item weight

[Lucas Sempe]

Dear Phil,

I've been working on creating a wealth index using MIRT. My syntax works well, although very simple.

irt4.1 = mirt(data = Assets4.1[4:13], model=1, itemtype="gpcm")

irt4.1f <- fscores (irt4.1)

> summary (irt4.1)

       F1    h2

Q1  0.675 0.456

Q2  0.446 0.199

Q3  0.685 0.470

Q4  0.780 0.609

Q5  0.750 0.563

Q6  0.537 0.288

Q7  0.551 0.304

Q8  0.693 0.480

Q9  0.486 0.236

Q10 0.683 0.466

Analysing the raw data, there is one item (!7) which should have a lot more weight than others (theoretically speaking). So, my question is: how to perform a "theta" modification in order to modify the weight of the specific item in the fscores?

I've tried introducing many customTheta to the command, such as:

Theta <- cbind(c(1, 1, 1, 1, 1, 1, 5, 1,1,1),c(1, 1, 1, 1, 1, 1, 5, 1,1,1))

irt4.1 = mirt(data = Assets4.1[4:13], model=1, customTheta=Theta, itemtype="gpcm")

But it didn't work.

Thanks! Lucas

[Phil Chalmers]

This is an interesting idea, I'd never really thought about weighting the scoring of the ability terms based on prior theoretical justification of items over their item properties. I don't think this is possible to do in the current code-base, though some modifications could be made to accomplish this (it's very similar in flavour to the survey.weights input in the mirt() function). 

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.

[smi...@rowan.edu]

Greetings,

I realize this conversation is quite old, but I would like to do a similar thing. I was wondering if Phil or anyone else has managed to do what Lucas inquired about: weighting different items to have more or less of an effect on theta.

Thanks so much!

Trevor Smith