Identification of Ineffective Team Members Using Normalized Peer Ratings

Abstract

Peer assessments are used in the senior process design course at the University of New Brunswick to determine individual grades, reduce free riding and improve team dynamics. The results of the monthly surveys are shared with students as formative feedback on their performance and used to calculate the relative performance of team members by dividing the overall average score of each student by the team average. This study examined whether the resulting normalized peer ratings could be used to identify ineffective team members. We found that the normalized ratings of effective team members follow the normal distribution with a mean of unity and a standard deviation of about 0.04. The small monthly variations in the standard deviation of the distribution are not statistically significant indicating that the characteristics of the distribution are time-invariant in our course. Students who obtain normalized ratings less than 0.9 are generally ineffective team members and their ratings should not be used in the calculation of the team average to avoid inflating the normalized ratings of their teammates.