1

I'm trying to do a factorial analysis of binary data (hit / miss) related to the results of a test that consisted of two types of items. My intention is to confirm the number of dimensions to later choose an IRT model to be applied. What is the best strategy to apply? Do I do a factor analysis with both types of items at the same time or by type of items?

The plan is to run these analyzes on R. I have already done several with the data, but I am never sure of the number of dimensions to hold. I don't find much information on how to assess the dimensionality of assessment tests.

Thanks

Filipe Falcão
  • 11
  • Please provide a bit more information about the nature of your data and the theoretical model for more than one dimension among the data. In brief, ¿is this an exploratory data analysis? or ¿are you trying to confirm a model? – Gregg H May 04 '21 at 14:02
  • Hi Gregg, thanks for your answer.

    I have student results data for a series of questions. The data is in binary form (0 - the item was wrong / 1 - the item was correct). The items are of two types: some were made by hand, others through appropriate software. My interest is to evaluate the process validity of both. For that, I wanted to do a factorial analysis of the data. I have tried several ways, but I never arrive at an adequate structure. After discovering the factorial structure, I wanted to apply IRT to evaluate each item.

    If necessary I can post an image of the data here.

     – Filipe Falcão May 04 '21 at 14:58
  • 1
    Does this answer your question? How do I appropriately examine the dimensionality of binary data? – Tom May 05 '21 at 07:04

0 Answers0