Why IQ Tests are Way Bogus



The above real IQ question appeared in New Scientist magazine (a British publication, hence the odd phrase "odd one out" to refer to the one that's most fundamentally different from the others). The following text paraphrases and extends a reader's complaint.

The leftmost image is the odd one out with respect to it being the only one that's drawn at a different scale than all the others. The second one is the odd one out with respect to it being the only one that has no curved lines. The third one is the odd one out with respect to it being the only one that's bilaterally symmetrical along only one axis rather than two. And the fourth one is the odd one out with respect to it being the only one that's not unique in some obvious way. So any of the four is a reasonable answer.

But from my experience with such tests, it's a very good bet that the "correct" answer is number 3, because I've learned that the organisms that concoct such tests regard bilateral symmetry along a vertical axis as being an extra special important thing. I expect that the reason for this bias is that these organisms are themselves roughly symmetrical along a vertical axis (or at least the outside parts of them that you can see are), and so they think very highly of this particular trait. This, therefore, is an example of an allegedly abstract test question primarily measuring familiarity with the biases of the people who created the test. So there.

(Perhaps the question is really supposed to detect whether the test taker has a "normal" bias toward this human trait. But I would hold that it's more a sign of ignorance of one's own biases to regard such a trait as inherently more important.)
More Random Cruft from Ken