Study	Reason for exclusion
August 2001	Author stated to be primarily a community based programme. The school component of a multicomponent intervention is a summer school intervention, i.e. not delivered during the normal school routine.
Blue 1982	Unable to confirm whether allocation to intervention and control groups was random, and whether the intervention was school based. Unable to contact author.
Dubow 1987	Randomisation method poor.
DuRant 1996	Cluster randomisation of 2 schools. N = 2 studies excluded.
Gottfredson 1986	Excluded due to primary focus on delinquency prevention and not aggression.
Gottfredson 1987	Excluded due to primary focus on delinquency prevention and not aggression.
King 1990	A randomised controlled trial to assess a social development program. Unable to contact author to determine eligibility as a violence prevention programme. Published data collapsed groups, losing randomisation.
Larson 1992	Cluster randomisation of 2 classes. N = 2 studies excluded.
Lochman 1985	Subjects were allocated to groups quasi‐randomly, using alternation.
Lochman 2002	Intervention primarily designed to reduce substance use.
Mason 1997	Subjects were allocated to groups quasi‐randomly, using alternation.
Mayer 1983	Study designed to reduce vandalism costs and secondly to assess effect on disruptive behaviour. Not designed to reduce violent behaviour.
Miller 1990	1) Aim of study was to improve moral reasoning, rather than reduction of violent behaviour per se. 2) Personal communication with author to CD indicated that 'randomisation' was in fact by alternation.
Smith1991	Does not contain the terms violence, aggression, bullying, conflict or anger in either the title, abstract or study aims/hypotheses. It is intended to reduce delinquency and disruptive behaviour through improved self esteem.