Priming for Novel Associations: Evidence for an Attentional Component

Two experiments investigated mechanisms underlying the priming effect for novel associations in stem completion, by manipulating attention at test. Experiment 1 demonstrated that dividing attention at test did not reduce overall priming, but abolished the novel association priming effect. Experiment 2 required subjects to report all completions generated for a stem, and showed that reinstating context enhanced the target output position, and that dividing attention eliminated this effect of context as well as reducing total output. These results were taken to argue that when stems allow multiple solutions, the reinstatement of context serves to promote the sampling of the word with which it was paired at study in preference to other words activated by the stem, and that this effect of context demands attentional resources. This interpretation also provides an explanation for the discrepancy in the novel association priming effects observed with amnesic subjects.