Electronic Thesis and Dissertation Repository


Master of Arts




Dr. Perry Klein


Handwriting and spelling present elementary students with significant sources of load on working memory as the various writing processes compete for cognitive resources (Kellogg, Whiteford, Turner, Cahill & Mertens, 2013). Several studies have shown that speech-to-text (STT) software can improve students' writing on a specific text (Higgins & Raskind, 1997; MacArthur & Cavalier, 2004; Quinlan, 2004); however, the question of whether STT can be used to teach writing strategies has been neglected. This pretest-post-test between groups study experimentally tested the effects of composition modality on learning a persuasive writing strategy. First, all students (N=45) completed a pretest of persuasive writing. They then, received instruction in Dragon NaturallySpeaking (version 11). Next, students were randomly assigned to participate in four lessons that emphasized dialectical elements of persuasive writing, in one of two modalities: STT or handwriting. Finally, all students completed post-tests of persuasive writing in both modalities (STT and handwriting). Writing samples were evaluated for word count, number of types of rhetorical moves, surface errors, and word errors. Students also completed measures of cognitive load for the pretest, each writing activity, and post-test. Both training conditions resulted in large, statistically significant, pre-to-post-test gains on word count, holistic quality and rhetorical moves. Students in the STT condition reported more effort compared to students in the handwriting condition for both post-tests. Students in both instructional conditions showed a high level of transfer from the trained modality to the untrained modality. Students who learned through handwriting, compared to students who learned through STT, showed more surface errors on the STT post-test. The results suggest that STT could be an equally effective alternative for teaching composition strategies.