# ANOVA claims that model is singular

I have a problem with a 3 way ANOVA with repeated measures. The design is the following: "localization" is the dependent variable, than I have the following within subjects factors "material" (4 levels), "diffusion technique" (6 levels) and "typology" (2 levels).

This is the formula I used:

aov_Material = aov(abs(Localization) ~ Diffusion_Type*Material*Typology +
Error(Subject/(Diffusion_Type*Material*Typology)), data=scrd)
summary(aov_Material)


And this is the output I get as result:

Warning message:
In aov(abs(Localization) ~ Diffusion_Type * Material * Typology +  :
Error() model is singular
> summary(aov_Material)

Error: Subject
Df Sum Sq Mean Sq F value Pr(>F)
Residuals 11   2117   192.5

Error: Subject:Diffusion_Type
Df Sum Sq Mean Sq F value Pr(>F)
Diffusion_Type  5   20.2   4.032   0.423   0.83
Residuals      55  523.7   9.522

Error: Subject:Material
Df Sum Sq Mean Sq F value   Pr(>F)
Material   3   1967   655.6    21.4 7.03e-08 ***
Residuals 33   1011    30.6
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Error: Subject:Diffusion_Type:Material
Df Sum Sq Mean Sq F value Pr(>F)
Diffusion_Type:Material  15  170.1  11.338   1.308  0.202
Residuals               165 1430.3   8.668

Error: Within
Df Sum Sq Mean Sq F value Pr(>F)
Residuals 288   2888   10.03


My questions are:

1. Why there is the warning (and, BTW, what does it means that the model is singular)?
2. Why there is not the effect for typology listed in the results? How can I understand if there is a difference between the two levels of typology?
• Sorry for migrating it; it is on-topic here. – user88 Feb 3 '13 at 0:12
• hello...can anyone help me please? – L_T Feb 3 '13 at 10:46