Statistical methods for Criminology (online)

Data and statistics form the basis of much political discussion about crime, provide the foundation for evidence-based research on justice interventions, and shape our understanding of possible biases in the justice system.
However, working with any social science data can be complex, and there are particular features of working with criminological data that can pose challenges for researchers.

Data and statistics form the basis of much political discussion about crime, provide the foundation for evidence-based research on justice interventions, and shape our understanding of possible biases in the justice system.
However, working with any social science data can be complex, and there are particular features of working with criminological data that can pose challenges for researchers.

Course Description

This is an intensive one-day workshop on criminological data and how data about crime and victimization can be understood using statistical methods. The workshop is specifically designed for criminologists and social scientists, and criminological data and examples will be showcased throughout the workshop. 

The workshop will introduce common types of criminological data and how this can be modelled in the generalized linear modelling framework. Participants will be introduced to common approaches to modelling criminological data sources, such as recorded crime and victimization data using count models, and how criminologically relevant quantities of interest (such as victimization inequality measures) can be simulated from fitted models. Participants will learn about issues that can arise analysing data from the criminal justice system, such as measurement error and selection bias. Finally, the workshop will discuss the ethics of working with crime data. 

The emphasis will be on interpreting outputs (e.g. from data analysis software packages) and understanding results (e.g. in published papers).

The event is intended to be engaging and informative and there will be audience participation and opportunities to ask questions. This is not a practical workshop and it does not provide training in the use of data analysis software. It will however provide a strong theoretical foundation for future engagement at practical workshops that are designed to provide hands-on training in data analysis.

A high level of mathematical ability is not required, but participants should ideally have undertaken an introductory statistics and data analysis course (e.g. a semester long module as part of a Masters degree) or have attended an NCRM workshop on Statistical Modelling.

Course Timings: 10:00 – 17:00 

This workshop does not require access to specialist software. Concepts will be illustrated using examples in the R programming language and the code for these examples will be provided for participants to follow along – but this is not a requirement.

Researchers at any career stage are welcome. The workshop focuses on criminological examples, but the issues raised will be relevant for researchers across the social sciences.

Course Contents:

• Types of criminological data (victimization survey data; administrative data)
• Review of the general linear model with applications in criminology (e.g. counts of crime or victimization)
• Accounting for measurement error and selection bias in data from the criminal justice system
• Using statistical models to describe criminologically relevant quantities of interest, such as victimization inequality measures
• The ethics of working with data from the justice system and presenting criminological results

By the end of this workshop you will:

• Understand the main types of criminological data
• Understand common types of general linear model used in criminology
• Be aware of how measurement error and selection biases can affect data from the criminal justice system
• Understand how criminologically relevant quantities of interest can be described from fitted regression models
• Be aware of ethical issues when working with criminological data and presenting results