Rigor in Behavioral Experiments: A Basic Primer for Supply Chain Management Researchers (Report‪)‬

INTRODUCTION Four basic elements are typically viewed as critical to traditional experimental design, regardless of research context: (1) random selection of subjects; (2) random assignment of subjects to the different treatment conditions; (3) experimenter manipulation of the treatments and (4) experimenter control over the conduct of the experiment. According to established research design, "random assignment of test units to treatment conditions facilitates causal interpretation by eliminating potential systematic differences across treatment conditions due to extraneous factors associated with characteristics of the test units" (Keppel 1982; Perdue and Summers 1986, p. 317). The effort to sidestep extraneous effects is supposedly furthered by the direct manipulation of treatments imposed during the experiment by the researcher. However, it is worth noting that discussions of what is referred to as "quasi-experimentation" in behavioral studies suggests that some of this rigor might be flexible in the interest of ensuring realism and robustness (e.g., using real workers in action-study experimental designs) (Cook and Campbell 1979). Yet, regardless of the extent of strict control, several steps remain common in experimental behavioral analysis: (1) conceptualizing the research question; (2) operationalization and design; (3) methodology and collecting data; (4) validity testing and interpretability and (5) effect and relationship testing. This essay explores each of these areas in turn.