Purpose and method. Crime reduction is a central feature of the current Government Home policy. This incorporates a wide variety of initiatives from crime-prevention strategies to o Vending-behaviour programmes. One key performance indicator for assessing crime reduction is reconviction. This article critically analyses the two main sources of criminal history data held in England and Wales from which reconviction is determined: e rst, the O Venders Index (OI), which is accessible to independent researchers; and second, the National Identie cation Service (NIS), available for limited Home O Yce personnel. Criminal history data were compared for 134 sexual o Venders using both sources of data. Results. Overall, there was a poor correlation for o Vence-based criminal history summaries between the two data sources. Further analysis revealed that neither source appeared more reliable, with each source contributing unique and additional information. Sentencing occasion-based summaries, however, showed a strong relationship between the two sources. Five-year reconviction rates for the sample were highest using a combination OI and NIS data than reconviction rates using one source alone. Conclusions. There is no evidence in this study to suggest that the NIS is more reliable than OI. OI users are urged to use sentencing occasion summaries over oVence-based summaries for control group matching purposes as these data appear to be more consistent between the two sources. In the future, when researchers have access to both sources of data, it is recommended that they use a combination of data sets for assessing past criminal history and deriving risk of reconviction in order to provide a more complete picture. A sound knowledge of the limitations of the data source being used is essential for researchers.