Wikipedia:Reference desk/Archives/Science/2021 February 17

= February 17 =

IQ and narrow-sense heritability question
If the average IQ of a couple is 110 and the group that this couple come from has an average IQ of 80, and IQ has a narrow-sense heritability of 0.6, are this couple's children going to have an average IQ of 92 or 98? I'm obviously taking regression towards the mean into account here. Futurist110 (talk) 07:01, 17 February 2021 (UTC)


 * How is "the group that this couple come from has an average IQ of 80" relevant to the IQ of their children?--Shantavira|feed me 09:26, 17 February 2021 (UTC)
 * See Heritability of IQ. It has a well studied and well established genetic component. Therefore genetic background is relevant. Fgf10 (talk) 12:06, 17 February 2021 (UTC)
 * First you need to show that IQ represents anything more than the ability to do well on an IQ test. --Khajidha (talk) 11:26, 17 February 2021 (UTC)
 * Why? The question is about IQ. How representative IQ performance is of intelligence and how you define intelligence are different (important) questions. Fgf10 (talk) 12:06, 17 February 2021 (UTC)
 * The only thing an IQ test measures is someone's ability to do well on an IQ test. ←Baseball Bugs What's up, Doc? carrots→ 12:17, 17 February 2021 (UTC)
 * https://twitter.com/stevestuwill/status/1076000663824461824?lang=en Futurist110 (talk) 23:10, 17 February 2021 (UTC)
 * I wouldn't say only. It also measures things like your stress level, your level of childhood nutrition, how much sleep you had the night before you took the test, whether or not you've experienced trauma, your socioeconomic status, and a whole slew of other things that affect the performance on the test (which are ultimately, of course, unrelated to the very thing that the IQ test is supposed to tell you about a person).  -- Jayron 32 13:39, 17 February 2021 (UTC)
 * Do IQ tests nowadays ask questions about those kinds of things? ←Baseball Bugs What's up, Doc? carrots→ 15:41, 17 February 2021 (UTC)
 * No, but all of those things can affect how you answer the IQ test questions on a given day, and have nothing to do with a person's inherent intellectual ability. -- Jayron 32 16:53, 17 February 2021 (UTC)
 * Yes, things like that will affect your score, as they will with most psychometric tests. But how important are they quantitatively in practice? Regarding "how much sleep you got", for instance, we can easily test people on two or more occasions and get data about this rather than opinion. According to our article the standard deviations are as low as 3 points (cf. population mean = 100, population standard deviation = 15). It is not so easy to measure the effect of things like socioeconomic status, but far from impossible to make an estimate based on statistical analysis. So it seems to me a bit silly to denigrate IQ just because the scores are influenced by factors that you might want to factor out (others might not want to factor them out; it depends on their questions). It is quite another issue how much IQ is useful for estimating "inherent intellectual ability", but it was you that brought that up, not the questioner. One can have different opinions about that issue yet still find IQ a valid psychometric measure. Jmchutchinson (talk) 17:51, 17 February 2021 (UTC)
 * I think the main disagreement is over the third to last word in your statement. -- Jayron 32 18:46, 17 February 2021 (UTC)
 * (ec)Sure, they can affect one's performance on the test, but they aren't measured by the test. In fact, some folks perform better under stress. ←Baseball Bugs What's up, Doc? carrots→ 17:54, 17 February 2021 (UTC)


 * You might get better answers if you ask the same question about height (for instance), where environmental influences are less important and the research is less controversial. Tigraan Click here to contact me 14:00, 17 February 2021 (UTC)
 * Environmental influences are not clearly less for height. Heritability of IQ in adults is about 0.8, very similar to that of height. You are correct though that IQ research often triggers a distracting knee-jerk response, as several responses have already demonstrated. One might justifiably object to some of the ways IQ has been interpreted, but it is a well defined and reproducible measure, so there is no reason why one cannot measure its heritability within a population. Ironically, that research actually provides some data about claims of environmental influences! Jmchutchinson (talk) 15:11, 17 February 2021 (UTC)
 * Since you were not triggered by my question here, can you please try answering it? Futurist110 (talk) 20:43, 17 February 2021 (UTC)


 * There's a melange of issues here that everyone seems to confuse though. The issue is whether a single score given by a single test is a reliable measure of something useful, which is a distinct and entirely unrelated question as to whether intellect (or more properly intellectual potential, since many factors can affect your intellect at any given moment) is a heritable trait.  The connection between the former and the latter is tenuous at best.  -- Jayron 32 16:56, 17 February 2021 (UTC)


 * A big thing I think is being missed here, as well, is that heritability of a trait does not explicitly mean that the trait is genetic. For example, a test for heritability of monetary wealth will show that it is a heritable trait, since comparing the Walton family through 3 generations to almost any other family will show that wealth is heritable. Spoken language is another trait that tests positive as a heritable trait; you are far more likely to speak the language that your parents spoke than any other, which means that you likely speak the same language as at least one set of grandparents, one set of great grandparents, etc. Probably more than that, since people are more likely to have relationships with and children with people who also speak the same language. Is the fact that I speak English and not Romansh a genetic trait? Of course not. It is, however, a heritable trait. This also applies in many other areas that we tend to think of as "genetic," rightly or wrongly. Certainly genetics has a role in height, for example, but if you are born into a family in poverty, there's a good chance that your grew up malnourished and with stunted growth. There's also a good chance that you spend your whole life in poverty, since lack social mobility and wealth inequality are big problems globally. So, chances are that your parents were malnourished and had stunted growth, and that your children will be malnourished and have a stunted growth. Even though you may have genes that suggest you will be tall, the heritable trait of poverty has meant you have heritable stunted growth. Barring epigenetic effects, if one of your descendent got out of poverty, they would no longer have stunted growth and would grow tall, but that doesn't change that stunted growth was a heritable trait for several generations. I'm bringing this up because it is a very common issue with the IQ test. People assume that heritability of IQ within a family explicitly points to IQ being genetic, but it may not. Sometimes, its that tests are written in a way that uses colloquial language that some in a population understand, and others do not. Sometimes it points to things like what type of problem solving was emphasized during primary schooling, or even access to primary schooling. Sometimes it merely represents literacy rate. There are so many factors that go into how well one performs on these tests that, to assign it a genetic component is not at all straightforward. Heritable, sure, but not necessarily genetic. That's not even getting into the issue on how well IQ tests actually measure intelligence at all (spoiler: they probably don't do that very well, as pointed out. --OuroborosCobra (talk) 21:32, 17 February 2021 (UTC)
 * I think this last contribution is fundamentally wrong in a technical sense. You are right that we usually inherit far more than just genes from our parents, so that these epigenetic effects may superficially mimic genetic inheritance. But heritability is specifically defined to measure only the genetic component of inheritance. That makes heritability more of a challenge to measure in humans, but there are some tricks involving adoptions, or comparing similarities between twins and siblings, etc. There is some information on methods in our article. Geneticists are very, very aware of the confounding of genetics and environment! How accurately they have managed to isolate the genetic component in the specific case of IQ, I don't really know, but you can't so simply explain away high heritabilities of IQ as due to epigenetic inheritance of home environment or socioeconomic status. Jmchutchinson (talk) 07:39, 18 February 2021 (UTC)
 * I think you're missing the point, or rather my point (can't speak to others), which is that IQ is basically modern-day phrenology. Proving that a person can inherit the pattern of bumps on their head from their parents doesn't make phrenology valid.  That IQ is heritable doesn't mean that IQ is useful or valid for anything.  -- Jayron 32 17:48, 18 February 2021 (UTC)
 * Right. Even if we want to use the strict definition that "heritable means inherited to a statistically significantly through genes," then I guess the issue is proving that IQ is heritable at all, which isn't a given or an assumption that can be made. Testing for heritable traits is not a simple task, and the more complicated/less understood the trait, the more difficult it is to test. IQ and the idea of intelligence in general is not a trivial trait to describe in the way that eye color or even height can be and, as I showed in my earlier comment, shortness can be mistaken as a genetically inherited trait when it is actually inherited malnutrition due to poverty within successive generations. Yes, a twin study could figure that out, but that means conducting an experiment (or being lucky enough to find identical twins separated at birth) where one is raised in poverty and the other is raised in abundance. This is not easily done on a complicated trait like IQ that, let's be honest, doesn't even have a great definition of itself. Building off your phrenology example artificial cranial reshaping was a not uncommon practice among certain mesoamerican communities, and since children with elongated skulls tended to have parents with the same, it was easy to assume some genetic link and possibly even link that to other traits that may not even exist (I think Stargate SG-1 had these connected with psyonic powers, or maybe that was Indiana Jones). Just as with phrenology, there was no genetic link to these shapes, and certainly no link to anything else, like psyonic powers. --OuroborosCobra (talk) 20:07, 18 February 2021 (UTC)
 * Jayron is interested in one aspect of validity, whether IQ measures anything useful and in particular something people want to call "general intelligence". My impression is that nowadays psychologists tend to deal with a multidimensional suite of measurable and specific intellectual abilities, not general intelligence. Their working assumption would be that IQ score reflects just some of these abilities. People obviously do differ in their various intellectual abilities and I don't know how you would intend to start investigating this phenomenon without obtaining data from performance tests, of which the IQ test (or refinements of it) is one example. A test of memory would be an analagous one that perhaps you find more palatable. What IQ does relate to is an empirical question; there are some pointers to the results in part of our IQ article. Your apparent opinion that it relates to nothing useful seems to me not to be based on empirical data.
 * OuroborosCobra, I think you are seriously underestimating the sophistication of geneticists in measuring the heritability of IQ, and other human traits. Twins separated at birth are just one source of data, which indeed are now a sparse resource. Another is the difference in similarities between monozygotic and dizygotic twins, or between full sibs and half sibs; these would normally share most aspects of their environment. Really, "proving that IQ is heritable at all" is not the issue. Everybody is surprised with how heritable IQ has been proved to be in multiple rigorously performed studies. Most of us find it uncomfortable to learn how little normal environmental variation affects adult IQ, which does not mean that extreme environments cannot affect it considerably. Again, I recommend looking at the empirical data, not relying on your intuition. Jmchutchinson (talk) 22:46, 18 February 2021 (UTC)
 * I think it is utterly impossible to suggest geneticists could measure the heritability of IQ if we don't even have a good idea of whether IQ is a measure of anything useful. If it isn't even a measure of anything beyond "I was capable of taking this test," then we have a problem if geneticists are saying that they've found a genetic link to the ability to pass this one specific test. --OuroborosCobra (talk) 22:57, 18 February 2021 (UTC)


 * So 98 comes from
 * $$ R = h^2 S $$

How did you get 92? Rmhermen (talk) 00:01, 19 February 2021 (UTC)

Why don't we develop two visions if we have two eyes?
I always wonder Why don't we develop two visions if we have two eyes? From Wikipedia, I can only tell eyes contains rod cells and cone cells. Rizosome (talk) 15:05, 17 February 2021 (UTC)


 * Some people have two eyes and see two images. In non-pathological cases, the brain (more precisely the thalamus) does some post-processing and blends both inputs. Binocular vision has more on the process and it's advantages. Bumptump (talk) 16:01, 17 February 2021 (UTC)


 * It depends on how well your brain works. Each eye does produce a different image.  You can check this yourself where you alternate closing one eye and then the other, and you can confirm that you do receive different images from each eye.  A classic test of this is to use parallax.  Hold a finger up in front of your face and close one eye.  Align the finger so it blocks a distant object.  Swap your eyes, so you close the open one and open the closed one.  The formerly obscured object will be visible, and it will appear that your finger has "jumped" to the other side.  The reason why you don't see two images when both eyes are open is that the Visual processing in your brain automatically stitches the two images together into a single image.  As Bumptump notes above, there are people whose brains don't do this correctly, and they have a condition known as diplopia or "double vision".  -- Jayron 32 16:51, 17 February 2021 (UTC)
 * Go to blind spot (vision) for another fun way to see how your brain processes what it gets from the eyes and fills things in without you being conscious of it. --47.152.93.24 (talk) 04:15, 18 February 2021 (UTC)
 * Something like that happens with a helmet-mounted display where pilots learn to see one system (e.g. avionics) through one eye, and another system (e.g. weapons) through the other. Might take some serious re-training of your brain, but appears to be doable.  85.76.75.69 (talk) 19:23, 17 February 2021 (UTC)
 * As a personal anecdote, back when I worked in VR, our pre-oculus headsets actually took two separate VGA inputs, for testing I would sometimes exploit that to display the simulation to one eye with the my debugger going to the other eye. At first I would close one eye, but I got to the point where I could look at both at once.  ApLundell (talk) 21:25, 18 February 2021 (UTC)


 * Generally speaking, predators have binocular vision while prey tends to have mostly monocular vision. Binocular vision helps the predator focus more accurately on their prey. Prey needs to see 360 degrees around to watch for predators. ←Baseball Bugs What's up, Doc? carrots→ 20:15, 17 February 2021 (UTC)


 * As an actual data point, after some face surgery my right eyeball was moved slightly, vertically. It took several days before my vision re-synched. Greglocock (talk) 20:34, 17 February 2021 (UTC)


 * See Ocular dominance. Most people's brains rely mainly on the image from only one eye. Alansplodge (talk) 13:00, 18 February 2021 (UTC)