What works to prevent youth violence: evidence summary

Annex C: Classification of Intervention Effectiveness

Drawing on definitions and terminologies used by NICE and DFiD a comprehensive classification system has been developed to categorise the effectiveness of interventions based on available evidence. This decision-making tool has been used to determine effectiveness ratings throughout this report on what works to prevent youth violence. It has been used alongside a purposively designed decision tree presented in Annex D.

Effectiveness ratings applied to interventions

Effective

Evidence that the intervention is associated with a positive impact on preventing violence, based on a moderate or strong evidence base. Due to the complexity of causality, an ‘effective’ intervention should be considered one that contributed towards violence prevention or mitigation rather than one that single-handedly accounts for a decrease in violence.

Promising

Findings were positive but not to the extent that they constituted evidence that an intervention was ‘effective’, this could be:

(i) in cases where an intervention has a positive impact on an intermediate outcome, rather than in reducing violence itself

(ii) where authors noted a positive change, but expressed doubts as to whether the intervention could confidently be said to have contributed to this (e.g. due to evidence being rated as “weak” or the other factors potentially having an impact).

Mixed

Findings of individual article -

(i) An individual article that finds varied impact of a single intervention across research sites, or populations.

(ii) An article examining multiple strands of an interventions that finds some were effective/promising and others not.

Findings from a number of studies -

(i) Where there have been a number of studies and the results contrast – e.g. some found positive effects and some did not.

(ii) Similarly, a body of evidence that is mostly comprised of individual articles finding a ‘mixed’ impact of interventions would be considered ‘mixed’ overall. 

No effect

No evidence of effect (positive or negative) of the intervention on reducing violence includes moderate or strong evidence found the intervention had no effect on reducing violence 

Negative effect/ Potentially harmful

Evidence that the intervention is associated with worse violence outcomes (e.g. worse than at the start of the intervention, or worse than for a control group). 

Inconclusive

Insufficient evidence to make a judgement on impact.