During the past decade research findings have suggested effective ways for probation officers to work with offenders. At the same time, the Criminal Justice Act 1991 changed the sentencing framework, as well as the manner in which probation officers write court reports and supervise offenders. Both of these changes led to an increase in the monitoring and evaluation of probation officers' work, and the appropriate targeting of sentence proposals, as well as take-up rates for proposals, took on greater significance and became more closely monitored than ever before. The present study examines the effects of these changes on probation officers' selection and presentation of community sentence proposals, and it looks at factors that affect court rates of take-up for these proposals, including the offence committed, gender of offender, risk of custody, the presence of social problems and risk of re-offending. The main focus of the study is upon the sentencing choice between community service and probation, which partly reflects the choice between punishment and rehabilitation. Data relating to sentence proposals and outcomes are analysed using cross-tabulation tables, which provide a readily understood method of examining outcomes relative to proposals. These quantitative analyses are combined with three qualitative studies of written reports, which provide insights into reasons for sentencing decisions not apparent from the statistical data alone. The study discusses ways in which probation officers and magistrates may take differing perspectives concerning the main alms of sentencing and hence may disagree over what constitutes the most appropriate sentence. The study concludes that despite the requirements of both effective practice guidelines and the CJA 1991, the probation service has been unsuccessful in convincing magistrates that the probation order is capable of achieving multiple sentencing objectives for more serious offenders.