Rationalizing having children on the ground that the children of a smart person will be smart and therefore an advantage to the world ... isn't very smart.
First, there's regression to the mean, which says the children of an above-average smart person won't be likely to be measurably above average (true also for a person less intelligent than average).
Second, as the planet becomes more crowded, the intelligence, wittiness or saintliness of the offspring stops mattering when they're all starving.
I'm not sure that's how RTM works. As I understand it, RTM is talking about the distribution of an uncorrelated variable and says that they'll trend towards the average on the second measurement. It doesn't seem to say that they'll be average with respect to a given correlate. Some schools are persistently better over time, for instance even if on the second measurement they trend closer to the average and then subsequently further away and then... that's just saying that it's not a perfect correlate and you've got noise in your data.
In this case the assumption seems to be that, measured by the intelligence of the first person as an outlier, IQ's going to be average on the second measurement - their kid. However, IQ doesn't necessarily trend that way in family lines over time. If IQ really does have a heritable genetic component, then, while people might trend up and down, over time the aggregate is going to be persistently higher for some than for others.
Actually, yes, that's how it works. An individual person who is smarter (or dumber) than his lineage isn't likely to have children equally or even significantly smarter (or dumber).
Think of a person as being the offshoot, not of his parents, but of 1000 predecessors including his parents. Now think of that person's children also being the offshoot of 998 predecessors plus his 2 parents. Work out the probability of inheriting any significant percentage of the high intelligence of the 2 compared to the 998.
This example is meant more to convey the idea than be exactly representative.
> while people might trend up and down, over time the aggregate is going to be persistently higher for some than for others.
Yes, but as explained above and accepting that premise, statistics argues against the idea that a smart person will have children significantly smarter than the population as a whole.
Also, bottom line and in the long term, intelligent starving children are still starving children.
> Think of a person as being the offshoot, not of his parents, but of 1000 predecessors including his parents. Now think of that person's children also being the offshoot of 998 predecessors plus his 2 parents. Work out the probability of inheriting any significant percentage of the high intelligence of the 2 compared to the 998.
Maybe the 998 were above average too. Regression to the mean would be the way to bet if you didn't have any other data - but how heritable is intelligence?
> Yes, but as explained above and accepting that premise, statistics argues against the idea that a smart person will have children significantly smarter than the population as a whole.
Wouldn't that only be the case if you take the population as a whole as their lineage?
> Also, bottom line and in the long term, intelligent starving children are still starving children.
I'd expect smart people to provide significant better for their children down the line than will stupid people. If you're looking at the future of humanity a small, dominant group that better fulfils your preferences is going to be a lot more worthwhile for some people.
> Maybe the 998 were above average too. Regression to the mean would be the way to bet if you didn't have any other data - but how heritable is intelligence?
Yes, all valid points, but to some extent set aside by the issue that, on an overcrowded planet, ten additional intelligent people isn't necessarily better than one additional intelligent person.
> Wouldn't that only be the case if you take the population as a whole as their lineage?
No, any grouping larger than two begins to show the effect of regression to the mean. My choice of 1000 was rather arbitrary, just to show the idea, and knowing the "right" number would require us to understand more about IQ heritability factors than we actually know.
I don't think assessing the entire population of a species is required for this idea to work as it does in nature. Indeed, on the contrary -- in this example, all one need do is group people by the genetic traits that lead to a particular IQ, and the heritability patterns of those genes.
> I'd expect smart people to provide significant better for their children down the line than will stupid people.
I might be tempted to agree on emotional grounds, but there's some circumstantial evidence that this isn't so, if we accept mental illness statistics and suicide rates at face value. Also, arguing that smart people should be allowed to have more children (an argument you aren't making) has some obvious logical difficulties, not least of which is that everyone thinks their children are above average.
> If you're looking at the future of humanity a small, dominant group that better fulfils your preferences is going to be a lot more worthwhile for some people.
Not very clear what you're saying here, but a wide genotype has been shown to produce better (healthier, fitter) results than a narrow one.
Comments
Rationalizing having children on the ground that the children of a smart person will be smart and therefore an advantage to the world ... isn't very smart.
First, there's regression to the mean, which says the children of an above-average smart person won't be likely to be measurably above average (true also for a person less intelligent than average).
Second, as the planet becomes more crowded, the intelligence, wittiness or saintliness of the offspring stops mattering when they're all starving.
I'm not sure that's how RTM works. As I understand it, RTM is talking about the distribution of an uncorrelated variable and says that they'll trend towards the average on the second measurement. It doesn't seem to say that they'll be average with respect to a given correlate. Some schools are persistently better over time, for instance even if on the second measurement they trend closer to the average and then subsequently further away and then... that's just saying that it's not a perfect correlate and you've got noise in your data.
In this case the assumption seems to be that, measured by the intelligence of the first person as an outlier, IQ's going to be average on the second measurement - their kid. However, IQ doesn't necessarily trend that way in family lines over time. If IQ really does have a heritable genetic component, then, while people might trend up and down, over time the aggregate is going to be persistently higher for some than for others.
> I'm not sure that's how RTM works.
Actually, yes, that's how it works. An individual person who is smarter (or dumber) than his lineage isn't likely to have children equally or even significantly smarter (or dumber).
Think of a person as being the offshoot, not of his parents, but of 1000 predecessors including his parents. Now think of that person's children also being the offshoot of 998 predecessors plus his 2 parents. Work out the probability of inheriting any significant percentage of the high intelligence of the 2 compared to the 998.
This example is meant more to convey the idea than be exactly representative.
> while people might trend up and down, over time the aggregate is going to be persistently higher for some than for others.
Yes, but as explained above and accepting that premise, statistics argues against the idea that a smart person will have children significantly smarter than the population as a whole.
Also, bottom line and in the long term, intelligent starving children are still starving children.
> Think of a person as being the offshoot, not of his parents, but of 1000 predecessors including his parents. Now think of that person's children also being the offshoot of 998 predecessors plus his 2 parents. Work out the probability of inheriting any significant percentage of the high intelligence of the 2 compared to the 998.
Maybe the 998 were above average too. Regression to the mean would be the way to bet if you didn't have any other data - but how heritable is intelligence?
> Yes, but as explained above and accepting that premise, statistics argues against the idea that a smart person will have children significantly smarter than the population as a whole.
Wouldn't that only be the case if you take the population as a whole as their lineage?
> Also, bottom line and in the long term, intelligent starving children are still starving children.
I'd expect smart people to provide significant better for their children down the line than will stupid people. If you're looking at the future of humanity a small, dominant group that better fulfils your preferences is going to be a lot more worthwhile for some people.
> Maybe the 998 were above average too. Regression to the mean would be the way to bet if you didn't have any other data - but how heritable is intelligence?
Yes, all valid points, but to some extent set aside by the issue that, on an overcrowded planet, ten additional intelligent people isn't necessarily better than one additional intelligent person.
> Wouldn't that only be the case if you take the population as a whole as their lineage?
No, any grouping larger than two begins to show the effect of regression to the mean. My choice of 1000 was rather arbitrary, just to show the idea, and knowing the "right" number would require us to understand more about IQ heritability factors than we actually know.
I don't think assessing the entire population of a species is required for this idea to work as it does in nature. Indeed, on the contrary -- in this example, all one need do is group people by the genetic traits that lead to a particular IQ, and the heritability patterns of those genes.
> I'd expect smart people to provide significant better for their children down the line than will stupid people.
I might be tempted to agree on emotional grounds, but there's some circumstantial evidence that this isn't so, if we accept mental illness statistics and suicide rates at face value. Also, arguing that smart people should be allowed to have more children (an argument you aren't making) has some obvious logical difficulties, not least of which is that everyone thinks their children are above average.
> If you're looking at the future of humanity a small, dominant group that better fulfils your preferences is going to be a lot more worthwhile for some people.
Not very clear what you're saying here, but a wide genotype has been shown to produce better (healthier, fitter) results than a narrow one.