Yeah, I’ve done many of those. I took the SAT when I was 12. I’ve taken a few probably-inaccurate online IQ tests. I’ve done a few cognitive testing suites at SIAI. I’m in pretty good shape. In general though, there are better frameworks for cognitive testing. It’s probable that one could make a neat suite out of PEBL, which is free and very customizable. Fluid g seems over-emphasized. The limiting factor for most rationalists tends to be strong metadispositions for thought, reflection, and drive.
We didn’t study long enough to get any statistically significant data. Like, not even close. And I think sending off the data (even without names attached) would sorta breach an implicit privacy agreement among those who took part in the tests.
Jaeggi 2008 didn’t necessarily study very long either, some around/less than a week.
(I wouldn’t be asking this question, by the way, if you had written more concretely and said something like ‘We only studied for 2 days, not long enough to get any statistically significant data’.)
Hmm, most people would be ok with that sort of data being sent out in an anonymized form. I’m surprised that you didn’t suggest that before hand. Is there any chance you can contact the people in question and get their permission to release the anonymized data?
Is there any chance you can contact the people in question and get their permission to release the anonymized data?
We could, but really, there’s no information there, no matter how much Bayes magic you use. It’s noise. If the data was at all significant then we’d send it out, of course. We might actually have gotten anonymized disclosure agreement from everyone; I don’t remember. But it didn’t end up mattering.
Yeah, I’ve done many of those. I took the SAT when I was 12. I’ve taken a few probably-inaccurate online IQ tests. I’ve done a few cognitive testing suites at SIAI. I’m in pretty good shape. In general though, there are better frameworks for cognitive testing. It’s probable that one could make a neat suite out of PEBL, which is free and very customizable. Fluid g seems over-emphasized. The limiting factor for most rationalists tends to be strong metadispositions for thought, reflection, and drive.
They have those?
They’re ad hoc, we’ve used one for a dual n-back study which ended up yielding insufficient data.
Any chance you could write up that study? I don’t believe I have seen any SIAI-related DNB study; certainly it’s not in my FAQ.
(Remember kids: only you can fight publication bias!)
We didn’t study long enough to get any statistically significant data. Like, not even close. And I think sending off the data (even without names attached) would sorta breach an implicit privacy agreement among those who took part in the tests.
Jaeggi 2008 didn’t necessarily study very long either, some around/less than a week.
(I wouldn’t be asking this question, by the way, if you had written more concretely and said something like ‘We only studied for 2 days, not long enough to get any statistically significant data’.)
Hmm, most people would be ok with that sort of data being sent out in an anonymized form. I’m surprised that you didn’t suggest that before hand. Is there any chance you can contact the people in question and get their permission to release the anonymized data?
We could, but really, there’s no information there, no matter how much Bayes magic you use. It’s noise. If the data was at all significant then we’d send it out, of course. We might actually have gotten anonymized disclosure agreement from everyone; I don’t remember. But it didn’t end up mattering.