This work contains the Bayesian model examples written by Simon Jackman and previously available on his website. These were originally written in WinBUGS or JAGS. I have translated these examples into Stan and revised or edited them as appropriate.

This work is licensed under the Creative Commons Attribution 4.0 International License

  1. Undervote: difference of two independent proportions; racial differences in self-reported undervoting
  2. Cancer: difference of two independent proportions; differences in rates of lung cancer by smoking
  3. Florida: learning about an unknown proportion from survey data; using survey data to update beliefs about support for Bush in Florida in the 2000 presidential election campaign
  4. Turnout: logit/probit models for binary response; voter turnout as a function of covariates
  5. Co-Sponsor: computing auxiliary quantities from MCMC output, such as residuals, goodness of fit; logit model of legislative co-sponsorship
  6. Reagan: linear regression with AR(1) disturbances; monthly presidential approval ratings for Ronald Reagan
  7. Political Sophistication: generalized latent variable modeling (item-response modeling with a mix of binary and ordinal responses); assessing levels of political knowledge among survey respondents in France
  8. Legislators: generalized latent variable modeling (two-parameter item-response model); estimating legislative ideal points from roll call data
  9. Judges: item response modeling; estimating ideological locations of Supreme Court justices via analysis of decisions
  10. Resistant: outlier-resistant regression via the t density; votes in U.S. Congressional elections, 1956-1994; incumbency advantage.
  11. House of Commons: analysis of compositional data; vote shares for candidates to the U.K. House of Commons
  12. Campaign: tracking a latent variable over time; support for candidates over the course of an election campaign, as revealed by polling from different survey houses.
  13. Aspirin: meta-analysis via hierarchical modeling of treatment effects; combining numerous experimental studies of effect of aspirin on surviving myocardial infarction (heart attack)
  14. Corporatism hierarchical linear regression model, normal errors; joint impact of left-wing governments and strength of trade unions in structuring the determinants of economic growth
  15. Bimodal: severe pattern of missingness in bivariate normal data; bimodal density over correlation coefficient
  16. Unidentified: the consequences of over-parameterization; contrived example from Carlin and Louis
  17. Engines: modeling truncated data; time to failure, engines being bench-tested at different operating temperatures
  18. Truncated: Example of sampling from a truncated normal distribution.
  19. Generalized Beetles: Generalizing link functions for binomial GLMs.
  20. Negative Binomial: Example of a negative binomial regression of homicides


The R packages, Stan models, and datasets needed to run the code examples can be installed with