Date of Award

2016

Degree Type

Thesis

Degree Name

Master of Arts in Psychology

Department

Psychology

First Advisor

Smith, Albert F.

Subject Headings

Psychology

Abstract

Many models of word identification suppose a hierarchical system in which units at increasing levels respond to features, letters, letter combinations, and words. Some models suppose units responsive to bigrams—letter pairs—that may not be adjacent in a letter-string stimulus. In such a model, a stimulus such as BIRD would activate, at the bigram level, bigrams representing adjacent letters BI, IR, and RD, and also bigrams representing nonadjacent letters BR and ID. Grainger, Mathot, and Vitu (2014) reported an experiment in which strings to be classified as words or pseudowords were flanked by bigrams from the target string or not; for flanking bigrams consisting of target-string letters, the order of the bigrams was as in the target string or switched, and the order of letters within the bigrams was as in the target string or switched. For example, BIRD could appear with these flankers: BI BIRD RD; RD BIRD BI; IB BIRD DR; DR BIRD IB; CE BIRD NT. Grainger et al. (2014) found, for words, better performance when flanking bigrams contained target-string letters (e.g., BI BIRD RD; RD BIRD BI; IB BIRD DR; DR BIRD IB) than when they did not (e.g., CE BIRD NT); and better performance when flanking bigrams contained letters ordered as in the target (e.g., BI BIRD RD; RD BIRD BI) than switched (e.g., IB BIRD DR; DR BIRD IB); but whether flanking bigrams were ordered as in the target did not affect performance. We investigated whether flanking open bigrams facilitate lexical decisions. Experiment 1 investigated performance in the conditions from Grainger et al. (2014). The results of Experiment 1 essentially replicated those of Grainger et al. (2014). Experiment 2 included four additional conditions in which the flanking bigrams consist of letters separated by one letter in the target (e.g., BR BIRD ID; ID BIRD BR; RB BIRD DI; DI BIRD RB). Importantly, results of Experiment 2 indicate that performance is better when flankers contain letters that are ordered as they are in the target, and this letter order effect does not depend on whether the flankers are adjacent-letter bigrams, or open bigrams.

Download

Included in

Psychology Commons

COinS