Semantic and spatial components of selective attention.

Semantic and spatial effects on selective processing were examined. A prime presented 250 ms or 1,000 ms before a pair of masked words was related to 1 of the words on 1/3 of the trials. Although Experiments 1 and 2 required report of both words, processing was selective; typically, just 1 word could be reported. In Experiment 1, reported words were more often top words, showing spatial selectivity, and more often words related to the prime, showing semantic selectivity. Experiment 2 included to-be-ignored peripheral cues 50 ms before the pair. Related words were reported more often at both delays, and cuing produced an additive benefit at 1,000 ms. Experiment 3 required report of just the cued word and included 100-ms cues. Cued words were reported more often, especially at 1,000 ms and with 100-ms cues, but semantic selectivity also occurred. Selectivity via semantic priming and via location reflect separate attentional mechanisms.