NORTHERN IRELAND DRINK DRIVING RECONVICTION ANALYSIS OF THOSE REFERRED ONTO A COURSE FOR DRINK DRIVING OFFENDERS (2001-2009) - September 2010
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
NORTHERN IRELAND DRINK DRIVING RECONVICTION ANALYSIS OF THOSE REFERRED ONTO A COURSE FOR DRINK DRIVING OFFENDERS (2001-2009) September 2010
Table of Contents Page Executive Summary 2 Background 4 Previous Research 5 Methodology 7 Results 10 Appendices Appendix A: Drink Driving Offence Codes 17 Appendix B: Technical Notes 19 1
Executive Summary In Northern Ireland between 1 January 2001 and 30 June 2006, 4,637 of those referred by a magistrate onto the ‘Course for Drink Drive Offenders’ were matched successfully with conviction data from the Causeway Criminal Record Viewer to enable a reconviction analysis to be carried out. Of those successfully matched, 1,452 (31%) completed the course, 69% did not complete a course. The offenders in the database were all convicted before 1 July 2006, providing a minimum period of three years between the initial conviction date for the criterion offence and 30 June 2009 (the cut off period). Key Findings • A greater proportion of females (35%) than males (31%) who were referred onto a course for drink drive offenders opted to attend and complete a course. • Among offenders who had completed the course, 7.0% were reconvicted of a subsequent drink drive offence compared to 11.4% who did not complete a course. In other words, the likelihood of reconviction was 1.6 times higher if the offender didn’t complete a course. • Whilst there was no gender difference in the likelihood of being reconvicted for all those who completed a course, males who did not complete a course were found to be 1.7 times more likely to reoffend than females who did not complete a course. • Up to two years (24 months) after their original conviction, 1.9% of offenders who had completed a course had been reconvicted of a subsequent drink/drive offence, compared with 6.2% of those who did not complete a course. This represents a reconviction rate that is 3.3 times higher for offenders who did not complete a course compared with those who did. 2
• Up to three years (36 months) after their original conviction, 3.4% of offenders who had completed a course had been reconvicted of a subsequent drink/drive offence, compared with 8.0% of those who did not complete a course. This equates to a reconviction rate that is 2.3 times higher for offenders who did not complete a course compared with those who did. • The effectiveness of the courses in NI appears to broadly reflect the experience in GB where it was similarly observed that the largest impact occurs in the first two years following initial conviction. 3
Background The Road Traffic Offenders (Northern Ireland) Order 1996 provided the opportunity to trial the use of rehabilitation schemes for drink drivers. These schemes were developed as an educational intervention aimed at reducing the likelihood of re-offending by individuals who have been convicted of a drink drive offence. Courts can be given powers to reduce the period of disqualification (up to a maximum of 25%) if the offender satisfactorily completes a rehabilitation course approved by the Department of Environment Northern Ireland (DoE NI). The Northern Ireland (NI) Courses for Drink Drive Offenders (CDDO) pilot scheme formally commenced on 1 April 1998, in two designated Magistrate Court areas - Belfast and Newtownabbey. Although Probation Board for Northern Ireland (PBNI) was the sole course facilitator, there was also a multi- agency element in the courses, with input from the Police Service of Northern Ireland (PSNI), DoE Road Safety Education Officers, Campaign Against Drink Drivers and Road Trauma Support. Legislation was introduced on 1 January 2001, to extend the existing pilot scheme until 31 December 2005, with the NI Court Service extending the authority to refer relevant offenders to all Magistrate courts in Northern Ireland. This scheme was made permanent following the successful evaluation in 2005. From the outset, the evaluation of the scheme has been core to its implementation, with the aim being; “to allow the use of such courses, as a sentence mechanism, to be evaluated in terms of their effectiveness in reducing drinking and driving”. This latest evaluation was commissioned by DoE NI, by way of the CDDO working group. The CDDO working group is chaired by an official from DOE Road Safety Division and has included representatives from the Department of Justice, the Public Prosecution Service, the Probation Board, the Driver and Vehicle Licensing Agency, Road Safety Division and Central Statistics and Research Branch. The purpose of this type of evaluation is to add to the qualitative information already gathered and to inform the CDDO working group of the effectiveness of the scheme by studying reconviction rates. 4
Previous Research Northern Ireland A reconviction study was carried out in Northern Ireland in 2005 to determine whether the CDDO had any effect on the relative reconviction rates of those who were referred onto and completed a course compared to those referred onto but who did not complete a course. Between 1 April 1998 and 31 December 2000, 370 people were referred by a magistrate onto a course. Thirty-two percent completed a course and 68% did not complete a course. Up to two years (24 months) after their original conviction, 0.9% of offenders who had completed a course had been reconvicted of a subsequent drink/drive offence, compared with 5.7% of those who did not complete a course. The reconviction rate of those who did not complete a course was 6.3 times the rate of those who did complete a course. Up to three years (36 months) after their original conviction, 2.6% of offenders who had completed a course had been reconvicted of a subsequent drink/drive offence, compared with 9.8% of those who did not complete a course. The reconviction rate of those who did not complete a course was 3.8 times the rate of those who did complete a course. Great Britain Department for Transport (DfT) (formerly DETR) commissioned the Transport Research Laboratory (TRL) to undertake research into calculating reconviction rates for this scheme in Great Britain (GB). Results of this were published in 1999 and can be found in TRL report 426, ‘Drink/driver rehabilitation courses in England and Wales’. In short, TRL studied 3 groups of people who were convicted between 1993 and July 1996: • Those convicted and who attended a course (at an experimental court). • Those convicted and not attending (at an experimental court). • Those convicted at a control court (where the scheme was not available). These groups where ‘followed’ until December 1998, allowing TRL to base reconviction rates on at least 29 months after, and in some cases 5 years after the first offence. 5
The purpose of the control group was to test whether any observed difference in reconviction rates between those attending and not attending courses could be due to ‘subject selection bias’, i.e., the possibility that offenders predisposed against reoffending were more likely to be offered, accept and complete training than offenders predisposed to reoffend. In the event, whilst no real difference in reconviction rates was found between all offenders convicted in the experimental courts and those in the control courts, further more detailed analysis concluded that subject selection bias only accounted for a small element of the differences in reconviction rates reported below. Twenty-four months after their original conviction, only 2% of offenders who had attended courses had been convicted of a subsequent drink/drive offence, compared with 6% of those who had not in the experimental courts. The reoffending rate of those who did not attend a course was 3 times the rate of those who did. Thirty-six months after their original conviction, only 3.4% of offenders who had attended courses had been convicted of a subsequent drink/drive offence, compared with 9.6% of those who had not in the experimental courts. The reoffending rate of those who did not attend a course was 2.8 times the rate of those who did, still very similar to the twenty-four month ratio. TRL carried out a further study in GB, the results of which were published in 2003 and can be found in TRL report 574, ‘Reconvictions for drink/drive course attenders: a six year follow up’. This report examines the reconvictions of the original sample of offenders tracked in the experimental courts, in report 426, only over a longer period of time providing at least 72 months of data for over 80 percent of the sample. TRL’s 6-year follow-up study found that after 72 months, nearly 2.5 times more non- course attenders than course attenders had reoffended (17.9% of non-course attenders compared with 7.6% of course attenders). This ratio of rate of reconvictions for non- course attenders to the rate of course attenders was found to be similar at 3, 4, 5 and 6 years after conviction. 6
Methodology In the original NI research, court data was not available for those people who were not offered the scheme from 1 April 1998 to 30 December 2000, and therefore it was not possible to have a control group. Whilst the GB study could not be replicated in its entirety, DRD statisticians followed its basic structure. The new research carried out in 2010 was set-up primarily to replicate the original NI research, therefore the same methodology has been applied, i.e, without a ‘no course offered’ control group. In GB it was concluded that ‘subject selection’ was not a significant source of bias. It is not unreasonable to assume that this conclusion would similarly apply in NI. However, it would be preferable that this actually be tested in the local context. It is anticipated that, with increased resources and availability of court data, further analysis will be carried out in this area which will allow this possibility to be properly investigated in future. The initial evaluation in 2005 sought to inform the CDDO Working Group of the effectiveness of the course, with the reconviction analysis supplementing the attendees’ pre and post course knowledge of alcohol and attitudes to drinking and driving. The analysis carried out in 2010 focuses solely on reconviction rates. For the purposes of the original first wave NI evaluation, the groups of drivers followed were those who had been convicted of a relevant drink driving offence between 1 April 1998 and 31 December 2000 (the pilot period) and were referred by a magistrate to attend one of the drink drive courses. The pilot period was chosen for the study so that reconviction rates could be examined over the maximum time period possible at that time. The drivers in this group were then tracked up to 30 June 2004. This allowed an investigation of the relative reconviction rates of drink/drive offenders who, once referred, had attended a course compared to those who had not. This new analysis is based on a much larger population of those convicted of a relevant drink drive offence during the period 1 January 2001 and 30 June 2006 and referred by a magistrate to attend a drink drive course. The referrals were tracked up to 30 June 7
2009. These time periods were selected to ensure that each person referred had a minimum of a three year period in which they could be reconvicted. Data The original evaluation used two main sources of data: i) The PBNI referral database (CDDO) - which holds details of all those persons convicted of a relevant drink drive offence (see Annex A) and who were referred onto a course by a magistrate. ii) Causeway Criminal Record Viewer (CRV). - This database records information of all offences and disposals in all NI courts. These data sources were utilised again for the 2010 analysis. Between 1 January 2001 and 30 June 2006, PBNI had provided DRD with details of 5,967 people who were referred by a magistrate onto a CDDO course. The CDDO referral database was used as the ‘population’ and the individual criminal records of those sentenced between 1 January 2001 and 30 June 2006 were accessed via an automated match with the CRV by DRD and Department of Justice (DoJ) (formerly Northern Ireland Office) statisticians. A ‘unique’ identifier for each referral was populated initially using the surname, forename and date of birth, as given on the CDDO referral database and the CRV. A further match with ’date of sentence’ was carried out. Despite some data issues, particularly with regard to incomplete and missing name/address information, a total of 4,637 individual referrals were matched successfully providing an overall match rate of 78%. Referrals that were not matched have been disregarded from this analysis. There is no reason to believe non-matches would be anything other than random and their exclusion therefore should not introduce any systematic bias into the study. Because of the exclusions, it is important to be aware that any references to resultant matched number of referrals do not represent the total number of all those referred onto CDDO courses. 8
A reconviction database was created by DRD and DoJ with the following additional information being added: - Date of any further reconvictions for relevant drink driving offence - Indication of whether individual has been reconvicted of relevant offence - Number of reconvictions - End date of 30 June 2009 (research cut off period) - Indication of whether CDDO course had been completed - Reconviction period between primary and subsequent offence Responsibility for the maintenance of the referral database during this period resided with Central Statistics and Research Branch, DRD. Statisticians from DRD and DoJ agreed the search and match methodology outlined previously, and the recording of additional information on the database from the CRV. Access to the CRV was controlled by DoJ statisticians. Permissions to match and access the new database was managed through an access agreement between DRD and DoJ. Once the referral and reconviction database was complete with all necessary information, DoJ transferred it to Central Statistics and Research Branch DRD so that the necessary analysis and evaluation could be carried out. The database is securely stored within the branch with access restricted to only those Road Safety statisticians who need to undertake any analysis. The DRD and DoJ staff involved with the study are on secondment from the Northern Ireland Statistics and Research Agency. This ensures that the data collection, matching, validation, analysis and evaluation were carried out by independent statisticians bound by the Code of Practice for Official Statistics. 9
Results Note that not all percentages presented in this section may sum to 100 due to rounding. Study Group - Profile Out of the 4,637 individuals matched and which formed the study population, 3,185 (69%) opted not to attend a course and 1,452 (31%) opted to attend and actually complete a course. See Figure 1. Figure 1: Number of people completing the course by gender 3500 3185 3000 2742 Number of People 2500 2000 1452 1500 1217 1000 443 500 235 0 Yes No Completed Course Male Female Male and Female Whilst the overwhelming majority of course referrals during the study period were males (85%), a greater proportion of females (35%) than males (31%) actually completed the course. The most common period of disqualification for all who were referred onto a course was 12 months, with 79% of offenders being disqualified from driving for this length of time. Thirteen percent were disqualified for more than 12 months and less than or equal to 24 months, 6% were disqualified for more than 24 months, but less than or equal to 36 months and 1% were disqualified for over 36 months. The offenders in the database were all convicted before 1 July 2006, hence the minimum time between conviction for the criterion offence and 30 June 2009 (research cut off period) was 36 months with the maximum being 102 months (or 8.5 years). See Figure 2 and Table 1 overleaf. 10
Figure 2: Number of months between criterion conviction offence and 30 June 2009 1800 1674 1600 1400 Number of People 1200 1027 1021 915 1000 800 600 400 200 0 36-48 49-60 61-72 73+ Number of Months Table 1 – Time Profile between criterion offence and 30 June 2009 Months Number % 36-48 1027 22.1 49-60 1021 22.0 61-72 915 19.7 73+ 1674 36.1 Total 4637 100 Note: Due to rounding percentages may not sum to 100. In order to ensure that any observed difference in reconviction rates, between those who completed courses and those who did not, was not simply due to a differential length of time available in which to reoffend, the time of first conviction profile of each group was also compared (See Table 2). Table 2 – Percentage of first convictions by year of conviction and course completion status Year of First Conviction Individuals Completing a Individuals Not Course (%) Completing a Course (%) 2001 10 12 2002 13 15 2003 21 19 2004 21 21 2005 23 21 2006 12 11 All First Convictions 1452 3185 11
Whilst there are small differences between profiles (typically 2 percentage points or less), statistical testing shows that there is no overall systematic difference between the groups. This provides additional evidence that any difference in the overall reconviction rates can be attributed to the influence of the course. Reconviction Rates Overall, between 1 January 2001 and 30 June 2009, 466 people out of the 4,637 study population were reconvicted of a further drink/drive offence – 10%. Among offenders who had completed the course, 102 out of 1,452 were reconvicted - 7.0%. The proportion of males being reconvicted after completing a course was the same as that for females. Among offenders who were referred, but who did not complete a course, 364 out of 3,185 were reconvicted - 11.4%. The percentage of males who did not complete a course and were reconvicted (12%) was 1.7 times greater than the percentage of females who did not complete a course and were reconvicted (7%). It would seem reasonable to conclude, therefore, that whilst there is no gender difference in the reconviction rate of males and females who have completed a course, non-completion would appear to have a bigger impact on males in terms of increased likelihood of reoffending. Over the entire study period, offenders who did not complete a course were 1.6 times more likely to be reconvicted for a subsequent drink/drive offence, compared with offenders who had completed a course. For the 466 persons in the study group who were eventually reconvicted, Figure 3 graphically illustrates the time between the primary and subsequent offence for completers versus non-completers while Table 3 provides further detail in terms of numbers and percentages. It should be noted that, apart from in the 60-72 month category, the apparent small differences between completers and non-completers in the proportions reconvicting after the first 2 years are not statistically significant. 12
Figure 3: Time between primary and subsequent offence by course completion status 3.0% 2.5% 2.0% 1.5% 1.0% 0.5% 0.0% 0-6 months >6 months >12 months >24 months >36 months >48 months >60 months >72 months
Up to 2 years (24 months) - the reconviction rate of those who did not complete a course was 3.3 times the rate of those who did complete a course. Up to 3 years (36 months) - the reconviction rate of those who did not complete a course was 2.3 times the rate of those who did complete a course. The reconviction rates for up to four years (48 months) were 4.9% and 9.5% for course completers and non-completers respectively while up to five years (60 months) they were 5.8% and 10.5%. For both time periods, the likelihood of reoffending is just under 2 times greater for those who complete the course. However, as the four and five year rates are based on an increasingly smaller subset of the total study group, they are not as robust as the estimates up to 3 years which comprised the total group. An alternative way of looking at this is presented in Figure 4 overleaf which illustrates that a larger proportion of course completers remained reconviction free across all time periods. It also shows the markedly increasing rate of reconviction within the first two years for non-completers compared to those who completed a course. The reconviction rate then appears reasonably consistent between the groups for the remainder of the study period. This suggests that the effectiveness of the course peaks in the first two years following conviction, a finding also highlighted in Figure 3 above. Figure 4: Proportion WITHOUT further drink drive convictions by course completion status 1.00 0.98 Proportion without reconvictions 0.96 0.94 0.92 0.90 0.88 0.86 0.84 0.82 0-6 months >6122436486072 months months months months months months months Time since criterion offence Course completed Course not completed 14
Of the 1,452 people who did complete a course, 1,350 (93.0%) had no further convictions during the study period, 91 had one further reconviction (6.3%) and 11 had 2-3 more convictions (0.8%). Of the 3,185 people who were referred but who did not complete a course during the study period: 2,821 (88.6%) had no further convictions; 319 (10.0%) had one further reconviction; 32 (1.0%) had two further reconvictions; 8 (0.3%) had three further reconvictions; and 5 (0.2%) had more than three further reconvictions. Comparison with NI First Wave Results and GB This research reports the reconviction rate of those completing a course to be 3.3 times that of those not completing a course, for the period up to 2 years after their criterion conviction. This ratio falls to 2.3 for the period up to 3 years after conviction. The original research carried out in NI reported the reconviction rate of those completing a course to be 6.3 times that of those not completing a course for the period up to 2 years after their criterion conviction. This ratio fell to 3.8 for the period up to 3 years after criterion conviction. These ratios for up to 2 years and up to 3 years are evidently lower in this new research. However, it should be borne in mind, that the significantly larger sample size in the current research will yield much more robust results. The research carried out in GB in 1999 by TRL found that 2% of course attenders and 6% of non-attenders had been reconvicted over the period up to 2 years after criterion conviction. For the period up to 3 years after criterion conviction TRL found that 3.4% of course attenders and 9.6% of non-course attenders had been reconvicted. This translates into a ratio of 3 for the period up to 2 years after criterion conviction and a ratio of 2.8 for the period up to 3 years. Apart from perhaps the year 3 reconviction rate for non-attenders, these results are strikingly similar to the current NI 2 and 3-year rates for attenders (1.9% and 3.4%) and non-attenders (6.2% and 8.0%) noted above. 15
TRL’s 6-year follow-up study found that after 72 months, nearly 2.5 times more non- course attenders than course attenders had reoffended (17.9% of non-course attenders compared with 7.6% of course attenders). This ratio of rate of reconvictions for non- course attenders to the rate of course attenders was found to be similar at 3, 4, 5 and 6 years after conviction. It fell from just under 4 in the period up to 2 years after criterion conviction to just under 3 in the period up to 3 years after criterion conviction with only a relatively smaller decline to 2.5 after 6 years. Whilst these latest findings might suggest a stronger impact of the courses in GB, it should be noted that the GB results are based on a much larger study group and hence will be more robust than the equivalent NI findings. The trend identified in NI is broadly comparable to that found in the GB research above i.e., that the greatest impact of the course occurs in the first 2 years followed by a levelling off in terms of course effectiveness. TRL, in 2003, highlighted that the impact of the courses appears to be more effective during the driver disqualification period. Whatever the reason, the positive impact noted in the original NI research has been reaffirmed in this latest study. 16
ANNEX A Drink Driving Offence Codes N.I.OFFENCE CODEOFFENCE DESCRIPTION 50 08 044 Failure to provide specimen of blood/urine - major 50 08 046 Driving with excess alcohol 50 08 048 In charge with excess alcohol 50 08 049 Driving with excess alcohol – blood /urine 50 08 052 Failing to provide specimen of breath - driving 50 08 053 Failing to provide specimen of breath – attempting to drive 50 08 123 Driving while unfit by drink or drugs 50 08 124 Attempting to drive while unfit by drink or drugs 50 08 125 In charge of a vehicle while unfit by drink or drugs 50 08 126 Driving with excess alcohol in breath 50 08 127 Attempting to drive with excess alcohol in breath 50 08 128 In charge of a vehicle with excess alcohol in breath 50 08 129 Driving with excess alcohol in blood 50 08 130 Attempting to drive with excess alcohol in blood 50 08 131 In charge of a vehicle with excess alcohol in blood 50 08 132 Driving with excess alcohol in urine 50 08 134 In charge of a vehicle with excess alcohol in urine 50 08 136 Failing to provide a specimen of breath for analysis 50 08 137 Failing to produce a specimen of blood or urine for analysis 50 08 138 Failing to produce a specimen when driving unfit 50 08 139 Failing to provide a specimen when attempting to drive unfit 50 08 140 Failing to provide a specimen when in charge unfit 50 08 141 Failing to provide a specimen when driving with excess alcohol 50 08 142 Failing to provide a specimen when attempting to drive with excess alcohol 50 08 143 Failing to provide a specimen when in charge with excess alcohol 50 08 204 Causing death by driving carelessly when unfit 50 08 206 Cause GBI by driving carelessly when unfit 50 08 207 Cause GBI by inconsiderate driving when unfit 17
N.I.OFFENCE CODEOFFENCE DESCRIPTION 50 08 208 Causing death by driving carelessly with excess alcohol 50 08 209 Causing death by inconsiderate driving with excess alcohol 50 08 210 Cause GBI by driving carelessly with excess alcohol 50 08 211 Cause GBI by inconsiderate driving with excess alcohol 50 08 212 Cause death by driving carelessly and failing to provide specimen 50 08 214 Cause GBI by driving carelessly and failing to provide specimen 18
ANNEX B Technical Notes Data Matching Due to incomplete data and missing variable information such as surnames, date of births, conflicting gender etc. a number of assumptions have been built into the data set which are detailed below: 1) Initial match of CDDO referral database with CRV was on a joined variable containing both DOB and Surname. Initially this match only identified over 3,000 cases. Further matches were then carried out adding in forename and CRV data of conviction to establish a final match of 4,637 cases. 2) A match was carried out to calculate the date of referral from the CDDO database against the conviction date on the CRV database. Based on the results a number of convictions that occurred prior to the CDDO referral date were removed from the database. Where there was a negative result all differences were inspected prior to any changes being made to the file. 3) During analysis there were @ 50 cases where gender conflicted between the two datasets. An individual check against forename was carried out and a gender marker set on this basis. 4) There were approx 90 cases were CRV conviction court date conflicted with the CDDO court date, each of these cases were individually inspected and dates amended as appropriate against other date information where available. For approximately 10 cases a date was selected by the statistician as it was not apparent which date was incorrect. 5) In 4 cases the match with the CDDO database and the CRV populated dates that were vastly apart. For the purposes of this analysis an additional row was added to the dataset for the original referral date and then the latter date was treated as a reconviction. 6) In 32 cases the number of months disqualified were missing from the CDDO database, these missing values were populated from the CRV length of time disqualified variable. 19
7) In the CDDO database there were a number of duplicate cases which were removed from the database and not included in the analysis. These were checked to ensure that they were not re-referrals. Significance Testing The results presented in this report cover a discreet study period and could be regarded as a sample of all persons who have been referred onto a course for drink drive offenders. As such the results will be subject to a degree of sampling error. However, any statements in this report regarding differences between groups, such as those who have completed a course and those who have not, are statistically significant at the 95% significance level. This significance level is the usual threshold for judging whether differences between groups might have arisen by chance. It means that there is only a 5%, or one in twenty, probability that an observed difference might have arisen by chance factors alone. 20
You can also read