CAUSAL INFERENCE WITH A CONTINUOUS TREATMENT AND OUTCOME: ALTERNATIVE ESTIMATORS FOR PARAMETRIC DOSE-RESPONSE FUNCTIONS WITH APPLICATIONS.

Causal inference with a continuous treatment is a relatively under-explored problem. In this dissertation, we adopt the potential outcomes framework. Potential outcomes are responses that would be seen for a unit under all possible treatments. In an observational study where the treatment is continuous, the potential outcomes are an uncountably infinite set indexed by treatment dose. We parameterize this unobservable set as a linear combination of a finite number of basis functions whose coefficients vary across units. This leads to new techniques for estimating the population average dose-response function (ADRF). Some techniques require a model for the treatment assignment given covariates, some require a model for predicting the potential outcomes from covariates, and some require both. We develop these techniques using a framework of estimating functions, compare them to existing methods for continuous treatments, and simulate their performance in a population where the ADRF is linear and the models for the treatment and/or outcomes may be misspecified. We also extend the comparisons to a data set of lottery winners in Massachusetts. Next, we describe the methods and functions in the R package causaldrf using data from the National Medical Expenditure Survey (NMES) and Infant Health and Development Program (IHDP) as examples. Additionally, we analyze the National Growth and Health Study (NGHS) data set and deal with the issue of missing data. Lastly, we discuss future research goals and possible extensions.

THERE GOES THE ELECTORAL NEIGHBORHOOD: LOCAL NETWORKS, ATTRIBUTION OF RESPONSIBILITY, AND THE (UN) FAIRNESS OF ELECTIONS

During the last two decades there have been but a handful of recorded cases of electoral fraud in Latin America. However, survey research consistently shows that often citizens do not trust the integrity of the electoral process. This dissertation addresses the puzzle by explaining the mismatch between how elections are conducted and how the process is perceived. My theoretical contribution provides a double-folded argument. First, voters’ trust in their community members (“the local experience”) impacts their level of confidence in the electoral process. Since voters often find their peers working at polling stations, negative opinions about them translate into negative opinions about the election. Second, perceptions of unfairness of the system (“the global effect”) negatively impact the way people perceive the transparency of the electoral process. When the political system fails to account for social injustice, citizens lose faith in the mechanism designed to elect representatives -and ultimately a set of policies. The fact that certain groups are systematically disregarded by the system triggers the notion that the electoral process is flawed. This is motivated by either egotropic or sociotropic considerations. To test these hypotheses, I employ a survey conducted in Costa Rica, El Salvador, Honduras, and Guatemala during May/June 2014, which includes a population-based experiment. I show that Voters who trust their peers consistently have higher confidence in the electoral process. Whereas respondents who were primed about social unfairness (treatment) expressed less confidence in the quality of the election. Finally, I find that the local experience is predominant over the global effect. The treatment has a statistically significant effect only for respondents who trust their community. Attribution of responsibility for voters who are skeptics of their peers is clear and simple, leaving no room for a more diffuse mechanism, the unfairness of the political system. Finally, now I extend analysis to the Latin America region. Using data from LAPOP that comprises four waves of surveys in 22 countries, I confirm the influence of the “local experience” and the “global effect” as determinants of the level of confidence in the electoral process.