Random generation in the Working Memory dual-task paradigm
Much research into the Working Memory model (Baddeley & Hitch, 1974) has involved secondary task experiments, where secondary tasks of known requirements are performed concurrently with primary tasks of interest. The thesis has explored the use of Random Generation, a novel secondary task, in this paradigm. The task requires subjects to generate a string of items from a given response set, by calling out the items in as random a fashion as possible. Random generation tasks are held to make heavy demands on the Central Executive component of Working Memory (Baddeley, 1986 & 1990), and would seem to reflect the involvement of this component in dual-task studies. The work has addressed the experimental design, administration, and analysis of performance when random generation is used in secondary task experiments. Standard procedures were developed and used throughout, so that the inter-comparison of experiments was possible. Performance was measured by calculating redundancy indices (Hsingle, Hpairs, H2, and Evans' RNG) from one-hundred responses. Random generation from a set of ten numbers (1 to 10) was chosen for further exploration and development. Random number generation proved to be sensitive to primary task loading in a variety of situations. The stability of some aspects of performance between experiments may be questionable, but the overall picture is encouraging. However, there is a clear need for basic research to support more detailed cognitive modelling of the task, before it can be used with confidence in the Working Memory dual-task paradigm.