Reconviction rates in Scotland: 2016-2017 offender cohort

Annex B – Sources of information, data quality and confidentiality, and revisions

Sources of information

B1 Information presented in this bulletin is based on data held in the Scottish Offenders Index (SOI), which is derived from the data used in the Criminal Proceedings in Scotland statistical bulletin. The Criminal Proceedings data is in turn derived from information held on the Criminal History System (CHS) which is owned by Police Scotland.

B2 The SOI currently contains a record of criminal proceedings against individuals (excluding companies) in Scottish courts as well as information on non-court disposals. The court convictions and non-court disposals are held in separate datasets by the Scottish Government and so are independent of each other.

B3 The data in the SOI on court convictions currently covers all convictions where a sentence was imposed since the beginning of 1989, and the main offence involved was either a crime in Groups 1-5, or some of the offences in Group 6, in the Scottish Government’s classification of crimes (see Annex D of the Criminal Proceedings Bulletin for further information about these classification groups). Minor offences (such as drunkenness, and almost all motor vehicle offences) are not included in the SOI.

B4 The distinction between crimes and offences is made only for statistical reporting purposes. Although the violations allocated under “crimes” tend to be more serious, there are some “offences” that have more severe punishments associated with them e.g. drink driving is classified under “offences” in the “motor vehicle offences” section rather than under “crimes”. Groups 1-5 of the Scottish Government’s classification covers non-sexual crimes of violence, sexual crimes, crimes of dishonesty, fire-raising, vandalism etc. and other crimes. The offences in Group 6 which are included in the court conviction SOI are: common assault, breach of the peace, threatening or abusive behaviour, offence of stalking, offensive behaviour at football and threatening communications (under the Offensive Behaviour at Football and Threatening Communication Scotland Act 2012), racially aggravated harassment or conduct, miscellaneous firearms offences, and social security offences. See Annex Table A2 for a more detailed list of the types of crimes in the SOI.

B5 The data in the SOI on non-court disposals covers all crimes and offences, including motor offences, where a non-court disposal was given, back to 2008.

B6 Each record on the SOI database includes information on the sex and date of birth of the offender, the dates of conviction and sentence, the main crime or offence involved and details of the sentence imposed. Information is also available on any crimes which were additional to the main crime. Each offender has a unique reference number, which allows individual convictions for that offender to be linked together. The SOI does not include the name and address of an offender, except the first half of an offender’s postcode. A privacy notice is available on the Scottish Government’s website.

B7 All but the most serious offences alleged to have been committed by children under the age of 16 are generally dealt with outside of the courts in the children’s hearings system, or by Early and Effective Interventions or other non-court disposals. The SOI does not hold data taken from the children’s hearings system.

B8 The Scottish Offenders Index (SOI) currently contains data on 584,000 offenders and 2,193,000 convictions since SOI records began in 1989. Eighty-five percent of the offenders are male and 15 percent are female.

Data quality

B9 The figures in the bulletin have been derived from administrative IT systems which, as with any large scale recording system, are subject to possible errors with data entry and processing. During the production of this bulletin we have put in place processes to ensure that the data are fit for purpose for this publication, which are listed below.

B10 Data standards are adhered to by organisations inputting data to the CHS in terms of the definitions of data items and their corresponding values. These standards are agreed under the Justice Digital Strategy and ensure there is consistency across the justice organisations in the information they collect. Further information on the data standards can be found in the Integration of Scottish Criminal Justice Information Systems (ISCJIS) data sharing manual.

B11 The following protocols also ensure consistency in the data collected:

• The Scottish Courts and Tribunal Service (SCTS) protocol for the handling of errors that may occur in the transmission of data between justice partners’ databases;
• The protocol for the investigation/resolution of disputed data between Police Scotland and the Crown Office and Procurator Fiscal Service (COPFS);
• The protocol for sharing electronic information between justice partners.

B12 The Scottish Government has representation on a data quality group and is kept informed of any data quality issues relating to the CHS. This group meets around three times a year and also has representation from Police Scotland, COPFS, the Scottish Prison Service, and SCTS.

B13 The analyses in this bulletin are based on the data published in the Criminal Proceedings in Scotland statistical bulletin, and the quality assurance process is described in the section on Data quality: Data validation during production of the statistical bulletin in Annex B of the Criminal Proceedings bulletin. In summary, the validation processes include automated and manual checks on the data. Any unusual or missing values are referred back to either Police Scotland, the SCTS, or COPFS. The figures are also checked against case processing information published by COPFS and management information provided by SCTS to ensure that the court volumes are consistent. Police Scotland, SCTS, COPFS and policy experts within the Scottish Government are also consulted to give insight on an operational level and provide insight into why any significant changes may have occurred. The figures are also checked by Scottish Government statisticians, who have not been involved in the production process, and they may highlight any issues that may have gone unnoticed.

B14 During the production of this bulletin, the data undergoes processing to calculate the frequency and prevalence of reconvictions. The numbers are manually checked to determine whether there are any unusual values, and if so, then the calculations are rechecked. Like the Criminal Proceedings in Scotland publication, the figures in this bulletin have undergone a further round of checks by Scottish Government statisticians, and policy experts within the Scottish Government are again consulted to provide insight and context to any significant changes in the figures.

Postcode data quality

B15 From the 2015-16 cohort publication, we have published reconviction rates for local authorities based on offenders’ home postcode (Table 13) alongside the reconviction rates for local authorities based on court location (Table 12).

B16 Previously we could only calculate reconviction rates for local authorities based on the court location of where offenders were convicted, as this was the only geographical information that we had. This approach has limitations, because some court areas cover multiple local authorities and offenders may be convicted for crimes committed in different areas to where they live. In particular, these figures are not fully suitable for local authority community justice partners, as they have a need for reconviction figures for offenders who live in their area for planning purposes, such as: schemes to reduce reoffending, or estimating the number of offenders that social workers need to supervise in their area.

B17 We started receiving data from Police Scotland on the first half of an offender’s postcode in 2014, and coverage was high enough from 2015-16 onwards to start using the data. With the first half of an offender’s postcode we can match offenders to their home local authority and calculate reconviction rates based on offenders’ home local authority.

B18 We are classifying the reconviction rates for local authorities based on the postcode data as Experimental Statistics: Data being developed, as the data are of insufficient quality to be labelled as National Statistics as the rest of the data in this publication are. We did not have postcodes for a fifth of offenders (20%) with an index conviction in 2016-17, and we are unsure of how this may affect the reconviction rates.

B19 Postcodes may be missing because offenders have no fixed abode, but it may also be a recording issue. We are also less likely to have the postcode of offenders released from a custodial sentence, as we would have received their conviction data before we started collecting the postcode. The missing category also includes a small number where postcodes were supplied, but they were not valid postcodes. We are working with Police Scotland to improve the coverage, and as the accuracy improves we will remove the experimental label.

B20 Annex Table B1 shows the percentage of offenders with missing postcodes in each local authority group based on location of the court they were convicted in. It also shows the percentage of offenders living in each local authority that make up the local authority group. This shows that there are significant percentages of offenders who are convicted in a court that covers a different local authority to where they live.

Table B1 Postcode coverage for each local authority group

Local authority group based on court location1	Offender's local authority2	Number	% of LA group
Aberdeen City and Aberdeenshire	Aberdeen City	1870	57%
	Aberdeenshire	949	29%
	other LA	259	8%
	Missing	182	6%
Angus	Angus	538	76%
	other LA	93	13%
	Missing	78	11%
Argyll & Bute	Argyll & Bute	206	62%
	other LA	51	15%
	Missing	73	22%
Ayrshire, East, North and South	East Ayrshire	710	20%
	North Ayrshire	712	20%
	South Ayrshire	462	13%
	other LA	238	7%
	Missing	1375	39%
Clackmannanshire	Clackmannanshire	298	71%
	other LA	60	14%
	Missing	63	15%
Dumfries & Galloway	Dumfries & Galloway	955	78%
	other LA	175	14%
	Missing	96	8%
Dunbartonshire, East and West	East Dunbartonshire	25	3%
	West Dunbartonshire	482	52%
	other LA	186	20%
	Missing	231	25%
Dundee City	Dundee City	1196	66%
	Fife	216	12%
	other LA	186	10%
	Missing	210	12%
East Lothian	East Lothian	6	67%
	other LA	1	11%
	Missing	2	22%
Edinburgh and Midlothian (and East Lothian)	East Lothian	442	10%
	Edinburgh, City of	2144	50%
	Midlothian	445	10%
	other LA	488	11%
	Missing	791	18%
Falkirk	Falkirk	831	69%
	other LA	213	18%
	Missing	166	14%
Fife	Fife	1931	80%
	other LA	124	5%
	Missing	345	14%
Glasgow City	Glasgow City	4095	53%
	East Dunbartonshire	221	3%
	North Lanarkshire	188	2%
	South Lanarkshire	534	7%
	other LA	765	10%
	Missing	1896	25%
Highland	Highland	1054	74%
	other LA	169	12%
	Missing	209	15%
Inverclyde	Inverclyde	349	50%
	Argyll & Bute	24	3%
	other LA	71	10%
	Missing	258	37%
Lanarkshire, North and South	North Lanarkshire	1755	38%
	South Lanarkshire	1369	30%
	other LA	424	9%
	Missing	1068	23%
Moray	Moray	473	84%
	other LA	54	10%
	Missing	34	6%
Na h-Eileanan Siar	Na h-Eileanan Siar	75	75%
	other LA	12	12%
	Missing	13	13%
Orkney Islands	Orkney Islands	72	77%
	other LA	8	9%
	Missing	14	15%
Perth & Kinross	Perth & Kinross	546	66%
	other LA	196	24%
	Missing	85	10%
Renfrewshire and East Renfrewshire	East Renfrewshire	134	9%
	Renfrewshire	683	45%
	other LA	238	16%
	Missing	473	31%
Scottish Borders	Scottish Borders	354	78%
	other LA	49	11%
	Missing	48	11%
Shetland Islands	Shetland Islands	107	69%
	other LA	25	16%
	Missing	24	15%
Stirling	Stirling	359	56%
	other LA	187	29%
	Missing	99	15%
West Lothian	West Lothian	788	74%
	other LA	137	13%
	Missing	134	13%
High Court	Postcode present	372	59%
	Missing	258	41%

1. Local authority group is based on the court location of where an

offender was convicted.

2. The offender's local authority is based on their postcode. Only authorities in the local authority group are shown. Those that live in other local authorities are grouped under "other LA". Missing is the percentage of offenders convicted in the local authority with a missing postcode.

Domestic abuse data quality

B21 Reconviction rates for domestic abuse crimes and offences were first published as an adhoc analysis of the previous 2015-16 offender cohort in April 2019. Therefore this is only the second time that reconviction rates for domestic abuse have been published.

B22 The domestic abuse identifier used to identify domestic abuse and non-domestic abuse crimes in these statistics has been used in the Criminal Proceedings in Scotland National Statistics publication for several years. Also the process to calculate reconvictions is the same as used to calculate reconvictions elsewhere in the bulletin. This means that relevant quality assurance has been carried out on all the data used in these figures. However, as these data have only recently started to be used in combination in these statistics, there may be some minor data quality issues that we may not have identified yet.

B23 A new statutory domestic abuse aggravation was created by the Abusive Behaviour and Sexual Harm (Scotland) Act 2016 and introduced in 2017-18 which would apply to some of the reconvictions for the 2016-17 cohort. This is also recorded by the police or COPFS. As this statutory aggravation is not currently applied to a conviction without the non-statutory identifier, this will not affect comparability before and after the introduction of the statutory aggravation.

Data confidentiality

B24 We have completed a data protection impact assessment to assess and mitigate any privacy issues with the collection, use, and publication of reconviction data.

B25 A privacy notice is available on the Scottish Government’s website covering the data used in this publication.

B26 The Scottish Government are legally able to collect the data for this publication as Section 84(1 & 2) of the Police and Fire Reform (Scotland) Act 2012 states that the “Scottish Police Authority must provide Scottish Ministers with reports, statistics or other information relating to the Police Service or functions, and the state of crime”.

B27 Court proceedings are held in public and may be reported on by the media, unless the court orders otherwise, for example where children are involved. While our aim is for the statistics in this bulletin to be sufficiently detailed to allow a high level of practical utility, care has been taken to ensure that it is not possible to identify an individual and obtain any private information relating to them.

B28 We have assessed the risk of individuals being identified in the tables in this bulletin and have established that no private information can be identified. Where demographic information is provided this is done in wider categories.

B29 Some of the additional data tables we provide alongside this publication have local authority information related to the offender. In the local authority tables, either demographic information is provided or offence-level information is provided, but not a combination of both.

B30 To maintain the security and confidentiality of the data received from the data suppliers, only a small number of Scottish Government statisticians and statistical support officers have access to the datasets outlined in the various stages of processing outlined above. The only personal details received by the Scottish Government in the data extract are those which are essential for the analyses in this bulletin and do not include the names of offenders.

Revisions

B31 The CHS is not designed for statistical purposes. It is dependent on receiving timely information from the SCTS, COPFS, and the police. It should also be noted that some types of outcome are removed from the system after a prescribed length of time. A pending case on the CHS is updated in a timely manner, but there are occasions when a slight delay may happen. Recording delays of this sort generally affects high court disposals relatively more than those for other types of court. The figures provided in this bulletin reflect the details of court proceedings recorded on the CHS and supplied to the Scottish Government by the end of September 2018 to allow later reconvictions in 2017-18 to be captured on the CHS.

B32 The CHS is regularly updated, so subsequent analyses will result in revised figures (shown in Annex Table B2) as late records are added. The first revision of the reconviction rate in the following year’s bulletin was typically around 0.4 percentage points higher than the figures published initially, although this has improved recently and the revised figure for last year was 0.2 percentage points higher. The first revision of the average number of reconvictions per offender is typically 0.01 to 0.02 higher than the initially published figure.

Table B2 Revisions to reconviction rates

Cohort	Initial published figures		1st revision of published figures		2nd revision of published figures		3rd revision of published figures
	Reconv. rate	Av. no. of reconvs. per offender	Reconv. rate	Av. no. of reconvs. per offender	Reconv. rate	Av. no. of reconvs. per offender	Reconv. rate	Av. no. of reconvs. per offender
2006-07	32.2	0.59*†	32.4	0.60*†	32.4	0.60†	32.4	0.60†
2007-08	30.9	0.56*†	31.2	0.57†	31.3	0.57†	31.2	0.57†
2008-09	31.0	0.58†	31.5	0.60†	31.5	0.60†	31.5	0.60
2009-10	30.1	0.54†	30.5	0.56†	30.6	0.56	30.6	0.56
2010-11	28.4X	0.50†X	30.1X	0.55X	30.1	0.55	30.1	0.55
2011-12	29.2	0.53	29.6	0.54	29.6	0.55	29.6	0.55
2012-13	28.6	0.51	28.9	0.53	28.9	0.53	28.9	0.53
2013-14	28.3	0.51	28.5	0.52	28.5	0.52	28.5	0.52
2014-15	28.2	0.50	28.3	0.50	28.4	0.51	-	-
2015-16	27.0	0.47	27.2	0.48	-	-	-	-
2016-17	27.0	0.47	-	-	-	-	-	-

* These figures were not published initially, but it is possible to determine their magnitude retrospectively.

† These figures have been previously reported as the reconviction frequency rate, which was the number of reconvictions per 100 offenders. Therefore these figures are the original figure divided by 100 to get the average number of reconvictions per offender.

X From the 2011-12 bulletin, some new offences that came into effect from 2010 were included in calculations for the reconviction rate and the average number of reconvictions. These offences were not included in calculations in the 2009-10 or 2010-11 cohort bulletins. The increase in the numbers after revision led to a slightly higher increase at the 1st revision for the 2010-11 cohort than it had been in previous and subsequent years.