Better Inference in Psychological Science: Test Less, Estimate More

The estimation approach to inference emphasizes effect sizes, confidence intervals, meta-analysis, and the use of Open Science practices. Emphasizing estimation over testing could improve inference in psychological science: by nudging researchers away from invalid interpretations of tests, by providing a more straightforward way to plan for sensible sample sizes, and by providing a framework for fair testing. In this seminar I’ll illustrate the benefits of estimation thinking with examples from the literature on oxytocin and trust, a research area that typifies some of the endemic problems that arise with poor statistical practice.