Psychological Medicine

Original Articles

Violent crime runs in families: a total population study of 12.5 million individuals

T. Frisella1a2 c1, P. Lichtensteina1 and N. Långströma1a2

a1 Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, Stockholm, Sweden

a2 Centre for Violence Prevention, Karolinska Institutet, Stockholm, Sweden


Background Etiological theory and prior research with small or selected samples suggest that interpersonal violence clusters in families. However, the strength and pattern of this aggregation remains mostly unknown.

Method We investigated all convictions for violent crime in Sweden 1973–2004 among more than 12.5 million individuals in the nationwide Multi-Generation Register, and compared rates of violent convictions among relatives of violent individuals with relatives of matched, non-violent controls, using a nested case–control design.

Results We found strong familial aggregation of interpersonal violence among first-degree relatives [e.g. odds ratio (OR)sibling 4.3, 95% confidence interval (CI) 4.2–4.3], lower for more distant relatives (e.g. ORcousin 1.9, 95% CI 1.9–1.9). Risk patterns across biological and adoptive relations provided evidence for both genetic and environmental influences on the development of violent behavior. Familial risks were stronger among women, in higher socio-economic strata, and for early onset interpersonal violence. There were crime-specific effects (e.g. ORsibling for arson 22.4, 95% CI 12.2–41.2), suggesting both general and subtype-specific familial risk factors for violent behavior.

Conclusions The observed familiality should be accounted for in criminological research, applied violence risk assessment, and prevention efforts.

(Received November 02 2009)

(Revised February 04 2010)

(Accepted February 06 2010)

(Online publication March 25 2010)


c1 Address for correspondence: T. Frisell, M.Sc., Centre for Violence Prevention, Karolinska Institutet, PO Box 23000, 104 35 Stockholm, Sweden. (Email: