Strategies or logical practices, often with iterative steps, used in scientific research. The process of doing science (Fig. 1) is not a "cookbook" set of rules. However, a valid scientific investigation typically involves systematic collection of data and, usually, formulation and testing of hypotheses based on that data. Indeed, the goal of a true scientific study is to ascertain whether a hypothesis is true to some degree.
Such enquiries into the natural world may include the following steps, often collectively referred to as "the scientific method":
- Identification of a problem or question about the natural world
- Precise formulation or reformulation of the problem
- Examination of background knowledge, including primary scientific literature, in a search for items or ideas that might help solve the problem
- Statement of a testable hypothesis about the solution to the problem or answer to the question that is compatible with the bulk of existing knowledge, including enumeration of testable consequences of this hypothesis
- Design of an empirical (observational or experimental) test of the hypothesis or a consequence of it
- Actual empirical test of the hypothesis, involving a search for both favorable and unfavorable evidence (examples and counterexamples)
- Critical examination and statistical processing of data (for example, calculation of average error and elimination of outlying data)
- Evaluation of the hypothesis in light of its compatibility with both the background knowledge and the fresh empirical evidence
- If test results are inconclusive, design and performance of new test, possibly using different techniques or procedures
- If test results are conclusive, acceptance, modification, or rejection of the hypothesis
- If the hypothesis is acceptable, checking whether its acceptance forces some change (enrichment or correction) to the background knowledge
- Identification and tackling of new problems raised by the confirmed hypothesis
- Repetition of the test and reexamination of its possible impact on existing knowledge
The steps outlined above are not a recipe for making original discoveries or inventions. They do not prescribe the pathway that scientists must follow to attain success, nor are they a substitute for creativity and resourcefulness. The actual pathway of scientific research is messy and partly at the mercy of unforeseen accidents, both lucky and unlucky. Moreover, this pathway depends on subject matter as well as on the investigator's imagination and experience. Thus, certain steps may become iterative, depending on the outcome at each milestone. The nucleus of the scientific method, however, is confrontation of an idea (hypothesis) with the facts it refers to, regardless of the source of the idea in question. In sum, the scientific method is a means for checking a hypothesis for truth rather than for finding facts or inventing ideas. See also: Hypothesis; Literature of science and technology; Philosophy of science; Science
The scientific method has not always been understood in this fashion. For instance, Francis Bacon (Fig. 2) and many others thought that the scientific method was a simple and guaranteed recipe for discovery and invention. However, no one has come up with any algorithms for having original ideas. Because the scientific method involves facts, not opinions, questions posed by investigators can only be resolved by empirical investigations. Once a promising problem has been identified, it must be stated or perhaps reformulated in a precise way, and a tentative plan to work on it must be formulated. The invention of new hypotheses and new methods also takes inspiration and luck, in addition to hard work and discussion with colleagues. The hypothesis must be checked to find out whether it is true to some degree, and the method must be tried out more than once to ascertain whether it accomplishes what it purports to do and, if so, whether that method is better than rival techniques. It is here, in checking, that the scientific method plays a decisive and distinctive role. In short, the scientific method is the way that scientists proceed to check ideas and techniques, not invent them. See also: Empirical method
An alternative interpretation of the scientific method is as a sequence of problems. The investigator asks questions that seek to find what the result implies or suggests; how the result could be corroborated independently, that is, by alternative procedures; whether the new result is more precise and plausible than results obtained by alternative procedures; if so, what that result implies or suggests; and, if not, what may have to be altered in the preceding operations. See also: Problem solving (psychology)
Whether the scientific method can legitimately and fruitfully be used in the social sciences, such as economics, psychology, and sociology, has been a subject of controversy. The history of these sciences shows conclusively that the scientific method has been fruitful wherever it has actually been employed.
Often, the scientific method involves experimentation—that is, deliberate, controlled modification of some properties (also called factors or variables). Such experiments involve design and operation of an investigational setup, which often includes one or more measuring instruments. When the objects, or variables, of an experiment are similar, the investigator(s) compare before and after the variable or variables in question have been altered. When the objects of an experiment exhibit significant variations, as is typically the case with living organisms, a large collection or group must be studied. Scientists divide such collections into two roughly equal parts: the experimental group and the control group. To avoid bias, the members of each collection are selected at random. In the case of humans, control groups are given placebos. (In more refined experiments, the control group is split into two, only one of which is given a placebo, and the experimenter does not know beforehand who will be given what. This is called a double-blind experiment.) Only the members of experimental groups are subjected to the stimulus whose effects the experimenter wishes to determine. See also: Experiment; Instrument science
The simplest experimental design involves the variation of one variable at a time. More complex experimental designs allow for simultaneous variation of two or more variables. After the stimulus has been applied, the variables of interest are observed or measured in the two groups. If a difference is observed, a statistical significance test is applied to find out whether the difference is genuine or due to small individual differences or to random errors. The procedure is usually repeated by the same observer or by an independent investigator to check for possible errors in design, execution, or interpretation. See also: Factor analysis; Statistics
Science versus nonscience
Because use of the scientific method is one characteristic of scientific research, its absence is a sure indicator of nonscience or pseudoscience. In other words, a discipline where the scientific method plays no role is not a science. Thus, such fields as theology, literary criticism, psychoanalysis, astrology, and palmistry can hardly be regarded as scientific.