Random number generation and working memory

Abstract

Abstract We demonstrate the close relationship that exists between random sequence generation and working memory functioning. We clarify the nature of this link by examining the impact of concurrent requirements for random sequence response quality. Experiments 1A and 1B show that marking specific response choices for differential treatment, either by requiring an ancillary behaviour or by suppressing these choices from output, impairs overall sequence quality. Contrasting with previous findings, these distinct concurrent tasks have comparable effects. We show that disruption is found only when concurrent demand is high. Experiment 2 demonstrates that increasing the dynamic working memory load by requiring the ancillary response to change during the task leads to additional disruption of randomisation. The results extend and refine our understanding of the contribution of active maintenance of representations in random generation. Acknowledgments This paper is dedicated to the memory of Derek Neil (1968–2001), who contributed software to the analysis of random sequences. More information about automated analysis of sequences can be found at http://www.lancs.ac.uk/staff/towse/rgcpage.html. We are grateful for comments on this paper from André Vandierendonck, Iring Koch, and an anonymous reviewer. Notes 1In cases where ANOVA sphericity assumptions were violated, Greenhouse-Geisser corrections were applied to the degrees of freedom, and these values are reported in the text.