[R] nlme versus aov with Error()

[Sven Garbade]

Hi all,

I computed a two way anova with two within-subjects fixed effects: diffi
(2 levels) and cond (4 levels) and one random effect subject, and one
dependent variable accuracy:

summary(aov(accuracy ~ diffi*cond + Error(sub/(diffi+cond)),
data=phd.df))

Error: sub
          Df  Sum Sq Mean Sq F value Pr(>F)
Residuals  4 136.549  34.137

Error: sub:diffi
          Df  Sum Sq Mean Sq F value    Pr(>F)
diffi      1 1321.12 1321.12  108.61 0.0004788 ***
Residuals  4   48.65   12.16
---
Signif. codes:  0 `***' 0.001 `**' 0.01 `*' 0.05 `.' 0.1 ` ' 1

Error: sub:cond
          Df  Sum Sq Mean Sq F value    Pr(>F)
cond       3 1952.62  650.87  14.295 0.0002887 ***
Residuals 12  546.37   45.53
---
Signif. codes:  0 `***' 0.001 `**' 0.01 `*' 0.05 `.' 0.1 ` ' 1

Error: Within
           Df  Sum Sq Mean Sq F value   Pr(>F)
diffi:cond  3 113.692  37.897  9.2825 0.001884 **
Residuals  12  48.992   4.083
---
Signif. codes:  0 `***' 0.001 `**' 0.01 `*' 0.05 `.' 0.1 ` ' 1

But when I did this with lme() from package lme, I got a different
result:

phd.lme <- lme(accuracy ~ cond*diffi, random=~1|sub, data=phd.df)
anova(phd.lme)
            numDF denDF  F-value p-value
(Intercept)     1    28 6759.188  <.0001
cond            3    28   28.298  <.0001
diffi           1    28   57.438  <.0001
cond:diffi      3    28    1.648  0.2009

I think I've missed something in the random argument, but can't figure
out what.

Thanks for any help, Sven

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._