Experimentation

Primary tabs

You are viewing a wiki page. You are welcome to edit.

Twitter icon
Facebook icon
Google icon
LinkedIn icon

Experimentation is a causal research method that can aid hospitality executives in decision making. Using experiments, decision makers can draw causal conclusions about the effects of various actions being considered. For example, experiments can be used to test the effects of changes in pricing, design, and other variables on actual consumption behaviors such as the amount of money consumers spend, or the number of visits to an establishment. However, despite its potential, experimentation is under used by hospitality firms because of the belief that experimentation is synonymous with laboratory research. While it is true that most basic research is experimental research conducted in a laboratory setting, it is possible to conduct true experiments investigating applied issues outside of the laboratory. Experimentation is a method, not a location.

True experiments have three characteristics. First, experiments have at least one treatment group and one comparison group. For example, one group of people can be exposed to a proposed advertisement for a product, while other people are exposed to the currently used advertisement. The comparison group can be either a different group of people, or the same group of people studied at a different time. Second, true experiments have at least one outcome measure. For marketers, this outcome measure should be sales or another measure of actual consumption behavior, not attitudes or reactions to the manipulated treatment. Finally, true experiments must have random assignment of subjects to the treatments. This means that each subject has an equal chance of getting into any treatment group. Random assignment distributes subjects’ demo graphic and other personal differences evenly across the treatment groups, thus allowing the researcher to conclude that any differences in the outcome variable must be caused by the differences in the treatments. While random assignment to treatments can be challenging outside of the laboratory, it is not impossible. For example, restaurateurs can randomly assign patrons to different menu designs. In addition, the Internet provides many opportunities to randomly assign people to different promotional messages.

When designing and interpreting the results of experiments, marketing researchers must consider three types of validity that can affect the conclusions drawn from the results. These types of validity are statistical inference validity, internal validity, and external validity. Statistical inference validity insures that chance does not explain the observed differences among the treatment groups. Statistical inference validity can be improved by using appropriate alpha levels (the ‘p value’) in statistical analyses, using large sample sizes, and decreasing the variability in the subjects and/or the way the treatments are delivered. Internal validity refers to the strength with which the researcher can conclude that the treatment caused the changes in the outcome variable. Confounded treatments (treatments that differ in more ways than those intended) are the threat to internal validity. Random assignment to treatments and keeping the people who deliver the treatments blind to the treatment are the best ways to increase internal validity. Finally, external validity is the extent to which the results of the experiment apply to the real world. External validity is threatened by differences between the real world and the people, treatments, outcomes, and context used in the research. The utility of marketing research is largely dependent upon the strength of external validity. Researchers can improve external validity by using samples, contexts, and outcome measures that are representative of the ones in the real world.

Descriptive and experimental research have a synergistic relationship. Ideally, decision makers can use descriptive research surveys, interviews, and focus groups to develop a variety of potential courses of action. Then, these potential courses of action can be evaluated using experimentation. Experimentation is the ideal research tool for drawing causal conclusions about the effects of managerial actions on employee and consumer behaviors.

Reference

Lynn, A., & Lynn, M. (2003). Experiments and quasi experiments: methods for evaluating marketing options. Cornell Hotel and Restaurant Administration Quarterly, 44(2), 75 84.

Add new comment