ICIAC 2011 Stream 4: Classes and Workshops

Abstracts and slides

Class 4C: Improving the explanatory content of problem profiles using hypothesis testing (G)

Spencer Chainey, Department of Security and Crime Science, University College London

Analysis is an integral part of police and community safety decision making – if a crime problem is clearly understood, it can help identify the solutions that will most likely be effective.  Although the profile of analysis has been raised in recent years, its routine production has often resulted in many analysis products often offering only a descriptive presentation of the problem that is being examined, rather than one that is more explanatory in its tone.  In this class we propose the use of a hypothesis testing methodology to improve the explanatory content of crime and intelligence analysis, and illustrate its use with example of serious violence in Sunderland and residential burglary in Oldham, Greater Manchester.  We argue that this approach produces analytical products that are richer in explanatory and interpretative substance, helps to improve commissioning dialog, and generates results that help to more specifically identify how a crime problem can be tackled.

Analysis is an integral part of police and community safety decision making – if a crime problem is clearly understood, it can help identify the solutions that will most likely be effective.  Although the profile of analysis has been raised in recent years, its routine production has often resulted in many analysis products often offering only a descriptive presentation of the problem that is being examined, rather than one that is more explanatory in its tone.  In this class we propose the use of a hypothesis testing methodology to improve the explanatory content of crime and intelligence analysis, and illustrate its use with example of serious violence in Sunderland and residential burglary in Oldham, Greater Manchester.  We argue that this approach produces analytical products that are richer in explanatory and interpretative substance, helps to improve commissioning dialog, and generates results that help to more specifically identify how a crime problem can be tackled.

Presenter's slides: ICIAC11_4C_SChainey

Class 4D: An introduction to the use of R for crime and intelligence analysis (A)

Lisa Wainer, Department of Security and Crime Science, University College London

This session aims to introduce delegates to the R programming language that was developed for data analysis, statistical modelling, simulation and graphics. R is a both a flexible and powerful tool. It can be used to run analytical methods typically found in other statistical packages - look for a statistical model or test that you want to run and you will find the tools available to do so in R. However, it can also be used to develop your own methods for analysing data that may work faster and be more relevant to your line of investigation. One big advantage lies in its ability to automate repetative but complex analytical tasks - a bit of effort devoted to learning the basics can pay off quickly. Automating complex tasks can free up time for more exploratory work - which can also be done with R! Many people (not just crime analysts) are put off by the perceived difficulty of learning R. This session aims to address some of those concerns and give you some basic concepts, hints and tips that should help you get started.

Presenter's slides: ICIAC11_4D_LWainer

Class 4E: Measuring repeat and near repeat victimisation - predicting the future locations of crime (I)

Shane Johnson, Department of Security and Crime Science, University College London

It is an understatement to say that the ability to perfectly predict when and where future crimes were to occur would be of considerable value to crime reduction agencies.  While we are not there yet, there is much that we already know.  For example, there are regularities associated with the timing and location of crime events that can inform methods of crime prediction.  In particular, our research has shown that the risk of burglary at a dwelling increases following a first offence, and that incidents of such “repeat victimization” tend to occur swiftly.  More generally, it appears that the risk of victimization clusters in space and time, with risk spreading much like patterns observed for contagious diseases.  That is, in the case of burglary, neighbours of burgled homes also tend to be at a temporarily elevated risk of victimization following after the offence occurs.  Where an offence occurs shortly after, and near to, a previous offence this is referred to as a “near repeat”.

Research demonstrates that near repeats are observed with a higher frequency than would be expected on a chance basis for burglary but also for crimes including theft from motor vehicle, cycle theft, shootings in the US, and even insurgency in Iraq.  Where patterns of near repeats are robust, it is possible to make predictions regarding the timing and location of future crimes that are superior to forecasts generated using existing methods of crime hotspot analysis.  In this session, I will start with a whistle stop of the research discussed above.  However, the focus of the session will be on metrics and methods of measuring repeat and near repeat victimization, including a discussion of freeware that can be used to quantify and establish the reliability of patterns observed.  Other types of analyses that may be useful in quantifying and understanding crime patterns will also be discussed.  The session will finish with a brief discussion of the direction of our current work.

Presenter's slides: ICIAC11_4E_SJohnson

Page last modified on 24 nov 11 15:23