Evolution vs Regression to the mean.

[labmath2]

I started thinking about this in relation to the IQ problem. Will blacks that are smarter, given enough time, produce descendants that are around the average? Wouldn't that also mean diaspora Jews have always been smarter, otherwise they would have regressed to the mean?

 

I looked up regression to the mean vs evolution on google and this is the best answer i found.

 

"TL;DR answer: Regression to the mean doesn't always occur (children can be taller than either parent, as can be easily observed at any high-school basketball game), so it doesn't affect the theory of evolution through natural selection.

The observation of regression to the mean was of great concern to Darwin and early thoughts on natural selection (don't confuse "evolution", which is the observed phenomenon, and "natural selection", which is one of the theories proposed to explain evolution). The problem has been effectively resolved since the rediscovery of Mendel's genetic work at the end of the 19th century.

"Regression to the mean" is inevitable if inheritance works through blending of features. That was how everyone believed inheritance to work in the mid-1800s, and Darwin understood that it presented an insuperable obstacle to his theory. That is, if blending inheritance was true, then natural selection could not be true. Darwin puzzled over this quite a bit, and came up with some unsatisfying suggestions to overcome it, but at the end of the day, one of the predictions that his theory of evolution through natural selection made, was that blending inheritance could not be true.

Of course, blending inheritance is not true. Mendel showed that in fact inheritance is quantal, not blending. We now know that many traits of interest look, superficially, as if they're blending (which leads to regression to the mean), but are actually the result of multiple traits interacting.

With blending inheritance, the variation that is required for natural selection to work is lost, generation after generation; regression to the mean in an inevitable consequence. With gene-based inheritance, variation (can be) preserved generation after generation; regression to the mean is a common but not inevitable consequence, because the variation in the original population is still present and can be re-created given appropriate selection pressure."

link: http://biology.stackexchange.com/questions/41982/regression-to-the-mean-and-evolution

This response can explain why we can still get taller kids from two tall parents. Given the fact that height is affected by multiple genes, aren't we getting a "blending" effect, even if the mechanism is not actual blending? I am curious if anyone has more info about this.

P.S. i would like to see Stefan do a video on it, but given the current events, i can see why he may be a little more busy with other things.

[st434u]

Regression to the mean is a phenomenon that happens because every time you have a new lifeform coming into being, you have two sets of genes that recombine. Because those genes have recombined a new time for each generation, they are used to working together in certain ways, to produce certain lifeforms as a result of their combinations.

 

People are used to thinking of features, such as tall, as being caused by a "tall gene", but in reality it's generally a combination of genes that produced a tall person, and another combination of genes won't, even if most of those genes are present.

 

You can see this very clearly in race horse breeding. Let's simplify it down to speed. These horses are generally bred from both a male and a female winning race horses, which are very fast, to increase the likelihood that some of the offspring will be winners (fast horses). But on average, most of the offspring will be slower than both the father and the mother, because of regression to the mean. However, some of the offspring will be just as fast, and even fewer still might be even faster than their parents. Both the mother and the father, because they are so fast, represent a significant deviation from the mean, which is to say that if you took all of their recent ancestors, and you averaged their speed, both the mother and the father would be much faster than the average. Regression to the mean means that there will be a tendency for their offspring to revert to a level that is closer to the mean, although never reaching it on average. That is, assuming you could breed thousands of horses from the same pair of a top level winning male and female race horses, and you took an average of their speed, what you would see is that even though the average offspring is slower than both their parents, the average offspring is also faster than the average for race horses, and faster than the average for both the father and the mother's recent ancestors, whichever way you want to look at it.

 

So it's not either regression to the mean or evolution; in fact it could be argued that regression to the mean helps evolution work better in the long term because it smoothes out drastic changes so that it's harder to get massive changes that pile on top of each other over few generations, which could render the lifeform unfit.

[labmath2]

Regression to the mean is a phenomenon that happens because every time you have a new lifeform coming into being, you have two sets of genes that recombine. Because those genes have recombined a new time for each generation, they are used to working together in certain ways, to produce certain lifeforms as a result of their combinations.

 

People are used to thinking of features, such as tall, as being caused by a "tall gene", but in reality it's generally a combination of genes that produced a tall person, and another combination of genes won't, even if most of those genes are present.

 

You can see this very clearly in race horse breeding. Let's simplify it down to speed. These horses are generally bred from both a male and a female winning race horses, which are very fast, to increase the likelihood that some of the offspring will be winners (fast horses). But on average, most of the offspring will be slower than both the father and the mother, because of regression to the mean. However, some of the offspring will be just as fast, and even fewer still might be even faster than their parents. Both the mother and the father, because they are so fast, represent a significant deviation from the mean, which is to say that if you took all of their recent ancestors, and you averaged their speed, both the mother and the father would be much faster than the average. Regression to the mean means that there will be a tendency for their offspring to revert to a level that is closer to the mean, although never reaching it on average. That is, assuming you could breed thousands of horses from the same pair of a top level winning male and female race horses, and you took an average of their speed, what you would see is that even though the average offspring is slower than both their parents, the average offspring is also faster than the average for race horses, and faster than the average for both the father and the mother's recent ancestors, whichever way you want to look at it.

 

So it's not either regression to the mean or evolution; in fact it could be argued that regression to the mean helps evolution work better in the long term because it smoothes out drastic changes so that it's harder to get massive changes that pile on top of each other over few generations, which could render the lifeform unfit.

 

At first when I read this response i wanted it to make sense, so i rationalized it in my head, but its not satisfactory. I already understood that many features of an organism is a result of multiple genes. The problem i am having is how you could get vastly different species given this regression. If most (if not all) significant variation is regressed in a few generations, then don't you simply get more variations from the average? By that i mean all possible variation of a species will be present in the group, with nature affecting the ratio. Take color for example, if we have a range of color for a species from red to yellow. It would seem you can only get more variations of color in the group (maybe darker red and lighter yellow becomes a possibility over time, and in rare cases green gets added to the possibility) and the ratio of the population color changes depending on environment. Lets assume the yellow members are more vulnerable to predation, then the ratio of yellow to red will be lower. Given the fact that their color is a combination of genes, it will be almost impossible to get rid of any color variation since all members have the potential to give birth to any color (to greater or lesser degree depending on their own colors). It would require a strong environmental change over a long period of time or a series of concerted changes that push the species in one evolutionary direction to produce significantly different species (by concerted i don't mean design). Even then, they might still not be that different from their ancestors since regression to the mean will still mitigate the change, though the mean will slowly shift over many generations. 

 

Maybe i am just intellectualizing, but i find this topic important for the discussion about the IQ gains of Jews. Unless the Ashkenazi Jews engaged in some form of eugenics (i don't mean this in a derogatory way, just in the sense that people were really picky about their mating choices and the less intelligent were significantly under-reproducing), i find the natural history account for the gain suspect.

[ObserveandReport]

At first when I read this response i wanted it to make sense, so i rationalized it in my head, but its not satisfactory. I already understood that many features of an organism is a result of multiple genes. The problem i am having is how you could get vastly different species given this regression. If most (if not all) significant variation is regressed in a few generations, then don't you simply get more variations from the average? By that i mean all possible variation of a species will be present in the group, with nature affecting the ratio. Take color for example, if we have a range of color for a species from red to yellow. It would seem you can only get more variations of color in the group (maybe darker red and lighter yellow becomes a possibility over time, and in rare cases green gets added to the possibility) and the ratio of the population color changes depending on environment. Lets assume the yellow members are more vulnerable to predation, then the ratio of yellow to red will be lower. Given the fact that their color is a combination of genes, it will be almost impossible to get rid of any color variation since all members have the potential to give birth to any color (to greater or lesser degree depending on their own colors). It would require a strong environmental change over a long period of time or a series of concerted changes that push the species in one evolutionary direction to produce significantly different species (by concerted i don't mean design). Even then, they might still not be that different from their ancestors since regression to the mean will still mitigate the change, though the mean will slowly shift over many generations. 

 

Maybe i am just intellectualizing, but i find this topic important for the discussion about the IQ gains of Jews. Unless the Ashkenazi Jews engaged in some form of eugenics (i don't mean this in a derogatory way, just in the sense that people were really picky about their mating choices and the less intelligent were significantly under-reproducing), i find the natural history account for the gain suspect.

I have little to no input on how natural selection and IQ interplay, but wouldn't the survival of a minority whose thriving is based on advances in maths, sciences, and finance necessarily increase the focus on educational development? Is that what you mean by the "history account'? How heavily do genetics influence IQ(I've seen estimates from 40% to 80%)? Most importantly, to what extent is this field of science in its infancy? I am not asking rhetorically.

[labmath2]

I have little to no input on how natural selection and IQ interplay, but wouldn't the survival of a minority whose thriving is based on advances in maths, sciences, and finance necessarily increase the focus on educational development? Is that what you mean by the "history account'? How heavily do genetics influence IQ(I've seen estimates from 40% to 80%)? Most importantly, to what extent is this field of science in its infancy? I am not asking rhetorically.

 

There are two main theories on significant changes to a population attribute. The first has undertones of use and disuse, which is the idea that there was an emphasis on education (a cultural change). The second is that there was a strong selection for one attribute over the other. 

 

Their thriving being dependent on intellectual achievement can only explain the successes they had in intellectual fields at best with a minor increase in IQ. Since this is a case of simply changing what a person does, and not who they are. A similar version to this is weight lifting. Any muscular gain is only accrued over your lifetime, its not genetically transferred. Unless culture/activities can actually change the genes you pass on to your children. 

 

The second which is the stronger case is that there was significant change in the genetic makeup of the Ashkenazi Jews. There are multiple explanations for why this is the case, including less intelligent members leaving the group and higher reproduction rates of more intelligent members (rabbis). My question is about how much selection for intelligence would have to exist to go up one standard deviation in IQ over 800-900 years given the push back from regression to the mean. 

 

The latest interview Stefan did with an expert on IQ suggests that genetic expression of IQ changes over an individuals lifetime so that at early ages it is around 20% genetic and by 65 its around 80% genetic. There seems to be a great deal to research in this field, but i am not particularly well read in this and i don't know to what degree the push-back against IQ research has prevented significant developments.

[st434u]

At first when I read this response i wanted it to make sense, so i rationalized it in my head, but its not satisfactory. I already understood that many features of an organism is a result of multiple genes. The problem i am having is how you could get vastly different species given this regression. If most (if not all) significant variation is regressed in a few generations, then don't you simply get more variations from the average?

 

Like I explained, all regression to the mean does is make it harder for big changes that pile on top of each other to occur. It doesn't make it impossible. Like in the race horse example, you do get some offspring who are faster than both of the parents, and the parents themselves were the fastest of their generation. In this way and over several generations, you get increasingly faster horses. Just like you get increasingly sweeter fruit, increasingly fatter cows, etc.

 

In the wild, it takes more time than through breeding, because the selection pressure is not as strong, not as directed towards maximizing a certain characteristic. Another important factor is genetic isolation. which is usually but not always accompanied and created by physical isolation. Without genetic isolation, the selective pressure needs to be stronger and fairly ubiquitous. Without genetic isolation, local selection pressures get diluted when the genes mix with the general population that is not faced with such selection pressure. When Darwin travelled to the Galapagos Islands, he found that giant turtles on each of the islands had evolved a different type of neck and shell according to the local vegetation in the island. The locals could immediately identify which island a turtle came from, based simply on looking at it's shell. However, birds, who could fly around and therefore mate with the birds in other islands, did not evolve any differences from each other, they were all the same, because there was no significant genetic isolation, which the giant turtles did have.

 

Not all differences define a new species. To give a more extreme example, polar bears are quite different to brown bears, because they've been in vastly different environments and with significant genetic isolation over a large number of generations. However, they are still technically the same species (even though most biologists categorize them as a different one). You could mate them together and they would produce fertile offspring. The same is true of dogs and wolves. However, it's not always an easy task to breed them together and produce fertile offspring. Once the two sub-species have drifted apart enough, you won't be able to produce fertile offspring by mating them together, and that's when speciation will effectively have taken place.

 

The Ashkenazi jews are a sub-species of the sub-species of humans of european origin. They became a sub-species not by being in physical isolation from other europeans, but because their religious culture led them to be genetically isolated to a fair extent. That is, jews typically only married other jews, for thousands of years, and they did not engage in proselytism, as christians and most other religious cultures do. The question about why there was a selective pressure in them in particular for higher IQ is an open one. There are several hypothesis.

[rosencrantz]

Regression to the mean is simple when you look at it the way it is supposed, from the perspective of standard deviation. 

Example 1:
Height is 100% genetically determined. There is only one gene that determines height.

Father: 4 deviations above the mean.
Mother: 2 deviation avove the mean.


The children of this couple are likely to be 3 deviations above the mean, they regress to the mean because of their mother. They will be taller than the rest of the population, but not as tall as their father. 

Example 2: 
Height is 80% genetically determined, 20% are enviroment. There is only one gene that determines height.

Father: 4 deviations above the mean.
Mother: 2 deviations above the mean.

Because of the enviromental factor, the children will be taller than the mean with a range from 2,4 (hostile enviromental conditions) to 3,6 deviations above the mean (positive enviromental conditions).

Because of genetic mixup, deviations from the mean are not transferred directly from father to offspring, rather you have a mix of both, mother and father that determine together with the enviroment where you will be on the bell curve.
The regression to the mean comes into play, because 4 deviations away from the average is very rare, so it is likely that there will be a partner that isn't that far high. And that's regression to the mean.
 

[mishochu]

This topic is particularly interesting to me (I'd like to say for obvious reasons).

 

The discussion of both height and IQ leave me feeling like one big anecdote.

 

As an aside, height is not 100% controlled by genetics as malnutrition can "lock away" one's full potential for height.

 

My grandfather on my father's side was 7 feet tall, my father is 6'. I am 6'3". On my mother's side I don't think anyone ever got taller than 5'7". Did I return to some sort of mean? I guess the probabilities of genetics don't make it easy to determine if the mean is my father's height legacy over my mothers.

 

I do not know the IQ levels of my father or mother or either grandfathers. I am of Ebira descent (a tiny inconsequential tribe in the middle of Nigeria that is not part of either of the main three tribes).

 

1. My grandfather on my father's side was a school headmaster and taught tribal leaders English so they could negotiate with the British (I'd venture to say he was more intelligent than some around him)
2. My father started an investment firm (that eventually became a bank) whose initial success was due to imports and exports of agricultural goods. Since the bank was "nationalized" (and Nigeria focused on oil) agricultural output dropped.
3. My mother, while indoctrinated as a Muslim early on, went to school (her father received some bad will from his peers about this). She converted to Christianity as a teenager graduated from a university and started a toy store flying frequently to the far east to import toys (to this day Nigerians I meet in the US remember the jingle from the ads she placed on television).
4. My brother and I are software developers who each run our own companies instead of simply being employees.
5. My sister is a lawyer and CPA.

I only know my IQ (as I tested at 127). I also have an ACE score of 0 (which I am sure is unusual for Africans).

 

If I am careful about who I marry (and I think I was) and nurture my son and peacefully parent how likely is he to regress to a mean?

 

Is the IQ mean in my lineage different than the mean in the rest of Nigeria (or Africa as a whole)?

 

 

Will blacks that are smarter, given enough time, produce descendants that are around the average? Wouldn't that also mean diaspora Jews have always been smarter, otherwise they would have regressed to the mean?

 

I would love to know the answer to this question (let's say I've got "skin" in this game  ).

 

How important is the regression to the mean idea? I think the statistics on race and IQ are important mostly to convince the collectivists to stop their meddling and let natural incentives for higher IQ take precedence (instead of subsidizing lower IQ).

 

I'm also a firm believer in assimilation (if you're going to move to western civilization it should be because you LIKE western civilization).

[st434u]

The regression to the mean comes into play, because 4 deviations away from the average is very rare, so it is likely that there will be a partner that isn't that far high. And that's regression to the mean.

 

 

No, that's not what regression to the mean is.

[st434u]

If I am careful about who I marry (and I think I was) and nurture my son and peacefully parent how likely is he to regress to a mean?

If your wife is of ebira descent, and has a similar high IQ as you, then your children will likely regress to the ebira mean. That doesn't imply that they will have the same IQ as the average for ebira people, only that their IQ will tend to move closer to the average for ebira people, which is lower than yours and your wives's. But they will still tend to be higher than the average for ebira people, just not as high as yours or your wives. Then again, some of your children will be as high as you and your wife, and some might be even higher.

 

Is the IQ mean in my lineage different than the mean in the rest of Nigeria (or Africa as a whole)?

I don't know about the ebira people in particular, but yes as far as Nigeria goes. If you want to maximize the chances that your children will have an IQ as high as you, you need to look not only at the IQ of your wife, but at the IQs of her family members as well. Then again, IQ in and of itself is not that important. It only measures a tiny part of what consitutes intelligence as a whole. So really what you want to be looking for is a smart wife with a smart family lineage. All of that said, intelligence isn't everything. I would much rather be with someone who is good and caring but not amazingly bright than someone who is bad and cold and super smart.

[rosencrantz]

Regression to the mean in statistics is different from regression to the mean in genetics. In statistics, you have random not connected events. In genetics the events are random, but connected. Hence, you have to use a different methodology. 

 

Now this is where we see the regression to the mean effect: the expected value of the child is 45 (0.25 * 20 + 0.25 * 50 + 0.25 * 40 + 0.25 * 70), which happens to be the parents' average (0.5 * 30 + 0.5 * 60). But this is only an expectation, and diversity is maintained with the child's value possibly ranging from 20 to 70.

 

http://statisticator.blogspot.de/2012/12/genetics-and-regression-toward-mean.html

You can use both absolute numbers, or deviations from the mean. Deviations from the mean are simpler because they'll tell you instantly about the likelyhood of combined elements.

[st434u]

I'm also a firm believer in assimilation (if you're going to move to western civilization it should be because you LIKE western civilization).

 

I would also add something that has become highly politically incorrect to mention, and that's that generally speaking, you have better chances of having healthy children if you marry someone who's the same race as yours, or somewhat similar. So for example, it's unlikely that marrying a nigerian who's not of ebira descent would make much difference versus marrying an ebira woman, similarly you could marry an african woman who's not nigerian and expect similar results; if on the other hand you're looking to marry, say, an european, or an asian woman, then while the regression to the mean as far as IQ goes may have less effect, you also have to consider that your genes and theirs haven't been mixed frequently for thousands of generations, and so they're not used to working together as well, in order to produce a healthy organism.

 

Genes co-evolved with other genes in the relevant gene pool. That's why, while you can get two sub-species to successfully mate together and produce fertile offspring, the more the two sub-species have drifted apart genetically, the more likely you are to run into problems. So to give an example, from African, to Arab, to European, to Indian, to Asian, to American Indian, you are likely to have increasing problems of genetic compatibility, as far as producing healthy offspring is concerned, because the genes of your ancestors have been mostly apart from those of theirs in increasing order of generations.

[Matthew Ed Moran]

I thought regression to the mean would exist from an evolutionary stand point because any trait has diminishing marginal utility towards reproduction.

[NoStringsOnMe]

I would also add something that has become highly politically incorrect to mention, and that's that generally speaking, you have better chances of having healthy children if you marry someone who's the same race as yours, or somewhat similar. So for example, it's unlikely that marrying a nigerian who's not of ebira descent would make much difference versus marrying an ebira woman, similarly you could marry an african woman who's not nigerian and expect similar results; if on the other hand you're looking to marry, say, an european, or an asian woman, then while the regression to the mean as far as IQ goes may have less effect, you also have to consider that your genes and theirs haven't been mixed frequently for thousands of generations, and so they're not used to working together as well, in order to produce a healthy organism.

 

Genes co-evolved with other genes in the relevant gene pool. That's why, while you can get two sub-species to successfully mate together and produce fertile offspring, the more the two sub-species have drifted apart genetically, the more likely you are to run into problems. So to give an example, from African, to Arab, to European, to Indian, to Asian, to American Indian, you are likely to have increasing problems of genetic compatibility, as far as producing healthy offspring is concerned, because the genes of your ancestors have been mostly apart from those of theirs in increasing order of generations.

 

Nah... I don't buy this. I'll have to see a scientific paper on this. What's your argument? Polar Bear/Brown Bear mix? Ligers? Statistically insignificant to use as any quantitative metric. You have to discuss a population, not individuals. And environmental factors can serve as a boon just as often as it can a bane.

[Mister Mister]

At first when I read this response i wanted it to make sense, so i rationalized it in my head, but its not satisfactory. I already understood that many features of an organism is a result of multiple genes. The problem i am having is how you could get vastly different species given this regression. If most (if not all) significant variation is regressed in a few generations, then don't you simply get more variations from the average? By that i mean all possible variation of a species will be present in the group, with nature affecting the ratio. Take color for example, if we have a range of color for a species from red to yellow. It would seem you can only get more variations of color in the group (maybe darker red and lighter yellow becomes a possibility over time, and in rare cases green gets added to the possibility) and the ratio of the population color changes depending on environment. Lets assume the yellow members are more vulnerable to predation, then the ratio of yellow to red will be lower. Given the fact that their color is a combination of genes, it will be almost impossible to get rid of any color variation since all members have the potential to give birth to any color (to greater or lesser degree depending on their own colors). It would require a strong environmental change over a long period of time or a series of concerted changes that push the species in one evolutionary direction to produce significantly different species (by concerted i don't mean design). Even then, they might still not be that different from their ancestors since regression to the mean will still mitigate the change, though the mean will slowly shift over many generations. 

 

Maybe i am just intellectualizing, but i find this topic important for the discussion about the IQ gains of Jews. Unless the Ashkenazi Jews engaged in some form of eugenics (i don't mean this in a derogatory way, just in the sense that people were really picky about their mating choices and the less intelligent were significantly under-reproducing), i find the natural history account for the gain suspect.

If a really smart or really tall person is born, like IQ 160 or 7 feet tall, and he has lots of children, that raises the collective IQ or height of the genepool, despite regression to the mean.  Does that make sense?

[labmath2]

If a really smart or really tall person is born, like IQ 160 or 7 feet tall, and he has lots of children, that raises the collective IQ or height of the genepool, despite regression to the mean.  Does that make sense?

 

That makes sense. The question is by how much does it change the mean and distribution?