>>11614284I'm OP and I went to take a nap cause i figured no one would reply to this thread anyway. Let me go post by post. I just finished reading this
>>11614044, and if this is the average level of argument you're gonna try to use, I'm already upset at the waste of time you're gonna put me through.
The premise that you can test the reliability of SAT scores by college performance is a flawed one, by infinitely many reasons. Only someone with absolutely 0 understanding of statistics, theory of measurements and errors, and basic scientific thinking could say some shit like this with a straight face. Again I wouldn't be surprised to know this was written by a psych*logist.
>https://en.wikipedia.org/wiki/Richard_J._HaierOH WOW what a surprise. A psych*logist.
OK so let me break this one down for you step by step, alright?
>Test bias has a specific meaning.If you are going to treat tests as actual measurements, you should stop using psychological jargon this very second and start using metrological one. Start talking about expanded uncertainties and that kind of shit. This won't happen for 2 reasons:
1 - Because psych*logists would shit their pants if they had to learn actual math and actual proper procedures to apply statistical models/calculate errors and uncertainties.
2 - Because it would probably show how weak and statistically insignificant their shit is.
>If people with high SAT scores consistently fail college course, the test is overpredicting success and it is a biased test.This is NOT what measurement bias means, and this is NOT what the test is testing for, and this is NOT the first conclusion anyone with half a brain would draw if a group of people with amazing SAT scores failed their classes on college. Only someone desperate to claim intelligence tests are amazing would say something like this. A more normal scientist would check if there is some extenuating circumstance happening in that particular college. Post is too long.
