To put it simply, would you say that every scientific study following the Scientific Method will produce true/correct facts? I was saying with sarcasm, NO, even faithful use of Scientific Method will sometimes produce bad science.
How do you judge the success or failure of the scientific method? Given it's an iterative optimisation method, it seems to me that the only valid measure of success or failure is the long term trend - the scientific method is successful if, over time, the trend is towards increasingly successful explanations for observed reality. I don't think it even has to be a monotonic improvement to justify being called a success - the biggest gains often come when particular theories 'fail', suggesting that both the immediate successes /and/ the immediate failures contribute to the success of the scientific method.
That measure of success suggests two things: firstly, that the success of individual theories is /not/ ultimately relevant to the success of the scientific method, and secondly, that 'stopping the clock' is /not/ a valid way to argue about the its success.
The scientific method is /not/ about 'outcomes', it's a process of continual improvement, and one that is astonishingly successful. The closest thing I can think of to a failure of the scientific method is the screwed up biological sciences that pertained in the early USSR, driven by political forces. Even there it did eventually correct itself when given the chance, and that localised failure did not affect the rest of the world. There are probably other examples of localised failure, but as with individual theories, that sort of failure doesn't indicate a failure of the scientific method.
It probably seems like I'm wiggling the goalposts around to make my argument easier, but I'm not really - the issue at hand is the success or failure of the /scientific method/, not 'science', or some particular set of theories or ideas. The state of the process at any instant in time does not reflect on the success of failure of the process - only the long term behaviour does. The process is as close to infallible as I can imagine, given enough time and diversity of practitioners.
himi