Perceptual recognition as a function of meaningfulness of stimulus material.

TLDR

The present study evaluates a class of models of human information processing made popular by Broadbent. A brief tachistoscopic display of one or two single letters, four‑letter common words, or four‑letter nonwords was immediately followed by a masking field and two single‑letter response alternatives chosen to minimize informational differences among the tasks. The study found that presenting response alternatives 5 s before or after the stimulus impaired performance, that word recognition outperformed single‑letter recognition, and that early processing occurs in parallel while subsequent scanning is serial.

Abstract

The present study evaluates a class of models of human information processing made popular by Broadbent. A brief tachistoscopic display of one or two single letters, four-letter common words, or four-letter nonwords was immediately followed by a masking field along with two single-letter response alternatives chosen so as to minimize informational differences among the tasks. Giving 5s response alternatives before the stimulus display as well as after it caused an impairment of performance. Performance on single words was clearly better than performance on single letters. The data suggest that the first stages of information processing are done in parallel, but scanning of the resultant highly processed information is done serially.