This paper is really thought provoking. In fact it organizes considerable research evidence, some of it known, some more recent under the comprehensive concept of ‘social capital’. As such it is a courageous attempt to unite the world of psychology/psychiatry with sociology in studying aggressive behaviour in a developmental perspective.

My main question is whether social capital can be used in the way the authors did, given the numerous definitions of social capital, and it is this aspect that I find somewhat problematic. The concept of social capital as it is used in the paper looks to me as a kind of unifying passepartout, a little like Freud’s concept of reaction formation, which can practically explain all kinds of behaviour. This does not imply that it is not useful. However, it seems to me that one should stick carefully to a clear definition, such as, for example, the one Coleman (1990)Footnote 1 or Lin (1999)Footnote 2 proposed, both cited by the authors.

A second observation is that there are practically no European studies cited. Since the paper is based on literature I do not understand why it is mainly based on US as well as some Canadian and UK literature. Was it really not possible to find some more Continental or Scandinavian (English) publicationsFootnote 3 relevant to the subject at hand? For a paper published in a European Journal this is regrettable.

However, returning to the main subject of this comment and keeping in mind the two definitions cited above, how does this work out?

The authors rightly make a distinction between social capital at the micro-level and at the macro-level and this should be kept in mind when considering the paper.

The first part of the study dealing with child welfare seems to me quite adequate. One might expect that investments in welfare -- in particular income support, healthcare and education -- would profit families by easing the conditions under which children grow up and that will give them more options in terms of further education and employment and thus a stake in conforming behaviour. Indeed it would have effects on crime later on (Esping-Andersen 2002; OECD 1999; Schweinhart et al. 1993, 2005).

Moreover, the International Self-report Delinquency Study (ISRD), in its first comparative volume of 11 countriesFootnote 4 (Junger-Tas et al. 2003a) found that the so-called ‘broken family’, which in most cases means father absence, was closely related to delinquency in Anglo-Saxon countries, but much less so in N-W Europe. One of the tentative explanations for this outcome might be that there are major differences between these two country clusters in income transfer system and social welfare. In N-W Europe the system provides for considerable support to lone mother households with small children so that these families have at least some stable socio-economic position, while the situation in Anglo-Saxon countries is less favourable: they have considerably more single teenage mothers, who are then forced to participate in the labour market but often without access to substitute child care. And all research evidence, some of which is also cited in the paper, shows indeed that is not mother employment per se, but the lack of supervision and the quality of available daycare that are related to delinquency.

I have more problems with the definition of social capital as it is used in the section on the family. I tend to agree with Sampson’s definition, which considers the family as a kind of entity having (or not) social relations and links to the larger community without including all kinds of intra-family structural and psychological characteristics in the definition. After all the term social capital refers to links with the community and the social environment.

I do not dispute the value of the cited studies: these are excellent studies produced by outstanding researchers. However, in the perspective of Sampson’s definition, many of the elements cited in the paper might be seen as favourable or unfavourable conditions for providing the family with more or less social capital rather than social capital itself. For example, this would be the case for family structure, family status, and delinquency of parents. It is also the case for such parental characteristics and behaviour as ‘time spent with their children’, ‘the strength of the bonds between parents and children’ and ‘parent’s critical attitude towards delinquency’. These characteristics are not independent variables, but they are highly correlated with family status, such as unemployment, poverty, ignorance (Sampson and Laub 1993; Sampson et al. 1997, 1999).

In this perspective the authors give a convincing explanation of how delinquency of parents reduces the family’s social capital, which may lead to delinquency of their children. The reasoning does perfectly confirm the definition of social capital cited above. Unfortunately and regretfully most of the studies that are cited, although perfectly valid, seem to belong to the category ‘black box’. For example, there is no explanation of the ways in which family structure operates to lead to teacher rated aggressive behaviour, or to convictions of violent crimes at age 18.

Let me explain this somewhat more. Together with the health authorities of the city of Rotterdam, I did a study of a representative school population of age 15 (9th grade (Junger-Tas et al. 2003b). The survey was conducted in 33 city schools and reached some 5,000 students. Some 60% belonged to one of the main ethnic minorities in the country.Footnote 5 The most delinquent groups in the city were Antilleans and Moroccans. The least delinquent were the Dutch, Turkish and Surinamese. If we look at family composition then it appears that 90% of the Moroccan and Turkish youth live in a complete family, 80% of the Dutch do so, 50% of the Surinamese and one third of the Antilleans.Footnote 6 However, social capital of the young people was seriously hampered by ethnic status. Their parents were mostly unemployed, illiterate, lived in the worst sections of the city and had no understanding whatsoever of how Dutch society is organized. The young children were badly prepared for primary school and -- according to the youths -- received little support from their parents while they pursued their school career. Many dropped out and did not find (unskilled) employment. In addition, relations with parents were not as good and trustful as those of Dutch youth. More important is that parents failed to control their sons, in particular in the Moroccan community, while they exercised strict control on girls. The Turkish community did exercise control on boys and girls and they were among the least delinquent group. Integration in Dutch society was essential showing that delinquency of better integrated Surinamese youth reduced delinquent behaviour. In fact this study is an illustration of the paper’s thesis, in that it clearly shows the harmful effects of a family’s lack of social capital on their children.

As for the Zolotor et al. (2006) study, I would have liked to know which neighbourhood’s characteristics as well as willingness to take personal action were measured. There might be a strong social status effect.

School is rightly considered as an important social institution creating social capital for students. In this respect the cited studies are a neat illustration of this statement. The same is true for the peer group, although this is rather complex, since joining a peer group is partly the result of simply joining those youths that one meets in the neighbourhood, partly a selection effect (Thornberry and Krohn 2003). However, it is a well-known fact that pro-social peer groups -- such as student groups (Eton boys, university students) growing up together, are extremely important for later career options in society, and mutual support among them often persists over the years. Of course the same may be true for anti-social peer groups.

The cited studies of regional social capital and violent crime are interesting and make the authors’ point quite clearly, similar to the conclusions which I can but support.

Concluding this commentary I would like to say that this is an interesting study opening up great possibilities for further research on the effects of social capital on social-economic options and on either conforming or delinquent behaviour of children, adolescents and adults. I think there is a need, however, to better specify and define the concept of social capital, so that the use of this concept in research will be improved, because if a concept is too large it becomes meaningless.

Finally, I would like to encourage those who wish to publish a paper in a European Journal, to make an effort to collect as many qualitatively sound European studies as possible.