This study examined the impact of collaborative filtering (the so-called recommender) on college students' use of an online forum for English learning. The forum was created with an open-source software, Drupal, and its extended recommender module. This study was guided by three main questions: 1) Is there any difference in online behaviors between students who use a traditional forum and students who use a forum with a recommender?; 2) Is there any difference in learning motivation between students who use a traditional forum and students who use a forum with a recommender?; 3) Is there any difference in learning achievement between students who use a traditional forum and students who use a forum with a recommender? This study was a one-way quasi-experimental design where the independent variable was the type of forum (two levels: traditional forum and forum with recommender). Students registering in four sessions of a college English course participated in the study and were randomly assigned into two groups. The total sample number was 144. The whole experiment lasted eight weeks. All students took a diagnostic test as a pre-test in Week One. From Week Two to Week Seven, students joined the class and wrote summaries, reflections and comments on the online forum. Students in different groups went on different forums. All students were asked to participate in a midterm exam in Week Four, and a final exam and online survey in Week Eight. Data collected in this study included pre-test scores, midterm exam scores (receptive and productive language test scores), final exam scores (receptive and productive language test scores), online survey (motivation and recommender perception), and Weblog data. The data were analyzed by using ANOVA procedure and Wilcoxon-Mann-Whitney "U"-test. The findings were as follows: 1) Students in the group with the forum recommender read online posts more frequently than the control group, and 2) students with the forum recommender outperformed their counterparts in their productive language test scores. However, there was no significant difference in learning motivation between the two groups. To enhance motivation for using the recommender, students offered their comments on how to revise the recommender, such as making the recommendation rating more personalized and explicit. This study is expected to provide empirical evidence to recommender research in education as well as broaden innovative insights into instructional recommender design. (Contains 6 tables and 4 figures.)