Learning about Learning: A Response to Wand

We welcome the opportunity to respond to Wand's careful and detailed analysis of our paper (Lewis and Schultz 2003). With the discipline's increasing inclination to move toward fully structural strategic choice models (e.g., Signorino 1999; Morton 1999), the issues that Wand raises are important to consider, as they bear on crucial questions of model construction and interpretation. Moreover, his work has allowed us to consider more carefully the properties of an estimator that we are in the process of applying to actual data.