How Big Is a Chunk?: By combining data from several experiments, a basic human memory unit can be identified and measured

Abstract

I have explored some of the interactions between research on higher mental processes over the past decade or two and laboratory experiments on simpler cognitive processes. I have shown that, by viewing experimentation in a parameter-estimating paradigm instead of a hypothesis-testing paradigm, one can obtain much more information from experiments-information that, combined with contemporary theoretical models of the cognitive processes, has implications for human performance on tasks quite different from those of the original experiments. The work of identifying and measuring the basic parameters of the human information processing system has just begun, but already important information has been gained. The psychological reality of the chunk has been fairly well demonstrated, and the chunk capacity of short-term memory has been shown to be in the range of five to seven. Fixation of information in longterm memory has been shown to take about 5 or 10 seconds per chunk. Some other "magical numbers" have been estimated-for example, visual scanning speeds and times required for simple grammatical transformations-and no doubt others remain to be discovered. But even the two basic constants discussed in this article-short-term memory capacity and rate of fixation in long-term memory-organize, systematize, and explain a wide range of findings, about both simple tasks and more complex cognitive performances that have been reported in the psychological literature over the past 50 years or more.