This February, Anna Wärnsby, Malmö University; Asko Kauppinen, Malmö University; Laura Aull, University of Wake Forest; Laura Anderson, University of South Florida; Joe Moxley, University of South Florida; Norbert Elliot, University of South Florida; Katie Walkup, University of South Florida; and Djuddah Leijen, University of
Tartu presented at the Conference on Writing Research Across Borders.
Below is a small sample from the presentation,”Politeness, Social and Intrapersonal Presence in Student Peer Reviews: A Cross-Cultural Analysis”:
While pedagogical essays have been published regarding best classroom-practices for conducting peer review, surprisingly little quantitative, replicable, aggregated, data-driven (RAD) research has been conducted on peer review in the discipline of writing studies. Regarding the paucity of empirical research in NCTE journals, Richard Haswell concluded that “peer critique seems to be one of the least studied of practices now very common in college writing classrooms” (211).
On Keeping Score, the largest quantitative study conducted on peer review, investigated the efficacy of having first-year composition students score one another’s intermediate drafts of essays using a five-trait rubric across 482 sections of two introductory composition courses (Moxley and Eubanks, in press with WPA: The Journal of Writing Program Administration). This study analyzed 46,689 reviews, which consisted of 16,312 reviews conducted by instructors and 30,377 reviews conducted by students. The researchers found low to modest correlations between peer ratings and instructor ratings on individual assignments. On average, peers assigned higher ratings than instructors although over time students’ scores were more highly correlated with instructors’ scores. The average differences in ratings between the students and instructors were smallest for Focus and Format and greatest for Evidence. Students who received higher ratings on their own writing from instructors provided scores that had a broader range of scores and were more highly correlated with instructors’ scores than students who received lower scores from instructors. Generally, peers had a smaller rating variance of scores than instructors.