Does the rise in IQ scores over the past century mean people are getting smarter? Since the beginning of the 20th century, IQ scores around the world have been increasing at a rate of around three points per decade, leaving intelligence researchers puzzling over whether historical gains in IQ — known as the “Flynn effect” — reflect an increase in general intelligence or something else, be it better education, better nutrition or even bigger brains.

A new paper published in the journal Learning and Individual Differences may have the answer:

We’re getting better at taking tests.

People today are not only staying in school longer, the authors point out, but are more than ever taught to the test: Students are trained in test-taking strategies and heuristics that, according to the paper, can be applied to IQ-type problems. “People are exposed to the formats of tests all the time — they are able to detect certain regularities, and they are able to exploit those regularities,” said Michael Woodley, one of the paper’s co-authors, in an interview. “You were probably taught in school, for instance, to guess on multiple choice tests.” Even outside the classroom, increasing exposure — often online — to cognitive games like Sudoku, Bridge and Go mean that people are more familiar with IQ-type problems when they sit down to an IQ test. “We live in a more cognitively intense environment than ever,” said Woodley.

Woodley and his co-author, Elijah Armstrong, ranked 14 IQ tests according to the extent to which they could be solved by applying and re-applying rules. Some tests, like Raven’s Progressive Matrices, ask test-takers to detect complex patterns — but figuring out and reapplying a few simple rules can result in a substantial boost in scores. Other tests, like vocabulary tests, depend on recalled knowledge; they cannot be solved by using rules. (Determining the rule-dependence of tests is the most subjective part of this paper.) When Armstrong and Woodley compared data on the Flynn effect for each of the 14 different IQ tests, their results were striking: The more rule-dependent a test, the more pronounced the Flynn effect — suggesting that the Flynn effect is not due to increases in general intelligence, but to a better ability to short-circuit the test by detecting and applying certain rules.

Playing Sudoku, then, probably doesn’t increase your general mental ability — but it could improve a specific, if narrow, cognitive function. Getting better at applying rules to new test questions is another skill that might not reflect general intelligence but could still have specific applications. “The gains in IQ are not meaningless,” Woodley explained. “The Flynn effect does not mean people are getting smarter, but it does reflect people developing a huge range of narrow cognitive specializations.”

Have something to add to this story? Share it in the comments.