mixed mirt

[zahra...@gmail.com]

Hi Phil,

I want to used mixed mirt to analysis my data. I have 2 levels but  I do not want consider item level and person level.  I have person and school levels (for example) item level is not important for me. Some things like fox 2007 with package mlirt.

1) Can I do this analysis in mirt?
2) If yes, would you please give me some sense of its syntax?

best
Zahra

[Seongho Bae]

Hi, Zahra.

I think it's hard to achieve. Because, It's not 'test response theory', this is 'item response theory'. So, The Multilevel Item Response Theory have to analyze item level.

Seongho

2016년 5월 18일 수요일 오후 12시 32분 9초 UTC+9, zahra...@gmail.com 님의 말:

[zahra...@gmail.com]

thank you for your response.
I want to analyze in item level.for example I want to detect DIF  across gender of student .but students are  cluster in school.I just want to take into account variance at the school level  in my DIF detection.
best
Zahra

[Phil Chalmers]

Detecting DIF with mixedmirt is possible, but really only for locating uniform DIF effects (non-uniform would require multiple-group methods, which the function currently does not support). What would be required is that for each item of interest, separate intercepts for each group would be estimated, and differences between these parameters would be indicative of uniform DIF. Wald tests are also possible here if the parameter covariance matrix is stable enough as well. But overall DIF detection with this function will be a bit more difficult than using something like multipleGroup(). 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.