An Introduction to (Bayesian) Statistics
This two hour workshop will be presented at the 31st EFPSA Congress in Qakh, Azerbaijan. There are two versions, a more extensive one, and a more concise one with about 40 slides less. I will present the latter, but you can view the former online here.
Most students dislike statistics not because it's hard, but because it's unintuitive or even confusing. And it's true: classical statistical concepts such as the p-value and confidence intervals are exceptionally difficult to grasp; for students and statisticians alike. However, statistical inference is the universal tool of science, and a good scientist must have a good command of it. Eschewing the standard way of teaching statistics in psychology, i.e., introducing loosely connected tests in a cookbook-oriented fashion, in this workshop, I provide an introduction to statistics from "first principles". I discuss its exciting history, controversies, and enigmatic key players. After introducing probability as the means to quantify uncertainty, I focus on the role of statistical modeling. On a real-life example, I illuminate the problems of parameter estimation, hypothesis testing, and model prediction both from a classical and Bayesian perspective. This allows me to (re)introduce you to concepts such as maximum likelihood, confidence intervals, and p-values, as well as outline a more intuitive and powerful approach to statistics --- the Bayesian approach. In the last, practical segment we use JASP (https://jasp-stats.org) to apply Bayesian principles to data sets from real research. All materials including slides, code, and further resources will be made available at https://github.com/fdabl/Intro-Stats. Note that there are no prerequisites for this workshop. All that you need to bring is a laptop and a focused mind!
- Bayesian Inference for Psychology. Part I and II (2017) by Wagenmakers et al.
- Bayesian Benefits for the Pragmatic Researcher (2016) by Wagenmakers, Morey, and Lee
- How to become a Bayesian in eight easy steps (2016) by Etz, Gronau, Dablander, Edelsbrunner, and Baribault
- The philosophy of Bayes factors and the quantification of statistical evidence (2016) by Morey, Romeijn, & Rouder
- Statistical tests, p-values, confidence intervals, and power: a guide to misinterpretations (2016) by Greenland et al.
- Popular Science Books
- The Etz-Files by Alexander Etz
- The 20% Statistician by Daniel Lakens
- Statistical Modeling, Causal Inference, and Social Science by Andrew Gelman
- Further resources
This work is licensed under a Creative Commons Attribution 4.0 International License.