User-specified probabilities of ratings in the case of disagreement in simulated rating matrix

Lines 47 to 50 in 5c8b158

    
           thedata[i, -((i %% k) + 1)] <- base::sample(l, 
        
           											size = k - 1, 
        
           											replace = TRUE, 
        
           											prob = response.probs)

For the case of disagreement, should the ratings be sampled from 1:nLevels, which will be specified in the same order as response.probs, rather than l, which is a random permutation of 1:nLevels? Unless I've missed something, this sample() will apply the probabilities in response.probs to a random ordering of the 1:nLevels ratings, rather than the user-intended ordering.

Thanks for this issue. This has been fixed on Github.

	thedata[i, -((i %% k) + 1)] <- base::sample(l,
	size = k - 1,
	replace = TRUE,
	prob = response.probs)