Arthur Jensen is a controversial figure in psychology, due in large part to claims about racial differences in intelligence. In his newest book, "Clocking the Mind," Jensen turns his attention to a more focused, yet still controversial topic: how is it that extraordinarily simple measures of reaction time can correlate so highly with intelligence?



To understand the importance of this question, consider the following. First, as Jensen notes, almost all reliable measures of cognitive performance are correlated. Across a large number of such tests, a single number - termed g, for "general intelligence" - can account for a large portion of individual differences on each task. Because no single test is "process pure," the correlations between g and scores on any given test are typically rather small; high correlations emerge from these measures only when they are considered in aggregate, with the following exception.



Despite the fact that g is commonly assessed with tests of vocabulary, memory for associations, reasoning ability on the Raven's Progressive Matrices (where subjects must discover a visual pattern within a matrix of stimuli, and select what the next pattern in the sequence would look like), and a wide variety of other very abstract and untimed tests, it appears that the variance they share can be reliably and accurately indexed by reaction time on a task where subjects must merely press a lighted button. The correlations between such simple tasks and g is around .62, which is higher than the correlation between many subscales of IQ tests and the g factor to which they contribute.



If you are skeptical of these results, you are not alone. Jensen notes a deep-seated bias against the idea that such simple measures could reveal important traits of the cognitive system, and reviews several historical reasons for this bias. However, in just over 200 pages, Jensen creates a persuasive argument for the RT-IQ correlation based on dozens of factor analyses, and both developmental and genetic work. In the process, he covers issues related to statistical methodology, procedural variations on simple RT tasks, and correlations between simple RT and Sternberg memory scanning, working memory, short-term memory, long term memory, and a variety of other cognitive constructs.



In the end, it appears that simple RT and g may be very closely related, if not indexing the same thing. Jensen advocates the "bottom-up" interpretation of the RT-IQ correlation, suggesting that individual differences in processing speed allow those individuals to think faster, accumulate more information per unit time, and provide other advantages that subsequently translate into g. Jensen notes that the "top-down" interpretation - for example, that increased IQ leads to better strategy-use, and for that reason result in lower RTs on simple tasks - is plausible but relatively uninteresting for those interested in mechanistic rather than merely descriptive accounts of intelligence. Whether or not you agree with Jensen's "neural oscillation" hypothesis of the RT-IQ correlation, these facts beg for a mechanistic explanation.



Jensen's writing is clear and concise, and every chapter is densely packed with information. The historical treatment of chronometry is perhaps most enjoyable, filled with personal anecdotes and unique insight into the politics of 20th century psychology and psychometrics. My only complaint is the index seems sparse for a book so rich in detail.



"Clocking the Mind" is not a popular science book; it is a scholarly work directed towards professionals and graduate students. Yet, anyone with a scientific interest in individual differences, intelligence, or executive functions will find much to consider here. After all, if Jensen is right, relatively simple and extremely reliable measures of reaction time might be a good replacement for some of the "fancy tasks" cognitive scientists have spent decades refining.