Michael Eisenring <eimic...@ethz.ch> wrote on 07/20/2012 09:35:03 AM:
> Dear Jean,
> thanks for this email. I'm grateful for this instruction Just to
> make sure that I understood you properly: something like this?:
Michael,
Yes, this is perfect. Very helpful.
> My data:
>
> dput(geo_anova_nested_input):
>
> structure(list(Site = c(1L, 1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L,
> 2L, 2L, 3L, 3L, 3L, 3L, 3L, 3L, 4L, 4L, 4L, 4L, 4L, 4L, 5L, 5L,
> 5L, 5L, 5L, 5L, 6L, 6L, 6L, 6L, 6L, 6L, 7L, 7L, 7L, 7L, 7L, 7L,
> 8L, 8L, 8L, 8L, 8L, 8L), Night = 1:48, Species = c(21L, 22L,
> 13L, 12L, 26L, 12L, 28L, 22L, 25L, 27L, 27L, 10L, 21L, 31L, 33L,
> 36L, 32L, 17L, 31L, 26L, 27L, 38L, 32L, 14L, 27L, 36L, 37L, 40L,
> 38L, 5L, 20L, 31L, 29L, 36L, 30L, 30L, 18L, 19L, 16L, 43L, 31L,
> 27L, 24L, 27L, 28L, 49L, 34L, 30L)), .Names = c("Site", "Night",
> "Species"), class = "data.frame", row.names = c(NA, -48L))
>
> RCode
>
> attach(geo_anova_nested_input)
> geo_anova_nested_input
> data<-anova(lm(Species~Site/Night))
> TukeyHSD(data,"Site", conf.level=0.90)
There are a number of issues with the line of code data<-anova...
First of all, look at the output from
lm(Species~Site/Night)
You are fitting a linear regression model with Site and Site x Night
interaction as the two independent continuous variables. You have no
categorical variables. I assume that you want Site to be a categorical
variable (it represents forests, right?). I'm not sure how you want to
deal with Night. You may want to talk over your design with a
statistician to be sure the model you are using is appropriate for the way
your data were collected.
Second, look at the output from
anova(lm(Species~Site/Night))
This is simply an ANOVA table for the regression model you just fit. But
TukeyHSD(), in your next line of code, wants a "... fitted model object,
usually an aov fit." This is clearly stated in the help page for
TukeyHSD().
?TukeyHSD
So, here's an example of some code that "works" (gives output without
warning/error messages) using your data. I have ignored Night, since I'm
not sure how it should be dealt with (again, talk to a statistician). But,
hopefully this will help you to understand how to make things work in R.
# create a new variable for Site, as a categorical variable (aka
factor)
geo_anova_nested_input$Sitef <-
as.factor(geo_anova_nested_input$Site)
# fit an ANOVA model to the data
fit <- aov(Species ~ Sitef, data=geo_anova_nested_input)
# compute Tukey honest significant differences for the Site factor
TukeyHSD(fit, "Sitef", conf.level=0.90)
Jean
> Michael,
>
> Use dput() to output your data (or perhaps a small subset). Then
> paste the result of that call and your R code (just those lines of
> code that are needed to reproduce the problem) right in your message
> to R-help. That makes it easier for the R-help list readers to help
> you troubleshoot.
>
> Jean
>
>
> Michael Eisenring <eimic...@ethz.ch> wrote on 07/20/2012 05:15:32 AM:
>
> > Dear r-help members.
> > I would like to compare species numbers of moths between eight
> > different forests (each sampled for six nights). I would like to do
> > a nested anova to compare species numbers between forests and nights.
> > For more site specific details I wanted to do a Tukey test
> > (TukeyHSD). Unfortunately this test is not working (message: "nicht
> > anwendbare Methode für 'TukeyHSD' auf Objekt der Klasse "c('anova',
> > 'data.frame" ) /(message: "method not appropriate for TukeyHSD for
> > objects of the class "anova", "data.frame").
> > I tried it also with the "aov" instead of the "anova" function but
> > neither the "aov" nor the "anova" approach is working.
> > Could anyone help me with this problem?
> > Attached you can find the txt.file and the r-script.
> > I'm working with a MAC OSX snow leopard, with R-Studio 0.96 316.
> >
> > Thank you very much.
> > Michael
[[alternative HTML version deleted]]
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
