

The scientific method refers to ways to get facts, correct errors and mistakes, and test theories.
The Oxford English Dictionary says that scientific method is: "a method or procedure that has characterized natural science since the 17th century". It is observation, measurement, and experiment, plus the formulation, testing, and modification of hypotheses.
A scientist gathers empirical and measurable evidence, and uses sound reasoning.
New knowledge often needs adjusting, or fitting into, previous knowledge.
What distinguishes a scientific method of inquiry is a question known as the criterion'. It is an answer to the question: is there a way to tell whether a concept or theory is science, as opposed to some other kind of knowledge or belief?
There have been many ideas as to how it should be expressed. Logical positivists thought a theory was scientific if it could be verified (proved true).
Karl Popper thought this was a mistake. He thought a theory was not scientific unless there was some way it might be refuted (shown to be wrong).
On the other hand, Paul Feyerabend thought there was no criterion. For him, "anything goes", or whatever works, works.
Scientists try to let reality speak for itself. They support a theory when its predictions are confirmed, and challenge it when its predictions prove false. Scientific researchers offer hypotheses as explanations of phenomena, and design experiments to test these hypotheses. Since big theories cannot be tested directly, it is done by testing predictions derived from the theory. These steps must be repeatable, to guard against mistake or confusion by any particular experimenter.
Scientific inquiry is generally intended to be as objective as possible. To reduce bias, scientists publish their work, and so share data and methods with other scientists.
Science and things that are not science (such as pseudoscience) are often distinguished by whether they use the scientific method. One of the first people to create an outline of the steps in the scientific method was John Stuart Mill.
There is no one scientific method. Some fields of science are based on mathematical models, such as physics. Other fields, such as social science, have rough theories. They rely more on patterns that emerge from their data. Sometimes scientists focus on testing and confirming hypotheses, but open-ended exploration is also important. Some scientific fields use laboratory experiments. Others collect observations from real-world situations. Many areas of science are quantitative, emphasizing numerical data and mathematical analysis. But some areas, especially in social science, use qualitative methods, such as interviews or detailed observations of human or animal behavior. Focusing too much one kind of method can lead us to ignore knowledge produced by other methods.
Some textbooks focus on a single, standard "scientific method". This idea of a single scientific method is based on experimental, hypothesis-testing. It doesn't apply very well to other areas of science. It is often written as a number of steps:
Although written as a list, scientists may go back and forth between different steps a number of times before being satisfied with the answer.
Not all scientists use the above "scientific method" in their day-to-day work. Sometimes the actual work of science looks nothing like the above.
Let's say we are going to find out the effect of temperature on the way sugar dissolves in a glass of water. Below is one way to do this, following the scientific method step by step.
Does sugar dissolve faster in hot water or cold water? Does the temperature affect how fast the sugar dissolves? This is a question we might want to ask.
One simple experiment would be to dissolve sugar in water of different temperatures and to keep track of how much time it takes for the sugar to dissolve. This would be a test of the idea that the rate of dissolving varies according to the kinetic energy of the solvent.
We want to make sure to use exactly the same amount of water in each trial, and exactly the same amount of sugar. We do this to make sure that the temperature alone causes the effect. It might be, for example, that the ratio of sugar to water is also a factor in the rate of dissolving. To be extra careful, we might also run the experiment so that the water temperature does not change during the experiment.
This is called "isolating a variable". This means that, of the factors which might have an effect, only one is being changed in the experiment.
We will do the experiment in three trials, which are exactly the same, except for the temperature of the water.
One way that makes it easy to see results is to make a table of them, listing all of the things that changed each time we ran the experiment. Ours might look like this:
| Temperature | Dissolving time |
|---|---|
| 1 °C | 30 min |
| 20 °C | 15 min |
| 50 °C | 4 min |
If every other part of the experiment was the same (we did not use more sugar one time than the other, we did not stir one time or the other, etc.), then this would be very good evidence that heat affects how fast sugar is dissolved.
We cannot know for sure, though, that there is not something else affecting it. An example of a hidden cause might be that sugar dissolves faster each time more sugar is dissolved into the same pot. This is probably not true, but if it were, it could make the results exactly the same: three trials, and the last one would be fastest. We have no reason to think that this is true at this time, but we might want to note it as another possible answer.
Good science is repeatable. but that can be difficult or impossible. Experiments on people is an example of this.
In psychology (in particular, social psychology) and in medicine, there can be failures of replication (cannot be repeated).
Elements of scientific method were worked out by some early students of nature.