Thanks for this post. I always found the Flynn Effect very suspicious, but didn’t have a good place to hang my skepticism.
OTOH, as someone who was a Norwegian conscript in 1994, it’s too early to tell to whether vanity will win out over confirmation bias in how this gets filed in my long-term memory. 😉
“Observed increases and subsequent decreases in intelligence scores do not reflect changes in latent intelligence.” Makes sense. But they do reflect a change in something - whatever it is that the test tests for. Is this something that someone might be interested in, or that might have an impact on the real world?
It definitely can be. People could be getting better at matrices tests because they're becoming better trained for dealing with certain types of manipulations that have real-world counterparts that are useful to be able to deal with more efficiently at any level of intelligence. That's possible! It's not shown yet, but it is possible.
The point of this post was just to note that the changes are not due to the rise and fall of intelligence.
This is interesting. I’ll be honest I wasn’t aware or how the layman was misunderstanding or abusing the Flynn effect before. I’m not as well read on the topic
Unrelated to the argument about loadings, but a more general question: just how latent is the latent construct to psychometricians?
Suppose most observable demonstrations of mathematical or verbal ability (test or real world) decrease in the next three decades due to lack of practice owing to constant access to AI. Suppose, by fiat, we know that individuals with this distribution of genomes would have equal observable performance to current humans in a counterfactual world where they were exposed to more calculation and writing. In this scenario has g gone down or is it constant?
I think it might be a joke about the more common phrasing "bipolar about", meaning that the subject vacillates between strong positive and steong negative feelings
I suspected this might have been something to do with "bipolar," either switching it to a "bisexual" as a joke, as you say, or the result of typing very quickly to make the 1 hour cutoff coupled with a mistaken autocorrect.
so, it's all just an illusion that our own measuring stick was getting shorter and longer as we measured, right? wouldnt have aaaannnyyything at all to do with disastrous changes in diet, with the sudden introduction of people , kids especially, living online 24/7 with their faces glued to screens, with ever increasing numbers of autism-inducing injections being done to children, or with an 'education' system whose main purpose is to suck down unfathomable amounts of money, and as a side effect grind children's souls into mush? wouldnt have anything to do with any of those factors, would it?
naaaahhhh it's our methodology of measuring intelligence that magically warped itself upside down.. got it.
So someone from 1957's "culture" is inaccurately rated compared to someone from 1990's "culture." Doesn't the lack of "comparability" between test takers of different generations/cultures lend some credence to the old complaint that IQ tests are "culturally biased"?
I also have to wonder how one can gauge the "accuracy" of a test's measurement of g without a circular reference to another test that would be subject to the same potential inaccuracy. But I am guessing that's discussed in the prior article.
No, incomparability over time doesn't lend credence to concerns over cultural bias at a given point in time. Thankfully, time travelers usually aren't an issue for IQ testing.
If I understand Alex’ point, he’s not worried about time travelers, but that differences over time would be analogous to differences between contemporary cultures.(?)
Thanks for this post. I always found the Flynn Effect very suspicious, but didn’t have a good place to hang my skepticism.
OTOH, as someone who was a Norwegian conscript in 1994, it’s too early to tell to whether vanity will win out over confirmation bias in how this gets filed in my long-term memory. 😉
"I always found the Flynn Effect very suspicious..."
Especially when I became more knowledgeable about humanity.
Unique opportunity to name an effect after what is discussed in the post 😄
Nnylf ("nilf") effect: reverse Flynn Effect
“Neuroscientist I'd Like to F…”
“Observed increases and subsequent decreases in intelligence scores do not reflect changes in latent intelligence.” Makes sense. But they do reflect a change in something - whatever it is that the test tests for. Is this something that someone might be interested in, or that might have an impact on the real world?
It definitely can be. People could be getting better at matrices tests because they're becoming better trained for dealing with certain types of manipulations that have real-world counterparts that are useful to be able to deal with more efficiently at any level of intelligence. That's possible! It's not shown yet, but it is possible.
The point of this post was just to note that the changes are not due to the rise and fall of intelligence.
Have you ever lost a post because the app deleted it?
If you allocated one hour to writing the post, and a half hour the next day revising, I bet you'd come up with something that's easier to understand.
I don't think it was too hard to understand, given that you're asking for 50% more effort!
This is interesting. I’ll be honest I wasn’t aware or how the layman was misunderstanding or abusing the Flynn effect before. I’m not as well read on the topic
Well, the math scores are simple enough; we're used to using calculators for more math.
(Cue Boomer relatives who use the calculator on their phone for anything more complicated than 2x2 complaining...)
In your opinion, has verbal comprehension specifically decreased or remained constant since the 1990s?
Would be interested to see a continued version of this article.
Why did you set app an aplet to delete them, if I may ask?
I don't want to spend too much time writing things. I have a full-time job that I don't want to detract from.
Unrelated to the argument about loadings, but a more general question: just how latent is the latent construct to psychometricians?
Suppose most observable demonstrations of mathematical or verbal ability (test or real world) decrease in the next three decades due to lack of practice owing to constant access to AI. Suppose, by fiat, we know that individuals with this distribution of genomes would have equal observable performance to current humans in a counterfactual world where they were exposed to more calculation and writing. In this scenario has g gone down or is it constant?
"Headlines and anecdotes are bisexual about these questions."
❓
They go both ways.
I think it might be a joke about the more common phrasing "bipolar about", meaning that the subject vacillates between strong positive and steong negative feelings
I suspected this might have been something to do with "bipolar," either switching it to a "bisexual" as a joke, as you say, or the result of typing very quickly to make the 1 hour cutoff coupled with a mistaken autocorrect.
Someone should administer tests from different periods to the same sample to check the results of the Intelligence study in a different way.
This has been done. The Flynn Effect does operate within cohorts.
so, it's all just an illusion that our own measuring stick was getting shorter and longer as we measured, right? wouldnt have aaaannnyyything at all to do with disastrous changes in diet, with the sudden introduction of people , kids especially, living online 24/7 with their faces glued to screens, with ever increasing numbers of autism-inducing injections being done to children, or with an 'education' system whose main purpose is to suck down unfathomable amounts of money, and as a side effect grind children's souls into mush? wouldnt have anything to do with any of those factors, would it?
naaaahhhh it's our methodology of measuring intelligence that magically warped itself upside down.. got it.
You might want to read some of the literature that Cremieux referenced before coming to a conclusion!
(If this comment looks familiar, Substack moved it to the wrong parent comment initially)
So someone from 1957's "culture" is inaccurately rated compared to someone from 1990's "culture." Doesn't the lack of "comparability" between test takers of different generations/cultures lend some credence to the old complaint that IQ tests are "culturally biased"?
I also have to wonder how one can gauge the "accuracy" of a test's measurement of g without a circular reference to another test that would be subject to the same potential inaccuracy. But I am guessing that's discussed in the prior article.
No, incomparability over time doesn't lend credence to concerns over cultural bias at a given point in time. Thankfully, time travelers usually aren't an issue for IQ testing.
If I understand Alex’ point, he’s not worried about time travelers, but that differences over time would be analogous to differences between contemporary cultures.(?)
I'm alluding to the fact that they're not. We can and do check within eras and bias is rare.
Oh. I’m a bit dense today. Sorry.
I had the Richard Lynn/Sub-Saharan IQ estimates in mind, and had to go back and reread your post on that.