[michaelt]

One of the most oft-cited examples is of an intelligence test asking questions like "Dish is to place mat as teacup is to _____" where the correct answer is saucer - making the test biased in favour of people who use place mats and saucers at home.

[krona]

This is a bit of a strawman because IQ tests today rarely rely on language comprehension.

[JohnBooty]

I suspect that a familiarity level with western-style IQ tests may be part of the story as well.

I was issued such tests several times as a young child. The first time, I cried and did miserably. It was all so strange and unfamiliar. Had never seen logic puzzles before and didn't understand what they were asking of me.

My mom took me back a few weeks later and paid for me to re-take another IQ test. I did so well they placed me in "gifted" classes where I remained throughout my primary education. I didn't magically get smarter; I just became more comfortable with the test format.

I had a similar experience as an adult in my late 20s. Bought a Mensa book containing two practice exams. Took one and fell about 15 points (a full standard deviation!!!) shy of my childhood score. A week later I took the second test and equaled my childhood score almost exactly. Again, I am assuming I didn't sprout a bunch of new synaptic connections.

I suspect that second chances like these and familiarity with IQ tests in general correspond tightly with socioeconomic background. The more privilege you have, the more likely you are to get the chance to become comfortable with these sorts of tests and score closer to your maximum personal potential.

If I, an American, take an IQ test right now... I'm probably going to score better than somebody from another part of the world with equal intelligence who's never seen one of those tests before.