摘要：Quasi-experimental techniques were developed to provide decision-making tools for documenting the impacts of developmental highways in rural areas. Regression discontinuity analysis (RDA) with limited observations was used to compare economic changes in highway counties to those in adjacent and non-adjacent control counties. The RDA models found statistically significant changes in population, per capita income, and taxable sales related to highway development. The study found that some counties benefited from developmental highways, some were unchanged, while some experienced economic decline. RDA models with adjacent controls had better explanatory powers while those with non-adjacent controls were more sensitive to highway-related changes in economic activity. When significant non-highway activities were present, adjacent control models may have understated highway-related impacts, while non-adjacent control models may have overstated these impacts. Arguments for using adjacent and non-adjacent experimental designs are discussed.
