R help - Nov 2009 - Help with multiple comparisons on a 2-way repeated measures ANOVA

Hi everyone,

I'm trying to do a 2-way repeated measures ANOVA with data that looks like
this:
subject block rep  day  light   response
       1     1       1    one    L1      5.5
       2     1       2    one    L1      4.5
       3     1       1    one    L2      4
       4     1       2    one    L2      5.1
       5     2       1    one    L1      5.3
       6     2       2    one    L1      4.8
       7     2       1    one    L2      4.6
       8     2       2    one    L2      4.9
       1     1       1    two    L1      5.2
       2     1       2    two    L1      5.0
       3     1       1    two    L2      3.2
       4     1       2    two    L2      3.4
       5     2       1    two    L1      4.6
       6     2       2    two    L1      4.9
       7     2       1    two    L2      3.3
       8     2       2    two    L2      2.9


I'm using lme():
m1<-light(buoy~light*day, random=~1|subject/day)
summary(m1)

and get this output:
Linear mixed-effects model fit by REML
 Data: NULL
       AIC      BIC    logLik
  24.94574 28.34008 -5.472868

Random effects:
 Formula: ~1 | subject
        (Intercept)
StdDev:  0.05400619

 Formula: ~1 | day %in% subject
                (Intercept)    Residual
StdDev:   0.2912773    0.06415494

Fixed effects: response ~ light * day
                        Value  Std.Error    DF  t-value       p-value
(Intercept)         5.025  0.1515545  6   33.15640   0.0000
lightL2              -0.200  0.2143304  6   -0.93314   0.3868
daytwo             -0.100  0.2109008  6   -0.47416   0.6522
lightL2:daytwo  -1.525  0.2982588  6   -5.11301   0.0022
 Correlation:
                        (Intr)     lghtL2   daytwo
lightL2             -0.707
daytwo             -0.696  0.492
lightL2:daytwo  0.492  -0.696   -0.707

Standardized Within-Group Residuals:
              Min                Q1              Med                Q3
       Max
-0.36863373 -0.15741132  0.03092751  0.14569214  0.32591770

Number of Observations: 16
Number of Groups:
         subject day   %in% subject
                    8                16

My FIRST problem is that I'm not quite sure how to analyse the above output.

My SECOND problem is when I try to use glht(), to get multiple comparisons:
summary(glht(m1,linfct=mcp(day="Tukey")))

I get an error message:
Error in as.vector(x, mode) : invalid 'mode' argument

I"m not sure where this comes from. I've tried everything I (and my
labmates, we are a bunch of field biologists, not statisticians) can think
of, but this keeps coming back.  Is it because glht() isn't appropriate for
2-way ANOVAs?  If that's the problem, what is the correct way to do this?
Or am I going about this completely wrong from the get-go?

Thanks for all the help you can provide!
K

	[[alternative HTML version deleted]]

install.packages("HH")

library(HH)

?MMC

Go to the bottom of the page and read the maiz example.
Read through the example that shows aovlist objects don't work
and how to construct the equivalent aov object.