Statistical tests were used to examine whether an offender's sex appeared to affect the likelihood of a prison sentence once criminal and sentencing history was taken into account. Penalties predicted by the statistical model were then compared to actual sentences men and women received. Women shoplifters were less likely than comparable males to receive a prison sentence and were also more likely to be sentenced to a community penalty or to be discharged. Women and men stood an equal chance of going to prison for a first violent offense. Among repeat offenders, however, women were less likely to receive a custodial sentence. Women first offenders were significantly less likely than equivalent men to receive a prison sentence for a drug offense, but recidivists were equally likely to go to prison. Among first and repeat offenders, women convicted of violent or drug offenses were more likely to be discharged and men were more likely to be fined. Magistrates said they found it hard to compare the way they sentenced women and men because they dealt with women offenders less frequently. Because they regarded women offenders as troubled, magistrates responded with measures to help women lead law-abiding lives rather than punishing them. References and tables