A quantitative study was conducted to determine if the amount of formulaic language influenced ESL teachers' perceptions (n=102) of non-native writing skill, as evidenced by composite and sub-scale scores on the ESL Composition Profile (Jacobs et al., 1981). Formulaic language was operationalized as 25 three-word strings sampled from the writing sub-list of the Academic Formulas List (Simpson-Vlach & Ellis, 2010) and further validated as frequent in the Michigan Corpus of Upper Level Student Papers. The target formulaic sequences were divided into three experimental groups representing a low, mid, and high amount of formulaic language. Four advanced non-native writers generated argumentative, timed writing samples that incorporated the target sequences. The writing samples were then assembled into data collection packets and distributed at eight Intensive English Programs across the southeastern United States. A repeated measures ANOVA indicated that there was a significant difference in composite score (p < .05) between the control and three experimental conditions; however, the essays that incorporated 16 and 25 formulaic sequences scored significantly lower than those with zero or eight target sequences. When the amount of syntactical and semantic errors were strictly controlled for, the composite scores also fell between the control and experimental conditions, but the decrease in score was not significant (p > .05). The content, organization, vocabulary, language, and mechanics sub-scales were also compared using a repeated measures MANOVA. In content, organization, and language, the control and low essays outscored the mid and high conditions (p < .05). For the vocabulary sub-scale, the control and low condition were not significantly different, but the control essays only outperformed the mid level essays. The low essays outperformed both the mid and high essays. In terms of mechanics, there was only a significant difference between the low and mid level essays. The results of the MANOVA were consistent when the amount of syntactic and semantic errors were controlled. Implications for teaching suggest that the Academic Formulas List would not benefit academically-oriented L2 learners preparing to enter a university. While corpus tools are valuable in helping teachers, material writers, and publishers improve vocabulary instruction in the English classroom, not all statistically salient lexical combinations are important for non-native writers to master and incorporate in their academic discourse.
Identifer | oai:union.ndltd.org:ucf.edu/oai:stars.library.ucf.edu:etd-5725 |
Date | 01 January 2014 |
Creators | Youngblood, Alison |
Publisher | STARS |
Source Sets | University of Central Florida |
Language | English |
Detected Language | English |
Type | text |
Format | application/pdf |
Source | Electronic Theses and Dissertations |
Page generated in 0.0015 seconds