Behavioural case linkage (BCL) is a procedure that can be used to identify linked crime series, which contain two or more crimes committed by the same person, thereby helping the police to detect and prosecute repeat offenders who are responsible for a disproportionate amount of crime. However, despite the potential benefits of BCL, there are also damaging consequences if crimes are incorrectly linked. Consequently, research has started to test if and how this procedure can work in the most efficient and reliable way. But, the extant literature has a number of important limitations, particularly in terms of (1) generalisability (i.e., there have been few attempts to replicate findings across geographical locations and time periods), (2) ecological validity (i.e., the methodology used to test BCL is not representative of how the procedure is used in practice), and (3) methodology (i.e., there is a lack of research to systematically compare the various methodological/statistical approaches to BCL). The primary aim of this thesis was to address these three important limitations. In terms of generalisability, this thesis has tested the extent to which previous BCL research on residential burglary, commercial robbery, and car theft can be replicated in new geographical locations and time periods. In terms of ecological validity, a number of new methodologies have been developed and tested that reduce the gap between research and practice in BCL by allowing both non-serial and unsolved offences (as well as solved, serial offences) to be included when testing the principles of BCL, and also for these principles to be tested with crime series that contain several different types of offence. In terms of methodology, novel methodological approaches have been compared with the ‘traditional’, status quo methodology for researching the BCL principles, thereby ensuring that the findings reported in this thesis can be compared with previous work. This thesis, therefore, has important implications for theory, research, and practice and the findings are discussed in the context of these. Future research directions are also outlined.