| This research project consisted of an 8-week, 11-hour series of poetry workshops developed, implemented, and evaluated by the researcher at a public charter school of the arts in Southern California. The study used a quasi-experimental design and mixed methods of data collection and analysis. The treatment group included 28 students (21 female, 7 male) in an intact 8th grade English class; the control group included 28 8th graders at the same school, matched to treatment students on baseline measures, who did not receive the workshops.; Results are reported for comparisons between treatment and control performance and for pre- and post-improvement on additional measures that were not collected from the control group. Additional analyses examined treatment versus control and pre- to post-improvement among students at the lowest baseline levels. Qualitative data from four case studies supplemented quantitative data. Outcome variables included low-level skills, writing fluency and vocabulary use, and high-level processes, metaphor and multiple perspectives use, revision, and overall writing quality. Based on the literature, it was expected that automatic low-level skills would free working memory resources for high-level composing processes.; Treatment students improved significantly more than controls in appreciation of multiple perspectives (p=.043) and in the vocabulary measure of type/token ratio, where the lowest treatment performers improved significantly more than their matched controls (p=.010). The strongest results were found for within treatment group pre- and post-performance comparisons on measures for which there were no control group data. At post treatment students wrote significantly longer journal entries (p<.001); they used more metaphors (p=.001) and showed greater appreciation of multiple perspectives in journals (p<.001). However, the average number of metaphors declined significantly (p=.001) from first to last poem written by treatment students. Treatment students performing lowest at baseline improved significantly on the same outcomes but did not show a significant decline on metaphor use in poetry.; The conceptual framework was grounded in the view that fluent low-level skills are a precondition for the development of high-level metacognitive processes. Limitations included length of intervention, lack of random sample, potential for bias, and lack of a control group on metaphor use. Future research should address these limitations, especially implementing a longer intervention and collecting control data on metaphoric competence. |
