Self-assessment and peer-assessment, as two alternative assessment procedures, have appealed to researchers in recent years and motivated L2 researchers to examine these two techniques. However, most of the studies have used them for summative purposes, and the formative dimension these two methods can have for learning has been neglected. This study was an attempt to find how they contributed to learning gains. To that end, sixty Iranian male and female intermediate language learners at a language institute were randomly assigned into three treatment conditions: Selfassessment and peer-assessment as experimental groups and teacher assessment as a control group. A language proficiency test was used to homogenize language learners, and a posttest was administered to measure the amount of gain language learners achieved after treatment sessions. We analyzed the test data using descriptive and inferential statistics as implemented in SPSS, a general-purpose computer program for data analysis. Results from a one-way analysis of variance showed statistically significant differences between the score means of three treatment groups. Post-hoc analyses revealed that language learners in the peer-assessment group outperformed those in the other two groups. The findings suggest that peer-assessment as a cooperative technique can be used in language classes to help students improve their writing abilities.