2050 words
There are a lot of conceptual problems with IQ tests that I never see talked about. The main ones are how the tests are constructed (to fit a normal curve, no less); to the fact that there is no construct validity to the tests (IQ tests aren’t calibrated against a biological model like breathalyzers are calibrated against a model of blood in the blood stream); and how the Raven’s Progressive Matrices test is actually biased despite being touted as the most culture-free test since all you’re doing is rotating abstract symbols to see what comes next in the sequence. These three assumptions have important implications for the ‘power’ of the IQ tests, the most important being the test construction and validity.
I) IQ test construction
IQ tests are constructed with the assumption that we know what IQ tests test (we don’t) and with the prior ‘knowledge’ of who is or is not intelligent. Test constructors construct the tests to reveal presumed differences between individuals.
It is assumed that 1) IQ scores lie on a normal distribution (they don’t) and 2) few natural bio functions conform to this curve. Another problem with IQ test construction is the assumption that it increases with age and levels off after puberty. Though this, like the other things, has been built into the test by choosing items that an increasing proportion of children pass. You can, of course, reverse this effect by choosing items that older people do well on and younger people don’t.
Further, they keep 50 percent of items that children get right while keeping a smaller proportion of items that children get right, which, in effect, presupposes who is or is not intelligent.
Though, you never see those who believe that IQ is a ‘good enough’ proxy for intelligence ever being this up. Why? This is very important for the validity of these tests. Because if how the tests are constructed is wrong and test scores are not to fit a normal distribution when no normal distribution actually exists for most human mental (including IQ scores) and physiological traits, then the assumptions and conclusions drawn from them are wrong. IQ tests are constructed with the prior idea of who is or is not ‘intelligent’ and this is done by how the items are chosen—50 percent of the items that people get right are kept while the smaller proportion of items people get right or wrong are kept. This is how this so-called ‘normal curve’ appears in IQ tests and is why the book The Bell Curve has the name it has. But bell curve don’t exist for a modicum of traits including IQ!!
Simon (1997: 204) writes (emphasis mine):
There is another, and completely irrefutable, reason why the bell-shaped curve proves nothing at all in the context of H-M’s book: The makers of IQ tests consciously force the test into such a form that it produces this curve, for ease of statistical analysis. The first versions of such tests invariably produce odd-shaped distributions. The test-makers then subtract and add questions to find those that discriminate well between more-successful and less-successful test-takers. For this reason alone the bell-shaped IQ curve must be considered an artifact rather than a fact, and therefore tells us nothing about human nature or human society.
The analysis and selection of items that go on the tests are biased since there is no cognitive theory on which the analysis and selection of items are based. Carpenter, Just and Shell (1990: 408) note how John Raven, the creator of the Raven’s Progressive Matrices, even discussed this in his personal notes, writing “He used his intuition and clinical experience to rank order the difficulty of the six problem types. Many years later, normative data from Forbes (1964), shown in Figure 3, became the basis for selecting problems for retention in newer versions of the test and for arranging the problems in order of increasing difficulty, without regard to any underlying processing theory.”
II) IQ test validity
Another problem with IQ tests are its validity. People attempt to ‘prove’ its validity with correlating job performance success with IQ scores, though there are huge flaws in the studies purporting to show a .5 correlation between IQ and job performance (Richardson, 2002; Richardson and Norgate, 2015). IQ tests are not like, say, breathalyzers (which are calibrated against a model of blood alcohol) or white blood cell count (which is a proxy for disease in the body). Those two measures have a solid theoretical basis and underpinning; as blood alcohol rises, the individual had increased alcohol consumption. The same is true for white blood cell count. The same is not true for IQ tests.
One of the biggest measures used in regards to job performance and IQ testing (people attempt to use job performance to attempt to validate IQ tests) is supervisor rating. However, supervisory ratings are hugely subjective and a lot of factors that would have a supervisor be said to be a ‘good worker’ are not variables that entail just that job.
The only ‘validity’ that IQ test have is correlations with other IQ tests and tests like the SAT. This is not validity. Say the breathalyzer wasn’t calibrated against a model of blood alcohol in the body, would breathalyzers still be a valid tool to test people’s blood/alcohol level? On that same note let’s say that white blood cells wasn’t construct valid. Would we be able to reliably use white blood cell count as a valid measure for disease in the body? These very same problems plague IQ tests and people accept them as ‘proxies’ for intelligence, they test ‘enough of intelligence’ to be able to say that one is smarter than another because they scored higher in a test and therefore tap into this mystical ‘g’ that they have more of which is like a ‘power’ or ‘energy’.
These tests, therefore, are constructed with the idea of who is or is not intelligent and you can see that by looking at how the items are chosen for the test. That’s not scientific. So a true test of ‘intelligence’ may not even exist since these tests have this type of construct bias already in them.
IQ tests have no validity like breathalyzers and white blood cell count, and the so-called ‘culture-free’ IQ test Raven’s Progressive Matrices is anything but.
III) Raven’s and culture bias
I specifically asked Dr. James Thompson about Raven’s being culture-fair. I said that I recall Linda Gottfredson saying that people say that Ravens is culture-fair only because Jensen said it:
Yes, Gottfredson made that remark, and I remember her doing it at an ISIR conference.
So that’s one thing about Ravens that crumbles. A quote from Ken Richardson’s book Genes, Brains, and Human Potential: The Science and Ideology of Intelligence:
It is well known that families and subcultures vary in their exposure to, and usage of, the tools of literacy, numeracy, and associated ways of thinking. Children will vary in these because of accidents of background. …that background experience with specific cultural tools like literacy and numeracy is reflected in changes in brain networks. This explains the importance of social class context to cognitive demands, but is says nothing about individual potential.
(This argument on social class is much more complex than ‘poor people are genetically predisposed to be dumb and poor’.
Consider a recent GCTA study by Plomin et al., who reported a SNP-based heritability estimate of 35% for “general cognitive ability” among UK 12 year olds (as compared to a twin heritability estimate of 46%) [8]. According to the Wellcome Trust “genetic map of Britain,” striking patterns of genetic clustering (i.e. population stratification) exist within different geographic regions of the UK, including distinct genetic clusterings comprised of the residents of the South, South-East and Midlands of England; Cumbria, Northumberland and the Scottish borders; Lancashire and Yorkshire; Cornwall; Devon; South Wales; the Welsh borders; Anglesey in North Wales; Scotland and Ireland; and the Orkney Islands [8]. Now consider the title of a study from the University and College Union: “Location, Location, Location – the widening education gap in Britain and how where you live determines your chances” [9]. This state of affairs (not at all unique to the UK), combined with widespread geographic population stratification, is fertile ground for spurious heritability estimates.
Still Chasing Ghosts: A New Genetic Methodology Will Not Find the “Missing Heritability”
I think this argument is interesting, and it throws a wrench into a lot of things, but more on that another day.)
Richardson continues:
In other words, items like those in the Raven contain hidden structure which makes them more, not less, culturally steeped than any other kind of intelligence testing items, like the Raven, as somehow not knowledge-based, when all are clearly learning dependent. Ironically, such cultural-dependency testing is sometimes tacitly admitted by test users. For example, when testing children in Kuwait on the Raven in 2006, Ahmed Abdel-Khalek and John Raven transposed the items “to read from left to right following the custom of Arabic writings.” (Richardson, 2017: 99)
Finally, we have this dissertation which shows that urban peoples score better than hunter-gatherers (relevant to this present article):
Reading was the greatest predictor of performance Raven’s, despite controlling for age and sex. Attendance was also strongly correlated with Raven’s performance. These findings suggest that reading, or pattern recognition, could be fundamentally affecting the way an individual problem solves or learns to learn, and is somehow tapping into ‘g’. Presumably the only way to learn to read is through schooling. It is, therefore, essential that children are exposed to formal education, have the motivation to go/stay in school, and are exposed to consistent, quality training in order to develop the skills associated with improved performance. (pg. 83)
This is telling: This means that there is no such thing as a ‘culture-free’ IQ test and there will always be something involved that makes it culture un-fair.
People may say ‘It’s only rotating pictures and shapes to get the final answer, how much schooling could you need??’, well as seen above with the Tsimane, schooling is very important to IQ tests since they test learned skills. I’ve seen some people claim that IQ tests don’t test learned ability and that it’s all native, unlearned ability. That’s a very incorrect statement.
So although the symbols in a test like the RPM are experience-free, the rules governing their changes across the matrix are certainly not, and they are more likely to be already represented in the minds of children from middle-class homes, less so in others. Performance on the Raven’s test, in other words, is a question not of inducing ‘rules’ from meaningless symbols, in a totally abstract fashion, but of recruiting ones that are already rooted in the activites of some cultures rather than others. Like so many problems in life, including fields as diverse as chess, science and mathematics (e.g. Chi & Glaser, 1985), each item on the Raven’s test is a recognition problem (matching the covariation structure in a stimulus array to ones in background knowledge) before it is a reasoning problem. The latter is rendered easy when the former has been achieved. Similar arguments can be made about other so-called ‘culture-free’ items like analogies and classifications (Richardson & Webster, 1996). (Richardson, 2002: pg 292-292)
Everyday life is also more complex than the hardest items on Raven’s Matrices, while the test is not complex in its demands compared to tasks undertaken in everyday life (Carpenter, Just, and Shell, 1990). They conclude that the cause is differences in working memory, but that is an ill-defined concept in psychology. They do say, though, that “The processes that distinguish among individuals are primarily the ability to induce abstract relations and the ability to dynamically manage a large set of problem-solving goals in working memory.” So item complexity doesn’t make Raven’s items more difficult for others, since everyday life is more complex.
I’ll end with a bit of physiology. What physiological process is does IQ mimic in the body? If it is a physiological process, surely you’re aware that physiological processes *are not* static. IQ is said to be stable at adulthood, what a strange physiological process. Let’s say for arguments’ sake that IQ really does test some intrinsic, biological process. Does it seem weird to you that a supposed real, stable, biological, bodily function of an individual would be different at different times?
Conclusion
There are a lot of assumptions about IQ tests that are never talked about. The most important being how the tests are constructed to fit a normal curve when most traits important for survival aren’t normally distributed. IQ tests are constructed with the assumption of who is or isn’t intelligent just on the knowledge of how the items are prepared for the test. When you look at how the tests are constructed you can see how they are constructed to fit the normal curve because most of their assumptions and conclusions rest on the reality of the normal curve. There is no construct validity to IQ tests, they’re not like breathalyzers for instance which are calibrated against a model of blood alcohol or white blood cell count as a proxy for disease in the body. Raven’s—despite what is commonly stated about the test—is not unbiased, it perhaps is the most biased IQ test of them all. This highlights the problems with IQ tests that are rarely ever spoken about, and should have you call into question the ‘power’ of the IQ test which assumes who is or isn’t intelligent ahead of time.
NotPoliticallyCorrect 
Realist, what are your ideas for improving this situation?
LikeLike
” IQ scores lie on a normal distribution (they don’t) ”
Yes they do, saying other Psychological traits aren’t normally distributed isn’t evidence that IQ follows that same pattern.
“Though this, like the other things, has been built into the test by choosing items that an increasing proportion of children pass. You can, of course, reverse this effect by choosing items that older people do well on and younger people don’t.”
And? What? You think Children are more intelligent than adults?
“The same is not true for IQ tests.”
It is, but you haven’t done enough research to know otherwise so……
LikeLike
They’re forced to fit a normal curve due to how the tests are constructed.
No. The point is the outcome is based on item choice.
So what type of construct validity do IQ tests have? Is it like a breathalyzer that’s calibrated against a model of blood in the blood stream? Or like how white blood cell count is a marker for disease in the body? There is no construct validity. The only ‘validity’ they have is correlations with other IQ tests and standardized tests. That’s not construct validity.
LikeLike
“They’re forced to fit a normal curve due to how the tests are constructed.”
But you haven’t provided evidence that they would have a non normal distribution.
“The point is the outcome is based on item choice.”
That is how the normal distribution is made, but that doesn’t make it irrelevant.
“So what type of construct validity do IQ tests have? ”
Brain size, Synaptic plasticity, Processing speed.
LikeLike
“But you haven’t provided evidence that they would have a non normal distribution.”
You provided no evidence that the bell curve from IQ test is accurate.
“That is how the normal distribution is made, but that doesn’t make it irrelevant.”
It does because it assume a lot of wrong things:
And see the article about testosterone/find the genes.
“Brain size, Synaptic plasticity, Processing speed.”
Brain size do not tell much about IQ, see the article about microcephaline and IQ. His 0.2-0.4 correlation with IQ is meaningless since it is tied to body size
Processing speed? Really? How is it scientific?
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2871325/
Fucking video games increase your processing speed.
LikeLike
“You provided no evidence that the bell curve from IQ test is accurate.”
I don’t have to. He made the assertion first, either way this will quell your bitching.
“It does because it assume a lot of wrong things:”
Like what? Most Psychometricians are aware of the built in Gaussian curve all it does is keep the test identical each year, it doesn’t manipulate results.
“Brain size do not tell much about IQ, see the article about microcephaline and IQ.”
I already debunked those claims. Microcephaly is not one disease and brain size is a direct effect of more neurons and larger cortical columns. Size is directly related to capacity and neuroplasticty.
Claiming IQ is not a meaningful measure of intelligence and the using that same proxy to conclude brain size ha no affect on intelligence is beyond stupid.
“Processing speed? Really? How is it scientific?”
It’s scientific in the sense that is an established biological and phsyiological construct, it affected by:
“the time for electrochemical signals to traverse axons and dendrites, axonal myelination, the diffusion time of neurotransmitters across the synaptic cleft, differences in synaptic efficacy, the coherence of neural firing, the current availability of neurotransmitters, and the prior history of neuronal firing. ”
“Fucking video games increase your processing speed.”
Your point LOL? So increased proficiency means what exactly? So suddenly Processing speed isn’t a real thing because you can get better at it?
LikeLike
I provided evidence (via test construction) that they’re forced to fit a normal curve. How the curve would look without the forced curve I don’t know, all i do know is that assumptions and conclusions drawn from the forced normal curve are seriously flawed.
Why? If you’re forcing the normal distribution due to choosing which items to or to not include in the test, that means that you’re attempting to force the outcome based on who is or is not intelligent. It makes assumptions on who is or is not intelligent and the test is constructed around those assumptions.
…That’s not construct validity! To give an example, blood alcohol rises as one drinks more alcohol. When they are then given a breathalyzer, they then get a specific value and that value determines one’s state. It’s calibrated against a biological model; how are IQ tests calibrated? Brain size is not ‘construct validity’, nor is ‘synaptic plasticity’ nor is ‘processing speed’. Processing speed, as John has cited, can be increased through playing video games so that’s a confound that is never tested for during RT studies. I’ve already cited evidence that RT is a flimsy variable and that due to the low correlations of .2 to .3 that shows that way more goes into RT than ‘processing speed’.
But let’s assume for the sake of the argument that those three things make IQ tests ‘construct valid’: As brain size increases linearly, does IQ? Synaptic plasticity is the ability for synapses to strengthen or weaken over time, if they get stronger does IQ increase? If someone plays a lot of video games and his processing speed increases because of that, is he more intelligent? Say that processing speed is measured in hertz. Is the person who uses 10 hertz in processing speed always more ‘intelligent’ or score higher on an IQ test than someone with 5 hertz? These are questions you have to answer and the answer ‘it’s not a perfect measure’ is wrong because if you’re trying to prove that IQ tests and its ‘correlates’ have the same construct validity as the breathalyzer or white blood cell count then you’re going to have to give a satisfactory answer to those questions other than ‘no measure is perfect all of the time’, because if that were the case for breathalyzers then we wouldn’t use them to catch drunk drivers.
A whole presentation on the false normal IQ distribution? Yawn.
So tossing out, say, 10 items because they don’t conform to what you’re attempting to do doesn’t manipulate results? Say the ‘hard’ items 50 percent get right. They’d then change the ‘hard items’, which shows that they’re deliberately constructing the tests to get the result they want.
I’m aware the microcephaly isn’t one disease. That doesn’t change the evolutionary claims from Skoyles and Sagan. Oh so you agree with me on cortical columns? Now do you agree on the expertise and cortical columns resulting in larger brains?
But differences in ‘processing speed’ don’t mean that person A is more intelligent than person B since there are numerous confounds to ‘processing speed’ like playing video games.
It’s a confound that’s not controlled for, that’s the point. Remember the problems with RT:
Relationship between Intelligence and Reaction Time; A Review Study
LikeLike
“If you’re forcing the normal distribution due to choosing which items to or to not include in the test, that means that you’re attempting to force the outcome based on who is or is not intelligent.”
If we had 100 people that we each gave tests to, whichever items had the majority of correct answer would be considered the average. From there the test is constructed as a normal distribution because If you use any natural mathematical metric for difficulty of questions, like “size of search space” in a chess problem, or “number of steps of deduction” in a mathematical problem, basically anything, in the natural metric, the distribution of humans would be a power-law like distribution with a heavy tail and different individuals would perform astronomically better at some tasks than others.
“Brain size is not ‘construct validity’, ”
Yes it is, the more neurons the larger the capacity.
“nor is ‘synaptic plasticity’”
Yes it is Synaptic plasticty is how you learn anything at all.
“Processing speed, as John has cited, can be increased through playing video games so that’s a confound that is never tested for during RT studies.”
A confound to what? The fact that Processing speed is physiological is not dependent on whether the extent of it’s effects are due to nature or nurture.
“I’ve already cited evidence that RT is a flimsy variable and that due to the low correlations of .2 to .3 that shows that way more goes into RT than ‘processing speed’.””
Processing speed is simply that, speed. It wont affect how deep a person’s intelligence is only how fast they are using it.
“A whole presentation on the false normal IQ distribution? Yawn.”
You have yet to demonstrate that it’s false. Try again.
“Now do you agree on the expertise and cortical columns resulting in larger brains?”
No. Expertise capacity and IQ are indistinguishable to me because I’m not being a semantic jackass.
LikeLike
What you are proposing is a way around it (I presume, after the tests are completed, to not do any item excising and keep it as is). However, how tests are constructed are not like that.
The test is then constructed with the assumption that scores should conform to a bell curve; as I already discussed (and provided evidence for), few natural biological functions conform to a bell-shaped distribution. Therefore, these tests are assuming that this natural biological function lies on a bell-shaped curve. I’ve provided evidence on traits important for survival (even reaction time).
Where is the biological evidence that, say, someone with 84 billion neurons has a lower IQ than someone with 90 billion neurons? Would it always be that case that someone with 84 billion and 90 billion neurons have the same IQs? If not, then your ‘measure’ is not ‘construct valid’.
So does someone who gets stronger synapses over time get smarter? If they get weaker do they get dumber?
Say you have 100 kids, 50 play action games and 50 do not. If it’s not controlled for then its confounded since people who play action games have better reaction times than those who do not. I’ve also been through it countless times that the low .2-.3 correlations with RT mean that other factors come into play and that the relationship between IQ and RT depends on numerous other variables than ‘processing speed’, that can be trained.
What? There is more to reaction time than processing speed.
misunderstanding instructions, familiarity with equipment, motivation to do the task, sensory acuity, learned response strategies, time spent on sensory processing and motor action rather than decision time, attention, arousal, task orientation, confidence and anxiety. (Richardson, 2017: 91)
Any differences in those potential confounds will change the RT speed. Therefore, with such low correlations, it shows that numerous other factors affect ‘speed of processing’ then how fast you react.
Few natural biological functions conform to a normal distribution (refer to the picture of the book I provided in my previous comment), the tests are constructed to assume the normal curve of the presumed differences between people. Therefore, since few natural biological traits lie on a bell curve (I showed evidence above), and since they do not, then IQ scores is one of them. I’ve already explained above that the test constructors create numerous items for testing, and excise the ones that don’t conform to the imaginary bell curve on human psychological traits (see the citations in my bell curve article). If IQ (‘g’) is a normal biological function, then it would not lie on a bell curve since, according to ‘g’/IQ theorists, it was important to survival in our evolutionary past. Most traits important to survival do not lie on a normal distribution. Therefore, “The falsity of the assumption must, therefore, raise serious questions about those analyses [few natural bio traits falling on a normal curve along with assuming a normal curve].” (Richardson, 2017: 81)
Well, I am semantic.
LikeLike
Test. For some reason my shit isn’t posting.
LikeLike
Just checked the spam folder and nothing. What’d it say when you tried to post your comment?
LikeLike
Nothing, it said it posted it but it never showed up. Fixed now though.
LikeLike
“Where is the biological evidence that, say, someone with 84 billion neurons has a lower IQ than someone with 90 billion neurons?”
Well obviously size of the brain is a factor and it has been shown to correlate with IQ, though only moderately, which has me interested in figuring out how well IQ measures actual capacity.
“Would it always be that case that someone with 84 billion and 90 billion neurons have the same IQs?”
Well it depends on what those Neruons are for really and the capacity says little of the actual Power. But all else equal yes I think it’s logical to conclude that.
“So does someone who gets stronger synapses over time get smarter? If they get weaker do they get dumber?”
Not necessarily. Stronger synapses implies more crystallized memory while weaker ones means higher adaptability to novel situations. Interestingly there is a theory that the brain can actually strengthen and weaken it’s plasticity at the SAME TIME, it’s one proposal for how Working memory functions. Amazing to say the least.
https://motherboard.vice.com/amp/en_us/article/yp3q75/neuroscientists-have-a-new-computational-model-for-memory-formation-in-the-brain
https://www.ncbi.nlm.nih.gov/pubmed/27694992
In regards to our “genes vs cell” argument I have seen new data suggesting that it isn’t synapses that form memories but instead the DNA withing the cell.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5112247/
“But when cells were exposed to a reminder of the shocks, the memory came back at full strength within 48 hours. “It was totally reinstated,” Glanzman says. “That implies to me that the memory wasn’t stored in the synapse.” The results were recently published in the online open-access journal eLife.
If memory is not located in the synapse, then where is it? When the neuroscientists took a closer look at the brain cells, they found that even when the synapse was erased, molecular and chemical changes persisted after the initial firing within the cell itself. The engram, or memory trace, could be preserved by these permanent changes. Alternatively, it could be encoded in modifications to the cell’s DNA that alter how particular genes are expressed”
” other factors come into play and that the relationship between IQ and RT depends on numerous other variables than ‘processing speed’, that can be trained.”
Again, i never stated at any point in time whether reaction time was due to environment or genetics. Get this through your head.
“misunderstanding instructions, familiarity with equipment, motivation to do the task”
Are the only things on that list that can actually not be considered a part of reaction time. Even then familiarity and comprehension both involve synaptic plasticity.
“Therefore, since few natural biological traits lie on a bell curve (I showed evidence above), and since they do not, then IQ scores is one of them. ”
Nope, that’s a fallacy, try again. You have to show that IQ doesn’t follow a normal distribution otherwise you must assume the null hypothesis. You don’t know so you can’t claim to know.
The forced distribution is there for a purpose it makes the comparison between subjects easier, you just have to use your brain and not take it at face value, Most Pychometricians are aware that IQ isn’t a perfect normal distribution, I already showed evidence for this, the excising of material does not actually make it less accurate because Psychometricians know there are more intelligent people that the curve lets on. You’re dressing this up as a purposeful manipulation of data to serve some underlying ideological purpose when mathematically it makes the system run more efficient, I’ve provided evidence of this as well.. But you don’t actually understand this and will choose to continue on with ad nauseam. All the while continuing to make yourself look stupid.
“Well, I am semantic.”
Really? it’s so hard to tell.
LikeLike
“you must assume the null hypothesis”
Then you’ll have to explain China’s state despite their “high IQ”, also data such as Mongolia having an IQ average based on Russia and China alone without direct sample…etc
In my opinion, your argument here sound like “Hey! God exist! You’ve to prove that god doesn’t exist to prove me wrong’.
LikeLike
“Then you’ll have to explain China’s state”
Their state? Oh please share with me how you think their way of life is so intellectually inferior to yours.
“also data such as Mongolia having an IQ average based on Russia and China alone without direct sample…etc”
Lynns methodology is irrelevant to the discussion at hand.
“In my opinion, your argument here sound like “Hey! God exist! You’ve to prove that god doesn’t exist to prove me wrong’.”
No. Taking the null hypothesis in this sense is neither agreeing with the alternate or the original. It’s more like Agnosticism not creationism. Plus I already provided evidence that Intelligence does form a bullcurve(though not a perfect one) and even the forced bellcurve is still a bellcurve, RR is relying soley on faith.
LikeLike
“Their state? Oh please share with me how you think their way of life is so intellectually inferior to yours.”
Well, they’re incredibly corrupt, and large portions of their population continues to live in abject poverty, despite having a supposed IQ of 106+ and a “racially/ethnically homogeneous” society. Same with Mongolia, for that matter.
“Lynns methodology is irrelevant to the discussion at hand.”
On the contrary, it’s completely relevant considering the fact that the overwhelming majority of so-called “objective” HBDers use his data as undeniable fact. Even if the many, many criticisms of Richard Lynn’s methodology and biases were non-existent, am I really supposed to believe that India, a country with many literary, scientific and philosophical achievements in history, for example, has an average IQ in the low 80s, while Moldova, a shithole that has accomplished virtually nothing has an IQ in the high 90s? Likewise the Bahamas, a rather functional, peaceful majority black country has an IQ of 84, possibly lower, while Russia, a country that for all intents and purposes is and has always been little more than a violent, corrupt, backwater shithole has an average IQ of 97, possibly higher?
LikeLike
Right, but that’s not construct validity. Hell, it’s not even a biological correlate.
Interesting stuff.
Though there is a pitfall:
Eric R. Kandel, a neuroscientist at Columbia University and recipient of the 2000 Nobel Prize in Physiology or Medicine for his work on memory, cautions that the study’s results were observed in the first 48 hours after treatment, a time when consolidation is still sensitive.
Interesting stuff nonetheless. I’ll keep an eye out for new papers on this.
Which would familiarize them with the equipment so they would then do better on the RT measures.
The whole point that I explained with test construction and the excising of items shows that it’s a forced curve (which you acknowledge). I showed that most important human survival traits aren’t normally distributed. I showed that even IQ correlates are not normally distributed. I’m literally explaining to you that they only fall on a normal curve because the tests are devised to do so. So you’re telling me that even if they weren’t devised to form a bell curve that they would anyway? Talk about bold claims…
Which fallacy did I commit?
Here’s something good:
A detailed analysis of test results obtained from a large sample of English children (4,665 in all), supplemented by a study of the meagre data already available, demonstrates beyond reasonable doubt that the distribution of individual differences in general intelligence by no means conforms with strict exactitude to the so-called normal curve.
IS INTELLIGENCE DISTRIBUTED NORMALLY? By CYRIL BURT
Where did you show evidence for this?
Ideology permeates both sides of this debate; if you say no you’re fooling yourself. Provide this evidence again; maybe I missed it. (You provided a slideshare that claims ‘IQ is theorized to be normally distributed’. Who theorizes this? Source? As far as I can tell, you’ve provided no evidence that excising existing material doesn’t make it less accurate.)
You’re aware that by manipulating item choice, you can have women score higher than men, blacks score higher than whites, and you can also change whatever ‘curve’ arises from these test scores. IQ tests are so scientific…
Which goes back to my claim that they devise these tests on the presumption of who is or is not intelligent.
You said:
A power-law distribution is not a bell curve. And you also didn’t provide any empirical evidence for your claims; I have.
See The log-dynamic brain: how skewed distributions affect network operations
“at many physiological and anatomic levels in the brain, the distribution of numerous parameters is in fact strongly skewed . . . suggesting that skewed . . . distributions are fundamental to structural and functional brain organization. This insight . . . has implications for how we should collect and analyze data.”
They review evidence against the bell-shaped curve for a lot of psychological traits.
LikeLike
rw95,
“Well…”
Again, you seem to not understand. China is not the only East Asian country. Ones like Japan with 1st world nutrition have been shown to have above average IQ’s. Corruption isn’t formed from differences in intelligence it’s a difference in morality that counts.
“On the contrary, it’s completely relevant considering the fact that the overwhelming majority of so-called “objective” HBDers use his data as undeniable fact. ”
I’m not the Majority, don’t straw man my views, we’re talking about IQ as concept and construct in general, not Lynn’s data. There are plenty of Psychometricians who denounced the opinions of Richard Lynn.
“am I really supposed to believe that India, a country with many literary, scientific and philosophical achievements in history, for example, has an average IQ in the low 80s, while Moldova, a shithole that has accomplished virtually nothing has an IQ in the high 90s?”
What exactly do you think an Intelligent country is supposed to look or behave like?
LikeLike
“Right, but that’s not construct validity. Hell, it’s not even a biological correlate
Brain size is biological. That’s not up for discussion.
“Though there is a pitfall:”
Yes I’m aware.
“Which would familiarize them with the equipment so they would then do better on the RT measures.”
The point is that they are still rooted in biological mechanisms.
“So you’re telling me that even if they weren’t devised to form a bell curve that they would anyway? Talk about bold claims…”
I don’t know, but neither do you.
“Where did you show evidence for this?”
The slide show plus the benefits of a normal distribution, but that’s okay, let me explain further.It’s sad that I have to explain why a normal distribution is not a bad thing.
“Another corollary is that the normal distribution makes math easy – things like calculating moments, correlations between variables, and other calculations that are domain specific. For that reason, even if a distribution isn’t actually normal, it is useful to assume that it is normal to get a good, first-order understanding of a set of data.”
“1) Usually the data distribution in Nature follows a Normal distribution ( few examples like – age, income, height, weight etc., ) . So its the best approximation when we are not aware of the underlying distribution pattern.
2) Most often the goal in ML/ AI is to strive to make the data linearly separable even if it means projecting the data into higher dimensional space so as to find a fitting “hyperplane” (for example – SVM kernels, Neural net layers, Softmax etc.,). The reason for this being “Linear boundaries always help in reducing variance and is the most simplistic, natural and interpret-able” besides reducing mathematical / computational complexities. And, when we aim for linear separability, its always good to reduce the effect of outliers, influencing points and leverage points. Why? Because the hyperplane is very sensitive to the influencing points and leverage points (aka outliers) – To undertstand this – Lets shift to a 2D space where we have one predictor (X) and one target(y) and assume there exists a good positive correlation between X and y. Given this, if our X is normally distributed and y is also normally distributed, you are most likely to fit a straight line that has many points centered in the middle of the line rather than the end-points (aka outliers, leverage / influencing points). So the predicted regression line will most likely suffer little variance when predicting on unseen data.”
The normal distribution is made is that comparisons between scored and other set of variables can be made, like how Pumpkin converts height or Sat scores to IQ, this impossible because of the normal distribution and even though it isn’t perfectly accurate it actually makes sense of data that isn’t highly variable.
“I showed that most important human survival traits aren’t normally distributed. I showed that even IQ correlates are not normally distributed.”
No, you cherry picked data, Pumpkin actually made a post displaying some normally distributed traits
“You’re aware that by manipulating item choice, you can have women score higher than men, blacks score higher than whites, and you can also change whatever ‘curve’ arises from these test scores. IQ tests are so scientific…”
All forms of measurement can be manipulated, this has nothing to do with normal distributions items are excised to make a normal distribution not, to discriminate.””Which fallacy did I commit?
“Here’s something good:”
The link says it’s a type 4 Pearson curve which still has an average it just isn’t perfectly symmetrical, semantics again.
“You said:
the distribution of humans would be a power-law like distribution with a heavy tail and different individuals would perform astronomically better at some tasks than others.
A power-law distribution is not a bell curve. And you also didn’t provide any empirical evidence for your claims; I have.
That’s my point, the bell curve makes the data feasible and scaled instead of purely additive.
“See The log-dynamic brain: how skewed distributions affect network operations”
Again not realizing what normal distribution are used for.
“‘IQ is theorized to be normally distributed’. Who theorizes this? Source? As far as I can tell, you’ve provided no evidence that excising existing material doesn’t make it less accurate.)”
You didn’t watch the slide obviously. It states the Gifted children are actually more numerous and it relates to the burt study you posted. Everyone theorizes this because most statistical traits are normally distributed.
LikeLike
rw95
“Likewise the Bahamas, a rather functional, peaceful majority black country has an IQ of 84, possibly lower…”
(not the main point I know of course, but) the Bahamas probably has an IQ somewhat higher, between 90 and the low 90s—as does Barbados, according to some more recent data (as with Monglolia Lynn did not have specific data on the Bahamas)
LikeLike
Whatever. I’ll grant that for this discussion. It’s still not construct validity.
No, the point is slower reaction times are about much more than just ‘processing speed.’
Right, and evidence I’ve provided makes the claim extremely dubious.
It’s not about you “explain[ing] why a normal distribution is not a bad thing”, it’s about the false assumption (refer back to the picture from the book above).
Income is not normally distributed. Going to need a citation for weight following a normal curve. And no, it’s not “the best approximation when we are not aware of the underlying distribution pattern”, because, for the millionth time the curve is constructed through item selection.
Yes it ‘makes sense’ of the data, but does it fit reality? Nope. Look at any stats textbook and it’ll state what I am saying now about the construction of the tests forcing a normal curve.
No I didn’t.
The examples he gave: height, birth weight, brain weight (physical variables) then a bunch of different IQ bell curves. I never said physical traits are not normally distributed. The claim is that ‘IQ’ is physiological. I’ve shown that physiological traits aren’t normally distributed. This makes sense since physiological traits are unstable which fluctuate based on what is occurring in the environment. I provided evidence that the curve is forced through item selection which, in effect presupposes who is or is not already intelligent. That’s how the curve is formed in the first place
Exactly. So if items weren’t excised…
And, again, the point is that this presupposes who is or is not intelligent (item selection), so if we presupposed women or blacks intelligent then we can select items they would score well on while excising items they do not. This also, of course, falls back to the social class/IQ test argument as well.
Still not normal. (Which would skew readings, which is the point.)
I’m aware, but I’ve provided a lot of evidence that there is no normal distribution, mainly through the fact of the forced curve.
Not realizing it was talking about physiological/psychological traits and anatomic brain structure not being normally distributed (using brain scans, not crude tests).
I was aware, I read the Burt paper.
Who is ‘everyone’?
Do you own any stats textbooks? Check a few out and you’ll see that they say what I’m telling you right now.
This book is stating what I am too.
It’d do you good to read this paper too: The Unicorn, The Normal Curve, and Other Improbable Creatures.
IQ scores are constructed to display a normal curve, but we can’t say there is a normal curve in nature or not. Micceria (1988) shows exceptions to the rule when measuring functions.
The Bell Curve in Psychological Research and Practice: Myth or Reality?
If IQ scores distribute normally, this does not mean that intelligence equally distribute normally in the population.” … “ In this way, a normal distribution in summated test scores, for example, would be seen as the sign of the presence of an error sufficient to give scores the characteristic bell shape, not as the proof of a good measurement.“
The Best and the Rest: Revisiting the Norm of Normality of Individual performance.
We conducted 5 studies involving 198 samples including 633,263 researchers, entertainers, politicians, and amateur and professional athletes. Results are remarkably consistent across industries, types of jobs, types of performance measures, and time frames and indicate that individual performance is not normally distributed—instead, it follows a Paretian (power law) distribution. Assuming normality of individual performance can lead to misspecified theories and misleading practices. Thus, our results have implications for all theories and applications that directly or indirectly address the performance of individual workers including performance measurement and management, utility analysis in preemployment testing and training and development, personnel selection, leadership, and the prediction of performance, among others.
LikeLike
“No, the point is slower reaction times are about much more than just ‘processing speed.’”
That is irrelevant to the discussion at hand, try not to repeat yourself. Ill give you another chance to respond to the comment properly.
“it’s about the false assumption”
It’s not a false assumption. The force is on purpose.
“And no, it’s not “the best approximation when we are not aware of the underlying distribution pattern”, because, for the millionth time the curve is constructed through item selection.”
Those two thoughts do not follow. We are aware of the distribution pattern that’s why we force it to a different one to begin with.
“but does it fit reality? Nope. Look at any stats textbook and it’ll state what I am saying now about the construction of the tests forcing a normal curve.”
LOL That’s the point. The curve creates sensible data and actually makes it more accurate.
“Still not normal.”
Do you know what normal means?
“and time frames and indicate that individual performance is not normally distributed—instead, it follows a Paretian (power law) distribution. ”
I literally just told you this. The paretian distribution is why we force it to begin with.
This is just cringe worthy all around. To avoid further misconceptions in the future I’m going to break each point of yours step by step
IQ is not normally distributed: We knew this.
The normal distribution is forced: We knew this.
This forced distribution makes data less accurate: Blatantly false, and only someone who doesn’t understand what Normal distributions are would ever say something like that.
So basically your argument is redundant. It assumes a lot of things, and tries to paint a “conspiracy”.
LikeLike
The power law distribution makes it impossible to meaningfully compare people so the normal distribution is made so that it can make a “rank order” it’s why the correlation with brain size is small.
LikeLike
the construct validity is that they’re designed so that blacks and italians score the lowest.
LikeLike
LOL RR you should make a response to this criticism of denis noble and his extended evolutionary synthesis.
https://whyevolutionistrue.wordpress.com/2013/08/25/famous-physiologist-embarrasses-himself-by-claiming-that-the-modern-theory-of-evolution-is-in-tatters/
LikeLike
Jerry Coyne is nonsensical when it comes to attacks on the Modern Synthesis.
Noble responded to Coyne.
LikeLike
I found it wanting, he didn;t fully respond to Coyne, but coyne also made another rebuttal as well.
https://whyevolutionistrue.wordpress.com/2016/11/23/once-again-a-call-for-a-revision-of-evolutionary-biology/
He raises great points, being to susceptible to environmental stress can be a bad thing, which would result in extinction more than adaption, if you live in a polluted city, epigenetics isn’t going to make you suddenly be able to breathe carbon monoxide, the only real benefit that could result would be maybe an increase in hemoglobin size resulting in an ability to take in more oxygen, but most adaptations would probably be detrimental. remember, Noble even said his views were still closer to Darwinian evolution than lamarkism
LikeLike
Those claims by Coyne I will address in the future (it’ll be too long to do here), but yes. Noble’s view is to extend the synthesis (while Laland and others call for extension as well), whereas Jabolonka and Lamb push to replace the Modern Synthesis.
Though Noble did have an interview on HuffPo:
I still don’t know where I lean here; both sides make good arguments that the MS neglects certain aspects of evolution.
Though, and I believe I have stated this before, evolutionary implications are highly theoretical (I will get you a quote from Evolution in Four Dimensions later for this claim), it’s fun to think about the possibilities in regards to DNA methylation, histone modification and acetylation of histones.
LikeLike
it may not be strictly a normal distribution, but it has to have two tails, unlike a power law, which has one.
Everyday life is also more complex than the hardest items on Raven’s Matrices, while the test is not complex in its demands compared to tasks undertaken in everyday life (Carpenter, Just, and Shell, 1990). They conclude that the cause is differences in working memory, but that is an ill-defined concept in psychology. They do say, though, that “The processes that distinguish among individuals are primarily the ability to induce abstract relations and the ability to dynamically manage a large set of problem-solving goals in working memory.” So item complexity doesn’t make Raven’s items more difficult for others, since everyday life is more complex.
it becomes more complicated when you put a time limit on it
Measuring human intelligence is more complicated than measuring blood alcohol concentration, but that does not render IQ tests useless. Is it a coincidence STEM people typically have IQs +2 SD higher than average? Or that high scorers learn material faster?
LikeLike
Even if it’s not ‘strictly’ a normal distribution, the assumptions and conclusions drawn from this imaginary distribution may be seriously flawed. (I’ve provided evidence directly against the contention.)
OK? The matrices are still 1) not complex and 2) Raven’s contains hidden structure which make the test more—not less—culturally steeped. Further, failure on Raven’s items may just be due to not being familiar with the specific relations. Goswami argues that to see if these tests do really test complex cognition that we should attempt to devise items familiar to all of the groups being tested.
Analogical Reasoning in Children
Not the point. The point is we have construct valid measures to measure blood alcohol levels in the blood; the same isn’t true for IQ tests and no, correlating IQ tests to other IQ tests and achievement tests don’t count.
LikeLike
I agree the matrix tests may not be totally culture-fair, but they are an attempt to make a test that does not rely on cultural norms, and second, to address the common left-wing objections that tests that have verbal and math components are biased in favor of those who have high educational attainment. Some IQ tests asks questions such as “Brother is to Sister as Niece is to: Mother—Father—Aunt—Uncle—Nephew ” that only English-speaking people may know.
As for the time limit, yes they are not that hard, but they are not supposed to be, due to the time limit and for practicality. The idea is if someone can solve a difficult matrix problem in 60 seconds, they can solve a more complicated real-word abstraction given more time.
LikeLike
Goswami argues that to see if these tests do really test complex cognition that we should attempt to devise items familiar to all of the groups being tested.
yes that would be ideal..but possibly also impractical if one has to administer tests to hundred+ countries. How would one contrast a special culture-far test for American Americans? How come Asians and Jews, that have their own unique cultures, do well on IQ tests for White, Anglo norms?
LikeLike
There are outstanding arguments for this. Read What IQ Tests Test.
A lot more goes into solving matrix problems than ‘g’ or ‘complex cognition’ as I’ve cited.
Because their cultures have some of that “deep underlying structure” that the Raven entails. The point with the Tsimane proves this.
LikeLiked by 1 person
[…] IQ Test Construction, IQ Test Validity, and Raven’s Progressive Matrices Biases […]
LikeLike
“Reading was the greatest predictor of performance Raven’s…[blah blah, rest of that quote…blah].”
Could be that reading is highly part of verbal IQ and since verbal IQ is highly g-loaded and the Raven’s is g-loaded, these predict each other by all going hand-in-hand of g. But that’s just my guess.
LikeLike
It’s because reading is correlated with schooling. (Whatever g is.) Raven’s isn’t culture-fair. It’s perhaps one of the most culture-un-fair tests due to hidden structures.
LikeLike
“It’s because reading is correlated with schooling. (Whatever g is.) Raven’s isn’t culture-fair. It’s perhaps one of the most culture-un-fair tests due to hidden structures.”
I think we could answer this question by looking to see if the reading-IQ correlation is in “g” since “g” seems to be the least environmentally malleable; see if it’s just from verbal IQ confounding them or if it’s truly school.
“Raven’s isn’t culture-fair.”
Don’t know if I would say that, only because I don’t think the Raven’s is taking in any specific X-factors between cultures. So far it seems to just be education effects and Flynn Effect that is hidden structures in it. So I’m pretty hesitant to say it’s not culture fair due to that (since it would be educational (outside of the FE) and not cultural… unless we can’t education as culture). Or here, let me put it another way: Group 1 naturally does X which leads to an IQ of 100; Group 2 does Y which leads to an IQ of 110. Is it cultural un-fairness towards Group 1 or is just Group 2’s natural-ness that leads to a higher IQ? I mean, they both naturally do X and Y, but the IQ differences stems from the environmental difference due to them doing X and Y. See what I’m getting at? It’s kinda hard to tell what’s “pure” environment and what I should be calling culture or not. Or least that’s how I currently see and maybe I’m just stupid or something.
“It’s perhaps one of the most culture-un-fair tests due to hidden structures”
Any suggested test that I should looking out for that’s cultural-fair?
LikeLike
“unless we can’t education as culture”
unless we count education as culture*
LikeLike