I'm wondering: what do the involved scientists think might be the source of error? GPS timing (namely, the synchronization of the two remote clocks) sounds likely to me, because I was initially surprised that they could even hope to have sufficiently accurate chronography to detect such a small duration.