Processing of geometrical dimensions in a binary classification task: Evidence for a dual process model

Abstract

Confirming the findings in search tasks with letters and digits, the typical RT same < RTdiff result was obtained in a matching paradigm requiring the classification of geometrical stimuli that were given in pairs. The study supports a dual process model that is based on an identity reporter for the faster "same" response and a difference detector for the slower "different" responses, both operating with equal accuracy. Subjects appeared to perceive outline aspects of figures, formed by size and form, holistically. An internal characteristic, such as an interior line, was apparently processed as a separate attribute. However, the outlines of the stimulus configurations appeared to be much more salient and interfered with the judgment of the orientation of the interior line. Moreover, the latter stimulus aspect could be easily ignored as the task required.

Holistic vs. attribute repetition effects in classifying stimuli

“…We now know that size of circle and tilt of diameter are nonintegral dimensions (Garner & Felfoldy, 1970), as are color and shape (Handel & Imai, 1972). Accordingly, Biederman's stimuli were at least partly nonintegral and thus might be expected to be processed an attribute (or object) at a time, just as were the attributes of Higgins and Getty's (1977) faces and some attributes of Keuss' (1977) geometric patterns. Consider, as an example, the first defined orthogonal task (called filtering) in Table 1 in Biederman's (1972) extensive classification study.…”

“…In some studies responses are speeded if all attributes of a stimulus repeat and do so in the same configuration (identity), but responses are not speeded if some characteristic of the stimulus, an attribute or the configuration, is changed on successive trials (nonidentity) (Felfody, 1974;Keuss, 1977; the present paper). In other studies response times are facilitated both when there is identity repetition and when only some attributes repeat.…”

“…In other studies response times are facilitated both when there is identity repetition and when only some attributes repeat. The latter results have been labeled as partial repetition effects (PREs) and have been reported in classification tasks (Biederman, 1972, Note 2), identification tasks (Higgins & Gettys, 1977), attribute detection tasks (Keuss, 1977), when the probability of occurrence of a subject-predicted attribute is manipulated (Geller, 1974), and when stimulus-response compatibility is different on successive trials (Shaffer, 1965). Thus, some studies report that responses are fast when the entire object repeats, but responses are not necessarily facilitated (depending on similarity) when an attribute repeats; other studies report that responses are speeded as long as relevant attributes repeat.…”

“…Other two-stage models, such that of Bamber (I969), make a similar prediction in this instance (cf. Keuss, 1977).…”

Storage mode of geometrical dimensions in a two-stimulus matching paradigm

“…Any repetition of Keuss' (1977) study, it is suggested, should include both sequential and simultaneous stimulus presentation modes. (All other procedural details of the present investigation are identical to Keuss, 1977. For a rationale, a reference is made to that study.) The main assumption of the present study is that the use of a retention interval following the first, or memory, stimulus may facilitate subject's analytical approach.…”