Testing Intelligence

Are we really asking for a measurement of smarts, or is it something different?

When we think of intelligence, most of us think of the following words: smarts, reasoning, creativity, and ingenuity. From the Binet-Simon Intelligence Test to Raven’s Progressive Matrices (and even the SAT!), intelligence testing is everywhere, and it comes in many different forms. Do all of these tests measure creative thinking or atypical reasoning? Do any of them?

Intelligence testing probably began with Alfred Binet, when he developed the first version of the Binet-Simon test to identify children who needed to be singled out in school for extra attention. Thus, from the beginning the test was a practical device for sorting students in terms of probable school performance. When we look at a similarly formulated test now, we don’t see it as what it was made for, and then we are surprised when it correlates with socioeconomic status (better schools, more involved parents) and level of education.

Beyond this issue of construct validity, we can also see that most traditional intelligence tests (such as the Binet-Simon, the SAT, and the Wechsler Scale) rely at least in part upon verbal reasoning. While this is a laudable goal and likely represents a very real avenue of inquiry, verbal reasoning questions overwhelmingly favor those individuals best acquainted with the English language specifically, and yet we are surprised that immigrants and those people in the lower classes perform worse than the higher-classed people who have lived their whole lives in Canada or in the United States. One solution to this problem of translatability to other cultures is non-verbal questioning, such as spatial reasoning in Raven’s Progressive Matrices. However, yet again we see stratification by socioeconomic status.

If we were testing brain capacity, flexibility, or efficiency, we might use brain scans, combined with creativity tests, social problems, and computational skill. We don’t do this. Instead, our intelligence tests predict job performance. Maybe we’re measuring perseverance, willingness to cooperate, or focus. Without a clear idea of what “intelligence” means, we cannot possibly test and quantify it in a series of multiple-answer questions.