07 May 2008
An explanation of changes to the way the government is measuring reoffending rates.
Up to now, the statistics have shown how many offenders commit further offences following commencement of a community sentence, or release from prison for adults. For juveniles it includes offenders commencing pre-court disposals, non-custodial court disposals or release from custody.
The figures will now show how many re-offences are committed - that is, the frequency of reoffending. This will be the new headline measure against which our targets will be set.
They will also show how many serious re-offences are committed. The definition 'serious' covers a small number of re-offences that are defined as the most serious such as serious violence (grievous bodily harm, murder, manslaughter), and serious sexual offences.
How the statistics are measured
The measure does not track all the offenders in a given year, but instead looks at a 'cohort'. The cohort is made up of all offenders discharged from a custodial sentence or starting a community sentence in the first three months of each calendar year (January to March). This cohort usually consists of between 40,000 and 50,000 offenders.
The results drawn from the cohort are not representative of reoffending over the whole year but provide a point of comparison which can be used to indicate progress in reducing reoffending. The same time period is used each year to ensure consistent and fair comparisons can be made over time.
Data is obtained from the Police National Computer on whether or not the offender is proven to have reoffended during a one-year follow up period, the number of offences they have committed and the number of serious offences committed.
The frequency of reoffending is produced by calculating the number of proven re-offences per 100 offenders. The same applies for re-offences classified as serious. Calculating a frequency per hundred offenders allows comparisons to be made from one year to another despite fluctuations in the size of the cohort.
Why the statistics are changing
The old measure can tell us whether or not an offender re-offended but not how many times, or whether or not a re-offence was serious. Under the new measures we will be able to distinguish between high volume and low volume offending, and serious re-offences. This will allow better targeting of resources.
This broader range of information on reoffending rates will allow for better, more timely analysis of reoffending and help to reduce crime and make communities safer.
What the old measure shows
The old measure looks at the proportion of offenders who reoffend, rather than the number of re-offences committed. This is measured within a given period (two years for adults, one year for juveniles) following a community sentence being given; or release from prison, or a pre-court disposal, non-custodial court disposal or release from custody for juveniles.
Under the old measure, the reoffending rate is compared against a 'predicted rate'. If the rate is lower than predicted, then reoffending has fallen.
The predicted re-offending rate is worked out based on how statistically likely offenders are to reoffend. This is based on factors such as age, gender, and criminal history.
This old measure has been helpful in providing an accurate picture of the difference that work to reduce reoffending has made to reoffending rates. We will continue to monitor progress using predicted rates and measure progress against the 2002-2006 target.
When new statistics will be published
On Friday 9 May the latest adult reoffending figures for 2005 will be published using the new reoffending measures. The National Statistics bulletin will also contain recalculated reoffending figures using the new metric from 2000-2004 for adults and 2000-2005 for juveniles (excluding the year 2001) so that meaningful comparisons can be made under the new measures.
We are reducing the time it takes to report adult reoffending data. Under the new measures, the amount of time after conviction or release from prison in which re-offences are by adult offenders are recorded has been reduced from two years to one year. The time allowed for convictions to be made and processed has also been reduced from nine to six months. This will provide a more timely measure of adult reoffending and allow us to align reporting of the adult and juvenile reoffending data.
Data calculated under the new measure will be available for the 2006 cohort in September 2008 and for the 2007 cohort in March 2009.
We will continue to publish a reoffending rate which compares the actual and predicted rates on the old two-year reoffending measure until the end of the Strategic Review 2002 target period which covered 2000-2006. Data for the 2006 cohort will be published in March 2009.