• 5 levels of priors
• Independence
• General principles
• How informative is the prior?
• Default prior for treatment effects scaled based on the standard error of the estimate
• Generic prior for anything
• Story: When the generic prior fails. The case of the negative binomial.
• Generic weakly informative prior
• Setting priors based on the prior predictive distribution (I. J. Good: "Method of imaginary results")
• Boundary-avoiding priors for modal estimation (posterior mode, MAP, marginal posterior mode, marginal maximum likelihood, MML)
• Prior for linear regression
• Prior for the regression coefficients in logistic regression (non-sparse case)
• Scaling
• Data-dependent scaling
• Sparsity promoting prior for the regression coefficients ("Bayesian model reduction")
• Prior for degrees of freedom in Student's t distribution
• Prior for the shape parameter in negative-binomial distribution
• Prior for elasticities (regressions on log-log scale)
• Prior for a single correlation parameter
• Prior for a covariance matrix
• Prior for scale parameters in hierarchical models
• Prior for cutpoints in ordered logit or probit regression
• Priors for Gaussian processes
• Priors for rstanarm
• Acknowledgements