Bruce G Charlton writes "In three studies looking at the best institutions for 'revolutionary' science, MIT emerged as best in the world. This contrasts with 'normal science' which incrementally-extends science in pre established directions." If you're interested in reading more about how this was determined, read more below.
"My approach has been to look at trends in the award of science Nobel prizes (Physics, Chemistry, Medicine/ Physiology and Economics — the Nobel metric) — then to expand this Nobel metric by including some similar awards. The NFLT metric adds-in Fields medal (mathematics), Lasker award for clinical medicine and the Turing award for computing science. The NLG metric is specifically aimed at measuring revolutionary biomedical science and uses the Nobel medicine, the Lasker clinical medicine and the Gairdner International award for biomedicine. MIT currently tops the tables for all three metrics: the Nobel prizes, the NFLT and the NLG. There seems little doubt it has been the premier institution of revolutionary science in the world over recent years. Also very highly ranked are Stanford, Columbia, Chicago, Caltech, Berkeley, Princeton and — in biomedicine — University of Washington at Seattle and UCSF. The big surprise is that Harvard has declined from being the top Nobel prizewinners from 1947-1986, to sixth place for Nobels; seventh for NFLT, and Harvard doesn't even reach the threshold of three awards for the biomedical NLG metric! This is despite Harvard massively dominating most of the 'normal science' research metrics (eg. number of publications and number of citations per year) — and probably implies that Harvard may have achieved very high production of scientific research at the expense of quality at the top-end."