Date of Award


Document Type


Degree Name



Quantitative Research Methods

First Advisor

Antonio Olmos


Bayesian methods, critical elections, critical realignment theory


Prior to this study, critical realignment theory, which presupposes eras of substantial and sustained swings in American political party dominance, had only been evaluated using the classical, frequentist approach to modeling. However, potential for more information concerning these electoral phenomena exists given a shift in the design and approach to realigning elections. This study sought to explore those options through one particular alternative to the classical approach to statistics--in this particular case, the Bayesian approach to statistics. Bayesian methods differ from the frequentist approach in three main ways: the treatment of probability, the treatment of parameters, and the treatment of prior information. This study sought to understand the effect of these differences as it applied to critical realignment theory: namely, what contribution is made in understanding the occurrence of these eras from each statistical approach? Does the Bayesian approach provide any improvements over the classical approach in terms of understanding critical realignment theory? This first set of research questions was asked from a political viewpoint, but a second set of research questions was also posed from a methodological viewpoint: What methods exist to formally compare these two statistical approaches, and what is the relative strength of each method? Using the most efficient method of comparison, is any further information gained concerning critical realignment theory, and is any further information gained concerning each statistical approach? Using multiple linear regression, results were similar across approaches. For the presidential data, critical elections were found in 1860 and 1932. This was replicated in the congressional models, with one additional realigning election found in 1996. As for additional information gained, Bayesian methods aided in understanding in some ways, but the classical approach also retained some benefit. Furthermore, these two statistical approaches were formally compared to one another, highlighting the comparison between credible intervals and confidence intervals. While these intervals are traditionally considered counterparts, this is not a direct comparison. These intervals represent different concepts, relating to underlying differences in the statistical approach. This, however, reiterates the strong role of correct interpretation as it pertains to results.


Recieved from ProQuest

Rights holder

Tara A. Rhodes

File size

237 p.

File format





Statistics, Political science