Repeated measures ANOVA for Latin Square designs

[Saoirse Connor Desai]

Following on from my post about using mixed effects for Latin square designs, I was also wondering if anyone had used aov_ez for a Latin square design.

It doesn’t appear to be possible (presumably because participants didn’t all see the same thing and therefore there aren’t an equal number of data points for each combination).

Does anyone know of a way to get this to work for Latin square design experiments? I have trawled the internet for solutions but am at a loss.

[henrik]

In general the afex ANOVA functions are for fully factorial designs. If your design is not fully factorial, I guess it is not possible to use the afex ANOVA functions.

For analyzing completely replicated latin square designs in R see: https://stat.ethz.ch/pipermail/r-help/2003-April/031783.html

[Saoirse Connor Desai]

Thanks for your help. I think there are several issues with my design that make it a little complicated to analyse. For instance, I have several replications but a number are not complete replicates.

One approach I have seen used is to analyse the residuals rather than the raw scores – so that I can report something in my thesis – with the proviso that I run a new study with a different design.

[Author]

Posts