So I Wrote A Book On The Scientific Method

Thank you. Yes, I'm sure I'll get a lot of help once it's accepted, but the question is whether it will be accepted. I think some of the comments I've got will be useful in this respect.I'm dealing, you see, with a difficult subject. And it is almost impossible to dumb it down. You can approximately explain the content of science to people, skipping the difficult bits, and they end up knowing more about the field than they did; but if you give someone an approximate pop-sci version of the scientific method they often actually end up dumber than they started. And yet I wish this to be a popular book! My only hope, then, is to tackle the difficult bits but to write it so lucidly that they don't strike the reader as difficult.The trouble is, I don't find any of it difficult. That's the biggest problem in exposition of a technical subject: by the time someone is expert enough to do it, they've forgotten what it's like to be a complete tyro. I need feedback from people who aren't me, and for whom my ideas don't seem completely obvious and natural.

Put all that in your cover letter. (I'm confident it will be accepted by the right publisher, though you may have to search some).Edited by Faith, : No reason given.

To use an analogy, it is very difficult to learn what a good golf swing looks like by reading the USGA rule book.It might be helpful to give examples of what good science looks like. More importantly, it might help the readers to understand the basic skepticism that underlies the scientific method. For me, the scientific method boils down to "What if I am wrong?". What if we are wrong about the existence of an aether that propogates light? Well, you do the Michelson-Morley experiment. What if we are wrong about mutations being random? Do the Luria-Delbruck fluctuation experiment, or the Lederberg plate replica experiment.The whole point is to do an experiment that will tell you if you are wrong. If you aren't wrong, you look for another way to prove yourself wrong. The art of science is determining the point at which you are probably not wrong. The next step is telling your colleagues about your idea so they can try and prove you wrong.What I think most lay people miss is how much work goes into validating methods and finding valid controls. For example, there was a ton of work done at the LHC to verify that it was producing valid observations before they were ready to conduct experiments. None of that shows up in subsequent papers. If your controls work they often only get one bar in a huge graph, or half a sentence in the Results section. However, they are the most important aspects of any experiment.That would be my advice. I don't know if it is helpful or not, but it's something.

Oh, I do that quite a lot. It annoys me when people don't. Have you read Popper's Logic of Scientific Discovery? It's like he took an oath never to use the words "for example". Oh yes, I talk about that a lot. Well, there I disagree, this is one of the things that I think people shouldn't say, it gives the wrong impression.I point out, first of all, that it is not psychologically true that scientists are trying to prove themselves wrong. Mostly, they're trying to prove themselves right. (Ask a cancer researcher what he's working on right now, he's not going to say: "I'm trying to prove that a method of treatment which I believe to be promising is in fact ineffective", that's not how he thinks of his job.)Does this matter? No. Because to try and prove yourself right, you put your ideas to a test which you hope they'll pass. Now, note that for this to be a test at all, it would have to cast doubt on your ideas if they failed this test. Which means that if you were trying to prove yourself wrong, you'd still want to perform the test, but hoping that your ideas would fail.So trying to prove something right, and trying to prove it wrong, would involve putting it to just the same tests. If I think that the boiling point of mercury is 150 C, and you think it isn't, how would you try to prove me wrong? You'd get some mercury, a means to heat it, a thermometer ... now, how would I try to prove that I'm right? Exactly the same way.(You mention the Michelson-Morley experiment. But it was not, in fact, done to disprove the existence of the ether, was it?)So (I say) trying to prove that you're right isn't the opposite of trying to prove you're wrong, because done properly these involve following just the same procedure. Rather, the opposite of both of them is apathy, laziness, and complacency.

I hate it when authors don't give examples. Maybe it's just the way I think, but I am able to understand a concept much quicker if I work through a problem. Perhaps I am too focused on the technical side of doing science. In my experience, the actual experimental samples are a tiny, tiny fraction of the work I do compared to validating equipment, assays, methodologies, and controls. I personally find it better to be your own critic and try to prove your own results wrong at every step. What you get in return is data that you can trust and data with the least amount of personal bias possible.As the saying goes, "Your mileage may vary".

Yes, but my point stands, you're actually trying to prove you're right by trying to prove you're wrong. This is why you call it, for example "validating equipment". You're doing a lot of tests that would prove the equipment works --- if it passes --- or doesn't work --- if it fails. You want, as it happens, to prove that the equipment works; but if you wanted to prove that it doesn't, you'd do the same tests.

Oh come on now, guys. You agree.It's the issue of finding a good way to word it so that the idiots can get a clue. Either side has good points, but what is the best way to phrase this important concept? Danged if i know. It was too obvious to me from the earliest i can remember.

I'd say that the scientist with the original idea is trying (mostly) to prove it right. It's (mostly) the job of other scientists to attempt to prove him wrong.(Although the fact that he knows that his colleagues/competitors are going to go through his work hoping to find a flaw is a good incentive for him to do the same first.)

I understand your position. What I am talking about is more of a mind set, an attitude you adopt when setting up experiments and evaluating data. To use my previous analogy, you are discussing the USGA rule book while I am talking about visualizing a good swing.To be fair, perhaps this mind set is wrong, or unique to me. Just thought I would share my own experience in the sciences in case it could be of help.

For what it's worth - probably not much - i'm currently doing some proper, scientific research uncovering and analysing primary data.I'm trying to prove something that I believe to be correct but I don't know which way the work will go. It may prove totally inconclusive (likely) or even that my hypothesis is wrong.I'm watching myself very closely whilst doing this.When I get a data set that is helpful to the hypothesis I'm enthusiastically entering it into my model but when I get one that goes the wrong way I'm trying to find some reason to exclude it. So far I've been able to resist the temptation to ignore negative results but I'm understanding better the pressures to do so.If I was a professional scientistTM and my career and reputation depended on this, day in day out, it would be a hell of a temptation to just loose some of this data - no-one would ever know. But maybe you get used to it. On the other hand, it must make you take fewer risks; why research something that you suspect might not work?

Well yes, we're not bickering.The question is, as you say, what an idiot might think. Now imagine an idiot who's heard people saying that scientists try to prove that they're wrong, and this is the essence of the scientific method. Now suppose he reads a statement such as "Eddington tried to prove the Einstein was right by measuring the deflection of starlight." Being an idiot, he's going to think "Oh, that idiot Eddington didn't understnd the scientific method at all ... unlike me ... I'm smart."

No, but it's the same thing as trying to prove you're right.If you think to yourself: "I really really want to prove that I'm right about X, incontrovertibly, with an iron-clad proof, so that the opposition will be crushed and no-one will argue against it", then being rational, your next thought would be "And how might people argue against it?" They might say you haven't measured W or calibrated Y or taken into account Z. Very well, you will measure W, calibrate Y, and take Z into account, so as to make your proof iron-clad. At that point you're behaving exactly like the most adamant and severe of your critics would behave, you've appointed yourself captain of the opposing team.

Well, we can assume that lying about stuff is outside the scientific method.If you publish an accurate account of your methodology then again my point stands. If your method involves cherry-picking your data, then the more you cherry-pick, the less chance it stands of proving that you're wrong. But for exactly the same reason, it becomes less and less convincing as a proof that you're right. If you wanted an iron-clad proof that you're right, you'd want to be able to point to a methodology that was scrupulously objective and gave you no chance to exercise any bias. The better the proof that you're right, the better the chance that it would have had to prove you were wrong if you were in fact wrong.

Correct. That is what you would do if you were a machine. I'm saying that it's a bit more complicated than that. Human emotion, ambition and politics may taint the pure nature of the scientific, objective vision more than we might hope or expect. Particularly in disciplines that aren't binary. Just musing.....

I do have a chapter on scientists being human and the problems this causes. I advocate replacing them with my patented line of titanium kill-bots ... wait, no, that was my plan for how England should win the World Cup. Mind you, you can't really go wrong with titanium kill-bots.