Recorded Crime in Scotland, 2014-15

4. Putting recorded crime in context - A comparison with the Scottish Crime and Justice Survey (SCJS)

Chart 19: Overall number of crimes in Scotland - Police Recorded Crime and the SCJS, 1992 to 2014-15

1. After 1994, the recording period for police recorded crime changed from calendar year to financial year.

2. The shift to the current survey design in 2008-09 has led to greater certainty around estimates.

3. Latest SCJS figures are available for 2012-13 while the latest recorded crime statistics cover 2014‑15. The results of the 2014-15 SCJS are due to be published in the first quarter of 2016.

The preceding sections of this bulletin contain information on the volume and types of crime recorded and cleared up by the police in Scotland in 2014-15. As highlighted earlier in this report, in addition to the information on police recorded crime, crime in Scotland is also measured by the SCJS, a national survey with adults (aged 16 and over) living in private households which asks respondents about their experiences of crime.

This chapter brings together these two complementary sources, police recorded crime and the SCJS, to present a fuller, more comprehensive picture of crime in Scotland^[9]. However bringing the two sources together in this way highlights that the SCJS and police recorded crime cover different populations and different timescales, and the SCJS does not cover the entire range of crimes and offences that the police are faced with. These and other differences mean that making direct comparisons between the two sources is not straightforward. Nevertheless, the SCJS and police recorded crime statistics do present complementary information on crime occurring in Scotland, so it is therefore helpful and informative to look at these sources together. This chapter will look at police recorded crime and SCJS findings in three ways:

• Firstly, it will look at national trends of overall crime captured by police recorded crime and by the SCJS.
• Secondly, it will look at crime in the two broad categories of crime captured by the SCJS (namely, property crime and violent crime). This section will also highlight how the SCJS captures more contextual information on the risk and characteristics of crime.
• Lastly, it will look at the comparable crime groups, a grouping of crimes specifically constructed to allow comparison of the estimates of all comparable crime and comparable sub-groups from the SCJS and police recorded crime statistics for a set of crimes that are covered by both sources.

At the time of this bulletin's publication, the latest published SCJS data still relates to 2012-13. As such this chapter remains broadly similar to what was included in the previous Recorded Crime in Scotland, 2013-14 bulletin. An update of the analysis, comparing the statistics for Police Recorded Crime and the SCJS, will be possible following the publication of the results for the 2014-15 SCJS, currently scheduled for the first quarter of 2016.

4.1. Overall number of crimes - Police Recorded Crime and the SCJS, 1992 to 2014-15

The 2012-13 SCJS estimates that there were 815,000 incidents of crime against adults in Scotland in 2012-13. This is 22% lower than in the 2008-09 survey when there were an estimated 1,045,000 crimes, meaning that in 2012-13, around one in six (16.9%) adults were the victim of at least one crime.

While Chart 19 presents overall estimated crime levels for surveys conducted since the 1990's, prior to the current SCJS methodology (employed in surveys since 2008‑09^[10]), crime survey estimates in Scotland were derived from smaller sample surveys and, in general, were subject to a higher level of uncertainty^[11]. No clear trend could be detected for changes to the overall number of crimes measured by the various Scottish crime surveys conducted prior to 2008-09. This is reflected in Chart 19 in the higher levels of fluctuation in the overall crime series prior to 2008-09. We have assessed that the greater levels of certainty around survey estimates since 2008-09 enable more consistency in comparisons between the SCJS and recorded crime series, and as such the analysis that follows in this chapter focuses on the period from 2008-09 onwards.

In comparison to the SCJS, as noted previously in this report, in 2014-15, the police recorded 256,350 crimes; this represents a decrease of 5% since 2013-14, and a decrease of 32% since 2008-09. Chart 19 shows that since the shift to the current survey design and increased sample sizes in 2008-09, survey estimates of the overall level of crime have fallen in line with similar reductions in overall recorded crime over the same period.

Chart 19 effectively highlights the scale of the difference between the number of crimes estimated by the SCJS, and the level recorded by the police. There are a range of reasons for this difference, however the main factor is that the SCJS captures crimes that do not come to the attention of the police, and therefore are not included in recorded crime figures. The 2012-13 SCJS estimated that, of the 815,000 incidents of crime, 39% came to the attention of the police. Therefore, while the SCJS is good for estimating the likely range of crime in the underlying population (and the level of uncertainty around such estimates), the police recorded crime data effectively highlights the level of crime with which the police are faced.

In summary then, the SCJS can help to identify the relative magnitude of crime not reported to the police and why crimes are not reported. In 2012-13, where crime was not reported to the police, the most common reasons SCJS respondents gave for not reporting crime were that the victim felt that the police could have done nothing (37%) or that the victim perceived the incident to be too trivial to involve the police (36%). Conversely, where crime was reported to the police, it was mostly because the victim felt that it was the right thing to do (49%) or in the hope that offenders would be caught and punished (27%).

4.2. Police Recorded Crime and the SCJS Crime Groups

As noted elsewhere in this report, recorded crime figures are grouped into five crime groups (Non-sexual crimes of violence, Sexual crimes, Crimes of dishonesty, Fire‑raising, vandalism etc. and Other crimes) and two offence groups (Miscellaneous offences and Motor vehicle offences). However, the SCJS presents information in two broad crime categories: Property Crime and Violent Crime (outlined in Box 1 below).

There are a number of reasons that the SCJS crime categories do not match the recorded crime groups; principally this is because the SCJS is a victimisation survey and does not collect data on all of the crimes and offences that the police are faced with (e.g. homicide, crimes against business i.e. shoplifting, and motor vehicle offences).

Box 1: SCJS Crime types

Property crime	Violent crime
Vandalism (including motor vehicle and property vandalism) All motor vehicle theft related incidents (including theft and attempted theft of and from a motor vehicle) Housebreaking (termed burglary in England and Wales) Other household thefts (including bicycle theft) Personal theft (excluding robbery).	Assault (includes serious assault, attempted assault, minor assault with no-negligible and minor injury) Robbery

However, the SCJS is able to collect more detailed information on crimes that are not reported to the police, as well as information on the characteristics of crime and the attitudes and perceptions of victims. Such information is vital in presenting a fuller picture of the nature of crime in Scotland, than can be gained from recorded crime figures alone.

This section provides an overview of the main findings from SCJS 2012-13 in the property crime and violent crime categories, however direct comparisons to police recorded crime data are limited, due to the differences in crime groups within the two sources, as outlined above. Comparisons with recorded crime results are made in Section 4.3 which outlines the comparable crime subset.

Property Crime

Property crime as measured by the SCJS involves theft or damage to personal or household property (including vehicles). In 2012-13, approximately 579,000 crimes (71% of all SCJS crime) were in this category, which means that it is estimated that around 15% of adults in Scotland were a victim of property crime. Between 2008-09 and 2012-13, there was a statistically significant decrease of 21% in property crime captured by the SCJS.

Of the 579,000 property crimes estimated by the SCJS in 2012-13, vandalism accounted for 27%, followed by other household theft (including bicycle theft) (21%), personal theft (excluding robbery)(13%), all motor vehicle theft related incidents (6%) and housebreaking (4%).

The types of property crime captured in the SCJS are mostly covered in two of the police recorded crime groups, Crimes of dishonesty and Fire-raising, vandalism etc. However, while the police recorded 126,857 crimes in the Crimes of dishonesty group in 2014-15 (an 8% decrease since 2013-14) and 52,091 crimes in the fire-raising, vandalism etc. group (a 4% decrease since 2013-14), it should be noted that these groupings are not directly comparable with the SCJS 'property' crime group as, for example, they cover some crimes (such as shoplifting (27,364 recorded crimes in 2014-15) and fraud (6,913 recorded crimes in 2014-15)) which are not captured in the SCJS.

In terms of crimes reported to the police, the 2012-13 SCJS estimates that over a third (35%) of property crimes were reported to the police, however, reporting rates were highest among victims of housebreaking (64%). The most common reasons given for not reporting property crime was that the incident was considered to be too trivial (43%) or that the victim felt that the police could not have done anything about it (41%). When property crime was reported, the most common reasons given were that reporting was considered to be the right thing to do/automatic (52%) or in the hope that offenders would be caught and punished (28%). Victims of property crime also experienced emotional responses, with annoyance (56%), anger (52%) and shock (16%) being the most commonly experienced.

Table 1 shows that property crime was experienced by near equal proportions of men (15%) and women (14%), however risk declined with age.

Table 1: The varying risk of property crime (SCJS 2012-13)

	All	Male	Female	16-24	25-44	45-59	60+
Property Crime (risk as a percentage)	14.8	15.3	14.4	18.1	18.7	15.2	8.5
Base	12,050	5,210	6,840	970	3,590	3,110	4,380

Violent Crime

The SCJS violent crime category includes attempted assault, serious assault, minor assault and robbery. Of the 815,000 crimes measured by the SCJS in 2012-13, 236,000 (29%) were violent crimes, which means that it is estimated that around 3% of adults in Scotland were a victim of violent crime in 2012-13. Between 2008-09 and 2012-13 there was a statistically significant decrease of 25% in violent crime captured by the SCJS.

Of the 236,000 violent crimes estimated by the SCJS in 2012-13, minor assaults (no-negligible and minor injury) made up the majority (23% of all crimes measured by the SCJS), followed by attempted assault (3%), serious assault (2%) and robbery (1%).

The latest data presented elsewhere in this report shows that in 2014-15, the police recorded 6,357 non-sexual crimes of violence. However, as noted already, this grouping is not directly comparable with the SCJS 'violent' crime group. Non-sexual crimes of violence (as used in police recorded crime) includes homicide, whilst, in addition, common assaults (which make up the majority of SCJS violent crime) are included in the Miscellaneous offences police recorded crime group. In 2014-15, the police recorded 58,178 common assaults.

We know from the SCJS that not all crime is reported to the police. The 2012-13 SCJS estimates that almost half (48%) of violent crimes were reported to the police. The most common reasons victims provided for why violent crime was not reported was because they felt that the police could not have done anything about it (24%), they dealt with the matter themselves (23%) or that they considered the incident to be too trivial to involve the police (17%). Conversely, where violent crime was reported, the most common reasons given for reporting were that it was considered to be the right thing to do/automatic (42%), and-or the victim hoped to avoid repetition of the crime to themselves (29%) or someone else (29%).

Table 2 shows that that risk of violent crime decreases with age from 8% for those aged 16-24, to less than 1% for those aged 60 or over, and that males (4%) are at a higher risk of violent crime than females (2%).

Table 2: The varying risk of violent crime (SCJS 2012-13)

	All	Male	Female	16-24	25-44	45-59	60+
Violent Crime (risk as a percentage)	3.1	4.2	2.2	8.2	4.3	1.9	0.4
Base	12,050	5,210	6,840	970	3,590	3,110	4,380

In addition, the 2012-13 SCJS estimates that almost one in three violent crimes (29%) happened between 9 pm and 3 am on the weekend and that victims thought that the offender was under the influence of alcohol in 59% of violent crime, and drugs in 29% of violent crime. Injuries were sustained by victims in almost three-fifths (58%) of violent crime. Where injuries were sustained, 60% received minor bruising or a black eye and 33% received scratches. Victims also experienced emotional responses to violent crime with anger (50%), annoyance (47%) and shock (38%) being the most commonly experienced.

4.3. Comparing SCJS estimates and Recorded Crime

4.3.1. Comparable Crime - Overall Comparison

As highlighted above, the two data sources cover different populations, time periods and crimes, which means that making direct comparisons is not straightforward. Comparisons can be made by examining a broadly comparable subset of crimes which are covered by each source and can be consistently coded in the SCJS in the same way as the police would do. Comparisons are made in the following three broad crime groups:

• Vandalism (other household crime including motor vehicle vandalism and property vandalism).
• Acquisitive crime (including bicycle theft, housebreaking and theft of motor vehicles).
• Violent crime (including serious assault, common assault and robbery).

On this basis, of the 815,000 crimes estimated by the 2012-13 SCJS, just under two-thirds (527,000) can be compared with police recorded crimes 2012-13.

The remainder of this section provides an overview of the level of crime and trends in the comparable subset from 2008-09 to 2012-13, the latest year for which data are available from both sources.

This analysis has been extended further in the Scottish Government report, Bringing Together Scotland's Crime Statistics^[12], available from http://www.gov.scot/Topics/Statistics/Browse/Crime-Justice/crime-and-justice-survey/publications/SCJSPRCanalyticalpaper.

Table 3: Comparable crime group estimates (2008-09 to 2012-13)

	2008-09	2009-10	2010-11	2012-13	% change 2008-09 to 2012-131	% change 2010-11 to 2012-131
Comparable Recorded Crime	215,901	195,728	183,117	144,662	-33%	-21%
Comparable SCJS Crime	731,000	630,000	556,000	527,000	-28%	-5%
Recorded Acquisitive Crime	27,527	26,146	26,478	21,834	-21%	-18%
SCJS Acquisitive Crime	64,000	61,000	61,000	73,000	14%	19%
Recorded Violent Crime	82,855	79,769	78,263	66,076	-20%	-16%
SCJS Violent Crime	317,000	266,000	220,000	236,000	-25%	7%
Recorded Vandalism	105,519	89,813	78,376	56,752	-46%	-28%
SCJS Vandalism	350,000	303,000	275,000	219,000	-37%	-20%

1. SCJS statistically significant changes (at 95% confidence interval) shown in bold.

As presented in Table 3, in 2012-13, the SCJS estimated that there were 527,000 crimes in the overall comparable crime category, while the police recorded around 145,000 crimes in this category. The extent of overall comparable crime, in both the recorded crime figures and SCJS estimates, decreased between 2008-09 and 2012-13 (-33% and -28% respectively). The reduction in estimated SCJS comparable crime from 731,000 in 2008-09 to 527,000 in 2012-13 is a statistically significant change.

SCJS respondents are asked whether the police 'came to know about' the crime, either by them or somebody else. This allows comparisons to be made between crime estimated to have been reported to the police in the SCJS, and police recorded crime data. Figures from the 2012-13 SCJS indicate that of the 527,000 crimes in the overall comparable subset, around 232,000 incidents (44%), were estimated to have been reported to police. In 2012-13, the police recorded 144,662 crimes in the comparable category, so overall, it can also be estimated that around 27% of the total comparable crimes estimated by the SCJS (that is reported and non-reported crime) were recorded by the police in 2012-13.

However, it should be noted that this 'comparable' series is broadly, rather than directly, comparable. As a survey, the SCJS can only provide estimates of crimes reported to the police, not precise figures. It is not possible, for example, to match SCJS microdata (i.e. the individual records of survey respondents) to police recorded crime records. Therefore, it is not possible to determine whether a crime that a respondent said they reported to the police actually appeared on a police incident log in the relevant time period (at all, or before/after the time period) and, if so, to identify how it was recorded.

Consequently, we would not expect estimates of the crime reported to the police and the level recorded by the police to be equal. In light of this, the methods of analysis presented in this section are more suitable for assessing this relationship and variation of each series over time, rather than, for example, assessing with confidence the absolute level of crime estimated to have been reported but not recorded within each survey.

Finally, it should also be noted that there are a range of other factors which may affect the comparability of these series, for example it is possible that a number of crimes reported to the police are not captured and recorded by the police. However auditing of incidents and crimes recorded by Police Scotland by HMICS (http://www.hmics.org/publications/hmics-crime-audit-2014) indicates that police compliance in recording is generally good overall and does not indicate that this accounts for the difference in our two series of crime data or changes over time. In addition, the SCJS also contains factors which are likely to affect the degree of comparability to recorded crime; for example non-quantifiable error around survey estimates (for example, error in the recall of respondents about the date of the incident which may have been outside the survey reference period); or a change in underlying survey sample design (from clustered to unclustered in 2012-13); or the switch to SCJS biennial design in 2012-13, although these factors are not thought to have introduced any bias to the SCJS results.

4.3.2. Comparable Crime - by Comparable Crime Sub-groups

This section summarises findings for the comparable crime sub-groups: acquisitive crime, violent crime and vandalism. When considering these comparable crime sub-groups over time (as shown in Table 3), police recorded crime data should be used to assess the level of crime with which the police are faced and SCJS results used as a barometer to estimate the underlying level of crime.

Acquisitive Crime

In 2012-13, the SCJS estimated that there were 73,000 acquisitive crimes (+/-10,000, meaning that the true number of acquisitive crimes in the underlying population is estimated to lie somewhere between 63,000 and 82,000^[13] ). Of these estimated 73,000 crimes, around 40,000 (55%) were said to have been reported to the police, while, in comparison, the police recorded 21,834 acquisitive crimes. Therefore, it can be estimated that 55% of the crime estimated to be reported to the police by the SCJS were captured in police statistics. This demonstrates and reinforces the finding that these two sources of crime data should be considered as broadly rather than directly comparable; for example, they do not cover the same populations or time periods, and the SCJS gathers information from a sample rather than the whole population so the results are always estimates not precise figures, and so are subject to a range of quantifiable and non-quantifiable error.

When considering this comparable crime sub-group over time, results show that recorded acquisitive crime decreased by 21% between 2008-09 and 2012-13, however, the SCJS estimates of acquisitive crime showed no statistically significant change between 2008-09 and 2012-13 (the apparent change from 64,000 to 73,000 is not statistically significant). Survey estimates on the comparable crime subgroups are based on small base sizes of respondents (272 for acquisitive crime in 2012-13) and as a consequence, there is less certainty around estimates.

The variable nature of making these type of comparisons is reflected, for example, in analysis of shorter term change in these series, which finds greater consistency in comparable acquisitive crime in the SCJS and recorded crime between 2008-09 and 2010-11 (respectively, a change of -5% (again non-statistically significant) and a decrease of 4%). Therefore, the uncertainty around the changes in this crime sub-group, both in the short and medium term, limit the conclusions which can be drawn when comparing the SCJS and recorded crime in the acquisitive crime sub-group at this stage.

Violent crime

In 2012-13, the SCJS estimated that there were 236,000 violent crimes (+/-40,000, meaning that the true number of violent crimes in the underlying population is estimated to lie somewhere within the range 196,000 to 275,000^[14]) and the police recorded 66,076 violent crimes. The SCJS estimates that around 113,000 (48%) violent crimes were reported to the police in 2012-13. Therefore, it can be estimated that 58% of the crime estimated to be reported to the police by the SCJS were recorded in police statistics, demonstrating, as mentioned above, that these two sources of crime data should be considered as broadly rather than directly comparable.

Between 2008-09 and 2012-13, recorded violent crime figures in the comparable category decreased by 20%, while, for the same period, the SCJS estimates of violent crime decreased by 25%, a statistically significant change in the SCJS results. As indicated by Table 3, the SCJS is often better able to detect changes in survey estimates with greater certainty over longer timescales, due primarily to the small numbers of victims identified in the survey (343 for violent crime in 2012-13).

Vandalism

In 2012-13, the SCJS estimated that there were 219,000 vandalism crimes (+/-20,000, meaning that the true number of vandalism crimes in the underlying population is estimated to lie somewhere between 199,000 and 239,000^[15]) of which around 79,000 (36%) were reported to the police, while the police recorded 56,752 vandalism incidents. From this, it can be estimated that 72% of the vandalism crime estimated to be reported to the police by the SCJS were recorded in police statistics, again demonstrating that these two sources of crime data should be considered as broadly rather than directly comparable.

Between 2008-09 and 2012-13, recorded vandalism figures fell by 46%, while SCJS estimates have shown a statistically significant decrease of 37%. Compared to the violent crime and acquisitive crime groups, vandalism estimates are based on larger samples of respondents (760 in the 2012-13 survey), leading to greater certainty around changes in the short and medium term.

4.4. Conclusion

This chapter has brought together the two main sources of crime statistics in Scotland: the Scottish Crime and Justice Survey and Police Recorded Crime statistics. Although each source covers different populations, crimes and offences and time periods, considering them together presents a more comprehensive picture of crime in Scotland. For example, the latest figures from the SCJS show that there were an estimated 815,000 incidents of crime against adults living in households in Scotland in 2012-13, while the police in Scotland recorded 273,053 crimes in 2012-13 (and 203,063 Miscellaneous offences, including 60,955 common assaults). This difference between survey estimates and recorded crime figures shows that, for many reasons, not all crime comes to the attention of the police. However, the SCJS helps to assess this and in addition is able to capture more information on the characteristics of crime and victims of crime, helping to provide a fuller picture of the nature of crime in Scotland.

Due to the differences between them, making direct comparisons between the two sources is not straightforward. However, a comparable subset of crime can be used to make some broad comparisons to assess the relationship between recorded crime figures and SCJS estimates.

Table 4: Strengths and limitations of Police Recorded Crime and the SCJS

	Recorded Crime	Scottish Crime and Justice Survey
Where do the data come from?	Administrative police records	Face to face interviews with residents from a nationally representative sample of the household population
Basis for inclusion	Crimes recorded by the police in Scotland, governed by the Scottish Crime Recording Standard and Counting Rules.	Trained coders determine whether experiences of victimisation in the last 12 months constitute a crime and assign an offence code.
Frequency	Collected by financial year. Statistics released in an annual publication.	Continuous survey with results currently published biennially.
Strengths	Covers the full range of crimes and offences. Provides data at a local level (and can be used for performance monitoring). A good measure of rarer, more serious crimes that are well reported. A good measure of rarer, more serious crimes that are well reported. Good measure of long-term trends.	Good measure of trends since 2008-09. Captures information about crimes that are not reported to the police (including sensitive issues such as domestic abuse or drug use). Provides information on multiple and repeat victimisation (up to 5 incidents in a series). Analyses risk for different demographic groups and victim-offender relationships. Provides attitudinal data (e.g. fear of crime or attitudes towards the criminal justice system).
Limitations	Partially reliant on the public reporting crime. Reporting rates may vary by the type of crime (e.g. crimes more likely to be reported include serious crime and crimes such as housebreaking where recording is required for insurance purposes). Trends can be affected by legislation; public reporting practices; police recording practices.	Does not cover all crimes (e.g. homicide or 'victimless' crimes such as speeding). Does not cover the entire population (e.g. children, homeless people or people living in communal accommodation). Unable to produce robust data at lower level geographies. Difficult to measure trends between survey sweeps in rarer forms of crime (such as more serious offences). Subject to quantifiable/non-quantifiable error.
What other data are collected?	Additional data on homicides, racist incidents, domestic abuse incidents and firearm offences.	Public perceptions about crime. Worry about crime and the perceived likelihood of being a victim. Confidence in the police and the criminal justice system. Prevalence estimates on 'sensitive' topics (partner abuse, sexual victimisation, stalking and drug use).