I assume you've written on this before, and if so please point me towards it, but: how is it that SAT scores can change so drastically with a bit of practice and still be considered an accurate measure of IQ? On your old site you said your score changed by about 200 points between your first and second attempts, I believe. So is your IQ 118 or 131?
That's because after some practice increase of score come to a halt and most of people studies for this exam. If you never study for this exam you would probably get a score under what you should get with your IQ. But if you study this turns basically into an IQ test.
With practice, IQ tests can also change drastically. All of these tests are a number based on how the general population performed on a specific day. So far, nobody has conceived a test that measures a person's specific intelligence accurately with only one variable. Intelligence can only be measured precisely, with limitations. If the average was what would be a 120, then their score would be 100, and same with a score of 80. I tend to avoid measuring people on a number, rather their effort. People who practice deserve every bit of their elevated score just as much as the effortless.
Can you do this for the GRE? I would assume the GRE is more accurate at the high end, since the average graduate entrant will be 1-2 SD smarter than average (depending on subject). Breaking out verbal and math GRE would be helpful. Thank you
The SAT changed in 2016. Is there a reason to think that the correlation of 0.8 mentioned here would not be true of the new version? Most of the frequently quoted papers on the correlation go to ~2004. 0.8 seems to be a very high correlation.
Regression to the mean. SAT and IQ correlate highly but not perfectly, somebody with a super high SAT score will not necessarily (and probably not, actually) have a super high IQ.
You might have used the wrong ACT-SAT conversion, because 29 ACT, for example, predicts around 1340 on the SAT according to the ACT official site, whereas you have 1370.
I assume you've written on this before, and if so please point me towards it, but: how is it that SAT scores can change so drastically with a bit of practice and still be considered an accurate measure of IQ? On your old site you said your score changed by about 200 points between your first and second attempts, I believe. So is your IQ 118 or 131?
It's not that good, correlation with IQ is like 0.8. Test-retest differences of 1SD are rare but should not be that unusual statistically speaking.
https://humanvarieties.org/2017/07/01/measurement-error-regression-to-the-mean-and-group-differences/
That's because after some practice increase of score come to a halt and most of people studies for this exam. If you never study for this exam you would probably get a score under what you should get with your IQ. But if you study this turns basically into an IQ test.
With practice, IQ tests can also change drastically. All of these tests are a number based on how the general population performed on a specific day. So far, nobody has conceived a test that measures a person's specific intelligence accurately with only one variable. Intelligence can only be measured precisely, with limitations. If the average was what would be a 120, then their score would be 100, and same with a score of 80. I tend to avoid measuring people on a number, rather their effort. People who practice deserve every bit of their elevated score just as much as the effortless.
Can you do this for the GRE? I would assume the GRE is more accurate at the high end, since the average graduate entrant will be 1-2 SD smarter than average (depending on subject). Breaking out verbal and math GRE would be helpful. Thank you
It's mentioned on the site you reference that Mensa considers SAT scores from after Jan 31, 1994 to no longer correlate with an IQ test.
The SAT changed in 2016. Is there a reason to think that the correlation of 0.8 mentioned here would not be true of the new version? Most of the frequently quoted papers on the correlation go to ~2004. 0.8 seems to be a very high correlation.
The test measures the same abilities.
Could you explain what you are doing and why with RTM?
Regression to the mean. SAT and IQ correlate highly but not perfectly, somebody with a super high SAT score will not necessarily (and probably not, actually) have a super high IQ.
I understand the concept, but how exactly do you calculate the adjustment for RTM?
(130-100)*0.8+100
Ah, is 0.8 pearson r or just professional judgement?
r
Thank you. Your analysis led me to Bland and Altman, and my eyes are now open.
Could this table be made available somewhere in tabular or textual format?
https://theuntangler.wordpress.com/2024/01/15/table/
You might have used the wrong ACT-SAT conversion, because 29 ACT, for example, predicts around 1340 on the SAT according to the ACT official site, whereas you have 1370.
the age is a little out of date for the plots online. i think the ones i saw were like 2018 or so.