The Partisan Vote: Shared Partisanship and Constituent Service in the U.S. Congress

Abstract

The theory of the “personal vote” assumes that constituent service, or the process by which members of Congress help their constituents navigate the federal bureaucracy, is a component of representation that transcends partisanship. However, recent research suggests that affective polarization and negative partisanship may discourage constituents from reaching out to ask for help from cross-partisan legislators. Using a conjoint experiment, I show that shared partisanship has a strong effect on constituents’ decisions about which of their legislators to contact for assistance. However, positive in-party affect is unrelated to these preferences, suggesting that negative partisanship is responsible for this result. Gender and seniority also matter for constituents’ preferences even when accounting for shared partisanship. The results suggest that legislators’ ability to build a “personal vote” by performing constituent service may be weakened in an era of strong partisanship.

Keywords

representation constituent service partisanship

Constituent service, or the process by which legislators help their constituents navigate the federal bureaucracy, is a key activity of congressional offices. Interactions with constituents help legislators build and maintain their “home styles” (Fenno, 1978), and constituent service, as an ostensibly nonpartisan component of representation that does not depend on shared policy preferences between legislators and their constituents, has long been a critical component of the “personal vote,” helping legislators build goodwill and support among constituents who may not have supported them otherwise (Cain et al., 1987). Legislators signal the importance of providing constituent service in various ways, such as highlighting casework accomplishments in communications with constituents and allocating nearly half of their total staff to district offices where most constituent service takes place. However, recent changes in partisanship among the electorate, including the decline of split-ticket voting (Jacobson, 2015) and the rise of both affective polarization (Mason, 2018) and negative partisanship (Abramowitz & Webster, 2018), raise questions about the contemporary power of the personal vote, especially as the size of legislators’ constituencies has grown in ways that have made legislators less accessible (Frederick, 2008). Is shared partisanship between constituents and their legislators now important enough to influence even a nonpartisan component of representation like constituent service?

In this paper, I examine the personal vote from constituents’ perspectives and ask if constituents are equally likely to reach out to all types of legislators, particularly those who do not share their partisanship. I answer this question by conducting a conjoint experiment that is designed to reflect the process by which constituents decide which of their federal representatives to contact for assistance once they have made the initial decision to ask a legislator for help (e.g. Lee & Oppenheimer, 1999). The experiment allows me to vary several legislator characteristics, such as partisanship, gender, seniority, and race and ethnicity of their staff members, simultaneously and to examine how the absence of information about partisanship, which constituents may not encounter during the process of contacting their legislators, affects constituents’ choices. Three main findings emerge from the analysis. First, constituents show a strong preference for seeking help from co-partisan legislators, but there is no evidence that positive in-party affect is driving these results. Second, legislators’ gender is also an important cue for respondents, although respondents’ partisanship tempers this relationship in the absence of information about legislators’ partisanship, as only female Democrats and male Republicans prefer to seek help from same-gender legislators. Finally, in all analyses, legislators’ seniority matters to respondents, suggesting that constituents value experience in the provision of constituent service.

The paper proceeds as follows. I begin by describing the historically non-partisan nature of constituent service and the “personal vote” before establishing the logic of my hypotheses. Next, I review my experimental design and methods of analysis. I then present the results of the condition of the experiment that includes information about the legislators’ partisanship, which show the strong influence of shared partisanship on constituents’ decisions, followed by the results of the condition without a partisan cue, which show the additional importance of information about legislators’ gender in the absence of information about legislators’ partisanship. I also present the analysis of the mechanisms through which constituents’ selections are operating, showing that positive in-party affect is unrelated to respondents’ preferences for asking for help from co-partisan legislators. The final section concludes with a discussion of the results and their implications for existing and future research.

Constituent Service and the Personal Vote

In the 1970s and 1980s, as the incumbency advantage among members of Congress increased and the importance of partisanship for voters in congressional elections declined (Romero & Sanders, 1994), scholars sought to explain these trends in terms of the “personal vote” (e.g. Fiorina, 1977). The personal vote, broadly understood to refer to electoral support based on attributes of individual legislators rather than partisanship, is often narrowly understood to refer to the electoral support that accrues to incumbents as the result of their providing benefits and service to their districts (Herrera & Yawn, 1999). Fenno (1978) famously observed the processes by which members of Congress cultivated their “home styles” to establish themselves as representatives of their districts and connect with individuals and communities there. Legislators sought to distinguish themselves electorally by emphasizing their connections and service to the district.

Political scientists and members of Congress have both assumed that constituent service, along with other perks of office such as the franking privilege, provided a vehicle for incumbents to build support among their constituents in a way that was entirely unrelated to partisanship. Cain et al. (1987) made the most notable contribution to this theory, finding that, in a time of relatively weak partisan attachments, legislators received electoral benefits from performing constituent service. Although some studies have found no evidence that the “personal vote” matters (Johannes & McAdams, 1981), research has supported the notion that casework pays off for incumbents (Serra & Moon, 1994; Serra & Pinney, 2004; Yiannakis, 1981).

However, as the importance of partisan attachments has grown in recent decades (Jacobson, 2015), contemporary research has begun to suggest that partisanship also matters for constituent service. Constituents who do not share their legislator’s partisanship express less satisfaction with constituent service and are less likely to view their legislator as helpful (Wagner, 2007). Looking beyond constituent service, Kaslovsky (2022) finds that district visits, another key component of legislators’ home styles, actually hurt senators’ reputations among constituents who disagree with them, likely because the visits make senators’ policy records more salient. Most notably, Broockman and Ryan (2016) show that constituents are more likely to contact co-partisan legislators to express opinions and to seek assistance, but their study does not explicitly test the mechanisms underlying these preferences or examine how other traits may work in concert with partisanship. Building on these findings, in the next section I identify several reasons why, according to recent changes in partisanship among the electorate, constituents may prefer to contact co-partisan legislators for assistance, as well as why constituents may prioritize factors other than shared partisanship in this context.

Partisanship as a Challenge to the Personal Vote

Recent research on partisanship and polarization has identified the emerging phenomenon of “negative partisanship,” or the tendency to express stronger negative feelings toward the opposing party than positive feelings toward one’s own party (Abramowitz & Webster, 2016, 2018; Iyengar et al., 2019). Negative partisanship makes it extraordinarily difficult for members of Congress to emphasize their individuality when representing their districts or campaigning for votes, as “Americans care less about what individual represents them and more about which party controls Congress” (Abramowitz & Webster, 2018, p. 130). Elected officials with an overwhelming incentive to attack and denigrate the other party have little reason to focus on the provision of constituent service. These changes to the nature of partisanship have significant implications for representation and the theory of the personal vote. Gone, according to recent findings on partisanship, are the days in which members of Congress could effectively solicit casework as a means of garnering electoral support, as partisans are increasingly less likely to support members of the opposite party (Bartels, 2000).

Existing work on partisanship points to several potential mechanisms through which shared partisanship may influence constituents’ decisions about which of their legislators to contact for assistance. If affective polarization originates in differences in ideology and policy preferences (Orr et al., 2023; Webster & Abramowitz, 2017), constituents may be drawn to co-partisan legislators because they believe they share their beliefs and values. Conversely, if partisan identity is driving affective polarization (Huddy et al., 2015; Iyengar et al., 2012; Mason, 2016), constituents’ preferences for co-partisan legislators may be grounded in perceptions of similarity that arise from a shared partisan identity. Partisanship’s function as a shared identity could also foster trust between constituents and legislators just as descriptive representation by race and gender does (Gay, 2002; Mansbridge, 1999), leading constituents to ask for help from co-partisan legislators on this basis. Finally, Broockman and Ryan (2016, p. 1095) speculate that constituents’ perceptions of “stubbornness or incompetence” of opposite-party legislators may prevent them from contacting these legislators, suggesting that constituents may ascribe expertise to legislators who share their partisanship. Thus, I expect that partisans of both parties will be more likely to ask for help from legislators who share their partisanship (Hypothesis 1), and I test four potential mechanisms—shared beliefs and values, perceived similarity, trust, and perceived expertise—through which shared partisanship and other characteristics may operate.

Alternatives to Partisanship

Even considering the strong influence of partisanship on individual political behavior, constituents may prioritize other characteristics, such as descriptive traits, when choosing which of their legislators to ask for help. In contrast to shared gender, which has largely been found to have an inconsistent relationship with representation that is often conditional on shared partisanship (Costa & Schaffner, 2018; English et al., 2019; Lawless, 2004), shared race has a clear impact on representation and is likely to affect constituents’ choices independently of partisanship. In particular, Black legislators’ responsiveness to the policy concerns of Black constituents (Ellis & Wilson, 2013; Gamble, 2007) and increased attentiveness to service requests from Black constituents, even without a potential electoral reward (Broockman, 2013), may signal to these constituents that legislators and staffers who share their racial identity are likely to be more responsive to their concerns.

More generally, members of Congress are more likely to perform constituent service on behalf of constituents who share their identities, including gender and race (Lowande et al., 2019). Constituents are also more likely to indicate that they feel represented by legislators of the same race (English et al., 2019), and both white and Black constituents are more likely to contact legislators who share their race (Broockman, 2014; Gay, 2002). Grose et al. (2007) find that Democratic and Black members of Congress, as well as members who represent majority-Black districts, hire more Black staff members in district office positions, which enhances the representation of Black constituents through responsive constituent service. Because district staff are a key means through which constituents experience contact with their legislators (Abernathy, 2018; Kaslovsky, 2022; Miler, 2010), the identities of staffers should play a role in constituents’ contact preferences. I therefore expect that constituents will be more likely to contact legislators with staff members who share their race (Hypothesis 2).

It is also possible that constituents do not possess information about their legislators’ partisanship when contacting their legislators to request assistance. Partisan affiliation is rarely obvious in places where constituents may seek contact information for their legislators, such as their official websites, and many constituents cannot recall the names and party affiliations of the legislators who represent them in Congress.¹ Instead, just as voters use heuristics like gender and race (Matson & Fine, 2006) or occupation (McDermott, 2005) to select candidates for office, constituents may use other characteristics to infer legislators’ partisanship or rely on other factors entirely when deciding which of their legislators to contact for help. Although constituents of both parties hold women legislators to a higher standard than men when evaluating their responses to constituent correspondence (Costa, 2021a), voters tend to view women candidates, and Democratic women in particular, as more liberal than men (Dolan, 2004), so respondents may assume that female legislators are more liberal and use gender to make choices that accord with their partisanship in this way. Similarly, voters view Black candidates as more liberal than white candidates (McDermott, 1998), suggesting that constituents may also use race as a substitute for a partisan cue. Therefore, I expect that in the absence of information about legislators’ partisanship, partisan respondents will use the race and ethnicity of legislators’ staff members (Hypothesis 3a) and legislators’ gender (Hypothesis 3b) to infer partisanship, with Democratic respondents being more likely to contact female legislators and legislators with non-white staff members.

As supplements to partisan cues, constituents may also rely on legislator qualities that signal competence, including committee membership. Members of Congress aim to serve on committees that allow them to best meet the needs of their constituencies, so committee membership may signal members’ policy interests or representational priorities to constituents. Specifically, constituents may make decisions based on legislators’ presence on prestige, policy, or constituency-focused committees (e.g. Heberlig, 2003). Constituents’ exact preferences will vary based on the type of representation that they prefer. A seat on a constituency-oriented committee, for instance, may signal to constituents that the legislator is invested in providing service to their constituents. Other constituents, however, may value the legislative power that a seat on a prestige or policy committee signals, reasoning that legislators need a major voice in policy to represent their constituents within the chamber.

Given that legislators’ personal characteristics and backgrounds affect their legislative preferences (Burden, 2007), constituents may also use personal characteristics of legislators to make decisions about whom to contact. One such characteristic is the legislator’s former occupation, which voters can use to infer expertise (McDermott, 2005). Voters have been found to prefer legislators with former occupations that signal competence and ties to their communities (Campbell & Cowley, 2014), as well as experience that is relevant for the elected position (Atkeson & Hamel, 2020). Constituents, therefore, may use occupational cues similarly when selecting a legislator to contact for assistance with a particular type of problem. Partisan differences may exist for such effects, as Kirkland and Coppock (2018) find that Republican voters value previous job experience among candidates, while Democratic voters value previous political experience. Like committee membership, legislators’ former occupations can signal trustworthiness or expertise, but constituents are likely to interpret these cues differently. From these findings, I expect that respondents will choose to contact legislators with more experience and relevant expertise, including tenure in Congress, committee membership, and occupation, but that respondents’ partisanship will inform these preferences (Hypothesis 4).

Data and Methods

Experimental Design

To assess how legislators’ characteristics affect how constituents decide which of their legislators to contact for help once they have a reason to do so, I conducted a conjoint experiment (Hainmueller et al., 2014b).² The conjoint design is advantageous for three reasons. First, it resembles the actual decision process that constituents face when they choose which of their elected representatives to contact when in need of help with a federal agency. Although most typically used to study voters’ preferences of hypothetical candidates for office (Carnes & Lupu, 2016; Kirkland & Coppock, 2018; Sung, 2022; Teele et al., 2018), here I apply the conjoint design to a study of constituents’ preferences of hypothetical incumbent legislators (Costa, 2021b). Lee and Oppenheimer (1999) show that constituents commonly consider contact with all three of their federal legislators—their House member and two senators—when seeking assistance with a problem, making it appropriate to examine how legislators’ characteristics may affect these choices. Second, it allows me to vary many legislator traits simultaneously and to analyze the effects of these individual traits on constituents’ selections. Finally, an experiment is appropriate theoretically because of the nature of constituent service and how it is captured in survey data. In any given year, the number of respondents in a survey who report having contacted their member of Congress to seek assistance in the past year is relatively low, but this group of people varies from one year to the next as individual needs change and new bureaucratic issues arise among different constituents.³ To understand the full picture of constituent service, we must consider potential contact in addition to observed contact. An experiment that asks constituents to imagine a hypothetical situation in which they are contacting a member of Congress to ask for help is an appropriate way to examine potential contact as it may take place once constituents have made the initial decision to contact their legislators for help.

As shown in Figure 1, participants in the experiment first read a vignette explaining a hypothetical situation in which they need help with a matter that relates to a federal agency. This scenario, in which the constituent is waiting to receive a tax refund check from the IRS, is designed to be broadly applicable to the American public, as most American adults have experienced the process of filing federal income taxes. Other common subjects of constituent service, such as veterans’ benefits, Social Security, and immigration matters, apply to narrower segments of the general public, and even requests involving passports or tours of buildings in Washington may seem economically inaccessible to some survey respondents.

After reading the vignette, respondents viewed information about two hypothetical incumbent legislators whom they could contact for help with this issue and indicated which legislator from each pair they would prefer to contact. Respondents evaluated six pairs of hypothetical legislators in total. An example of two profiles that the conjoint experiment produced is shown in Figure A1 in the Online Appendix. Although many constituents may not always use such a broad array of information to decide which of their legislators to contact, information about most of the attributes included in the experiment is typically made available on members’ websites and would often be encountered when navigating those sites to submit casework requests.

In addition to the legislators’ partisanship, my main independent variable of interest, the experiment varied the legislators’ chamber (House or Senate), gender (male or female), length of service in Congress (2, 12, or 20 years), current committee membership (Agriculture, Budget, Judiciary, or Veterans’ Affairs), and former occupation (lawyer, public school teacher, small business owner, or state legislator).⁴ The latter three traits allow me to assess how legislators’ expertise affects constituents’ choices. Specifically, I varied committee membership between two policy-focused committees (Budget and Judiciary) and two constituency-focused committees (Agriculture and Veterans’ Affairs) that exist in both the House and the Senate to determine whether constituents may choose to contact legislators based on their perceived orientations toward representation.

Because overt partisan cues are often absent from biographies on members’ websites, I followed Sen (2017) and randomly assigned half of the sample to receive a version of the experiment without information about the legislators’ partisanship. This allows me to explore how respondents may use other cues to infer legislators’ partisanship or how respondents may make choices based on other criteria altogether. Finally, to understand how descriptive representation at the staff level may influence constituents’ choices of whom to contact, I also varied the names of the primary staff contacts of each legislator, with the names designed to cue the race (white, Black, Hispanic, or Asian American) of the staff members (Butler & Broockman, 2011). Staff in district offices and constituent service positions are more likely to be Black or Hispanic (Ziniel, 2009), and the unique interpersonal nature of constituent service contact may make the race of the staff member particularly important for constituents (Grose et al., 2007).⁵ I chose to vary the names of staff members both to capture the specific effects of constituents’ identifications with staff and because creating names of hypothetical legislators would sacrifice a significant amount of realism in the experiment.

Following the experiment, respondents answered five additional questions about the legislators that they selected. The first four questions were designed to uncover the mechanisms—shared beliefs and values, perceived similarity, trust, and perceived expertise—through which respondents’ selections of legislators may operate. I followed Tomz and Weeks (2013) and used post-experiment questions to assess these mechanisms because the mechanisms involve respondents’ perceptions of these legislators. These questions asked respondents to assess how well the following statements described the legislators they selected in the experiment: “They are someone like me,” “They seem like someone I can trust,” “They share my beliefs and values,” and “They have the expertise required to solve the problem.” The response scale ranged from 1 (“Not at all well”) to 4 (“Very well”). The fifth question asked respondents how confident they were that they would receive the assistance that they requested from their legislators, regardless of whether the issue could be resolved. The response scale ranged from 1 (“Not at all confident”) to 4 (“Very confident”). This question allows me to assess the extent to which these respondents would feel represented by the process of seeking assistance from the legislators they selected in the experiment.

Sample

I obtained respondents for the experiment through Lucid Theorem, a firm that connects researchers to a nationally representative sample of survey respondents that has been shown to provide reliable estimates of experimental effects (Coppock & McClellan, 2019). After programming the conjoint experiment in Qualtrics, I provided Lucid Theorem with a link through which respondents could access the survey. Responses were collected on August 24 and 25, 2020. Lucid Theorem provides demographic information about the respondents, including their age, gender, race, ethnicity, level of education, region of residence, ZIP code, and political party. In all analyses that follow, I include only respondents who spent more than 1 minute completing the survey to minimize effects that may occur due to inattentive survey responses. This robust sample consists of 1768 American adults, which ensures that each of the two conditions of the experiment contains enough respondents to facilitate analysis of the data. The partisan condition ultimately included 885 respondents who each viewed six pairs of legislators, resulting in a total of 10,620 observations; the non-partisan condition included 883 respondents who also each viewed six pairs of legislators, for a total of 10,596 observations.

Analysis

To analyze the results from the conjoint experiment, I follow Hainmueller et al. (2014b) and estimate the average marginal component effects (AMCEs) of the legislator attributes using Hainmueller et al.'s (2014a) “cjoint” package. The AMCE estimates show the probability of a respondent’s selecting a legislator with the value of a specific attribute relative to a baseline level when all of the other attributes’ effects are averaged. It allows me to identify the impact of specific legislator traits on respondents’ choices. Because the unit of analysis is the conjoint profile, I cluster standard errors at the respondent level to account for the lack of independence that exists between respondents’ profile selections.

To assess the mechanisms—shared beliefs and values, perceived similarity, trust, and perceived expertise—through which respondents’ selections of legislator attributes may operate, I use ordered logistic regression to examine how shared partisanship, shared race and ethnicity, and shared gender affect respondents’ evaluations of their chosen legislators on these four outcomes. I also examine how these factors affect respondents’ confidence that they would receive the assistance they requested. The dependent variables in each ordered logit model are respondents’ assessments of the legislators that they chose in the experiment, as measured using the four-point scale described above. The independent variables are the number of profiles that a respondent selected in which the legislator or their staff shared the respondent’s partisanship, race or ethnicity, or gender. The models also control for respondents’ age, gender, household income, and level of education. As in the AMCE analyses of the conjoint profiles, I cluster standard errors at the respondent level. This analysis allows me to determine if selection of legislators based on shared traits is grounded in trust, perceived similarity, a sense of shared values, or confidence in the legislators’ expertise. I expect that respondents who chose a greater number of legislators who share their characteristics will evaluate their selected legislators more positively.

Results

Partisan Condition

I first present the results from the condition of the experiment in which respondents received information about each legislator’s partisanship. Figure 2 shows the results of the average marginal component effect (AMCE) estimates for this condition. The most striking result is the strong effect that legislators’ time in office exerts on respondents’ preferences. Respondents are 8 and 10.3 percentage points more likely to prefer contacting legislators who have been in Congress for 12 and 20 years, respectively, than they are to prefer contacting legislators who have only been in Congress for two years. These results accord with my expectations in Hypothesis 4 that respondents will prefer to ask for help from legislators with more experience in Congress. Respondents are 4.3 percentage points more likely to select members of the Budget Committee compared to members of the Judiciary Committee, and they prefer former lawyers and small business owners over former state legislators by 3 and 2.9 percentage points, respectively. Respondents also show a bias of 2.3 percentage points against male legislators and a slight preference for House members over Senators, although this effect just fails to reach statistical significance. There are no clear effects for the racial identities of legislators’ staff members.

To provide a direct test of Hypothesis 1 and show how partisan differences between respondents may be driving the results contained in Figure 2, Figure 3 shows the results from the partisan condition when respondents are separated by political party. In accordance with Hypothesis 1, shared partisanship exerts a strong effect on respondents’ preferences, with Democrats being 24.4 percentage points more likely to seek help from Democratic legislators and Republicans being 22 percentage points less likely to seek help from Democratic legislators. Democrats, who are 6.8 percentage points less likely to contact male legislators and 5.5 percentage points more likely to contact members of the Budget Committee, are driving the preferences for these legislators that appear in the full results. Republicans show a nearly significant preference (4.9 percentage points) for former small business owners compared to former state legislators, and Democrats show a slight preference (5.7 percentage points) for former lawyers. Partisans of both parties prefer legislators who have been in office for longer periods of time.

The results for shared race, shown in Figure A2 in the Online Appendix, do not provide strong support for Hypothesis 2. While Democrats who are white are 6.2 percentage points less likely to ask for help from legislators with Asian American staff members and 6.9 percentage points less likely to ask for help from legislators with Hispanic staff members than legislators with white staff members, there is little evidence of effects for shared race among other groups. Hispanic Democrats are slightly more likely to select legislators with Hispanic and Asian American staff members, although these effects are not statistically significant. Republicans, meanwhile, show no clear preferences regarding race, although Hispanic Republicans, like Hispanic Democrats, show a positive but statistically insignificant preference for legislators with Hispanic and Asian American staff members.

Analyses interacting respondents’ gender with legislators’ gender show that the effect of shared gender is strongest among women who identify as Democrats, with these respondents being 11.9 percentage points less likely to seek assistance from male legislators. Similarly, Republican men are 6.8 percentage points more likely to contact male legislators for help. While Republican women are less likely (4.3 percentage points) to select male legislators, the effects are not statistically significant. Democratic men are also no more or less likely to contact male legislators. That this effect for shared gender appears even when respondents possess information about legislators’ partisanship is striking, as it suggests that Democratic women and Republican men are not merely using gender as an additional marker of partisanship, but rather that they prefer to seek help from legislators who share their gender regardless of party affiliation. However, these effects are muted by partisan affiliation, with Republican women forgoing selections based on shared gender to instead select legislators in accordance with their parties’ stereotypes about gender.

Non-Partisan Condition

Figure 4 displays the AMCE results from the condition of the experiment that did not provide explicit information about legislators’ party affiliations. As in the partisan condition, respondents in the non-partisan condition express a preference of 8.4 and 13.4 percentage points for legislators who have been in Congress for 12 or 20 years over legislators who have been in Congress for two years, providing additional support for Hypothesis 4. While respondents also express a strong bias against members of the Agriculture Committee compared to members of the Judiciary Committee (8 percentage points), they no longer show a preference for members of the Budget Committee. Similarly, they no longer show a preference for seeking help from female legislators, nor do they display preferences based on the legislator’s chamber, former occupation, or staff member race.

In Figure 5, I show the results of the non-partisan condition separated by respondent partisanship. Both Democrats (5.7 percentage points) and Republicans (10 percentage points) show a bias against members of the Agriculture Committee, although this effect is nearly twice as large for Republicans. Republicans also show a nearly significant bias of 4.2 percentage points against former public school teachers, while Democrats show no clear preferences with respect to legislators’ former occupations.

Also evident in these results is a gender bias that is stronger and more symmetric than that found in the partisan condition of the experiment. In accordance with their parties’ respective stereotypes about gender, Democrats are 7.5 percentage points less likely to contact male legislators for help, while Republicans are 5.4 percentage points more likely to contact male legislators. These results accord with my expectations in Hypothesis 3a. Analyses of shared gender by party reveal that Republican men and Democratic women are 11.4 percentage points more likely and 11.5 percentage points less likely, respectively, to select male legislators, while null effects exist for Democratic men and Republican women. Thus, the preferences of Democratic women and Republican men are driving partisans’ apparent gender preferences, and respondents’ partisanship tempers the role of shared gender in the absence of a partisan cue. I explore the mechanisms that may be underlying these results in the next section.

There is little evidence to suggest that shared race influences respondents’ selection of legislators when a partisan cue is absent. Figure A3 in the Online Appendix suggests that Black Democrats are slightly more likely to select legislators with Black staff members and Asian American Democrats are more likely to select legislators with non-white staff members, but these effects are not significant. Similarly, none of the effects for shared race are significant for Republicans. Although Asian American Republicans appear to be less likely to select legislators with Asian American and Black staff members, these results are based on just seven respondents and therefore must be interpreted with caution. The race of legislators’ staff members, in addition to not cueing a sense of shared racial identity among respondents, is also not functioning as a substitute for partisan cues. These results do not provide support for Hypothesis 3b.

Mechanisms and Robustness Checks

To assess the mechanisms through which shared partisanship, race, and gender operate, I rely on post-experiment questions that asked respondents to assess how well four statements—“They are someone like me,” “They share my beliefs and values,” “They seem like someone I could trust,” and “They have the expertise required to solve the problem”—described the legislators they chose to contact for help. I also examine responses to a question asking how confident respondents were that they would actually receive the assistance they requested. I examine how selecting a greater number of legislators with the same partisanship, race and ethnicity, or gender as the respondent affects respondents’ evaluations of the legislators they selected. I first turn to the results for shared partisanship, which are displayed in Table 1.

Table 1.

Shared Partisanship and Respondents’ Post-experiment Assessment of Chosen Legislators.

	Dependent variable
	Values	Like Me	Trust	Expertise	Confident
Number of same-party	−0.006	−0.010	−0.005	−0.100^*	−0.088
profiles selected	(0.057)	(0.053)	(0.054)	(0.056)	(0.054)

Note. ^*p < .1; ^**p < .05; ^***p < .01.

If shared partisanship were operating through any of these four mechanisms, then respondents who selected a greater number of co-partisan legislators would provide positive evaluations of their selected legislators using these criteria. However, the ordered logit models reveal no significant relationship between the number of co-partisan legislators that respondents selected and their post-experiment evaluations of their selected legislators. Shared partisanship, while clearly an important factor in constituents’ choices, does not necessarily mean that constituents feel closer to these legislators or evaluate them more positively, suggesting that negative partisanship, rather than positive in-party affect, may actually be driving this result. To account for the possibility that respondents used only the most recent legislator profile to form their evaluations, I conducted separate analyses with an independent variable indicating whether the final profile that the respondent chose was a legislator who shared their partisanship. Like the count variable presented here, the effects of this variable on respondents’ evaluations of their chosen legislators, shown in Table A3 in the online appendix, are not statistically significant.⁶

Although shared race has no relationship with constituents’ selections of legislators in the conjoint experiment, it is possible that receiving assistance from a staff member of the same race may leave constituents with more positive impressions of the legislator. To explore this possibility, I also conducted analyses of these mechanisms for shared race and ethnicity, the results of which are shown in Table 2. Despite the null effects found for shared race and ethnicity in the AMCE analyses, the number of shared race and ethnicity profiles that respondents selected is related to four of their post-experiment legislator evaluations. In the partisan condition, respondents who selected more legislator profiles with staff members who shared their race or ethnicity were more likely to see those legislators as sharing their values, being similar to them, and possessing more expertise. Respondents were also more likely to feel confident that they would receive the assistance that they requested. These results do not hold, however, for the non-partisan condition: although the signs of the coefficients in the models for shared values, trust, expertise, and confidence are in the expected direction, the coefficients are small and statistically insignificant. Even though shared race is not operating as a substitute for shared partisanship when respondents lack information about legislators’ partisanship, this analysis suggests that contact with staff members of the same race through constituent service can positively affect constituents’ evaluations of their legislators.

Table 2.

Shared Race and Respondents’ Post-experiment Assessment of Chosen Legislators.

	Dependent variable
	Values	Like Me	Trust	Expertise	Confident
Number of same-race profiles	0.126^**	0.120^**	0.084	0.182^***	0.166^***
selected (with partisan cue)	(0.057)	(0.054)	(0.056)	(0.056)	(0.058)
Number of same-race profiles	0.029	−0.012	0.043	0.050	0.025
selected (without partisan cue)	(0.052)	(0.057)	(0.057)	(0.055)	(0.058)

Note. ^*p < .1; ^**p < .05; ^***p < .01.

Table 3 presents the results from ordered logit models assessing the relationship between the number of same-gender legislators that respondents selected and their post-experiment evaluations of their selected legislators. Respondents in the partisan condition who selected a greater number of same-gender legislators viewed those legislators as possessing more expertise (p < .05) and as being more trustworthy and similar to themselves (p < .1). However, in the non-partisan condition, respondents who selected more same-gender legislators viewed the legislators they chose as being more likely to share their values (p < .05) in addition to being more trustworthy (p < .1). Thus, while shared gender seems to function as a partisan cue by signaling shared values to constituents, it also functions as an important piece of information on its own by signaling expertise, similarity, and trust.

Table 3.

Shared Gender and Respondents’ Post-experiment Assessment of Chosen Legislators.

	Dependent variable
	Values	Like Me	Trust	Expertise	Confident
Number of same-gender profiles	0.033	0.087^*	0.086^*	0.107^**	0.044
selected (with partisan cue)	(0.049)	(0.049)	(0.050)	(0.050)	(0.052)
Number of same-gender profiles	0.139^***	0.087^*	0.010	0.030	0.051
selected (without partisan cue)	(0.049)	(0.049)	(0.052)	(0.049)	(0.052)

Note. ^*p < .1; ^**p < .05; ^***p < .01.

Finally, as a robustness check on the external validity of the experimental design, I conducted additional analyses using only respondents from states with eight or fewer congressional districts. Lee and Oppenheimer (1999, p. 62) find that residents of smaller states have more contact with their senators and are more likely to view them as “two extra House members,” and so it may be more reasonable to assume that these constituents would actively choose among their three federal legislators when seeking assistance. Although my sample does not allow me to conduct the analyses for shared race and gender on this subgroup because of the extremely small sample size, my main findings about shared partisanship, including the null results for the shared partisanship mechanisms, remain among constituents from smaller states, suggesting that there are no major differences between the decision processes of constituents from smaller and larger states, despite the fact that they may perceive the accessibility of their legislators differently. The results of these additional analyses are shown in Tables A4 and A5 and Figures A4 and A5 in the Online Appendix.

Discussion and Conclusion

The results of the conjoint experiment suggest that, while constituents value shared partisanship in interactions with members of Congress involving constituent service, other factors are also relevant for constituents’ decisions. As I hypothesized, respondents are more likely to select legislators who share their partisanship. These results accord with those of Broockman and Ryan (2016) and are a clear departure from decades of research assuming that incumbent members of Congress can build a “personal vote” independent of partisanship by providing service to their constituents. Rather than contacting their incumbent House member, regardless of their partisanship, to ask for help with a problem, constituents who are represented by House members and senators of different parties may “shop around” and deliberately seek help from a legislator who shares their partisanship.

However, shared partisanship does not appear to be working through any of the hypothesized mechanisms of shared values, perceived similarity, trust, or perceived expertise. Respondents who selected a greater number of co-partisan legislators were no more likely to evaluate their selected legislators positively on these four counts; in fact, the effects, while not statistically significant, are negative, contrary to my expectations. Because respondents’ choosing to contact co-partisan legislators does not appear to produce positive evaluations of the legislators that they selected, my results raise the possibility that negative partisanship, rather than positive in-party affect, is responsible for the co-partisan affinity that is apparent here. Partisan respondents may be driven more by a desire to avoid contact with legislators who do not share their party than by a positive preference for legislators of their own party.

Other results suggest that constituents also value expertise in their legislators when contacting them for assistance. Constituents may value expertise that is specific to the context of the bureaucratic issue or more general expertise that speaks to other representational priorities. Respondents in the partisan condition of the experiment prefer legislators with a seat on the Budget Committee, perhaps because respondents assume that the work of the Budget Committee is related to the IRS and other tax matters, which was the subject of the vignette in the experiment. And although consistent with previous research showing that legislators’ seniority is positively correlated with their casework loads (Johannes, 1980), it is notable that legislators’ seniority exerts such a strong effect on constituents’ preferences in this context, especially because legislators who have had more time in office to build their “home styles” and establish connections with constituents (Fenno, 1978) are least likely to need the electoral benefits of constituent service. Respondents’ preferences for more senior legislators also offer a surprising contrast with public support for term limits and age limits for members of Congress, which stand at 87% and 79%, respectively, as of July 2023 (Pew Research Center, 2023). Despite calls for these reforms, constituents of both parties appear to view longevity in Congress and the expertise that it implies as assets when seeking assistance from legislators. Scholars should continue to explore why and under what circumstances constituents may prefer representation from legislators who have spent more time in office.

Legislators’ gender also serves as an important cue for respondents, particularly when information about legislators’ partisanship is not present. Without information about partisanship, Democrats opt to contact women and Republicans opt to contact men, and Democrats prefer to ask for assistance from women legislators even in the presence of a partisan cue. Shared gender, however, has an additional effect on respondents’ choices, with Democratic women driving the preference for female legislators among Democrats and Republican men driving the preference for male legislators among Republicans. Democratic men and Republican women are cross-pressured by their gender and partisanship such that they express no clear gender preferences in their selection of legislators. These results accord with evidence from observational research showing that shared partisanship between constituents and legislators matters more than shared gender for encouraging political engagement among constituents (Lawless, 2004). However, they also speak to previous studies about the relationship between partisanship and gender stereotyping (Sanbonmatsu & Dolan, 2009) and partisan cues and gender affinity (Badas & Stauffer, 2019) and suggest that these processes also occur in the context of constituent service. While partisans use legislators’ gender to infer partisanship, gender affinity appears to be layered on top of these partisan concerns, suggesting that shared gender may be particularly important to constituents when seeking assistance from their legislators.

In contrast to gender, the race and ethnicity of legislators’ staff members do not appear to be cueing a sense of shared identity among respondents with the same race or ethnicity. Shared race is, however, related to constituents’ evaluations of the legislators in the partisan condition of the experiment. While other factors may be more important determinants of constituents’ initial choices of contact, constituents appear to evaluate their legislators more favorably if the legislator’s constituent service staff share the constituent’s race or ethnicity. These results speak to the important role that staff members play in the provision of constituent service, given that they are the individuals who actually interact with constituents in this context. While the results of the experiment do not provide evidence that constituents might go out of their way to reach out to a legislator with staff members who share the constituent’s race or ethnicity, the mechanism analyses suggest that constituents may ultimately have more positive experiences with constituent service when this is the case.

Respondents’ prioritization of legislators’ partisanship has the potential to diminish the role that staff members play in constituents’ experiences with constituent service. Even as shared partisanship emerges as the most dominant factor in constituents’ choices of legislators to contact for assistance, the lack of a relationship between these choices and constituents’ positive impressions of legislators raises doubts about legislators’ and staffers’ abilities to gain the trust of their constituents through constituent service. That shared partisanship does not appear to have any bearing on constituents’ evaluations of legislators in the context of constituent service suggests that legislators and their offices may need to work to build trust when providing constituent service even to constituents who share their partisanship and support them otherwise. More broadly, these findings suggest that negative partisanship has permeated even this ostensibly nonpartisan element of representation.

While these findings clearly suggest that constituents’ preferences for co-partisan legislators have the potential to mute the impact that constituent service has on the personal vote, constituents’ choices of which legislators to contact for help have two additional implications for constituent service and representation. First, constituents’ grounding their contact decisions in partisanship may exacerbate the already uneven demand for constituent service that exists across legislators (e.g. Johannes, 1980), particularly within state delegations. For instance, constituents whose representative in the House does not share their partisanship may instead opt to contact one of their senators for help if one of these senators shares their partisanship. Thus, senators who represent states in which their party also dominates among congressional districts may be particularly unlikely to receive requests for help, while senators who are members of split-party delegations in states that are more evenly divided or dominated by the other party may receive an unusually high volume of requests from constituents of their party. Such patterns of contact, of course, do not preclude constituents from reaching out to cross-partisan legislators to share their opinions directly, but they would limit the nature of these interactions to expressions of disapproval on matters related to policy.

Second, when exercised, constituents’ preferences for seeking help from co-partisan legislators can skew the pool of constituent voices from whom legislative offices hear. Congressional offices use interactions with constituents to determine constituent opinion and needs (Abernathy, 2018; Hertel-Fernandez et al., 2019; Miler, 2010), and if legislators do not hear from cross-party constituents in the relatively neutral context of constituent service, their views of these constituents may only be informed by more negative interactions involving constituents’ expressions of policy preferences. By refusing to ask for help from legislators who do not share their partisanship, it is possible that constituents could be inadvertently denying themselves the opportunity to convey their needs to their legislators.

These results point to two useful areas of study for future research. First, scholars should continue to build on Broockman and Ryan (2016) to explore how partisanship may affect earlier stages of the constituent-legislator contact process when constituents are seeking assistance from their legislators, particularly if constituents who are not represented by a co-partisan legislator may forgo asking for help at all. Second, future research should explore constituents’ understandings of constituent service and their expectations of what occurs when they reach out to legislators for help. It is striking that constituents’ preferences are driven so strongly by legislator characteristics that should have little bearing on the legislators’ provisions of constituent service. How, if at all, do constituents believe that partisanship affects the type or quality of service that legislators provide? To what degree do constituents believe staff are involved in constituent service? How might other characteristics of legislators, such as their committee assignments, occupational backgrounds, and other experience in Congress, actually affect their orientations toward constituent service? In other words, what makes a constituent service operation both effective and broadly appealing? Addressing these questions would allow scholars of representation to understand the full extent of partisanship’s impact on constituent service.

Supplemental Material

Supplemental Material - The Partisan Vote: Shared Partisanship and Constituent Service in the U.S. Congress

Supplemental Material for The Partisan Vote: Shared Partisanship and Constituent Service in the U.S. Congress by Rochelle Snyder in American Politics Research

Footnotes

Acknowlegdments

The author thanks Barry Burden, Eleanor Powell, David Canon, Kathy Cramer, Levi Bankston, Mike DeCrescenzo, Jessie Munson, Matthew Shor, participants of the UW-Madison American Politics Workshop and Experimental Politics Workshop, and the anonymous reviewers for comments.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research is supported by funding from the UW-Madison Department of Political Science Summer Initiative Program and has been approved by the Institutional Review Board at the University of Wisconsin-Madison.

ORCID iD

Rochelle Snyder

Supplemental Material

Supplemental material for this article is available online.

Notes

References

Abernathy

(2018). Understanding and responding to constituent opinion on Capitol Hill. Congress and the Presidency, 45(3), 254–278. https://doi.org/10.1080/07343469.2018.1482385

Abramowitz

A. I.

Webster

(2016). The rise of negative partisanship and the nationalization of US elections in the 21st century. Electoral Studies, 41, 12–22. https://doi.org/10.1016/j.electstud.2015.11.001

Abramowitz

A. I.

Webster

S. W.

(2018). Negative partisanship: Why Americans dislike parties but behave like rabid partisans. Political Psychology, 39(S1), 119–135. https://doi.org/10.1111/pops.12479

Abramson

S. F.

Koçak

Magazinnik

(2022). What do we learn about voter preferences from conjoint experiments? American Journal of Political Science, 66(4), 1008–1020. https://doi.org/10.1111/ajps.12714

Ansolabehere

Schaffner

B. F.

(2017). CCES common content , 2016 . Harvard University. https://doi.org/10.7910/DVN/GDF6Z0

Atkeson

L. R.

Hamel

B. T.

(2020). Fit for the job: Candidate qualifications and vote choice in low information elections. Political Behavior, 42(1), 59–82. https://doi.org/10.1007/s11109-018-9486-0

Badas

Stauffer

K. E.

(2019). Voting for women in nonpartisan and partisan elections. Electoral Studies, 57, 245–255. https://doi.org/10.1016/j.electstud.2018.10.004

Bartels

L. M.

(2000). Partisanship and voting behavior, 1952-1996. American Journal of Political Science, 44(1), 35–50. https://doi.org/10.2307/2669291

Broockman

D. E.

(2013). Black politicians are more intrinsically motivated to advance blacks’ interests: A field experiment manipulating political incentives. American Journal of Political Science, 57(3), 521–536. https://doi.org/10.1111/ajps.12018

10.

Broockman

D. E.

(2014). Distorted communication, unequal representation: Constituents communicate less to representatives not of their race. American Journal of Political Science, 58(2), 307–321. https://doi.org/10.1111/ajps.12068

11.

Broockman

D. E.

Ryan

T. J.

(2016). Preaching to the choir: Americans prefer communicating to copartisan elected officials. American Journal of Political Science, 60(4), 1093–1107. https://doi.org/10.1111/ajps.12228

12.

Burden

B. C.

(2007). Personal roots of representation. Princeton University Press.

13.

Butler

D. M.

Broockman

D. E.

(2011). Do politicians racially discriminate against constituents? A field experiment on state legislators. American Journal of Political Science, 55(3), 463–477. https://doi.org/10.1111/j.1540-5907.2011.00515.x

14.

Cain

Ferejohn

Morris

(1987). The personal vote: Constituency service and electoral independence. Harvard University Press.

15.

Campbell

Cowley

(2014). What voters want: Reactions to candidate characteristics in a survey experiment. Political Studies, 62(4), 745–765. https://doi.org/10.1111/1467-9248.12048

16.

Carnes

Lupu

(2016). Do voters dislike working-class candidates? Voter biases and the descriptive underrepresentation of the working class. American Political Science Review, 110(4), 832–844. https://doi.org/10.1017/s0003055416000551

17.

Comenetz

(2016). Frequently occurring surnames in the 2010 Census. United States Census Bureau.

18.

Coppock

McClellan

O. A.

(2019). Validating the demographic, political, psychological, and experimental results obtained from a new source of online survey respondents. Research & Politics, 6(1), Article 205316801882217. https://doi.org/10.1177/2053168018822174

19.

Costa

(2021a). He said, she said: The gender double bind in legislator-constituent communication. Politics and Gender, 17(4), 528–551. https://doi.org/10.1017/s1743923x19000862

20.

Costa

(2021b). Ideology, not affect: What Americans want from political representation. American Journal of Political Science, 65(2), 342–358. https://doi.org/10.1111/ajps.12571

21.

Costa

Schaffner

B. F.

(2018). How gender conditions the way citizens evaluate and engage with their representatives. Political Research Quarterly, 71(1), 46–58. https://doi.org/10.1177/1065912917722235

22.

Dolan

(2004). The impact of candidate sex on evaluations of candidates for the US House of Representatives. Social Science Quarterly, 85(1), 206–217. https://doi.org/10.1111/j.0038-4941.2004.08501015.x

23.

Ellis

W. C.

Wilson

W. C.

(2013). Minority chairs and congressional attention to minority issues: The effect of descriptive representation in positions of institutional power. Social Science Quarterly, 94(5), 1207–1221. https://doi.org/10.1111/ssqu.12023

24.

English

Pearson

Strolovitch

D. Z.

(2019). Who represents me? Race, gender, partisan congruence, and representational alternatives in a polarized America. Political Research Quarterly, 72(4), 785–804. https://doi.org/10.1177/1065912918806048

25.

Fenno

R. F.

(1978). Home style: Representatives in their districts. Little Brown & Co.

26.

Fiorina

M. P.

(1977). The case of the vanishing marginals: The bureaucracy did it. American Political Science Review, 71(1), 177–181. https://doi.org/10.2307/1956961

27.

Frederick

(2008). Constituency population and representation in the U.S. House. American Politics Research, 36(3), 358–381. https://doi.org/10.1177/1532673x07309740

28.

Gamble

K. L.

(2007). Black political representation: An examination of legislative activity within US House committees. Legislative Studies Quarterly, 32(3), 421–447. https://doi.org/10.3162/036298007781699663

29.

Gay

(2002). Spirals of trust? The effect of descriptive representation on the relationship between citizens and their government. American Journal of Political Science, 46(4), 717–732. https://doi.org/10.2307/3088429

30.

Grose

C. R.

Mangum

Martin

(2007). Race, political empowerment, and constituency service: Descriptive representation and the hiring of African- American congressional staff. Polity, 39(4), 449–478. https://doi.org/10.1057/palgrave.polity.2300081

31.

Hainmueller

Hopkins

Yamamoto

(2014a). cjoint: Causal inference in conjoint analysis: Understanding multi-dimensional choices via stated preference experiments, volume 22 number 1. R package. version 2.1.0. https://CRAN.R-project.org/package=cjoint

32.

Hainmueller

Hopkins

D. J.

Yamamoto

(2014b). Causal inference in conjoint analysis: Understanding multidimensional choices via stated preference experiments. Political Analysis, 22(1), 1–30. https://doi.org/10.1093/pan/mpt024

33.

Heberlig

E. S.

(2003). Congressional parties, fundraising, and committee ambition. Political Research Quarterly, 56(2), 151–161. https://doi.org/10.2307/3219894

34.

Herrera

Yawn

(1999). The emergence of the personal vote. The Journal of Politics, 61(1), 136–150. https://doi.org/10.2307/2647778

35.

Hertel-Fernandez

Mildenberger

Stokes

L. C.

(2019). Legislative staff and representation in Congress. American Political Science Review, 113(1), 1–18. https://doi.org/10.1017/s0003055418000606

36.

Huddy

Mason

Aarøe

(2015). Expressive partisanship: Campaign involvement, political emotion, and partisan identity. American Political Science Review, 109(1), 1–17. https://doi.org/10.1017/s0003055414000604

37.

Iyengar

Lelkes

Levendusky

Malhotra

Westwood

S. J.

(2019). The origins and consequences of affective polarization in the United States. Annual Review of Political Science, 22(1), 129–146. https://doi.org/10.1146/annurev-polisci-051117-073034

38.

Iyengar

Sood

Lelkes

(2012). Affect, not ideology: A social identity perspective on polarization. Public Opinion Quarterly, 76(3), 405–431. https://doi.org/10.1093/poq/nfs038

39.

Jacobson

G. C.

(2015). It’s nothing personal: The decline of the incumbency advantage in US House elections. The Journal of Politics, 77(3), 861–873. https://doi.org/10.1086/681670

40.

Johannes

J. R.

(1980). The distribution of casework in the US Congress: An uneven burden. Legislative Studies Quarterly, 5(4), 517–544. https://doi.org/10.2307/439572

41.

Johannes

J. R.

McAdams

J. C.

(1981). The congressional incumbency effect: Is it casework, policy compatibility, or something else? An examination of the 1978 election. American Journal of Political Science, 25(3), 512–542. https://doi.org/10.2307/2110817

42.

Kaslovsky

(2022). Senators at home: Local attentiveness and policy representation in Congress. American Political Science Review, 116(2), 645–661. https://doi.org/10.1017/s0003055421001088

43.

Kirkland

P. A.

Coppock

(2018). Candidate choice without party labels. Political Behavior, 40(3), 571–591. https://doi.org/10.1007/s11109-017-9414-8

44.

Lawless

J. L.

(2004). Politics of presence? Congresswomen and symbolic representation. Political Research Quarterly, 57(1), 81–99. https://doi.org/10.1177/106591290405700107

45.

Lee

F. E.

Oppenheimer

B. I.

(1999). Sizing up the Senate: The unequal consequences of equal representation. University of Chicago Press.

46.

Lowande

Ritchie

Lauterbach

(2019). Descriptive and substantive representation in Congress: Evidence from 80,000 congressional inquiries. American Journal of Political Science, 63(3), 644–659. https://doi.org/10.1111/ajps.12443

47.

Mansbridge

(1999). Should blacks represent blacks and women represent women? A contingent ‘yes. The Journal of Politics, 61(3), 628–657. https://doi.org/10.2307/2647821

48.

Mason

(2016). A cross-cutting calm: How social sorting drives affective polarization. Public Opinion Quarterly, 80(S1), 351–377. https://doi.org/10.1093/poq/nfw001

49.

Mason

(2018). Uncivil agreement: How politics became our identity. University of Chicago Press.

50.

Matson

Fine

T. S.

(2006). Gender, ethnicity, and ballot information: Ballot cues in low-information elections. State Politics and Policy Quarterly, 6(1), 49–72. https://doi.org/10.1177/153244000600600103

51.

McDermott

M. L.

(1998). Race and gender cues in low-information elections. Political Research Quarterly, 51(4), 895–918. https://doi.org/10.2307/449110

52.

McDermott

M. L.

(2005). Candidate occupations and voter information shortcuts. The Journal of Politics, 67(1), 201–219. https://doi.org/10.1111/j.1468-2508.2005.00314.x

53.

Miler

K. C.

(2010). Constituency representation in congress: The view from capitol hill. Cambridge University Press.

54.

Orr

L. V.

Fowler

Huber

G. A.

(2023). Is affective polarization driven by identity, loyalty, or substance? American Journal of Political Science, 67(4), 948–962. https://doi.org/10.1111/ajps.12796

55.

Pew Research Center . (2023). American’s dismal views of the nation’s politics. URL: https://www.pewresearch.org/politics/2023/09/19/americans-dismal-views-of-the-nations-politics/

56.

Romero

D. W.

Sanders

(1994). Loosened partisan attachments and receptivity to incumbent behaviors: A panel analysis, 1972-1976. Political Research Quarterly, 47(1), 177–192. https://doi.org/10.1177/106591299404700109

57.

Sanbonmatsu

Dolan

(2009). Do gender stereotypes transcend party? Political Research Quarterly, 62(3), 485–494. https://doi.org/10.1177/1065912908322416

58.

Sen

(2017). How political signals affect public support for judicial nominations: Evidence from a conjoint experiment. Political Research Quarterly, 70(2), 374–393. https://doi.org/10.1177/1065912917695229

59.

Serra

Moon

(1994). Casework, issue positions, and voting in congressional elections: A district analysis. The Journal of Politics, 56(1), 200–213. https://doi.org/10.2307/2132353

60.

Serra

Pinney

(2004). Casework, issues and voting in state legislative elections: A district analysis. Journal of Legislative Studies, 10(4), 32–46. https://doi.org/10.1080/1357233042000322472

61.

Sung

Y.-H.

(2022). How U.S. Voters elect prosecutors: Evidence from a conjoint experiment. Political Research Quarterly.

62.

Teele

D. L.

Kalla

Rosenbluth

(2018). The ties that double bind: Social roles and women’s underrepresentation in politics. American Political Science Review, 112(3), 525–541. https://doi.org/10.1017/s0003055418000217

63.

Tomz

M. R.

Weeks

J. L. P.

(2013). Public opinion and the democratic peace. American Political Science Review, 107(4), 849–865. https://doi.org/10.1017/s0003055413000488

64.

Tzioumis

(2018). Demographic aspects of first names. Scientific Data, 5(1), Article 180025. https://doi.org/10.1038/sdata.2018.25

65.

University of Michigan and Stanford University . (2017). American National Election Studies (ANES) cumulative data file. URL: https://electionstudies.org

66.

Wagner

M. W.

(2007). Beyond policy representation in the US House: Partisanship, polarization, and citizens’ attitudes about casework. American Politics Research, 35(6), 771–789. https://doi.org/10.1177/1532673x07299867

67.

Webster

S. W.

Abramowitz

A. I.

(2017). The ideological foundations of affective polarization in the US electorate. American Politics Research, 45(4), 621–647. https://doi.org/10.1177/1532673x17703132

68.

Yiannakis

D. E.

(1981). The grateful electorate: Casework and congressional elections. American Journal of Political Science, 25(3), 568–580. https://doi.org/10.2307/2110819

69.

Ziniel

C. E.

(2009). Faces in the office: Racial employment segregation among congressional staff. Social Science Research Network. Available at SSRN: https://ssrn.com/abstract=1434290 or https://doi.org/10.2139/ssrn.1434290

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.95 MB