Fatal flaws

Chris Beckett is a senior lecturer in social work at
Anglia Polytechnic University, Cambridge, and previously worked as
the manager of a local authority children and families social work
team. He is the author of three books, including Child
Protection: An Introdcution
, published by Sage in

“Since Sweden banned smacking three decades ago child deaths at
the hands of parents have fallen to zero. In Britain, they average
one a week,” – so declared a leading article in a national
newspaper earlier this year.1

The NSPCC made a similar comparison in a press release in 2002. “At
least one child is killed every week by a parent or carer in
England and Wales,” it declared, contrasting this with Sweden, “the
first country to change the law to protect children from being
hit”, where “only four child abuse deaths were recorded between
1981 and 1996”.2 And the same sort of thing – with minor
variations in the figures – has been said by, among others, a House
of Commons select committee and an umbrella organisation for the
anti-smacking lobby (“in the decade after Sweden fully outlawed
smacking and every other form of physical punishment… not one
single child in Sweden died of physical abuse at the hands of his
or her carer” states the submission to the Victoria Climbie Inquiry
from the Children are Unbeatable Alliance). The message is clear:
by the simple act of banning smacking, Sweden has solved the
problem of fatal child abuse. There is just one difficulty with
this. It isn’t true.

The claim that only four child abuse deaths occurred in Sweden in
the 16 years from 1981 (a claim which some seem to have generously
reinterpreted as “no deaths at all”) originates, as far as I can
tell, in a study carried out by Joan Durrant in 1999.3
What is not explained by those who quote it is that this study
adopted a very narrow definition of “fatal child abuse” taken from
an earlier Swedish study.4

This definition excluded all intentional killings, all deaths
resulting from neglect or occurring in the context of a sexual
assault, all killings of newborn babies, all killings by women
suffering from post-natal depression and all killings where the
motive of the perpetrator was not clear. Using this definition,
“fatal child abuse” accounts for only about 5 per cent of Swedish
child homicides – and only 6 per cent of within-family child
homicides. But what the NSPCC and others have done is compare these
figures with UK child homicide figures as a whole. It is a bit like
taking Sweden’s bicycle accident statistics and comparing them with
Britain’s statistics for traffic accidents of every kind.

To confuse things further, what is also sometimes disregarded is
the fact that Sweden’s population of eight million is closer to the
population of Greater London than to that of the UK as a whole.
Looking at child homicide in general – rather than at narrowly
defined “fatal child abuse” – a Swedish government report concluded
that “fewer than 10 children under the age of 15 are killed each
year”.5 Although this is a lot higher than four children
in 16 years, it still sounds very low until we remember that “fewer
than 10” a year would translate into “fewer than 70” if Sweden had
the same population as the UK. Fewer than 70 remains lower than the
UK child homicide rate – a Unicef report suggests that the UK rate
is 50 per cent higher – but it would average out at well over one
child homicide a week.6

The child homicide rate in Sweden is relatively low – as indeed one
would expect from a country with the world’s lowest level of child
poverty, a teenage birth rate which is less than a quarter of
Britain’s and excellent public services – but it is not uniquely so
in the way that some authorities have suggested. In fact, according
to Unicef, six other industrialised countries have lower rates,
with the lowest being Spain, Greece, Italy and Ireland, none of
which have legal bans on smacking. But it would still be possible
to argue that banning smacking had helped to reduce child
homicides, if one could demonstrate that these had fallen in Sweden
in the quarter-century since the 1979 ban on corporal

Although Joan Durrant’s statistics seem to be the ones quoted in
these claims, she herself did not suggest that a fall in child
abuse deaths had occurred since the ban: she simply asserted that
there was no evidence that child deaths had significantly risen. As
to the Swedish government report mentioned earlier, it thinks the
figures may have fallen a bit but is not sure: “Child homicide has
probably decreased during the past 20 years, but the statistics are
not perfect.”

But we can judge the evidence for ourselves. The main Swedish study
of child abuse deaths leading up to the corporal punishment ban in
1979 found 96 child homicides between 1971 and 1980 (85 per cent of
which occurred within the child’s family). This averages out at 9.6
deaths a year. Compare this with the “fewer than 10 a year” which
the 2001 Swedish government report estimated to be the current
child homicide rate and draw your own conclusions.

There are good arguments for banning corporal punishment.
Protecting children, after all, is not just about preventing
fatalities. But a decision on whether to introduce a ban here
should be based on a realistic assessment of the pros and the cons.
Let’s be clear that, whatever other pros might exist, the abolition
of child maltreatment deaths cannot be included among them.


This article looks at the statistics that lie behind the
frequently made assertion that Sweden’s 1979 ban on all forms of
corporal punishment has abolished, or drastically reduced, fatal
child abuse in that country. It concludes that statistics have been
used in a misleading way and that there is no clear evidence that
rates of fatal child abuse have been affected one way or the other
by the ban.


The Observer, 7
March, 2004 

2 NSPCC, Response to the
Report of the UN Committee on the Rights of the Child,

3 J Durrant, “Evaluating the
success of Sweden’s corporal punishment ban”, Child Abuse and
, 23 (5), 435-448, 1999  

4 L Somander, L Ramner,
“Intra- and Extrafamilial Child Homicide in Sweden, 1971-1980”,
Child Abuse and Neglect, 15, 45-55, 1991  

5 S Janson, Children and
Abuse – Corporal Punishment and Other Forms of Child Abuse in
Sweden at the End of the Second Millennium
(English summary),
Stockholm: Ministry of Health and Social Affairs,

6 Unicef, A League Table
of Child Maltreatment Deaths In Rich Nations
. Innocenti Report
Card No 5, Florence: Unicef Innocenti Research Centre,

Further information   

1 S Creighton, Child Protection Statistics, 5, NSPCC
inform, www.nspcc.org.uk/inform/

2 A Hjern, “Child Abuse and Sexual Assault” in Stockholm 
Council, Evidence-Based Health Promotion for Children and
Adolescents in Stockholm County
, (at www.cbu.dataphone.se),

3 N Trocm‚, D Lindsey, “What can child homicide rates tell
us about the effectiveness of child welfare services?”, Child
Abuse and Neglect
, 20 (3), 171-184, 1996 

4 Unicef, A League Table of Child Poverty In Rich
. Innocenti Report Card No 1 (www.unicef icdc.org)




More from Community Care

Comments are closed.