Temporal dynamics of public perceptions toward water infrastructure in US shrinking cities before and after the Flint Water Crisis



Journal Title

Journal ISSN

Volume Title



Understanding public perceptions allows decision makers to assess public priorities for actions to pursue sustainable infrastructure management. However, since public views change based on new information or events, the cross-sectional sample from previous studies does not account for the temporal dynamics of perceptions. In 2014, Flint, Michigan switched water sources resulting in lead leaching into drinking water and impacts to public health. This study evaluated public perceptions toward water infrastructure providers at two different time periods before (November 2013) and after (June 2016) the Flint Water Crisis, using surveys deployed to residents in 21 US shrinking cities. Two questions of interest from these surveys were statistically modeled to assess the temporal dynamics of public perceptions toward local water providers, specifically: (1) whether individuals trust his/her water provider to make decisions in his/her best interest (Trust) and (2) if individuals want to partake in participatory processes with local utilities (Active). Random parameter binary probit models were the resulting best-fit models, and used to identify demographic (e.g., age, gender, income) and geographic parameters (cities) influencing Trust and Active. A likelihood ratio test was conducted to evaluate the independence of two data sets to determine if the public perceptions have changed from 2013 to 2016, which resulted in a 99% confidence that the perceptions should be modeled separately. Shrinking cities comprise the survey sample, a classification of cities that are typically fiscally constrained, experiences infrastructure underutilization, and to which Flint (the location of the water crisis) belongs. This study illustrates the limitations of cross-sectional surveys regarding infrastructure perception in light of new events or information. Additionally, the statistical modeling found that most geographic parameters had homogeneous impacts (i.e. minimal statistical variability across respondents), while demographic parameters had heterogeneous impacts (i.e. statistically varies across respondents) on Trust and Active. The homogeneous impact of geographic parameters demonstrates the localized significance of utility-customer relationships in shaping perceptions.