Scottish Government

Measurement of the Extent of Youth Crime in Scotland

4 Estimating levels of youth crime

Method

4.1 There are three basic stages in our method for producing estimates of the number of crimes committed by young people.

• Recorded crime data. The most consistent and complete source of data that we have on crime is police force data on recorded crime. This police force data was therefore the starting point for our exercise.
• Adjustment for unrecorded crime. Not all crime is reported to the police or recorded by them, therefore we needed to use other sources to adjust recorded crime data so as to take account of unreported and unrecorded crime.
• Estimating youth crime. The last stage of the method involved making estimates of crimes committed by young people using data from the Children's Hearings system and the courts.

4.2 Following this three-stage approach led to the production of the headline figures in Tables 4.1-4.3 and the approach is further explained below.

Note: Figures may not sum to 100% or to their totals due to rounding.

Recorded Crime

4.3 The first step was to request police recorded crime data. We used the same crime and offence classification system used by the police for recorded crime.

4.4 However, the offences for which we provide a breakdown of the data take into account the types of offences that young people tend to commit. Therefore, we provide more details on those types of crimes most strongly associated with young people, e.g. crimes of dishonesty (which can be further broken down into housebreaking, shoplifting, etc.), than on those crimes that are not particularly associated with young people, e.g. crimes of indecency. We chose to exclude the "other, other crime" category from the analysis as there were less than 150 of such offences, which includes crimes such as conspiracy to cause explosions, in Scotland as a whole in 2002.

4.5 The recorded crime data that we are using is for the last full year of data (2002). As already highlighted, police force data does not record all crime committed, therefore we need to make an adjustment for unrecorded crime.

Adjusting for unrecorded crime

4.6 Police force data was adjusted, or grossed up, using various sources to give a more accurate reflection of the number of actual crimes. These sources included the Scottish Crime Survey (SCS), for estimates of unrecorded crime across a range of personal and household crime types committed against adults. However, given the assumptions involved in deriving these multipliers, some of them unavoidably crude, we felt it more appropriate to provide a range (between two whole numbers) in which these multipliers lie. The exception is for crime categories where the multiplier is very large.

4.7 There is SCS data available for violent crime, acquisitive crime and vandalism that is comparable with recorded crime. We used this data to derive a multiplier for each of these crime types in much the same way as the Home Office did in its study on estimating the costs of crime. Since the latest SCS was conducted in 1999, we compared the SCS crime figures with recorded crime in that year to derive the multipliers.

4.8 However, the SCS does not take into account crimes committed against children under-16 or crimes committed against commercial or public sector interests. Ideally, we would have preferred to separate out crimes and offences against those under-16 and against commercial premises from recorded crime statistics and produced three versions of Table 4.1 for adult victims of crime, young victims of crime and commercial crime, before pooling them back into one table. An appendix to the initial findings from the 2000 SCS compared recorded crime against adults with SCS findings ¹⁸, but it does not differentiate between commercial crime and crime against the under-16s.

4.9 "Counting the Cost", a survey undertaken on crime against business in Scotland provides a measure of unreported commercial crime across a range of crime categories, including vandalism, theft and violence. However, results from this survey could not be used as we did not have the basic recorded commercial crime figures to multiply using results from this survey.

4.10 There is no survey of crimes against those under-16. It was our intention to derive multipliers of crimes against the under-16s using young people's reporting rates for various categories of crime from studies such as Anderson et al (1994) and the 2000 SCS, but, again, since we do not have separate recorded crime figures for those committed against under-16s, we have no basis on which to multiply.

4.11 The only solution is to use the multipliers that we have for the number of crimes committed against adults (from the comparable exercise that was done between 1999 SCS data and recorded crime) and assume that they are the same for under-16s and commercial crime. This is a major assumption, but unavoidable given the limitations of the data. The derivation of these multipliers is provided in Table 4.4.

4.12 These figures provide us with multipliers for violent crime, serious assault (assume the same as for all assault), robbery, other violent crime (assume the same as for assault and robbery together), housebreaking, and vandalism (assume the same for fire-raising).

4.13 These are the only comparable categories of crime between the SCS and recorded crime statistics. To produce multipliers for the other categories of crime, we had to examine other sources.

4.14 For crimes of indecency, we used the multiplier used by the Home Office in its study of the costs of crime in 2000. Neither the SCS nor the British Crime Survey (BCS) publish details on the level of sexual victimisation due to concerns about the accuracy of the results, but a multiplier was included in the Home Office exercise on measuring the costs of crime, based on BCS estimates, even though it was accepted that the results were tentative.

4.15 With the exception of housebreaking, crimes of dishonesty are not comparable between the SCS and recorded crime statistics, so we had to use other sources. We used the Home Office multipliers, derived from the BCS, for theft from a motor vehicle and "other theft". The multiplier for "other crimes of dishonesty" was assumed to be similar to that derived by the Home Office using data from the National Economic Research Associates (NERA). ¹⁹ For "theft by opening lockfast places", we used the same multiplier as that for housebreaking, as there is no similar category in the BCS. Using data from Farrington (1999), we used the same multiplier as the Home Office of 100 shoplifting offences for each recorded crime.

4.16 For each of the relevant crime types, we applied the multipliers to the recorded crime figures to provide a range of the number of incidents of crime. These are rounded up to the nearest thousand in Table 4.1.

4.17 The high estimates for the number of shoplifting and "other dishonesty" incidents (such as reset, fraud and embezzlement) due to young people is very much dependent on the high multipliers for these offence types. As explained above in 4.15, these multipliers are based on discrete pieces of work and were also used by the Home Office in its attempt to estimate the social and economic costs of crime. However, the results are not suggesting that Scotland has over a million shoplifters. Many of these types of offences will being committed by much smaller numbers of offenders who are committing serial offences.

4.18 For all "other crimes" and both major offence types, we chose not to use a multiplier because of the absence of any reliable evidence as to what are the exact number of incidents in each of these categories. Only a small proportion of people committing offences like speeding and dangerous driving are ever caught and, as they do not involve personal or household crime, they will not be picked-up in sources such as the SCS. They are also unlikely to be picked up in self-report studies as people are unlikely to be able to remember every time they have driven dangerously or gone over the speed limit. We can assume that very few of these crime/offence types are recorded, but a multiplier to take account of the unreported crimes and offences would be a complete guess. In the circumstances, we thought it better to simply report the number of recorded crime and offence incidents in these categories and then attempt to estimate the proportion due to young people. The Home Office also did not attempt to calculate incidents of these crimes and offences in its study on the costs of crime.

4.19 Surveys like the SCS produce fairly accurate estimates of certain types of crime at a national level, but are much less accurate at a local level due to the much smaller sample sizes. To produce local authority total crime figures, we therefore just simply weighted the results according to each local authority's share of police recorded crime.

Estimating the number of offences committed by young people

4.20 To provide estimates of the proportions of different types of crime committed by young people, we looked at a profile of crimes and offences committed by young people that are dealt with by Children's Hearings and by the courts. Both these sources provide details on the type of crime, the age and gender of the offender and the local authority area in which the proceedings are taking place. We compared this profile of young people to that of adults who are dealt with by the courts and derived from this simple ratios of charges against young people relative to charges against adults. Using this information, we arrived at estimates for different types of crimes committed by young people by crime type, age, gender and local authority area.

4.21 We would also have liked to have used data from the Crown Office and Procurator Fiscal Service on cases that were referred to it, but were not proceeded with in court. However, this data was not readily available.

4.22 The base year for this part of the exercise is 2001, as this is the latest year for which full court statistics are available. A problem is that although court statistics by crime type are available by crime type for 2001, the equivalent SCRA data is only available for 2002. To make the courts and the SCRA data comparable, we had to combine the total number of offences proceeded against young people in the courts for 2001 (73,248) and the total number of SCRA offence referrals for 2001 (43,281) to provide a total of 116,529 recorded crime incidents attributed to young people in 2001. As a proportion of cases proceeded against in court and all SCRA offender referrals (276,207), this provides a rough estimate of 42% of all crimes and offences being the responsibility of young people of 21 or under. Given the limitations of the data, in estimating the proportion of crimes that are caused by young people in each crime category, we need to weight the numbers in the SCRA data (for 2002) to reflect total SCRA offender referrals in 2001. Adjustments were also made to account for offences where the crime/offence was unknown. The unknowns were distributed among the crime/offence types according to their share of total crimes and offences.

4.23 The weighted SCRA offence referrals are then added to the offence court proceedings data for those 21 and under and the proportion of incidents that these constitute as a proportion of all SCRA offence referrals and all offence proceedings in court (for all ages) is calculated to derive estimates of the proportion of offences caused by young people by crime type. Data from the High Court was excluded from the analysis as it was not possible to assign offences proceeded against here to a local authority area.

4.24 The SCRA data by local authority area, age and gender is added to the courts data in the same way and estimates are made of the proportion of offences committed by young people in each of these categories based on their relative shares of total offences data. The results in this report are presented by gender and broad age band. The results by local authority area have been made available to the Scottish Executive, but are not reproduced in this report because the figures in some local authority areas broken down by crime type are sufficiently small for individuals to be recognised.

4.25 There are a number of other limitations with the method that are important to bear in mind when using the results obtained. This research does not claim to provide definitive counts of youth crime in Scotland. Some of these limitations are described more fully below.

Few Crimes Result in any Kind of Proceedings

4.26 Figure 4.2 outlines the stages involved between a crime being committed and formal action being taken against an offender. At each stage of the process, offenders are filtered out of the system and a relatively small proportion of offences result in formal action being taken against an offender. The relatively low clear up rate for crimes ²⁰ (46%) is a major reason why the numbers are dramatically reduced, and in the courts there are very few actual convictions relative to the number of crimes. Not all offences by young people are dealt with by any formal proceedings.

4.27 We had to make the assumption that the clear up rates and charge rates for youth and adult crimes are the same. It was necessary to assume that young people are equally likely as adults to avoid detection, charging and conviction as we could not find a source of data that provided information on the relative likelihood of being detected, charged and convicted for young people and for adults.

4.28 Our method also makes the assumption that offences that are dealt with by the courts or Hearings systems have the same profiles (in terms of age of offender, local authority area in which proceedings take place, etc.) as offences where no proceedings take place.

Crimes can be committed outwith the area in which the offender resides

4.29 The area in which proceedings take place will not always necessarily be the area in which the crime has been committed: an offender may have committed an offence outwith their home area, but have been dealt with within their home area. However, there is no reliable information on where each offender committed each of their offences and we will have to accept this limitation of the available data and use local authority area in which proceedings take place in estimating youth crime by area.

There will be some double-counting

4.30 Some cases are referred jointly to the Reporter and to the Procurator Fiscal, for example serious offences involving a youth under-16, and a youth over 16 who is placed under supervision. The Procurator Fiscal can also make the decision to refer a young person to the Reporter. However, there is no reliable information on the number of these cases by type of crime, gender, age, etc., therefore there will be some double-counting in our method for estimating the levels of youth crime, but we do not believe that this problem is of such a significant extent to question the validity of this exercise.

Figure 4.2. The criminal justice process in Scotland.

4.31 An alternative method would have been to use self-reporting data of studies of the whole population to estimate the proportion of crime committed by young people relative to the rest of the population. We could then compare the proportion of offences committed by young people relative to the whole population for these crime types and apply this proportion to the total crime figures.

4.32 A major difficulty with this method, however, is that there is no one survey that could provide this sort of information. The National Centre for Social Research is currently undertaking a similar exercise for the Home Office, but the results are not as yet available.

4.33 However, as a check on our calculations, we can compare our results to other work that has been undertaken. For example, we know from previous research that:

young people under-18 account for between 25% and 33% of all offences in England and Wales and the US ²¹;

a range of international studies has shown that most juvenile crime is theft-related and only a small proportion of juvenile crime is violent ²²;

young men commit a far higher proportion of crime than young women and in the UK in 1995 young men under-21 accounted for 80% of recorded offences cleared up and attributed to this age group ²³; and

the peak age for offending is usually in the late teens, with official statistics in the UK suggesting that it is 18 for young men and 15 for women. ²⁴

4.34 But it has to be accepted that there are few genuine sources of comparison given that few studies attempt to measure incidence rates for the age group and all the types of crime in which we are interested.

Estimates of youth crime

4.35 The headline results in Table 4.1 show that young people, 21 and under, are responsible for over 40% of crimes and offences in Scotland. Results from Table 4.3. The under-15 account for 16-17% of all crime, the under-18 for about 22-23% and the 18-21 for 21%. The estimates therefore for those under-18 is slightly lower than the range previous research indicates as being between one-quarter and one-third of all crime. As expected, the peak age for offending seems to be the late teens and early twenties. Also as expected, most youth crime is theft-related, particularly shoplifting.

4.36 Young men account for a very high proportion of youth crime, at around 87% of all crime attributable to the under-21 age group, this is slightly higher than previous research suggests. Females only account for around 13% of all youth crime, but relatively large proportions of shoplifting offences and other types of violent crime (which includes cruel and unnatural treatment of children).

4.37 As anticipated, young people account for high proportions of types of theft, vandalism, fire-raising and offences involving handling offensive weapons and a lower proportion of crimes involving indecency. Perhaps the one very surprising result from Table 4.1. is the relatively low proportion of drug crime being the responsibility of young people (one-third), but, as argued earlier in the report, very few drug offences are detected and result in any form of proceedings. The proportion of shoplifting offences due to young people (42%) is also lower than anticipated, but it may be a reflection of young people being less likely to be formally charged with a shoplifting offence.

4.38 Among the detailed results in Table 4.3., it was expected that the under-15s would account for relatively high proportions of fire-raising and vandalism and relatively low proportions of violent crime and other crimes. What was not expected was the high proportion of crimes of indecency attributable to this age group (60% of all youth crime). Crimes of indecency includes rape, attempted rape, lewd and indecent behaviour and offences such as prostitution. It is not clear why those under-15 appear more likely to commit such offences compared to the rest of those under-21.

4.39 Those 16-17 commit relatively high proportions of motor vehicle theft and robbery. However, nearly half youth crime is attributable to those aged 18-21, particularly violent crime, other crimes of dishonesty such as fraud and reset, crimes against public justice such as bail offences and motor vehicle offences. This age group commits relatively few crimes of vandalism.

4.40 The research does appear to show the scale of the numbers of crimes and offences involving young people in Scotland, with young people being responsible for around 30,000 incidents of violent crime, over 48,000 break-ins, around 20,000-30,000 motor vehicle thefts, over 1 million counts of shoplifting and around 300,000 acts of vandalism.