Research utilizing experimentation is increasingly being conducted in venues outside the research laboratory. Such projects, when they involve the manipulation of an independent variable in realistic circumstances, are called “field experiments.” Natural experiments, involving research conducted in realistic circumstances where the researcher does not manipulate the independent variable, are discussed elsewhere.
Distinction From Laboratory Experiments
Field experiments provide an excellent approach for studying both theoretically derived hypotheses and problems of practical application. Conceptually, the differences between the laboratory experiment and the field experiment are slight; ideally, both are structured on one of the true experimental designs and consequently incorporate randomization and manipulable experimental treatments or interventions (i.e., independent variables) as fundamental components.
Similarities in the two experimental contexts diverge considerably, however, in terms of experimental control, with the researcher’s ability to guard against circumstances that might provide alternative or rival explanations, threatening the validity of causal explanations, frequently challenged in field experiments. Yet, field experiments, because they are undertaken in circumstance not radically different from everyday life, can afford the researcher greater external validity. Following the initial experimental treatment, for example, research participants in field experiments typically continue functioning in their everyday social settings with little investigator interaction until outcomes assessment (i.e., dependent measures). This can significantly reduce the reactive or interactive influence on subsequent outcomes resulting from participants’ awareness of the research procedure and enhance external validity.
At the same time, however, undertaking experimentation in the field can involve complications rarely seen in laboratory experiments. Field experiments, for instance, typically entail a significantly longer time frame (e.g., weeks and months rather than hours and days) and often engage a substantially larger number of research participants. They frequently require a variety of exigencies and preliminaries, such as establishing contacts, gaining cooperation, and securing permissions, which can strain research resources and researcher patience. The process of identifying eligible research participants in field experiments, for example, can be difficult, and the failure to retain research participants for follow-up (i.e., outcomes assessment) can be a serious threat to the generalizability of research results. Additionally, field experimentation commonly occurs in settings permeated with systematic and random noise where achieving an adequate degree of measurement precision or accuracy can be difficult. Some threats to internal validity (e.g., compensatory rivalry and the Hawthorne effect) can be particularly problematic in field experiments.
Examples Of Field Experiments
Field experiments, although not prominent in the communication research literature, have appreciably informed our understanding of important phenomena across diverse areas, including interpersonal communication (Fennis et al. 2006), language and communication disorders (Cohen et al. 2006), political communication (Gerber & Green 2001), and media effects (Robinson & Borzekowski 2006). Generally, field experiments appear most commonly in health communication research, with such projects typically operationalized as randomized controlled trials. Guidelines and practices incorporated in the randomized controlled trial, which is a refinement of the basic pre-test/post-test control group design, assist the researcher in overcoming many of the common limitations encountered in field experiments (Moher et al. 2001). Consequently, the randomized controlled trial, if effectively implemented, can yield the strongest evidence of causality of all research undertaken in realistic environmental and situational circumstances.
An informative example of the application of key communication concepts in a field experiment is provided by a study of an HIV sexual risk-reduction intervention by DiClemente and Wingood (1995). This single-blind, randomized controlled trial employed a communication and social skills enrichment program in an intervention (i.e., experimental treatment) to enhance the consistency of condom use for HIV prevention. A sample of 128 sexually active, heterosexual, African-American women aged 18 through 29 years were recruited using street outreach techniques in an economically disadvantaged neighborhood of San Francisco, CA. The women were randomly assigned to one of three groups: (1) a communication and social skills intervention, conducted over five sessions, which emphasized communication skills, ethnic and gender pride, proper condom use skills, and development of partner norms supportive of consistent condom use; (2) an HIV risk-reduction education program administered in a single session; or (3) a control condition in which administration of the HIV riskreduction education program was delayed three months until after outcomes assessment (i.e., dependent variables).
The findings revealed that, compared with the delayed HIV education control condition, women in the communication and social skills intervention demonstrated increased consistent condom use, greater sexual communication, assertiveness, and self-control, and increased partners’ adoption of norms supporting consistent condom use. No significant differences in the outcome variables were observed between the HIV risk reduction education condition and the control condition. Overall, this field experiment demonstrates that community-based HIV risk-reduction programs that provide communication and social skills training within a gender and culture-sensitive framework can effectively enhance consistent condom use and promote HIV prevention.
Another instructive example of a field experiment is provided by the work of Vaughan and his colleagues (2000; Rogers et al. 1999), who explored the impact of an entertainment education radio program on HIV/AIDS prevention in Tanzania. The project employed a radio soap opera, Twende na Wakati (“Let’s Go with the Times”), as the communication intervention. The program used negative, transitional, and positive behavioral role modeling to promote HIV prevention, family planning, gender equity, and other health themes, and was broadcast in Swahili, the national language of Tanzania, twice per week for 30 minutes from July 1993 through the end of 1999.
The field experiment was structured on a variant of the pre-test/post-test nonequivalent control group design, with restricted access to the program within a large geographic region of Tanzania during the first 2 years of the project providing the control group. Outcome measures were assessed through five personal interview surveys conducted annually. Each survey, the first of which was initiated just prior to the first program broadcast, involved a unique randomly selected sample. The survey instrument tapped personal characteristics, exposure to and perceptions of Twende na Wakati, knowledge and attitudes about HIV/AIDS, and HIV/AIDS preventive behaviors practice.
The findings revealed that Twende na Wakati positively influenced behaviors facilitating HIV/AIDS prevention, significantly reducing the number of sexual partners reported by both men and women and significantly increasing condom adoption. The radio soap opera influenced these behavioral changes through several intervening variables, including increased self-perception of HIV/AIDS contraction risk, enhanced self-efficacy for HIV/ AIDS prevention, and more extensive interpersonal communication about HIV/AIDS.
- Cohen, W., Hodson, A., O’Hare, A., et al. (2005). Effects of computer-based intervention through acoustically modified speech (Fast ForWord) in severe mixed receptive–expressive language impairment: Outcomes from a randomized controlled trial. Journal of Speech, Language, and Hearing Research, 48, 715 –729.
- DiClemente, R. J., & Wingood, G. M. (1995). A randomized controlled trial of an HIV sexual riskreduction intervention for young African-American women. Journal of the American Medial Association, 274, 1271–1276.
- Fennis, B. M., Das, E., & Pruyn, A. T. H. (2006). Interpersonal communication and compliance: The disrupt-then-reframe technique in dyadic influence settings. Communication Research, 33, 136 –151.
- Gerber, A. S., & Green, D. P. (2001). Do phone calls increase voter turnout? A field experiment. Public Opinion Quarterly, 65, 75 – 85.
- Kerlinger, F. N. (1986). Foundations of behavioral research, 3rd edn. New York: Holt, Rinehart and Winston.
- Moher, D., Schulz, K. F., & Altman, D. (2001). The CONSORT statement: Revised recommendations for improving the quality of reports of parallel-group randomized trials. Journal of the American Medical Association, 285, 1987–1991.
- Robinson, T. N., & Borzekowski, D. L. G. (2006). Effects of SMART classroom curriculum to reduce child and family screen time. Journal of Communication, 56, 1–26.
- Rogers, E. M., Vaughan, P. W., Swalehe, R. M. A., Rao, N., Svenkerud, P., & Sood, S. (1999). Effects of an entertainment-education radio soap opera on family planning behavior in Tanzania. Studies in Family Planning, 30, 193 –211.
- Vaughan, P. W., Rogers, E. M., Singhal, A., & Swalehe, R. M. (2000). Entertainment-education and HIV/AIDS prevention: A field experiment in Tanzania. Journal of Health Communication, 5(suppl.), 81–100.
- Wimmer, R. D., & Dominick, J. R. (2003). Mass media research: An introduction, 7th edn. Belmont, CA: Wadsworth / Thomson Learning.