Distinguishing Constructs from Variables in Designing Research

Current research practices often conflate theoretical constructs and explanatory hypotheses with va riables and predicted effects, to the detriment of research progress. This has led to the use of procedu res such as manipulation checks, mediation analysis, and boundary conditions predicated on the i dea that matching constructs to variables is necessary to validate that a theory corresponds to an effect . An alternative perspective, Inference to the Best Explanation (IBE), calls for designing researc h to exploit the power of distinguishing constructs from variables, hypotheses from predict i ons, and theory from effects. IBE calls for stating hypotheses (Hs) about construct-to- constru ct relationships and, separately, the predictions (Ps) about variable-to-variable effects that are explai n ed by the hypotheses. In addition, articles should include disparate effects, a single explanation coverin g all studies, and a discussion of the use of the research in specific problem contexts. The applic a tion of IBE is illustrated with research investigating when judgments are based on a feeling about t h e ease of information retrieval versus the information content itself.