How is science really done?

Step 1: recognize and describe a problem.
Step 2: form a hypothesis
...

That's the way we would all like it to be done, by following a simple and clear 6 step recipe.  However, the truth is far messier.  Scientific American has an article of value to all but specifically for elementary school science  teachers.

"We are teaching the scientific method too simplistically," says Judy Scotchmoor, of the University of California Museum of Paleontology at Berkeley, where she develops tools for training STEM teachers. An overemphasis on rote learning and an under-emphasis on critical thinking yield students who lack an interest in science. She quoted surprising statistics showing that the longer students study science, the more negatively they view it.

To replace the linear approach to teaching the scientific method, Scotchmoor designed a flow chart, which Stefanski has implemented in his classes, that shows the messiness of scientific inquiry, with arrows running in multiple directions among several steps: exploration and discovery, testing ideas, benefits and outcomes, community analysis and feedback. She's had a number of working scientists diagram their own discoveries on her flow chart and post them on the Web as examples. Her goal is to show that "science is dynamic and creative and far from linear."

If the link in the quote doesn't work, here is the flow chart.  The articles describe these enigmatic tubes.