You are a clown. I am sure you will not answer me. It is now irrefutably demonstrated and replicated in widely varying research designs that genetics is tens to hundreds of times more important than genetics for almost everything. For example, we have a study of adoptive siblings that compared biological siblings and estimated that heritability is 40 times more important than shared environment. We have a study of Swedish twins raised apart that indicates that heritability is 40 times more important than shared environment for IQ. We have a meta-analysis of five independent studies indicating that shared environment contributes zero to criminal behavior. We have a study of adoptees that reported zero correlation between adoptive siblings for criminal behavior, indicating a genetic impact potentially hundreds of times greater than environmental. I can go on and on.
If it is not genetics, what causes biological relatives to be similar and adoptive relatives to be completely dissimilar? Magic? Explanations such as prenatal environment, epigenetics and postnatal environment are completely refuted.
Real and stable heritability is close to 100%, measurement error and transient events explain the non-shared environmental variance.
The general public tends to equate “heritability” with “inheritability.” I would argue that “inheritability” is a biological term, and that “heritability” is more accurately understood as a statistical term.
There is money, prestige, fame and a hand some to be made with phrasings such as these, which at some point will be read not by the technical some, but by the easily impressed some, and here I am contributing to its fame or infamy, some, how?:
"These results overhaul the state of scientific knowledge on the genetic etiology of variation in human personality, establishing a rigorous basis for genetic inference and a fundamental role of personality genetics in the human condition."
So, I don´t think it is just questionable behavior by researchers which would be unneeded?.
Because if as you elaborated it is not needed, then it is a choice.
Unless it is not a choice and you are either wrong in your approach, or incomplete.
Or neither: Contradictory behavior would not be outlandish, it is part and parcel of Human Behavior: love and hate are present at the same time in the same person in the same circumstance.. or not!, consistency in Personality is truly literally overrated and overstated.
In emotions and personalities contradictions do not seem to make the perception of the whole field of studying behavior as illogical, merely the presence of contradictions renders something irrational, illogical, unempirical and unscientific. By the strictest definition of consistency, having no contradictions. Which explains, truly?, the other sense of consistency, the popular one!: non-replicable, non-explanatory, non-predicting, and unpredictable results...
Except for the predictable more is coming, we can do better!: that´s just Tradition. Creatures of Habits, good, back and Monastic...
Or am I wrong on that one?
And if I am not?: why applying something ostensible Scientific to something ostensible irrational, specially without Causal Models without enough explanatory and predicting power?.
--"With or without you, without you.. Logic.. I can live..."
--"No, no, no, just multiply the inconsistencies, square them even!, use a power law! wherever needed!, and in the long haul we are hoping someone or they will cancel out..."
Dear Eric. I am trying to wrap my head around the "phenotypic null hypothesis". I even tried to clarify it with ChatGPT. It suggested an analogy, that I found somehow helpful but I am not sure if this is what you mean by it. The idea is "imagine studying crime in a city and finding that neighborhoods with more broken streetlights have higher crime rates. Some lightbulbs might even be especially predictive. But of course, the broken lights don’t cause the crime, they just reflect deeper conditions like poverty or neglect.". So, in the same way, maybe SNPs correlate with personality traits (not as false positives), but not because they’re mechanistically shaping the trait, but because the trait already exists as stable, emergent patterns. And genetic variation among individuals just happens to statistically tag them. Does that get at what you mean by genes and environments “going along for the ride” of the phenotype?
Yes, I think that is a reasonable example. I always like a version from a great philosophical paper by Craver and Bechtel (https://tinyurl.com/4jy22h9b). They imagine Iggy, a hotdog vendor with a cart full of hot dogs. Iggy decides business will be better on the other side of the street, so he crosses over. The question is, why did the hot dogs in the cart cross the street? And the answer is, Because Iggy made the decision. Trying to explain Iggy's movement in terms of the hotdogs would be hopeless. The point is that our intuition that small parts always explain the behavior of large parts is incorrect. Sometimes it is the other way around. So if the SNPs associated with conscientiousness are correlated with the SNPs associated with grooming, it doesn't show that the "same genes" cause both of them. It shows that being conscientious causes better grooming. The SNPs are hotdogs.
“By the way, I don’t really buy inflating the heritabilities by correcting them for unreliability. This is an old method that Arthur Jensen used to use a lot to produce more impressive looking heritabilities. What I was taught in grad school is that it makes sense to “correct” a measure if there is some plausible possibility of actually measuring it more accurately in the future— you are predicting a future result, not attributing reality to some idealized abstraction.”
This is interesting, can you say more? In a draft of my paper with Tobias Edwards we looked at corrected coefficients (on number of children) from (eg) EA2 and how well they predicted EA3. It didn’t make it into the final version. But isn’t the future more accurate measure going to be polygenic scores from whole genome sequencing? That doesn’t seem unreasonable… a more relevant word might be, do the assumptions of errors in variables hold, or is the noise correlated with “other stuff”?
They aren't correcting for errors in the polygenic scores, they are correcting for error in the measurement of the FFM. The idea of "correcting for attenuation" was developed years ago, mostly as a way of anticipating changes in reliability related to making the test longer, using the Spearman-Brown formula. (I'm really showing my age here.) The problem with this kind or correction is that it replaces an observed score on the FFM, with a latent variable. Nothing wrong with that, but it can produce unrealistic results. See for example my thread about the genetics of executive function (https://bsky.app/profile/ent3c.bsky.social/post/3lcmhfja6w725). Another way to say essentially the same thing is that once you analyze personality as a latent rather than an observed variable, you lose the ability to assign an individual score to each participant-- that is what it means for a variable to be "latent". If you could assign a number corresponding to each person's "true score" of course you would do so, but then there would be no need for the whole exercise. Finally, getting back to your strategy of anticipating polygenic validity based on whole exome sequencing, as I understand it the relation between heritability and performance of a polygenic score can be understood in terms of an analytic formula. That's why I mentioned that the less-than-dramatic heritabilities put a limit on our future expectations. More generally, I think it is a mistake for behavioral genomics to always be pointing to what they are sure to be able to accomplish in the future. If whole exome sequencing eventually allow us to predict a whole lot of variance in extraversion, well, I'll deal with it when it happens. There have been a lot of unfulfilled promises.
Nice summary, Eric. Thanks!
You are a clown. I am sure you will not answer me. It is now irrefutably demonstrated and replicated in widely varying research designs that genetics is tens to hundreds of times more important than genetics for almost everything. For example, we have a study of adoptive siblings that compared biological siblings and estimated that heritability is 40 times more important than shared environment. We have a study of Swedish twins raised apart that indicates that heritability is 40 times more important than shared environment for IQ. We have a meta-analysis of five independent studies indicating that shared environment contributes zero to criminal behavior. We have a study of adoptees that reported zero correlation between adoptive siblings for criminal behavior, indicating a genetic impact potentially hundreds of times greater than environmental. I can go on and on.
If it is not genetics, what causes biological relatives to be similar and adoptive relatives to be completely dissimilar? Magic? Explanations such as prenatal environment, epigenetics and postnatal environment are completely refuted.
Real and stable heritability is close to 100%, measurement error and transient events explain the non-shared environmental variance.
The general public tends to equate “heritability” with “inheritability.” I would argue that “inheritability” is a biological term, and that “heritability” is more accurately understood as a statistical term.
No?
There is money, prestige, fame and a hand some to be made with phrasings such as these, which at some point will be read not by the technical some, but by the easily impressed some, and here I am contributing to its fame or infamy, some, how?:
"These results overhaul the state of scientific knowledge on the genetic etiology of variation in human personality, establishing a rigorous basis for genetic inference and a fundamental role of personality genetics in the human condition."
So, I don´t think it is just questionable behavior by researchers which would be unneeded?.
Because if as you elaborated it is not needed, then it is a choice.
Unless it is not a choice and you are either wrong in your approach, or incomplete.
Or neither: Contradictory behavior would not be outlandish, it is part and parcel of Human Behavior: love and hate are present at the same time in the same person in the same circumstance.. or not!, consistency in Personality is truly literally overrated and overstated.
In emotions and personalities contradictions do not seem to make the perception of the whole field of studying behavior as illogical, merely the presence of contradictions renders something irrational, illogical, unempirical and unscientific. By the strictest definition of consistency, having no contradictions. Which explains, truly?, the other sense of consistency, the popular one!: non-replicable, non-explanatory, non-predicting, and unpredictable results...
Except for the predictable more is coming, we can do better!: that´s just Tradition. Creatures of Habits, good, back and Monastic...
Or am I wrong on that one?
And if I am not?: why applying something ostensible Scientific to something ostensible irrational, specially without Causal Models without enough explanatory and predicting power?.
--"With or without you, without you.. Logic.. I can live..."
--"No, no, no, just multiply the inconsistencies, square them even!, use a power law! wherever needed!, and in the long haul we are hoping someone or they will cancel out..."
--"Eventually, in any Event..."
Dear Eric. I am trying to wrap my head around the "phenotypic null hypothesis". I even tried to clarify it with ChatGPT. It suggested an analogy, that I found somehow helpful but I am not sure if this is what you mean by it. The idea is "imagine studying crime in a city and finding that neighborhoods with more broken streetlights have higher crime rates. Some lightbulbs might even be especially predictive. But of course, the broken lights don’t cause the crime, they just reflect deeper conditions like poverty or neglect.". So, in the same way, maybe SNPs correlate with personality traits (not as false positives), but not because they’re mechanistically shaping the trait, but because the trait already exists as stable, emergent patterns. And genetic variation among individuals just happens to statistically tag them. Does that get at what you mean by genes and environments “going along for the ride” of the phenotype?
Yes, I think that is a reasonable example. I always like a version from a great philosophical paper by Craver and Bechtel (https://tinyurl.com/4jy22h9b). They imagine Iggy, a hotdog vendor with a cart full of hot dogs. Iggy decides business will be better on the other side of the street, so he crosses over. The question is, why did the hot dogs in the cart cross the street? And the answer is, Because Iggy made the decision. Trying to explain Iggy's movement in terms of the hotdogs would be hopeless. The point is that our intuition that small parts always explain the behavior of large parts is incorrect. Sometimes it is the other way around. So if the SNPs associated with conscientiousness are correlated with the SNPs associated with grooming, it doesn't show that the "same genes" cause both of them. It shows that being conscientious causes better grooming. The SNPs are hotdogs.
Thanks a lot! I will read the paper!
“By the way, I don’t really buy inflating the heritabilities by correcting them for unreliability. This is an old method that Arthur Jensen used to use a lot to produce more impressive looking heritabilities. What I was taught in grad school is that it makes sense to “correct” a measure if there is some plausible possibility of actually measuring it more accurately in the future— you are predicting a future result, not attributing reality to some idealized abstraction.”
This is interesting, can you say more? In a draft of my paper with Tobias Edwards we looked at corrected coefficients (on number of children) from (eg) EA2 and how well they predicted EA3. It didn’t make it into the final version. But isn’t the future more accurate measure going to be polygenic scores from whole genome sequencing? That doesn’t seem unreasonable… a more relevant word might be, do the assumptions of errors in variables hold, or is the noise correlated with “other stuff”?
They aren't correcting for errors in the polygenic scores, they are correcting for error in the measurement of the FFM. The idea of "correcting for attenuation" was developed years ago, mostly as a way of anticipating changes in reliability related to making the test longer, using the Spearman-Brown formula. (I'm really showing my age here.) The problem with this kind or correction is that it replaces an observed score on the FFM, with a latent variable. Nothing wrong with that, but it can produce unrealistic results. See for example my thread about the genetics of executive function (https://bsky.app/profile/ent3c.bsky.social/post/3lcmhfja6w725). Another way to say essentially the same thing is that once you analyze personality as a latent rather than an observed variable, you lose the ability to assign an individual score to each participant-- that is what it means for a variable to be "latent". If you could assign a number corresponding to each person's "true score" of course you would do so, but then there would be no need for the whole exercise. Finally, getting back to your strategy of anticipating polygenic validity based on whole exome sequencing, as I understand it the relation between heritability and performance of a polygenic score can be understood in terms of an analytic formula. That's why I mentioned that the less-than-dramatic heritabilities put a limit on our future expectations. More generally, I think it is a mistake for behavioral genomics to always be pointing to what they are sure to be able to accomplish in the future. If whole exome sequencing eventually allow us to predict a whole lot of variance in extraversion, well, I'll deal with it when it happens. There have been a lot of unfulfilled promises.