<p><b>Scientific classification, although often underemphasised, has a profound impact on subsequent scientific tasks such as explanation, prediction, and in applied fields, treatment. It also shapes the way that we interact with the world, and guides the kinds of decisions that we make, be they empirical or non-empirical. In correctional psychology, it has been acknowledged that classification practices lack validity and are fraught with conceptual weaknesses. However, there has been little scholarship addressing these weaknesses, offering alternative approaches, or critically considering what would be required to improve correctional classification practices. In this doctoral thesis I begin to address this gap in the literature and argue it is time that classification be taken seriously in psychological research. In doing so, I adopt a theoretical perspective known as epistemic pluralism, and from this lens I develop a novel approach to correctional classification and provide two conceptual frameworks that are necessary for this approach to be implemented in practice. The implications of this novel approach are explored, and it is argued that a theoretical shift from a unified to a pluralistic account of classification provides a more appropriate and profitable approach to classifying correctional phenomena, and helps to address the conceptual problems that existing practices display.</b></p>
<p>Continuing to establish research and practice around conceptually impoverished categories is not a scientifically defensible direction for correctional psychology. Rather, it is time that the complexity of correctional psychology is acknowledged and accounted for in our classificatory work. Failing to do so results in poor scientific progress and treatment programmes/modalities that display, at best, modest outcomes. This thesis provides the methodological and conceptual foundations for advancing correctional classification practices, and ultimately, improving correctional research and practice.</p>