Electronic Thesis and Dissertation Repository


Doctor of Philosophy




Dr. Stephen Lupker


The goal of the present research was to examine the nature of masked priming with an emphasis on the influence of stimulus-response (S-R) associations. In Chapter 2, both the magnitude of the category congruence (priming) effect and the nature of the priming distance effect were assessed in two number classification tasks. Participants made either magnitude (i.e., is the target larger or smaller than ‘5’?) or identification judgments (i.e., press one button if the target is a ‘1’, ‘2’, ‘3’ or ‘4’ or the other button if the target is a ‘6’, ‘7’, ‘8’ or ‘9’). Priming distance effects in both tasks indicated that, regardless of task instructions, semantic activation played a key role in producing masked priming even when the identification task showed evidence of priming from S-R associations. In Chapter 3, category learning tasks were used involving stimuli having neither the ability to activate semantic information nor any a priori response associations. Participants learned, through feedback, artificially-defined category sets. Each category contained a prototype item that served as a masked prime. In the single-session tasks, responses were faster when a stimulus was primed by its own prototype versus the prototype of the other category, however, this (priming) effect only increased in size when participants performed the task over multiple sessions, indicating that it takes considerable time for the S-R associations to develop and impact priming. The research in Chapters 2 and 3 was based on a prospective view of masked priming. There is, however, an alternative retrospective view supported by numerous demonstrations of prime validity effects (i.e., larger priming effects when the proportion of congruent trials is high). In Chapter 4, an arrow classification task with free choice stimuli was used to examine this debate. The prime-target relationship for the arrow targets was either: a) always congruent, b) always incongruent, or c) unpredictive. Prime validity effects for the free choice trials (i.e., trials involving “either way” stimuli, e.g., < >, for which either response was acceptable) occurred using both 77 ms and 165 ms prime-target intervals. The results in the unpredictive conditions support a prospective view of masked priming since they indicate that it was response bias suppression when the proportion of incongruent trials was high that produced the prime validity effects.