Same—differentjudgments of multiletter strings: Insensitivity to positional bias and spacing

Abstract

Several recent studies of multiletter matching have included pairs of strings that have the same letters in different positions (rearranged pairs). The task can be defined such that these rearranged pairs are correctly classified as different (i.e., subjects respond "same" only if the strings have the same letters in the same positions-the order task) or as same (i.e., subjects respond "same" if the strings have the same letters regardless of their positions-the item task). The order task produces left-to-right serial-position effects, whereas the item task produces U-shaped serialposition effects. Because these differences suggest that subjects may be able to exert strategic control over the comparison process, two sets of experiments were designed to test whether or not subjects can change the relative weightings devoted to the respective serial positions. In Experiments 1 and 2, the probability that a mismatch occurred in the different positions was manipulated. In Experiments 3 and 4, the physical spacing between letters, as well as whether or not the spaces were filled with neutral noise characters, was varied. None of the manipulations had much influence on the serial-position effects. Thus, the distinct serial-position effects for the order and item tasks apparently are mandatory and not due to any voluntary comparison strategy.Numerous studies have investigated same-different matching for pairs of multiletter strings. In many of the models proposed to account for a variety of findings from these studies, it has been presumed that accumulation of pooled perceptual information feeds into a response process (Krueger, 1978; Ratcliff, 1981;Vickers, 1979). The response process includes criteria that determine the amount of similarity or difference to which the information must build before one of the two responses is selected. Consistent with such accumulation models, subjects can trade speed for accuracy when a deadline is imposed (Krueger & Chignell, 1985), suggesting a drop in the criterial amount of information required to select a response. Moreover, subjects can adjust the relative accuracies and latencies for the same and different responses

Letter-position coding in random consonant arrays

“…These two extensions were incorporated to allow the model to capture the variation in the size of position-independent priming effects as a function of the degree of displacement of the critical letters in the prime and target strings of the present experiments. Since similar effects are also obtained in the multi letter matching task (Angiolillo-Bent & Rips, 1982;Proctor & Healy, 1985, 1987Proctor et al, 1991;Ratcliff, 1981), the proposed extension of the PSLD hypothesis could usefully be invoked to explain these results. One could hypothesize that subjects are matching patterns of activity in either the PSLD or the PILD system, depending on whether or not they are required to take order information into account.…”

“…Moreover, since degree of displacement directly affects PSLD activity and only indirectly affects PILD activity, we are led to correctly predict that the effects of this factor are larger in the position-relevant task than in the position-irrelevant task (Proctor et al, 1991). Also, since PSLD units are hypothesized to be activated before PILD units, this interpretation also correctly predicts that different responses to rearranged pairs in the orderrelevant task are faster and more accurate than same responses to the same stimuli in the order-irrelevant task.…”

“…Depending on whether subjects are instructed to respond on the basis of both identity and position information or identity information alone, the correct response to the second category of stimuli is either same or different. The main finding with respect to this particular class of stimuli is that reaction time (RT) and accuracy vary systematically as a function of the degree of displacement of the letters in each string (Angiolillo-Bent & Rips, 1982;Proctor & Healy, 1985, 1987Proctor, Healy, & Van Zandt, 1991;Ratcliff, 1981). Same responses are slower and less accurate as degree of displacement increases, while different responses follow a mirror-image pattern.…”

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

“…More recently, Proctor, Healy, and Van Zandt (1991) clarifIed these fIndings. They conducted a series of experiments with the order-relevant and order-irrelevant matching tasks, in which the probability of the location of a mismatching item varied across display positions, and the spacing between the display items was varied.…”

Task-Invariant Aspects of Goodness in Perceptual Representation

“…Here, the task requires a different response to pairs that are members of the same ES. We may expect that these conditions will become particularly difficult to respond to (Proctor & Healy, 1985, 1987Proctor, Healy, & Van Zandt, 1991). Another effect that we might expect in a physical matching task is that of similarity between individual items.…”