This is more a question of engineering. Doing modern physics experiments require extremely complicated machines with tolerances so fine that the tiniest amount of noise in the results can throw off the accuracy of the measurements.
The experimenters were not some quacks making spurious claims. They very, very, very carefully checked and rechecked their results before deciding to talk to the outside scientific community and asking for help.
Given the complexity of these machines, it doesn't surprise me at all that a single cable might be the source of the errors. If you write code, you find many times throughout your career you get errant results. In the process of tracking down the cause, you may spend hours, days, or longer going line by line and missing the problem. In the end, it may turn out to be a mis-named variable or the wrong kind of comparator (== vs === in js). Simple things that are easy to miss.
There was a story recently about a commonly used algorithm that dates back decades that had a bug in it. When the algorithm was originally developed, it was never thought that it would still be in use 30 years later and disregarded the size of data sets available. When the algo was used on a very large, modern set of data an integer would flip and go negative. The point is, its easy to miss things in very complicated systems and very hard to make sure there are little or no bugs at all. I don't remember what type of algo it was, but it was an interesting read.