I used to have the same problem, but really they're following the spirit of the scientific process, just not the rigour. "Think of something and try it out" is the fundamental essence; "try it out rigourously" is the follow-on.
My favourite lack of rigour is their lack of statistical knowledge. They had a toast dropper, and were testing the randomness before the mythbusting run. The test toast dropped 7/10 drops on one side. "This is not random enough!", cried Savage, "If it was truly random, it would have been 5 on each side!". Betrays a startling lack of statistical knowledge :) But yes, they have done wonders for promoting the spirit of inquiry.
> Testing is often edited due to time constraints of a televised episode. It can often seem as if the teams draw results from fewer repetitions and a smaller data set than they actually have. During the Outtakes Special, they specifically stated that while they are, in fact, very thorough in testing myths and repeat experiments many times in many different configurations, it is simply impossible to display all of it during a program. Beginning in the fifth season, episodes typically contain a prompt for the viewer to visit the show's homepage to view outtake footage of either additional testing, or other facets of the myths being tested. However, Savage himself has acknowledged that they do not purport always to achieve a satisfactorily large enough set of results to overcome definitively all bias.
(I enjoyed[1] mythbusters, and thought it was good.)
But the way they had that big metal plate that they'd drop in slomo with a clang saying "BUSTED!" gave the impression that this was now FACT and TRUTH.
For a fun tv show that's fine, but for science education (which is what a lot of people thought MB was) it's sub-optimal.
[1] although I prefered the remixed versions where each item was run in a single block, rather than the broadcast version which flicked from item to item and back again, with many catchups and coming nexts.
Apparently the ridiculous amount of catchups and re-updating is something the presenters don't like either, but their producer listens to the cuts in her car, and basically demands the catchups so she knows what the current cut is about.
My favourite lack of rigour is their lack of statistical knowledge. They had a toast dropper, and were testing the randomness before the mythbusting run. The test toast dropped 7/10 drops on one side. "This is not random enough!", cried Savage, "If it was truly random, it would have been 5 on each side!". Betrays a startling lack of statistical knowledge :) But yes, they have done wonders for promoting the spirit of inquiry.