Problems simulating data

[Luis Manuel Lozano]

Hi Phil,

I ama new R user and of course, new on dirt package, so maybe the question is an easy one because of my inexperience.

I am trying to simulate a multidimensional IRT model with two factors (a correlation of .35 between them). The answers have to be in a Likert scale of 5 points. The code I use is:

library(mvtnorm)

library(mirt)

Theta <- rmvnorm(1000, sigma = matrix(c(1, .35, .35, 1), 2))

a <- matrix(c(rlnorm(40, .2, .3),rlnorm(40,.7,.4)),2)

diffs <- t(apply(matrix(runif(40*4, .3, 1), 40), 1, cumsum)); 

diffs <- -(diffs - rowMeans(diffs)); 

d <- diffs + rnorm(40)

datos <- simdata(a, d, 1000, Theta=Theta, itemtype = 'graded')

datos

modelo<-mirt(datos,2)

When I run the program I obtain negative coefficients in the a parameters and I do not know what is wrong?

Do you think that the program is correct? If not where can I read something to make it properly?

Thanks in advance and congratulations for the packages

Luis M. Lozano

[Phil Chalmers]

The model is exploratory, so negative coefs are likely to occur and require rotation after convergence (see topics on exploratory factor analysis for more information on this phenomenon). Cheers.

Phil

--
You received this message because you are subscribed to the Google Groups "mirt-package" group.
To unsubscribe from this group and stop receiving emails from it, send an email to mirt-package...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.