Controls and Boundary Reinstatement

Abstract

The imposition of additional constraints to improve network accuracy has often been empirically investigated. One recent example is Stock, Taylor and Hannigan’s (1998) study on rural reinstatement. Yet, the problem of induction remains since no amount of empirical cases can lead one to logically conclude that the imposition of additional control is justified. The purpose of this paper is to derive the theoretical relationship between the imposition of additional control and network accuracy to shed some light on the issue. The main conclusion is that, apart from network design, the imposition of additional control may bias the estimates and, more importantly, accuracy is improved even though the bias exists.