On the nature of the proportion effect in semantic priming

The purpose of the present study was to investigate the nature of the proportion effect in semantic priming. The experiment employed a lexical decision task under conditions of low, medium, and high proportion of related pairs with an SOA of 550 msec. The results indicated that the higher proportions of related trials had a facilitatory effect on targets preceded by related primes as compared with a neutral priming condition. In comparing the unrelated trials with the neutral trials there is an increase in the inhibitory effect associated with unrelated pairs as the proportion of related pairs increases. The proportion of related items did not affect the nonword responses. In addition, practice was not found to have a significant effect in any of the conditions. It was concluded that the proportion effect could best be explained by a post-lexical coherence checking mechanism within the constraints of the present experiment.