Regarding Mankind Quarterly and the Pioneer Fund: The relationship between genes, IQ, race, and GDP is very controversial. Prestigious journals are hesitant to publish articles about these topics. Using the beliefs of the founding members in the 1930s to dismiss an article published in 2022 is an extremely weak heuristic. The US government funds a lot of research but it committed unethical acts in the name of eugenics. Sam Bankman-Fried, a fraudster, funded a lot of EA projects. If I linked to some research that was performed using FTX money, I would not consider it worthy of dismissal. Furthermore, bad people can fund and conduct good research. I cited a lot of more mainstream journals for less controversial claims. I don’t consider instrumental variables that strong of evidence, but it felt worth mentioning. I will email you the PDF if you are interested.
Regarding cultural factors: If there are national differences in genotypic IQ, then measures of quality of education and culture will be genetically confounded. I do not doubt that schooling increases scores on tests of mental ability, but the gains appear “hollow.” Hollow is a technical term meaning it is not increasing g. I am not concerned about what “true intelligence” is because “intelligence” is an ordinary term without a precise definition. We can call what IQ scores are trying to measure “GMA” and not care that it’s not “intelligence” but still care that GMA is correlated with good outcomes and we have a means of increasing GMA. The benefits of increases in general mental ability generalize to other areas (career, academic success, good life choices), whereas non-general gains will be limited. As an extreme example, it is obvious why giving children Raven Progressive Matrices is not going to make them drive better. But evidence suggests having higher IQ, reduces risk of traffic accident.
Regarding Richard Lynn: I address this within the article. Until his death, Lynn and colleagues where updating the NIQ scores. Looking at the most recent version on ViewOnIQ from Becker, I see that Lynn has excluded all those samples for the Nigerian estimate and incorporated the Maqsud estimates. He has also included several other more recent estimates, but arrived at a similar estimate still. You can view the samples used and estimates here if you download the file. More importantly, focusing on Lynn is a mistake as I mention in the article. Other less controversial researchers estimate “universal basic skills” or “harmonized learning outcomes” and produce estimates which correlate highly with the NIQ estimates. See the chart from Warne 2023. A side point, but Wikipedia is politically biased. I intentionally capitalized the L to give credit as Richard Lynn’s discovery preceeded Flynn’s first publication. Although, his discovery was preceeded by Runquist.
On The Bell Curve: You say “That book has so many problems that instead of typing it all out I would like to direct people to this video which points out a lot of them.” I don’t plan on watching the 2 hour 39 minute video just to respond to you. At the time, a large number of claims (like the one I make) was not particularly controversial among intelligence researchers (see Gottfredson and APA response). I discuss this in the article. Furthermore, the more recent Rindermann et al. (2020) found in a survey among intelligence experts that many believed SES was substantially explained by intelligence (see in the article). I am also going to make a point about isolated demands for rigor. You are dismissive of some the researchers and journals for lack of academic quality, but in response to a book co-authored by a Harvard psychologist, you give me a YouTube video with a psuedonymous guy with a skull avatar that is part of LeftTube or “Bread Tube”. I am not suggesting this means that I can merely dismiss anything that he’s saying, but I will admit that this feels like a double standard given the Mankind Quarterly critique.
On environmental influence: You say “I don’t think anyone thinks the environment explains 100%, but given that it’s much larger and has many more variables it seems reasonable to assume it can explain more of it.” That doesn’t make sense as an argument in my view. We don’t know that it is larger. Within the USA, the heritability is over 50% in adulthood indicating that genetic differences are the largest driving force among all known influence. Whether that is true of international differences is not entirely clear. The existence of more possible environmental explanations doesn’t mean that it has more explanatory variables. Analogously, this argument would be wrong: There are over a billion base pairs in the genome, therefore genetic explanations are better. The issue is that the environmental influence largely, although not entirely, consists of unsystematic differences. Iodine is an exception and should be given in areas with deprivation. But the gains from embryo selection in large batches is beyond 9 points, and could be used on the non-iodine deficient population.
I have responded. I don’t know if you will be brigaded. I have not personally downvoted you.
It seems I didn’t get brigaded [tap on wood], but I still feel uneasy answering this. You got some downvotes on this comment initially which means the karma system pushes you to not reply, in the same way it pushed me to not reply to the HBD-proponents I was debating. This voting-power-by-popularity system doesn’t incentivize having conversations, so feel free to answer in the comment section on your substack instead. I will edit in a link to it at the end of this comment if you do so. This comment is going to be shorter anyway.
Firstly, I wanted to say that I also didn’t downvote your post because while I disagree I do sympathize with the amount of effort that went into it, and this karma system would punish future unrelated posts and comments in a guilt-by-association-fashion, not just this post (although I did give it a disagreement vote since that influences nothing):
Secondly, to clear up any confusion I don’t think a journal being made by horrible people allows you to conclude that their conclusions are false, but I do think it allows you to not give them any money.
Thirdly, I think we run into the same issue with g as we do with intelligence. If g is just correlation between different cognitive tasks, then the natural question is, which tasks? And which tasks are considered ‘cognitive tasks’? Because the results will differ based on what you choose.
Fourthly, I think the difference with linking a video and linking the mankind quarterly is primarily money, one costs 75 dollar and the other is free. For the record I have watched the video in it’s entirety and am not just throwing something at you while I myself don’t know of any counterarguments. I could’ve typed them out, but I just don’t think I have much to add both in terms of information nor presentation. For those who are familiar with the subject you can skip to 1:02:11 of the video at which point he really starts diving into their methodology instead of giving a general overview. Which ties into...
Fifthly, I think the video points out a general pattern of Lynn and his colleagues using a clearly cherrypicked dataset then being called out on it, at which point they switch to another slightly less clearly cherrypicked dataset which people then call them out on etc. Now maybe this latest dataset they use is genuinely good but I think this is a ‘boy who cried wolf’ scenario where I just no longer think it’s prudent to trust them or trust that reading their work is a productive use of time.
Lastly, maybe at some point we run out of good environmental interventions (like iodine) and maybe then (assuming some premises) it becomes prudent to switch to genetic interventions. But until that time we should focus on those environmental interventions, not just because of their immediate cost effectiveness but also because of one of my points you didn’t address, namely that those environmental interventions are way more egalitarian/emancipatory, which produces better results in the longterm.
I will respond here because it’s important for everyone to see.
You don’t need to give the journal money. I am offering to email you the pdf if you are that interested.
Cognitively demanding tasks. These require puzzle-solving, reasoning, drawing on past knowledge, connecting ideas, etc. As long as the test has a wide range of tasks like this, estimates will be similar. Provided they are cognitively demanding and diverse, results are not particularly sensitive to the actual content of the test for native speakers. Spearman called this the “indifference of the indicator.” You can read more in Chapter 7 of Rusell Warne’s book In The Know.
Another interpretation of what Lynn is doing is improving his estimates when people critique him. As I mentioned in the article, many other researchers estimating national averages on mental ability tests produce moderate to highly correlated estimates with Lynn’s. Why do you think that is? And who do you think has more accurate estimates? If you have a specific objection to the scores, I can respond with estimates of correlations after making adjustments. Ee can use the ViewOnIQ data to drop scores/samples from countries you find are bad. Or we can Winsorize the scores and check the strength of the relationship. Or we can look at Rindermann’s estimates.
What do you think, in your view, is the correlation between average cognitive ability and log(GDP/c)? What are you basing this estimate on, and why is it better than Lynn, Becker, Rindermann, Angrist, etc?
Iodine deficiency will not work for everyone, only those who are iodine-deprived. Where people are iodine deficient, we should try to help them. I focus on genetic enhancement because it is under-considered. The possible returns from genetic enhancement will be unevenly distributed but have the potential to be absolutely massive. Since parents will adopt it voluntarily and IVF is largely legal worldwide (and hopefully IVG will be), there is a plausible means of improving humanity immensely merely through funding research in a narrow area to accelerate certain discoveries.
Regarding Mankind Quarterly and the Pioneer Fund: The relationship between genes, IQ, race, and GDP is very controversial. Prestigious journals are hesitant to publish articles about these topics. Using the beliefs of the founding members in the 1930s to dismiss an article published in 2022 is an extremely weak heuristic. The US government funds a lot of research but it committed unethical acts in the name of eugenics. Sam Bankman-Fried, a fraudster, funded a lot of EA projects. If I linked to some research that was performed using FTX money, I would not consider it worthy of dismissal. Furthermore, bad people can fund and conduct good research. I cited a lot of more mainstream journals for less controversial claims. I don’t consider instrumental variables that strong of evidence, but it felt worth mentioning. I will email you the PDF if you are interested.
Regarding cultural factors: If there are national differences in genotypic IQ, then measures of quality of education and culture will be genetically confounded. I do not doubt that schooling increases scores on tests of mental ability, but the gains appear “hollow.” Hollow is a technical term meaning it is not increasing g. I am not concerned about what “true intelligence” is because “intelligence” is an ordinary term without a precise definition. We can call what IQ scores are trying to measure “GMA” and not care that it’s not “intelligence” but still care that GMA is correlated with good outcomes and we have a means of increasing GMA. The benefits of increases in general mental ability generalize to other areas (career, academic success, good life choices), whereas non-general gains will be limited. As an extreme example, it is obvious why giving children Raven Progressive Matrices is not going to make them drive better. But evidence suggests having higher IQ, reduces risk of traffic accident.
Regarding Richard Lynn: I address this within the article. Until his death, Lynn and colleagues where updating the NIQ scores. Looking at the most recent version on ViewOnIQ from Becker, I see that Lynn has excluded all those samples for the Nigerian estimate and incorporated the Maqsud estimates. He has also included several other more recent estimates, but arrived at a similar estimate still. You can view the samples used and estimates here if you download the file. More importantly, focusing on Lynn is a mistake as I mention in the article. Other less controversial researchers estimate “universal basic skills” or “harmonized learning outcomes” and produce estimates which correlate highly with the NIQ estimates. See the chart from Warne 2023. A side point, but Wikipedia is politically biased. I intentionally capitalized the L to give credit as Richard Lynn’s discovery preceeded Flynn’s first publication. Although, his discovery was preceeded by Runquist.
On The Bell Curve: You say “That book has so many problems that instead of typing it all out I would like to direct people to this video which points out a lot of them.” I don’t plan on watching the 2 hour 39 minute video just to respond to you. At the time, a large number of claims (like the one I make) was not particularly controversial among intelligence researchers (see Gottfredson and APA response). I discuss this in the article. Furthermore, the more recent Rindermann et al. (2020) found in a survey among intelligence experts that many believed SES was substantially explained by intelligence (see in the article). I am also going to make a point about isolated demands for rigor. You are dismissive of some the researchers and journals for lack of academic quality, but in response to a book co-authored by a Harvard psychologist, you give me a YouTube video with a psuedonymous guy with a skull avatar that is part of LeftTube or “Bread Tube”. I am not suggesting this means that I can merely dismiss anything that he’s saying, but I will admit that this feels like a double standard given the Mankind Quarterly critique.
On environmental influence: You say “I don’t think anyone thinks the environment explains 100%, but given that it’s much larger and has many more variables it seems reasonable to assume it can explain more of it.” That doesn’t make sense as an argument in my view. We don’t know that it is larger. Within the USA, the heritability is over 50% in adulthood indicating that genetic differences are the largest driving force among all known influence. Whether that is true of international differences is not entirely clear. The existence of more possible environmental explanations doesn’t mean that it has more explanatory variables. Analogously, this argument would be wrong: There are over a billion base pairs in the genome, therefore genetic explanations are better. The issue is that the environmental influence largely, although not entirely, consists of unsystematic differences. Iodine is an exception and should be given in areas with deprivation. But the gains from embryo selection in large batches is beyond 9 points, and could be used on the non-iodine deficient population.
I have responded. I don’t know if you will be brigaded. I have not personally downvoted you.
It seems I didn’t get brigaded [tap on wood], but I still feel uneasy answering this. You got some downvotes on this comment initially which means the karma system pushes you to not reply, in the same way it pushed me to not reply to the HBD-proponents I was debating. This voting-power-by-popularity system doesn’t incentivize having conversations, so feel free to answer in the comment section on your substack instead. I will edit in a link to it at the end of this comment if you do so. This comment is going to be shorter anyway.
Firstly, I wanted to say that I also didn’t downvote your post because while I disagree I do sympathize with the amount of effort that went into it, and this karma system would punish future unrelated posts and comments in a guilt-by-association-fashion, not just this post (although I did give it a disagreement vote since that influences nothing):
Secondly, to clear up any confusion I don’t think a journal being made by horrible people allows you to conclude that their conclusions are false, but I do think it allows you to not give them any money.
Thirdly, I think we run into the same issue with g as we do with intelligence. If g is just correlation between different cognitive tasks, then the natural question is, which tasks? And which tasks are considered ‘cognitive tasks’? Because the results will differ based on what you choose.
Fourthly, I think the difference with linking a video and linking the mankind quarterly is primarily money, one costs 75 dollar and the other is free. For the record I have watched the video in it’s entirety and am not just throwing something at you while I myself don’t know of any counterarguments. I could’ve typed them out, but I just don’t think I have much to add both in terms of information nor presentation. For those who are familiar with the subject you can skip to 1:02:11 of the video at which point he really starts diving into their methodology instead of giving a general overview. Which ties into...
Fifthly, I think the video points out a general pattern of Lynn and his colleagues using a clearly cherrypicked dataset then being called out on it, at which point they switch to another slightly less clearly cherrypicked dataset which people then call them out on etc. Now maybe this latest dataset they use is genuinely good but I think this is a ‘boy who cried wolf’ scenario where I just no longer think it’s prudent to trust them or trust that reading their work is a productive use of time.
Lastly, maybe at some point we run out of good environmental interventions (like iodine) and maybe then (assuming some premises) it becomes prudent to switch to genetic interventions. But until that time we should focus on those environmental interventions, not just because of their immediate cost effectiveness but also because of one of my points you didn’t address, namely that those environmental interventions are way more egalitarian/emancipatory, which produces better results in the longterm.
I will respond here because it’s important for everyone to see.
You don’t need to give the journal money. I am offering to email you the pdf if you are that interested.
Cognitively demanding tasks. These require puzzle-solving, reasoning, drawing on past knowledge, connecting ideas, etc. As long as the test has a wide range of tasks like this, estimates will be similar. Provided they are cognitively demanding and diverse, results are not particularly sensitive to the actual content of the test for native speakers. Spearman called this the “indifference of the indicator.” You can read more in Chapter 7 of Rusell Warne’s book In The Know.
Another interpretation of what Lynn is doing is improving his estimates when people critique him. As I mentioned in the article, many other researchers estimating national averages on mental ability tests produce moderate to highly correlated estimates with Lynn’s. Why do you think that is? And who do you think has more accurate estimates? If you have a specific objection to the scores, I can respond with estimates of correlations after making adjustments. Ee can use the ViewOnIQ data to drop scores/samples from countries you find are bad. Or we can Winsorize the scores and check the strength of the relationship. Or we can look at Rindermann’s estimates.
What do you think, in your view, is the correlation between average cognitive ability and log(GDP/c)? What are you basing this estimate on, and why is it better than Lynn, Becker, Rindermann, Angrist, etc?
Iodine deficiency will not work for everyone, only those who are iodine-deprived. Where people are iodine deficient, we should try to help them. I focus on genetic enhancement because it is under-considered. The possible returns from genetic enhancement will be unevenly distributed but have the potential to be absolutely massive. Since parents will adopt it voluntarily and IVF is largely legal worldwide (and hopefully IVG will be), there is a plausible means of improving humanity immensely merely through funding research in a narrow area to accelerate certain discoveries.