Addressing unmeasured confounding in nonexperimental psychological research: A guide to computing and interpreting E-values

Randomized experiments remain the gold standard for establishing causality, yet ethical and practical constraints in certain fields often require researchers to rely on observational data. Although psychologists recognize that correlations do not imply causality, the conventional cautionary statements regarding correlation typically found at the end of articles have not sufficiently advanced psychological science, particularly in subfields such as developmental and personality psychology that predominantly rely on observational data. Sensitivity analyses commonly used in biostatistics and epidemiology offer powerful tools to address unmeasured confounding in observational data analysis. This tutorial explores the frequently overlooked but critical issue of unmeasured confounding in psychological research and introduces psychologists to the E-value, a novel and straightforward method for assessing the robustness of treatment outcome associations to unmeasured confounding. We demonstrate the application of E-value in common psychological research scenarios using R. The strengths and limitations of E-value are discussed, along with recommended best practices for its implementation in psychological research. By more explicitly discussing unmeasured confounding and incorporating sensitivity analysis techniques like the E-value into their analytical toolkits, psychologists can more accurately assess and transparently report research findings, particularly in subfields that primarily rely on observational data.