I think linear regression is very hard to read in this sort of study. I personally would group number of children into: 0; 1; 2; 3; 4 or more. I would also be careful about having the adults be age 40 or more in the year surveyed (maybe even 45 or more). Then I would look at transition probabilities. How many people who came from a family with 1 child ended up with 0 children, 1 child, 2, children, 3 children, or 4 or more children? The same with people who came from a family of 2 children; of three children; of 4 or more children.
Then you might look at how that changes by birth year.
I just finished reading the Eric Turkheimer’s new book, called “Understanding the nature-nurture debate”, and I would strongly recommend it to you or anyone. I would be interested to hear you react to its arguments. He is also on substack.
I came away from that book thinking that arguments like the one you make in this blog post: neatly partitioning effects into “genes”/“nature” and “environment”/“nurture” dont really make a lot of sense from a causal, scientific perspective. I could say more, but i think my summary would be inadequate compared to Turkheimer’s short, well-written book.
Anyway, hope you can give it a look sometime and see if it changes your perspective, or at least gives you more material to understand a contrary perspective to yours.
An addendum to be just a little more on-target in responding to your actual post:
What i interpret to be your general point about nature/nurture and childrearing is not directly contradicted by Turkheimer’s perspective. We may well have less control over our children’s behavioral outcomes than we like to think, and genetics are likely an important reason for that fact.
But it would also be wrong to conclude from twin studies that somehow genetics dominates “nurture” or environment in determining complex behaviors like fertility. Similarly, it would also be wrong to conclude that parenting in particular has little to no effect on fertility outcomes because genetic effects dominate. The right answer is that we just don’t know and may never know, because 1. There is vanishingly little evidence the particular genes associated with any complex behavioral outcome (even with increasing GWAS samples), and 2. we cannot perform the experiment(s) that are necessary to make progress on many such questions (e.g. systematically and permanently altering childrearing environments in particular ways).
So the massive drop in teen pregnancy in the last couple decades in the US represents a genetic shift and not something different in the environment? That seems highly implausible to me
Doesnt this look like people looking at what kind of social envirement seems good or natural, and then mirroring it?
If they dont see any big families, then thats alien. If they see big families or grew up in one and feel that is good/natural, then one is more likely to do it.
One way to say this would essentially just be culture.
(I have met some people who dislikes high fertility because they look at some examplenof massive familyes in history and find them bad. They havnt done any cost calculations or moral considerations, just going off appearences and the emotions they get from them.
But hey could get defensive if i would discuss this all with them)
Unfortunately, it's not possible to measure how many children the offspring of people who have no chidren typically have. Does that distort the statistics at all? All the people in the survey are the offspring of people who decided to have at least one child.
I think linear regression is very hard to read in this sort of study. I personally would group number of children into: 0; 1; 2; 3; 4 or more. I would also be careful about having the adults be age 40 or more in the year surveyed (maybe even 45 or more). Then I would look at transition probabilities. How many people who came from a family with 1 child ended up with 0 children, 1 child, 2, children, 3 children, or 4 or more children? The same with people who came from a family of 2 children; of three children; of 4 or more children.
Then you might look at how that changes by birth year.
My father would tell us that he wants 121 grandchildren, 11 from each of his 11 children. He only got 52.
wow
Hi Bryan,
Interesting post.
I just finished reading the Eric Turkheimer’s new book, called “Understanding the nature-nurture debate”, and I would strongly recommend it to you or anyone. I would be interested to hear you react to its arguments. He is also on substack.
I came away from that book thinking that arguments like the one you make in this blog post: neatly partitioning effects into “genes”/“nature” and “environment”/“nurture” dont really make a lot of sense from a causal, scientific perspective. I could say more, but i think my summary would be inadequate compared to Turkheimer’s short, well-written book.
Anyway, hope you can give it a look sometime and see if it changes your perspective, or at least gives you more material to understand a contrary perspective to yours.
Noah
An addendum to be just a little more on-target in responding to your actual post:
What i interpret to be your general point about nature/nurture and childrearing is not directly contradicted by Turkheimer’s perspective. We may well have less control over our children’s behavioral outcomes than we like to think, and genetics are likely an important reason for that fact.
But it would also be wrong to conclude from twin studies that somehow genetics dominates “nurture” or environment in determining complex behaviors like fertility. Similarly, it would also be wrong to conclude that parenting in particular has little to no effect on fertility outcomes because genetic effects dominate. The right answer is that we just don’t know and may never know, because 1. There is vanishingly little evidence the particular genes associated with any complex behavioral outcome (even with increasing GWAS samples), and 2. we cannot perform the experiment(s) that are necessary to make progress on many such questions (e.g. systematically and permanently altering childrearing environments in particular ways).
So the massive drop in teen pregnancy in the last couple decades in the US represents a genetic shift and not something different in the environment? That seems highly implausible to me
Probably with birth control, it's more and more of a choice to have lots of kids, so the people who do it really go all-in.
Doesnt this look like people looking at what kind of social envirement seems good or natural, and then mirroring it?
If they dont see any big families, then thats alien. If they see big families or grew up in one and feel that is good/natural, then one is more likely to do it.
One way to say this would essentially just be culture.
(I have met some people who dislikes high fertility because they look at some examplenof massive familyes in history and find them bad. They havnt done any cost calculations or moral considerations, just going off appearences and the emotions they get from them.
But hey could get defensive if i would discuss this all with them)
I must be dense but... if high fertility is a heritable trait, why is overall fertility declining?
If high fertility IS heritable , then the simplest explenation is simply that it takes a lot of time for evolution to take hold
Unfortunately, it's not possible to measure how many children the offspring of people who have no chidren typically have. Does that distort the statistics at all? All the people in the survey are the offspring of people who decided to have at least one child.
What's the difference between B and Beta in the regression results tables?