Farid Qamar

The diffusion of urban policy may occur through multiple mechanisms with complex interactions that obscure the interpretation of independent effects. In this work, we present a case study to explore the potential impact of covariance on the identification of the independent effects of the explanatory variables. We explore the influence of multiple diffusion mechanisms: learning, imitation, and coercion; policy mobility factors; and internal determinants on the likelihood of hyperlocal air quality monitoring program (HAMP) adoption by large cities in the US with a population >300,000 over the past decade. In general, results imply the adoption of HAMPs over the past decade has been motivated by sociopolitical rather than environmental goals. However, comparing the outcomes of a common, but limited, method to detect covariance and those from an expanded systematic variable selection exploration, shows that a third of the variables have covariance effects that can alter their impact and level of significance. The absence of coercion is found to lead to half of the variables exhibiting different results. These results highlight the importance of variable selection methods in determining a variable's significance and impact on the likelihood of policy adoption.