The EPIC Survey relies on a stated preference empirical approach to data collection. As opposed to revealed preference approaches, which make use of data on observed behaviours, stated preference approaches use data gathered by asking individuals to either report their actual behaviour, or report how they would behave in a given hypothetical situation. There are advantages and drawbacks to both revealed and stated preference approaches (OECD, 2018[1]). Although revealed preference approaches have high reliability and validity because they reflect the real-world constraints faced by individuals, this also constitutes a limitation insofar as analyses are limited to addressing only those choices and conditions that are available in real-world contexts. The main challenges of stated preference approaches, on the other hand, include response bias and sample representativeness.
Generally speaking, limitations of analyses based on survey data arise from the extent to which reported responses may differ from actual behaviours (i.e. hypothetical bias) as well as the extent to which the characteristics of survey respondents may diverge from those of the actual population. Hypothetical bias is a well-known issue in stated preference methods and a number of strategies have been employed to mitigate it, including informing respondents that their responses will be used to help develop public policies, and informing them about hypothetical bias and encouraging them to reflect on their choices carefully in light of this tendency. A number of ex-ante and ex-post strategies pertaining to survey design and statistical methods, respectively, can also be used to mitigate other biases (e.g. anchoring, order effects).
Despite the challenges identified above, stated preference approaches offer a number of significant advantages over revealed preference approaches when it comes to ex-ante policy evaluation (OECD, 2018[1]). Discrete choice experiments, for example, are well-suited to analysing choice in the context of relatively complex, multi-dimensional issues (Bateman et al., 2002[2]; OECD, 2018[1]). Flexibility to define decision scenarios allows for an evaluation of the impact of hypothetical policy interventions. Stated preference approaches also generate valuations of changes in health status and environmental quality that provide critical input into cost-benefit analyses.
Development of the survey was guided by a Steering Committee comprised of WPIEEP Delegates who provided input regarding policy issues of interest and relevant contextual considerations in their respective countries. A Scientific Advisory Committee of methodological and thematic area experts provided input regarding methodological best practice in survey design in light of the analytical objectives of the work. Finally an internal coordination group involving the IEA, ITF, and TAD also provided policy and technical feedback during the development of the survey instrument.