Automatic-Spreading Activation Effects following Children's Reading of Complete Sentences

Abstract

Third-, sixth-, and college-grade level students participated in an on-line reading task that incorporated both a naming latency and a subsequent cued-recall memory requirement. Following reading of a complete sentence, latency for naming a target word was measured in which the target word was either (a) a repetition of a word in the previous sentence, (b) an associate of a word in the previous sentence, (c) inferable from the integrated meaning of words in the previous sentence, or (d) unrelated to words in the previous sentence. Increased speed was found for naming words presented in the repeated, associated, and inferred target conditions as compared to the unrelated word condition. In each case, the observed facilitation effect was of greater magnitude for the younger readers. In the cued-recall task, single-word cues resulted in better recall memory performance when the cue had been explicitly presented in a prior sentence as compared to cues which were only inferable from previously read material. Results were interpreted in terms of context effects which extend beyond sentence completion boundaries and in terms of developmental differences in automatic expectancy and semantic integration effects.