I think, it works for all subject areas but the study modules should be aligned with the assessment questionnaires or the data.
Depends what you mean by work. If you mean fit the data well (i.e., are the data really unidimensional and other assumptions), then no, it usually does not work. If you mean something else, please specify this.
We are able to know about that only after testing. While IRT packages are able to provide us with fit indexes, giving the information about how well estimated curves approximate and can explain the observed data, it is important to also think about the available models. You might need other parameters that could make more sense into educational context such as discrimination and a guessing parameter. You'll have to think also about how to deal with item scoring - is it right/wrong, a partial credit system, rating system, etc? I'm not sure about your question to as you refer to scores. Please, remember that processing data for IRT models is in the form of answers to items, not scores. I think you need to expand your question a lot in order for us to help you.
@Jimmy Also for most data a unidimensional model is too simply, depending on the purpose, so again, it depends what the person who asked the question means by works. It the questioner just means can you run the model, then the answer is trivial.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Applied Statistics
- Survey Design
- Questionnaire