The Value of Bayesian Methods for Accurate and Efficient Neuropsychological Assessment

Abstract

Abstract Objective: Clinical neuropsychology has been slow in adopting novelties in psychometrics, statistics, and technology. Researchers have indicated that the stationary nature of clinical neuropsychology endangers its evidence-based character. In addition to a technological crisis, there may be a statistical crisis affecting clinical neuropsychology. That is, the frequentist null hypothesis significance testing framework remains the dominant approach in clinical practice, despite a recent surge in critique on this framework. While the Bayesian framework has been put forward as a viable alternative in psychology in general, the possibilities it offers to clinical neuropsychology have not received much attention. Method: In the current position paper, we discuss and reflect on the value of Bayesian methods for the advancement of evidence-based clinical neuropsychology. Results: We aim to familiarize clinical neuropsychologists and neuropsychological researchers to Bayesian methods of inference and provide a clear rationale for why these methods are valuable for clinical neuropsychology. Conclusion: We argue that Bayesian methods allow for a more intuitive answer to our diagnostic questions and form a more solid foundation for sequential and adaptive diagnostic testing, representing uncertainty about patients’ observed test scores and cognitive modeling of test results.