Neuroskeptic breaks down for us here just how "practical" lab experiments in school really isn't all that practical at all. It reminds me of physics experiments I did in high school, timing cars and balls rolling down inclines, graphing the data, fiddling around to make the data look like what we knew it should be, generally learning experimental techniques that was never useful, etc.
In particular, I remember this one time when I was filling out the lab report with "observations" that I knew would've been observed had I done the experiment on the lab bench. My science teacher came by asking what I was doing, and I explained how I knew the physics behind the experiment already so I was filling in what I knew would happen based on sound physical laws I knew. He wanted me to do the experiment anyway to see what would happen, which is of course what I knew would happen.
The point is, given practical lab experiments in grade school is really not at all practical, as Neuroskeptic explains, and if the student already understands the physics or scientific principals the experiment is designed to keep students entertained long enough to learn, then what's the point of wasting time doing the experiment?
Of course not all students know the scientific principals being taught, some even after all the experiments and notes are over and done with. Perhaps lab experiments help those students learn the science, impractical as the labs actually are to teaching students skills that they'll need if they want to do actual research, as explained by Neuroskeptic.
In that case, the lab experience ought to be designed without the false notion that it's to teach actual science research skills. Instead, focus on making the lab experience entertaining perhaps, supposing that helps learning, and focus on making the experience about feeling the spirit of scientific inquiry perhaps, again supposing that helps learning.
It would be more effective to focus on optimizing the lab experience to those things that enhance learning, however it's supposed to help do that, rather than optimizing the lab experience for a goal it doesn't even fulfil: giving students skills for doing real science research.
Practically speaking, this might mean doing lab experiments with less structure, less of the problem, hypothesis, test, results, conclusion, rinse and repeat dead procedure. Instead, maybe it'd help for students to do more exploratory experiments with completely unknown outcomes, then make observations that generates a thousand different hypothesis, and maybe testing just one of those hypothesis later.
The vast majority of experiments students get to do have problems that are set out for them by the teacher or textbooks, problems that are either obvious as to its outcome or problems that are boring. Then students have to write up the problem, hypothesis, etc, in a prescribed format that often doesn't even fit the kind of experiment being done. I've seen more exploratory experiments jammed into a problem/single-hypothesis framework, that would've been more interesting, fun, observation requiring, multiple hypothesis generating, etc, if not for a textbook or curriculum made to enforce that kind of notion of "experiment".
I suppose they enforce that kind of notion because they think science essentially boils down to writing up clever questions with neat solutions, writing up a neat hypothesis that can easily be proved or rejected with the equipment available in the lab. No exploration required. No messy "I don't know"s. No hypothesis for which we might not even know how to test. No possibility that even the question may be unclear at first, and for a while longer. No real chance for failure or mistakes.
In education, it's taboo to set students up for failure or mistakes. Odd. A mentor once told me that "within the scope of safety...mistakes are the best tools to educate." Sometimes, even often times, students learn best when placed in situations where there's a genuine chance of some kind of failure. Real education is a tough sell.