- Navigation menu list for the main sections in this web site
- What we do
- Get involved
- Help and advice
- Under 18?
- Donate
There has been a spate of national newspaper reports which claim that the number of child homicides in England and Wales has halved since the 1970s, that Britain has one of the lowest child homicide rates in the developed world, and that Britain has the second largest reduction in baby and infant intentional killings in the last 25 years1.
These claims appear to contradict statements put out by the NSPCC, most recently as part of its Child Abuse Deaths initiative, that the number of child killings in England and Wales has remained constant for almost 30 years and that the rate of infant killings is higher here than in several other European countries.
This paper outlines how the NSPCC comes to the conclusions it does about child killings and challenges the reliability of these media reports.
Reliable estimates of the number of children killed in England and Wales can be derived from two government publications - Criminal Statistics and Mortality Statistics. Both publications include figures of the number of children who are victims of homicide. Child homicide is defined as the infanticide, manslaughter or murder of babies or children.
The Home Office publishes police records of all homicides reported to it in Criminal Statistics. The criteria for recording homicides have not changed since 1974. Criminal Statistics is published annually around November and provides homicide figures for the previous year and the nine years preceding that.
Criminal Statistics are broken down by sex and the following age groups for children: under one year; one and under five years; five and under 16 years. The report provides data on the annual number of child homicide offences and the number per million population.
Criminal Statistics includes in its most recent figures reported homicide cases which have not yet resulted in a conviction. These figures are adjusted in subsequent reports to take into account cases for which there is no homicide conviction. The adjustment is usually small.
The Office of National Statistics (ONS) provides figures in Mortality Statistics on the number of children who die in any one year including those who are killed from specified causes like strangulation or poisoning. Publication of Mortality Statistics is more erratic than Criminal Statistics.
Figures on child deaths are taken from death certificates provided by local Registrars and Coroners reports. They are broken down by sex and several age groups up to 16 years2.
Mortality Statistics not only includes data on the numbers of children dying in one year from 'homicide and injuries purposely inflicted by other persons'. It further includes figures on the numbers of children dying from injuries where it is 'undetermined whether they were accidentally or purposely inflicted'.
In 1979, the ONS began to assign all deaths awaiting the outcome of criminal proceedings to a sub-category of these 'undetermined deaths', effectively halving the number of child deaths recorded as homicides3.
Although nearly all such cases do result in a conviction, homicide figures in Mortality Statistics are not subsequently amended to take this into account (although relevant information is provided in an appendix).
ONS child homicide figures prior to 1979 cannot therefore be reliably compared to child homicide figures in subsequent years without taking into account the supplementary information on child homicides provided in the appendices.
The number of child homicides fluctuates every year. However, the annual number of child homicides recorded in Criminal Statistics has averaged out at 79 a year for the last 28 years. This provides the substantiation for the NSPCC statement that on average one to two children are killed every week.
The Home Office also provides unpublished figures on the relationship between the child victims of homicide in any one year and the principal suspect. Latest figures for 2000/2001 show that parents were the principal suspect in 78 per cent of child homicides.
Media reports that the number of child homicides has dropped in England and Wales are derived by taking homicide figures in Mortality Statistics from selected years in the 1970s and comparing them to selected years in the 1990s. This is invalid for two reasons.
First, the changes in recording criteria introduced by ONS in 1979 effectively halved the number of child homicides recorded in Mortality Statistics.
Second, the numbers of child homicides in any one year are small and chance variation from year to year can be large. Because of this, you cannot reliably compare the numbers of child homicides in selected years with the number of child homicides in selected later years. This runs the risk of comparing particularly low years with particularly high ones.
There is therefore no basis for media claims that child homicides have dropped since the 1970s.
To reliably establish a trend in child homicide figures over a period of time one needs to look at the figures over the whole period. Analysis of figures from Criminal Statistics shows that the overall rate of child homicide has, with some fluctuation, remained constant for 28 years. There is no trend either upwards or downwards.
Mortality Statistics are more difficult to analyse than Criminal Statistics. However, research shows that the data on child homicides from Criminal Statistics is comparable with those from Mortality Statistics, if you include in the latter deaths awaiting the outcome of criminal proceedings (nearly all of which result in a conviction). As with Criminal Statistics there is no significant trend in rates of child homicide in the ONS data between 1979 and 1997 - the last year for which figures have been analysed in detail4.
Both Criminal Statistics and Mortality Statistics provide homicide figures for England and Wales but not for Northern Ireland or Scotland. Neither Criminal Statistics nor Mortality Statistics provide homicide figures for young people aged 16 and 17 - who face a relatively high homicide risk.
The World Health Organisation (WHO) provides information on all deaths recorded in different countries. International comparisons on childhood deaths from accidental and deliberately inflicted injury are not as robust as comparisons on deaths from childhood diseases. There are important differences in national laws and practice on how accidental and violent deaths are investigated, certified and registered which affect the completeness, accuracy and timeliness of the data.
WHO figures on child homicides in England and Wales are derived from Mortality Statistics but do not include those deaths awaiting the outcome of criminal proceedings. They are a serious underestimate of child homicides in England and Wales.
There is therefore no basis for the kind of international comparisons on child homicides reported in the media.
WHO also records the number of deaths 'undetermined whether accidentally or purposely inflicted' for different countries.
A realistic international comparison of the number of babies killed in England and Wales can be derived by including 'undetermined' infant deaths along with the homicide figures, since nearly all the 'undetermined' infant deaths do eventually result in homicide convictions. This shows that, during the 1990s, England and Wales had a high rate of infant homicides and undetermined deaths compared with several other Western European countries5.
As noted above, there has been no reduction in the rate of homicides of under one year olds in England and Wales since 1979. Recent figures in Criminal Statistics show that both the numbers and rate of homicides of under one year olds increased over the most recent five year period from 1996 to 2000/2001.
There is therefore no basis for media reports that Britain has the second largest reduction in baby and infant intentional killings in the last 25 years.
For older age groups, it is less valid to include 'undetermined' deaths along with the homicide figures as more of the 'undetermined' deaths are likely to result from other causes such as suicide and suspicious injuries.
The NSPCC has accordingly confined its international comparisons of child homicide to the WHO data on homicides and undetermined deaths of babies under one.
The NSPCC has called for the following changes:
It is clearly important to establish good statistical information on child homicides. But it is also essential to establish meaningful measures of effectiveness for the child protection system as a whole.
A Government objective for Children's Social Services is 'to ensure that children are protected from emotional, physical, and sexual abuse and neglect (significant harm)'. A sub-objective of this is 'to reduce the number of deaths of children, where abuse or neglect is a factor'. The proposed measure of this sub-objective - the collection of information from an annual survey of serious case reviews to the Department of Health - will not meaningfully measure the reduction of these deaths.
There is also a need for a process that better identifies the cause of death of children. That is why the NSPCC is campaigning for Child Death Review Teams to be established to identify all the factors that may have contributed to the death. This will enable strategies to be put in place to prevent many future child deaths.
1. The Sun, leader, August 21 2002; The Guardian leader 14 October 2002; Guardian Society, 16 October 2002; Guardian On-line 24 October 2002).
2. The information on child homicides is recorded in two publications DH3 "Mortality Statistics: Childhood, infant and perinatal " and DH4 "Mortality Statistics: Injury and poisoning". Recently DH3 has stopped recording a detailed breakdown of post-neonatal (28 days - 15 years) causes of death, so data on neonatal (0 -28 days) homicides has to be extracted from one publication and post-neonatal homicides from the other.
3. The ONS changed their reporting criteria following the Coroners (Amendment) Rules 1977 and the Registration of Births, Deaths and Marriages (Amendment) Regulations 1977 coming into force. These assigned all deaths awaiting the outcome of criminal proceedings to a separate category, (E988.8) 'injury by other specified means' - known as 'accelerated registrations'). This effectively halved the number of child deaths recorded as homicides from 94 in 1979 to 47 in 1980.
4. See paper by Rooney and Devis (1999) in Health Statistics Quarterly on 'Recent trends in deaths from homicide in England and Wales'.
5. Out of Sight: NSPCC report on child deaths from abuse 1973-2000. 2nd Edition. January 2001. NSPCC.
This research briefing is based on a review of research and literature. It reports the findings and views of a range of authors. These views are not necessarily the views of the NSPCC.
Although the websites listed here are checked regularly, the constantly changing nature of the internet means that some sites may alter after we have viewed them. The NSPCC is not responsible for, nor does it necessarily endorse, the content of these external websites.
Copyright © 2006 NSPCC - All rights reserved. National Society for the Prevention of Cruelty to Children, Weston House, 42
Curtain Road, London EC2A 3NH. Incorporated by Royal Charter. Registered charity number 216401. NSPCC, charity registered in Scotland, charity number SC037717.