# repeated measures Manova: error E matrix singlar?

#### silkevos

##### New Member
I tried to do a rp MANOVA but for some reason R is saying that it is impossible to do a MANOVA and I should do a ANOVA? The error seens to tell that the E matrix is singlar? Anyone knows how I can fixt this? What did I do wrong or is it really true I cannot do an MANOVA on this dataset?

> ### Longformat data import dataset
subject learning session target present search accuracy
1 1 s one A yes onetarget 0.9923077
2 1 s one A no onetarget 0.9923077
3 1 s two B yes onetarget 0.9769231
4 1 s two B no onetarget 0.9846154
5 1 s three A yes twotargets 1.0000000
6 1 s three A no twotargets 1.0000000

> ########## rp MANOVA
> data.melt <- melt(data_longformat[,-(3:4)] , id=c("subject","learning","present","search"), measured=c("accuracy"))
> data.melt
subject learning present search variable value
1 1 s yes onetarget accuracy 0.9923077
2 1 s no onetarget accuracy 0.9923077
3 1 s yes onetarget accuracy 0.9769231
4 1 s no onetarget accuracy 0.9846154
5 1 s yes twotargets accuracy 1.0000000
6 1 s no twotargets accuracy 1.0000000

> ######### from long to wide data (Analyses)
> data.wide <- cast(subject +learning ~ present + search, data=data.melt)
Aggregation requires fun.aggregate: length used as default
> data.wide
subject learning no_onetarget no_twotargets yes_onetarget yes_twotargets
1 1 s 4 2 4 2
2 2 d 2 6 2 6
3 3 s 4 2 4 2
4 4 d 2 6 2 6
5 5 s 4 2 4 2
6 6 d 2 6 2 6
7 7 s 4 2 4 2
8 8 d 2 6 2 6
9 9 s 4 2 4 2
10 10 d 2 6 2 6
11 11 s 4 2 4 2
12 12 d 2 6 2 6
13 13 s 4 2 4 2
14 14 d 2 6 2 6
15 16 d 2 6 2 6
16 17 s 4 2 4 2
> ## all tests at onces
> present <- factor(c(1,1,2,2))
> search <- factor(c(1,2,1,2))
> idata <- data.frame(present, search)
> idata
present search
1 1 1
2 1 2
3 2 1
4 2 2
> fit <- lm(cbind(no_onetarget ,no_twotargets ,yes_onetarget ,yes_twotargets )~. , data=data.wide)
> Anova(fit, type="III", test= "Wilks", idata=idata, idesign= ~present*search)
Error in print.Anova.mlm(list(SSP = list((Intercept) = 741.152141802068, :
singular error SSP matrix; multivariate tests unavailable
try summary(object, multivariate=FALSE)