Predicting Medical Students' Specialty Choices Using Geographic Variables: Findings From Data Spanning 30 Years (Poster 50)

• In Event: AERA Poster Session Nine
  In Poster Session: Division I Poster Session

Abstract

Understanding students’ specialty choices are imperative for medical schools to better align their support, learn the strength of their schools, and improve their teaching and learning. A total of 913 medical students who graduated from 1986 to 2015 from Michigan State University are included in this study. Students’ geographical and background variables at matriculation were used to predict their specialty and primary care choices. Machine learning methods, Tree methods and Random Forest, were conducted to compare and evaluate the predicting results. Results identified important factors in predicting specialties and showed a shift of focus of students in specialties during the 30 years.

Authors

• Chi Chang, Michigan State University
  Presenting Author

• Hope O. Akaeze, Michigan State University
  Non-Presenting Author