Registration of colour as a trademark in Qualitex v Jacobsen 1995 paved the way for other companies to register colours they believe have a distinctive association with their brand. Consumer survey evidence is often used to demonstrate distinctiveness, but frequently receives little weight due to criticisms of the methodology. In particular, the questions used and their ability to provide insights into the legal question of interest are heavily criticised. In terms of colour trademark applications, multiple methods have been employed to provide evidence of colour brand associations, however, to date there are no standard measures. The overall objective of this study was to test various methods of estimating colour-brand associations with the focus of assessing the similarity in the estimates and developing a robust methodology. This study tested four methods of estimating brand-colour distinctiveness in the chocolate product category, where Cadbury has recently sought to register the colour purple, and the rice category where Effem Foods had unsuccessfully applied to register orange in relation to Uncle Ben's rice. The research in this thesis replicated and extended a study conducted in the chocolate category in 2003. The first method explored associations with brands, while the second used a sequence of increasingly specific questions to explore the same associations. The third method required respondents to match a range of shades with brands and the fourth method employed a choice modelling experiment designed to examine respondents' choice behaviour and the interaction between brands and colours. Overall, it was found that all methods revealed a strong association between Cadbury and purple, although Cadbury was also associated with other colours. The findings also suggested that orange was not inherently associated with Uncle Ben's. The key implication that arises from these findings is that where a high level of association is detected the methods had strong convergent validity. However, the choice modelling method is recommended as the more robust methodology as it focused on the effect of colour-brand interactions on behaviour rather than cognitive constructs.