Dear all,
I have follwing factors in my data set (G)
names(G)
relative ht, biomass, treatment, variety, soil group
relative height
biomass
treatment (4 levels)- control+lime, control+ash, contaminated soil+lime,
contaminated soil+ash
variety (4 levels)- Salix swherinii, Salix mursinifolia, Klara, Karin
soilgroup (2 level)- control and contaminated
if I fit aov function this is ok
anova1<-aov(ht~treatment*variety,data=G)
summary(anova1)
R- output
Df Sum Sq Mean Sq F value
Pr(>F)
treatment 3 14015 4672 65.021 < 2e-16
***
variety 3 5883 1961 27.295
2.95e-12 ***
treatment:variety 9 5474 608 8.465 8.93e-09 ***
Residuals 80 5748 72
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05
'.' 0.1 ' ' 1
above I can see my treatment and variety interaction effect
anova3<-aov(ht~treatment*variety*soilgroup,data=G)
summary(anova3)
R-ouput
Df Sum Sq Mean Sq F value
Pr(>F)
treatment 3 14015 4672 65.021 < 2e-16
***
variety 3 5883 1961 27.295
2.95e-12 ***
treatment:variety 9 5474 608 8.465 8.93e-09 ***
soilgroup -missing
treatment:variety:soilgroup - missing
Here, I am missing value of soil group and interation of treatment: variety:
soilgroup- Can any one please tell me why?
Thanks for help
Best regards
Salam
[[alternative HTML version deleted]]
I suspect that your "treatment" levels are perfectly nested inside the "soilgroup" levels. If this were the case you can either use one of the factor in your analysis or the other, depending on what you want to analyze. Il giorno Wed, 11 Feb 2015 14:47:18 +0000 Mir Salam <mir.salam at uef.fi> ha scritto:> Dear all, > >
> R-ouput > > Df Sum Sq Mean Sq F value Pr(>F) > treatment 3 14015 4672 65.021 < 2e-16 *** > variety 3 5883 1961 27.295 2.95e-12 *** > treatment:variety 9 5474 608 8.465 8.93e-09 *** > soilgroup -missing > treatment:variety:soilgroup - missing > Here, I am missing value of soil group and interation of treatment: variety: > soilgroup- Can any one please tell me why?Do the lines> soilgroup -missing > treatment:variety:soilgroup - missingActually appear in your output? If so, I have no idea. But if the terms are simply absent, one possibility is that soilgroup is completely confounded with an earlier factor in the model and therefore unidentifiable. aov seems fairly tolerant of that in one sense - instead of throwing an error and stopping it tells you what it can identify and leaves out anything it can't. S Ellison ******************************************************************* This email and any attachments are confidential. Any use...{{dropped:8}}