Quote:
Originally Posted by gina_re
I understand this concept since I work in PsycTESTS at the APA, but this is a way too complicated. At least to understand the psychometrics described.
|
All the statistics aside (that will be done later and not all that rigorously), there are two rounds after I gave all questions (just a few, not more than 10).
The first round is for giving predictions and takes 24 hours. Just predict scores just like you would score your own answers.
The second is for giving your own score.
No "real"/common statistics will be used: that's not the point!
Just (like) variance per item
and respondent and if one wants to score extra points, three types of correlation (does a increase for one score mean that an increase/decrease of another is likely or not).
I might give standard deviations to make it sound classy and to have a way of determining normalcy, but that's just "sugar on the cake" and doesn't matter: you can't lose. This will not be ranked.
Hope this helps. Thanks for letting me know it might "sound"/be a bit complicated.
Edit:
And to reduce the likelihood of confusion later, I will prepend to the numbers of my questions Icare_.