AQMeN - Applied Quantitative Methods Network

Presenter: Dr. David Lucy

 Click Here To Book

This course introduces students to the use of Bayesian methods for data analysis in the social sciences. This also provides the basic concepts of the Bayesian approach to statistics such as the subjective interpretation of probability, types of prior distributions, the use of Bayes theorem in updating information and inference procedures such as Bayesian estimates. These will include incorporating classical likelihood within the Bayesian framework, and fitting linear regression models and generalized linear models including the binary response model, the multinomial model, and the Poisson model for counts. The main focus of the course will be the application of Bayesian models in social science modelling and related disciplines.

Sample Notes

Cost

• Lancaster University staff and postgraduates - £50
• External participant from an academic institution - £270
• External participant from a non academic institution - £390

 The course fees include all supporting documentation, refreshments and lunches.

 Students will acquire a knowledge of: 

• the fundamental notion of Bayes theorem
• the relationship between Bayesian methods and classical likelihood methods
• the use of Bayesian methods to combine prior information with data
• the basic concepts of Bayesian inference, including posterior conditioning, credible sets, prior distributions, and the likelihood principle
• the estimation of statistical models using Bayesian methods

 and develop skills to: 

• apply theoretical concepts
• examine model fitting in practice using Bayesian principles
• explore applied Bayesian modelling

Topics

Topics covered will be:

An introduction to Bayesian analysis, single parameter Bayesian modelling, informative priors, noninformative priors, posterior and predictive distributions, conjugate distributions, Bayesian forms of confidence intervals, Bayesian regression and analysis of variance models, Bayesian analysis of generalized linear models including the binary response model, the multinomial model, and the Poisson model for counts.