It has long been known that people who have high scores on intelligence tests are also relatively fast at simple laboratory tests of ‘mental speed’ (Woodrow, cited in Woodworth, 1946). When first observed these correlations attracted little interest because, though ubiquitous, they are very modest, accounting for only 3–9% of variance between individuals (Sternberg, R. J., 1977). Correlations between intelligence test scores (TSs) and speed of performance on simple tasks were seen as less interesting than equally strong associations between TSs and performance on simple tests of memory and learning (Woodrow, 1938). The idea that intelligence can be defined as the capacity to learn new skills seemed a more fruitful guide to the development of individual differences in competence (Woodrow, 1946). Further, since scores on most intelligence tests represent the number of problems correctly solved within a time limit, it did not seem remarkable that people who can solve complex problems quickly should also be able to make simple decisions fast. Finally, evidence accumulated that, speed and accuracy of problem solving can be independent; for example, the times that individuals take to solve complex spatial problems do not predict their error-rates (Egan, 1978). Consequently mental speed was seen as only one of many different demands that solving intelligence test problems make on the cognitive system. Many distinguished investigators believe that better insights can be gained into the nature of individual differences in intellectual ability by studying correlations between TSs and performance on complex tasks that demand the encoding, organisation and transformation of information in working memory or successive stages in syllogism solution (e.g. Carpenter & Just, 1989; Daneman & Carpenter, 1980; Kyllonen & Crystal, 1990; Necka, 1992; Sternberg, R. J., 1977).