Led by Oxera, the project team included University of East Anglia, Taylor Nelson Sofres and the Buildings Research Establishment. Oxera reviewed the available trend data and literature, and talked to government and others involved in energy efficiency promotion to establish the focus areas and to identify hypotheses for what influences householders' decisions to adopt energy efficiency measures. There was, as expected, a lack of behavioural models and data from which to model the effect of policies on the take-up of energy efficiency measures. To fill this gap, we collected information about consumers' knowledge of energy efficiency, preconceptions, preferences, and decision-making patterns, through a pilot survey followed by 1,000 face-to-face interviews using CAPI. The survey collected data on consumers' preferences using a discrete choice experiment to minimise bias. This generated a wealth of intuitive, descriptive statistics, as well as a large dataset that Oxera analysed using advanced statistical models, such as the conditional logit model. The results were fed into a series of models that simulated the take-up of measures over time in response to a range of policy instruments. The most sophisticated of these was a pseudo-agent model tracking the energy efficiency purchasing decisions of 5,000 representative households over a 15-year period. This was coded in Visual Basic in Excel. In addition, we worked to refine a long list of candidate policies, by establishing their coverage of the population and their likely effectiveness in altering decisions by householders and landlords.