Scientific Method: Steps, Types of Research, and Examples

What Is the Scientific Method

Doing science, in the words of the Stanford Encyclopedia of Philosophy, means observing systematically and reasoning inductively and deductively. It also means building models and testing hypotheses and theories, as far as the data allow for generalizing.

Anyone studying how the mind works, as in psychological science, applies the same framework to a harder object to observe directly: behavior and mental processes.

The Steps of the Scientific Method

Every scientific path starts with an observation. From there a question arises, trying to explain the phenomenon observed, and the hypothesis is the expected answer to that question: a provisional, testable explanation.

Next comes designing an experiment built to test the hypothesis and check whether the results match the predictions. The data collected then get analyzed, and the final step — the conclusion — states whether that data supports the starting hypothesis or not.

Practical example: an observation, such as a plant growing less in one corner of a garden than elsewhere, opens the path. The question is why it happens; one possible hypothesis is that the corner gets less light. The experiment is to move the plant to a brighter spot, keeping every other condition the same, and compare its growth after a few weeks. If the plant grows more, the data analysis supports the hypothesis; if nothing changes, the conclusion is that light, in that case, was not the cause.

Different Interpretations Through History

Francis Bacon, in the Novum Organum of 1620, proposed a method built on the methodical collection of observations and the systematic correction of the mistakes the senses can cause. Isaac Newton, in the Principia, held that his explanations came strictly from the observations he gathered, a view he summed up in the phrase hypotheses non fingo, following the same path Bacon had opened.

Karl Popper, in The Logic of Scientific Discovery of 1959, moved the center of the method to the asymmetry between confirmation and disproof: a theory earns the label scientific when an experiment could, at least in principle, prove it wrong. This approach is called falsificationism. The physicist Steven Weinberg noted, in 1995, that scientists interpret this shared set of practices each in their own way, which is why the procedure looks different from one scientist to the next.

Anyone reconstructing Darwin’s theory of evolution from fossils and field observations works along the same inductive path Bacon opened: collecting systematic observations before generalizing. Newton’s three laws of motion remain instead the best-known example of his own way of inferring from observations.

Three Types of Research

Descriptive research observes a phenomenon without manipulating it. The survey is its typical tool: it collects data by asking people to answer questions about their own behaviors and opinions. It captures well what people think and feel, but not as well how they actually behave in social situations.

Experimental research instead investigates relationships between variables to test a hypothesis. In laboratory experiments, researchers create artificial situations where they can manipulate variables, unlike descriptive research, which observes without intervening. Observing volcanoes and earthquakes is an example of descriptive research on a large scale: the phenomena get recorded just as they happen, outside any laboratory.

Qualitative and quantitative instead describes the type of data collected, an axis separate from the descriptive-or-experimental design. The in-depth comparison between qualitative and quantitative research goes into that distinction, with the criteria used to judge its rigor.

Qualitative and Quantitative Data

Qualitative data is, in the glossary of statistics, an observation whose value is expressed by a label — a category — instead of a number. Quantitative data is an observation whose value is a number, split into discrete, when it is countable like the number of correct answers, or continuous, when it sits on a scale, like a measurement of time or weight.

Qualitative data is harder to organize and tabulate than quantitative data, and it is often observed in a natural setting instead of being produced in a lab.

Anyone setting up an effective study method uses both types of data to figure out what works: a test score is quantitative, the account of how a student tackled an exercise is qualitative.