experiment

An experiment is a systematic procedure carried out to test a hypothesis, establish causal relationships, or demonstrate a principle by deliberately altering one or more variables and observing the effects. Experiments are characterized by the manipulation of an independent variable, measurement of a dependent variable, and attempts to control or randomize extraneous variables. They contrast with observational studies, where researchers do not intervene.

Common types include controlled experiments (often conducted in laboratories) with a comparison group; field experiments conducted in real-world settings; natural experiments where conditions approximate an experiment but are not fully controlled; and randomized controlled trials, in which participants are randomly assigned to treatment or control conditions. Quasi-experiments resemble true experiments but lack random assignment.

Key design features include operational definitions, replication, and randomization. Blinding or double-blinding reduces bias. Controls, such as placebos or sham interventions, help isolate the effect of the independent variable. Validity concerns include internal validity (causal inference within the study) and external validity (generalizability).

The typical workflow comprises formulating a testable question, proposing a hypothesis, planning and conducting the experiment, collecting and analyzing data, and reporting results. Ethical considerations may involve informed consent, safety, privacy, and welfare, particularly in research involving humans or animals, often overseen by an ethics review board.

History traces experiments to early science, with notable contributions from figures such as Galileo and Bacon, and they remain central to the scientific method as a means of testing ideas and building evidence.