Sweden: data does not support success claims

Campaigners against the physical correction of children frequently cite Sweden as a model. They argue that the 1979 ban on smacking has had thoroughly positive outcomes. Here, Dr Robert E Larzelere of the University of Nebraska Medical Center provides a critique of some recently-published papers.

 

 

This article was previously published in Families First, Issue 2, Autumn 2001.

It is published here with the kind consent of the author and Families First.

 

 

 During the past couple of years, Joan Durrant, of the University of Manitoba in Canada, has published two overlapping evaluations of the welfare of Swedish children during the two decades since Sweden banned smacking by parents in 1979.¹ There is undoubtedly a need for rigorous, objective evaluations of such major policies, particularly as other countries are considering placing limits on the right of parents to employ physical correction. Unfortunately, however, Durrant's conclusions seem to reflect her unconditional commitment to an anti-smacking perspective more than an objective appraisal of the data available from her sources.

Few people outside Sweden can check out her sources since they are written primarily in Swedish and sometimes unavailable in other languages. This critique therefore highlights some of the inconsistencies between her conclusions and data sources. Durrant's four main conclusions concern the effect of Sweden's smacking ban on: (a) attitudes and practices about smacking; (b) physical child abuse; (c) assaults by young people raised after 1979; and (d) safety networks of support for child-rearing in Sweden.        

 

Public attitudes towards smacking

According to Durrant, "the corporal punishment ban and on going public education campaigns appear to have been extremely effective in altering the social climate with regard to corporal punishment.² She supported this with a table showing that the percentage of the Swedish public supportive of corporal punishment dropped from 35% in 1971 to 26% in 1981, and down to 11 % in 1994.

Unfortunately, she not only compared survey questions that were very different in 1981 and 1994, but she used only one of the two responses to the 1994 question that indicated qualified support for corporal punishment. The original survey item used from 1965-1981 was: "A child has to be given corporal punishment from time to time." The percentage in Sweden agreeing with this statement dropped from 53% in 1965 to 26% in 1978 the year prior to the ban on smacking. It then stayed at 26% in 1979 and 1981, 4 which was apparently the last year that item was used in a Swedish national survey.

 

The 11% cited by Durrant in 1994-5 were "positively inclined to milder forms of physical punishment", while a further 22% were "in principle against all form of physical punishment, but can use such punishment if upset enough."5 Only 56% were against all forms of physical Punishment, and the remaining 10% did not choose any of the three options. The survey also included the following item, which was closer to the wording used between 1979 and 1981: "Mild or moderate physical punishment is sometimes necessary as a child-rearing method, but should be carefully considered and not the result of anger." Thirty-four per cent agreed partly or fully with this item, an increase from the 26% support in 1978, just before the 1979 ban. 6

 

"Durrant's conclusions seem to reflect her unconditional commitment to an anti-smacking perspective more than an objective appraisal of the data"

 

This same Swedish survey found that the actual use of physical punishment had dropped very little. Thirty-two per cent of respondents aged 13-15, born during or after 1979, had received physical punishment from their fathers, compared with 34% in the next oldest generation. (3 1 % of those born after the ban on smacking had been physically disciplined by their mothers, compared with 36% in the next oldest age-group). Physical punishment of teenagers changed even less. For example, 17% of 13-15 year-olds reported corporal punishment by their fathers "when a teenager", compared to an average of 16% in the three older generations. Corporal punishment of 13-15 year-olds by their mothers remained constant at 16% across the four age groups.

 

So, in direct contrast to Durrant's conclusion, changes in attitudes towards physical punishment occurred before, the 1979 legislation and have changed very little since then.7 This raises questions about everything else. If the ban on

 smacking failed to change attitudes or practices concerning physical punishment, how could it influence anything else? One possibility is that the ban affected the frequency of physical punishment, even though the percentage of parents using it dropped very little. Another possibility is that physical punishment against the law may have different effects from its use within the law.

 

For example, parents may have become less inclined to use a mild physical sanction when they are still in control and more, likely to use physical punishment when "upset enough". This could easily increase the risk of child abuse.

 

Physical child abuse

One of the major motivations behind the ban on smacking was to reduce child abuse. There have been few evaluations of whether it achieved that goal and no available study has convincingly documented a resulting decrease in child abuse. 8 It is noteworthy then that Durrant focuses only on mortalities classified as due to child abuse in Sweden. Sweden had one of the lowest mortality rates for young children both before and after the 1979 ban on smacking. As Durrant has noted, the change from before to after 1979 failed to achieve statistical significancc.9 However, the same source that she used for her statistics on assaults against children indicates sharply increasing rates of physical child abuse, at least in criminal records of assaults by relatives against children under the age of seven. This frequency increased from 99 in 1981 to 583 in 1994, a 489% increase.10 This could reflect a change in reporting mechanisms, an actual increase, or other factors. Other countries need an unbiased, objective way of deciding among these alternative explanations before emulating Swedish policies.

 

Violence by Young Perpetrators

Durrant concluded that those raised after the 1979 ban on smacking were less likely to be perpetrators of assaults against children, relative to overall societal trends. 11 Table l summarizes the percentage increases in criminal assaults against 7-14 year-olds by the age of perpetrator, from the very source that Durrant used to test this hypothesis. 12 This shows that the largest increases occurred for perpetrators under 15 years of age, who were born after the ban on smacking. The second largest percentage increase occurred for 15-19 year-old perpetrators, who were aged 0-4 when the law was passed in 1979. Thus those raised after the ban on smacking are increasingly likely to be perpetrators of such assaults as they grow up.

 

How then did Durrant arrive at the opposite conclusion? First, she reported the percentage of all perpetrators who were in a particular age, cohort rather than the actual frequency of criminal assaults. This distances the readers from the actual descriptive data and obscures the fact that assaults against children were increasing for all age group& second, she featured adults in their twenties as the youngest cohort. Table l shows that the smallest percentage increase in assaults against 7-14 year-olds was by 25-29 year-olds. People in this age group were at least 10 years old when the ban on smacking was introduced in 1979. By combining them with 20-24 year-olds, she obtained a group that was increasing its criminal assaults against children less rapidly than the older cohorts. Since this category of assaults was increasing less rapidly, it was decreasing as a percentage of all assaults against children.

 

In the journal article, but not in the paper published, by Save the Children, Durrant acknowledged that the number of assaults by 15-19 year-olds against 7-14 year-olds increased significantly from 1984 to 1994, but she claimed that as a proportion of all perpetrators, there was no significant increase among 15-19 year-olds. Nevertheless Table l shows that the percentage increase in assaults against 7-14 year-olds was larger for 15-19 year-old perpetrators than in any older group. Its percentage increase is exceeded only by perpetrators born after the ban on smacking, whose assaults increased by 519% from 1984 to 1994. This trend was reported by Durrant, but in a separate section on youth crime. In her published article she combined data on assaults against children under 7 as well as assaults against 7-14 year-olds. Assaults by 7-14 year-olds against children under 7 are infrequent and relatively stable, so combining the two ages of victims serves to dampen the increase of 519% shown in Table 1.

 

Table 1: Frequency of Criminal Assaults Against Children from 7-14 Years of Age (Wittrock 1995)

 

 

Age of suspect

1984(Birth Year)

1994(Birth Year)

Percentage Increase

under 15

116 (1970+)

718 (1980+)

519%

15 -19

107 (1965-69)

354 (1975-79)

231%

20 - 24

12 (1960-64)

28 (1970-74)

1:33%

25 - 29

19 (1955-59)

29 (1965-69)

53%

30 - 39

68 (1945-54)

151 (1955-64)

122%

40 - 49

47 (1935-44)

116 (1945-54)

147%

50+

25 (<1935            )

57 (< 1945)

128%

 

 

 

Safety Networks of Support for Child-Rearing in Sweden

Durrant claims that interventions on behalf of children have become much more preventive and supporting, occurring earlier in the escalation of violence. I cannot critique her conclusions on this as confidently because I do not have access to her data. I do, however, have access to some other statistics that suggest a different perspective on her statistics, but the entire picture is not completely clear to me. Secondly, whether the current Swedish system is supportive or intrusive towards families is seen very differently by critics of the current Swedish social system than it is by its supporters. After hearing the two perspectives, it is difficult to figure out how the Swedish system is experienced by families, especially by less privileged families, such as immigrants and the poor.

For example, the critics' perspective on the nature of voluntary vs. compulsory social services is quite different from Durrant's. In the compulsory programme, parents are allowed to see their children only once a month for a closely supervised visit. 13  Most parents therefore choose the voluntary programme as the lesser of two evils. Their child might be taken away from their home (37% of new cases in the 'voluntary' programme in 1995, down from 54% in 1982), 14 but at least they will be able to visit the child more often and have a greater chance of being reunited.

 

"those raised after the ban on smacking are increasingly likely to be perpetrators of assaults as they grow up."

 

The number of new children in the compulsory programme for out-of-home care actually increased by 7% between 1982 and 1995,15 although you would not know that from Durrant's journal article. Neither would you know that what Durrant calls "support and care measures" consist of removal from the home for 46% of new children in the system, down from 60% being removed from the home when the programme started in 1982. 16 So her hypothesis that early identification of problems was "intended to lead to earlier, more supportive intervention" turns out to mean removal from the home in at least 46% of the new cases receiving "support and care measures". This is not the kind of "increasingly preventive" child welfare measure that I would welcome.

 

Two sources have claimed that children are much more likely to be removed from their homes in Sweden than in other European countries, though I cannot vouch for their statistics. Using 1981 data, Ivarsson records that around 22,000 children in Sweden were removed from their homes in 1981, compared with 1,900 in Germany, 710 in Denmark, 552 in Finland, and 163 in Norway.17 Durrant corroborates this by showing that there were 22,807 Swedish children in out-of-home care in 1982, 4,839 for the first 'time that year. 18 According to Westerberg, Sweden has about 15,000 children in care, compared with 40,000 in Great Britain, which has a population more than seven times the size of Sweden.19

 

Durrant also implies that some of the increases, such as those in criminal assaults against children, are due to increased reporting because of increased awareness and because more minor incidents are considered reportable. It is difficult to tell the extent to which increased statistics on child abuse and on assaults by minors reflects a genuine increase or a change in reporting mechanisms. Several facts suggest the increases are actual and not only due to changes in reporting mechanisms.

 

Firstly, Gelles and Edfeldt compared the Swedish rate of physical child abuse in 1980 with figures from two American surveys conducted in 1975 and 1985. 20 As Durrant pointed out, the 1975 American response rate was lower than the Gelles-Edfeldt survey undertaken in Sweden in 1980. This was probably because that American survey used face-to-face interviews, whereas the Swedish survey used telephone calls. Fortunately, the 1985 American survey used telephone calls and had an even higher response rate than the Swedish survey. Considering a variety of factors, the fairest and most conservative comparison was to compare the Swedish child abuse rate with the average of the two American rates. By this method the Swedish child abuse rate was 49% higher in 1980 than the average of the 1975 and 1985 American rates.21

 

Note that a supposedly anonymous survey would not be influenced by the procedures by which physical child assaults get recorded in criminal records. Yet 3% of Swedish parents reported beating up their child. in 1980 compared with 1% of American parents in 1975 or 1985. I first thought that these 1980 statistics might reflect a temporary increase in child abuse as Swedish parents were adjusting to child-rearing without smacking. I am now doubting that because I have seen no evidence of a decrease in Swedish child abuse rates since then.

 

Secondly, Durrant's view that more minor incidents are being recorded as criminal assaults would suggest that the most serious category (aggravated assaults, punishable by 1-10 years in prison) should be increasing more slowly than criminal assaults in general. However, serious aggravated assaults against all children increased 388% from 1984 to 1994, whereas assault suspects in general increased 277%.22 Durrant maintains that this difference is not statistically significant, but it provides strong evidence against criminal assault records; becoming increasingly predominated by mild incidents.

 

Thirdly, the timing and suddenness of the increase does not support a reporting interpretation. Durrant implies that the ban on smacking and the 1982 changes in social services had the commendable purpose of enhancing an early warning system for violence before it got more serious. That would suggest a sharp increase during the 1980s. If this was in fact preventive, then criminal statistics for physical child abuse and assaults by minors should level off or decrease subsequently. However, both statistics increased relatively little during the 1980s and then increased sharply at an accelerating rate in the 1990s. Children whose preschool years from 2-6 were entirely under the ban on smacking first became teenagers in 1990. From 1984-1989 the average annual increase in assaults by minors against minors was 3.4%. From 1990-1994, the average annual increase was 17.9%.

 

Fourthly, as noted by Durrant, if the increase is entirely due to reporting differences, the same increase would not be reflected in victimization statistics. This would have been a strong point in Durrant's favour if her sources had reported trends in victimization statistics for 7-14 year-olds. Instead, von Hofer showed that the victimization statistics for 16-20 year-olds were stable over this time period. 23 Indeed, von Hofer’s article corroborates my Table 1 in that he shows that the rate of criminal assaults by minors (7-14 years) has increased much more dramatically than older groups. His statistics are based on a wider age range of victims than the child victims that Durrant and I are interpreting differently. Further, von Hofer showed that the incidents requiring medical attention doubled for 16-20 year-olds. The latter trend suggests that the average victimization incident is getting more severe and not less severe as Durrant implies.

 

 

Conclusion

Durrant draws conclusions about four major trends since Sweden’s 1979 ban on smacking. On the first three trends, the very sources she cites strongly suggest conclusions opposite to hers. Attitudes and practices about corporal punishment have changed very little since 1979. In fact such changes were far more dramatic before 1979.

 

Secondly, the best indicators of physical child abuse showed a 489% increase in physical child abuse cases classified as criminal assaults in Sweden from 1981-1994. Child abuse fatalities have been infrequent in Sweden both before and after the 1979 legislation, though not as low as Durrant claims.

 

Thirdly, the best evidence suggests that perpetration of criminal assaults against 7-14 year-olds is increasing most rapidly in age groups raised after the law against smacking was passed. This directly contradicts Durrant's conclusion based on selected evidence from the same data source.

 

On the fourth issue, it is difficult to evaluate the extent to which the Swedish social services have achieved an optimal balance between a preventive approach to protecting children on the one hand, and becoming overly intrusive on the other. Most of the evidence suggests that the large increase in assaults by minors and in physical child abuse is not entirely explained by changes in reporting mechanisms. Critics of the law do not think the increase has been caused entirely by the ban on smacking. Rather, the critics say that the influence of parents has been inadvertently compromised by the entire set of overly intrusive Swedish policies. Because parents have been disempowered, the police must intervene in many more incidents than was previously the case.

 

Sweden has historically been a very non-violent country, especially compared to the United States. Perhaps Sweden can afford a sixfold increase in criminal assaults by minors against minors. Other countries cannot. Accordingly, we need to get a much more convincing explanation of this increase in Sweden before other countries regard Sweden as an example to emulate. We also need objective, unbiased evidence that their policies have reduced physical child abuse. Such evidence does not currently exist. 24

 

To repeat the main conclusion of my 1999 article, we need timely, rigorous and unbiased evaluations of these kinds of policy changes in the future. The degree of bias in Durrant's evaluations increases my suspicions that the success of Sweden’s ban on smacking is uneven at best and counterproductive at worst. Successful policy changes do not need their evaluations to be biased to document their success.

 

Notes:

1 Durrant, J E (1999) Evaluating the success of Sweden’s corporal punishment ban. Child Abuse & Neglect, 23, 435-448.

Durrant, J E (2000) A Generation Without Smacking: The impact of Sweden’s ban on physical punishment. London: Save the Children.

2 Durrant (1999), op cit.

3 Solheim, J S (1982) A cross-cultural examination of use of corporal punishment on children: a focus on Sweden and the United States. Child Abuse and Neglect, 6, 147-154.

4 Ziegert, K A (1983) The Swedish prohibition of corporal punishment: a preliminary report. Journal of Marriage and the Family, 45, 917-926.

5 Statistics Sweden (1996) Spanking and other forms of physical punishment. (Demography, the Family and Children 1.2) Stockholm, Sweden: Statistics Sweden.

6 Sanden, A & Lundgren, L (1997) Spanking of children much less common. Statistika centralbryan. http://www.scb.se/scbeng/vhtm/barnaga.htm

7 This analysis of the relevant data is corroborated by Roberts, J (2000) Changing public attitudes towards corporal punishment: The effect of statutory reform in Sweden. Child Abuse & Neglect, 24, 1027-1035.

8 Larzelere, R E & Johnson, B (1999) Evaluation of the effects of Sweden’s spanking ban on physical child abuse rates: A literature review. Psychological Reports, 85, 381-392.

9 Durrant (1999), op cit.

10 Wittrock, U (1992) Barnmisshandel i kriminalstatstiken 1981-1991 [Violent crimes against children in criminal statistics, 1981-1991]. KR Info, 1992, 7; Witrock, U (1995) Barnmisshandel, 1984-1994 [Violent crimes against children, 1984-1994]. KR Info,

1-6. These papers are available in English at: http://people.biola.edu/faculty/paulp/

11 Durrant (1999), op cit.

12 Wittrock (1995), op cit.

13 Westerberg, S (1999), How parents were deprived of their rights in Sweden. London: Family Education Trust.

14 Durrant (1999), op cit.

15 Durrant (2000), op cit.

16 ibid.

17 Ivarsson, M (1984) Sverige 1984. Malmo, Sweden: Lehmanns Forlag.

18 Durrant (2000), op cit.

19 Westerberg (1999) op cit.

20 Gelles, R J & Edfeldt, A W (1986) Violence towards children in the United States and Sweden. Child Abuse and Neglect, 10, 501-510.

21 Larzelere, R E & Johnson, B (1999) op cit.

22 Wittrock (1995) op cit.

23 von Hofer, H (1995) Criminal violence and youth in Sweden in a long-term perspective. Paper presented at the Tenth Workshop for Juvenile Criminology, Siena.

24 Larzelere, R E & Johnson, B (1999) op cit.

 

 

Smacking and the Law - a European Perspective

 

Families First

 

Corporal Punishment of Children

 

Back to Article Index