|Title||Voting Behavior in Local Elections|
|C1 Background and Explanation of Rationale||
How do citizens decide to vote in local elections? This is an important question in a political climate which partisanship characterizes national politics. Partisan disagreements at the federal level, can often lead to citizens’ interests not being realized. However at the same time, only about 10% of elected officials work at the federal level. If we change our perspective, much of the day- to-day operations of people’s lives are governed through local administrative bodies, run by local officials. To understand how citizens decide to elect such officials is increasingly an urgent and relevant research agenda.
While there are scholars who argue that state politics has become more a function of national politics (Hopkins, 2018), we focus on even more local, sub-state politics. Here, we believe there are grounds to believe that citizen’s voting considerations are not simply a function of federal partisanship. Kuriwaki (2018), using ballot image records from South Carolina, finds people tend to vote less along party lines in super-local electoral positions such as County Council, Judicial, or Sheriff. Moskowtiz (2018) also suggests that local news coverage encourages voters to vote separately from their national party identification.
|C2 What are the hypotheses to be tested?||
I hypothesize that citizens in local elections are voting in an at-random fashion. As such giving citizens information about their candidates’ policy positions will affect how well they think the candidate voted for represents their interests. We are interested in the following subgroups for analysis and expect heterogeneous treatment affects in different combinations of the following groups:
My main variable of interest is the measured sense of regret for vote choice. We ask respondents, “How well do the candidates you voted for today represent your interests?”. Answers are given in 5 options: Perfectly, Well, Neither, Not well, Not at all
Please refer to Pre Analysis Plan to for further detail.
|C3 How will these hypotheses be tested? *||
First, I will look at the cross table with percentages of different responses. I will use a chi- squared test to test the difference in distribution of answers among treatment and control groups in various subgroups.
Second I will look at the bivariate relations between the treatment and outcome. This will be estimated by transforming the outcome into an indicator variable of whether a given category was chosen or not. I will use OLS, logit and probit with robust standard errors to estimate the change in probability that the outcome would be in a certain category k, when given treatment.
Third, I will conduct regression analysis, but with controls. I will use the same OLS, Logit and Probit models. The motivation here is to generate results with smaller standard errors. The controls will be as shown in Table 2 of the Pre Analysis plan.
Finally I will conduct an ordered logit and probit regression, treating the outcomes as ordered categorical variables. Like before, I will look at both bivariate relations and regression with controls. Here, the estimand is the direction of the coefficient(s). I want to know whether the treatment causes a positive or negative change in the predicted probability of moving from one category to another.
|C4 Country||United States|
|C5 Scale (# of Units)||About 1400 responses (non-responses included). Formal numbers wait digitization of results.|
|C6 Was a power analysis conducted prior to data collection?||Yes|
|C7 Has this research received Insitutional Review Board (IRB) or ethics committee approval?||Yes|
|C8 IRB Number||11045|
|C9 Date of IRB Approval||2018-11-05|
|C10 Will the intervention be implemented by the researcher or a third party?||Researchers|
|C11 Did any of the research team receive remuneration from the implementing agency for taking part in this research?||No|
|C12 If relevant, is there an advance agreement with the implementation group that all results can be published?||Yes|
|C13 JEL Classification(s)||not provided by authors|