[R] Nested ANOVA residuals error

Daniel Malter daniel at umd.edu
Thu Jul 16 22:18:07 CEST 2009


It is because the nesting structure perfectly explains the data (i.e., there
is only one observation and, therefore, no variation for each Ind in each
Treatment).

e=rnorm(90,0,1)
x=rep(1:3,30)
y=rep(1:30,each=3)
z=x+y+e

ano=aov(z~factor(y)/factor(x))
ano
residuals(ano)

Best,
Daniel


-------------------------
cuncta stricte discussurus
-------------------------

-----Ursprüngliche Nachricht-----
Von: r-help-bounces at r-project.org [mailto:r-help-bounces at r-project.org] Im
Auftrag von Jeff DaCosta
Gesendet: Thursday, July 16, 2009 3:37 PM
An: r-help at r-project.org
Betreff: [R] Nested ANOVA residuals error

I am having trouble setting up a nested anova model.  Here is a truncated
version of my data:

 > data
      Ind Treatment         PC1
1  PER14        SC  1.14105282
2  PER14        SH  1.45348615
3  PER14        AC  2.45096904
4  PER25        SC  1.23687887
5  PER25        SH  4.54797450
6  PER25        AC -0.73391131
7  PER30        SC  1.15899433
8  PER30        SH  2.83273736
9  PER30        AC -0.86272186
10 PER31        SC  1.90495285
11 PER31        SH  5.68514429
12 PER31        AC -2.25110314
13 PER35        SC  2.16463279
14 PER35        SH  1.69053389
15 PER35        AC  1.87528942
16 PER36        SC -2.44795781
17 PER36        SH -3.59415298
18 PER36        AC  1.90756192
19 PER41        SC -2.83086573
20 PER41        SH  4.26803170
21 PER41        AC  2.42548862
22 PER48        SC -1.86403368
23 PER48        SH -0.08713013
24 PER48        AC -1.98056853
25 PER52        SC  2.35012436
26 PER52        SH -2.19384500
27 PER52        AC  1.38686223

 > str(data)
'data.frame':	27 obs. of  3 variables:
  $ Ind      : Factor w/ 9 levels "PER14","PER25",..: 1 1 1 2 2 2 3 3  
3 4 ...
  $ Treatment: Factor w/ 3 levels "AC","SC","SH": 2 3 1 2 3 1 2 3 1
2 ...
  $ PC1      : num  1.14 1.45 2.45 1.24 4.55 ...

PC1 is my response variable, and Treatment is my factor with three levels.
Each  of the nine individuals (Ind) in the study was tested for all three
factor levels, so I want to run the model with Ind nested in Treatment.  I
think that I want to set up the model like this:

 > m1<-aov(PC1~Treatment/Ind))

but I get the following error with anova(m1), and am not sure why I get zero
for all the residuals.
 > anova(m1)
Analysis of Variance Table

Response: PC1
               Df  Sum Sq Mean Sq F value Pr(>F)
Treatment      2   9.215   4.607
Treatment:Ind 24 142.140   5.923
Residuals      0   0.000
Warning message:
In anova.lm(m1) :
   ANOVA F-tests on an essentially perfect fit are unreliable

Any input would be appreciated, thanks for your time and consideration.
-Jeff
	[[alternative HTML version deleted]]

______________________________________________
R-help at 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.




More information about the R-help mailing list