How does IQ test actually work?
In: Biology
Brief background, I am a PhD researcher in psychology and I have published papers on intelligence, and particularly the Flynn effect which is the increase in measured intelligence found in most countries.
This is long so I’m putting the most important thing first: your IQ is not your worth. People have an inherent dignity that is equal and inviolable regardless of how smart you are. Albert Einstein does not have more value as a person than someone who is incapable of tying their own shoelaces. I think people get really defensive about IQ and intelligence because our society values intelligence to an extraordinary degree. If IQ tests do what they purport then (1) people are not equal on this valuable trait and (2) we can objectively determine who does and does not have more or less of this valuable trait. People then start to think that we have a test that we might try using to determine someone’s worth, but your IQ *does not determine your worth.* Your IQ determines your value as a person as much as your height does, which is not at all.
IQ tests today are typically either something like Ravens progressive matrices, which are a series of pictorial puzzles of increasing difficulty, or they are somewhat more traditional tests that include a variety of problems centered broadly around “reasoning”. Modern tests are highly sophisticated instruments subjected to very rigorous statistical methods to ensure a few things (1) that the measure what they say they measure (2) that they do so in an unbiased way and (3) that they do so accurately. “How do IQ tests actually work?” Well, after the test is developed you take the test, the test is scored (this can be either a simple summary, or for more sophisticated tests, a score that takes into account the difficulty of the specific questions you answered correctly, how well they tend to distinguish high from low IQ individuals, how well they measure IQ etc.). This score is then compared to some “norm”. A norm is simply the distribution of scores for some group of people (say 20-30 year olds, measured in 2020). Your score lies somewhere in that distribution and we tell you where you stand compared to everyone else. Usually this score is adjusted so that the average person has a score of 100 and the standard deviation (kind of like the average difference from the average) is usually either 15 or 16 points.
How do we decide that the tests measure intelligence? Well, do they predict outcomes that we would expect to occur based on differences in intelligence? For example, if you have a job that requires a “smart person” do people who have high IQ’s tend to do better in that job? (The answer is yes.) IQ tests are predictive of a number of things that we tend to associate with “intelligence” as a concept. Higher IQ is generally predictive of higher levels of education (i.e., *before you get the education you have a higher IQ*). Higher IQ is generally predictive of better job performance in jobs that require critical thinking and an ability to solve complex problems. It is predictive of maintaining your health better, etc. *This is not to say that IQ is the only predictor of these things*. However, IQ is one of the best psychological predictors of these things, generally speaking the only other psychological construct that comes close to having the same kind of predictive ability is Conscientiousness (which is, roughly, your ability to act in a way that is considerate of others). IQ is also predictive *above and beyond* things that people commonly raise as being what IQ really measures (particularly socieoconomic status).
You’re going to get a lot of comments to the effect of “we don’t really know what IQ tests measure” or “IQ tests don’t really predict anything.” That’s pretty much categorically false, and not a position held by the vast majority of intelligence researchers. It’s a fairly anti-science position, bluntly. Most of it appears to come from Stephen Gould’s “The mismeasure of man.” That book was pretty widely criticized by pretty much the entire community of intelligence researchers. The issues he raised were either (1) his own misunderstandings of the science, (2) out of date, or (3) flatly wrong. You will see a lot of people say “well you take a standardized test with multiple choice answers, but life doesn’t have multiple choice answers, so really that’s meaningless.” No, it’s not. The tests are designed to test your ability to use information and solve problems, that you can choose from a variety of answers doesn’t change that you’re solving the problem, it’s just far more convenient from a test creation perspective.
Again though, because I can’t say it enough, these tests do not, will not, and cannot, determine your worth as a person. A smart person can be a monster, and a dumb person can be a saint, which one you are really doesn’t depend on how smart you are.
Latest Answers