A Study of Peer-Review Grading Within Graduate Courses in Biostatistics

• Activity Number: 408
• Type: Topic Contributed
• Date/Time: Tuesday, August 11, 2015 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistical Education
• Author(s): Elizabeth Colantuoni* and Zhenke Wu and Jeffrey T. Leek and Hongkai Ji
• Companies: The Johns Hopkins University and The Johns Hopkins University and Johns Hopkins Bloomberg School of Public Health and Johns Hopkins Bloomberg School of Public Health
• Keywords: peer-review grading ; statistical education ; graduate level statistical education ; innovative teaching methods

Abstract

We completed a study of peer-review grading within two graduate-level Biostatistics courses using a freely available peer-review grading tool, Crowdgrader (www.crowdgrader.org). The first course focuses on both the theory and implementation of linear models (enrollment 30-40 students). The second course focuses on applying statistical methods for analyzing longitudinal data (enrollment 80 students). The goal of the study was to assess the accuracy of peer-review grading relative to traditional teaching assistant grading. For each assignment, each student (grader) peer-reviewed 5 submitted assignments and teaching assistants graded each submitted assignment. We estimated the intra-grader reliability by decomposing the variation in the assigned grades into between- and within- grader variance. We assessed if the between-grader variation could be explained by degree, department or assignment order. For each assignment, we compared the peer-review grade to the grade assigned by the teaching assistant (teaching assistant). In the presentation, we will discuss what we learned and make recommendations for faculty considering implementing peer-review grading within their courses.