Every college wants to know who its peers are. Discussions about the peer group often bog down in confusion around the peer / aspirant distinction. One way to address the question is by listing some factors (endowment, faculty salaries, board scores, etc.) all could agree on. We are “like” those institutions if we are close to them on a reasonable statistical measure of closeness.
I looked at 383 institutions from the U.S. that could be broadly considered similar under a definition provided by IPEDS. I gathered data for all 383 on full time enrollment, endowment per student, net tuition revenue per student, the six year graduation rate, faculty salaries, tuition and fees, retention, the student to faculty ratio and ACT 25th and 75th percentile measures. I then constructed what economists call a “gravity” model, measuring distance from Wittenberg on those 10 factors for each of the other 382.
Here are the top 22 on the closeness measure:
|Lake Forest College **|
|Roanoke College **|
|Juniata College **|
|Knox College **|
|Birmingham Southern College **|
|Washington & Jefferson College **|
|Saint Anselm College|
|Ohio Wesleyan University **|
|Susquehanna University **|
|Luther College **|
|Mount St. Mary’s University|
|Hiram College **|
|Saint Johns University|
|Marietta College **|
|Allegheny College **|
The highlighting (**) represents my musing…. I had never heard of McDaniel College (my apologies to our friends out there). The others not highlighted seem to be outside our geographic area.
So If I were the peer group czar, I would choose our peer group by picking the first 12 colleges on the list that pass the laugh test. A general approach would be to start at the top and uses a consensus method. Lake Forest (most like us on the gravity measure) is either in or out by consensus vote…. etc.
Who is at the bottom of the list, using the above described model? (These schools are apparently not like us.)
|Claremont McKenna College|
|Cooper Union for the Advancement of Science and Art|
How would you pick an aspirant group, using this method? Drop down to say, the 100th position on the list, slice off the next 25 or so names, remove those who are unlike us because they are “lesser” institutions, and pick the dozen or so who you would really like to be.
Perhaps a different list of factors, or unequal weighting of the 10 factors used would generate a different peer list. But the bloodless approach I have described here removes the impressionistic factors often used to construct peer lists.