A chronometric analysis of strategy preparation in choice reactions

Abstract

In this research, we investigated the processof preparing strategies for performing choice-reaction tasks. Before each choice-reaction trial, subjects were shown a cue that indicated features of the stimulus-response mapping to be used on the upcoming trial. Subjects used this cue to specify their strategy for responding to the stimulus. The time needed for specifying the strategy was measured by allowing subjects to control the cue presentation and surreptitiously recording how long they spent looking at the cue. The experiments demonstrated that the time to prepare a strategy was a function of the number and nature of the strategy features that had to be specified; simple uncertainty about the possible strategies had little direct effect. The results disconfirmed a serial model in which the time to prepare a strategy is the sum of the times to specify the individual strategy features. A mixed serial-parallel model was proposed as an alternative.In the present article, we investigate the process by which people prepare strategies for performing simple speeded tasks. The term strategy is used to refer to the internalrepresentation of the perceptual, cognitive, and motor operations needed to perform an information-processing task. The representation specifies the kind of perceptual analysis needed, the way in which the perceptual information is to betransformed, the decision rules to beused, and the responses to begenerated (cf. Logan & Zbrodoff, 1982;Logan, Zbrodoff, & Fostey, 1983). The present notion of a strategy is similar to the concept of a motor program in that both specify the response to be made (see Keele & Summers, 1976;Rosenbaum, 1980). However, a strategycontains information aboutthe association of each of the responses with the various possible stimuli, in addition to information about how each responseshouldbe performed. In the experiments described here, we focused on how decision rules and responses are specified in strategies for choice reactions.In this investigation, we used a strategy cuing paradigm, as developed by Dixon (1979Dixon ( , 1981. The paradigm is similar in some ways to the movement precuing technique de-The authors wish to thank Patricia Carpenter, David Klahr, David Rosenbaum, Jeff Bisanz, Renee Elio, Gordon Logan, and Robert Proctor for comments on previous drafts. This research was supported by the United States National Institute of Mental Health (Grant No. MH-29617 to the second author) and by the Natural Science and Engineering Research Council of Canada (Grant No. A8263 to the first author). Experiments 2 and 3 were reported previously at the 1978 Psychonomic Society Meeting in San Antonio, and in the first author's doctoral dissertation.Requests for reprints should be sent to Peter Dixon, Department of Psychology, University of Alberta, Edmonton, Alberta, Canada, T6G 2E9.vised by Rosenbaum (1980), and to the strategy construction task used by Logan and his colleagues (Logan & Zbrodoff, 1982;Logan et al., 1983). The technique also has a historical antecedent in the partia...

On measuring selective attention to an expected sensory modality

“…More importantly for our purposes, it is reduced (although not completely eliminated) when subjects are informed in advance that a task switch is about to occur. Rogers and Monsell (1995) have suggested that this reduction with advance information is caused by the endogenous reconfiguration of task set (see also Dixon & Just, 1986). Since task-preparation effects can arise even when all possible tasks in the experimental situation arise within the same sensory modality, it could be suggested that our present cuing effects have absolutely nothing to do with preparation for one modality versus another, but simply reflect preparation for one task versus another.…”

“…However, the subjects may have used the cue simply in order to reconfigure their task set (cf. Allport et aI., 1994;Dixon & Just, 1986;Rogers & Monsell, 1995) rather than to direct their attention to one or another sensory modality per se. This seemed particularly likely because very different stimulus-response mappings were used for the auditory versus visual tasks.…”

On the representation of task information in task switching: Evidence from task and dimension switching

“…This corresponds to the contrast between be performed on the number of digits. Note that the cue information was not only specific with regard to the task goal, but also indicated the specific response rules that were required on the trial; that is, the left-right order of the cue information was consistent with the order of the response mappings (see, e.g., Dixon & Just, 1986;Logan & Bundesen, 2003;Sudevan & Taylor, 1987). Since there were two stimulus dimensions and two tasks, there were 4 such compound cues.…”

Mixing costs in task shifting reflect sequential processing stages in a multicomponent task

“…Therefore, two stimulus categories had to be taken into account: the level (global/local) by which a numeral had to be selected and the judgment categories (odd/even, less/greater) by which a numeral had to be translated into a response. Both task components were announced by a cue, and the cue-stimulus interval was self-paced-that is, on each trial, the subjects could prepare for the upcoming task as long as they wanted (Dixon, 1981;Dixon & Just, 1986). Since the levels and judgments were either mixed or held constant, four mixing conditions could be realized.…”