Closed AlexSwiderski closed 3 years ago
Is there any particular reason we might want to do this?
The rational for this based on a courtesy to the authors for previous work in this area and also to present a short form option which may provide a greater variety of words (length, frequency of use, etc). On the latter point, we should look at the test to test correlations of the predictors for item difficulty. Might be interesting, perhaps I can coax one of our undergrads to do this.
I had two reasons for suggesting adding Grant's short forms to the app: 1. Professional courtesy (though, it occurs to me now that perhaps we should ask him if he even wants us to do this), and 2. Assuming that some clinicians may prefer or only be able to give pencil and paper tests, they could use the app to get T-scores from these static short forms administered offline.
Given the results in Hula, et al. 2015, Figure 5, that show that the static forms are associated with more error, I am not sure I see a strong reason for including them as a professional courtesy. Also, I am not sure how many clinicians would do what Will describes. Having said that, I am happy to go either way. However, if we choose to include the static forms, we might want to consider adding some language pointing out some of the limitations of the short forms.
I will see how difficulty this is to implement and if its just a few tweaks we can include it, and if its more involved we can table it for now.
These have been tentatively added. they seem to be working fine. I probably need to write a test for them...
For now, maybe we can give a few of the undergrads some fake 30-item walker scores and have them enter them in and make sure they get the same results.
Also...I'm still showing an ability estimate for the walker short forms, is this right?
Yes, ability estimates (and 95% CrI) can be provided.
From end of July meeting: W.H. to reach out to Grant to see what his preference is.
@william-hula any update from Grant?
go ahead from Grant. closing this issue.