The "I Didn't Comment in Another Thread" Thread


log in or register to remove this ad

Cadence

Legend
Supporter
Way off tangent:

Large standardized testing companies do quite a bit of work to make sure they don't have biased test items. Part of this is pretty qualitative, but there are also statistical tests for "differential item functioning (DIF)". Had an assignment in a class once where we were asked to do something like guess which items on an operational test had shown different levels of estimated DIF and, if so, against which group. After seeing the answers It was pretty hard to figure out why some of them showed the way they did. I think in some cases we spent quite a bit of time speculating without suggesting it was just a Type I or II error. Accompanying the exercise with examples of the (presumably) problematic ones that were screened out earlier in the process would have been nice too. Only seeing the ones on the edges feels like it would miss part of the big picture. A blurb about the awful stuff that used to be common in the big tests would be a nice addition to the text books too.
 
Last edited:

Asisreo

Patron Badass
Well, that thread went in a different direction...

It's been an awful and stressful week for me. I usually take short rests, but I'll take a long one this time.
 

Cadence

Legend
Supporter
Remember that great Finland diversion earlier in the "didn't post thread" (with a Monty Python video even)?
Want something on the other side of the world from New Zealand?
Vote Mielikki!
Don't set yourself up for a bad Nuutipukki visit by picking someone else or not voting!
 
Last edited:







Remove ads

Top