In much scientific work, researchers construct an experiment in which they assign subjects to experimental or control groups. Sometimes, instead of constructing an experiment, researchers may find one that was created by natural or social events beyond the researchers' control. The study by Kark et al1 in this issue of JAMA is one example of a "found" experiment. In this study the authors document an abrupt increase in mortality after the first Iraqi missile attack on Israel in January 1991. Found experiments of this sort have three characteristics: (1) the independent variable (eg, the presence or absence of a missile attack) shows a marked change in a short time; (2) the dependent variable (eg, mortality) also displays a large, abrupt change in level; and (3) the investigation seeks a short-term link between independent and dependent variables. For example, the present study found a sharp increase
See also p 1208.
Phillips DF, Halebsky SC. The Epidemiology of Found Experiments. JAMA. 1995;273(15):1221. doi:10.1001/jama.1995.03520390081038