Memory overload or expectancy effect? ‘Hysteresis’ revisited*

Abstract

The effects of expectancy and variation in task demand on the rate of human information transmission were studied. Subjects performed an eight-choice key-pressing task, attempting to match characters presented singly under both increasing and decreasing demand. The results support previous research indicating that the relationship between the rate of information an individual is able to transmit and task demand depends, at least in part, upon the temporal history of demand. When a relatively high level of demand was imposed, performance failed to recover at the expected rate as demand was reduced. However, this 'hysteresis' effect occurred even when a cue was provided to indicate clearly that a reduction in task demand was imminent, suggesting that an overload of short-term memory, rather than an individual's erroneous expectations regarding demand, is primarily responsible for the effect.

Effect of Workload History on Task Performance

“…L. Matthews (1986) conducted a set of studies to further explore the effects of workload history on visual task performance. His results replicated the findings of Cumming and Croft (1973) and Goldberg and Stewart (1980): Higher decrements in performance result when workload decreases as compared with when workload increases. In addition, M. L. Matthews (1986) tested the theory that the performance decrement was attributable to a failure of STM.…”

“…Even when participants were given this cue, performance was still worse when moving from high to lower workload as compared with moving from low to higher workload, suggesting that it is not a person's inaccurate expectations that are responsible for the effect. Rather, Goldberg and Stewart (1980) suggested post hoc that the decrements following a decrease in work demand resulted from a temporary overload of short-term memory (STM). In other words, they proposed that the rapid presentation of information (at the higher workload level) overloaded STM, resulting in a performance decline.…”

“…As a result, Cumming and Croft (1973) suggested post hoc that this inability to match expectancies regarding changes in task difficulty explained why their participants showed a much larger performance decrement when they moved to a low workload level after performing at a high workload level, compared with when they experienced an increase in task difficulty following a low workload level. Goldberg and Stewart (1980) reported a study designed to test whether expectancies were responsible for the decline in task performance when workload moved from high to lower levels. They employed a visual performance task.…”

“…The studies of Cumming and Croft (1973) and Goldberg and Stewart (1980) neglected to provide participants with training sessions. This is of special concern, given that a minimum level of training is essential to bring participants to a reasonably stable performance level and consequently minimize contamination of test data from learning effects (Schlegel & Gilliland, 1990).…”

“…Previous research has suggested that the type and duration of work performed prior to a given point in time (i.e., workload history) may have a strong influence on work performance following that point (Cumming & Croft, 1973;Goldberg & Stewart, 1980;M. L. Matthews, 1986).…”