Data concerning all young (15-20 years, n = 56) sex offenders (YSOs) subjec
ted to forensic psychiatric investigation in Sweden during 1988-1995 were u
sed in an attempt to construct and validate an introductory YSO typology ba
sed solely on offence characteristics. A 5-cluster solution received optima
l support from cluster analysis of 15 offence-related variables. A few hist
orical and clinical characteristics varied across clusters. Survival analys
es revealed that the clusters differed with respect to sexual but not to vi
olent or general reconviction rates. (C) 2000 The Association for Professio
nals in Services for Adolescents.