Purpose: When self and peer assessment methods become commonly used in the teaching process, the most important problem turns out to be the reliability of the ratings acquired from these sources. Increasing the rater reliability has great importance in the performance evaluation for the reliability of the measurement. This study aimed to determine rater behaviors university students display in the process of self and peer assessment. The research was based on a descriptive model. The participants were 58 students at the Guidance and Psychological Counseling Program in 2017-2018 academic year at a foundation university in Ankara.
Findings: Many Facet Rasch Model (MFRM) analysis was applied, and no statistically significant difference of raters’ severity and leniency behaviors in the ratings was observed in terms of gender, but there was a statistically significant difference based on the rater types (self and peer). The raters seemed to be more lenient in self-assessments. The study also showed that while raters showed central tendency behavior on individual level, they did not show such tendency at the group level. It was concluded that individuals’ ratings are more biased than group ratings when they evaluate group performance.
Implications for Research and Practice: Some of the raters had differentiating rating behaviors based on the groups. The teacher candidates made systematic mistakes in the performance evaluation process and showed behaviors that had negative effect on the validity of the rating. It is important for the raters to conduct studies to reduce the scoring bias of the raters.
Peer assessment, self-assessment, rater bias, alternative assessment, Many-Facet Rasch Model