In the United Kingdom, the planning process requires applicants to submit an air quality impact assessment wherever an impact on national limit compliance is likely, and this factors into the resultant decision. We identify flaws in the current methodological frameworks and policies associated with this process that in the worst cases could lead to poor decision making. We give examples of how inaccurate data is certified as good through unsuitable pre-processing, how these errors are then amplified by poor modelling practice, and how the final data is judged against metrics that are evidence impaired to arrive at potentially unsound decisions. We then discuss the implications and propose a way forward.