COOCHE@QUCDN.BITNET (Evan G. Cooch) (01/15/90)
Using SAS in ANCOVA, it is fairly straightforward to determine significance of
heterogeneity of slopes.
PROC GLM;
CLASS A;
MODEL Y=A X A*X;
If the interaction term (A*X) is significant, then the slope of the
relationship between Y and X differs for different levels of A.
It is also fairly straightforward to have PROC GLM output the estimated linear
equations for each level of A (using the ESTIMATE option). But what I want to
be able to do is ask the question "which slope, or subset of slopes, is
significantly different from the rest". I.E I am looking for an a posteriori
pairwise comparison of slopes for each pair of levels of A. Zar desscribes a
number of proceudres for doing this (p. 302-303 2nd edition), but this would
involve doing things by hand, which is an affrontery to my personna which
demands that everything be done on the computer! :-)
Does anyone have SAS code out there to do what I described? Pairwise
comparisons of intercepts I've managed (using the CONTRAST statement in GLM).
Can this be easily (and correctly) extended to comparing slopes?
Please respond rpivately, and I will summarize if desired.
Thanks...
_///_
-. .-
( L ) \//// +---------------------------+
--(\=/)-- ( ) | Evan G. Cooch |
/ \ / / | Department of Biology |
I V / | Queen's University |
I / | Kingston, Ontario, Canada |
I / | (613-545-6129) |
+---------------------------+