With urban development at the heart of crime prevention policy – which led to what is sometimes sceptically called a carousel of small-scale pro-jects for and with inhabitants, especially in deprived neighbourhoods – community prevention has flourished over recent decades (icpc 2014). The lion’s share of the evidence presented here regarding community interven-tions is borrowed from a recent and detailed international review by Lub (2013). It is important to note that Lub bases his conclusions on a broad range of empirical studies, not just studies that meet the (quasi-)experi-mental criteria of impact evaluations. He distinguishes between several types of interventions, with varying degrees of scientific support.
The first is encouragement of social cohesion among local residents, for which policy theory often posits a three-way effect. First, social cohesion is
thought to encourage social control, which in turn reduces social incivili-ties and crime. Second, social cohesion is expected to mobilise inhabitants to become actively involved in their neighbourhood, and subsequently improve local safety conditions. Lub concludes that there is little scien-tific basis for these intervention theories and that there is more evidence for a reversed causality: the presence of active inhabitants and those who exert social control encourages local social contacts. Activism and social control strongly depend on individual characteristics, such as personal competences, motivation, sense of duty and neighbourhood attachment.
Third, social contacts between deprived social groups and middle-class inhabitants are thought to enhance the cultural and social resources of the former. Along with their prospects, the prospects of the neighbourhood itself improve. Once again, Lub finds that social contacts between these two groups – ‘bridging social capital’ in Robert Putnam’s terminology (2000) – hardly occur in reality. Neighbourhood improvements are more often due to a changed composition of the population with the influx of middle-class families, than to bridging social contacts. Hence, according to Lub’s review of international evaluations, the preventive impact of local social cohesion is overrated.
A second type of intervention revolves around policy involvement of inhabit-ants. Here, the evidence is slightly more persuasive regarding the under-lying logic of the assumptions, but almost as critical regarding the actual effects on crime and livability. First, it is assumed that policy interventions become more focused and effective when citizens provide authorities with local information. Second, the same is assumed if citizens are allowed to decide on neighbourhood priorities. Both policy assumptions have been poorly tested, but the little evidence that is available points to weak or no effects on crime and incivilities. Sherman and Eck (2002) confirm the lack of effect of community meetings and information points where citizens can meet police officers. The results of contacts between citizens and authorities tend be more promising for feelings of safety. Several studies have however observed that, if an active group of inhabitants happens to be successful in conveying their concerns to authorities, the latter may disproportionately focus on issues that may not be shared or may even be contested by other groups. This is undesirable from the perspective of social cohesion and local democracy.
A more promising strategy involving social contacts between citizens and authorities is community policing, also known as reassurance policing.
To strengthen their legitimacy and establish trust, the police invest in long-term relationships within the community. They do this by treating inhabitants, victims and offenders alike with respect and by lending them a listening ear. Once trust in the police has been established, citizens are more willing to conform to local norms as well as the law (Sherman and Eck 2002).
The results for informal surveillance through neighbourhood watch schemes are quite ambiguous. An extensive meta-analysis by Sherman and Eck (2002: pp. 315-317) is very clear about the lack of effect of neighbourhood watch schemes on crime. They even seemed to incite fear of crime, rather than reduce it. By contrast, the majority of evaluations in a later meta-anal-ysis by Bennett (2006, in: Lub 2013) found positive and substantial effects for neighbourhood watch in reducing crime. This positive finding is more in line with visible human surveillance in general (see situational prevention), whether by the police or other public officials (Van Noije and Wittebrood 2008).
Sports and other leisure activities for youngsters are popular neighbourhood interventions. Such projects are thought to contribute to a safer communi-ty in several ways. First, such interactive activities are thought to enhance social competences and promote prosocial behaviour. Little or no effects on community safety have been found. Causality is a major issue here, as it is often socially competent youngsters who engage in sports and games.
Moreover, the competitive nature of sports can also trigger aggression.
Lub concludes that it is not sport per se, but the combination with moral and pedagogic guidance, that may make a difference. Second, sport and game facilities are intended to function as venues for new social contacts within the community, enhancing social cohesion and neighbourhood safety. In practice, however, these venues are used by existing networks of youngsters, excluding others. They are even known to become a source of nuisance for other inhabitants, with negative effects on social cohesion and neighbourhood safety. Third, and most promising, engaging young people in leisure activities has been shown to substantially reduce crime and nuisance, due to the simple fact that time spent on sports cannot be spent on offending. Hence, sport may serve as a substitute, but is unlikely to fundamentally change behaviour (Lub 2013).
Social projects in the form of self-regulating codes of conduct have proved unable to improve neighbourhood safety. This is most likely because the neighbourhood – or even the street –no longer functions as a community in which effective group norms can be enforced. Hence, often only limited and selective groups of inhabitants are willing to comply. Note, however, that this critique applies specifically to codes of conduct within neigh-bourhoods; there is evidence to suggest that they might work in schools to reduce criminal and aggressive behaviour among students, if embedded in a broader and well-structured school policy (Gottfredson et al. 2002).
Redevelopment, a drastic physical neighbourhood intervention, has been evaluated to a considerable extent. The objectives of redevelopment are not so much physical as social. By replacing old rented housing with new homes, not only are better living conditions realised for existing resi-dents, but often a new population with a better socioeconomic position can be attracted. Mixing inhabitants with weaker and stronger cultural,
social and economic capital dilutes the concentration of social problems.
At best very small effects (Wittebrood and Van Dijk 2007; Wittebrood and Permentier 2011) or no effects (Ouwehand and Davis 2004; Bolt and Torrance 2005; Van Beckhoven and Van Kempen 2002, 2006; Van Bergeijk et al. 2009) should be expected from this mixed population on social cohesion (as was mentioned at the start of this section) due to the fact that different social groups tend to prefer to keep to themselves. Nonetheless, several quasi- experimental evaluations have reported effects of redevel-opment on crime (most consistently on violence), feelings of safety and/
or neighbourhood satisfaction (Wittebrood and Van Dijk 2007; Wittebrood and Permentier 2011; Permentier, Kullberg and Van Noije 2013). Effects seem larger when social housing is replaced by new owner-occupied homes than when social housing is replaced by social new-builds (Wittebrood and Permentier 2011). Effects were also found to be larger for large-scale and long-term interventions (Permentier, Kullberg and Van Noije 2013).
Overall, current research seems at odds with the zest with which social projects have been launched in the neighbourhood over the past years.
This is confirmed by a technically advanced impact evaluation of a recent government perennial investment programme in the 40 most deprived urban areas in the Netherlands, which – overall – hardly seems to have made any difference in these intervention areas compared to other deprived areas (Permentier, Kullbert and Van Noije 2013). A note of cau-tion is in order, however, as the review evidence is mostly disappointing with respect to impact on actual crime and incivilities, less so on feelings of safety. Limited effects of community prevention on objective safety and greater potential for subjective safety are also found in Van Noije and Wittebrood (2008), especially as far as social activities by and for inhabit-ants are concerned. This effect of social mobilisation on subjective safety may be explained by the fact that activities in the neighbourhood are open for everyone to see and may give the impression of an active social network that is making an effort to improve local safety, apart from the actual improvements. Similarly, it is quite possible that social community projects make a valuable difference for other subjective indicators, such as neighbourhood satisfaction or even the level of perceived incivilities (Welsh and Hoshi 2002).