What would you do? An investigation of stated-response data

When analysing choices or policy impacts, economists generally rely on what people actually do, rather than what they say they would do. The "stated response" approach is treated with scepticism due, for example, to concerns regarding the effect of strategic or social considerations on what people say, and a belief that people may not adequately consider such a hypothetical question. This paper evaluates an example of this approach; the direct questioning of parents as to whether they would withdraw their children from school if the Familias en Accion education subsidies were withdrawn. Our results suggest that these concerns are not entirely invalid but that the stated responses do provide important information and correlate in the expected manner with child and household characteristics. We conclude by emphasising the importance of good question design, which may allow researchers to use the "stated response" method as a complement to more typical quantitative methodologies.