Inherently every college ranking has a bias. So by aggregating these lists together in total the result can be more meaningful and less biased.
So I combined together five major best college lists that rank a significant number of schools to see what the overall top colleges are based on a combined ranking. Here are the lists I combined together with equal weight and simply the schools with the lowest total score wins. For example the top school didn’t finish first in any list but was #2, #2, #3, #4, #4 = 15, while the #20 school came in at 143:
With Rice it’s easy, Rice doesn’t have that national/global name recognition, so it is an easy university for people to sleep on. Even though it’s been there, was there when US News first ranked their top handful of colleges in 1983.
I’m not saying it isn’t a good school, it is in fact very good. All I am saying is that it does not have the same recognition and prestige as schools like Columbia, Duke and UChicago, which it is ranked above. Also, you can’t really average the rankings since they each use different metrics to calculate the rankings.
You can’t just add up the ranks. You have to normalize the scores from all the rankings, then create a new score which will produce the new ranking. There might be a huge difference in the scores between let’s say rank 3 and rank 6 in one ranking and very little difference between them in another. If you don’t normalize, you will get terrible answers.
So instead of using the rank in Usnews, you have take the score of 98/100 for Harvard. Now you have to determine on the same 100 scale what Harvard scored in let’s say the Money ranking, decide what weight each ranking will have and then add up the scores for a final score.
That will be a Lot of work for basically very little value if you ask me
Btw you also ignored two of the ranking systems that are probably the most numbers driven, namely the ARWU and the CWUR rankings. Your results would be very different if you used them.
This may be my bias, but unless you can differentiate between majors and industries and have a reliable salary database of all graduates, output based rankings that try to measure success after college don’t really tell you anything useful. Garbage in Garbage out.
For me personally, if I were doing this, I would use three rankings. Usnews 50%, ARWU 25%, CWUR 25% to get a nice blend of undergraduate experience and university preeminence. But again that’s just me. To each his own.
I did what I recommended and here are the rankings for the Universities I looked at.
I gave USNews 50% weight and ARWU and CWUR 25% each, since I wanted to emphasize undergraduate experience at these universities. I Think most folks here are familiar with USNews and ARWU.
For those who are unfamilar with CWUR, here is their criteria
Here is the ranking with the final scores for the universities I included.
Harvard 99.00
Stanford 90.74
MIT 88.58
Princeton 87.68
Columbia 85.21
Chicago 84.73
Yale 83.16
Caltech 79.89
Berkeley 79.36
Penn 76.53
Cornell 75.19
Johns Hopkins 74.61
Duke 71.43
Northwestern 71.04
UCLA 69.18
Michigan 63.85
Vanderbilt 63.00
Brown 62.94
Rice 61.84
Dartmouth 61.29
CMU 57.79
Illinois 55.88
UVA 55.47
Just goes to show you that you can get whatever ranking you want but to me this list looks reasonable based on my experiences. If you want me to add any Universities to this, let me know and I will do my best
If one comes into a ranking with a bias and then weights the factors to achieve the results they want to produce (say US News), one produces a biased ranking.
The more one blends more and more rankings each with their inherent biases, one will start to get a better picture of the college landscape. FYI Cornell finished right outside the Top 20 when the 5 rankings were aggregated. I can share the next 20 if people are interested.
Of course @CollegeAngst , The difference is @ClarinetDad16 's composite used rankings that were focused on the undergraduate experience (4/5 rankings were about undergrads). Your rankings are weighted heavy on a universities research output (the post-graduate programs), as 50% was measured by that.
Neither way of looking at things is wrong, it just depends on how you want to look at it.