学术讲座: Using Bayes to get the most out of null results

Speaker:  Prof Zoltan Dienes 
Deputy Head of School
School of Psychology, University of Sussex
Time:　　  14:00pm - 15:30pm
Date:　　  Sep 6 (Thu), 2012
Venue:　  Lecture Hall, North Building

Abstract:
User's of orthodox statistics, including psychologists, have typically not interpreted null results in a principled way, resulting in mistaken conclusions and choices of research direction based on non-significant findings. A non-significant result does not distinguish no evidence for an effect from evidence for no effect, radically different states of affairs. The distinction falls out naturally from a Bayesian analysis, in a consistent way that cannot be accomplished with orthodox statistics, even when statistical power is taken into account. I will discuss how confidence intervals (the best orthodox solution available) can be used to test theories in ways researchers seem unaware, and then indicate the crucial inferential limitations they nonetheless have. I will show how simple free online software for calculating Bayes Factors can be used to determine what a null result is actually telling us. The new tools I introduce in this talk, and now gradually appearing in the literature, should enable publishing null results on an equal footing with significant results. I hope to make people aware of a genuine practical hole in current practice in any domain of psychology and offer a genuine practical solution. The talk will be useful to anyone who uses inferential statistics. No statistical background is assumed other than knowledge of how to do a t-test.