This paper presents the multilevel model approach to analyzing contingent valuation surveys of individuals’ willingness to pay for reductions in the level of air pollution.
It is likely that individuals living in the same area are exposed to the same level of air pollution, and accordingly these individuals’ valuations of a reduction may be correlated. Thus, the data have a hierarchical structure with individuals clustered within regions, and this structure violates the general assumption of independence among observations. Multilevel models allow for this type of data structure. In this paper we analyze individuals’ stated willingness to pay in an open-ended contingent valuation survey for a reduction in the local level of air pollution in Sweden. The results suggest that most variations are among individuals. However, our results indicate that there are also variations at higher levels, which may be explained by homogeneous preferences for a reduction in air pollution among individuals living in the same household or region with a similar level of air pollution.
