I'm trying to do an ANCOVA with two factors (clipping treatment with two
levels, and plot with 4 levels) and a covariate (stem diameter). The
response variable is fruit number. The minimal adequate model looks like
this:
model3<-lm(fruit~clip + plot + st.dia + clip:plot)
I'd like to get some multiple comparisons like the ones from TukeyHSD, but
TukeyHSD doesn't work with the covariate. I've tried using glht() in the
multcomp package, but I'm not sure how to get it to give me the TukeyHSD for
all the interactions (i.e. clipped:plotA vs. unclipped:plotA, etc). It
seems as if I can only specify one factor or the other, and it gives me a
warning about there being interactions.
glht(model3, linfct=mcp(plot="Tukey"))
General Linear Hypotheses
Multiple Comparisons of Means: Tukey Contrasts
Linear Hypotheses:
Estimate
151Nat - 151Trans == 0 -1.924
HC - 151Trans == 0 -7.942
HE - 151Trans == 0 -4.637
HC - 151Nat == 0 -6.018
HE - 151Nat == 0 -2.712
HE - HC == 0 3.305
Warning message:
In mcp2matrix(model, linfct = linfct) :
covariate interactions found -- default contrast might be inappropriate
How can I get it to give me the same sort of output that TukeyHSD gives with
just a regular two-way ANOVA?
Thanks for your help
-Eric Scott
[[alternative HTML version deleted]]
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
