Abstract

Context: Previous research has shown that Americans with disabilities turn out to vote at significantly lower levels than people without disabilities, even after accounting for demographic and other situational factors related to political involvement. The authors examined the potential mechanisms underlying their low turnout. They asked whether people with disabilities exhibit participatory attitudes and behaviors at levels commensurate with their other individual-level characteristics.

Methods: The present study conducted descriptive and predictive analyses on data from the 2012 and 2016 American National Election Studies.

Findings: Despite low levels of turnout in recent elections, people with disabilities were just as participatory, if not more so, when considering alternative forms of political engagement. The authors' analyses indicate that, while disability status had no bearing on political efficacy or partisan strength, those with disabilities reported being even more interested in politics than those without disabilities. Evidence is provided that depressed turnout rates among those with disabilities may be due in part to lower levels of attentiveness to the news, political knowledge, and negative perceptions of government.

Conclusions: The psychological impacts and behavioral consequences that emerge from possessing a disability and the broader role of disability in the American political context are multifaceted. This area of research would benefit from future studies that examine a variety of electoral contexts.

In April 2018 Tammy Duckworth, Democratic senator for the state of Illinois, made headlines as she cast a vote on the Senate floor while holding her 10-day-old baby, Maile (Stolberg 2018). Not only was Ms. Duckworth the first senator to give birth during her tenure in office, but she was also the first disabled woman in the Senate, after losing both legs in Iraq (Biography.com2018). Although Senator Duckworth occupies a highly visible position, she represents just one of millions in the United States living with a disability. In fact, the 2010 Census estimates that 56.7 million Americans, or 18.7% of the noninstitutionalized population, live with a disability (Brault 2012). This population is important for the study and practice of American politics, particularly as the number of people with disabilities continues to rise due to aging (National Institutes of Health 2010). Due to population growth, the political status of people with disabilities is in flux not only in the United States but also in other countries around the world. The World Health Organization estimates that about 15% of the global population lives with a disability, making them members of the “world's largest minority” (United Nations 2017).

We examined what scholars know about the voting habits of those with disabilities and assessed whether these patterns are in line with other forms of political engagement. Specifically, we asked whether attitudinal measures can be leveraged to shed light on the behavioral outcomes of individuals with disabilities. The goal of this research was multifaceted, with results contributing to the study of American democracy, political behavior, minority political incorporation, and identity politics. Theoretically, by offering more evidence about the engagement of the large, diverse population of persons with disabilities, this research further refines theories of political behavior. Empirically, this project tested how disability relates to a variety of traditional predictors of behavior and preferences in the United States. The broader scope of this research speaks to the role of health in facilitating or impeding political participation.

Prior Literature

Existing research paints a clear, if not bleak, portrait of political participation among Americans with disabilities. Scholars have documented sizable voter turnout gaps between people with and without disabilities. Schur and Kruse (2000), focusing on voters with spinal cord injuries in New Jersey, found that people with disabilities were 10% less likely to vote in 1992. In an analysis of data from the 1998 midterm election, Schur et al. (2002) found a 20% voter turnout gap between those with and without disabilities. Using data from the Currently Population Survey (CPS), Hall and Alvarez (2012) report that, relative to people without a disability, people with disabilities were 7% less likely to vote in 2008 and 3% less likely in 2010. Also drawing on data from the CPS and other surveys, including the 2006 General Social Survey (GSS), Schur and Adya (2013) and Schur and Kruse (2014) corroborated these findings. Schur, Ameri, and Adya (2017) added to this line of research by presenting a 5.7% voting disparity in 2012, again noting less political involvement among those with disabilities.

When examining eight different types of political activities, Schur, Shields, and Schriner (2005) again found low participation among people with disabilities. However, as with previous studies, the authors omitted key explanatory factors. Particularly, exposure to political news, strength of partisanship, and strength of ideology all tend to be stronger and better developed with age, factors that were not given much consideration. More current scholarship on the topic from Schur and Adya (2013) examined multiple data sets to analyze the participatory practices of Americans with disabilities, concluding people with disabilities to be significantly less likely to vote. They also showed that differences between people with and without a disability diminish once controlling for education (Schur and Adya 2013). However, as in Schur, Shields, and Schriner (2005), the Schur and Adya (2013) study omitted news exposure, partisanship, and ideology as key control variables. In their report to the US Election Assistance Commission, Hall and Alvarez (2012) also provided recent data showing people with disabilities were less likely than those without a disability to participate in politics, though regression analyses were not reported.

How do behavior scholars make sense of these findings? Motivational factors of political participation have always been of interest to political scientists. With Downs's (1957) paradox of voting, Riker and Ordeshook's (1968) addition of the “D term,” and Campbell et al.'s (1960) funnel of causality, social scientists have seemingly always sought to explore the internal origins of external behavior. Yet the root cause of the turnout disparity between individuals with and without disabilities remains largely elusive to disability scholars. Earlier research identified self-sufficiency as a mechanism behind the lower voting rates of people with disabilities, emphasizing the types of employment and mobility barriers that people with disabilities face (Schur and Kruse 2000). Potential policy solutions that might help with self-sufficiency problems are ones that increase the employment of people with disabilities. This is because “along with enhancing economic self-sufficiency and social integration, employment may also help this important segment of the population become more active citizens” (Schur and Kruse 2000: 586). Subsequent research from Schur et al. (2002) pointed toward a similar direction and, related to self-sufficiency, also called for more research into how “major life-transitions” affect people with disabilities differently than people without disabilities. Indeed, more recent work from Haselswerdt and Michener (2018) indicates that large-scale changes in health insurance policy, particularly the loss of one's insurance coverage, have a negative impact on political involvement.

In addition to suggesting ways to increase self-sufficiency via increases in the employment rates of those with disabilities, scholars have also focused on election administration solutions. Schur et al. (2002) revealed that voter turnout might be depressed by actual and expected problems with polling place accessibility. In response to problems with voting technology in 2000, the Help America Vote Act was passed in 2002 to update voting machines. Notably, the Help America Vote Act contains a number of provisions relating to polling place accessibility for people with disabilities, though in practice it needs more rigorous enforcement (Schur and Adya 2013). Independent of voting accessibility on Election Day, other research has recommended better options for people with disabilities in terms of convenience voting, convenience registration, and ballot simplification (Hall and Alvarez 2012; Miller and Powell 2016; Schur and Adya 2013; Schur and Kruse 2014). These measures are particularly important because, as Schur, Ameri, and Adya (2017) found, when those with disabilities experience voting difficulties they are more likely to perceive group stigmatization or to hold negative perceptions about their group's political influence, which in turn affects willingness to vote. Although accessibility and administrative issues are outside the scope of the current project, they certainly may play a role in shaping political efficacy and/or attitudes toward government more broadly.

Current Project

The present project drew on data from the 2012 and 2016 American National Election Studies (ANES). Analysis of the 2012 ANES showed that people without disabilities reported voting at a rate of 81%, while people with a disability reported a voter turnout rate of 69%, thus resulting in a 12% voter turnout gap for that year. The 2016 ANES reveals a turnout gap of 12% again, with those reporting disabilities and those not reporting disabilities voting at rates of 87% and 75%, respectively.

While self-reported voter turnout rates are likely exaggerated (Harbaugh 1996), a wealth of aforementioned empirical evidence in the voting behavior literature consistently suggests people with disabilities to be a group with untapped political potential. What are the causal mechanisms that links having a disability to lower levels of voter turnout? To what extent do those with disabilities engage with politics, beyond measures of voting? Conclusive answers to these questions remain difficult to obtain. Here we explore whether, commensurate with the pattern of low voter turnout, people with disabilities also exhibit low levels of political involvement across a variety of attitudinal and engagement measures. Our hypotheses propose several paths of influence.

Hypotheses

Psychological Resource Hypothesis

Ojeda (2015) argues that, beyond the traditional resource model—time, money, and civic skills (Brady, Verba, and Schlozman 1995)—political participation necessitates both physical and mental exertion. That is to say, the political resources needed to successfully participate in a representative democracy are just as much (if not more so) physical and psychological as they are material. Following political events requires a significant amount of attentiveness and psychological entrenchment with the topic, a task that may not be easily attainable for those dealing with ill health on a daily basis. Living with a physical disability is taxing on one's mental well-being, much in the same way that living with a mental disability is taxing on one's physical health. Research has shown physical and mental health to have reciprocal effects on each other (Lenze et al. 2001; Schreurs, de Ridder, and Bensing 2002). To the extent that health-related concerns take precedence in day-to-day life, we should expect individuals with disabilities to report lower levels of political attentiveness, political interest, and political knowledge than individuals without a disability.

Political Conviction Hypothesis

A wealth of empirical research has demonstrated the powerful relationship between strength of political conviction and political engagement. In this project, we conceived of political conviction as strength of political ideology and strength of partisanship. Stronger ideologues, or those who report a strong conservative or strong liberal leaning, are not only more likely to vote (Palfrey and Poole 1987) but are also more likely to participate in a variety of political activities (Converse 1964) than are their more moderate counterparts. The authors of The American Voter Revisited (Lewis-Beck et al. 2008: 207) noted that “ideology summarizes a person's overall stance toward the political world. . . . An ideology can also give political meaning to an enormous variety of observations, events, and experiences that fall outside the immediate realm of politics.” Similarly, canonical models of voting have argued stronger partisans to experience greater levels of political engagement than those with weaker or moderate partisanship inclinations (Campbell et al. 1960). In fact, more recent studies have found that partisanship not only asserts influence on participation but also structures one's political identity (Bartels 2000; Green, Palmquist, and Schickler 2004; Huddy, Mason, and Aarøe 2015).

We investigated whether these relationships hold for individuals with disabilities. Our political conviction hypothesis expects that individuals with disabilities have lower documented levels of political involvement due to weaker (i.e., more moderate) political convictions. Living with a disability and identifying as a disabled individual introduce a unique multidimensionality to one's sense of self and therefore to one's political convictions.1 If ideology is a summary judgment (Lewis-Beck et al. 2008), then ideological strength is the degree of confidence in that judgment. As a sense of self or social identity, disability status might bolster or attenuate one's political convictions depending on the salience of those conditions and how they fit with one's overall worldview. For example, a disabled individual who generally supports small government ideals might view providing government benefits for those with documented chronic illness, visual impairments, or limited mobility as an exception. In short, disability status presents another layer of one's social identity, a layer that may complicate political decision making. Such crosscutting identities among those with disabilities, we believe, could induce conflicting policy or candidate preferences that in turn lend themselves toward moderate political convictions (Treier and Hillygus 2009).

Political Efficacy Hypothesis

Political behavior scholars are well familiar with the influential nature of political efficacy. Campbell, Gurin, and Miller (1954: 187) identified efficacy as “the feeling that individual political action does have, or can have, an impact upon the political process, i.e. that it is worthwhile to perform one's civic duties. It is the feeling that the individual citizen can play a part in bringing about change.” Political efficacy then evolved as conceptually two-dimensional: internal and external. Lane (1959: 149) referred to internal efficacy as “the image of the self as effective” and external efficacy as “the image of democratic government as responsive to the people.” Scholars have since presented empirical evidence supporting a two-dimensional notion of efficacy (Converse 1972; Coleman and Davis 1976; Balch 1974). The more politically efficacious individuals feel, the more likely they are to engage with politics, perhaps because efficacy conveys a sense of personal control.

Relatedly, Schur, Shields, and Schriner (2003) found that people with disabilities have significantly lower levels of both external and internal political efficacy compared to those without disabilities. In line with this and other previous findings (Schur 1998; Papadopoulos, Montgomery, and Chronopoulou 2013), we anticipated that people with disabilities—which are generally not bestowed on individuals by choice—to feel a minimized sense of control. One's experience with disability status, therefore, is expected to manifest in lower levels of both internal and external political efficacy.

Perceptions of Government Hypothesis

While lower levels of external political efficacy might indeed indicate more negative perceptions of government and public officials, we analyzed each as separate constructs. As stated, we expected political interest, attentiveness, and political knowledge to be lower among those with disabilities than those without disabilities. We expected that this disinterest, or “exiting of the system,” may be precipitated by (a) negative experiences with government and/or (b) a lack of perceived governmental representation. Throughout the process of securing legal disability status or filing for disability benefits, those with disabilities may have direct experiences with the unpleasantness of governmental red tape. Alternatively, those with disabilities might simply feel like government officials do not represent them or their interests. That is to say, elements of both descriptive representation, whereby elected officials possess physical traits similar to their constituency, and substantive representation, whereby elected officials pursue interests pertinent to their constituency, may be perceived as missing among individuals with disabilities (Wright 2016). Consider, for example, recent work by Ojeda and Slaughter (2018) demonstrating that the negative relationship between depression and turnout is attenuated in the presence of a coethnic representative, particularly for black men.

Although we are not able to examine specific causal pathways, prior research has noted lower levels of trust in government and more cynical assessments of government performance among those with disabilities, especially in the area of managing unemployment (Schur and Adya 2013). In our study, we expected individuals living with disabilities to report lower levels of presidential and congressional approval and to report higher levels of perceived government corruption than individuals living without disabilities.

Sources of Data and Measures

The data for this project come from the 2012 and 2016 ANES, two nationally representative, cross-sectional data sets. In this article we report a series of quantitative analyses in which disability served as the primary independent variable of interest. In line with our four hypotheses, key dependent variables included news attentiveness, political knowledge, interest in politics, strength of ideology, strength of partisanship, political efficacy, and perceptions of government. Further, the data were drawn from both the pre- and postelection waves and were weighted by the full sample weight. Note that all of our analyses were restricted to US citizens and those 18 years of age or older. Please see  appendix A for complete question wording and coding of all variables.

Measuring Disability

With regard to our independent variable, there is no consensus on how to define and quantify who has a disability, even among those who have made it their life's work. However, in her seminal work The Disabled State (1984), Stone noted that pressures for expanding the concept of disability have come for years from the citizens who seek aid, the workers who make eligibility decisions, and the policy makers who set standards related to disability programs. Legal definitions of disability in the United States vary by state and also between state law and federal law. Further, international organizations, such as the United Nations, have constructed yet additional ways of defining the population of people with disabilities. Laws that require a definition of the population of people with disabilities share one important commonality: the definition presents disability as a binary concept. That is, either one has a disability, and is perhaps eligible for benefits under the law, or one does not have a disability and thus does not have access to such benefits.

In attempts to measure the population of people with disabilities, surveys present a multitude of indicators of disability, both objective and subjective. To measure the population eligible for reasonable accommodation under the 1990 Americans with Disabilities Act, for example, one could ask whether or not the respondent has a record of a “physical or mental impairment that substantially limits one or more major life activities” (2019: n.p.). Other survey measures gauge disability by asking questions related to one's employment status or by inquiring how many days per month mental or physical disabilities interfered with one's routine activities. More detailed surveys may even incorporate questions that allow respondents to specify disability by type and severity. The GSS, for example, did this in a specialized module in 2006, and the CPS also regularly asks about types of disabilities. However, the GSS has not implemented the same module since 2006, and attitudinal measures related to politics in the CPS are limited.

In line with previous research (Miller and Powell 2016; Schur, Ameri, and Adya 2017), in this study we used a binary operationalization of disability as our main independent variable. In both surveys we used respondents' preelection current employment status to gauge disability. Those who indicated being “permanently disabled” were coded as 1, and all other employment statuses were coded as 0.2 Using this operationalization, we found nearly 7% of the 2012 ANES sample (n = 394) and roughly 4% of the 2016 ANES (n = 182) to have a disability.

While standard in the field of disability research, identifying disability according to employment status is conceivably fraught with measurement error. To gain some leverage on this potential for error and to differentiate the effects of employment from disability status, we included three additional comparison groups in our analyses: (a) employed persons, (b) retired persons, and (c) other unemployed persons. Employed individuals were those who indicated their employment status as “working now” (n = 3,095 in 2012, n = 2,547 in 2016). Retired individuals indicated their current employment status as “retired” (n = 1,315 in 2012, n = 922 in 2016). Other unemployed individuals were those who selected “unemployed,” “temporarily laid off,” “homemaker,” or “student” as their current employment status (n = 1,097 in 2012, n = 604 in 2016).

Dependent Variables

To explore the psychological resource hypothesis, measures of political interest, news attentiveness, and political knowledge were required. Political interest asks individuals how often they pay attention to what is going on in government and politics, with responses ranging from never to always. Although question wording and response options varied somewhat between the 2012 and 2016 ANES, our measures of news attentiveness generally gauged the extent to which respondents reported following politics/campaigns through four primary mediums: Internet, television, radio, and newspapers. Political knowledge was assessed using an index of seven questions in the 2012 data set and four questions in the 2016 data set related to both current and general knowledge about US government. Our political conviction hypothesis was tested based on a folded, four-point ideological scale ranging from moderate to extremely liberal/conservative and a folded 4-point partisanship scale ranging from independent to strong Democrat/Republican (see  appendix A).

To investigate our political efficacy hypothesis, we included internal efficacy, which measures the extent to which one understands what is going on in government, and external efficacy, which measures the extent to which one feels that they can affect government. In 2012, the ANES offered four different question wordings of both internal and external political efficacy (each asked to half the sample); the 2016 ANES efficacy measures were composed of just two questions (one for internal and one for external) asked of the entire sample. Finally, we relied on three measures to evaluate the perceptions of government hypothesis: approval of Obama, approval of Congress, and evaluations of government corruption. Questions related to approval asked how each is handling his or her job, whereas measures of government corruption asked respondents to assess how many people in government are corrupt, ranging from none to all (see  appendix A).

Demographic and Political Controls

Standard demographic controls were employed within all of our inferential models: age, gender, marital status, race, ethnicity, religiosity, education, income, and geographic location (the South). Age is widely known to be an important predictor of political behavior (Verba and Nie 1972; Rosenstone and Hansen 1993), and indeed, disability status specifically addresses some of the reasons that an elderly person may be less likely to participate. Schur and Adya (2013) found age to be a significant predictor of participation in their study of the political participation levels of people with disabilities. Research has shown, as with age, one's gender influences one's relationship with politics. With regard to political efficacy in particular, Verba, Burns, and Schlozman (1997) showed men to display significantly higher levels, and Schur, Shields, and Schriner (2003) observed the same in their study of people with disabilities. Relatedly, research shows that marital status influences political engagement: married people tend to be more interested in politics relative to unmarried people, and married people tend to engage with politics together (Verba, Burns, and Schlozman 1997; Wolfinger and Rosenstone 1980; Leighley and Nagler 2014). As a multitude of empirical research will attest, political deliberation and participation largely take place among those with higher socioeconomic status (Schattschneider 1960; Verba and Nie 1972; Wolfinger and Rosenstone 1980; Rosenstone and Hansen 1993; Verba, Schlozman, and Brady 1995; Schlozman, Verba, and Brady 2012; Leighley and Nagler 2014). Due to this effect, and particularly due to our operationalization of disability according to employment status, we included household income and education as controls.

To account for race and ethnicity, we included two binary variables: whether or not the respondent identified as black, and whether or not the respondent identified as Hispanic. Previous research has shown that, on the whole, blacks tend to exhibit stable and cohesive levels of political conviction, generally identifying as Democrats and liberals (Tate 1993, 2010; Black 2004). Hispanic political behavior research uncovers a similar pattern: many Hispanic Americans identify as Democrats and liberals and tend to vote along these lines (Lopez and Taylor 2012). Relatedly, we included religiosity as a control variable (a) because of the well-documented connection between it and race/ethnicity (Cox and Jones 2017) and (b) because religiosity is also a predictor of political engagement (Smith and Walker 2012).

An additional demographic item we controlled for was whether or not the respondent lived in the South (see  appendix A for coding). We included this variable due to the South's history of voting discrimination, once necessitating special coverage under the Voting Rights Act (Overton 2006; Hasen 2012; Wang 2012); the more recent removal of such coverage (Shelby County v. Holder, 570 U.S. 529, 2013) (Blacksher and Guinier 2014); and empirical evidence demonstrating the relationship among southern states, voter identification laws, and decreased levels of voter turnout (Hajnal, Lajevardi, and Nielson 2017). Additionally, we controlled for the South due to its relationship with disability status itself: 46% of permanently disabled individuals reported living in the South in the 2012 ANES, and 28% in the 2016 ANES.

As described earlier, party identification and ideological orientation are paramount in predicting an individual's political involvement and their perceptions of government. Thus, all model specifications accounted for these two variables. Other politically relevant controls in our models included strength of partisan identification, strength of ideological orientation, political interest, and political knowledge. See  appendix A for question wording and coding details.

Results

Our results first begin with the replication of extant patterns of turnout and political engagement among those with and without disabilities. As previously noted, turnout among individuals with disabilities was rather dismal in both the 2012 and 2016 general elections (fig. 1). Disabled individuals were 11% less likely to vote than employed persons in both years, but only about 3% and 7% less likely to vote than other unemployed individuals (i.e., nonemployed, nondisabled) in 2012 and 2016, respectively. With regard to voting and campaign-related activities beyond voting (e.g., attending a political rally, donating to a candidate) (fig. 2), retired persons demonstrated the most political engagement of any employment group. While patterns of campaign participation among people with disabilities were relatively on par with employed and other unemployed individuals (fig. 2), average engagement among disabled individuals slightly surpassed these groups in 2016.

Table 1 provides two specifications—demographics only and demographics plus political variables—for models predicting voter turnout and campaign participation. Dummy variables for disability status, retired, and other unemployed were included in each model. Various model specifications in table 1 allow us to gauge the degree to which significant effects for disability status were altered when accounting for different types of variables, such as education, age, political interest, and partisan strength. When controlling only for demographic factors, disability status was not a statistically significant predictor of turnout, though in 2016 disabled individuals were significantly more likely than employed individuals (the excluded category) to participate in other ways. The same can be said of other unemployed individuals in both 2012 and 2016. While the addition of demographic factors largely had no effect on retired persons' inclination to vote or participate, we did observe that the positive, statistically significant effect on voting in 2012 for these individuals disappears in our 2016 models. In other words, retired persons were no more or less likely to vote in 2016 than employed persons.

In contrast with demographic only models, fully controlled analyses (i.e., demographic and political models in table 1) indicated statistically less turnout in 2012 among disabled individuals (p = 0.019). While the same relationship remains negative in 2016, the effect is only marginal (p = 0.099). With regard to campaign-related activities, however, data from the 2016 ANES show those with disabilities were significantly more likely than employed persons to become engaged with campaign-related activities beyond turnout (p = 0.037); all else held constant. Interestingly, while retired persons remained more participatory than employed persons in both general elections, the positive and statistically significant relationship between this group and turnout dissipated as political controls were factored into the model (p = 0.081 in 2012, p = 0.492 in 2016). Likewise, those who were unemployed were no more or less likely to vote, relative to employed individuals, though this group did indicate engaging in other types of electoral behavior (p = 0.005 in 2012, p = 0.017 in 2016).

We aimed to identify changes in the magnitude and direction of disability status on turnout and participation, given various model specifications outlined in table 1. Yet we believed it was most fruitful and, indeed, most accurate to analyze how disability status “works” in the context of both demographic and political variables. That is to say, we wished to explore the direct effect of disability status on political attitudes and behaviors, all else being equal. Therefore, we emphasized fully specified models as we evaluated each of our hypotheses.

Psychological Resources

Our psychological resource hypothesis contends that the gap in voter turnout among people with and without disabilities might be explained by lower levels of attentiveness to, interest in, and knowledge about politics. The results are presented in figure 3 and table 2. Figure 3 displays the main effect of disability status on each dependent variable, controlling for retired and other unemployed dummy variables, as well as individual differences in demographic and political factors. Across news attentiveness, political knowledge, and political interest, disability status had a more muted effect in the 2016 election than in 2012. We also observed that the effect size and directionality of disability on political knowledge changed dramatically between 2012 and 2016. Unstandardized coefficients in table 2, as well as the additional comparison categories, put these changes into perspective.

In general, those with disabilities were less attentive to political news than were those who were employed (p = 0.000 in 2012, p = 0.295 in 2016), despite showing increased levels of political interest (p = 0.001 in 2012, p = 0.094 in 2016). While coefficient directionality of disability status on news and interest remained unchanged across ANES years, results for other categories of employment showed volatility (table 2). For instance, retired individuals were significantly less likely than employed individuals to be attentive to political news in 2012 (p = 0.000), while we found no such effect for 2016. Compared with those who were presently working, individuals identifying themselves as disabled showed less knowledge of politics in 2012 (p = 0.001). Curiously, in 2016 all employment categories but one failed to present a statistically significant effect on attentiveness, knowledge, or interest. Based on these results we might speculate that in this election year employment status of any type was largely unrelated to one's psychological involvement with the election, its candidates, and with politics more broadly.

In total, results of the tests of our psychological resource hypothesis are somewhat muddied. There is evidence suggesting that the voting gap between disabled and nondisabled individuals rests on lower levels of news attentiveness and, at least in 2012, lower levels of political knowledge. However, disability status was consistently predictive of higher levels of political interest, which may indeed factor into this group's involvement in campaign activities aside from voting. We added to these results by next turning to our expectations of disability status and political efficacy.

Political Efficacy

Results of our analyses of our political efficacy hypothesis are shown in table 3. (An accompanying figure containing the effect size of disability on efficacy is not presented due to lack of statistical significance across all models.) Our political efficacy hypothesis expected disabled individuals to experience lower levels of both internal and external political efficacy, which may explain dampened turnout rates. Despite failing to reach statistical significance, patterns of unstandardized regression coefficients indicate higher levels of internal efficacy among disabled persons, compared to employed persons, in both 2012 and 2016 (p = 0.437, p = 0.562, respectively). With regard to external efficacy (e.g., that government is responsive one's preferences), having a disability exhibited a positive relationship in 2012 (p = 0.725) and a negative relationship in 2016 (p = 0.389). As indicated in table 3, other unemployed individuals reported feeling significantly more internally efficacious (e.g., that one understands and is well qualified to participate in politics) than employed individuals in 2016 (p = 0.011). On the whole, however, these results suggest that political efficacy does not vary greatly between those with and without disabilities—or between any employment category, for that matter. Self-assessments of internal and external efficacy do not appear to be a primary motivator of the voting gap.

Political Conviction

Our political conviction hypothesis suggests that people with disabilities may experience lower levels of political involvement due to weak ideological orientations and weak partisan affiliations. Figure 4 shows the main effect of disability status on partisan and ideological strength across 2012 and 2016 ANES respondents. Coefficient estimates from these fully specified models are shown in table 4. The effect of disability on partisan strength is minimal, though its effect on ideological strength is substantial and subject to fluctuation (fig. 4). Such opposing findings might not be altogether surprising, given that traditional theories of political attitudes (Converse 1964) assert that partisan affiliation and ideological orientation are not equivalent constructs. Compared with employed individuals, those who had a disability and those who were otherwise unemployed were generally less partisan in 2012 and 2016. On the other hand, despite minimal effects on strength of partisanship, retired individuals (p = 0.245 in 2012, p = 0.036 in 2016) and other unemployed individuals (p = 0.002 in 2012, p = 0.018 in 2016) showed stronger ideological convictions than did employed individuals.

Of considerable importance is that disability status relates to strong liberal or conservative attitudes in 2012 and weak liberal or conservative attitudes in 2016 (fig. 4, table 4). Retired and other unemployed categories do not experience such volatility across years. Compared with those presently working, retired and other unemployed individuals consistently exhibit strong ideological orientations. A breakdown of ideological strength by disability status showed that approximately 40% of disabled persons reported moderate leanings, 24% reported weak leanings, 26% reported liberal or conservative, and 10% reported strong leanings within both 2012 and 2016 ANES data. While these raw percentages remained relatively unmoved from year to year, other demographic and political controls within our models exhibited volatility (see appendix table B1). Specifically, controls for female, married, black, and education—all statistically significant predictors of ideological strength in 2012 models—were no longer significant in 2016 (p = 0.158, p = 0.746, p = 0.820, and p = 0.454, respectively). Such dramatic changes in the predictive power of our control variables may have some bearing on the opposing coefficient estimates observed in figure 4.3 Indeed, when bivariate analyses of disability status on ideological strength are performed (see appendix table B2), coefficients in both 2012 and 2016 models revealed a clear negative relationship, though only 2016 results were significant (p = 0.027).

In sum, when demographic and political variables are unaccounted for, individuals with disabilities appear to report weaker or more moderate ideological orientations. Other explanations for our findings also likely lie within the dynamics of the 2012 and 2016 campaigns and candidates themselves, discussed next. The results of our political conviction hypothesis, therefore, suggest that partisan strength was altogether not a significant component in explaining low turnout rates among those with disabilities, though ideological strength may matter depending on the electoral context.

Perceptions of Government

Finally, we considered the relationship between disability status and negative perceptions of government. The perceptions of government hypothesis expected persons with disabilities to report lower approval ratings of government and greater perceptions of government corruption. In general, ANES data showed that views of government between 2012 and 2016 grew increasingly pessimistic among those with disabilities. While in 2012 disability status was related to greater approval of Obama (p = 0.110) and Congress (p = 0.030) and fewer perceptions of government corruption (p = 0.038), in 2016 approval ratings fell and perceptions of corruption increased (fig. 5). In 2016, individuals with disabilities were significantly less approving of Obama (p = 0.028) and significantly more likely to view government as corrupt (p = 0.011) than those currently working. Comparisons across other unemployment categories did not yield such instability (table 5). Though the results were not statistically significant, retired individuals tended to approve of Obama in both 2012 and 2016, while other unemployed individuals tended to disapprove of Obama during this time. Likewise, other unemployed individuals were generally more disapproving of the way Congress handled its job in 2012 and 2016, compared with the excluded category of employed persons.

When it comes to making sense of the perceptions of government hypothesis, our results indicate that the effect of disability on all three dependent variables changed direction and statistical significance across years (table 5). As in table 4, estimates of disability status on these three variables were heavily influenced by (volatility among) accompanying controls. With regard to perceptions of government corruption, for instance, blacks were less likely to view government as corrupt in 2012 (p = 0.009), though no relationship existed between these two variables in 2016 (p = 0.983) (see appendix table B3). Similarly, in 2012 higher-income individuals were less likely to view government as corrupt (p = 0.000), a finding that did not replicate in 2016 (p = 0.448).

As with our examination of ideological strength, bivariate analyses are somewhat useful in clarifying the relationship between disability and perceptions of government (see appendix table B4). Disability status predicted perceptions of greater government corruption in both 2012 and 2016, though this effect was significant only in the latter (p = 0.686 in 2012, p = 0.000 in 2016). Again, we suspect that fluctuations in the predictive power of traditional demographic and control variables may be due in part to the nature of the 2016 general election. Taken as a whole, the perceptions of government hypothesis is not cleanly supported. The 2012 and 2016 ANES data clearly demonstrate increasingly negative sentiments toward government and governmental actors among those with disabilities. Still, given tenuous findings across years and model specifications, we remain hesitant to conclude that perceptions of high corruption and low governmental approval rates are historically responsible for the turnout gap among those with disabilities.

Discussion

Our findings diverge somewhat from previous research on disability status and political behavior. Although it appears that traditional forms of engagement (i.e., voting) by and large disenfranchise those with disabilities, this group is just as likely to display political signage, to donate money to candidates/campaigns, to advocate for political causes, and so forth. When it comes to political engagement in 2016, our fully controlled results indicate that individuals with disabilities were notably more involved than other employment groups. While participation in rallies and other political volunteer work remains low among persons with disabilities, 48% of all disabled individuals surveyed in 2016 reported talking to another person about voting for or against a party/candidate, up from 40% in 2012. This statistic may tie in with the fact that disabled individuals were more politically interested (especially in 2012) than employed persons, despite tuning out of traditional political news sources (e.g., television, newspapers, radio).

Based on these results, we might further speculate that disability status influences not only the frequency with which one attends to political news but also the medium by which that information is sought. For example, physical or cognitive limitations may make it easier to get one's news from a television screen rather than a handheld newspaper. Likewise, individuals who are employed might be more likely to get their news from radio, particularly as they commute to work. Cursory results from the present ANES data provide some support for this conjecture. People with disabilities were less likely than those without disabilities to gather political news via newspaper, internet, or radio. We believe that, given varying news quality across mediums and outlets, coupled with lower levels of political knowledge among those reporting disabilities, subsequent empirical study of the media habits of those with disabilities should be a priority.

In contrast to Schur, Shields, and Schriner (2003), we found no differences in external or internal efficacy among those with and without disabilities. Schur, Shields, and Schriner (2003: 121) hypothesized that “people with disabilities may have lower levels of political efficacy because of discrimination, prejudice, and negative social constructions. They may perceive themselves as less able to perform various politically relevant skills . . . and they may believe that they have less influence in politics and do not receive equal treatment from public officials.” Within the 2012 and 2016 ANES datasets there are no measures of perceived discrimination in relation to one's disability status, though we believe this theoretical rationale is a valuable starting point for future study. Specifically, perceptions of discrimination may play into the construction and maintenance of one's social identity (Johnstone 2004) and identity politics more broadly. Still, our findings here may serve as a positive indication that efficacy levels of people with disabilities have improved overall since 1998, when Schur, Shields, and Schriner (2003) collected their data.

Ideological strength is traditionally a powerful determinant of voting and thus a key factor in our analysis of low turnout rates among individuals with disabilities. With the exception of respondents reporting a disability, employment status exhibits consistent effects on ideological strength. Moreover, we found the correlation between partisanship and ideological orientation to be particularly low for those with a disability (r = 0.32 in 2012, r = 0.57 in 2016). Individuals with disabilities may not have crystallized attitudes toward politics and/or their own political identity, a finding that we believe relates to lower levels of political knowledge and attentiveness to political news, particularly in 2012.

Given inconsistent support for our political conviction hypothesis, future research must consider whether ideology and partisanship are constructed similarly for individuals with and without disabilities. Pertinent to modeling antecedents of the voting gap, we might also explore the extent to which disability status takes precedence over partisan or ideological affiliation as one considers their individual identity. We assert that the salience of one's disability plays a large part in shaping ideological preferences and thus the relationships between disability, identity, and political involvement. Still, it is imperative to note that in models predicting participation, as well as several of our other dependent variables, we observed that those reporting a disability were not like those who were either retired or unemployed. That is to say, there is genuinely some aspect of possessing a disability, rather than possessing “free time,” that exerts an influence on political attitudes. The ability to identify such differences between these employment groups is a benefit of the present research and an advancement to the study of disability and political behavior more broadly.

Although partisanship was not a primary dependent variable in the current set of analyses, we did unearth several results related to disability, engagement, and partisan affiliation. Individuals with disabilities largely tend to identify as Democrats, though the degree to which they vote along party lines shows some instability. In the ANES data, of those disabled persons who cast a ballot in 2012 (n = 256), 72% voted for Obama, 22% voted for Romney, and 3% voted for a third-party candidate; of those who voted in 2016 (n = 101), 51% voted for Clinton, 30% voted for Trump, and 7% voted for a third-party candidate. Additional research confirms that physical and mental health impact not only if one votes but also for whom one votes. For example, in the 2016 presidential election, counties with poorer public health were significantly more likely to shift their vote in favor of Donald Trump, relative to Mitt Romney in 2012 (Wasfy, Stewart, and Bhambhani 2017). On one hand, such shifting loyalties help position our somewhat inconsistent results, particularly with regard to ideological strength. On the other hand, the nature of Donald Trump's 2016 candidacy as both a celebrity and a political outsider adds potential noise to our findings. How election contexts and other exogenous factors moderate partisan affiliations, and therefore behavioral outcomes, among those with disabilities is an avenue ripe for empirical study.

Despite disabled individuals' tendency to identify as Democrats, our results with regard to the perceptions of government hypothesis indicate that this group was significantly less approving of Obama than employed persons in 2016. This may in part reflect the tenuous relationship between partisan strength and disability status. It is also possible that increased perceptions of government corruption bled over to or were conflated with attitudes toward Obama and Congress. One might also speculate that those with disabilities hoped the Obama administration would do a better job of representing disability rights, starting in 2008. At the outset, tenets of the 2010 Affordable Care Act (ACA) seemed promising to those with disabilities, specifically policies that bridged coverage for those with preexisting conditions (Collins et al. 2012). It is quite possible, however, that benefits afforded to disabled individuals were overlooked or “submerged” (Chattopadhyay 2018) by negative public sentiment, as private insurers left the ACA marketplace (Khazan 2017) and states continued to withdraw from ACA Medicaid expansion (Young 2017). In all, the unique nature of the 2016 election makes it difficult to draw definitive conclusions about partisan preferences and disability. Future research might explore when and under what circumstances (e.g., midterm elections, diversity of candidates) partisanship exerts influence on the voting decisions of those with and without disabilities.

For scholars researching disability, operationalization of this construct has proved difficult and subject to constant criticism (Burkhauser, Houtenville, and Tennant 2014). As such, the way that we are able to measure disability in this project provides for several limitations and/or caveats of our results. Within large-sample datasets, like the ANES, disability has typically been measured according to employment status. While this operationalization potentially conflates physical or mental limitations with employment status, our present use of employed, retired, and other unemployed (e.g., laid off, homemakers, students) individuals as comparison groups seeks to disentangle the true nature of disability status on political behavior and attitudes. Still, it is quite possible that respondents who fall into one category are not precluded from another, muddying the raw effect of disability status.4

We also noted compositional differences between disabled persons surveyed in the 2012 ANES and in the 2016 ANES. We found that in 2012 nearly 7% of the ANES sample indicated living with a disability (n = 394), but in 2016 only 4% indicated the same (n = 182). Additionally, a breakdown of demographic factors (appendix table B5) indicated that these samples contain disabled individuals with varying characteristics. Notably, blacks and Hispanics made up 30% and 18%, respectively, of the 2012 sample, whereas they made up 10% and 20%, respectively, of the 2016 sample. We also observed that a larger percentage of disabled individuals reported living in the South in the 2012 ANES sample (46%) than in the 2016 ANES sample (28%). Average income of disabled respondents, which hovered around $20,000/year, remained fairly stable across ANES samples, as did average age (approximately 50–54 years) and average education level (high school graduate or high school graduate plus some college). While these demographic differences should be considered in the present analysis, we express more broadly a concern regarding discrepancies between the 4–7% of survey respondents who report disability status and the estimated 18% of Americans who live with a disability (Brault 2012). Disabled populations are often hard to reach and may require additional assistance to participate in survey research (e.g., transcription services, assisted listening devices), a circumstance we encourage disability researchers and survey methodologists alike to contemplate.

Disability is similar conceptually to the notion of pan-ethnicity in race and ethnic studies, meaning not all disabilities are the same. To be sure, there is additional variance within the disabled population that cannot be addressed here. For instance, people with disabilities also differ from one another in important ways, one of which is time since the onset of disability and the severity of disability. One limitation of the current operationalization is that all disabilities were aggregated into one category. Though it matters indisputably for predicting political outcomes, we are not able to examine the effects of different types of disabilities, such as paraplegia, multiple sclerosis, or schizophrenia. We should expect each particular disability, as well as its onset and severity, to influence (a) the extent to which one identifies as a person with a disability, (b) the effect of this identification on political attitudes and behaviors, and (c) individuals' ability to report such attitudes and behaviors within surveys. Consider, for example, that surveys are not generally able to access the sample of people with the most severe of disabilities, as they may be institutionalized in health facilities or incarcerated. Federal efforts toward better measures are ongoing, particularly as light is shed on disparities between people with disabilities and those without them (Brucker and Houtenville 2015).

Despite these limitations, shared by all disability scholars, the analytical approach presented here advances the study of people with disabilities. Beyond democratic ideals of inclusiveness, those with disabilities should be particularly encouraged to engage within the political sphere, as this type of activity contains potential healing properties. Bergstresser, Brown, and Colesante (2013) found that participation in politics is an important recovery tool for those suffering from mental illness, as engagement imparts a sense of empowerment and feelings of social connectedness. As we demonstrate here, the results of this project present implications for political inclusion, for partisan coalition building, for disability representation and policy in government, and for subsequent electoral outcomes.

For the greater part of a century, scholars of American political behavior have given precedence to individual-level demographics such as education, income, and race as predictors of political engagement. While these factors certainly remain useful in explaining gaps in attentiveness and turnout, we implore scholars to consider a more holistic approach. We, along with a handful of contemporary researchers (Ojeda 2015; Pacheco and Fletcher 2015; Burden et al. 2016; Schur, Ameri, and Adya 2017, Ojeda and Slaughter 2018), are beginning to make strides in incorporating physical and mental health into the conversation on political outcomes. It goes without saying that both the American electorate and American political institutions are composed of living beings, with varying levels of physical mobility and cognitive functioning. With this in mind, we deem it essential that the foundation for political inquiry begin with considerations of physical and mental health.

Notes

1.

For more on narrative construction and social identification among individuals with disabilities, see Galvin 2005.

2.

The 2012 ANES allows for the selection of multiple employment categories (e.g., laid off and homemaker, student and working now). Our operationalization categorizes employment status by respondent's first mention.

3.

One might also suspect that ideology itself disproportionately influences coefficient estimates in these models. No changes, either in coefficient direction or statistical significance, were observed within our results when ideology was removed as a control variable.

4.

Indeed, in the 2012 ANES nine respondents indicated being permanently disabled (first mention) and currently working (second mention).

References

References
Americans with Disabilities Act National Network Website
.
1990
. “
What Is the Definition of Disability under the ADA?
adata.org/faq/what-definition-disability-under-ada.
Balch
George I.
1974
. “
Multiple Indicators in Survey Research: The Concept ‘Sense of Political Efficacy.’
Political Methodology
1
, no.
2
:
1
43
.
Bartels
Larry M.
2000
. “
Partisanship and Voting Behavior
.”
American Journal of Political Science
.
44
, no.
1
:
35
50
.
Bergstresser
Sara M.
,
Brown
Isaac S.
, and
Colesante
Amy
.
2013
. “
Political Engagement as an Element of Social Recovery: A Qualitative Study
.”
Psychiatric Services
64
, no.
8
:
819
21
.
.
2018
. “
Tammy Duckworth
.”
April
10
. www.biography.com/people/tammy-duckworth-21129571.
Black
Merle
.
2004
. “
The Transformation of the Southern Democratic Party
.”
Journal of Politics
66
, no.
4
:
1001
17
.
Blacksher
James
, and
Guinier
Lani
.
2014
. “
Free at Last: Rejecting Equal Sovereignty and Restoring the Constitutional Right to Vote, Shelby County v. Holder
.”
Harvard Law and Policy Review
39
, no.
8
:
39
69
.
Brady
Henry E.
,
Verba
Sidney
, and
Schlozman
Kay Lehman
.
1995
. “
Beyond SES: A Resource Model of Political Participation
.”
American Political Science Review
89
, no.
2
:
271
94
.
Brault
Matthew W.
2012
. “
Americans with Disabilities: 2010
.”
Washington, DC
:
US Department of Commerce, Economics and Statistics Administration, US Census Bureau
.
Brucker
Debra L.
, and
Houtenville
Andrew J.
2015
. “
People with Disabilities in the United States
.”
Archives of Physical Medicine and Rehabilitation
96
, no.
5
:
771
74
.
Burden
Barry C.
,
Fletcher
Jason M.
,
Herd
Pamela
,
Jones
Bradley M.
, and
Moynihan
Donald P.
2016
. “
How Different Forms of Health Matter to Political Participation
.”
Journal of Politics
79
, no.
1
:
166
78
.
Burkhauser
Richard V.
,
Houtenville
Andrew J.
, and
Tennant
Jennifer R.
2014
. “
Capturing the Elusive Working Age Population with Disabilities: Reconciling Conflicting Social Success Estimates from the Current Population Survey and American Community Survey
.”
Journal of Disability Policy Studies
24
, no.
4
:
195
205
.
Campbell
Angus
,
Converse
Philip E.
,
Miller
Warren E.
, and
Stokes
Donald E.
1960
.
The American Voter
.
New York
:
Wiley
.
Campbell
Angus
,
Gurin
Gerald
, and
Miller
Warren E.
1954
.
The Voter Decides
.
Evanston, IL
:
Row, Peterson
.
Chattopadhyay
Jacqueline
.
2018
. “
Why Health Insurance Regulations Struggle to Generate Citizen Constituencies: Adding Policy Interdependence to the List of Design Features That Shape a Policy's Feedback Potential
.” Paper presented at the
Journal for Health, Politics, Policy and Law Special Issue on Health and Political Participation
Workshop,
Columbia, MO
,
February 2
.
Coleman
Kenneth M.
, and
Davis
Charles L.
1976
. “
The Structural Context of Politics and Dimensions of Regime Performance: Their Importance for the Comparative Study of Political Efficacy
.”
Comparative Political Studies
9
, no.
2
:
189
206
.
Collins
Sara R.
,
Robertson
Ruth
,
Garber
Tracy
, and
Doty
Michelle M.
2012
. “
Gaps in Health Insurance: Why So Many Americans Experience Breaks in Coverage and How the Affordable Care Act Will Help
.” Issue brief.
New York
:
Commonwealth Fund
.
Converse
Phillip
.
1964
. “
The Nature of Belief Systems in Mass Publics
.” In
Ideology and Discontent
, edited by
Apter
David
,
206
61
.
New York
:
Free Press
.
Converse
Philip E.
1972
. “
Change in the American Electorate
.” In
The Human Meaning of Social Change
, edited by
Campbell
Angus
and
Converse
Philip E.
,
263
337
.
New York
:
Russell Sage Foundation
.
Cox
Daniel
, and
Jones
Robert P.
2017
. “
America's Changing Religious Identity
.”
Public Religion Research Institute
,
September
6
. www.prri.org/research/american-religious-landscape-christian-religiously-unaffiliated/.
Downs
Anthony
.
1957
.
An Economic Theory of Democracy
.
New York
:
Harper
.
Galvin
Rose
.
2005
. “
Researching the Disabled Identity: Contextualising the Identity Transformations Which Accompany the Onset of Impairment
.”
Sociology of Health and Illness
27
, no.
3
:
393
413
.
Green
Donald
,
Palmquist
Bradley
, and
Schickler
Eric
.
2004
.
Partisan Hearts and Minds: Political Parties and the Social Identities of Voters
.
New Haven, CT
:
Yale University Press
.
Hajnal
Zoltan
,
Lajevardi
Nazita
, and
Nielson
Lindsay
.
2017
. “
Voter Identification Laws and the Suppression of Minority Votes
.”
Journal of Politics
79
, no.
2
:
363
79
.
Hall
Thad
, and
Alvarez
R. Michael
.
2012
. “
Defining the Barriers to Political Participation for Individuals with Disabilities
.” Information Technology and Innovation Foundation Accessible Voting Technology Initiative, Working Paper No. 1.
Washington, DC
:
Information Technology and Innovation Foundation
.
Harbaugh
W. T.
1996
. “
If People Vote Because They Like to, Then Why Do So Many of Them Lie?
Public Choice
89
, no.
1–2
:
63
76
.
Haselswerdt
Jake
, and
Michener
Jamila
.
2018
. “
Disenrolled: Retrenchment and Voting in Health Policy
.” Paper presented at the
Journal for Health, Politics, Policy and Law Special Issue on Health and Political Participation
Workshop, Columbia, MO, February 2.
Hasen
Richard L.
2012
.
The Voting Wars: From Florida 2000 to the Next Election Meltdown
.
New Haven, CT
:
Yale University Press
.
Huddy
Leonie
,
Mason
Lilliana
, and
Aarøe
Lene
.
2015
. “
Expressive Partisanship: Campaign Involvement, Political Emotion, and Partisan Identity
.”
American Political Science Review
109
, no.
1
:
1
17
.
Johnstone
Chris
.
2004
. “
Disability and identity: Personal Constructions and Formalized Supports
.”
Disability Studies Quarterly
24
, no.
4
.
Khazan
Olga
.
2017
. “
Why So Many Insurers Are Leaving Obamacare
.”
Atlantic
,
May
11
. www.theatlantic.com/health/archive/2017/05/why-so-many-insurers-are-leaving-obamacare/526137/.
Lane
Robert E.
1959
.
Political Life: Why People Get Involved in Politics
.
New York
:
Free Press of Glencoe
.
Leighley
Jan E.
, and
Nagler
Jonathan
.
2014
.
Who Votes Now? Demographics, Issues, Inequality, and Turnout in the United States
.
Princeton, NJ
:
Princeton University Press
.
Lenze
Eric J.
,
Rogers
Joan C.
,
Martire
Lynn M.
,
Mulsant
Benoit H.
,
Rollman
Bruce L.
,
Dew
Mary Amanda
,
Schulz
Richard
, and
Reynolds
Charles F.
III
.
2001
. “
The Association of Late-Life Depression and Anxiety with Physical Disability
.”
American Journal of Geriatric Psychiatry
9
, no.
2
:
113
35
.
Lewis-Beck
Michael S.
,
Jacoby
William G.
,
Norpoth
Helmut
, and
Weisberg
Herbert F.
2008
.
The American Voter Revisited
.
Ann Arbor
:
University of Michigan Press
.
Lopez
Mark Hugo
, and
Taylor
Paul
.
2012
. “
Latino Voters in the 2012 Election
.”
Pew Research Center
,
November
7
. www.pewhispanic.org/2012/11/07/latino-voters-in-the-2012-election.
National Institutes of Health
.
2010
. “
Disability in Older Adults Fact Sheet
.” report.nih.gov/nihfactsheets/Pdfs/DisabilityinOlderAdults(NIA).pdf.
Miller
Peter
, and
Powell
Sierra
.
2016
. “
Overcoming Voting Obstacles: Convenience Voting by People with Disabilities
.”
American Politics Research
44
, no.
1
:
28
55
.
Ojeda
Christopher
.
2015
. “
Depression and Political Participation
.”
Social Science Quarterly
96
, no.
5
:
1226
43
.
Ojeda
Christopher
, and
Slaughter
Christine
.
2018
. “
Intersectionality, Depression, and Voter Turnout
.” Paper presented at the
Journal for Health, Politics, Policy and Law Special Issue on Health and Political Participation
Workshop,
Columbia, MO
,
February 2
.
Overton
Spencer
.
2006
.
Stealing Democracy: The New Politics of Voter Suppression
.
New York
:
Norton
.
Pacheco
Julianna
, and
Fletcher
Jason
.
2015
. “
Incorporating Health into Studies of Political Behavior: Evidence for Turnout and Partisanship
.”
Political Research Quarterly
68
, no.
1
:
104
16
.
Palfrey
Thomas R.
, and
Poole
Keith T.
1987
. “
The Relationship between Information, Ideology, and Voting Behavior
.”
American Journal of Political Science
31
, no.
3
:
511
30
.
Papadopoulos
Konstantinos
,
Montgomery
Anthony J.
, and
Chronopoulou
Elena
.
2013
. “
The Impact of Visual Impairments in Self-Esteem and Locus of Control
Research in Developmental Disabilities
34
, no.
12
:
4565
70
.
Riker
William H.
, and
Ordeshook
Peter C.
1968
. “
A Theory of the Calculus of Voting
.”
American Political Science Review
62
no
1
:
25
42
.
Rosenstone
Steven
, and
Hansen
John
.
1993
.
Mobilization, Participation, and Democracy in America
.
New York
:
Macmillan
.
Schattschneider
E. E.
1960
.
The Semi-Sovereign People: A Realist's View of Democracy in America
.
New York
:
Holt, Rinehart, and Winston
.
Schlozman
Kay Lehman
,
Verba
Sidney
, and
Brady
Henry E.
2012
.
The Unheavenly Chorus: Unequal Political Voice and the Broken Promise of American Democracy
.
Princeton, NJ
:
Princeton University Press
.
Schreurs
Karlein M. G.
,
de Ridder
Denise T. D.
, and
Bensing
Jozien M.
2002
. “
Fatigue in Multiple Sclerosis: Reciprocal Relationships with Physical Disabilities and Depression
.”
Journal of Psychosomatic Research
53
, no.
3
:
775
81
.
Schur
Lisa A.
1998
. “
Disability and the Psychology of Political Participation
.”
Journal of Disability Policy Studies
9
, no.
2
:
3
31
.
Schur
Lisa
, and
Adya
Meera
.
2013
. “
Sidelined or Mainstreamed? Political Participation and Attitudes of People with Disabilities in the United States
.”
Social Science Quarterly
94
, no.
3
:
811
39
.
Schur
Lisa
,
Ameri
Mason
, and
Adya
Meera
.
2017
. “
Disability, Voter Turnout, and Polling Place Accessibility
.”
Social Sciences Quarterly
,
98
, no.
5
:
1374
90
.
Schur
Lisa
, and
Kruse
Douglas
.
2000
. “
What Determines Voter Turnout? Lessons from Citizens with Disabilities
.”
Social Science Quarterly
81
, no.
2
:
571
87
.
Schur
Lisa
, and
Kruse
Douglas
.
2014
. “
Disability Election Policies and Practices
.” In
The Measure of American Elections
, edited by
Burden
Barry C.
and
Steward
Charles
III
,
188
222
.
New York
:
Cambridge University Press
.
Schur
Lisa
,
Shields
Todd
,
Kruse
Douglas
, and
Schriner
Kay
.
2002
. “
Enabling Democracy: Disability and Voter Turnout
.”
Political Research Quarterly
55
, no.
1
:
167
90
.
Schur
Lisa
,
Shields
Todd
, and
Schriner
Kay
.
2003
. “
Can I Make a Difference? Efficacy, Employment, and Disability
.”
Political Psychology
24
, no.
1
:
119
49
.
Schur
Lisa
,
Shields
Todd
, and
Schriner
Kay
.
2005
. “
Generational Cohorts, Group Membership, and Political Participation by People with Disabilities
.”
Political Research Quarterly
58
, no.
3
:
487
96
.
Smith
Lauren E.
, and
Walker
Lee Demetrius
.
2012
. “
Belonging, Believing, and Group Behavior: Religiosity and Voting in American Presidential Elections
.”
Political Research Quarterly
66
, no.
2
:
399
413
.
Stolberg
Sheryl Gay
.
2018
. “
‘It's About Time’: A Baby Comes to the Senate Floor
.”
New York Times
,
April
19
. www.nytimes.com/2018/04/19/us/politics/baby-duckworth-senate-floor.html.
Stone
Deborah A.
1984
.
The Disabled State
.
Philadelphia
:
Temple University Press
.
Tate
Katherine
.
1993
.
From Protest to Politics: The New Black Voters in American Politics
.
New York
:
Russell Sage Foundation
.
Tate
Katherine
.
2010
.
What's Going On? Political Incorporation and the Transformation of Black Public Opinion
.
Washington, DC
:
Georgetown University Press
.
Treier
Shawn
, and
Hillygus
D. Sunshine
.
2009
. “
The Nature of Political Ideology in the Contemporary Electorate
.”
Public Opinion Quarterly
73
, no.
4
:
679
703
.
United Nations
.
2017
. “
Factsheet on Persons with Disabilities
.” www.un.org/development/desa/disabilities/resources/factsheet-on-persons-with-disabilities.html.
Verba
Sidney
,
Burns
Nancy
, and
Schlozman
Kay Lehman
.
1997
. “
Knowing and Caring about Politics: Gender and Political Engagement
.”
Journal of Politics
59
, no.
4
:
1051
72
.
Verba
Sidney
, and
Nie
Norman H.
1972
.
Participation in America
.
New York
:
Harper and Row
.
Verba
Sidney
,
Schlozman
Kay Lehman
, and
Brady
Henry E.
1995
.
Voice and Equality: Civic Voluntarism in American Politics
.
Cambridge, MA
:
Harvard University Press
.
Wang
Tova Andrea
.
2012
.
The Politics of Voter Suppression: Defending and Expanding Americans' Right to Vote
.
Ithaca, NY
:
Cornell University Press
.
Wasfy
Jason H.
,
Stewart
Charles
III
, and
Bhambhani
Vijeta
.
2017
. “
County Community Health Associations of Net Voting Shift in the 2016 U.S. Presidential Election
.”
PLoS One
12
, no.
10
:
e0185051
.
Wolfinger
Raymond E.
, and
Rosenstone
Steven J.
1980
.
Who Votes?
New Haven, CT
:
Yale University Press
.
Wright
Kenicia
.
2016
.
Power and Minority Representation
. In
Global Encyclopedia of Public Administration, Public Policy, and Governance
, edited by
Farazmand
Ali
,
1
7
.
Switzerland
:
Springer International Publishing
.
Young
Jeffrey
.
2017
. “
In States That Didn't Expand Medicaid, It's as If Obamacare Doesn't Even Exist for the Poor
.”
Huffington Post
,
December
6
. www.huffingtonpost.com/2014/07/09/obamacare-medicaid-uninsured_n_5572079.html.

Appendix A: Variable Question Wording and Coding

Notes: Refused, skipped, don't know, not asked, and no data responses were recoded as missing for all variables, with the exception of political knowledge. Unless otherwise indicated, question wording and coding are identical for 2012 and 2016.

  • Disability 2012: Permanent disability mentioned as first response to employment status of respondent. Coded 0 = permanent disability not mentioned first; 1 = permanent disability mentioned first.

  • Disability 2016: Permanent disability mentioned in response to employment status of respondent. Coded 0 = permanent disability not mentioned; 1 = permanent disability mentioned.

  • Employed 2012: Working now mentioned in response to employment status of respondent. Coded 0 = working now not mentioned; 1 = working now mentioned.

  • Employed 2016: Working now mentioned as first response to employment status of respondent. Coded 0 = working now not mentioned first; 1 = working now mentioned first.

  • Other unemployed 2012: Any category besides permanent disability or working now mentioned as first response to employment status of respondent. Coded 0 = working now or permanent disability; 1 = retired, unemployed, student, homemaker, or temporarily laid off.

  • Other unemployed 2016: Any category besides permanent disability or working now mentioned in response to employment status of respondent. Coded 0 = working now or permanent disability; 1 = retired, unemployed, student, homemaker, or temporarily laid off.

  • Political interest: Question wording: “How often do you pay attention to what's going on in government and politics?” Coded 1 = never; 2 = some of the time; 3 = about half the time; 4 = most of the time; 5 = always.

  • Follows news 2012: Combines responses to four questions about following news and national politics on the Internet, television, in printed newspapers, and on the radio. Values range from 1 = none at all for each question to 20 = a great deal for each question.

  • Follows news 2016: Question wording: “From which of the following sources have you heard anything about the presidential campaign?” Combines data from yes/no response options presented about the Internet, television, newspaper, and radio news. Values range from 0 = none selected to 4 = all selected.

2012 Political Knowledge

  • Index of responses to 7 questions about American politics. Values range from 0 to 7, with 7 being most knowledgeable. Each question coded 0 = incorrect; 1 = correct. Question wordings:

  • “Do you happen to know how many times an individual can be elected president of the United States under current laws?”

  • “Is the U.S. federal budget deficit, the amount by which the government's spending exceeds the amount of money it collects, now bigger, about the same, or smaller than it was during most of the 1990s?”

  • “For how many years is a United States senator elected, that is, how many years are there in one full term of office for a U.S. senator?”

  • “What is Medicare?”

  • “On which of the following does the U.S. federal government currently spend the least?”

  • “Do you happen to know which party had the most members in the House of Representatives in Washington BEFORE the election [this/last] month?”

  • “Do you happen to know which party had the most members in the U.S. Senate BEFORE the election [this/last] month?”

2016 Political Knowledge

Index of responses to 4 questions about American politics. Values range from 0 to 4, with 4 being most knowledgeable. For each question, 0 = incorrect; 1 = correct. Question wordings:

  • “For how many years is a United States senator elected—that is, how many years are there in one full term of office for a U.S. Senator?”

  • “On which of the following does the U.S. federal government currently spend the least?”

  • “Do you happen to know which party currently has the most members in the U.S. House of Representatives in Washington?

  • “Do you happen to know which party currently has the most members in the U.S. Senate?

Internal Efficacy

The 2012 index combined four questions, each asked to half the sample, to produce values from 2 to 10, with 10 being most internally efficacious. The 2016 index had the same values, but only the latter two questions were asked, and were asked of the entire sample. Question wordings:

  • “Sometimes, politics and government seem so complicated that a person like me can't really understand what's going on. Do you agree strongly; agree somewhat; neither agree nor disagree; disagree somewhat; disagree strongly with this statement?” Coded 1 = agree strongly; 2 = agree somewhat; 3 = neither agree nor disagree; 4 = disagree somewhat; 5 = disagree strongly.

  • “I feel that I have a pretty good understanding of the important political issues facing our country. Do you agree strongly; agree somewhat; neither agree nor disagree; disagree somewhat; disagree strongly with this statement?” Coded 1 = disagree strongly; 2 = disagree somewhat; 3 = neither agree nor disagree; 4 = agree somewhat; 5 = agree strongly.

  • “How often do politics and government seem so complicated that you can't really understand what's going on?” Coded 1 = always; 2 = most of the time; 3; about half of the time 4; some of the time; 5 = never.

  • “How well do you understand the important political issues facing our country?” Coded 1 = not well at all; 2 = slightly well; 3 = moderately well; 4 = very well; 5 = extremely well.

External Efficacy

The 2012 index combined four questions, each asked to half the sample, to produce values from 2 to 10, with 10 being most internally efficacious. The 2016 index had the same values, but only the latter two questions were asked, and were asked of the entire sample. Question wordings:

  • “How much do public officials care what people like you think?” Coded 1 = not at all; 2 = a little; 3 = a moderate amount; 4 = a lot; 5 = a great deal.

  • “How much can people like you affect what the government does?” Coded 1 = not at all; 2 = a little; 3 = a moderate amount; 4 = a lot; 5 = a great deal.

  • “Public officials don't care much what people like me think. Do you agree strongly; agree somewhat; neither agree nor disagree; disagree somewhat; disagree strongly with this statement?” Coded 1 = agree strongly; 2 = agree somewhat; 3 = neither agree nor disagree; 4 = disagree somewhat; 5 = disagree strongly.

  • “People like me don't have any say about what the government does. Do you agree strongly; agree somewhat; neither agree nor disagree; disagree somewhat; disagree strongly with this statement?” Coded 1 = agree strongly; 2 = agree somewhat; 3 = neither agree nor disagree; 4 = disagree somewhat; 5 = disagree strongly.

Campaign Activity

Index combines “Yes” responses to seven questions. Values range from 0 to 7, with 7 being most involved. Question wordings:

  • “We would like to find out about some of the things people do to help a party or a candidate win an election. During the campaign, did you talk to any people and try to show them why they should vote for or against one of the parties or candidates?” (Yes, No)

  • “Did you go to any political meetings, rallies, speeches, dinners, or things like that in support of a particular candidate?” (Yes, No)

  • “Did you wear a campaign button, put a campaign sticker on your car, or place a sign in your window or in front of your house?” (Yes, No)

  • “Did you do any (other) work for one of the parties or candidates?” (Yes, No)

  • “During an election year people are often asked to make a contribution to support campaigns. Did you give money to an individual candidate running for public office?” (Yes, No)

  • “Did you give money to a political party during this election year?” (Yes, No)

  • “Did you give money to any other group that supported or opposed candidates?” (Yes, No)

Party Identification

Coded 1 = strong Democrat; 2 = not very strong Democrat; 3 = independent leans Democrat; 4 = independent; 5 = independent leans Republican; 6 = not very strong Republican; 7 = strong Republican. Question wordings:

  • “Generally speaking, do you usually think of yourself as a Democrat, a Republican an Independent, or what?”

  • If responded Democrat or Republican: “Would you call yourself a strong Democrat/Republican?”

  • If responded Independent, No Preferences, or Don't Know: “Do you think of yourself as closer to the Republican Party or to the Democratic Party?”

Strength of Party Identification

  • Coded 1 = independent; 2 = independent leaner; 3 = not very strong Democrat, not very strong Republican; 4 = strong Democrat, strong Republican. Question wording same as for party identification.

Ideology

  • Question wording: “Where would you place yourself on this scale, or haven't you thought much about this?” Coded 1 = extremely; 2 = liberal; 3 = slightly liberal; 4 = moderate (middle of the road); 5 = slightly conservative; 6 = conservative; 7 = very conservative.

Strength of Ideology

  • Coded 1 = moderate; 2 = slightly liberal, slightly conservative; 3 = liberal, conservative 4 = extremely liberal, extremely conservative. Question wording same as for ideology.

Congressional Approval

  • Question wording: “Do you approve or disapprove of the way the U.S. Congress has been handling its job?” Coded 1 = approve; 2 = disapprove.

Presidential Approval

  • Question wording: “Do you approve or disapprove of the way Barack Obama has been handling his job as President?” Coded 1 = approve; 2 = disapprove.

Voted

  • Summary variable of whether or not respondent voted in the November general election. Coded 0 = did not vote; 1 = voted.

Vote Choice

  • Question wording: “How about the election for president? Did you vote for a candidate for president?” If yes, asked: “Who did you vote for?” For 2012 coded 0 = Obama; 1 = Romney; for 2016 coded 0 = Clinton; 1 = Trump.

  • Question wording: “How about the election for president? Did you vote for a candidate for president?” If yes, asked: “Who did you vote for?” For 2012 coded 0 = Obama or Romney; 1 = Other; for 2016 coded 0 = Clinton or Trump; 1 = Johnson, Stein, Other

Corruption in Government: 2012

  • Question wording: “How many of the people running the government are corrupt?” Coded 5 = all; 4 = most; 3 = about half; 2 = a few; 1 = none.

Corruption in Government: 2016

  • Question wording: “How many in government are corrupt?” Coded 5 = all; 4 = most; 3 = about half; 2 = a few; 1 = none.

Demographic Variables

  • Age 2012: Respondent age in categories of years. Coded 1 = 17–20; 2 = 21–24; 3 = 25–29; 4 = 30–34; 5 = 35–39; 6 = 40–44; 7 = 45–49; 8 = 50–54; 9 = 55–59; 10 = 60–64; 11 = 65–69; 12 = 70–74; 13 = 75+

  • Age 2016: Respondent age in years coded 18–90 years.

  • Female: Gender of the respondent coded 0 = male; 1 = female.

  • Married: Question wording: “Are you now married, widowed, divorced, separated or never married?” Coded 0 = widowed, divorced, separated, never married; 1 = married.

  • Black: Respondent race and ethnicity coded 0 = nonblack; 1 = black.

  • Hispanic: Respondent race and ethnicity coded 0 = non-Hispanic; 1 = Hispanic.

  • Education 2012: Respondent's highest level of education coded 1 = less than high school; 2 = graduated high school; 3 = some college; 4 = graduated college; 5 = graduate degree.

  • Education 2016: Respondent's highest level of education coded 1 = less than first grade; 2 = first, second, or third grade; 3 = fifth or sixth grade; 4 = seventh or eighth grade; 5 = ninth grade; 6 = tenth grade; 7 = eleventh grade; 8 = twelfth grade no diploma; 9 = high school graduate; 10 = some college; 11 = associate degree in college—occupational; 12 = associate degree in college—academic; 13 = bachelor's degree; 14 = master's degree; 15 = professional school degree; 16 = doctorate.

  • Household income: Family income coded 1 = under $5,000; 2 = $5,000–$9,999; 3 = $10,000–$12,499; 4 = $12,500–$14,999; 5 = “$15,000–$17,499; 6 = $17,500–$19,999; 7 = $20,000–$22,499; 8 = $22,500–$24,999; 9 = $25,000–$27,499; 10 = $27,500–$29,999; 11 = 30,000–$34,999; 12 = “$35,000–$39,999; 13 = $40,000–$44,999; 14 = $45,000–$49,999; 15 = $50,000–$54,999; 16 = $55,000–$59,999; 17 = $60,000–$64,999; 18 = $65,000–$69,999; 19 = $70,000–$74,999; 20 = $75,000–$79,999; 21 = $80,000–$89,999; 22 = $90,000–$99,999; 23 = $100,000–$109,999; 24 = $110,000–$124,999; 25 = $125,000–$149,999; 26 = $150,000–$174,999; 27 = $175,000–$249,999; 28 = $250,000 or more.

  • Religiosity: Question wording: “Do you consider religion to be an important part of your life, or not?” Coded 0 = not important; 1 = important.

Appendix B: Supplemental Analytical Models