Hawthorne Effect

From 1924 to 1927, Elton Mayo of the Harvard Business School, together with Fritz J. Roethlisberger and William J. Dickson, carried out a series of experiments with the level of illumination in a factory called the Hawthorne works of the Western Electric Company in Illinois. After informing the workers that they would be observed to assess their level of productivity, the level of illumination was varied in the factory. In some cases, during the experiment a person associated with the research remained on the factory floor. The researchers expected to find that illumination correlated with increased productivity; however, worker productivity increased throughout the experiment regardless of the level of light. In fact, in one experiment, the level of illumination was decreased steadily; productivity increased until the lights were so low that the workers complained that they were unable to see well enough to work.

The findings of these studies have been widely reported as proof that people change their behavior whenever they know that they are being observed. Furthermore, there is a suggestion that when people feel included in the decision process (as they were in some of the Hawthorne studies), they are empowered and tend to work harder.

The Hawthorne studies have been criticized for poor experimental design. Not all the experiments included a control group, some studies involved very small numbers of workers, and worker turnover may have influenced the results. Portions of the Hawthorne effect have also been attributed to an observer effect, because of the experimenter who remained on the factory floor. While these criticisms are valid, the Hawthorne effect itself continues to be observed in a variety of settings. It has been cited as an explanation for results of studies in widely divergent areas, including patients’ perceptions of postsurgical recovery and quality of life, effectiveness of training on reduction of infection rates in day care centers, and the impact of repeated assessments of smoking on the rates of smoking among adolescents.

The Hawthorne effect should certainly be considered when designing an epidemiologic trial. For example, randomization to experimental and control groups will help control the tendency of people to behave differently while in the study. The schema for the control group should also be as similar to the experimental groups as possible; if a placebo is used in this group, it should be formulated to look similar to and be administered in the same way as the experimental drug. These components of the study design would help address the tendency of patients in the trial both to be more compliant in their medication use during the study than they would be in their daily lives and to report improvements in their symptoms—both of which are possible results of a Hawthorne effect and thus would be observable even in the control group. Because of this tendency, the Hawthorne effect is often equated with the widely publicized ‘placebo effect.’