This work-in-progress paper explores the development of a research-based application for finding an effective mentor-mentee match. Mentorship is an important part of professional success, especially for underrepresented populations in STEM as it can further their education and careers. Mentoring experiences exist in a variety of ways, such as formal or informal. The most successful mentorships occur when the relationship evolves organically without official agreements in place. Mentoring programs exist, but there are limited programs addressing lasting connections. Moreover, finding an ideal mentoring relationship remains challenging. We hypothesize that for an ideal mentoring relationship to occur, there should be a percentage of matching between four dimensions: personality type, demographics, career aspiration, and interests. The combination of dimensions that are more impactful for mentor-mentee compatibility is unknown. Hence, we designed an experimental Phase 1 algorithm to test a preliminary combination using a tier-distance based approach. We used the algorithm to develop an application to automate the mentor-mentee matching process. We asked one hundred (100) participants to sign up as either a mentor, mentee or both, depending on their career stage for the Phase 1 experiment. The application then used the algorithm to score participants' responses and generate mentor-mentee pairs. The pairs with the highest and lowest matching percentages entered into mentoring relationships without knowing their scores to prevent bias. We conducted a survey on the mentoring experience after two months. Out of twenty-one (21) participants, eighteen (18) completed the post-experience survey. Results showed that 88.8% of the participants felt neutral or agreed with their match, while 11.1% disagreed and none strongly disagreed. Of the 18 respondents, 38.9% believed that all dimensions were important for a mentorship experience, and 22.2% believed that career aspiration was the most important. Results are extremely preliminary, so we cannot yet conclude whether the current algorithm is optimal or not. In the future, we will increase the sample size to complete the Phase 1 experiment and adjust the algorithm for greater user experience in Phase 2. We expect that with a well-tested algorithm, the research-based application will be an effective approach that can be widely accessible.
Are you a researcher? Would you like to cite this paper? Visit the ASEE document repository at peer.asee.org for more tools and easy citations.