This case study - conducted as part of the ESRC grant The use of interactive electronic-books in the teaching and application of modern quantitative methods in the social sciences - was undertaken with the kind involvement of Paul Norris.

Paul is a Lecturer in the School of Social and Political Science in the University of Edinburgh, UK. Paul's research interests focus on the relationship between social context, political economy and crime and justice. In particular, he has a keen interest in peoples’ attitudes towards the criminal justice system and how criminal justice policy relates to a country’s wider social and political context. Paul uses, especially, latent variable models, event history analysis and multilevel modelling to help answer his substantive questions of interest.

The case study we look at with Paul investigates inequality, poverty and homicide rates in European countries. It explores and compares a range of covariates which can be used to try to control for poverty when investigating effect of economic inequality on homicide rates. Using Bayesian estimation methods, Paul also contrasts a variety of modelling approaches, including negative binomial models.

This research used data from the publicly-accessible EU-SILC (European Union - Statistics on Income, Social inclusion and living Conditions) database.

Paul predominantly used the statistical software package R and the Bayesian estimation engine JAGS to analyse this data.

You can download the Stat-JR workflow, together with a Stat-JR template called by the workflow (the workflow calls additional templates distributed with the standard Stat-JR release) we prepared from this link: case_study_neg_binom (zip, 4 kb)

Note that you will need JAGS installed (and will further need to tell Stat-JR where to find it), and to have downloaded the relevant EU-SILC dataset. The screenshots below show the Stat-JR workflow as it appears in the Stat-JR:LEAF interface: