Item keying

[Hynek Cígler]

Dear Phil,
I suggest a minor and not-so-important improvement to several mirt functions regarding item keying.

In many cases, reverse-keyed items are present in the data. They do not need to be reverted before analysis in mirt, which results in negative discrimination parameters. I actually prefer this option in most cases.
However, in some situations, initial item reversals is necessary – for example, when plotting test information or total test score functions, or when estimating EAPsum factor scores (or reliability). This limitation is quite inconvenient.

The implementation should be fairly straightforward: a new parameter, such as key = c(1, 1, -1, -1, …), could be introduced.

I also suppose that item_weights parameter in fscores function could be used, but I am not sure and this doesn't help in plotting functions.

Thank you for the consideration!

Best Regards,
Hynek

[Phil Chalmers]

Hi Hynek,

I like the idea, but find the extra cognitive load may be a bit harder for users to swallow. I would generally prefer the user to manipulate the data in some way that makes the item mapping orientation something that matches with the model they are trying to work with (e.g, higher scores due to higher trait values), rather than internally doing sign flippings. Perhaps it would be convinient to have a reverse.score() function or something added to the package doing something like 

newdat <- reverse.score(dat, which=c('item1', 'item'2' ...))

mirt(newdat)

If this would solve the issue I'd be happy to add something like this.

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.
To view this discussion visit https://groups.google.com/d/msgid/mirt-package/e0a0ddc3-2077-4135-bc4d-b07464655d32n%40googlegroups.com.