Highlights of findings include: (1) evidence to indicate that once adjustments are made for inappropriate counting of crimes or detections, significant differences in force level conviction and detection rates do exist; (2) victims were found to withdraw in just under 4 in 10 of criminal cases with a high proportion of withdrawals taking place without an arrest being made, while only 10 percent resulted in a suspect being charged; (3) 16 variables (excluding police force area) were found to be significant predictors of whether a case got to court or resulted in a conviction; and (4) evidence to support the claim that the combined impact of National Crime Recording Standard (NCRS) and changes to the detection counting rules have between them accounted for most of the reduction in recorded detection rates since the late 1990s. This study attempted to build upon a growing body of research around attrition in the investigation of rape offenses. The study was specifically designed to examine why detection and conviction rates varied markedly across different force areas in England and Wales and explored the nature of changing detection rates over time. The study also examined the factors closely associated with the detection and conviction of rape cases. The study applied logistic regression techniques to identify those factors which appeared to be central in determining the success of a case getting to court and those factors that led to a victim withdrawing from the process. Tables, figures, appendixes A-C, and references