They initially did the analysis with intentionally inaccurate assumptions, while the various groups in OPERA made improvements on these aspects. For example, the biggest factors are
- The distance between CERN and the experiment ("geodesy")
- The measurement of the time the protons hit the target at CERN
- The synchronization of the clocks between CERN and the experiment.
- eight or ten other smaller effects
The thorough analysis with bad assumption led to neutrinos that were 1048 nanoseconds too fast. Then they "opened the box" on the more accurate systematic measurements. The corrections to the accurate results were 988 nanoseconds, leading to the 60 nanosecond too-fast result.
They "opened the box" six months ago, and they were clearly stunned. This collaboration of hundreds of scientists have considered possible mistakes in the analysis for this time.
An aside on the distance between the source and the experiment: They show a huge step function at the time of the Italian earthquake in 2009. This huge change in the distance was 7 centimeters!
This is an uncomfortably reasonable result--very precisely determined.