Diagnosing the False Faster-than-light Reading
Despite poring over their results for months before publishing the faster-than-light readings of the OPERA project in September 2011, the CERN scientists missed the cause of the inaccurate readings. They kept working to verify their experiment, and in November announced that a new test confirmed the accuracy of its timing measurement. The results still showed faster-than-light travel.
They didn't hit on the problem until February 2012, when OPERA discovered two potential issues with its measurements. An oscillator used to provide GPS time stamps could have been overestimating the neutrino travel time. If that was the issue, it could mean the neutrinos were actually traveling even faster. The other potential issue meant the opposite: An optical cable bringing GPS signals to the OPERA master clock may have been malfunctioning, which would cause the system to underestimate the neutrino travel time.
With the optical cable fixed, the 60 nanosecond advantage over light speed disappeared, and all subsequent neutrino tests at CERN have matched up with the expected speed of light. Mistakes happen all the time in the scientific community, and they're treated exactly as this one was: Testing, re-testing and external studies seek to confirm a result. Faster-than-light travel just happened to be a big enough deal to attract a whole lot of attention.