Can faculty productivity be predicted with an algorithm?
- Professors at the Massachusetts Institute of Technology say they have developed a predictive analytics model that can forecast the future publishing and research activity of faculty members based upon previous research and writings.
- The data is published in Operations Research, and suggests that bibliometrics based upon citations and co-authorship and scholarly networking helped their system to make more statistically sound, virtual tenure decisions for more than 50 Ph.D. earners since 1995.
- The system does not account for service, and skeptics say that scholars could cheat the system to meet its basic standards of performance, and that any use in actual tenure decisions would be an incomplete notion of promotion.
While publishing is perhaps the most critical element of tenure and promotion review, too many universities are built on cultures which also value student feedback, peer collegiality and the exposure of scholarly work in non-traditional academic media like social networks and traditional news outlets for sharing expertise.
In the future, colleges should consider using analytics as a faster means of reviewing the productivity of faculty members, but until robots completely take over teaching in higher education, inspiration and scholarly exposure will still likely require a human touch in the important review process.