Request for Item Response Models (IRT)

[darcysantor]

Item Response Models in JASP?

* Enhancement: I wanted to ask if you have plans to implement methods to fit an item response model (IRT) for questionnaire, survey and multiple choice test data. In its simplest form, this is essentially involves fitting a curve that models the probability of answering a question correct as a function of total score. Once the curve is fit, the parameters associated with the curve provide an estimate of item difficulty and discrimination (slope) as function of ability. This model can be extended to model multiple choice tests. Multilog is an example of proprietary software that models a 3-parameter model. However, I get this type of analysis whenever I have the multiple choice tests scored at my university. * Purpose: The purpose is to analyze the performance of individual questions from questionnaires that measure some kind of knowledge (e.g., multiple choice exams at university), personality characteristics (e.g., amount of neuroticism that you have), or the amount of condition (e.g., self-report depression tests). I will be using JASP in my psychometrics course. Would be great to be able to include these kinds of analyses as well. SPSS and SAS do not do this, as far as I know. This would be a marvelous enhancement for JASP and make it unique I think.

[fplatz]

This would be an major contribution to JASP and was articulated as question for enhancement several times in the past ( eg. stated in #2174 for the Bayes' version). It would be a great improvement!

[sukhoi713]

Absolutely agree with this enhancement, as a psychometrics student from Indonesia, IRT model have been widely use here but not everyone can use RStudio or can buy Winsteps, so please add this feature to the next update. Great job by the way with the JASP. Cheers.

[EJWagenmakers]

This is clearly a valuable addition. We'll explore the possibilities...

[darcysantor]

This should be relatively easy, as this is a model-fitting exercise. I think you could (a) provide the LAVAAN code to fit this model, (b) build a stand-alone feature that utilizes LAVAAN, or (c) both.

[vankesteren]

the mirt package might a good start if we want to do it well, with rasch, 2pl, and 3pl models.

[darcysantor]

This sounds promising. I did not realize that JASP was built on R packages. Definitely the way to go!

[DocGavin]

Totally would like IRT in JASP. i agree mirt package is the way to go @vankesteren

[ShungoSuzuki]

I still wait for the addition of IRT and Rasch!

[darcysantor]

You have my vote too!

From: ShungoSuzuki @.> Sent: March 9, 2022 8:30 PM To: jasp-stats/jasp-issues @.> Cc: Darcy Santor @.>; Author @.> Subject: Re: [jasp-stats/jasp-issues] Request for Item Response Models (IRT) (#160)

Attention : courriel externe | external email

I still wait for the addition of IRT and Rasch!

— Reply to this email directly, view it on GitHubhttps://github.com/jasp-stats/jasp-issues/issues/160#issuecomment-1063561298, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AJKRFWDR73IGDJ2JTNEVWZDU7FGA7ANCNFSM4GDVQUYQ. Triage notifications on the go with GitHub Mobile for iOShttps://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Androidhttps://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub. You are receiving this because you authored the thread.Message ID: @.***>

[tomtomme]

related to

120 bayesian IRT

50 descriptive part of IRT

2072 mentions DIF analysis R-Packages that are part of IRT:

• lordif (Choi, Gibbons & Crane, 2011)
• difR (Magis, Beland & Raiche, 2013)
• TAM (Kiefer, Robitzsch & Wu, 2013)

[juliuspfadt]

I think this would warrant its own label. Would be a module called jaspIrt. It is on my toDo list, although that list is long. And I will need some expert help with this.