I am studying mixed models and have a doubt about nested random effects..
In the example, we tested subjects variable X and outcome Y and want to see if X is correlated with Y. The data were collected on 8 different mountains in 3 different site (ie mountain A was collected on site 1 2 and 3, mountain B site 1 2 and 3, etc). Site 1 of mountain A is not related at all with site 1 of mountain B. Then we want to analyse the relationship Y ~ X controlling for mountain and site.
Since that site and mountain are nested, we created a variable "sample" that is
sample <- factor(mountainRange:site)*
Now, here is my doubt. The course suggest to do the analysis as following
mixed.lmer2 <- lmer(Y~ X + (1|mountain) + (1|sample), data = data)
But I don't understand why we include (1|mountain)
. I thought I don't need to include it as it is already included in the variable (1|sample)
What I mean is that I would run the analysis as following:
mixed.lmer3 <- lmer(Y~ X + (1|sample), data = data)
Could anyone explain why am I wrong? Thanks
You simply embedded mountain
and site
within the same term sample
, which makes it nested. Alternatively, you could have done
mixed.lmer2 <- lmer(Y~ X + (1|mountain/site), data = data)
OR
mixed.lmer2 <- lmer(Y~ X + (1|mountain) + (1|mountain:site), data = data)
to represent the same thing. Again, you just did it outside the model formula.
Do see https://stats.stackexchange.com/q/228800/238878 for a more thorough answer.
The technical post webpages of this site follow the CC BY-SA 4.0 protocol. If you need to reprint, please indicate the site URL or the original address.Any question please contact:yoyou2525@163.com.