Age, speed of information processing, recall, and fluid intelligence

On three occasions, 63 adults, ranging in age from 26 to 80 years, all in good health, were tested with three speed of information-processing paradigms (the Sternberg, the Posner, and the Hick), two long-term free-recall tasks, and, as a measure of fluid intelligence, the Raven Advanced Progressive Matrices (RAPM) test. Whereas within-condition latencies for the three of the information-processing tasks and recall scores were found to be reliable and consistently correlated with age and RAPM, individual differences in within- condition accuracies and between-condition slopes produced by the three information- processing tasks were found to be unstable over time and unrelated to age and RAPM. As suggested by Salthouse (1985), a large portion of the age-related differences in fluid intelligence was found to be accounted for by age-related declines in a general latency factor (cognitive speed). Furthermore, in agreement with Salthouse, this general latency factor appeared to reflect more than what can be accounted for by the simplest of information-processing tasks (simple reaction time). Finally, given that free recall had a substantial independent effect on RAPM when age and latency were held constant, the results called into question the assumption that cognitive speed can account for all individual differences in IQ.