| |
| |
List of Figures | |
| |
| |
List of Tables | |
| |
| |
Preface | |
| |
| |
Acknowledgments | |
| |
| |
Introduction | |
| |
| |
| |
Introducing Bayesian Analysis | |
| |
| |
| |
The foundations of Bayesian inference | |
| |
| |
| |
What is probability? | |
| |
| |
| |
Probability in classical statistics | |
| |
| |
| |
Subjective probability1 | |
| |
| |
| |
Subjective probability in Bayesian statistics | |
| |
| |
| |
Bayes theorem, discrete case | |
| |
| |
| |
Bayes theorem, continuous parameter | |
| |
| |
| |
Conjugate priors | |
| |
| |
| |
Bayesian updating with irregular priors | |
| |
| |
| |
Cromwell's Rule | |
| |
| |
| |
Bayesian updating as information accumulation | |
| |
| |
| |
Parameters as random variables, beliefs as distributions | |
| |
| |
| |
Communicating the results of a Bayesian analysis | |
| |
| |
| |
Bayesian point estimation | |
| |
| |
| |
Credible regions | |
| |
| |
| |
Asymptotic properties of posterior distributions | |
| |
| |
| |
Bayesian hypothesis testing | |
| |
| |
| |
Model choice | |
| |
| |
| |
Bayes factors | |
| |
| |
| |
From subjective beliefs to parameters and models | |
| |
| |
| |
Exchangeability | |
| |
| |
| |
Implications and extensions of de Finetti's Representation Theorem | |
| |
| |
| |
Finite exchangeability | |
| |
| |
| |
Exchangeability and prediction | |
| |
| |
| |
Conditional exchangeability and multiparameter models | |
| |
| |
| |
Exchangeability of parameters: hierarchical modeling | |
| |
| |
| |
Historical note | |
| |
| |
| |
Getting started: Bayesian analysis for simple models | |
| |
| |
| |
Learning about probabilities, rates and proportions | |
| |
| |
| |
Conjugate priors for probabilities, rates and proportions | |
| |
| |
| |
Bayes estimates as weighted averages of priors and data | |
| |
| |
| |
Parameterizations and priors | |
| |
| |
| |
The variance of the posterior density | |
| |
| |
| |
Associations between binary variables | |
| |
| |
| |
Learning from counts | |
| |
| |
| |
Predictive inference with count data | |
| |
| |
| |
Learning about a normal mean and variance | |
| |
| |
| |
Variance known | |
| |
| |
| |
Mean and variance unknown | |
| |
| |
| |
Conditionally conjugate prior | |
| |
| |
| |
An improper, reference prior | |
| |
| |
| |
Conflict between likelihood and prior | |
| |
| |
| |
Non-conjugate priors | |
| |
| |
| |
Regression models | |
| |
| |
| |
Bayesian regression analysis | |
| |
| |
| |
Likelihood function | |
| |
| |
| |
Conjugate prior | |
| |
| |
| |
Improper, reference prior | |
| |
| |
| |
Further reading | |
| |
| |
| |
Simulation Based Bayesian Analysis | |
| |
| |
| |
Monte Carlo methods | |
| |
| |
| |
Simulation consistency | |
| |
| |
| |
Inference for functions of parameters | |
| |
| |
| |
Marginalization via Monte Carlo integration | |
| |
| |
| |
Sampling algorithms | |
| |
| |
| |
Inverse-CDF method | |
| |
| |
| |
Importance sampling | |
| |
| |
| |
Accept-reject sampling | |
| |
| |
| |
Adaptive rejection sampling | |
| |
| |
| |
Further reading | |
| |
| |
| |
Markov chains | |
| |
| |
| |
Notation and definitions | |
| |
| |
| |
State space | |
| |
| |
| |
Transition kernel | |
| |
| |
| |
Properties of Markov chains | |
| |
| |
| |
Existence of a stationary distribution, discrete case | |
| |
| |
| |
Existence of a stationary distribution, continuous case | |
| |
| |
| |
Irreducibility | |
| |
| |
| |
Recurrence | |
| |
| |
| |
Invariant measure | |
| |
| |
| |
Reversibility | |
| |
| |
| |
Aperiodicity | |
| |
| |
| |
Convergence of Markov chains | |
| |
| |
| |
Speed of convergence | |
| |
| |
| |
Limit theorems for Markov chains | |
| |
| |
| |
Simulation inefficiency | |
| |
| |
| |
Central limit theorems for Markov chains | |
| |
| |
| |
Further reading | |
| |
| |
| |
Markov chain Monte Carlo | |
| |
| |
| |
Metropolis-Hastings algorithm | |
| |
| |
| |
Theory for the Metropolis-Hastings algorithm | |
| |
| |
| |
Choosing the proposal density | |
| |
| |
| |
Gibbs sampling | |
| |
| |
| |
Theory for the Gibbs sampler | |
| |
| |
| |
Connection to the Metropolis algorithm | |
| |
| |
| |
Deriving conditional densities for the Gibbs sampler: statistical models as conditional independence graphs | |
| |
| |
| |
Pathologies | |
| |
| |
| |
Data augmentation | |
| |
| |
| |
Missing data problems | |
| |
| |
| |
The slice sampler | |
| |
| |
| |
Implementing Markov chain Monte Carlo | |
| |
| |
| |
Software for Markov chain Monte Carlo | |
| |
| |
| |
Assessing convergence and run-length | |
| |
| |
| |
Working with BUGS/JAGS from R | |
| |
| |
| |
Tricks of the trade | |
| |
| |
| |
Thinning | |
| |
| |
| |
Blocking | |
| |
| |
| |
Reparameterization | |
| |
| |
| |
Other examples | |
| |
| |
| |
Further reading | |
| |
| |
| |
Advanced Applications in the Social Sciences | |
| |
| |
| |
Hierarchical Statistical Models | |
| |
| |
| |
Data and parameters that vary by groups: the case for hierarchical modeling | |
| |
| |
| |
Exchangeable parameters generate hierarchical models | |
| |
| |
| |
�Borrowing strength� via exchangeability | |
| |
| |
| |
Hierarchical modeling as a 'semi-pooling� estimator | |
| |
| |
| |
Hierarchical modeling as a 'shrinkage� estimator | |
| |
| |
| |
Computation via Markov chain Monte Carlo | |
| |
| |
| |
ANOVA as a hierarchical model | |
| |
| |
| |
One-way analysis of variance | |
| |
| |
| |
Two-way ANOVA | |
| |
| |
| |
Hierarchical models for longitudinal data | |
| |
| |
| |
Hierarchical models for non-normal data | |
| |
| |
| |
Multi-level models | |
| |
| |
| |
Bayesian analysis of choice making | |
| |
| |
| |
Regression models for binary responses | |
| |
| |
| |
Probit model via data augmentation | |
| |
| |
| |
Probit model via marginal data augmentation | |
| |
| |
| |
Logit model | |
| |
| |
| |
Binomial model for grouped binary data | |
| |
| |
| |
Ordered outcomes | |
| |
| |
| |
Identification | |
| |
| |
| |
Multinomial outcomes | |
| |
| |
| |
Multinomial logit (MNL) | |
| |
| |
| |
Independence of irrelevant alternatives | |
| |
| |
| |
Multinomial probit | |
| |
| |
| |
Bayesian analysis via MCMC | |
| |
| |
| |
Bayesian approaches to measurement | |
| |
| |
| |
Bayesian inference for latent states | |
| |
| |
| |
A formal role for prior information | |
| |
| |
| |
Inference for many parameters | |
| |
| |
| |
Factor analysis | |
| |
| |
| |
Likelihood and prior densities | |
| |
| |
| |
Identification | |
| |
| |
| |
Posterior density | |
| |
| |
| |
Inference over rank orderings of the latent variable | |
| |
| |
| |
Incorporating additional information via hierarchical modeling | |
| |
| |
| |
Item-response models | |
| |
| |
| |
Dynamic measurement models | |
| |
| |
| |
State-space models for [pooling the polls] | |
| |
| |
| |
Bayesian inference | |
| |
| |
| |
Appendices | |
| |
| |
| |
Working with vectors and matrices | |
| |
| |
| |
Probability review | |
| |
| |
| |
Foundations of probability | |
| |
| |
| |
Probability densities and mass functions | |
| |
| |
| |
Probability mass functions for discrete random quantities | |
| |
| |
| |
Probability density functions for continuous random quantities | |
| |
| |
| |
Convergence of sequences of random variables | |
| |
| |
| |
Proofs of selected propositions | |
| |
| |
| |
Products of normal densities | |
| |
| |
| |
Conjugate analysis of normal data | |
| |
| |
| |
Asymptotic normality of the posterior density | |
| |
| |
References | |
| |
| |
Topic index | |
| |
| |
Author index | |