Recommender systems for teachers: The relation between social ties and the effectiveness of socially-based features

Open Educational Resources (OER) repositories provide teachers with a wide range of learning resources (LRs), enabling them to design various learning sequences. However, search & select in large OER repositories can be a daunting task for teachers. Incorporating peer recommendations, as is common in online marketplaces, is becoming a popular solution that seeks to exploit the wisdom of the crowd for this task. However, teachers are often reluctant to take a contributory role and provide social recommendations. In addition, little is known about the actual value of social recommendations as a search aid. In this research, we implemented a “light-weight” socially-based recommender system (RS) within a large OER repository that includes social network features. We examined two aspects of the socially-based recommendation mechanisms. First, their utility as search aids that assist teachers in searching and selecting suitable LRs, and second, their impact on teachers' incentives to share recommendations that can assist fellow teachers. To study these two aspects, we examined two science teacher communities using this repository. The results demonstrated the incentivising power of social rewards, and the value of social recommendations as means for search & select. However, we also observed a heterogeneous effect of social features on teachers' behaviour. To explore the factors that may explain these differences, we employed a mixed-method approach, combining qualitative, quantitative, and Social Network Analysis methods. Triangulation of the findings underline the relation between the strength of the social ties within the teachers’ community and the effectiveness of socially-based features.