Interested in Bayesian statistics and want to know more?

Some suggested articles below. Not a systematic review but ones that I found useful or might be of general interest. Some start off easy but suddenly get hard - dip in and out of several to see which interest you most.

Problems with p values, confidence intervals and stepwise methods

• Scientists rise up against statistical significance - a short article outlining some of the issues
• 1,500 scientists lift the lid on reproducibility - the reproducibility crisis
• An investigation of the false discovery rate and the misinterpretation of p-values - excellent article by David Colquhoun, who is not a Bayesian but clearly explains the problems with p values - probably the best long article to read, but for a shorter more accessible version see he has done a more accessible version here.
• Statistical Thinking - My Journey from Frequentist to Bayesian Statistics - worth reading if you think frequentist statistics is “objective” and Bayesian statistics is “subjective”
• The ASA Statement on p-Values: Context, Process, and Purpose - if you still want to use p values then do read this to understand best practices.
• Significance, truth and proof of p values: reminders about common misconceptions regarding null hypothesis significance testing - we all need reminding
• The P Value and Statistical Significance: Misunderstandings, Explanations, Challenges, and Alternatives - PMC - interpret your p value as a continuous variable and don’t use thresholds
• Beyond psychology: prevalence of p value and confidence interval misinterpretation across different fields - it’s endemic
• How Confident are Students in their Misconceptions about Hypothesis Tests? - strangely worded title; some students thought that a significant p value was mathematical proof
• Redefine statistical significance - a proposal to change the p value threshold
• HARKing: Hypothesizing After the Results are Known - why you should publish your analytical protocol to give confidence that your results were not due to data dredging
• The Extent and Consequences of P-Hacking in Science - “p hacking”, selective reporting and publication bias
• The fallacy of placing confidence in confidence intervals - “Frequentist CI theory says nothing at all about the probability that a particular, observed confidence interval contains the true value; it is either 0 (if the interval does not contain the parameter) or 1 (if the interval does contain the true value).”
• Step away from stepwise - “standard statistical tests assume a single test of a pre-specified model and are not appropriate when a sequence of steps is used to choose the explanatory variables”
• Moving to a World Beyond “p < 0.05” - don’t say “statistically significant”
• Abandon Statistical Significance- “We recommend dropping the NHST paradigm - and the p-value thresholds intrinsic to it - as the default statistical paradigm for research, publication, and discovery in the biomedical and social sciences.”

Bayesian concepts, methods and software

• BBC Radio 4 - More or Less: Behind the Stats, Bayes: the clergyman whose maths changed the world - story of Thomas Bayes
• Bayes theorem, the geometry of changing beliefs - an introductory video re understanding Bayes’ theorem
• The Bread and Butter of Bayes with Ghassan Hamra | SERious EPI - a podcast episode on Bayesian statistics
• Four reasons to prefer Bayesian analyses over significance testing - four scenarios where Bayesian methods offer more
• An Introduction to Data Analysis - one of the gentler introductions to Bayesian methods, which are introduced alongside frequentist statistics for beginners (which is probably the best way to learn statistics)
• Get Started with Bayesian Analysis - why you should consider the Bayesian approach and an R package to help
• Bayesian perspectives for epidemiological research: I. Foundations and basic methods - anything by Sander Greenland is worth reading but this is good on concepts and the misconceptions about Bayesian statistics which may discourage their use.
• Introduction to Bayesian Statistics - Yi’s Knowledge Base - lots of R code, though it gets increasingly mathematical
• Bayesian Basics - possibly not the first thing you should read about Bayesian statistics, but a good practical intro once you are starting to get the basic ideas.
• An Introduction to Bayesian Reasoning and Methods - online book covering Bayesian methods - lots of exercises
• What does a Bayes factor feel like? - if you Google Bayesian statistics you will find a lot about Bayes factors, but I haven’t yet decided what place they have
• Bayes Rules! An Introduction to Applied Bayesian Modeling - a whole online book going through a lot of the foundations of Bayesian statistics
• Stepwise regression for Bayesian models - Cross Validated - why Bayesians don’t use stepwise regression.
• Bayesian Workflow - a very general approach to fitting Bayesian models
• rstanarm: Bayesian Applied Regression Modeling via Stan - the package you should use for Bayesian multivariable regression
• A Shiny app for fitting Bayesian regression models - intended for econometrics, not so useful for epidemiology - perhaps adoption of Bayesian methods would increase if someone developed a Shiny app for Bayesian epidemiological model fitting (it has been on my list of ideas for a long time)
• User-friendly Bayesian regression modeling: A tutorial with rstanarm and shinystan - the shinystan Shiny app allows you to explore and check a fitted Bayesian regression model
• Introduction to Bayesian Regression Modeling in R using rstanarm - another example of using the rstanarm package
• Bayesian Regression Models: Choosing informative priors in rstanarm - good presentation with examples of code and use of priors, e.g. for Bayesian logistic regression
• Bayesian Data Analysis demos for R - lots of example Bayesian analyses with R code
• Tidy Data and Geoms for Bayesian Models - the Bayesian tidyverse
• JASP - A Fresh Way to Do Statistics - easy to use statistical software (based on R) that can do most commonly-used frequentist and Bayesian statistical methods (unfortunately no Bayesian logistic regression yet, though they were apparently working on it and this would be a killer feature)
• Prophet - a Bayesian time series forecasting tool useful if you have seasonal data from over several years - created by Facebook
• causalimpact - an R package estimating the causal effect of an intervention on a time series using Bayesian methods - created by Google
• The Rapid Inquiry Facility was a tool for fitting Bayesian spatial models - it is now an ArcGIS plugin
• SPSS Statistics - even SPSS has Bayesian methods nowadays
• Introduction to Bayesian Data Analysis | openHPI - this course looks good