Predictive modelling is commonly used to support decision-making in planning. This paper investigates the use, reliability, and effect of this to better understand the ways in which models influence planning policy, practice, and outcomes. The methods comprise of a literature review, model scoping phase, and a systematic analysis of case law. The paper demonstrates the importance and reach of predictive modelling as a DST in planning and how processes and outcomes are influenced by multiple pressures across the science-policy nexus. These pressures can be classified as scientific practices, political imperatives, and the spaces in-between, such as the practices of communication and translation. First, the analysis identifies a range of issues which may affect the quality of the evidence base relied upon for decision making. These issues largely relate to the substance of models including their scientific components, evaluative techniques and modes of application. Adoption of rigorous scientific method and techniques of model evaluation are identified as strengthening the evidence base, as is clear communication of model limitations to model users and decision makers. Second, consistency, clarity and communication of key terms underpinning a predictive model are vital to effective planning decisions. Third, reduction in qualitative assumptions and connection of model development to statutory mandates and policy contexts may also increase integrity of the evidence and avoid models assuming de facto decision-making power or interfering with the role of decision makers. Fourth, modelled evidence should not be viewed in isolation but in the context of the wider environment, cumulative effects and the intent of the policy settings. The paper provides the foundation for ongoing research with stakeholders who develop, use or analyse predictive models as decision support tools in urban planning. The next stage interrogates more deeply the key themes identified in this interim report, such as model fitness for purpose, reliability, the operation of different logics and hidden power, and the policy interface and translation of predictions.