Working memory and strategy use during mental comparison tasks
This thesis describes a series of eight experiments that examined working memory involvement and strategy use during mental comparison tasks. The main conclusions drawn from the results of the eight experiments are as follows: Firstly, the evidence suggests that the spatial 'inner scribe' component of working memory (Logie, 1995) is involved during mental comparisons based on relative size. This conclusion is based on the finding that concurrent spatial tapping selectively interferes with subjects' performance on mentally comparing the angles between the hands of imaged analogue clocks (Experiment 3), the comparison of the size of a shaded section of a circle with one presented previously (Experiment 4), and the comparison of objects' relative sizes based on information stored in long-term memory (Experiments 7 and 8). It is argued in the general discussion chapter of this thesis that the functions attributed to the 'inner scribe' working memory system are comparable with the collection of spatial subsystems proposed by Kosslyn in his computational model of imagery (Kosslyn, 1980; 1991; 1994). Secondly, the results of this thesis suggest that during mental comparison tasks the majority of subjects combine the generation of visual images with additional verbal strategies. This finding is in contrast to previous literature (e.g., Paivio, 1975; 1978b), which has tended to assume that subjects' performance of mental comparisons is dependent purely on the utilisation of a visual imagery strategy. Post-task strategy questionnaires administered in Experiments 4 to 8 revealed that consistently around 25% of subjects did not report being aware of relying to any great extent on visual images while making the mental comparisons. In Experiments 2 and 3 it was also demonstrated that the adoption by subjects of a combined imagery/verbal strategy was less effective for comparing mental clocks than the utilisation of a pure imagery strategy, and moreover that reliance on verbal strategies during the mental clocks task could be disrupted by concurrent relevant speech or oral random generation. Conversely, in Experiment 4 it was demonstrated that when a variant of the clock task termed the 'wedge task' was used as a primary task, the adoption of a pure verbal strategy corresponded to the absence of any significant effect of angular separation on response times, in contrast to those subjects who did report using visual images as a basis for making the comparison. In addition, the utilisation of a pure verbal strategy was shown to be more effective than an imagery strategy for the smaller wedge comparisons, with the opposite pattern of results occurring with larger wedge comparisons. Finally, in the general discussion chapter of this thesis an attempt is made to clarify the relationship between the subsystems of working memory and the generation and manipulation of visual images. It is argued that the computational model of imagery developed by Kosslyn and the current two-component model of visuo-spatial working memory are not directly comparable. Instead, the subsystems of VSWM are deemed to be functionally separate from the visual buffer, though they are assigned a vital role during cognitive operations such as image transformations, and during the transference of information from long-term memory into the visual buffer. Finally, an attempt is made to integrate aspects of both working memory and imagery theories, particularly in the areas of visual rehearsal and the maintenance of images within the visual buffer.