Increases in land value, known as value uplift, follow from improvements in accessibility arising from new transport infrastructure. This paper investigates how different approaches in evaluating property value uplift could lead to different results. The case study of the light rail transit system in the Gold Coast, Queensland, Australia, is the context for this study. This paper addresses two of the critical aspects of the value uplift literature—how many of the results are determined by the method adopted and, perhaps more importantly, if there is a method which provides the best results. Historical data on private property sales are used to evaluate the timing, shape, and conditions for increases in land value or value uplift employing two of the most widely used modeling approaches: difference-in-differences and multilevel regression models. The more recent literature has also identified that the choice of control area is pertinent and so this study uses two different approaches for the selection of catchment areas: conventional distance-based methods and propensity score matching. The model results do confirm the increases in property prices because of better accessibility to Gold Coast light rail transit, but the amount of uplift does appear to depend both on the model approach and the method to select catchment and control areas. The paper discusses the implications of this for literature which has a variety of different methods established for research.