Hi, I am running the non-parametric rasch model tests using eRm. I have a reasonably large dataset for this type of exercise (110 items, 248 persons). I run:> allb2=as.matrix(allb) > rsample <- rsampler(allb2, ctr) > t102<-NPtest(rsample, method="T10") #global test, subgroup invand receive error message as follows. "Error in m[idx1, idx2] <- 1 : subscript out of bounds" Using the same code for a different dataset does not cause this problem. Any suggestions much appreciated! (ctr spec: burn_in = 100 n_eff = 1000 step = 16 seed = 0 tfixed = FALSE ) -- View this message in context: http://r.789695.n4.nabble.com/eRm-raschsampler-error-message-tp3770265p3770265.html Sent from the R help mailing list archive at Nabble.com.
I'm not familiar with eRm but subscript out of bounds often occurs when the data isn't the structure you think it is. Perhaps your input is not what is expected - might not be as long/wide as required. Depending on how allb is stored, you might be flipping rows/columns around? try checking the dimensions of the object with dim(objname) and view the top part of the data to see if the people and objects are correctly oriented. objname[1:10,1:10] Do you get the error still if you reduce the data set to something smaller? I'd leave the people at 248 and reduce the items - so you can have many counts of each item. If so, it might be worth posting a sample of the allb data so people here can see what is happening. Try reducing the data gradually until you get to small example that still causes the problem. I'd initially leave the people at 248 and reduce the items - so we can have many counts of each item. Then once you reduce the number of objects, if you're still getting the errors, then try reducing the people down to a reasonable number and then take a look at the data you've got and see if it really makes sense to you. Then consider posting here the data so we can see what's happening. Without the actual numbers you're putting in to the equations, it's hard for me at least to tell what's happening. -- View this message in context: http://r.789695.n4.nabble.com/eRm-raschsampler-error-message-tp3770265p3771197.html Sent from the R help mailing list archive at Nabble.com.
Hi, this error message is due to a bug in RaschSampler. I corrected it and sent it to CRAN. RaschSampler version 0.8-5 should be available from there soon. Thanks for the hint and please apologize for any inconveniences. Reinhold -- View this message in context: http://r.789695.n4.nabble.com/eRm-raschsampler-error-message-tp3770265p3773171.html Sent from the R help mailing list archive at Nabble.com.