While I would not say that science is driven by consensus, I think it is important to have some baseline to begin with, on which you can verify, criticize and improve. If you had for example a 100 equally supported competing theories instead of one widely accepted one it would be much harder to work out the truth, because not one scientist would be able to study them all in his lifetime.
There is a list of all the superseded theories here:
http://en.wikipedia.org/wiki/Superseded_scientific_theories
its interesting to see how they evolved.
Regarding the code or data analysis that is done by GW scientists I would not be surprised to learn that it was sloppy work, because they are not specialists in this area. It would be better if such work was left to the professionals – software engineers and data analysts, but that of course requires a different level of funding. Still sloppy work is better than no work IMHO, because you can criticize, improve and work out the issues.