Same goes for IQ testing, use the same one each time and it should keep going up well into your 80s despite the thing you are trying to measure probably starting to degrade 50 years earlier.
Based on my own experience of doing a lot of Quantified Self and talking to other people about Quantified Self I’m doubtful about people self motivating to do the same IQ test for 50 years.
The existing IQ tests are also not designed to be used the way you propose. There are likely training effects that make them a bad measurement.
Long-term I would expect that you can mine existing data like Anki for a measurement of cognitive ability.
re: Anki, I think you might find some of the data the creator of SuperMemo (what anki is based on) very interesting. He’s used a combination of repetition data + a thing in SuperMemo called sleepchart to look at how alertness as the day goes on varies with different sleep things.
First graph here shows how recall varies as a function of hours from waking.
As a long-term SM user, Idon’t think you could use anki/supermemo to measure cognitive ability (easily). Compared to when I first started, my cards are much better with higher optimum interval. That hasn’t been do to anything but just me getting better at the skill of formulating my cards. Changes in skill would create too much noise long-term to see cognitive ability changes. Short term though, I wonder if you could daily recall % as a proxy for cognitively enchanting drugs as long as the time period isn’t so long that skill improvements mess things up.
As a long-term SM user, Idon’t think you could use anki/supermemo to measure cognitive ability (easily).
I don’t claim that you can measure it easily. I think you would need a more complicated mathematical model.
I think it would be great to know the percentage of cards answered at a certain day correctly that get answered correctly the next time the card is shown.
>I think it would be great to know the percentage of cards answered at a certain day correctly that get answered correctly the next time the card is shown.
Sorry but what would you do with that? It’s not immediately obvious to me
I should have specified different IQ tests meant to give similar results, of you can’t take the same one twice.
Long-term I would expect that you can mine existing data like Anki for a measurement of cognitive ability.
Personally that’s what I’m doing, simpler cognitive tests + other metrics such as my WPM while doing various things and getting the data from that. But that’s simply because I have a silly point of pride for never getting an IQ test, and I thought the IQ test is the safer thing to recommend for “long term” measurement.
Short term, obviously, things like short cognitive tests work best.
Based on my own experience of doing a lot of Quantified Self and talking to other people about Quantified Self I’m doubtful about people self motivating to do the same IQ test for 50 years.
The existing IQ tests are also not designed to be used the way you propose. There are likely training effects that make them a bad measurement.
Long-term I would expect that you can mine existing data like Anki for a measurement of cognitive ability.
re: Anki, I think you might find some of the data the creator of SuperMemo (what anki is based on) very interesting. He’s used a combination of repetition data + a thing in SuperMemo called sleepchart to look at how alertness as the day goes on varies with different sleep things.
First graph here shows how recall varies as a function of hours from waking.
As a long-term SM user, Idon’t think you could use anki/supermemo to measure cognitive ability (easily). Compared to when I first started, my cards are much better with higher optimum interval. That hasn’t been do to anything but just me getting better at the skill of formulating my cards. Changes in skill would create too much noise long-term to see cognitive ability changes. Short term though, I wonder if you could daily recall % as a proxy for cognitively enchanting drugs as long as the time period isn’t so long that skill improvements mess things up.
I don’t claim that you can measure it easily. I think you would need a more complicated mathematical model.
I think it would be great to know the percentage of cards answered at a certain day correctly that get answered correctly the next time the card is shown.
>I think it would be great to know the percentage of cards answered at a certain day correctly that get answered correctly the next time the card is shown.
Sorry but what would you do with that? It’s not immediately obvious to me
I should have specified different IQ tests meant to give similar results, of you can’t take the same one twice.
Personally that’s what I’m doing, simpler cognitive tests + other metrics such as my WPM while doing various things and getting the data from that. But that’s simply because I have a silly point of pride for never getting an IQ test, and I thought the IQ test is the safer thing to recommend for “long term” measurement.
Short term, obviously, things like short cognitive tests work best.