Is the Scientific Method Flawed?

I love this thread. Thank you for the thoughtful comments, guys.

My input: the scientific method is good to advance science incrementally (or, as Ernest Rutherford once said, "All science is either physics or stamp collecting"). To get a breakthrough that advances science by leaps and bounds, you need something else. (At least to start).

I think the issue is a confusion between free-form experimentation and formalized proof. You need both approaches, but if you want your conclusions to be respected, you need to formalize your methodology at least when you present your findings.

Just do not include "If the results don't fit your expectations-fudge your results"-- then I'm OK with traditional research.

The scientific model as spelled out in grade school text books shouldn't be taken to literal or specific, and should be considered a pedagogical analogy. It still covers a lot of stuff as is. And you can easily shoehorn every thing else to fit the textbook method depending on how you interpret each step, or vis-a-versa like here, and force the method to be more vague description of what is actually done. But either way is kind of missing the point. It is about the back and forth between model and reality, that the model must be a description of reality and that reality is the final discriminator. And his revision does seem to not illustrate that by using unidirectional arrows, and not separating testing from model building.

He does have a point though, as memorizing the steps probably doesn't really do much good. Maybe it works well for students that don't have an intuition about such things or limited exposure, that it is one of those things you need to learn the rules before you can understand when to break them. But that could probably be all wrapped up together, instead of just making you memorize the steps in grade school... and then letting you figure out the practical world is different if you ever end up in an undergraduate science program. I do think I would have a few words with teachers that spent too much time on the formality of my science fair projects in grade school instead of substance, although formality is a lot easier to teach ( and I would have plenty to say to those teaching too much formality of how to cite works...).

There are times that some of that formality shines through. Astronomers and particle physicists that have their data before they do tests or look for patterns know better than to just fish for patterns or results until they find something, as they will likely do the statistics incorrectly and by luck find something that is not actually significant. So they split the data up, play around with only part of it until they think they found something, then use the rest of the data for a more formal test. That method is rather important and distinct from just messing around, and fits the more formal idea of needing a question first, and research into how to test it, etc.

At some point though, talking in extensive detail about the method itself starts to diverge from introductory science, and go down the route of philosophy of science. Not to say that makes it useless, quite the opposite, but there are volumes written on that topic and a lot of well laid out thoughts that get more into the meat of things without dwelling on what may be merely superficial.