| Summary: | 博士 === 高雄師範大學 === 英語學系 === 99 === The research employed both quantitative and qualitative meta-analysis methods to systematically synthesize empirical studies investigating peer feedback activities in an ESL/EFL writing class between the years of 1990-2010. Ninety-five individual studies concerning ESL/EFL peer feedback were included in the final body of research for the synthesis. Of these 95 studies, 49 studies were considered eligible for quantitative meta-analysis, and 11 studies provided sufficient data to estimate the effectiveness of peer feedback, i.e. effect size via Cohen’s d-index on ESL/EFL writing improvement. Constant comparison method, which originated from Glaser and Strauss’ Grounded Theory (1967), was used to analyze the 54 of the 95 studies that contained qualitative data, later organized into four strands of ESL/EFL peer feedback research.
For the quantitative meta-analysis, two models, Fixed Effects Model and Random Effects Model, were employed to analyze the effect sizes. The results of the analysis using the Fixed Effects Model indicated that peer feedback had positive and near medium effect on ESL/EFL students’ writing improvement (d=0.48). However, the effect size estimated for each of the samples of the 11 studies did not turn out to be homogeneous. Therefore, the hypothesis that sampling error explains the difference in these effect sizes was rejected. Due to the heterogeneity in the sample, the approach of the analog to the Analysis of Variance was employed to look for systematic effects of influence from study-level descriptors. The results of the further analysis suggested that participants’ English proficiency, the length of peer feedback treatment, and the provision of specific peer feedback demonstration or training might be potential moderating variables for excessive effect size variance. Under the analysis of the Random Effects Model, the random variance component was taken into consideration and added into a series of calculations. The weighted mean effect size (d) was enhanced from 0.48 to 0.85, which suggested that after controlling for the random variability, the effect size was enhanced. In addition, the homogeneity hypothesis was not rejected, suggesting the excessive variance among effect sizes would not exist when taking the random variance component into calculation. This further analysis suggested that peer feedback had medium to large effect on students’ writing improvement.
Four main themes emerged from the qualitative meta-analysis through the constant comparison on the findings and discussions of the 54 studies. The four main themes were: (1) The success of peer feedback was established on participants’ appropriate attitude and teacher’s sufficient involvement and timely intervention; (2) The discovery type of stances in peer reviewing should be encouraged and cultivated through training and a teacher’s modeling; (3) Peer feedback training was a key factor in deciding the success of peer feedback activities in the writing instruction, and teachers played a critical role in designing successful peer feedback training; and (4) Appropriately incorporating technology into peer interaction and peer feedback training in the writing classrooms helped enhance classroom dynamics which might help the peer feedback process. Overall, the four main themes suggested that writing instructors should consider the interrelation among teachers, learners, training, and technology when they employ peer feedback activities in the writing classrooms.
With using both the quantitative and qualitative meta-analysis methods, the study made several contributions. First, it directs readers to perceive ESL/EFL peer feedback research from multiple aspects. Second, it benefits different audiences, such as writing instructors, researchers and learners, by providing pedagogical implications, research directions, and peer feedback strategies. Third, this research also informed instructional designers, technology programmers, and even software developers to consider the capabilities or affordances of technology to develop a better platform to enhance peer discussion in a writing classroom. Additionally, the research reduced the gaps identified in previous reviews (Ferris, 2003; Ferris & Hedgcock, 2005) and added to our understandings of overall findings and research trends in the field of ESL/EFL peer feedback in a writing classroom.
|
|---|
