Local Government Responsiveness and Discrimination Regarding Citizen Diversity

• In Event: Experiments on Political Behavior in Advanced Democracies

Abstract

Local public officials should serve citizens without prejudice or discrimination. But do they? Social science has learned much about discrimination by elected officials, yet few studies have focused on appointed city and county officials, whose actions arguably most directly affect the citizen welfare. In this study, we examine whether government agencies discriminate (according to gender, race or ethnicity, immigration status, religion, and/or socioeconomic status) in communication with their constituencies by conducting an array of correspondence-audit field experiments in which actual citizens contact local bureaucrats to request information on public services in the United States. Benefitting from the variation within confederates in terms of their gender, race or ethnicity, immigration, religious, and socioeconomic background, we match them on covariates and statistically maximize differences on the key treatment dimensions between individuals in pairs. We then randomly assign individuals from the pairs as treatments to government agencies. Our main outcome variable is responsiveness, particularly whether the agency responds, how helpful the response is, and how much time elapses for the response to be received. To analyze the data on how long it takes our confederates to get a response, we use survival analysis techniques (also known as time-to-event analysis). We anticipate that women, racial or ethnic minorities, immigrants, religious minorities, and economically disadvantaged individuals will be subject to more discrimination, manifested by a decreased likelihood in response, poorer quality responses, and increased length of response time.

In conducting our audit experiment, we cooperate with real confederates who need services from their governments. Unlike previous studies that have used aliases, our method alleviates ethical concerns particularly from a utilitarian point of view (Humphreys 2015; Findley, Nielson, and Desposato, 2016), because we avoid using deception and wasting public officials’ time. Moreover, we add to the public good not only with knowledge of a general phenomenon but also by helping actual citizens get answers to their questions about their local governments. In recruiting real citizens with real questions about local government policies, we utilize the Nextdoor online community platform (a popular neighborhood-based community online platform). Using Nextdoor for confederate recruitment also provides methodological benefits since Nextdoor’s zipcode information allows us to utilize naturally divided treatment conditions in our audit experiment. In other words, we are able to take advantage of methodological benefits from the natural field experimental setting. Then, we match these citizens statistically and help them reach out to local officials by email to get their questions answered. We randomly assign which citizens contact which officials and from that can estimate treatment effects regarding discrimination against different classes of citizens – say, from minority racial or ethnic groups or lower income levels. We track civil servants’ responses in cities, counties, and other local governments in the United States.

Authors

• Jiseon Chang, Princeton University

• Batul Mabruk

• Daniel L. Nielson, University of Texas at Austin