Weighing People Rather Than Food: A Framework for Examining External Validity

Abstract

Research training in psychological science emphasizes common threats to internal validity, with no comparably systematic or rigorous treatment of external validity. Trade-offs between internal and external validity are well known in some areas (e.g., efficacy vs. effectiveness studies in clinical psychology), less so in others (e.g., forensic research on eyewitness identification, false memories, or confessions). We present a framework for examining external validity grounded in four domains—populations, settings, outcomes, and timeframes—that can be used to enhance the generalizability of findings. We discuss this framework and then illustrate its use by reviewing mindless eating interventions intended to help people lose weight. Research in this published literature seldom samples from appropriate populations (e.g., overweight or obese individuals) or measures appropriate outcomes (e.g., weight change) in appropriate settings (e.g., the home) over appropriate timeframes (e.g., sustained interventions with follow-up) to determine whether practical advice is empirically supported. In their applied work, we encourage psychological scientists to design studies, analyze data, and report findings with greater attention to external validity to demonstrate, rather than assume, the generalizability of findings to the intended populations, settings, outcomes, and timeframes. Editors and reviewers can hold investigators accountable for doing so.