Over a billion people—around 15% of the world’s population—live with a disability (WHO, 2018). As defined in Article 1 of the United Nations Convention on the Rights of Persons with Disabilities (CRPD), this includes people who “have long-term physical, mental, intellectual or sensory impairments which in interaction with various barriers may hinder their full and effective participation in society on an equal basis with others.” Disabled people ¹ have historically experienced discrimination, stigma, and exclusion from various spheres of society and continue to do so today (Ryan, 2019; Nario-Redmond, 2020). As a key step to improving their rights and living conditions, the disability community has consistently called for greater inclusion in political decision-making processes—in other words, “Nothing About Us Without Us” (Charlton, 1998). Yet, political representatives who are known to be disabled, either because they have a visible impairment or because they publically identify as such, remain few and far between (House of Commons, 2010). Anonymous candidate survey data from the United Kingdom confirm that disabled people are underrepresented among candidates seeking election to parliament, constituting only around 10% compared to 20% in the general public (Reher, forthcoming 2020b). This underrepresentation might be one reason for the lower levels of political trust and electoral participation of disabled people, which scholars have observed in different countries (e.g., Schur and Adya, 2013; Mattila and Papageorgiou, 2017; Powell and Johnson, 2019; Reher, 2020a).

The numerical underrepresentation of disabled people in politics is likely to be jointly explained by a range of different factors. While studies have pointed out various barriers in the political recruitment process, from inaccessible buildings to higher costs of campaigning and a political culture that is often not inclusive (Evans and Reher, 2020; Waltz and Schippers, 2020), our understanding of how voters view disabled candidates remains extremely limited (but see Loewen and Rheault, 2019). Yet, based on what we know about voter attitude formation and decision-making, on the one hand, and the pervasiveness of disability stereotypes, on the other, we have strong reasons to suspect that candidate disability does influence voter perceptions.

As voters rarely have the motivation, capacity, and resources to gather a large amount of information to evaluate candidates in an election, they often rely on cognitive shortcuts (Lau and Redlawsk, 2001). Stereotypes about social groups are a powerful basis of information that we use to form our impression of others (Fiske and Neuberg, 1990), and this also applies to the political context: studies have shown that voters make assumptions about the character traits as well as the policy interests, positions, and competencies of candidates based on their gender (e.g., Sapiro, 1981/1982; Alexander and Andersen, 1993; Huddy and Terkildsen, 1993; Koch, 2002; Sanbonmatsu, 2002; Lawless, 2004; Dolan, 2010; Schneider and Bos, 2014) and their race and ethnicity (Sigelman et al., 1995; Schneider and Bos, 2011; Jones, 2014).

The rarity with which we encounter disabled politicians implies that an impairment is likely to be a characteristic that stands out to voters, and this “contextual novelty” may strengthen the extent to which they use group generalizations to evaluate disabled candidates (Koch, 2002). Cultural stereotypes about disabled people are predominantly negative and often result in avoidant reactions (e.g., Weinberg, 1976; Park et al., 2005; Louvet, 2007; Nario-Redmond, 2010). However, drawing on insights from social psychology and disability studies, I argue that there is another image that might influence voter perceptions, namely, that of an inspirational person who has proven to be capable of overcoming hurdles and persevering in the face of adversity. When it comes to the policy-specific beliefs and competencies of candidates, voters are likely to draw inferences from their assumptions about the experiences and interests of disabled people as well as their trait perceptions.

The expectations are tested via an original survey experiment conducted among a representative sample of the British public (N = 1,500). The conjoint design allows identifying the effects of three common impairment types—a visual, a hearing, and a mobility impairment—among candidates on voter perceptions. The findings suggest that candidate disability influences how voters perceive the traits, beliefs, and competencies of candidates more consistently and strongly than most other characteristics and identities tested, even though the effects are modest in size. Remarkably, the effects are almost exclusively positive, with disabled candidates being perceived as more compassionate, hard-working, and honest than nondisabled candidates. Furthermore, they are seen as more concerned about and competent in handling issues likely assumed to be of particular interest to disabled people: healthcare, minority rights, and social security and welfare. In line with these perceptions, voters also perceive disabled candidates to be situated further to the left on the ideological spectrum.

The study thus yields a number of important insights. First, it firmly places disability on the list of candidate characteristics that we know matter to voters’ attitudes, therefore deserving further research. Second, it suggests that voters perceive disabled candidates as having specific strengths, and using this knowledge strategically might help disabled people gain greater representation in politics. Third, the patterns are not dissimilar to those for gender and ethnicity, encouraging further research into the electoral consequences of the intersections of different identities among candidates.

The hypotheses are tested through an experiment with a conjoint design embedded in an online survey of a representative sample (based on age, gender, and regional quotas) of 1,500 respondents in Britain, conducted through Qualtrics in May and June 2020. ² In the experiment, respondents are presented with descriptions of fictional candidates and asked to evaluate them on a number of dimensions. The United Kingdom is a very suitable case for the study since citizens are used to evaluating and voting for individual political candidates in its general elections, as well as in some subnational elections, under a first-past-the-post system. Moreover, British voters are unlikely to be completely unfamiliar with the idea of disabled candidates considering that the country has had a handful of prominent disabled politicians in the past and currently has a few high-profile representatives with disabilities in the House of Commons (e.g., Robert Halfon MP and Marsha de Cordova MP).

The existence of disability groups within the major political parties, government funding to support disabled candidates’ campaigns ³ , and the regular inclusion of disability in political debates on representation (House of Commons, 2010) also suggest that British voters are less likely to be surprised by candidate profiles that mention disability than voters in countries where the topic is largely ignored. As a result, there are unlikely to be significant demand effects, where respondents guess the purpose of the experiment and respond in a way that confirms the hypotheses. At the same time, Britain might also be a least-likely case to find negative effects of candidate disability on voter support: cross-national data from a 2012 Eurobarometer survey show that Britons express high levels of comfort with the idea of a disabled political leader as compared to most other European countries (Figure 1).

In the survey experiments, respondents are presented with a pair of short vignettes describing two fictional candidates, A and B, standing for election to the House of Commons in the respondent’s constituency. The descriptions contain information about several attributes of the candidates, which can take on a set of predefined values: gender (either male or female), minority ethnic background (yes/no), age (35–65), profession (doctor/lawyer/teacher/business owner/factory worker) (cf. Carnes and Lupu, 2016), number of children (0–3) (cf. Campbell and Cowley, 2018), years of political activity (4–17), and previous experience of elected office (local Councillor/none). Most importantly, the descriptions either mention no disability or state that the candidate 1) is paralyzed below the waist and uses a wheelchair to get around; 2) is blind and reads using text-to-speech software; or 3) is deaf and communicates mostly in British Sign Language. Details on the attributes and their values are provided in Supplementary Table S1 in the Supplementary Information; Supplementary Figure S1 contains the text of the vignettes. ⁴

The experiment has a conjoint design, meaning that the values of the attributes listed above are, independently from one another, randomly assigned to each respondent. This results in a large number of different combinations of characteristics that a candidate can have. The random assignment allows identifying the effects of each characteristic on the outcomes and comparing their effect sizes. The method enables us to estimate the average marginal component effect (AMCE) of candidate disability over all values of the other attributes, for instance, across male and female candidates and candidates of different ages (Haimueller et al., 2014). This is important, first, because the effects of disability on voter perceptions might vary with other characteristics of the candidates. For instance, if candidate gender and disability interacted, presenting respondents only with male candidates would lead to false conclusions if generalized to female candidates. Second, by manipulating other characteristics of the candidates that voters might associate with disability, we can more accurately identify the effect of disability. For instance, voters might assume that disabled candidates are less educated and less likely to be employed and professionally successful, in line with the lower education, employment, and income levels of disabled people. Alternatively, they might assume that disabled candidates have required more socioeconomic resources in order to stand for election due to the higher barriers. By experimentally manipulating candidates’ profession, I control for such assumptions.

Some attributes have different probabilities of appearing (e.g., ethnic minority background and impairments; Supplementary Table S1) in order for the distribution of candidate profiles to appear more realistic. After being presented with the vignettes, respondents are asked a set of questions about their impressions of the traits, beliefs, and issue competencies of both candidates, listed in Table 1. The response scales of these dependent variables have been normalized to range from 0 to 1 in order to facilitate interpretation of the effect sizes.

For the disability treatments, I selected three impairment types with which most people will be at least superficially familiar: a mobility impairment (paraplegic), a visual impairment, and a hearing impairment. Respondents are moreover provided with information about the nature of key barriers—mobility, reading, and communicating, respectively—and the adjustments the candidates use to address them. Thereby, respondents have some additional cues to form an image of what the disability might mean for candidates' election campaign and work as a representative. Deaf persons often communicate in sign languages, such as British Sign Language (BSL), which means that interpreters may be required in face-to-face interactions with other actors involved in the policy-making process, voters or journalists. Since BSL is sometimes their first language, some deaf people also have difficulty reading and writing in English, which could affect, for instance, the speed at which they are able to process policy documents in preparation for parliamentary debates or communication from constituents. In the election campaign, deaf candidates may face barriers communicating with constituents, participating in debates, etc. A range of adjustments can be made to address these barriers, such as providing interpreters and scribes as well as subtitles during speeches and debates. These adjustments require some organizational effort and financial resources, and candidates have reported occasional issues with the provision of these adjustments by political parties (Evans and Reher, 2020).

Individuals with (severe) visual impairments also face barriers related to communication, primarily reading and writing. They often use adjustments such as text-to-speech and speech-to-text software, which can be crucial for officeholders to read policy documents, etc. Additional barriers in this respect include documents in formats that cannot be read by the software and some financial costs of providing hardware and software (Evans and Reher, 2020). Other barriers that voters might be concerned about include difficulties with moving around, for instance, during canvassing but also as elected representatives. By contrast, for persons with mobility impairments who use a wheelchair, communication, reading, and face-to-face interactions are generally not affected as much as accessing the built environment and getting around. These physical barriers can have important implications during election campaigns, depending on the candidate’s financial means, support, and access to transport as well as the accessibility of stages at hustings. Other key concerns that voters might have include the inaccessibility for wheelchair users of parts of the Palace of Westminster and other buildings that MPs need to access, as well as the costs of improving accessibility. Travel, for instance, between London and the representative’s constituency, may also pose a range of barriers.

While the three impairment types and associated barriers are quite different, none of them stands out in terms of the amount or severity of barriers that voters likely anticipate, nor with respect to societal stigma. In Tringo's (1970) hierarchy of disability types, attitudes towards paraplegics are slightly more negative than towards deaf and blind people. At the same time, physical and sensory impairments are generally less stigmatized than intellectual or learning disabilities and mental health conditions (Stone and Colella, 1996; Bell and Klein, 2001). Accordingly, Loewen and Rheault (2019) found that voters have more negative views of, and are less likely to vote for, candidates with mental health conditions as compared to those with physical illnesses. By contrast, we would not expect large differences in voter perceptions for the three impairment types included here. Hence, rather than exploring variation, the aim of this study, and the rationale for including different disabilities, is to establish whether candidate disability has robust effects on voter perceptions.

To analyze the effects of candidate disability, I regress the outcome measures on the randomly assigned candidate characteristics using linear regression analysis. All models include fixed effects indicating the position of the candidate in the vignettes (i.e., Candidate A or B in the first or second experiment) and standard errors clustered by survey respondent, as each respondent evaluated two candidates.

Despite significant progress in the rights of disabled people over the past few decades, including in the United States with the American Disability Act from 1990, in Britain with the Equality Act 2010, and with the United Nations Convention on the Rights of People with Disabilities from 2006, disabled politicians remain a rare sight. Following the 2019 general election in Britain, there were as few as five MPs in the British House of Commons who identify as disabled ⁶ , and the numbers are not higher elsewhere either (Waltz and Schippers, 2020). It is certainly possible that there are also politicians with hidden disabilities who have chosen not to disclose them. Yet, this could be another symptom of the same issue that might be partly at the root of the low numbers: fear of negative reactions from voters. Given that disabled people are stigmatized across societies and often described as weak, incapable, dependent, and vulnerable, traits considered highly undesirable for political leaders, downplaying or ignoring an impairment or refraining from nominating disabled people as candidates would appear to be understandable reactions by (aspiring) candidates and party selectorate.

However, the findings of this study suggest that such fears may be largely unfounded, as voters overall do not appear to apply these negative stereotypes when evaluating disabled candidates. They do not see them as less competent and strong, as theories of stereotypes (Fiske et al., 2002) and previous research on perceptions of disabled people suggest (Nario-Redmond 2010), but instead as working harder than nondisabled candidates. These findings do not go as far as to suggest that disabled candidates are seen as political “supercrips,” an image of disabled people being exceptionally strong, capable, heroic, and inspiring. Yet, they indicate that standing for office with a disability is perceived by voters as evidence of being willing to work hard and able to persevere in the face of obstacles and challenges. Furthermore, voters also perceive disabled candidates as more compassionate and honest, in line with the stereotype content model prediction (Fiske et al., 2002). Thus, disabled candidates are able to enjoy the positive implications of being perceived as a group that has experience with low status and, as a result, a heightened sense of empathy, while avoiding the “flipside” of being seen as incompetent. Although all of the disability effects are rather small, they still tend to be more important than the few effects observed for other characteristics such as gender, ethnic minority identity, age, and even political experience. Only candidates’ profession had larger effects than disability in some cases. Whether and how these views positively influence voters’ willingness to vote for disabled candidates remains to be explored in future research.

That voters associate disabled candidates with healthcare, minority rights, and social policy and welfare does not come as a surprise, considering that studies have consistently shown that voters infer competence in and concern about policy issues from their assumptions about the interests and experiences of the social group to which a candidate belongs. Nevertheless, it is worth pointing out these findings because they have important implications. Voters who consider social policy and particularly issues related to health and disability as important may be inclined to vote for a disabled candidate. Contexts where these issues are salient may thus be particularly promising for disabled candidates. Accordingly, emphasizing these issues and framing their link to disability experience might be an effective campaign strategy for candidates with disabilities and their parties.

This first exploration into the effects of candidate disability on voter perceptions generated a range of further questions for future research. Some of them concern the role of political parties, which serve as a particularly powerful heuristic in voter decision-making (Rahn, 1993). As with candidate gender and race, party affiliation might mitigate the effects of candidate disability or interact with it (e.g., Koch, 2002; King and Matland, 2003; Hayes, 2011). For instance, in cases where disability stereotypes conflict with party stereotypes—i.e., a disabled candidate standing for a right-wing party—voters might either discount the disability stereotypes, or perceive the candidate as more liberal than his or her nondisabled peers (resulting in higher support from liberal voters), or punish the candidate for violating their expectations (Sigelman et al., 1995).

Voters’ own policy preferences are likely to play a key role, too (King and Matland, 2003), as might their personal experience with disability. Given that the disability gaps in the assumed political views of candidates are similar to those observed among citizens (Gastil, 2000; Schur and Adya, 2013; Reher, 2020b), disabled voters might be inclined to support disabled candidates to improve both their descriptive and substantive representation. However, it is also possible that disabled voters are less likely to perceive disabled candidates as a homogenous group and apply stereotypes when evaluating them, as they are more aware of the variation in experiences that disabled people have (Linville et al., 1996). Moreover, the extent to which a coherent disability identity exists remains contested (Watson, 2002; Jenks, 2019; Deal, 2003). Especially members of the deaf community tend to think of themselves as a cultural and linguistic group and often do not identify as disabled (Dolnick, 1993). Therefore, exploring how not only disabled citizens in general but also people with different impairments perceive disabled candidates is crucial in advancing our understanding of this issue. In this context, we should also expand the variety of disabilities among candidates we investigate, including hidden disabilities and chronic illnesses, learning disabilities and mental health conditions, and congenital and acquired impairments. Like the present study, this endeavor will greatly benefit from integrating knowledge and approaches from political science, social psychology, and disability studies.

Another objective for further research will be to expand this work to other contexts. Voter perceptions of disabled politicians are likely to vary with the prevalence and nature of stigma associated with disability in a society, for which we lack systematic comparative evidence as of yet. Moreover, differences in political systems and institutions will shape our research questions, design, and findings. In systems with multimember districts, representatives have more incentives and freedom to speak for social groups with which they identify rather than the geographical area they represent (Tremblay, 2003). This might imply that candidates’ personal characteristics play a more prominent role in shaping voters’ expectations about candidates’ policy priorities in these settings compared to systems with single-member districts (SMD), like the United Kingdom. However, how such considerations affect the vote choice is also likely to vary between electoral systems.

If voters perceive a candidate to represent the interests of a minority group (to which they do not belong), they might be inclined to support this candidate as part of a group of representatives who together represent a diverse group of citizens (i.e., in multimember districts). Yet, they might be less willing to vote for this candidate as the sole representative of their constituency (i.e., in SMD). Thus, we might find stronger support for disabled candidates in proportional representation (PR) systems with open lists, where voters give preference votes for candidates within a party rather than in SMD systems. In contrast, in PR closed-list elections, voters are not asked to evaluate individual candidates. Therefore, these contexts will require a different experimental design, and it is likely that voter perceptions of candidates—however strongly these perceptions are affected by disability—play a weaker role in their vote choice (cf. Chin and Taylor-Robinson, 2005).

The study also lays a foundation for further research into the causes of the low numbers of disabled politicians in power. If it emerges that voters do not discriminate against disabled candidates at the ballot box, as the findings here would lead us to expect, then other factors must be responsible. As previous research shows, candidates report a number of barriers that they experience in the selection and election processes, including a lack of accessibility and insufficient funding to make processes accessible, fewer networks and mentors, a risk of losing disability benefits, and a political culture that can be exclusive and alienating (Evans and Reher, 2020; Waltz and Schippers, 2020). However, there may be other factors, such as societal stigma, fewer resources, or a lack of role models, which can make it difficult or off-putting for disabled people to become involved in political parties and seek nomination as a candidate in the first place (D’Aubin and Stienstra, 2004; Sackey, 2015; Langford and Levesque, 2017; Levesque, 2016). What is more, parties might be hesitant to recruit and select them for various reasons, including prejudice and fear of a backlash at the ballot box. To determine the relevance of these various factors, scholars will need to trace the role of disability in the entire recruitment process, from initial political participation at the citizen level (Schur and Adya, 2013; Mattila and Papageorgiou, 2017; Reher, forthcoming 2020b) to the decision of running for reelection.

Finally, it would be valuable to examine whether the findings travel to perceptions of disabled people in other professional groups, especially those where having experience with adversity in general and/or disability in particular could be seen as an advantage, like the care sector, humanitarian work, or the legal profession. In this context, a key question is whether individuals need to provide some form of evidence for previous achievement, equivalent to being nominated as a candidate, to counter perceptions of low competence and weakness.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by the University Ethics Committee, University of Strathclyde. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

The author confirms being the sole contributor of this work and has approved it for publication.

The research was supported by the Economic and Social Research Council (ES/S015469/1) and the Carnegie Trust for the Universities of Scotland (RIG007430).

The author declares that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.