Three essays on political polarization in the United States
Abstract
Is the United States polarized? In order to address this question, this dissertation explores three dimensions of political polarization. Three related, but independent, essays on political polarization provide information on questions concerning polarization, First, the studies on polarization themselves appear to be polarized. Chapter 2 focused on reviewing past studies on polarization. Still, one side argues that American citizens are severely polarized, while the other side argues that polarization is an illusion (Abramowitz and Saunders 2008; Abramowitz 2010; Fiorina 2014, 2017; Mason 2016). A sample is carefully chosen from the sociological and political science SCIMAGO rankings. Since almost all measures of polarization used in earlier works can be classified as a single type of polarization, all measures of polarization are coded into four types: issue consistency, issue divergence, affective polarization, and perceived polarization. Based on the systematic review and meta-analysis in Chapter 2, no strong evidence supports the idea of a "polarized America." One interesting finding is that the majority of the studies that produced significant results used ANES data, with a secondary group using the GSS. While there are many differences between the two surveys, the key distinction is the timing of the field surveys. Second, Chapter 3 begins with the premise that political interests will be deeply related to political action, and so will political polarization. For example, recent studies present evidence that political interest can vary depending on the political context (Prior and Bougher 2018). Political interest also ebbs and flows with politically salient events, especially federal elections. In Chapter 3, the focus is whether polarization is a stable characteristic of the electorate, like political interest or partisanship. Previous scholarship emphasizes the "situational" characteristics of interest for change and adaptation to a new environment as opposed to the "static" characteristics of interest (Featherman et al. 1994; Prior and Bougher 2018). Therefore, the research question centers on political polarization, as under the same presumption, the degree of political polarization within the general public is influenced by political interest and electoral circumstance. Using data from two nationally representative surveys, the ANES and the GSS, the level of issue polarization increases similarly in both sets of data, but the patterns of sorting are different. The level of sorting in the ANES fluctuates, whereas instability shows a more dramatic increase in the GSS. In contrast, the overall level of sorting in the ANES is higher than in the GSS. Thus, the difference in the level of polarization between two surveys comes from the timing of the surveys. While the ANES is typically conducted from August to December in the midst of the election campaign, the GSS is typically conducted from February to May. Since the GSS and the ANES surveys do not conduct surveys in non-election years, it is difficult to isolate the effect of elections on the level of polarization. The Pew Research Center conducts annual political polls. Using Pew political surveys, I investigated the difference in the level of polarization between election years and non-election years and between presidential elections and midterm elections. The findings demonstrate that the timing of a survey has a significant effect on the level of polarization (i.e., sorting) in general. This finding suggests that elections are a key determinant of the intensity of the level of polarization, partisan strength, and political views. The level of polarization, like measures of political interest and party affiliation, is not stable. There are fluctuations in the level of polarization that are associated with points in time within the electoral cycle, and citizens are more likely to show more polarized attitudes as a function of the proximity of an election. Chapter 4 focused on the level of polarization at the state level. Studies of polarization have received more attention at the national level due to the difficulty of collecting enough samples. The CCES provides enough samples of between 30,000 and 50,000 individuals every year to make it possible to examine the state level of polarization. In particular, Chapter 4 investigates the relationship between state swing in a presidential election and state polarization. The fourth chapter examined the potential relationship between the phenomenon of state swing or state competitiveness in presidential elections and the state's level of polarization. By concentrating on sorting, the connection between polarization at the state level and swing states, including those that are competitive in presidential elections, is analyzed. Swing states should have more polarization than safe states like California and Texas. Using an independent sample t-test, the degree of polarization between swing states and safe states is significantly different. The effects of polarization on the potential for state swing, or the likelihood that the state will be competitive in presidential elections, are examined. Sorting has a sizable magnitude and significant impact on the probability that a state will change its support in presidential elections, using a logistic model. States are more likely to change their support in presidential elections as the degree of polarization within those states rises or falls. The impact of partisan composition within a state is one of the chapter's more intriguing findings. The findings show that a more balanced partisan composition within a state increases the likelihood of switching support from one party candidate to another in presidential elections and makes the state's presidential elections more competitive. In addition, possible connections might exist between a swing state and a person's degree of polarization. Using the same four categories of swing states, there might be differences in which groups of citizens are more likely to be polarized than others, depending on the state in which they live and the circumstances surrounding the state election. Based on an OLS model, mixed results were obtained regarding the prediction of the impact of the state's electoral circumstances (competition, TV ad spending, and campaign events) and the swing experience on the degree of polarization. The three essays on political polarization in the United States suggest several implications. First, based on meta-analysis in chapters 2 and 3, I concluded that Americans are still not polarized as Fiorina (2018) recently contended. While some types of polarization are, arguably, on the rise and pervasive in the mass public, scholarly findings still do not reach a consensus. It is too early to conclude that the United States is polarized. One of the most challenging aspects of previous and recent research on polarization is that it is almost entirely based on survey data collected in the midst of presidential or midterm elections (i.e., the ANES, the GSS). The campaign season accentuates partisan polarization for the average citizen in a way other times do not. The findings in Chapter 3 follow this notion and suggest that the level of polarization among the mass public can be influenced by external factors such as electoral circumstances. Future research should evaluate the reasons behind each type of polarization and how the political circumstances and the survey instrument may influence measures of political polarization. Also, it should be examined how these conditions affect different political outcomes. Second, scholars should be careful when evaluating polarization at the sub-national level, as shown in Chapter 4. It is easy to conflate the concepts of geographical polarization and state polarization. There are two distinct ideas here. This dissertation shed some light on the distinction between state-level and geographic polarization. It also contributed to clarifying the difference between the state-level concept of polarization and the national one. In conclusion, polarization is a complex concept that demands careful discussion. This dissertation presents several original perspectives and ideas that will be helpful for future studies on political polarization in the United States and other democratic societies.
Degree
Ph. D.