you an estimate of mean increase in Measure2 (from Measure 1) by group. Thats eally what paired t-tests are doing, anyway.

IRTFM
  • 258,963
  • 21
  • 364
  • 487