The results came from ICARUS, and they looked at what is essentially the "dispersion" of the neutrino energies via looking at the dispersion of the created muons in a "neutral-currents weak-interaction" radiation. Tommaso Dorigo has a wonderful explanation for this whole process which you should read.
Essentially, what ICARUS found is that the muon spectrum is very much similar to what is expected for neutrinos moving at c, not at the speed that OPERA claimed.
They find that the energy spectrum of the detected neutrino interactions in ICARUS shows a very nice agreement with the expectation for well-behaved light-speed-moving neutrinos. A very dramatic distortion of that spectrum would instead be expected for the speed measured by OPERA, such that indeed ICARUS can place a very tight constraint on the superluminal speed of the CERN neutrinos: consistent with the speed of light, and not larger than that by more than four part in ten billionths. An order of magnitude looser than the limit obtained with the neutrinos from SN1987a, but still quite tight -and certainly excluding without argument the value of 50 millionths measured by OPERA.The preprint on the ICARUS result can be found here. I'm sure there will be other experiments to be conducted that will try to verify the OPERA results, but this one certainly doesn't look good for OPERA.
If you are unfamiliar with millionths and billionths, I can make it easier for you: the ICARUS result says that the difference between the speed of neutrinos and the speed of light cannot be as large as that seen by OPERA, and is certainly smaller than that by three orders of magnitude, and compatible with zero.