Games and Phone Numbers: Do Short Term Memory Bounds Affect Strategy Behavior?

Research in experimental and behavioral game theory has revealed a substantial and persistent degree of heterogeneity in the strategic behavior of real individuals. While the prevailing theoretical explanations of the observed heterogeneity typically invoke underlying differences in beliefs among the population of players, we argue that a further source of heterogeneity may consist in the individuals' different ability to process information, of which short term memory capacity provides a measurable proxy. Research in cognitive psychology has shown that individuals typically differ in their short term memory capacity; furthermore, short term memory capacity provides a fundamental cognitive bottleneck to our ability to process information efficiently and hence seems correlated with performance in a variety of problem solving and reasoning tasks. In this paper we conduct experiments on a set of well-known games whose solution concepts require the application of some paradigmatic forms of strategic reasoning, such as iterated dominance, reasoning about common knowledge and backward induction. We separately conduct standard short term memory tests on our subjects to detect the presence of a correlation between individuals' behavior in the games - here defined in terms of degrees of conformity to the standard game-theoretic prescriptions - and their short term memory score. Our results show the presence of a significant and positive correlation between subjects' short term memory score and conformity to standard game-theoretic prescriptions in the games, thus confirming our hypothesis. While the robustness of our conjecture awaits to be confirmed by further data gathering in more interactive experimental settings, our preliminary results suggest a promising line of inquiry on the interconnections between information processing capacity and strategic behavior