Displaying 1 result from an estimated 1 matches for "otrend".
Did you mean:
trend
2010 Jan 12
0
post-hoc after ancova
...n the dependent
variable.
I would like to do a post-hoc test to see which groups in the categorical
variable differ.
I have explored Tukey test in multcomp package. My study is similar to the
"litter data". In the code it's mentioned that the contrast matrix also has
some trends like otrend, atrend and ltrend.
otrend = c(-1.5, -0.5, 0.5, 1.5),
atrend = doselev - mean(doselev),
ltrend = log(1:4) - mean(log(1:4)))
Here are my questions:
Are this trends absolutely essential for conducting the Tukey test?
If yes, how can I set these trends?
thanks,
mahua
[[alternative HTML version...