1
$\begingroup$

I am hoping someone can help with the following:

  1. I have been trying to use the ltm package in R to get person fit estimates using a 3PL model. The problem I have been having is that I start off with 4947 response patterns (dichotomous) but when I run tpm() I end up with 4851 ability estimates and 4851 person fit estimates. I tried with another data set and ended up with a similar problem.

    Here is the line of code I've used---gr3 is the dataset with 4947 lines of data containing zeroes and ones.

     threePL <- tpm(gr3, na.action = NULL)
    
  2. In addition, in order to run the 3PL model, I have to get rid of the person IDs, but I would like to retain the person IDs so that I can accurately map the ability and person fit estimates onto the right person.

$\endgroup$
  • 1
    $\begingroup$ Welcome to the site, @nikki0111. Are you thinking of this as a conceptual statistics / psychometrics question, or an R programming assistance question? I'm not sure, but I wonder if it is the latter. If so, this question would be off-topic for CV (see our help page). It could be on-topic on Stack Overflow, but it would need to have a reproducible example. $\endgroup$ – gung - Reinstate Monica Dec 2 '13 at 15:06
  • $\begingroup$ If you want help with the ideas, please edit to clarify; if you just want help w/ the coding, add a reproducible example & flag your Q. We can migrate it for you. (Please don't cross-post, though.) $\endgroup$ – gung - Reinstate Monica Dec 2 '13 at 15:07
1
$\begingroup$

rows that have the same response pattern will be combined to save space, try,

    threePL <- tpm(gr3, na.action = NULL)
    ability.est <- factor.scores(threePL, gr3)

and you will get the ability estimates that correspond to each of your rows.

| cite | improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.