[R] How to solve empty cells in the contingency table?

[David Duffy]

>
> Hi,Dear all R experts,
>
>    I am trying to do the 2-way contingency table analysis by fitting 
> the loglinear models. However, I found my table has several empty 
> cells which are theoretically missing values.I have no idea of how 
> to solve them coz we cannot compute the simulated p-value with zero 
> marginals.Does someone have some suggestions? Please help me out, 
> thanks a lot!
>
> Cheers,
> Yan

One approach: if they are *structural zeroes*,

tab <- as.data.frame(table(x,y))

then use the exactLoglinTest package's mcexact function

David Duffy.