Validity of IQ

[Alice Amell]

I have been wondering about the validity or usefulness of IQ for a while. I have not studied it very much but am looking up studies through PsycInfo database. My basic question is: Is IQ a valid construct, and how rigid and predictive is it? Even if most people do not change, and do not make an effort to change IQ or overall standing in life, does not mean having a low IQ predetermines you for low success. Personally at the moment I see it as irrelevant and it is more productive to focus on improving my life, ignoring IQ. Maybe this stance is wrong though.

 

I came across one study which looked at childhood IQ (at age 13) and looked at later life outcomes. The study says

 

"Most children with high IQ scores came from the families of professionals and other non-manual workers. These children were raised in much more advantageous conditions than the children with low IQ scores. Thus, it is difficult to assess to what extent our respondents' intellectual functioning in adulthood and their own life careers resulted from their genuine capabilities and to what extent it was actually stimulated, or limited, by the environment in which they were brought up..."

 

Basically they can't say whether its their background or their IQ that predicts outcomes because those are highly correlated.

 

Those with higher IQ got on average more education, better jobs (professionals, technicians vs. service workers, manual workers) and made more money.

 

In the discussion it says

 

"In the cases of persons with identical IQ scores but different social backgrounds, we discovered that the predictive power of IQ was limited."

 

In AP psych in high school I learned intelligence is about half nature half nurture. So I wonder to what degree it can change later on too or if someone can escape their upbringing so to speak, similar to how many here deal with their childhood and try to improve their lives. In that sense IQ might predict where you are in a sense but not where you have to end up. It also doesn't capture all types of intelligence (like social).

[Wuzzums]

IQ is not deterministic, there's still choice, not every tall man becomes an NBA player and not every high IQ person becomes a genius. For instance, Einstein had an IQ of 160 and Marilyn vos Savant has an IQ of 228. Having the capability/potential of doing something does not mean you're already able to do it. An actor with a 200+ IQ is not capable of beating a chess grand master. Skill is not innate.

 

Skill however requires some ingredients, and those ingredients don't always depend on IQ. I need to be tall to make it in the NBA regardless of how many hours I pour into playing basketball. No matter how hard I try and how bad I want it I must accept my limits, and this is the true importance of IQ I believe. 

 

If IQ has a predictive value in job outcomes this does not imply that having a certain IQ will land you a certain job. You must look at things in reverse, if a job requires a high IQ then people with high IQs will get that job. I have a working set of hands, I'm not determined to be a guitar player but I can safely say that all guitar players are determined to have a working pair of hands. IQ is not the ticket to success in certain jobs, it's the currency you need to buy that ticket.

[Existing Alternatives]

I need to be tall to make it in the NBA regardless of how many hours I pour into playing basketball. No matter how hard I try and how bad I want it I must accept my limits, and this is the true importance of IQ I believe. 

 

It wouldn't be prudent to spend all the time necessary to become an NBA player, if you are short, but it is not impossible...

- Tyrone Bogues at 5'3" is the shortest NBA player

- Jim Abbott was born without a right hand and played as a pitcher in MLB

- Oscar Pistorius lost both of his legs and run in Olympics

 

I don't know what the distribution of IQs among rocket scientists looks like, but I am sure it has a pretty long left tail...

 

There is predisposition and then there is will. 

[Josh Takacs]

I think a better comparison is teaching a dog tricks. There is only so much you can teach a dog and only so much they can truly understand. Some dogs can learn more then others just like some are just stronger or faster then others. Its the same basic concept with people. Now I can't say if IQ tests are a valid way to see how "smart" a person is because I too have not studied them. 

[TheFifthApe]

My basic question is: Is IQ a valid construct, and how rigid and predictive is it?

 

Great topic. I think Wuzzums summed it up quite well. It is valid if you apply it inversely to the proposition at hand. For example, let's say you are black and tall and most NBA players are black and tall thus you have a better chance at playing in the NBA. But obviously because you are black and tall that does not mean you will play in the NBA. This is where many other important factors come into play; socioeconomic status, likeability, passion, IQ, willingness, determination, forunate circumstance, location and so on.

 

Predictability is based on a high or low probability and probability is based on a wide range of factors (IQ is just one of them). An individual has complete control over themselves (to a degree), but one's destiny or the effect that one's environment has on oneself is out of one's direct control.

[Pepin]

If IQ has some predictive power it is useful. It is like grades in college, if it gives some indication of how successful a student will be in their profession, then it has value. The predictive ability doesn't need to be high, it just needs to be high enough to matter.

 

IQ isn't the only factor, but many studies demonstrate that it is a significant factor. If it contributes more than 20% to any particular outcome, such as income, it is a useful predictor.

 

Also keep in mind that we are always talking about statistics when we talk about IQ. Someone who has a higher IQ will on average make more money, but this does not mean that every person with a higher IQ will make more money.

 

For similar reasons, single studies aren't that predictive either. Where people get confused with science is that in some fields, 5-10% of studies are expected to contradict the all of the other studies. This would be true with the same researchers carrying out the same methodology. It is a result of statistics and can't really be avoided in softer sciences. I'm saying this because you are likely to find a lot of studies that don't find IQ to be useful, but you have to weigh them against all the studies that do.

[Better Future]

Are their examples of high IQ children who have absolutely horrible parents and teachers? Is there any evidence showing a wide range in natural biological IQ? How much can a human brain vary?

 

When I see "high IQ" I read "great parents and teachers". This is the reason I agree with Stefan. Parenting skills are the closest thing we have to a silver bullet when it comes to improving the wellbeing of the species. If there is a wide range in natural IQ then maybe we could include genetic selection.

[Pepin]

IQ is measured on the bell curve, with the median IQ always being 100. There is a wide range that is observed and it can be validated. For instance, if you test 1000 random people, about 20 of them will have an IQ from 130-140. If you find that this correlation doesn't hold up, then there is likely some adjustment needed for the model. Now if IQ doesn't fit on a bell curve, that would provide a lot of challenges to the model, but like everything else it seems to be.

 

Most studies tend to find little correlation between environment and IQ.

 

Wiki is a great source.

 

http://en.wikipedia.org/wiki/Intelligence_quotient#Genetics_and_environment

 

With that said, IQ isn't everything.

[Better Future]

There is a wide range that is observed and it can be validated.

 

I am aware of that but I'm talking about natural IQ. I've never seen any evidence that shows variation existing at birth.

 

IQ is associated with parental social status and biological parental IQ. This suggests to me that IQ is mostly culture based. I would say that the bell curve just represents the variation in parenting skills. Exceptions would be conditions like down syndrome.

 

Children with high IQ/great parents are likely to be more successful in life.

[Alice Amell]

Thank you all for the replies. I think I understand it more and should look into more research to understand how predictive it is.

[thelion]

A quick summary of the literature over lunch.

 

IQ (and things like the SAT or GRE) are primarily based on working memory performance (as opposed to general memory).

 

And working memory performance significantly varies whether it is tested under pressure. Such as a time limit.

 

Although I am not a big fan of the nobel prize winning economist (he's paternalistic socialist) whose experiment this is, it's a nice demonstration of the issue. (For if you have no time limit, you can chunk random data and chunk the chunks and memorize arbitrarily long random sequences. See George Miller's famous review article, or his book with Pribram and Galanter.) The experiment

 

Mentally, take a number, say eight digits, all different. Now, every four seconds, add one to one of the numbers, then after four more seconds, add one to the next order of magnitude down, and so on, until you get to the ones, then in the next four seconds return to the beginning of the sequence, the hundred millions, and keep going.

 

Periodically you are quizzed as to what the number you have remembered is. When you make an error, the test ends.

 

Notice how your ability to succesfully keep this up decreases if the adding is done every two seconds rather than four seconds. Or each second. Or each half second.

 

This experiment however highly predicts IQ and SAT and GRE scores. Often, if you do poorly on this experiment, you will know in advance what your IQ or GRE score shall be without taking that particular test, the said experimenter found out. But is that really general intelligence, or is it solely working memory, which a very small part of what cognitive and social scientists comprehend as intelligence (below)?

 

Working memory can be improved by exercise. Computers will always beat you at this however; it's ultimately pointless in a tool-using division of labor society, in our society.

 

The excercise is a predictor of social conformity. The remaining minor deviation is due to vocabulary knowledge, most determined by daily social environment, since this too is tested. The natural genetic variation predicts natural ability to work under pressure, since the real key is the time limit on problems, not the mental challenge of the task itself.

 

The brain is mostly an experience based model constructor, trying to predict the future state of its environment. It uses working memory to process afferent feed into the extrinsic nervous system; this is a small part of what determines adult general intelligence.  But differences in experience, outside working memory, are "unfair" test material. I go to the library and read a book by accident, and my model updates based on what I learn. The kinds of inferences I can do are outside the scope of what another can do if they neither read the book also nor rediscover the material.

 

So working memory, a very small subset of brain function, depending as it does primarily on how many hours you train it with pointless in themselves exercises, otherwise varying randomly in a bell curve for each species without further training, is "fair" to test. Because experience based gains are solely due to putting effort into generally boring and pointless tasks like mental arithmetic -- which have become magic tricks instead of something useful in an age of computers. (See Wolfram.)

 

Computers do working memory better than humans. Humans have general memory, and can do what computers cannot do, but this is due to experience and experience varies nonrandomly. Whoever read more would win, having a more coherent model through which lens to interpret future experience and make predictions with less error. Garbage in garbage out is true, but "unfair".

 

Testing working memory means really testing that part of intelligence which varies in bell curve and exceptions are what Darwin called costly signalling, a show of willingness to do useless work to signal sunk investment and good intent by such means -- conformity. This desired in lowlevel and midlevel workers by government and employers. (But not desirable in high level workers and researchers, but these are a minority of students.)

 

IQ (and similar test like SAT and GRE, proxies) are used because they DO NOT measure not intelligence in general. Competent physicists or mathematicians who self selected through training will beat anybody on logical questions without a time limit.

 

The standardized tests measure daily performance under pressure, which is what employers and governments are looking for. They are not measuring what the cognitive science community calls general intelligence, which is the ability to make true inferences and anticipate cause and effect truly in the natural environment. General intelligence doesn't even require effort, for people can learn by gaining experience they like to gain. So it's a bad proxy for work under pressure, since it's self selected.

 

A nonhuman species cannot learn calculus but humans, who can, are not born knowing calculus. Yet the ability to make inferences is what human intelligence is about. Hebb considered that only a couple percent of human intelligence traits are genetic beyond the fact that they are what any trained person in the human species could do with proper conditioning. Another few percent are genetic disorders and result in retardation. The rest is experience. But again, most institutions don't need to test intelligence. They require prediction of goodwill and rule following plus ability to work under pressure ... and indeed this does predict performance at the undergraduate and graduate level plus entry level work performance.