This study used propensity score matching techniques to examine whether long good behaviour bonds and suspended sentences are more effective than short bonds and suspended sentences in reducing re-offending. Using data from all appearances finalised in the NSW Local Court between 2006 and 2008, rates of re-offending among offenders who received a bond of 24 months or more were compared with rates of re-offending among a matched group of offenders who received a bond of less than 24 months. Re-offending outcomes for offenders who received a suspended sentence of 12 months or more were also compared with a matched group of offenders who received a suspended sentence of less than 12 months. After adjusting for other factors, the probability of reconviction and the time to reconviction were found to be lower for offenders placed on bonds of 24 months or longer compared with offenders placed on shorter bonds. A significant effect of bond length on reoffending was apparent for both supervised and unsupervised orders. The probability of reconviction and the time to reconviction were also lower for offenders given long (12 month plus) suspended sentences compared with offenders given short suspended sentences, after adjusting for other factors. No significant effect of suspended sentence length was found when supervised and unsupervised suspended sentences were analysed separately. These results suggest that long bonds and long suspended sentences are more effective in reducing re-offending than short bonds and short suspended sentences.