Self-terminating versus exhaustive processes in rapid visual and memory search: An evaluative review

Abstract

A major issue in elementary cognition and information processing has been whether rapid search of short-term memory or a visual display can terminate when a predesignated target is found or whether it must proceed until all items are examined. This study summarizes past and recent theoretical results on the ability of self-terminating and exhaustive models to predict differences in slopes between positive (target-present) and negative (target-absent) set-size functions, as well as position effects. The empirical literature is reviewed with regard to the presence of slope differences and position effects. Theoretical investigations demonstrate that self-terminating models can readily predict the results often associated with exhaustive processing, but a very broad class of exhaustive models is incapable of predicting position effects and slope differences typically associated with self-termination. Because position effects and slope differences are found throughout the rapid search literature, we conclude that the exhaustive processing hypothesis is not tenable under common experimental conditions.Since the earliest work in experimental psychology, researchers have been attempting to uncover the architecture of various simple mental processes through the measurement of reaction time (RT; Cattell, 1886Cattell, /1947 Donders, 1868Donders, /1969Wundt, 1894). In RT tasks, an observer is presented with a stimulus, and the time that elapses between its onset and the execution of a response is recorded. The changes in RT under different modes of stimulation or different response requirements are assumed to provide information about the component mental processes and how these components are linked together. Luce (1986), Posner (1978), Townsend and Ashby (1983), and Welford (1980) provide reviews and somewhat differing perspectives on empirical and theoretical research in which RT has been used to study mental processes. This paper focuses on two simple paradigms used extensively to study simple mental processes: rapid visual and memory search. In the search task, an observer is We wish to express our appreciation to one anonymous reviewer, as well as to C. W. Eriksen, W. K. Estes, R. W. Proctor, R. Schweickert, and ]. Theios for helpful comments on an earlier draft of this paper. This research was supported by NSF Grant BNS-831-9377, awarded to]. T.T. Correspondence should be addressed to]. T. Townsend, Department of Psychology, Indiana University, Bloomington, IN 47405 (e-mail: jtownsen@ucs.indiana.edu). 563presented with a list of items, called the "search set." The search set is usually composed of letters or numbers, but pictures, words, and tones have also been used. The observer is required to indicate whether a "target" item is present in the search set by executing a positive ("yes") or negative ("no") response. In memory search tasks, the search set is presented to the observer prior to the target item. When the target is presented, the observer must "search" the items held in memory to decide whether or not t...

Redundancy gain for semantic features

“…Although we have favored a parallel race model to account for the present results, it should be noted that the data are also consistent with a random-order serial-decision model (see van Zandt & Townsend, 1993). According to this model, participants perform a serial self-terminating search and check for the color target and the category target in a random order.…”

“…For instance, it has been argued that searching for a visual target is guided by the same mechanism that underlies memory search (Gilford & Juola, 1976). Yet, search slope ratios obtained in visual, as compared with memory, search often deviate from each other, suggesting that search mechanisms differ between external stimuli and internal memory representations (for a review, see Van Zandt & Townsend, 1993).…”

Near and far space: Understanding the neural mechanisms of spatial attention

“…Target absent trials involve an added cognitive and neurological component related to the decision to terminate the search (Van Zandt & Townsend, 1993). Therefore, in order to examine the involvement of our regions of interest in the response to targets, uncomplicated by extra substrates (see Ashbridge et al, 1997), only target present responses were analyzed.…”

The influence of irrelevant location information on performance: A review of the Simon and spatial Stroop effects

“…This issue has been investigated primarily by means ofvisual search tasks in which a target stimulus in an array of distractors must be detected or identified. Reaction time (RT) in such tasks is typically an increasing function ofarray size, and accuracy a decreasing function, when the target is defined by a conjunction of features (e.g., Treisman & Gelade, 1980;Van Zandt & Townsend, 1993), leading some authors to propose that attending to location is necessary for feature integration (e.g., Nissen, 1985;Treisman & Gelade, 1980). As expected if location must be attended in such situations, a stimulus for a second task shows a processing benefit when it occurs in a location adjacent to the target stimulus of the search task rather than in a more remote location (e.g., Hoffman & Nelson, 1981), and precuing the location in the array in which the target stimulus will occur facilitates its processing (e.g., Eriksen & Hoffman, 1972).…”