q**t
发帖数: 36 | 1 Hi All,
I am using lme in R to estimate standard error the random effect.
The output of the StdDev of lme is the SE of the random effects?
lme(y~1, random = ~1| s/p/k )
Random effects:
Formula: ~1 | s
(Intercept)
StdDev: 0.02100801
Formula: ~1 | p %in% s
(Intercept)
StdDev: 0.001980935
Formula: ~1 | k %in% p %in% s
(Intercept) Residual
StdDev: 0.01581869 0.006076241
Number of Observations: 240
Number of Groups:
s
5
p %in% s
10
k %in% p %in% s
240
or I have to find the degree freedom of each groups stdDev/df , where df=(5-
1)/(10-1)?(240-1)?
Waiting online for response, Thanks very much!!! | c***z
发帖数: 6348 | | a******r
发帖数: 706 | 3 Std.Dev are the estimates of the random effects variance (the \sigma^2's).
You do need to divide them by the corresponding df to make proper inference.
Check any introductory linear model or anova book, you'll find the answer
about how to find the df. |
|
