IQ often gets used to promote racism on the internet but it's also highly mathematical so people don't really understand it.
In this thread, I will explain:
1. How the math of IQ works at a high level (don't worry no formulas!)
2. Why IQ is partly socially construct... 🧵👇
In this thread, I will explain:
1. How the math of IQ works at a high level (don't worry no formulas!)
2. Why IQ is partly socially construct... 🧵👇
An IQ test is basically a list of questions that a psychometrician (kind of a cross between a psychologist and a statistician) thinks might measure intelligence. If you give such a test to a lot of people, you will get a range of scores.
The first thing you might notice is that the distribution of scores isn't a nice shape. So psychometricians adjust the scores so that a certain number fall within certain percentiles so you get a nice bell curve like this: 
It's kind of like when a professor curves the class test scores so a certain number of students get As, Bs etc.
The second thing you might notice is that men and women score differently. Psychometricians re-weight the questions men tend to get right and the ones women tend to get right so that men and women get the same average scores.
When I say "re-weight" what I mean is you essentially change how much points test-takers get for different questions until you get a situation where men and women are scoring about the same.
The third thing you might notice is different IQ tests give different IQ scores so psychometricians do some more re-weighting of the questions within each test to make all the tests give similar numbers.
The core assumption of the re-weighting procedure is all tests are measuring the same thing and there's some single common factor (intelligence???) that causes a person to score highly across all tests (kind of a big assumption in my opinion).
Finally, remember in the first step we had to give the test to a lot of people to get a baseline? Turns out different populations will have different score distributions.
If you pick 50-year-old white Americans in 1900 you will get one baseline. If you pick black Canadian teens in 2021, you will get a different baseline.
Selecting a sample of people to create a good baseline is extremely important. That baseline population becomes the yardstick by which everybody else is measured.
As you can see, there's a lot of complicated statistics that goes into creating the IQ construct. It's a statistical construct but it's also a social construct based on an agreement among psychometricians about how to adjust the raw test scores.
To summarize, IQ tries to capture how much doing well on one test designed by psychometricians would predict that you will do well on other tests designed by psychometricians.
Personally, I don't find it surprising that IQ has a (very weak) relationship to life success. Obviously, tests measure something useful. That's why we use them to assess how well students are doing in school.
Many people seem to think IQ is a completely objective concept. This is clearly not true. I don't think IQ tests are completely useless but there are a huge number of subjective design decisions that go into how these tests are constructed.
Hope you found this thread useful. This kind of long-form content takes extra work so if you like it and want to show support, like and retweet the thread, and give me a follow! 👍
• • •
Missing some Tweet in this thread? You can try to
force a refresh
