I wonder if anyone can help me with this question.
When conducting an ANCOVA (using pre-test as covariate) on the results of a
reading test (experimental and control groups), I find no significant
differences. HOWEVER, when I compute gain scores (post-test - pre-test) and then
conduct an ANOVA on the gain scores, the difference between the groups DOES
attain significance. Can anyone explain to me, both conceptually and
statistically, why this is so? I have heard "there are problems" with gain
scores but no one has been able to articulate for me what these are exactly.
Thanks!
------------------------------
Mark R. Gover
2312 Barritt Street
Lansing, MI 48912
http://www.msu.edu/user/govermar
Phone: 517.371.1997
Fax: 1-630-604-4977
email: [log in to unmask]
ICQ: 46928284
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|