Introduction to applied Bayesian statistics and estimation for social scientists.
The author states in the preface that the purpose of this book is to describe in great detail the Bayesian approach through Markov chain Monte Carlo methods to readers from social sciences, recognizing the appeal of such methods in applied social sciences and the lack of a book that explains the approach in a simple language. The book, indeed, contains a very detailed and comprehensive description of MCMC methods useful for applied researchers. The first three chapters contain a short introduction to the problem (Chapter 1). The classical approach to statistics and the Bayesian approach are treated in Chapters 2 and 3, respectively. Not many mathematical details are given but only the necessary ones to proceed. The author avoids philosophical discussions for and against the Bayesian approach; such issues hardly take two pages in the book. Chapters 4 and 5 describe the Gibbs sampling and the Metropolis-Hastings algorithm, respectively, while Chapter 6 contains more advanced issues related to MCMC, such as diagnostics of convergence and model fit. The four next chapters present particular models widely applicable in social sciences, like the regression model in Chapter 7, generalized linear models in Chapter 8 (while the title is unlucky since only the dichotomous and ordinal probit models are discussed and not the general GLMs), hierarchical models in Chapter 9 and multivariate regression models in Chapter 10. Concluding remarks can be found in Chapter 11. R codes are provided throughout the book to facilitate the computations. Some short mathematical exposition is contained in the Appendix to help mathematically unexperienced readers. Undoubtedly the book is interesting and the topic that is tried to be covered and presented in a rather simplified way is huge and hard to be simplified. The reader will gain an extensive knowledge of the issues covered even if details about the origin of the methods and related material will not referred. For example, while the book describes MCMC methods, Markov chains are totally absent of the book. The MCMC approach is not fully motivated. Never in the book the huge problem of evaluating multivariate integrals involved in Bayesian computations is shown and the reader may not see the real reasons for them. Some issues are treated only superficially, and some others mentioned in the relevant chapters are never used later on (e.g., diagnostics). The author chooses to present certain parts of the existing theory ignoring other, more relevant issues (e.g., in the convergence diagnostics only the R diagnostic is referred). To the extend that the book is directed to scientists from social science this is excusable; on the other hand neither the social scientists nor any other readers will feel happy to check that only a part of the existing methodology has been presented and thus only a portion of the methodology has been given. The book contains also a series of simplifications which can be annoying to many readers. Many integrals do not have differentials, the definition of cumulative functions is misleading (page 13), the random variables are not in capitals as they are usually defined, while in the description of the MCMC the initial values are invalid and hence one calculates a conditional distribution that does not exist. Perhaps all those problems do not cause serious problems for applying the methodologies; however they shed ambiguity on the book. More importantly, some of the examples are wrong and they should not be used in a textbook, like the use of Poisson likelihood for binary data, or the use of a normal regression for ordinal outcome, since they give a rather wrong signal to the user, especially if he is somebody that expects from the book to help him understand the whole issues. I am sure that the author could select more appropriate examples to present the methodologies. Concluding, the book may have funs on the social sciences but the readers must consider this as just a starting point in order to go deeper into the subject. Students may use the book with caution as it contains certain misconceptions. The book itself may throw not enough light on the researchers needs to work with Bayesian approaches and perhaps they will realize that still a series of problems remains after reading the book. The book only partially fulfills its aims.
- Inequalities between others do matter: evidence from multiplayer dictator games
- Unsupervised human activity analysis for intelligent mobile robots
- On a signal detection approach to \(m\)-alternative forced choice with bias, with maximum likelihood and Bayesian approaches to estimation
- Bayes model selection with path sampling: factor models and other examples
- Bayesian inference in the social sciences
- Bayesian reliability models of Weibull systems: state of the art
- scientific article; zbMATH DE number 6388052 (Why is no real title available?)
- Applied Bayesian hierarchical methods.
- scientific article; zbMATH DE number 1912149 (Why is no real title available?)
- Bayesian hierarchical factor analysis for eficient estimation across race/ethnicity
- Network meta-analysis: a statistical physics perspective
- scientific article; zbMATH DE number 7247632 (Why is no real title available?)
- Time-varying threshold regression model using the Kalman filter method
- Bayesian methods. A social and behavioral sciences approach
- Bayesian analysis for social sciences
- Prediction of the nonsampled units in survey design with the finite population using Bayesian nonparametric mixture model
- On E-Bayesian analysis of the hierarchical normal and inverse gamma model using different loss functions and its application
- Random effects multinomial processing tree models: a maximum likelihood approach
- Model selection and identifiability analysis of HIV and SARS-CoV-2 co-infection model with drug therapy
- Classical and Bayesian estimation of entropy for Pareto distribution in presence of outliers with application
- Clustering longitudinal ordinal data via finite mixture of matrix-variate distributions
- A latent variable mixed-effects location scale model that also considers between-person differences in the autocorrelation
- Inference of improved adaptive progressively censored competing risks data for Weibull lifetime models
- Small area estimation of receiver operating characteristic curves for ordinal data under stochastic ordering
- Inference for two Weibull populations under joint generalized progressive type-I hybrid censoring with a simulation study and applications
- Assessing respiratory virus co-infections using an identifiable model: the case of influenza and SARS-CoV-2 in Italy
- An efficient PG-INLA algorithm for the Bayesian inference of logistic item response models
- Statistical inference and optimal design for multiple constant-stress accelerated life tests under ordered ranked set sampling with progressive Type-II censoring
- Estimation of process capability index ℂ s based on multiple interval type I censored samples using generalized inverse Lindley distribution
- A novel nonvisual procedure for screening for nonstationarity in time series as obtained from intensive longitudinal designs
- Evaluating manifest monotonicity using Bayes factors
This page was built for publication: Introduction to applied Bayesian statistics and estimation for social scientists.
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2641628)
