Annaële,
I was unable to remember how to get this directly but, my toy example below
demonstrates one alternative.
DaveT.
#########################################
### the data:
> SDF1
blk rep res
1 1 1 45
2 1 2 98
3 2 1 35
4 2 2 57
5 3 1 84
6 3 2 61
### the model:
> junk <- aov(res~., data=SDF1)
### the output:
> junk
Call:
aov(formula = res ~ ., data = SDF1)
Terms:
blk rep Residuals
Sum of Squares 902 451 1460
Deg. of Freedom 2 1 2
Residual standard error: 27
Estimated effects are balanced
### the available model components:
> attributes(junk)
$names:
[1] "coefficients" "residuals" "fitted.values" "effects" "R"
"rank"
[7] "assign" "df.residual" "contrasts" "terms"
"call"
$class:
[1] "aov" "lm"
$xlevels:
$blk:
[1] "1" "2" "3"
$rep:
[1] "1" "2"
### the desired residual standard error:
> sqrt(sum(resid(junk)^2)/junk$df)
[1] 27
#########################################
DJT: -----Original Message-----
DJT: From: asanquer@virbac.fr [mailto:asanquer@virbac.fr]
DJT: Sent: January 28, 2004 11:44 AM
DJT: To: s-news@wubios.wustl.edu
DJT: Subject: Extract residual standard error form aov
DJT:
DJT:
DJT: Hi,
DJT:
DJT: My question is certainly very simple, but I cannot find
DJT: the solution.
DJT:
DJT: I would like to extract the residual standard error from
DJT: the result of a
DJT: simple analysis of variance performed with the function
DJT: aov. In the example
DJT: below, I would like to extract the value 0.5610507.
DJT:
DJT: > aov(y ~x, data=SDF)
DJT: Call:
DJT: aov(formula = y ~ x, data = SDF)
DJT:
DJT: Terms:
DJT: Diametre Residuals
DJT: Sum of Squares 229.2976 10.7024
DJT: Deg. of Freedom 1 34
DJT:
DJT: Residual standard error: 0.5610507
DJT: Estimated effects may be unbalanced
DJT:
DJT: Thanks,
DJT:
DJT: Annaële Sanquer
|