Awhile back, Payscale.com published a report that included salary data for more than 300 colleges. I developed a model to examine the variation in starting salary among colleges, controlling for a set of institutional variables, including school type and size, geographic location, demographics, financial need of students, admission and graduation rates, endowment size (adjusted for enrollment), net tuition and fees, and athletic team success.
The institutional characteristics that are statistically significant and associated with higher average starting salaries include:
Having attended a medium or large college (in terms of enrollment), as opposed to a small one, with the premium being highest for graduates of large schools.The institutional characteristics that are statistically significant and associated with lower average starting salaries include:
Having a large percentage of African American and Asian students enrolled, with the Asian affect more than twice that of the African American one.
Having a large endowment, adjusted for enrollment
Having a higher net price to attend
Having a successful athletics program, as measured by director’s cup score
Being a public schoolIt should be noted that the results of this statistical model are correlation, and should not be considered as causative.
Being located in the southeast or the midwest, with graduates of schools in the southeast earning the lowest starting salaries
Having a large percentage of females enrolled
Being less selective, measured by the percent of applicants admitted
The fun part is next – comparing the graduate’s actual starting salaries versus that predicted by the model. There were 298 colleges used in the model. The actual starting salary was greater than that predicted by the model for 145 of the schools, versus 153 schools whose actual starting salaries were less than that predicted. The two schools whose actual salaries were closest to that predicted are the University of North Dakota and Princeton University, both within +/- $10. The 25 schools whose graduate’s earned the most over their predicted salary were: