The Scientific Method is a method of research in which a problem is identified, relevant data are gathered, a hypothesis is formulated from these date, and the hypothesis is empirically rested.
The Scientific Method was used even in ancient times, but it was first documented by England's Sir Francis Bacon (15-61-1626) who set up inductive methods for scientific inquiry.
Scientific method is often spelled out in a simplistic stepwise fashion in textbooks as if it were some sort of recipe or set of instructions that scientists follow. It’s somewhat mythical, not something a working scientist necessarily thinks of and consciously follows step by step. But textbook authors have very limited space and must simplify.
Briefly, scientific method is not a set of instructions for research, but a habit of thought and investigation by which we gain the most reliable and objective information on how the world works—from the world of subatomic particles to the world of galaxies and beyond.
A scientist is, above all, an endlessly curious and highly disciplined person. A scientist has a question about nature. He or she conceives of a possible answer to it framed in the context of what we already know about the subject. That possible answer is called a hypothesis.
A hypothesis is useless and not really scientific at all unless there’s some conceivable way of empirically testing it—i.e., by direct observation. (Stephen Hawking even went so far as to say the string theory of particle physics isn’t really science at all, because there’s no conceivable way of putting it to an empirical test.) So a scientist’s next step is to formulate a way of observationally testing his or her hypothesis.
The test might be simple qualitative observation (not manipulating nature but just observing it with a trained and careful eye—as in much field research in animal behavior and paleontology), or it might be by experimentation (manipulating some variables to see what effect that produces on a system, as medical research, most laboratory science, and some field science).
The test should generate information—data, usually in numerical form. Next, one must ask whether those data really mean anything. Do they truly show an effect of your manipulated (independent) variables on the behavior of a system (dependent variables). This calls for tests of statistical significance—mathematical procedures of great variety, depending on the kind of data one is testing (Student’ t-test, analysis of variance, Person’s product-moment correlation coefficient, Tukey’s range test, and many others). These tests are meant to minimize subjective bias in the interpretation of data. The product of such tests is probability statements. Experimental scientists speak in probabilities, not certainties. Scientists will not say, for example, This proves that disease X is caused by virus Y, or that this new drug controls hypertension or produces leukemia remission better than the old one. Rather, they will make such statements as, There’s at least a 99.5% chance that variable X caused the observed change in variable Y. A careful scientist doesn’t assert “My findings prove ___,” but “My findings are consistent with the hypothesis that ___.”
Then, to have truly “done science,” one must make one’s findings available to the scientific community and the public—that is, publish them. Some say, if you don’t publish it, it’s not science. You’ve contributed nothing to the world’s understanding if you don’t publish. Public knowledge is the ultimate goal of science, not just the satisfaction of personal curiosity.
To publish one’s findings, one must first demonstrate that the work was important enough to warrant a share of the limited and expensive space in a recognized scientific journal; then show that one’s hypothesis and methods were sound; then show that the results were meaningful—often, but not always, by demonstration of statistical tests of confidence.
All of this entails surviving a process of peer review, in which an editor or grant agency sends your draft publication out to other experts in your field, whose job is basically to find fault (if there is any) with what you’ve done: to closely scrutinize your logic, assumptions, methods, findings, and tests of significant to see if they can find anything wrong with it, serious enough that you need to revise your paper or the journal editor should outright reject it for publication. The prestigious journals Cell, Science, and Nature reject 97% of all papers that scientists submit for publication, for various reasons including insufficient importance to warrant publication, inappropriateness of subject matter to those journals, unclear hypotheses, fallacious methods, poor analysis of the results, violation of research ethics, or poor writing.