Background: This paper examines the effects of deviant and non-deviant partner involvement at age 21 on patterns of continuity and change in offending between the ages of 18 and 21 years in a birth cohort of 983 young men and women. Results: Results showed that those involved with a non-deviant partner had lower rates of offending at age 21 than those with no partner, whilst those without a partner had lower rates of offending at age 21 than those involved with a deviant partner. Associations between deviant/non-deviant partner involvement and offending risk best fitted a main effects model in which both offending at age 18 and young people's partnership choices at age 21 made independent and additive contributions to the prediction of offending at age 21. There was no interaction between offending at age 18 and partner choice at age 21 in determining offending risk in early adulthood. In addition, the effects of deviant/non-deviant partner involvement on patterns of offending were the same for men and women, and were found to persist even after extensive control for the confounding effects of a wide range of selection factors measured during childhood and adolescence. Conclusion: These results highlight the importance of partnership choices during early adulthood in determining young people's risk of offending as adults.