@GMTplus7 Thanks for your detailed comments. And you must be a proud parent to have such kids 
I understand your logic: a great swimmer can earn many more medals than a great soccer player or similar, just like Michael Phelps and Usain Bolt. Therefore, a school with a successful swimmer is more likely to rank high in the Olympic medal list than a school with a successful sprinter/soccer player, and so the list can’t give a fair measure of the overall athletic performance for a school.
However, as I said in other replies, the list is positive correlated with the overall athletic performance of the universities. The problem is how strong the correlation is (i.e., if it is 1, then the list perfectly reflects the overall performance). But the first problem is how to define “overall performance”. Do we view all kinds of sports equally? Or do we put different weights on different sports according to their popularity and so on? I guess you would agree with me that we do actually put weights on different sports, since you mentioned that “China focused on some sports that most people don’t care about”.
OK, let’s now assign an intuitive distribution of weights: aquatics 0.25, track & fields 0.3, ball games 0.3, and others 0.15. This can be combined with the Olympic medals of a university to give a score measuring the overall athletic performance of that university. For example, if a university has 15 aquatics medals, 10 track & fields, 4 ball games, and 3 others, then the overall score is 8.4. So then we have a ranking of universities according to their calculated scores, in addition to the ranking of universities according to their total Olympic medals. Finally, the problem is whether the two lists match well - the degree of such “match” is the correlation I mentioned above (from 0 to 1).
Now, it is apparent that the ranks of a university in the two lists would differ a lot only if its neighbors in the Olympic medal list and itself are seriously “lopsided” in sports (i.e., they are strong only in some sports). This is true for most low-ranked universities, but untrue for most high-ranked universities. For instances, Berkeley, Florida and Stanford all have very strong swimming teams, but they are also pretty strong in many other areas. Therefore, the correlation is close to 1 if we only consider high-ranked universities, and would gradually decrease as we take account of more and more low-ranked universities.
So, the conclusion is: the upper half of the list (i.e., high-ranked universities) should well reflect the overall athletic performance of the universities, while the lower half of the list should be viewed with extra caution. This can be “confirmed” by the NACDA Directors’ Cup rankings: https://en.wikipedia.org/wiki/NACDA_Directors%27_Cup. It is obvious that the high-ranked Olympic universities are pretty consistent with the high-ranked NACDA ones.