TI Methods Speaker Series
WHEN: Wednesday, March 27th, 2019 from 12:00 noon to 1:00 PM
WHERE: Rudy North Lecture Theatre (CBH 101) in the Centre for Brain Health, UBC Point Grey Campus. Remote attendance by videoconference is possible but you will need to register.
TITLE: Navigating the p-value swamp for meta-analysts
SPEAKER: Dr. Ed Kroc, Assistant Professor of Measurement, Evaluation, and Research Methodology within UBC’s Department of Educational and Counselling Psychology, and Special Education
About the topic: The statistical concept of the p-value has enjoyed a long and contentious history in applied quantitative research. Recent proclamations of “replicability crises” in the social and health sciences have inflamed the modern debate, yet while methodologists argue, applied researchers are left with less confidence than ever in their quantitative mettle. In this talk, designed for applied researchers, I will unpack these issues and discuss their specific implications for meta-analytic inferences. In particular, I will explain the interplay between p-values, statistical power, and non-normality of random effect sizes. Key meta-analytic inferences about a mean effect size and the dispersion of true effects in a meta-population are particularly sensitive to these issues under the usual random effects meta-analytic model. I will argue that the best way to extract accurate information from your p-values is to always interpret them within the context of the statistical power of the studies that have generated them.
About the speaker: Dr. Ed Kroc is an Assistant Professor of Measurement, Evaluation, and Research Methodology within UBC’s Department of Educational and Counselling Psychology, and Special Education. A graduate of UBC’s Department of Mathematics and a former post-doc in UBC’s Department of Statistics, he has long been interested in bridging the gap between statistical rigour and practical scientific research. Dr. Kroc’s particular research interests focus on ways to address this gap, as well as on statistical measurement, spatio-temporal modelling, and urban ecology.
About the TI Methods Speaker Series: The TI Methods Speaker Series are offered free of charge and everyone is welcome, but space is limited, registered participants will be given priority. Remote attendance is possible by videoconference, you will receive the link after you register.