Tell me more ×
Cross Validated is a question and answer site for statisticians, data analysts, data miners and data visualization experts. It's 100% free, no registration required.
up vote 2 down vote favorite

Suppose A is a fixed factor and B is a random factor. A and B are crossed. Another random factor C is nested in B. Taking into account the AB and AC interactions, how should I fit the model using lme function in the nlme package? Thanks.

share|improve this question
You mention the lme package on the title, and nlme on your question. Which is the correct package? – João Daniel Apr 22 '12 at 22:07
1  
AND .... where's the code? – DWin Apr 22 '12 at 23:22
I said the lme function in the nlme package. lme is not a package. And I'm just wondering how to specify the arguments in this function for my problem. – Knightgu Apr 26 '12 at 19:59
Have you tried lmer() in the lme4 package? See e.g. this book chapter – guest May 7 '12 at 0:05
Thank you so much. Yes, the lmer() function can deal with crossed random effects easily. That is why the lme4 package is a big improvement for the nlme package. I realized later that it was really hard to use lme() function to fit crossed random effects. Since we've got the newer package, why not use it？ – Knightgu May 16 '12 at 22:27

migrated from stackoverflow.com May 6 '12 at 18:22

1 Answer

up vote 3 down vote

The tricky part here is to specify the right formula.

library(nlme)
# Assume y is your response variable.
lme(fixed = y ~ A, random = ~ A | B/C)

Here B/C stands for C nested in B and is short for B + B:C. The interaction with A is present because A is repeated in the random part.

For a basic introduction to nlme, I recommend this document.

share|improve this answer
Thank you very much for your solution. I think you are right. I will use a practical example to check it. By the way, really nice reference. Thank you. – Knightgu Jun 4 '12 at 16:27
1  
I have tried the formula you mentioned using R. Unfortunately, the data set I used did not justify the formula. The lme() function applied simply did not converge to a numerical solution. Seems that the lmer() function in the lme4 package does the right thing. – Knightgu Jun 15 '12 at 21:34

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.