Project related to Research & Statistics for manager decision making

Camille Victoria Delgado

Monroe College Table of Contents

Part 1 2 Introduction 2 Background of the Study 2 Purpose of the Study 3 Definition of Key Terms 4 Theoretical Framework 5 Research Questions and Hypothesis 7 Summary 8 Part 2 8 Review of Literature 8 Related Research 9 Part 3 11 Research Methodology 11 Data 12 Part 4 14 Results 14 Table 1: Frequency Distribution 14 Tables 2-4: Percentage Distribution Tables 15 Tables 5: Descriptive Statistics 17 Table 6: Regression Statistics 18 Tables 7-9: ANOVA 19 Part 5 20 Conclusion 20 Table 10: Class Total Percentage Distribution 22 References 23

Following the November 2016 Elections, a nation-wide census was published that measured the demographics of the voting population (U.S. Citizens above the age of 18). According to the United States Census Bureau, of a survey population of 224,059,000 United States (“US”) Citizens, 157,596,000 were registered, but 53,860,000 did not vote in the November 2016 elections. This is 24% of the US Citizen Population who did not vote. Considering the nature of a Democratic government, where the power is held by the people, it is significant that 24% of the voting population did not have a say in the election of their leader.

The US President is decided by the Electoral College, not the Popular Vote, and these non-voting numbers are suggestive of deeper issues. The Electoral College in 2016 consisted of 538 members, and the leading presidential candidate[footnoteRef:1]—in this case, either Donald Trump or Hillary Clinton—needed 270 electoral votes to win the presidency. Except for Maine and Nebraska, all other states have a “winner-take-all” (Shugerman, 2016) standpoint on electoral votes, which means that the most popular candidate in that state receives all the electoral votes of that state. In November 2016, Donald Trump won 304 of the Electoral College votes. Considering the ultimate outcome of the election, where Hillary received 65,844,954 (48.2%) of the popular vote and Trump only received 62,979,879 (46.1%) (Krieg, 2016), this system—as the primary electoral system in a nation that praises itself for being democratic—is flawed. The validity of the US Electoral College, however, is not the topic of this study. It is significant that 24% of the US Citizen Population did not vote, because in states where the percent differential between the two candidates was 0.3% (Michigan), 1% (Wisconsin), or 1.2% (Pennsylvania and Florida) (Catanese, 2016), and the electoral votes won by Trump in those states were 17, 10, 21, and 27 (“US Electoral College System,” n.d.) respectively, the 24% of non-voters becomes substantial. [1: To keep this study uncomplicated, the Vice Presidential candidates will not be named as candidates. The Presidential candidates ran with their Vice Presidential candidates, so to save space and a constant list of names, the Vice Presidential candidates are implied.]

Now that the context and importance of the 24% have been established, it stands to follow that there may be some correlation or relationship between certain social or economic factors and this percentage, and if those factors can be addressed, the representational deficit of 24% may be lessened by the next election.

In a world dependent on the world wide web, where information—true, false, or simply misleading—is available to the vast majority of the population, the ability to reason, deduce, and verify is, arguably, one of the most important abilities to have. People are heavily influenced by what they believe, and unfortunately, it is easier to believe something already aligned with one’s views, even if it is not necessarily true.[footnoteRef:2] This is where the value of education becomes a factor. In schools, people are taught analytical skills and how to determine a trustworthy source—“people with higher levels of knowledge may be better able to resist incongruent information” (Flynn, Nyhan, & Reifler, 2017, p. 127). Learning these skills does not equate to one using them, but not developing them at all, whether in school or out, does result in them not being used by those people. Putting this theory simply: having deductive reasoning abilities and knowing how to fact-check may lead to one using those skills when reading the news and informing oneself on the elections. Those who lack those skills altogether cannot, and therefore will not, use them. Not to belittle the self-taught, but most people tend to learn these methods and learn to value this way of internalizing information in school, particularly those attaining higher levels of education—they would have had more time for such reasoning to superimpose itself on their habitual tendencies. [2: There is also the question of morality when discussing viewpoints shared online and accepted, but that is a completely different variable affected by many different factors and would broaden this discussion farther than can be holistically considered in this particular study.]

As formal education does the heavy-lifting of teaching the population these skills, this study investigates a potential correlation between higher education and the decision not to vote. Hopefully, by proving a correlation, an appropriate focus on education can be made. While this does not mean that future results will definitively be in one side’s favor—the Electoral College system itself complicates that question—it could mean that more of the population is represented in the nation’s vote, thus rendering the Electoral College system more representative of a democratic vote.

Candidates: Specifically, the presidential candidates (i.e. the persons nominated for US Presidency, actively running, and in this study, on 40+ states’ ballots). With regards to the November 2016 election, these would be Former US Secretary of State Hillary Clinton, Donald J. Trump, Gary Johnson, and Jill Stein (“Presidency 2016,” n.d.).

Census: “A usually complete enumeration of a population” (“Census,” 2017). The census used in this study is the United States Census Bureau, the principal agency of the US Federal Statistical System, whose mission is “to serve as the leading source of quality data about the [United States of America’s] people and economy” (“About the Bureau,” 2017)

Educational Attainment: The “highest level of education that an individual has completed...distinct from the level of schooling that an individual is [currently] attending” (“Educational Attainment,” n.d.).

Deductive Reasoning: A logical process using which a person can come to a conclusion through an investigation of facts and by reasoning with multiple premises.

Democracy: A type of government in which “the supreme power is vested in the people” and is exercised “through a system of representation usually involving periodically held free elections” (“Democracy,” 2017).

Electoral College: A body of electors who, in the United States of America, elects the President and Vice President.

Higher Education: As High School education is generally considered Secondary Education, this would be a level of education following High School, therefore college or university.

Due to the pervasiveness of social media and the influence of “fake news” (Allcott & Gentzkow, 2017, p. 212), the months—even years—before the November 2016 kept the American public inundated with vast amounts of information, much of it highly biased, and a lot of it simply false. H. Allcott and M. Gentzkow argue that education “should increase people’s ability to discern fact from fiction [and] gives people better tools to counterargue against incongruent information” (2017, p.228). Their theory is can be supported by D.J. Flynn, B. Nyhan, and J. Reifler’s investigation on false beliefs about politics, where they investigate how bias “limits the effectiveness of corrective information about controversial issues and political figures” (2017, p. 127). If this study finds a correlation between educational attainment and voting practices, steps can be made to address the issue of the non-voting population. An analysis of voters’ and non-voters’ levels of educational attainment will allow people to focus on certain aspects of societies, potentially leading to positive structural and systemic change that could result in a more democratic form of elections.

Independent Variable: Educational Attainment. This will be split into the following classes: Less than 9th Grade; 9th - 12th Grade, no diploma; High School Graduate; Some College or Associate Degree; Bachelor’s Degree; Advanced Degree. The independent variable will be used to predict results (e.g. Are those with an advanced degree more or less likely to vote than those with no High School Diploma?), and will affect the dependent variable.

Dependent Variable: Voting Status. This will be measured as Reported Voted or Reported Did Not Vote. The population measured will be US Citizens 18 years and older. The dependent variable is affected by the independent variable, and these are the results that are measured in this study.

Moderating Variable: Graduation Year. Particularly in regards to Higher Education and Advanced Degrees, the point at which a person attained that level of education is not available in this study. While when they received the degree does not affect the fact that a person achieved the same level of study as someone many years their senior or junior, an increased age may lead to less impressionable minds and more fixed political views. A younger graduate would have many more years as adults to form their political beliefs in the context of everything they learned while attaining their higher degrees. A much older graduate would have spent much of their adult lives with the level of study they previously achieved, and would only later be able to utilize the skills learned when pursuing their advanced degrees. A moderating variable affects the strength of the relationship between the independent and dependent variables.

Intervening Variable: Economic Status. The study does not show the economic status of the people surveyed, especially in the period of time when the person was at schooling age. A person’s economic status may be a large factor in determining their level of educational attainment. If the study shows that changes in the educational system should be made in order to promote nationwide voting, economic factors will play a huge role in the assessment. An intervening variable cannot be observed in the experiment, but helps explain links between variables.

This study will test the hypothesis that the more highly-educated individuals are, the more likely they are to vote. The alternate hypothesis is that there is no correlation between the level of educational attainment a person has and their voting practices.

Research Question: Does a person’s level of educational attainment directly affect his or her decision to vote or not to vote?

H01: There is no relationship between a person’s level of educational attainment and his or her voting practices.

Ha1: There is a relationship between a person’s level of educational attainment and his or her voting practices.

Are people who have attained a higher level of educational attainment more or less likely to vote than those at a lower level of educational attainment? What do these results suggest? The purpose of this study is to examine the correlation between education and voting practices, which will then provide more concrete reasoning behind education reform in America. With this information, policymakers and voters will have the tools to make more informed decisions regarding education in their communities and, ultimately, America as a nation.

Using Regression Analytics, Descriptive Statistics, and Frequency Distribution Tables, the study will investigate the existence of a relationship, if any, between increasing levels of Educational Attainment, and the act of voting. The data used will be taken from the US Census Bureau’s collection and nationwide statistics. Depending on the results of the research, certain suggestions can be made that would ideally help reduce the percentage of non-voters and therefore render the system of election more representative of the American population, and therefore more democratic. Hopefully, the study can be of use for those who read it—not just the general public, but particularly those in positions of power who can streamline the process of making structural change.

The essence of Democracy lies in the people’s choice, voice, and vote. The structure of the Electoral College aside,[footnoteRef:3] an individual’s choice, in addition to an individual’s choice to vote, is affected by a host of factors. While these factors include geographic region, economic upbringing, and potentially even biological considerations, the primary factor that this exploration will be limited to is the level of educational attainment. Since the exponential growth of technology and technological communication, the volume of information surrounding the presidential candidates of the November 2016 election has increased, and the veracity of all the information has become progressively more questionable. Multiple studies have shown the effect of biased and questionable news surrounding the elections, the ways in which a voter is likely to take in the news, and the effect of education on both counts. In order to address the educational system of the United States, it must be acknowledged that those in political power tend to control the trajectory of public schools’ success. A “[l]arge-scale redesign of a state’s higher education governance system occurs in a distinctively and decidedly political context” (McLendon, & Ness, 2003, p.69). Considering the fact that political candidates have the power to directly affect the American educational system, understanding the potential relationship between education and voter preferences is imperative in knowing where to begin in the pursuit of equitable elections. [3: See supra Part 1, Background of the Study.]

New technology has shaped the political environment for centuries—newsprint, the radio, television, the World Wide Web, and now, social media. Due to the pervasiveness of social media, content “can be relayed among users with no significant third-party filtering, fact-checking, or editorial judgment” (Allcott & Gentzkow, 2017, p. 211). Hunt Allcott and Matthew Gentzkow’s article, “Social media and fake news in the 2016 election,” explores the current debate that Donald Trump “would not have been elected president were it not for the influence of fake news” (2017, p. 212). D.J. Flynn, Brendan Nyhan, and Jason Reifler (2017) investigate the effect of “directionally motivated reasoning, which limits the effectiveness of corrective information about controversial issues and political figures” (Flynn, Nyhan, & Reifler, 2017, p. 127), on political misperceptions. They evaluate different studies in order to discern factors affecting the prevalence of this type of reasoning, and assess survey strategies measuring such misperceptions. In their article, Flynn, Nyhan, and Reifler argue that “people with higher levels of knowledge may be better able to resist incongruent information and maintain alignment between their factual beliefs and predispositions” (2017, p.136). They claim that education leads to a greater ability to “counterargue incongruent information” (Flynn, Nyhan, & Reifler, 2017, p.136).

Allcott and Gentzkow’s theory responding to the Donald Trump debate comes in four parts: (1) social media can be compared to sources of political news and information; (2) the most widely shared fake news was “heavily tilted in favor of Donald Trump” (Allcott, & Gentzkow, 2017, p.212); (3) the average American may have seen at least one fake news story before the election; and (4) both political sides are more likely to believe “ideologically aligned headlines” (Allcott, & Gentzkow, 2017, p.213). Ultimately, they confirm all of these points but do not provide an assessment of the debate. In their view, fake news was quite pervasive, but there are too many mitigating factors involved to definitively respond one way or another (e.g. a large amount of pro-Trump fake news was viewed by voters already predisposed to elect him, or that they could only measure the amount of stories read and remembered by those in their survey). While Allcott and Gentzkow conclude that there are too many mitigating factors to respond to the debate, they prove that fake news was pervasive as regards the November 2016 election. On a related note, Flynn, Nyhan, and Reifler argue that a higher level of educational attainment helps reconcile incongruent information, like fake news.

In the years preceding this election, voting rates had “downsized” (Kamens, 2009, p.101) democracy, and the worldwide expansion of higher education has resulted in the rise of “social movements, nongovernmental organizations, and other kinds of citizen organizations” (Kamens, 2009, p.99). David Kamens is of the opinion that the “[d]emobilization of the less educated[ created] a rising relative gap in civic participations” (Kamens, 2009, p.100, citing Brody 1978; Crenson and Ginsberg 2004; Nie et al. 1996; Powell 1986; Skocpol 2003; Wuthnow 2002), and concludes that the expansion of higher education has driven out the less educated, and the resulting, radical, political system with “narrower boundaries and lower participation rates” (Kamens, 2009, p. 120) have driven the upper classes from society as well. Edward B. Foley appears to hold a related view in his article, which argues that “American democracy is plagued by excessive partisanship,” and “constitutional law thus far has been incapable of redressing this ill” (2017, p. 655). In his view, the November 2016 proves that “[t]here is no expectation that long-standing geographical subdivisions within the polity will be consistent with both political parties’ equal ability to translate raw popular votes into effective electoral power” (Foley, 2017, p.671). What Kamens fails to realize, however, is that his interpretation—that the lowest and highest classes have been driven out of the voting pool—may have been describing something akin to a bubble. This may have been true at the time he was writing (pre-2009), but the latest elections show that the cognitive gap between the middle and working class hasn’t yet eroded. In fact, the effects of social media and fake news may have contributed to a widening of the gap. The change in voting rates may actually be more of a re-distribution of the voting body, rather than a simple increase or decrease of voters at both ends of the spectrum.

The US Census Bureau published a spreadsheet containing basic voting data of the US population over eighteen (18) years old and older. As secondary data released by a third-party source, these numbers were not personally collected, and therefore, the exact specifications of the project and sample set are not definitively known. The collection required that the individuals were of voting age and were U.S. citizens (“Voting and Registration,” 2017). The entire survey included information on race, sex, age groups, and multiple other social, demographic, and economic factors (“Voting and Registration,” 2017). For the purposes of this study, however, the variables taken from the data will be the level of educational attainment, as the independent variable, and voting status, as the dependent variable. This information was made public in light of the US Census Bureau’s mission to “serve as the leading source of quality data about the [United States of America’s] people and economy” (“About the Bureau,” 2017), and was conducted openly. The data collected by the Bureau is used to determine the state distribution of Congressional seats, to plan new roads and schools, to plan the locations of job training centers, to distribute federal funds[footnoteRef:4] to local, state, and tribal governments annually, and more (“About the Bureau,” 2017). [4: Over $675 billion (“About the Bureau,” 2017).]

In order to best determine a correlation between educational attainment and voting practices, this study will use descriptive statistics, frequency distribution tables, and regression analytics. Descriptive analyses use central tendency, variation, and the shape of a variable to describe data, and utilize the covariance and coefficient of correlation in order to describe the strength of the association between numerical values (Levine, Stephan, & Szabat, 2017, p. 96). Frequency distribution collects and displays values as part of one of a set of numerically ordered classes (Levine, Stephan, & Szabat, 2017, p. 38). Regression analyses help “uncover relationships between variables” through the use of models that express how the value of the dependent variable is affected by the independent variable (Levine, Stephan, & Szabat, 2017, p. 428). These analyses will help synthesize the data provided by the US Census Bureau.

After isolating this study’s independent and dependent variables from the US Census Bureau’s data set, the third-party secondary data collected is as follows:

Using descriptive analyses and regression analyses to process this data, the hypotheses will be tested and the research question, answered:

Research Question: Does a person’s level of educational attainment directly affect his or her decision to vote or not to vote?

H01: There is no relationship between a person’s level of educational attainment and his or her voting practices.

Ha1: There is a relationship between a person’s level of educational attainment and his or her voting practices.

Due to the nature of the research question and hypotheses, the distinctions between genders will not be made, and the opinions of two sexes will not be investigated separately.

This table represents the number of instances of each dependent variable in a set of ordered classes. Each value is only assigned to one class, and “every variable must be contained in one of the class[es]” (Levine, Stephan, & Szabat, 2017, p. 38). In this case, the classes—the independent variable—are representative of the various levels of educational attainment possible in the United States of America: Less than 9th Grade; 9th - 12th Grade, no diploma; High School Graduate; Some College or Associate Degree; Bachelor’s Degree; Advanced Degree. For the purpose of this study, the dependent variable is voting status (i.e. voted, did not vote), so the lack of response will not be considered as closely, but must remain in order for the data to be representative of the sample. According to these results, the majority of voters have attained some college or associate degree, and the majority of non-voters, high school graduate degrees.

Percentage distributions present the percentage of the total for each class, which is actually its proportion, multiplied by 100%. The proportion, also known as relative frequency, is equal to the number of values per class divided by specific total numbers of values. Table 2 represents the proportion percentage of the total number of all values in the population. Table 3 displays the results as a proportion percentage of the total results per class. Table 4 shows the results as a proportion percentage of the total number of instances of each dependent variable.

According to the data, the majority of the population held some college or associate degree and voted (18.88%), and the minority of the data fell at the two extremes: advanced degree holders who did not vote (0.94%), and those who attained less than 9th grade and both voted (0.80%) and did not vote (1.22%).

For those who attained less than 9th grade or 9th to 12th grade, no diploma, the most common response was “Did Not Vote” (48.41% and 45.84% respectively) Adversely, for those with advanced degrees or bachelor’s degrees, the most common response by an almost overwhelming amount was “Voted” (80.27% and 74.19% respectively). The other classes also displayed a majority of voting responses, but not by the same margin that in the highest levels of educational attainment. The percentage of people who chose to vote within each level of educational attainment also increases the higher the level. Those who attained less than 9th grade only had 31.69% of their population who did vote, and the numbers increase—35.35%, 51.55%, 63.31%, 74.19%, and ultimately 80.27% for those with advanced degrees.

Of all the people who did vote, those between holding a high school graduate degree and a bachelor’s degree were the most likely demographic (80.30% of the total voters), while out of the non-voters, the most likely demographic was determinedly a high school graduate (39.67%), followed by someone with some college or associate degree (27.96%). The voters’ least likely demographic to vote were those who attained less than 9th grade (1.30%), while the non-voters’ least likely demographic was almost split between those who attained less than 9th grade (5.07%), and those with advanced degrees (3.9%).

As these all relate to the same data population, it is important to consider the results together. Those who reported voted have the largest range of the three variables (40,508 over 19,267 and 9,256), which means that the vast majority of the population reported voting in the November 2016 election. The count in this data represents the different levels of educational attainment used as the independent variable. The mean of this data does not actually say very much about the nature of the data itself, as it simply represents the average value of votes in six classes. It is important, however, in that it is the central point from which the variance and standard deviation are measured. Variance and standard deviation measure how the values are distributed around the mean, and in this study, they depict how the amount of voters, non-voters, or non-responders in each educational attainment level varies between levels. As the option least chosen by the population (14.58%), the results of those who did not respond to voting had a standard deviation of 3,833.04. In contrast, the results of those who reported not voting had a standard deviation of 7,634.53, while the results of those who reported voting had a standard deviation of 16,688.75.

The Multiple R value is the absolute value of the coefficient of correlation, which describes the relationship between two numerical values (Levine, Stephan, & Szabat, 2017, p. 96). The R Square value is simply the square of the correlation coefficient and represents the percentage of variation caused by the independent variable. The Coefficients of Correlation for those who reported voted, reported did not vote, and had no response to voting are 60.10%, 8.49%, and 24.63% respectively. The R Square values are 36.12%, 0.72%, and 6.06% respectively, and they represent the coefficient of determination, which indicates the extent to which the dependent variable (response to voting) is predictable. The Standard Error here represents the average distance of a variable from the regression line, and similar to the standard deviation, it also displays how the amount of voters, non-voters, or non-responders in each educational attainment level varies between levels. The higher this value is, the larger the spread between variables in the data set. The Observations of each are 6, representing the 6 classes of the independent variable.

The confidence level for this data is 95%. This means that the significance level is 0.05 is the furthest possible distribution from H01. The area within 0.05 (0.025 on each end of the normal distribution curve) represents the area in which the H01 should be rejected. The Significance of F value determines the significance of the regression analysis, and in this case, the Significance of F values are 0.21 (responded voted), 0.87 (responded did not vote), and 0.64 (did not respond). These values are all larger than 0.05. The P-value helps determine the observed level of significance, and looks at the probability of obtaining a test statistic equal to or more extreme than the observed sample value given that H01 is true. If the P-value is less than 0.05, the H01 should be rejected, but if the P-value is greater than or equal to 0.05, H01 should not be rejected. In each of these cases, the P-value is larger than 0.05.

The November 2016 Elections brought a drastic new change to the face of American politics, yet according to the census taken from the US population eligible to vote, 24% reported that they did not have a direct say in the election of their president. 2016 is also in the middle of the social media revolution, when information is relayed with little third-party checking, fact verification, or editorial judgment. The pervasiveness of social media is such that informed decisions made from shared content require a certain personal resistance to incongruent information, and even more importantly, require an awarenesss of “factual beliefs and predispositions” (Flynn, Nyhan, & Reifler, 2017, p. 136). In the debate of whether or not Donald Trump would have been elected without the influence of fake news and social media, Allcott and Gentzkow decided that there are too many mitigating factors to conclusively pick one side, but Flynn. Nyhan, & Ryfler state that education leads to higher levels of knowledge, a value which is key to resisting incongruent information.

This study was conducted to determine whether or not there was a relationship between a person’s level of educatioal attainment and his or her voting practices. Following the regression analyses, the data suggests that the null hypothesis, H01, should be rejected. The Significance of F values for the dependent variable are larger than the 0.05 significance level, and the P-values were also larger than the significance level. These suggest that Ha1, “there is a relationship between a person’s level of educational attainment and his or her voting practices,” should be accepted.

The values of central tendency in the descriptive statistics approach display that in general, more people chose to vote regardless of educational attaiment, and the large standard deviation and variance values displayed a large spread of the data.

Following the exploration of the percentage distribution of the data, it may first seem like the working hypothesis cannot be verified, but after a closer look, the data also suggests that the null hypothesis should be rejected. It appears that the most common demographic of voters was not the highest tier of educational attainment, but the central tier—those between holding a high school graduate degree and a bachelor’s degree. In fact, there were almost as many people who attained advanced degrees who did not vote (2,098), as there were people who only attained less than 9th grade (2,732). An important note, however, is that there were only 5,643 members of the population who had only attained less than 9th grade, while there were 25,056 members of the population who received advanced degrees. When looking at the percentage distribution, the numbers who “reported did not vote” actually represent 48.41% of the less than 9th grade population, but 8.37% of the advanced degree population. In fact, for those who attained 9th to 12th grade, no diploma, 6,746 of 14,715 (45.84%) also “reported did not vote.” Those with bachelor’s degrees only had 5,862 of 46,317 (12.66%) reporting that they did not vote. While the numbers may seem initially misleading, as certain values in the higher classes almost match those in lower classes when looking at responded did not vote—suggesting the same amount of people chose not to vote regardless of educational attainment, or, that there is no relationship between educational attainment and voting practices. When you consider Table 10, and the percentage distribution of the classes themselves, the results are placed into context.

Ultimately, through the use of percentage distribution, descriptive statistics, and regression analysis, the study proved that the null hypothesis should be rejected. As such, it can be seen that there is, in fact, a relationship between educational attainment and voting practices, wherein those with lower levels of educational attainment are more likely not to vote, while those with higher levels of educational attainment are more likely to vote.

About the Bureau. (2017, January 30). United States Census Bureau. Retrieved September 26, 2017, from https://www.census.gov/about/what.html

Allcott, H., & Gentzkow, M. (2017). Social media and fake news in the 2016 election. The Journal of Economic Perspectives, 31(2), 211-235.

Brody, R. 1978. The puzzle of political participation in America. In A. King (Ed.), The new political system. Washington, DC: American Enterprise Institute.

Catanese, D. (2016, November 14). The 10 closest states in the 2016 election. US News. Retrieved on September 28, 2017, from https://www.usnews.com/news/the-run-2016/articles/2016-11-14/the-10-closest-states-in-the-2016-election

Census. (2017). Merriam-Webster.com. Retrieved October 1, 2017, from https://www.merriam-webster.com/dictionary/census

Crenson, M., & Ginsberg, B. (2004). Downsizing democracy. Baltimore, MD: Johns Hopkins University Press.

Democracy. (2017). Merriam-Webster.com. Retrieved October 1, 2017, from https://www.merriam-webster.com/dictionary/democracy

Educational Attainment. (n.d.). United States Census Bureau. Retrieved September 26, 2017, from https://www.census.gov/topics/education/educational-attainment.html

Flynn, D. J., Nyhan, B., & Reifler, J., (2017). The nature and origins of misperceptions: Understanding false and unsupported beliefs about politics. Advances in Political Psychology 38(S1): 127-50.

Foley, E.B. (2017). Due process, fair play, and excessive partisanship: A new principle for judicial review of election laws. The University of Chicago Law Review 84(2), 655-756.

Kamens, D. H. (2009). The expanding polity: Theorizing the links between expanded higher education and the new politics of the post‐1970s. American Journal of Education 116(1), 99-124.

Krieg, G. (2016). It's official: Clinton swamps Trump in popular vote. Central News Network. Retrieved September 26, 2017, from http://www.cnn.com/2016/12/21/politics/ donald-trump-hillary-clinton-popular-vote-final-count/index.html

Levine, D.M., Stephan, D.F., & Szabat, K.A. (2017). Statistics for managers using Microsoft Excel (8th Ed.). Pearson Education, Inc.

McLendon, M.K., & Ness, E.C. (2003). The politics of state higher education governance reform. Peabody Journal of Education 78(4), 66-88.

Nie, N., Junn, J., & Stehlik-Berry, K. (1996). Education and Democratic Citizenship in America. Chicago, IL: University of Chicago Press.

Powell, E. B. (1986). American voter turnout in comparative perspective. American Political Science Review 80(1), 17–45.

Presidency 2016. (n.d.) Politics 1. Retrieved on October 1, 2017, from http://www.politics1.com/p2016.htm

Shugerman, J. (2016). The winner-take-all state rules for the Electoral College are unconstitutional. ShugerBlog. Retrieved September 26, 2017, from https://shugerblog.com/2016/12/05/the-winner-take-all-state-rules-for-the-electoral-college-is-unconstitutional/

Skocpol, T. (2003). Diminished democracy: from membership to management in American civic life. Norman, OK: University of Oklahoma Press.

US Electoral College system. (n.d.). World Atlas. Retrieved on September 26, 2017, from http://www.worldatlas.com/webimage/countrys/namerica/usstates/electorl.htm

Voting and Registration in the Election of November 2016. (2017, May). United States Census Bureau. Retrieved September 26, 2017, from https://www.census.gov/data/tables/time-series/demo/voting-and-registration/p20-580.html

Wuthnow, R. (2002). United States: Bridging the privileged and the marginalized? In R. D. Putnam (Ed.), Democracies in flux. Oxford, UK: Oxford University Press.