|
|
Different group sizes multiple group |
|
Message/Author |
|
Gerine L posted on Thursday, June 04, 2015 - 8:34 am
|
|
|
I have a very simple model: Y = b0 + b1 X. Both Y and X are scale variables and non-normally distributed (thus I use the MLR estimator). I want to examine whether results for this model differ between group A (young participants, around 1100) and group B (old participants, around 175). I want to test whether differences occur in b0 and b1. Can I reliably conclude that X is related to Y for group B but not for group A, if my sample size for group A is so much smaller than for group B? |
|
|
Having very different group sizes will affect the results. The small group has less power. There is noting you can do to compensate for this. |
|
Back to top |
|
|