Journal of Experimental Psychology: General. 140(4):622–636, NOVEMBER 2011

DOI: 10.1037/a0024230

,

PMID: 21707205

Issn Print: 0096-3445

Publication Date: November 2011

# Dynamic Adaptation to History of Trial Difficulty Explains the Effect of Congruency Proportion on Masked Priming

Sachiko Kinoshita;Michael Mozer;Kenneth Forster;

+ Author Information

1Macquarie Centre for Cognitive Science and Department of Psychology, Macquarie University, Sydney, Australia2Department of Computer Science and Institute of Cognitive Science, University of Colorado3Department of Psychology, University of Arizona

### Abstract

In reaction time research, there has been an increasing appreciation that response-initiation processes are sensitive to recent experience and, in particular, the difficulty of previous trials. From this perspective, the authors propose an explanation for a perplexing property of masked priming: Although primes are not consciously identified, facilitation of target processing by a related prime is magnified in a block containing a high proportion of related primes and a low proportion of unrelated primes relative to a block containing the opposite mix (Bodner & Masson, 2001). In the present study, this phenomenon is explored with a parity (even/odd) decision task in which a prime (e.g., 2) precedes a target that can be either congruent (e.g., 4) or incongruent (e.g., 3). It is shown that the effect of congruence proportion with masked primes cannot be explained in terms of the blockwise prime–target contingency. Specifically, with masked primes, there is no congruency disadvantage in a block containing a high proportion of incongruent primes, but there is a congruency advantage when the block contains an equal proportion of congruent and incongruent primes. In qualitative contrast, visible primes are sensitive to the blockwise prime–target contingency. The authors explain the relatedness proportion effect found with masked primes in terms of a model according to which response-initiation processes adapt to the statistical structure of the environment, specifically the difficulty of recent trials. This account is supported with an analysis at the level of individual trials using the linear mixed effects model.