Multidimensional Within Item Partial Credit Model

[yutwer]

Hello,

I try to describe Multidimensional Within-Item Partial Credit Model (the text of this model and a part of the data file in the attachments).

So I need to get the theta estimation, but I don't see this estimation in the report-file. I think that I do someting wrong describing the model. Can you hint me where mistakes are?

Also, do you know where I can find out the likelihood function of the model in the attachments (file - model.jpg) and how ConQuest estimate the latent parameters mathematicly?

[dan_c]

for ability parameters (theta), see page 113 of the Command Reference - you should use the show command, e.g.:

Code: [Select]

show cases ! estimates=latent, filetype=xlsx >> example.xlsx;
Page 87 of the Command Reference gives examples of how to interpret your model statement in terms of the estimate of the odds of given person responding to an item/category.

[yutwer]

Hi, Dan! I tried to use the show command as in your example. So, I've got a matrix, what it shows?

On 87 page of the Command Reference I see the mh command description.

[dan_c]

page 87 of the Command Reference, here:

https://www.acer.org/files/Command-Reference.pdf

If yours is different, you may be using an old version of ConQuest. Go get the latest version here:

https://shop.acer.edu.au/acer-conquest-4

Your show file, the columns are:

Code: [Select]

sequence,
pid,
PV 1 for dimension 1,
PV 1 for dimension 2,
...,
PV 1 for dimension n,
PV 2 for dimension 1,
PV 2 for dimension 2,
...,
PV 2 for dimension n,
...
PV m for dimension 1,
PV m for dimension 2,
...,
PV m for dimension n,
EAP for dimension 1,
SD for dimension 1,
reliability dimension 1,
EAP for dimension 2,
SD for dimension 2,
reliability dimension 2,
EAP for dimension n,
SD for dimension n,
reliability dimension n
* wehre n is the number of dimensions, and m is number of PVs (default is 5)

See page 116 of the Command Reference.

[yutwer]

Hi, Dan!

Would you consult me please about the model that we had discussed on April 13, 2018? If you know any articles where is shown how to express the unknown parameters, I mean theta and difficulty of tasks (b) it would be great!

Thanks!

[dan_c]

check out:

Adams, R. J., Wilson, M., & Wang, W. (1997). The Multidimensional Random Coefficients Multinomial Logit Model. Applied Psychological Measurement, 21(1), 1–23. https://doi.org/10.1177/0146621697211001

[yutwer]

Thanks a lot, I will!