# ANCOVA with repeated measures in R

I am a novice to program R and have been trying to perform a repeated measures ANCOVA with Temperature as the dependent variable, Site as the independent variable, Date as the covariate and Year as the repeated measures.

My dataset consists of temperatures from 4 sites, over 20 days, during 2 different years. There appears to be a significant effect of site, date and year on temperature, but no significant interaction between site and year or site and date. I originally performed two separate ANCOVAs, but my coauthor wants them in a single model. As far as I can tell, there is no simple way to do a repeated measures ANCOVA as there is for ANOVA.

I have tried using: lm(formula = Temp~Site*Date + Year) , but wouldn't this just put year as an interaction that doesn't affect the model?

-
1) From the text it sounds like you want to fit the Date:Year interaction, but not Site:Year or Site:Date; however, the lm code has the Site:Date interaction but not the others. 2) Why do you think there's no Site:Year interaction if you've only fit two separate ANCOVAs? – Aaron Nov 28 '11 at 15:37
Also, sample data can greatly help in describing your setup and comparing possible solutions. I've posted an example as an answer; consider modifying (if necessary) and editing your answer to include it. – Aaron Nov 28 '11 at 15:46

There is a list of tutorials on this subject here:

http://www.r-statistics.com/2010/04/repeated-measures-anova-with-r-tutorials/

Good luck.

-

Sample data?

set.seed(5)
d <- expand.grid(Site=LETTERS[1:4], Date=1:20, Year=factor(1:2))
d\$Temp <- round(rnorm(nrow(d), mean=60, sd=15))

-

Maybe you can try

lm(formula = Temp~Site*(Date + Year))


In this way you will have two interactions with the Site, and there will be no interaction between Date and year

-