19.07.2013 Views

Murder and Serious Sexual Assault - Lancaster EPrints - Lancaster ...

Murder and Serious Sexual Assault - Lancaster EPrints - Lancaster ...

Murder and Serious Sexual Assault - Lancaster EPrints - Lancaster ...

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

APPENDIX C<br />

In total, therefore, 13 variables contributed to the risk score for murder. Eight of the<br />

summary measures produced an increased risk (shown by a positive estimate in<br />

column 2 <strong>and</strong> a relative risk contribution above one in column 4); five summary<br />

measures lowered the risk (shown by a negative estimate in column 2 <strong>and</strong> a relative<br />

risk contribution below one in column 4). Some of the unusual offences identified<br />

in Table 6 are not present in the risk score. Kidnapping <strong>and</strong> blackmail are both<br />

missing from the final score, as such activity does not further significantly change<br />

the risk of murder once the chosen thirteen variables have been allowed for.<br />

However, manslaughter increases the risk of murder by a factor of over eleven times,<br />

even after controlling for the two offence codes of violence. Similarly, attempting to<br />

pervert the course of justice (offence code 79) significantly reduced the risk of<br />

murder by over seven times, even after controlling for other variables.<br />

Interpretation<br />

The estimates in the above table can be used to build a relative risk for subsequent<br />

murder for any offender with a prior conviction. In order to consider what such a<br />

score is relative to, it is necessary to define a typical ‘baseline’ offender with none of<br />

the criminal career characteristics appearing in the above table. Such an offender is<br />

allowed to have one or more prior convictions for any offences not appearing in<br />

Table 17 (the ‘murder-neutral’ offences) which have not led to a custodial sentence<br />

on the last sentencing occasion, but should have no convictions for any of the<br />

offences which do appear in the above table. These ‘murder-neutral’ offences would<br />

include, for example, burglary in a dwelling, shoplifting <strong>and</strong> car crime. We can<br />

imagine this baseline offender to be a typical petty offender, with a string of<br />

relatively minor offences that have not led to a custodial sentence.<br />

Compared with this typical petty offender, it is possible to calculate a relative risk of<br />

murder for any other offender. If any of the risk factors above appear in the criminal<br />

history of the offender, then we multiply the relative risks together. If none of the risk<br />

factors appear, then the relative risk would be 1.0. One example is considered below:<br />

Example A. A 25-year-old active offender with prior offences for arson, actual<br />

bodily harm <strong>and</strong> indecent assault of a female. This last offence leading to a<br />

custodial sentence of six months.<br />

The offence of arson has a relative risk contribution of 1.987; the offence of actual<br />

bodily harm belongs in offence category 8 – other wounding etc. – which has a<br />

relative risk contribution of 1.329. The offence of indecent assault of a female is<br />

58

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!