ABSTRACT
In Chapter 2, Ted McConnell lays out the theory behind combining behavioral data from various sources. In simplest form, the practical task of drawing inferences from joint stated and revealed preference models boils down to estimating appropriate models for the data collection efforts at hand while imposing consistency in the underlying preference structure. In this chapter, we explore some of the practical speed bumps that arise when different types of preference data are combined in a single estimation attempt. Herein, we assume a basic familiarity with standard models for non-market valuation and maximum likelihood estimation, at least to the extent of estimating preprogrammed models from econometric packages such as SAS and LIMDEP. 1
