Target-noise similarity in the visual field: Masking effects as a function of target-noise similarity

Ronald Jackson Evans, Purdue University

Abstract

The purpose of the present research was to test the hypothesis of feature-specific inhibition in visual displays. This hypothesis was tested through the use of two tachistoscopic identification experiments where displays consisted of a target stimulus, to be identified, flanked by either one or two noise characters. In Experiment I, stimuli consisted of the presence or absence of each of three separate line features. Using this design, the effect of the presence of any line feature in the flanker on the perception of any line feature in the target could be measured. Models employed in the analysis of this data included Luce's (1963) Similarity-Choice Model and variations, Signal Detection Theory, and Wolford's Feature Perturbation Model. In virtually all instances, the analyses rejected the notion of feature-specific inhibition. The data were more consistent with models such as Treisman's (1977) Feature Integration Theory, Wolford's (1975) Feature Perturbation Theory, and Eriksen's (1979) Continuous Flow Model which all pose some type of process including confusion of the flanker with target information. Experiment II was designed in a manner similar to Santee and Egeth (1982) with the purpose of generating results consistent with a feature-specific inhibition effect. Displays consisted of tachistoscopic presentations of two letters and included pre and post stimulus masks. After offset of the two letters, an arrow designated which of the letters to report. Results from Experiment II did not support feature-specific inhibition. An analysis of variance and posthoc tests showed that the four stimulus letters (D, O, M, and W) differed in their ability to mask one another and that the letters were not equally discriminable. Although previous studies relied upon decreased accuracy when repeated letters were presented in a display (e.g., DD, OO, MM, or WW), only one of the four letters (MM) showed this effect in Experiment II. Results from the analyses were indicative of effects related to the letters and their well-learned relationship (e.g., letters in the display DO being more perceptible). Overall, then, the attempt to replicate feature-specific inhibition in Experiment II also failed.

Degree

Ph.D.

Advisors

Townsend, Purdue University.

Subject Area

Psychology|Experiments

COinS