Peer assessment has been implemented in schools as both a learning tool and an assessment tool. Earlier studies have explored the effectiveness of peer assessment from different perspectives, such as domain knowledge and skills, peer assessment skills, and attitude changes. However, there is no holistic model describing the effects of cognitive and affective feedback, grading, and prior knowledge, and earlier studies often discuss learning outcomes without ascribing them to particular causes. Moreover, few studies have differentiated between the effects of peer assessment on both assessors and assessees. This study used a path model to investigate how two online peer-assessment activities--rubric-based assessment and peer feedback--affected the learning performance of assessors and assessees. One hundred and eighty-one high school students engaged in peer assessment via an online system--iLAP. Several path models were tested and we found that the original model did not fit when the variable of cognitive feedback from peers was included. The best fit model was the one in which direct paths from cognitive feedback from peers and student exam scores in a prior Humanities course were removed. (Contains 3 tables and 4 figures.)