Crime and Criminal Justice Research Findings No. 36
Homicide in Britain: A Comparative Study of Rates in Scotland and England & Wales
Keith Soothill, Brian Francis, Elizabeth Ackerley and Samantha Collett
For some time it has been recognised that Scotland appears to have a homicide rate around double that of England & Wales 1. This research investigates whether the overall difference is an artefact of the different recording policies and practices of the two countries, where any differences seem to be more precisely located and the beginnings of an explanation.
- Using the homicide databases for Scotland and England & Wales over a 10 year period (1985-94), the average yearly homicide victimisation rates are much larger in Scotland than in England & Wales (19.15 per million population compared with 11.28). It is estimated that differences in statistical recording practices only contribute about 0.6 to this difference. Hence, a very real difference in the homicide rates between Scotland and England & Wales remains.
- The male homicide rate in Scotland (28.84) is over twice the rate for England & Wales (13.40), while the female homicide rate in the two jurisdictions is very similar (10.07 and 9.25 respectively).
- Homicide rates are statistically significantly higher in Scotland for all age groups aged 15 years or over. For each age group, the rates in Scotland are about twice as high as those in England & Wales. In contrast, for those victims aged 14 years or below, the rates are similar.
- Homicide rates in Scotland are significantly higher than those in England & Wales in circumstances of a 'rage or quarrel', 'theft / other gain', or a 'feud', although in the latter category in particular, the numbers are very small.
- Homicides using the methods of 'sharp instrument', 'blunt instrument' and 'hitting or kicking' occur more often in Scotland than in England & Wales. The rates of homicides using other methods are similar in the two jurisdictions.
- It is the homicide rate in Scotland of 12.91 per million population for 'male victims by acquaintance' (compared to a rate of 4.72 for this type of homicide in England and Wales) that needs to be the pivot of concern. The homicide rate of males by acquaintances is higher in Scotland than in England & Wales across the age categories and is not just a special problem among young people.
- Males are about one and a half times more at risk from homicidal attack by strangers in Scotland than is the case in England & Wales. The rates for females being killed by strangers are similar in the two jurisdictions.
- For males there is less danger of being a victim of homicide in England & Wales than in Scotland, but for other kinds of violence, males in England & Wales are at greater risk.
The origin of this research study was an expressed concern about the apparently much higher homicide rate in Scotland compared with England & Wales. The work focused on the practices and processes of recording homicides in the two jurisdictions, and a statistical analysis of the material on the two homicide databases was undertaken. Finally a discussion of the homicide rates in relation to violence victimisation rates was developed.
The data sources consisted of the Homicide Index for Scotland and the Homicide Index for England & Wales. For the 10-year period 1985-94, 1037 homicide victims were initially recorded in Scotland, and 6712 in England & Wales.
The process of recording homicides
The aim of the investigation of the process of recording statistical information on homicide in both The Scottish Office and the Home Office was to establish whether the difference in homicide rates arises from differences in recording practices or in fact whether there is evidence of a real difference.
It was established that, despite differences in criminal law, the type of offence included on (or excluded from) the Homicide Index in each jurisdiction was satisfactorily similar, and would be unlikely to lead in itself to any significant bias.
Over the 10 year period 1985-94, 58 (5.6%) of the 1037 initially recorded homicide victims in Scotland were removed from the statistics, becoming 'no longer recorded' as victims of homicide; the equivalent figure in England & Wales was 983 (14.6%) of the 6712 initially recorded victims. The difference in these proportions initially recorded but later reclassified is highly statistically significant (p < 0.0001).
By far the largest difference occurred in respect of cases where all suspects had been acquitted; these were the circumstances for over half (596) of those victims reclassified in England & Wales, but for only 4 victims in Scotland, a highly significant statistical difference (p < 0.0001). Thus, far more of the cases that result in an acquittal in England and Wales are removed from the homicide figures than is the case in Scotland. If the proportion of acquittals reclassified as 'no longer homicide' for England & Wales was made equivalent to that of Scotland, the overall reclassification rate in England & Wales would decrease from 14.6% to 6.1%, much closer to Scotland's reclassification rate of 5.6%.
If the Scottish procedures for reclassifying cases had been applied in England & Wales, an estimated 297 deaths over the 10 year period 1985-94 might have remained as 'currently recorded' homicides instead of being reclassified as 'no longer homicide'. However, the effect of reinstating these estimated 297 deaths would be minimal, increasing the homicide victimisation rate for England & Wales by just 0.58 per million population from 11.28 to 11.86, still substantially less than the rate of 19.15 per million in Scotland. It is therefore evident that, although the difference in homicide rates does appear to reduce once the differences in statistical recording practices have been accounted for, a very real difference in the homicide rates between Scotland and England & Wales remains.
Analysing the currently recorded homicides
Having established that the difference in the homicide victimisation rates of the two jurisdictions is not simply a result of the different recording practices, analyses were undertaken to determine where the differences were more precisely located. For the 10 year period 1985-94, the focus was on the 979 victims in Scotland and 5729 victims in England & Wales 'currently recorded' as victims of homicide, making no adjustments in the light of the results from the investigation of the differences in recording procedure.
When controlling for a number of variables such as sex, method and motive of homicide, it was found that Scotland does not have consistently higher rates across all categories of homicide than England & Wales. The sex of the victim is crucial. The male victimisation rate in Scotland is over twice the rate for England & Wales (28.84 per million population compared with 13.40); the female victimisation rate in Scotland is very similar to the rate for England & Wales (10.07 per million compared with 9.25).
The homicide rates are statistically significantly higher in Scotland for all age groups aged 15 years or over. For each of the age groups from 15 years to 65 years and over, the rates in Scotland are about twice as high as those in England & Wales. In contrast, for those victims aged 14 years or below, the rates are virtually identical. Further analysis suggests, however, that the differences are accounted for almost entirely by the difference in male homicide rates, with the rates for females in each age group being largely similar between the two jurisdictions.
The two jurisdictions display different homicide rates for particular circumstances. Compared with England & Wales, Scotland has twice the rate in circumstances of a 'rage or quarrel' and 11/2 times the rate for 'theft / other gain', while in circumstances of a 'feud', although the numbers concerned are small, Scotland's rate is nearly eight times that of England & Wales.
Homicides using the methods of 'sharp instrument', 'blunt instrument' and 'hitting or kicking' occur more often in Scotland than in England & Wales. The rates of homicides using other methods (strangulation/asphyxiation/drowning, fire, shooting and 'other', including drugs) are similar in the two jurisdictions. Compared with England & Wales, 'hitting and kicking' is used disproportionately in Scotland when males are killed by other relatives, by acquaintances and by strangers, and when females are killed by their partners. The only context in which people in England & Wales are shown to employ a particular method more than those in Scotland is in using a 'sharp instrument' in the killing of females by acquaintances; in Scotland, a 'blunt instrument' is used more often in this type of homicide than in England & Wales.
While rates for the various types of homicides involving a male victim - whether killed by partners, relatives, acquaintances or strangers - are statistically significantly higher in Scotland than in England & Wales, it is the homicide rate in Scotland of 12.91 per million population for 'male victims by acquaintance' - nearly twice the rate of any other category - which needs to be the pivot of concern. The homicide rate of males by acquaintances is higher in Scotland than in England & Wales across the age categories and is not just a special problem among young people. Above the age of 15 years, the rate in Scotland of this type of homicide is around three times the rate in England & Wales.
Homicides by strangers also reveal considerable gender differences. Comparing male and female victimisation rates, males are over six times as likely in Scotland and over three times as likely in England & Wales to be killed by a stranger than is the case for females. Comparing jurisdictions, males are about one and half times more at risk from homicidal attack from strangers in Scotland than is the case in England & Wales. The rates for females being killed by strangers are similar in the two jurisdictions.
Rates for different types of homicides involving female victims are similar in the two jurisdictions. The exception is the homicide of females by relatives where the homicide rate for Scotland is significantly higher. While some of Scotland's rates may be higher than those for England & Wales, they do not alter the overall picture of a female being at considerably more risk than a male in a partnership in both jurisdictions, while males are at greater risk in other situations.
In the decade under study (1985-94) Scotland always had a significantly higher homicide rate than England & Wales, but it became a much greater difference in the 1990s with an especially high year in 1992 (See Table 1). There are also substantial differences according to the gender of the victim; the victimisation rates year by year for females are almost identical in Scotland and England & Wales except for 'females killed by relatives', where the overall ten year difference is largely generated by the higher rates in Scotland in the late 1980s and early 1990s. For males the situation is different; for each type of homicide, Scotland has a higher rate than England & Wales year by year (except for 1985 and 1986 for the homicide of males by partners where the rates for the two jurisdictions are roughly equal).
Table 1: Yearly Homicide Victimisation Rates (per million population)
England & Wales
Including all types of homicide in an analysis by standard regions, Greater London has the highest rates, Scotland has the next highest, the northern regions and Wales the next, while the rest of England which includes the midlands and the south and south-west of England has the lowest set of rates. This pattern of homicide rates gradually getting lower as one moves southwards is fairly consistent for various kinds of homicide. However, Greater London is always among the higher rates.
Of the eight Scottish police authorities, the rates for Strathclyde are by far the highest with the rates in the other police authorities significantly lower. The homicide rate for Strathclyde is higher than for Greater London, although geographic and population differences between the two areas should be considered in making this comparison.
Homicide in the wider sphere of violence
Contrasting sharply with the pattern found when considering recorded homicide figures, the violence victimisation rates in England & Wales (as measured by recent national crime surveys) are statistically significantly higher than those in Scotland for domestic, acquaintance and stranger violence, for both male and female victims. The disparity in the rates is greater for males for each type of violence than for females, and thus people living in England & Wales seem to be at much greater risk of experiencing violence than those in Scotland, and males particularly so. Crime victimisation surveys show the increasing importance of acquaintance violence in terms of incidents against both men and women.
The links between homicide and violence are complex. For males there is more danger of being a victim of homicide in Scotland than in England & Wales and this is true whether it is in the domestic context or involving acquaintances or strangers. For other kinds of violence, however, males in England & Wales are at greater risk. For females it is only in the domestic sphere in Scotland that more of the violent incidents involving female victims 'convert' into homicides.
The need for further research
It is necessary to pinpoint with greater accuracy the crucial ingredients that produce the higher 'conversion' of violence into homicide in Scotland for males in such a wide variety of contexts and for females in the domestic sphere.
The criminological literature suggests some key enabling behaviours that have been identified as potential contributors to violence and, thus, homicide. Such behaviours include: physical aggression which often starts early in life and can lead to homicidal violence later in life; illegal weapon acquisition, transport and display - the presence of a weapon in the vicinity of a person experiencing anger, rage, humiliation or fear increases the likelihood that a weapon will be used against the source of that emotion; drug and alcohol abuse; and criminal gang membership where within-group norms can support violence and criminal activity. Rather more detailed research needs to be carried out in order to go beyond the stage of speculation in seeking a full explanation of the causes of homicide.
1: The term 'homicide' encompasses murder and culpable homicide in Scotland and murder, manslaughter and infanticide in England and Wales.
If you wish copies of this Research Findings or have any enquiries about the work of CRU, please contact us at:
Scottish Executive Central Research Unit,
Room J1-0, Saughton House,
EDINBURGH, EH11 3XA.
If you wish a copy of 'Homicide in Britain: A Comparative Study of Rates in Scotland and England & Wales', the research report which is summarised in this Research Findings, please send a cheque for £15.00 made payable to The Stationery Office to:
The Stationery Office Bookshop
71 Lothian Road,
Tel 0131 228 4181 Fax 0131 622 7017
This document (and other CRU Research Findings and Reports) and information about the work of CRU may be viewed on the Internet at http://www.scotland.gov.uk/cru/