This article examines results from one pilotstudy and two empirical studies of acollaborative filtering (CF) system applied ineducational settings. CF is a populartechnology in electronic commerce, whichleverages the interests of entire communitiesto provide targeted, personalizedrecommendations of interesting products orresources to individuals. In electroniccommerce, entertainment, and related domains,CF has proven an accurate and reliable tool;yet educational applications remain limited.From analyses of data from these three studies,we believe that CF holds promise in educationnot only for the purposes of helping learnersand educators find useful resources forlearning, but as a means of bringing togetherpeople with similar interests and beliefs, andpossibly as an aid to the learning processitself.