A very common method for testing causal relations by social scientists, especially psychologists, is the experimental method. An experiment may be carried out in a laboratory or a field setting, and it ideally begins with two or more equivalent groups, with an experimental variable introduced into only the experimental group. The researcher measures the phenomenon under study before the introduction of the experimental variable and after, thus getting a measure of the change presumably caused by the variable.
There are two common ways of setting up experimental and control groups. One is the matched-pair technique. For each person in the experimental group, another person similar in all important variables (age, religion, education, occupation, or any other variable important to the research) is found and placed in the control group. Another technique is the random-assignment technique, in which statistically random assignments of persons to experimental and control groups are made—such as assigning the first person to the experimental group and the next to the control group, and so on.
Experiments in sociology face certain difficulties (Babbie, 2017). An experiment involving thousands of people may be prohibitively expensive, and the cost factor is often the decisive issue whether or not to embark on a project. It may take years to complete such an experiment. Ethical and legal considerations prohibit the use of people in any experiments that may injure them. When people are unwilling to cooperate in an experiment, they cannot be forced to do so. Moreover, when individuals realize that they are experimental subjects, they begin to act differently and the experiment may be spoiled. Almost any kind of experimental or observational study upon people who know they are being studied will give some interesting findings, which may vanish soon after the study is terminated. Experiments with human subjects are most reliable when the subjects do not know the true object of the experiment. But the use of deception in social research poses the ethical question of distinguishing between harmless deception and intellectual dishonesty.
In law and society research, experimental methods have been used to study jury deliberation (Hans, 1992, 2006; Jonakait, 2003; Simon, 1975), the evaluation of objections in the courtroom (Koehler, 1992), allocation of scarce criminal resources (Nagel and Neef, 1977), the impact of increasing or decreasing police patrols on crime (Zimring, 1989), and the determination of the effectiveness of pretrial hearings (Zeisel, 1967). In the last example, a controlled experiment was done to find out whether pretrial hearings were time savers or time wasters. Sociologists developed a design calling for a random assignment of cases by court clerks to one of two procedures: obligatory pretrial hearing in one group of cases and optional pretrial in the control group, where it was held only if one or both of the litigants requested it. The conclusion was that the obligatory pretrial hearing did not save court time; in fact, it wasted it (Zeisel, 1967). Persuaded by the experiment, the state of New Jersey changed its rules and made pretrial hearings optional.
Many experiments, such as those dealing with juror and jury behavior (Diamond, 1997; Jonakait, 2003; Kramer and Kerr, 1989), are conducted in a laboratory situation. The widely publicized National Commission on the Causes and Prevention of Violence (1969), for example, relied heavily on the results of laboratory experiments for its final report. In one group of experiments, young children who were shown acts of violence and then later observed at play committed more acts of violence in their play than children who did not witness acts of violence. In another group of experiments, college students were told that they were participating in a “learning experiment” in which they must apply mild electric shocks at whatever level of intensity they wished to other “learners” if the “learners” made a mistake. The “learning experiment” was interrupted, and some students were shown a violent film while others were shown a nonviolent one. When the “learning experiment” was resumed, the students who saw the violent film used slightly stronger shocks on their “learners” than those who had watched the nonviolent film.
Laboratory experiments, as important as they may be in revealing insights into human behavior, achieve rigorous and controlled observation at the price of unreality. The subjects are isolated from the outside and from their normal environment. The laboratory experiment has been criticized for its unnaturalness and questioned as to its generalizability. By contrast, experimental methods that are used in nonlaboratory settings increase the generalizability of results and lend greater credence to the findings, but concomitantly increase the difficulty of controlling relevant variables.