I'm not sure if this is known or not, but I'm having issues with the residual method failing when the N of cases gets above about 100K.
dat<-mirt::simdata(a=rnorm(21),d=rnorm(21),N=140533,itemtype="dich")
mod <- mirt(dat,1)
residuals(mod)
Above gives error:
Error in if (s == 0) s <- 1 : missing value where TRUE/FALSE needed
In addition: Warning message:
In c + d :
Show Traceback
Rerun with Debug
Error in if (s == 0) s <- 1 : missing value where TRUE/FALSE needed
But the following runs fine (only difference is N decreases by factor of 10).
dat<-mirt::simdata(a=rnorm(21),d=rnorm(21),N=14053 ,itemtype="dich")
mod <- mirt(dat,1)
residuals(mod)
Note that the original error I received was an integer overflow error on the same line of the residual function. My data is ordinal and not dichotomous, which appears to push s out of the allowable range for integers. I cannot seem to generate a minimum working example for this issue though...
Thanks. Another user reported this a while back too, and the fix is available on the dev version here. Feel free to install to get the fix, or wait until the next formal CRAN release. Cheers.
I'm not sure if this is known or not, but I'm having issues with the residual method failing when the N of cases gets above about 100K.
Above gives error:
But the following runs fine (only difference is N decreases by factor of 10).
Note that the original error I received was an integer overflow error on the same line of the residual function. My data is ordinal and not dichotomous, which appears to push
s
out of the allowable range for integers. I cannot seem to generate a minimum working example for this issue though...