The recommended gender fair teaching materials are sorted by a personalized recommender system. The Gender4STEM recommender system computes the results of the Gender4STEM self-assessment tool to generate a level of gender fairness in teaching (the user profile) and to generate recommendations of teaching materials. The final result is then uploaded to the Gender4STEM Teaching Assistant where the user can access his/her personalized recommendations.
But, how does it work?
The core of the recommender system is on a graph-based semantic recommender system. This system relies on a a graph model mapping that establishes relationships between the Gender4STEM competency model , the teaching materials and the questions in the Gender4STEM self-assessment tool.
Based on the user profile that has been generated, the recommender system is able to calculate the best matching teaching materials that will improve the level of gender fairness of the user teaching practices. A matching user profile –><– teaching materials score is determined, where the materials with the highest scores will be the most appropriate for the users. This is known as the appropriateness of the teaching materials to the level of gender fairness of the user teaching practices.
The recommendations of teaching materials are grouped into sets for each teaching practice and ranked according to their appropriateness score. For each sub-list individually, only the teaching materials achieving an appropriateness score above 50% of the maximum score are recommended. Indeed, for efficiency reasons, the Gender4STEM Teaching Assistant only recommends materials with such significant appropriateness scores. However, the user can also browse other materials at will for inspiration.
It is worth mentioning that the Gender4STEM recommender system works with user pseudonyms only, thus avoiding data privacy issues.