Saturday Physics for Everyone 2015: Peter Abbamonte, "How Laws, Sausages, and Science are Made..."

> I hate the scientific method, and the reason is because that is not what happens. [...] The scientific method is based on the assumption that all of your assumptions are true. The speaker then goes on to provide an example of a cartoonishly inadequate application of modus tollens. I'm already having a bad feeling about this talk. I am well aware that the scientific method is regularly misapplied in the world of science. This is why we have to actively confront problems like p-hacking. However, I was expecting some explanation of how we can try to do better, perhaps even ways to frame the process that already takes place as a special case of the scientific method, and how to correctly analyze it in that light. Instead, I'm now expecting to be persuaded to just not worry about it: "p-hack or don't, who cares?" Side note:A proper application of modus tollens is to test an actual implication of a hypothesis; testing whether the earth is square or has an edge as a proxy for whether the earth is flat is improper because *the earth being flat does not imply that it is square or has an edge*. An implication just serves as a proxy hypothesis for the purposes of hypothesis testing, and the manner in which its' refutation leads to the refutation of the original hypothesis should be logically demonstrable, i.e. *a proof*. This is why proof-based mathematics is important to science, and I think, at least in physics, we undervalue that. The speaker is appearing to explain that the scientific method doesn't work because a particular experiment of his was bogged down in the hypothesis testing step; they failed to formulate an adequate null hypothesis. This is what he is referring to when he talks about "looking at the background", being "lost". He's lost because they failed to establish a controlled test by failing to observe the outcome in the absence of the factor under test. Hypotheses must be expressed in terms of how the absence or presence of some factor changes the outcome, this cannot be tested in the absence of a control. This is not the scientific method not working. This is just a challenging hypothesis test to formulate. This fits my previous observations about misapplication of the scientific method. Students and researchers, more often than not, operate according to some process which can be framed as a special case of the scientific method, but fail to correctly analyze the results because formulating some aspect of it is difficult and requires abstract thought, whether it be the original hypothesis, the prediction or proxy hypothesis, or the distribution of the factor under test. The main reason for ordering the formulation of the hypothesis before the test is that a hypothesis which is formulated based on the measurements taken during the test cannot be tested by those measurements. You end up testing how likely it is that those particular measurements would lead a human to that alternative hypothesis instead of testing how likely the measurements would be given the null hypothesis. The correct thing to take away from the speaker's observations is not that the scientific method is bunk, it is that testing is not the same as exploratory research, and most of the real-world process called doing science consists of exploratory research, not testing. The complex process which the speaker describes on the final slide is really just a portion of the hypothesis formulation step of the scientific method.