Our understanding of several communication phenomena has benefited extensively from informative reports of “natural” experiments. Natural experiments are infrequently detailed in the social and behavioral sciences, however, because they result only when naturally arising circumstances make it possible to separate and examine typically confounded phenomena. Like field experiments, natural experiments occur in venues where the researcher has limited control over environmental or situational conditions. Distinctive in the natural experiment, moreover, randomization is impossible and the researcher exercises no influence over the independent variable. That is, the independent variable is not manipulated but rather varies (often presence vs absence) due to a naturally occurring event or situation.
Natural experiments, consequently, are not actually experiments; but, instead, are observational studies. In a natural experiment, the researcher identifies and anticipates the progression of an evolving phenomenon, isolates and defines unique characteristics of the incident or event to articulate different levels of the independent variable, and develops systematic techniques to observe outcomes (i.e., dependent variables). Quasi-experimental designs typically provide the logical structure for natural experiments. The strongest evidence emerges from variants of the pre-test/post-test nonequivalent control group design, with systematic observations recorded both before and after the onset of the naturally occurring phenomenon (i.e., independent variable). It is important to recognize that causation can never be unequivocally determined through natural experiments. Nevertheless, this alternative research method can inform theory development by providing data unattained through other means.
A classic example of a natural experiment is a study initiated in 1973 by Williams (1986) and her colleagues. They conducted an elaborate investigation of the impact of television – or more precisely, the absence of television – on a non-isolated working-class community in Canada. “Notel,” the name given the town, received no television programming. Two other communities that were demographically and socio-economically matched with Notel, but which received either only a single Canadian Broadcasting Corporation (Unitel) television channel or television content from four channels (Multitel), were identified. The three communities were operationalized as three levels of a TV-content independent variable.
Using the pre-test/post-test nonequivalent control group design, Williams and her associates first surveyed (pre-test or O1) the residents of the three communities about one year prior to the introduction of television in Notel and then surveyed (post-test or O2) the communities two years later. In each community adults and children were assessed on a number of dependent variables including aggressive behavior, personality characteristics, sex-role perceptions, cognitive and creative task performance (e.g., intelligence, reading ability, and vocabulary), and leisure-time activities. Television programming was also content analyzed for aggressive behavior and sex-role portrayals.
Williams and her associates found that television viewing significantly affected other everyday activities concluding that “television affects viewers negatively in a variety of areas via displacement” (p. 395) since each day viewers spent several hours of their limited discretionary time watching TV. They observed that activities that could not be multitasked with TV use (e.g., reading and creative activities), especially those conducted outside the home (e.g., participation in community activities), were most adversely displaced. Williams and her colleagues also found, consistent with prior laboratory experiments, that television was a prominent source of social learning, particularly for children, with significant content effects observed in both aggressive behaviors and sex-role perceptions. Finally, they found that the presence of television, not the number of channels available, was the origin of television effects. The effect of television on residents of Notel appeared significantly different from that observed in Unitel and Multitel; but residents in Unitel and Multitel, despite variations in the source and number of television channels available, tended to be similar.
More recently, Weimann (1996) conducted a natural experiment exploring how the rapid diffusion of cable television access affected television viewers in Israel. Specifically, he employed the pre-test/post-test nonequivalent control group design to compare a sample of residents with newly established cable access (cabled residents) with a matched sample unreached by cable television (uncabled residents). Weimann found significant behavioral, cognitive, and affective differences. Residents with cable reported substantial increases in television viewing, and more positive feelings toward television, despite the fact that increased TV viewing displaced time spent with other activities. Cabled residents also reported considerable uneasiness, discomfort, and shock as “family feuds,” conflicts, and disagreements emerged following adoption of cable television in their homes.
Finally, Snedeker et al. (2007) provide an excellent example of a natural experiment structured on the time series design in their work examining language development among children of international adoption. Parental reports and speech samples were collected from 27 preschool children repeatedly for 3 to 18 months after US parents adopted them from China. These children, the authors observed, showed the same developmental patterns in language production as much younger monolingual infants (matched for vocabulary size). Initially, their vocabularies were dominated by nouns and their utterances were short. The children, at later stages, displayed vocabularies that were more diverse and produced longer utterances with more grammatical morphemes.
- Schutt, R. K. (2004). Investigating the social world: The process and practice of research, 4th edn. Thousand Oaks, CA: Pine Forge.
- Snedeker, J., Geren, J., & Shafto, C. L. (2007). Starting over: International adoption as a natural experiment in language development. Psychological Science, 18, 79 –87.
- Weimann, G. (1996). Cable comes to the Holy Land: The impact of cable TV on Israeli viewers. Journal of Broadcasting and Electronic Media, 40, 243 –257.
- Williams, T. M. (1986). The impact of television: A natural experiment in three communities. Orlando, FL: Academic Press.