ericmazur / PnPbook

Tracking of typos, errors, and improvements for "The Principles and Practice of Physics"
0 stars 1 forks source link

Scientific method/hypothesis #76

Open ericmazur opened 9 years ago

ericmazur commented 9 years ago

From @JohnDenker: Concerning "the scientific method" -- The discussion in general and the diagram in particular greatly overstate the role of hypothesis testing.

It is a fundamental principle of sound planning that one should plan for all of the plausible scenarios. This is standard procedure in all branches of business, government, et cetera. When the Boy Scouts plan a campout, they plan for the possibility that it will rain, even though they hope and expect that it won't.

In any experiment worthy of the name, there are multiple possible outcomes. Otherwise it's a demonstration or an exercise, not really an experiment, not informative, and certainly not research.

Also I'm not at all sure I agree that "hypothesis" should be defined as a "tentative explanation". In its broadest and most useful meaning, it is just something to be considered. It might be probable, improbable, known true, or possibly even provably untrue. For example, in a proof by contradiction, the hypothesis is expected to be proven untrue.

Since the definition of "hypothesis" is sometimes disputed, I tend to avoid it and speak about /scenarios/ or /outcomes/ instead.

Constructive suggestion #1: In the diagram, there should be a lead-in arrow, leading from empty space to the "observations" box. This is to indicate that hypotheses and predictions are absolutely not a prerequisite for making scientific observations. The text is reasonably clear on this point, but the diagram is not as supportive as it could be. This point is often taught wrongly, so it is worth emphasizing.

By way of illustration, when Lewis and Clark set out, they were not testing the hypothesis that they would find bighorn sheep in the Rocky Mountains ... obviously not, because they had never heard of either of those things.

One should be careful to distinguish how science is /done/ from how science is /published/. Scientific publications often list which hypotheses are consistent with the data, and which are ruled out. However, these lists are generally prepared with the benefit of 20/20 hindsight, long after the initial discoveries were made.

Focusing on testing hypotheses (or worse, a single hypothesis) gives students a false and constrictive notion of how science is done. It leaves no room for creativity, originality, or serendipity.

Hitting a home run is hard enough. Let's not require students to predict exactly where in the bleachers the ball will land. It is /not/ necessary to make such a prediction. Let's do the experiment and see where it lands. Let's consider all of the plausible outcomes.

In my all-time favorite science-fair project, there was no hypothesis at all, not even in hindsight. It was a survey. Details available on request.

Also note that for reasons of safety if nothing else, it is vital for the experimental plan to consider all of the plausible outcomes. This absolutely requires /not/ focusing on a single hypothesis.

Constructive suggestion #2: Whenever discussing how science is done, the singular form of the word "hypothesis" should be a red flag. It is almost always better to write plural hypotheses, scenarios, and/or outcomes.