Prior research on peer assessment often overlooks how much students learn from providing feedback to peers. By practicing revision skills, students might strengthen their ability to detect, diagnose, and solve writing problems. However, both reviewer ability and the quality of the peers' texts affect the amount of practice available to learners. Therefore, the goal of the current study is to provide a first step towards a theoretical understanding about why students learn from peer assessment, and more specifically from providing feedback to peers. Students from a large Introduction to Psychological Science course were assigned four peers' papers to review. The reviewing ability of each student was determined, and to whom the students provided feedback was manipulated. The features and focus of the comments from a sample of 186 participants were coded, and the amount of each type was analyzed. Overall, reviewer ability and text quality did not affect the amount of feedback provided. Instead, the content of the feedback was affected by reviewer ability. Low reviewers provided more praise than high reviewers, whereas high reviewers provided more criticism than low reviewers. This criticism from high reviewers described more problems and offered more solutions, and it focused more often on high prose and substance. In the only significant reviewer ability × text quality interaction, high reviewers described more problems in the low-quality texts than in the high-quality texts, whereas low reviewers did not make this distinction. These results suggest that high reviewers and low reviewers may utilize different commenting styles, which could significantly impact the benefits of peer assessment.