The study investigated the effects of anonymity on online peer assessment and compared three different conditions. Fifty-eight preservice teachers at a mid-size US university engaged in a series of online peer assessments during fall 2017. Peer assessment was embedded in a blended course as a required asynchronous activity using the Canvas learning management system. Students were randomly assigned to three different peer assessment conditions: anonymous, partially anonymous, and identifiable. They were asked to provide feedback comments and rate the quality of peers' work. The researcher examined to what extent three different conditions had influenced the quality of feedback comments, measured quantitatively through the number of words and negative statements. At the end of the semester, a survey that included a 5-point Likert scale and several open-ended questions was also distributed to analyse students' perceptions about peer assessment and anonymity. The results indicate that although students prefer anonymity, it may not be a necessary condition for increasing student engagement. Implications for practice or policy: (1) Instructors should provide several opportunities for online peer assessment to reinforce the skill of writing effective feedback throughout a course; (2) Students may be given an option for anonymous peer feedback to ease their anxiety; and (3) Providing specific grading criteria for feedback quality is strongly recommended.