Electrical and computer engineering professor creates new graduate school ranking algorithm

Reddy picIn March, U.S. News and World Report (USNWR) releases the latest version of its graduate school rankings, which includes engineering programs. While the USNWR ranking is widely followed, some researchers believe that it may not be the most accurate indicator.

For USNWR graduate engineering rankings programs at 215 engineering schools that grant doctoral degrees are surveyed. Engineering specialty rankings are based solely on peer assessments by department heads who rate the other schools that offer a doctoral degree in the specialty on a five-point scale.

Other rankings have used different metrics including number of publications and citation counts to rank graduate programs.

Dr. Narasimha Reddy, the J.W. Runyon, Jr. '35 Professor I in the Department of Electrical and Computer Engineering at Texas A&M University, believes a more accurate ranking could be based on faculty hiring from peer institutions.

Reddy, who is also the assistant agency director for national and global initiatives at the Texas A&M Engineering Experiment Station (TEES), said that the best people to measure a program’s value are the people in the program because they know about their field more than anyone else.

“If I assume that universities are going to try and hire from programs that are better than them, or at least as good as them, this metric we’re using would say that the more people you place the better program you are,” he said.

Reddy and his research team devised a statistical and mathematical approach to rank graduate programs using algorithms deployed on a mutual “hiring graph” among universities. They collected faculty data from top 50 departments across the United States (according to USNWR) to construct their hiring graph.

Reddy said the new rankings were generally consistent with U.S. News rankings, but they did expose some new observations about some particular programs.

“The interesting thing is our results vary from U.S. News in a few categories,” he said. “U.S. News seems to favor private programs over public programs. The programs that are rated higher by them and lower by us are private programs and the programs that are rated higher by us and lower by them are public programs.”

In addition to the discrepancies between public and private institutions, Reddy said USNWR rankings tend to favor colleges with a reputation that may no longer exist. Reddy’s algorithm can shed light on changing reputations based on the history of hiring patterns.

“U.S. News has more of a lag effect — once you have a reputation it takes a long time to lose it,” he said. “Similarly newer programs are doing very well now in placing their students, but they’re not ranked as high as they should be ranked. That’s something you can measure with this data.”

While getting data for his rankings may prove difficult and has the possibility of becoming skewed if programs start concentrating on getting their graduates placed in peer institutions, Reddy believes it’s a good technique to supplement UNSWR rankings and reveals interesting insights.

For more information please visit here.