What Do Citizens Want from Their Member of Congress?

Abstract

What do citizens want from their members of Congress? Do they expect them to be constituent servants? Do they expect them to work on local problems? Or do they expect them to represent them on the national issues of the day? While citizens expect members of Congress to perform all of these roles, we argue that, in the contemporary political environment, citizens especially value members who represent them on the salient national issues of the day. We also argue that such behavior will be especially pronounced among those who are the most educated and partisan. We show, using several recent nationally representative surveys, that citizens prioritize this sort of issue representation, and that such evaluations shape member approval and vote choice. We conclude by discussing the implications of this pattern for related trends such as elite polarization and the nationalization of elections.

Keywords

Congress representation constituency service

What do ordinary citizens want from their member of Congress? Do they want someone who will go to Washington to serve the district, passing bills to help the local area, performing constituency service, and bringing back federal projects and monies to the district? Do they want someone who will go and serve as a watchdog over federal agencies? Or do they want someone who represents their views on the key national issues of the day? We know a great deal about how members think about balancing these various aspects of their job, dating back to foundational works in the 1970s and 1980s (e.g., Cain, Ferejohn, and Fiorina 1987; Fenno 1978; Mayhew 1974). While this “supply side” of the problem is well studied, the demand side—what constituents want—is somewhat less so (though see Griffin and Flavin 2011; Harden 2016).

We take up the question of what citizens want from their members. This is particularly timely given that Congressional approval is near an all-time low: since 2010, Gallup’s Congressional approval has hovered in the teens, and occasionally dipped below 10 percent.¹ For the public to restore its faith in Congress, an important first step is to better understand what constituents want from their individual representatives.² We argue that while voters want members to do all of the jobs above (e.g., help the district, oversee government, and represent them on the issues), in the contemporary environment, citizens especially prioritize a member who reflects their views on the issues, particularly salient national issues. Given elite polarization, and beliefs about how members will behave in D.C., voters see issue representation as perhaps the core function of a member. To be clear, this is not to say that they do not value service to the district or other functions: our data show clearly that they do. Instead, our argument is that most voters—and especially the most educated and partisan—value issue representation in the contemporary context.³ In turn, our findings have important implications for understanding the related changes in Congress and American politics more broadly, such as the nationalization of elections and increasing levels of polarization in Congress.

What Do Voters Want?

Representation is a multi-dimensional affair (Eulau and Karps 1977). Members can not only represent their district on the issues of the day (Doherty 2013; Miller and Stokes 1963), but they can also work to serve the district in other ways, by (say) bringing federal monies to the district or by providing constituency services (Cain, Ferejohn, and Fiorina 1987) or by regulating the bureaucracy or providing symbolic representation.Indeed, there are many different “spheres of representation” (Griffin and Flavin 2011), and different legislators prioritize these spheres differently (Harden 2013).

While we have a strong sense of how legislators perceive the costs and benefits of emphasizing these different dimensions (Harden 2013), we have a less well-developed sense of how voters evaluate these trade-offs (though see Griffin and Flavin 2011; Parker and Goodman 2009). In particular, we do not know how voters weigh the relative importance of having a member who represents them on the issues versus one who provides constituency service.⁴ Undoubtedly, voters want members both to serve the district and represent them on the issues, but there is a question of degree: on which factor do voters put more emphasis?

We argue that in the current political environment, voters will, on average, put more weight on issue representation over service, both when evaluating a member generally and at the ballot box. This stems from three related aspects of contemporary American politics. First, taking credit for bringing federal money to the district is more difficult and potentially more politically risky in the era of the Tea Party and its focus on federal debt and spending. Traditionally, members saw bringing federal spending to their district as a key nonpartisan benefit they could provide to their constituents (Cain, Ferejohn, and Fiorina 1987). But after Congress responded to outside pressure by banning earmarks in January 2011, particularistic spending became more difficult, though certain types have remained possible (Joachim 2014). Voters are much less positively predisposed (at least in the abstract) toward federal spending, and as a result, members are far less likely to claim credit for spending in their districts (Grimmer, Westwood, and Messing 2014, chap. 7).

Second, voters have come to see the provision of district services as mechanistic. That is, voters do not necessarily think that there is a benefit having a member from one party or the other in terms of providing services to them or in helping the district more generally (Lockerbie 1999). Providing services to the district—helping someone with a lost social security check or helping to fund a community center—are simply part of the job of the member, and anyone in the office, regardless of party or expertise, would carry it out.

Finally, and most importantly, there is the dramatic polarization of the parties in Congress (McCarty, Poole, and Rosenthal 2006). Today the elite party differences are as sharp as they have been in nearly a century, and the policies supported by a Democrat and a Republican in the same district are quite distinct (Lee, Moretti, and Butler 2004). In large part stemming from this polarization, party has returned to the forefront of voters’ decision-making calculuses in a way not seen since midcentury (Bartels 2000). Indeed, party has come to replace incumbency as the key heuristic for congressional voting (Jacobson 2015a). But this partisan loyalty is not simply blind loyalty to a party. Voters have increasingly aligned their issue positions with their partisanship (Levendusky 2009). For many ordinary voters today, then, to elect a member from their party is to elect someone who shares their views across a range of issues.⁵

Given these factors, we argue that constituency service, while still important, is somewhat less important to voters today than is issue representation. Selecting a member from one party or the other gives voters a real choice because of the vastly different policy positions of the parties. So when voters evaluate what they want from a member, especially at the ballot box, they will emphasize issue considerations. Again, to be clear, our argument is not that service is unimportant or not valued by citizens: our data below show that voters still value it. Our claim is simply that, in today’s political environment, issue considerations are more prominent in voters’ minds and decision-making processes.

Such a desire for issue-based representation should be especially pronounced for two groups of voters. First, it should be stronger among those who are sorted partisans—those whose ideological self-identification matches their partisanship (i.e., liberal Democrats and conservative Republicans; see Levendusky 2009). While many voters have only vague and somewhat inchoate views on the issues, that is much less true of sorted voters, who have more developed and consistent preferences (Jewitt and Goren 2016). As a result, even relative to most voters, they are much more poorly represented on the issues by a member from the other party, and they therefore prioritize issue/partisan representation by their member of Congress even more heavily.

Second, those who are better educated should also have a stronger preference for issue representation. Education gives citizens the cognitive skills to be politically active, and to hold solid issue positions that can drive behavior (Nie, Junn, and Stehlik-Berry 1996). Better educated voters will have stronger, more established issue positions, and they will want to select a member who represents those beliefs.

Data and Analyses

Our data to test these hypotheses come from two recent, large-scale nationally representative surveys: The American Panel Survey (TAPS), conducted by GfK/Knowledge Networks for the Weidenbaum Center at Washington University, and the 2014 Annenberg Institutions of Democracy (AIOD) Congress Survey, conducted by SRBI, Inc. for the Annenberg Public Policy Center at the University of Pennsylvania. TAPS is a monthly panel study of approximately two thousand households conducted between November 2011 and December 2013 (our data come from August and November 2013).⁶ The AIOD study is a rolling cross-sectional design with 21,185 total interviews⁷ and was fielded between June 2 and July 16, 2014 (excluding July 4), and between September 2 and November 3, 2014, with an average of 198 interviews per day over the sample period. All analyses with both datasets are weighted to make the data nationally representative. Furthermore, given the rolling cross-sectional nature of the AIOD data, analyses with that dataset are presented with week fixed effects to account for changes during the survey period.

The first part of our argument is that the public has come to expect members to be issue representatives more than providers of constituency services. To examine this claim, we use an item from the AIOD study that asked respondents to rate how much several factors mattered to their vote choice: constituency service, how well the candidate represents them on the issues, partisan control of Congress, and bringing projects back to the district. Respondents could say that these factors mattered “a great deal,” “somewhat,” “not too much,” or “not at all.” These ratings allow us to assess how respondents prioritize various dimensions of a member’s job at the ballot box. We also use a second item from the TAPS study that asks respondents to rate how important various aspects of a member of Congress’s job are: keeping in contact with constituents, working on national issues (even if they do not matter much locally), working on local issues that matter to the district but not the nation, oversight of government agencies, bringing federal monies and projects to the district, and constituency service; respondents could say these were “very important,” “somewhat important,” or “unimportant.” While the surveys ask different items with slightly different response options, they both get at a similar question: what do voters value in a member? Table 1 gives the breakdown of both sets of items (the TAPS item is in the top panel, the AIOD item is in the bottom panel).

Table 1.

The Factors That Matter to Voters When Selecting a Member of Congress.

Characteristic	Very important	Somewhat important	Unimportant
TAPS data
Keeping in touch with constituents	76	23	1
Working on national issues	57	39	4
Working on local issues	51	44	5
Bringing federal monies to the district	50	44	6
Oversight of agencies	55	40	5
Constituency service	32	51	17
Factor	Not at all	Not too much	Somewhat	A great deal
AIOD data
Issue positions	11	8	22	59
Constituency service	19	16	33	33
Spending in the district	18	15	35	32
Party control of the chamber	24	14	24	39

The upper panel gives the importance ratings of various aspects of a member’s job from the TAPS data. The lower panel shows respondents’ answers to the question of how important each factor is to their vote from the AIOD data; the cell entries give the percentage selecting each option. TAPS = The American Panel Survey; AIOD = Annenberg Institutions of Democracy.

Table 1 supports our argument that issues are highly important for contemporary voters. If we begin in the top panel with the TAPS data, we see that the most important activity is keeping in contact with constituents. This is outside the bounds of our theoretical argument, but makes good intuitive sense. Citizens want members who keep them informed about what’s happening in D.C., and how they are working for the district, but this says little about what members want in terms of a focus on issues versus service. Looking at the remainder of the table, the next highest rated item is working on national issues, even if those issues are not especially relevant locally. While a majority of voters rate both national and local issues as very important, if we look at the average rating assigned to each item, voters assign a significantly higher rating to national issues (p < .05). So national issues do carry somewhat more weight than local ones. And further, both types of issues are significantly more important than providing constituency service, which ranks lowest in this set of issues. While service is not unimportant—one-third of voters say it is very important—overall it ranks below these other concerns.

We see another manifestation of this pattern in related data from the TAPS study. In another item, respondents were asked if they had contacted a member, and if so, why. While such contact is relatively uncommon (only approximately 13 percent of respondents reported any contact with a member), of those who had contacted them, a policy issue was the dominant rationale. Approximately 12 percent of those who had contacted a member did so with a service request, a quarter requested more information about an issue, while 34 percent did so to request that a member change his or her position on an issue, and 45 percent did so to support or oppose a member’s position on an issue (another 20% requested something else; individuals could select more than 1 option, so the percentages do not sum to 100). So when constituents contact members, they do so considerably more frequently about issues than about service. This is indirect evidence, but it supports our argument about how contemporary voters rank these factors.

In the bottom panel with the AIOD data, we see an even clearer prioritization of issue concerns. Here, when voters evaluate a member of Congress, one factor stood out from the rest: how well the member represented voters on the issues. Nearly 60 percent of respondents said that this mattered a great deal, and only 11 percent said that it did not matter at all (while 8% said that it did not matter too much and 22% said that it mattered somewhat). So, for more than eight in ten citizens, a member’s issue positions are either “a great deal” or “somewhat” important to their vote. Furthermore, note that 39 percent of voters said that party control of Congress also mattered a great deal to their vote, which also indirectly supports our argument about issue-based representation, given the tight party/issue connection in the contemporary era. Consistent with our findings above, service-based considerations are ranked somewhat lower by voters. Overall, then, both datasets support our basic claim that while voters still want members to perform constituency service, in the contemporary political environment, they prioritize issue-based representation.

The data in Table 1 show that citizens value issue representation from members. But another part of our argument is that voters see service as more mechanistic: that is, it does not matter which member is in the office, any member will help them get a lost social security check or interact with an agency. To evaluate this claim, we use an item from the AIOD study that asks respondents which party would do a better job at performing various aspects of the job of a member of Congress: bringing federal money back to the district, performing constituency service, standing up to special interests, saying what she or he truly believes, doing whatever it takes to win reelection, and representing citizens’ personal views on the issues. Respondents could say they thought that a Democratic member would do each task better, that a Republican member could do them better, or that there would be no difference between a Democrat and a Republican. Our expectation was that respondents would perceive the biggest difference between the parties on representing their issue positions. In an era of polarized elite parties, the parties stand for quite different policy platforms, so there should be a large partisan dimension to representation—whereas other tasks are less clearly partisan. Other tasks—especially factors such as bringing money to the district and providing service—should see much more muted differences, with most voters saying it makes no difference which party the member is from. Table 2 assesses which party respondents felt would better perform each task. Here we pool Democrats and Republicans together and analyze whether respondents selected their party, selected the opposite party, or thought there was no difference between the parties. We pool Democratic and Republican respondents here because there were no substantive differences between partisans in their responses to these items.⁸

Table 2.

Perceived Partisan Differences.

Function	Opposite party	No difference	Same party
Representing issue positions	14	26	60
Securing federal spending	13	51	36
Constituency service	8	61	32
Standing up to special interests	9	55	35
Tells voters their true beliefs	8	61	31
Does whatever is needed to win reelection	21	64	14

Source. Annenberg Institutions of Democracy data.

Respondents were asked which party would do a better job at each function of a member of Congress; the cell entries give the percentage who indicated a party or indicated there would be no difference between the parties.

The most striking finding here is that for all but one item, the dominant response is that there would be no difference between the parties. For most voters, the job of a member is largely mechanistic: a member from either party would do it just as well (or depending on your level of cynicism, just as poorly). The only item that breaks this pattern is how well a member represents the respondent’s issue positions: more than 60 percent of respondents think their own party would do a better job. This sheds new light on the findings in Table 1: part of the reason issue positions have become more important is that most voters do not see differences between the parties on other items.

Which Voters Prioritize Issue Positions?

The findings reported in Tables 1 and 2 show that most voters place a large weight on issues when evaluating a member of Congress. Yet, while this is true of most voters, it is equally true that considerable heterogeneity is likely to characterize which voters put more emphasis on a member’s issue positions. In particular, we expect that those who are sorted partisans (liberal Democrats/conservative Republicans), as well as those who are better educated (here, operationalized as having a college degree, which is the most educated third of our sample), should be especially likely to prioritize issue-based representation from their member of Congress.

To test these hypotheses, we turn to another item in the AIOD survey. In that question, respondents were asked to rank order the importance of four factors they could weight when evaluating a member of Congress: the member’s provision constituency service, how well he or she represents them on the issues, the member’s character, and the member’s party. This item is especially nice because it forces respondents to explicitly choose which factor was more important to them—issues or service—in assessing their member of Congress? This helps us avoid a natural limitation of the data shown in Table 1: many voters say that all factors are very important. We focus our attention on the constituency service and issue representation items, as these correspond most closely with our theoretical construct of interest.

We examine three different operationalizations of this ranking item. We analyze the rankings voters give to both services and constituency service, and see if sorted partisans and the better educated assign higher rankings to issues and lower rankings to constituency service. Furthermore, we can also look at which voters rank issue representation more highly than service rankings, and ask whether being a sorted partisan or better educated predicts such a ranking. In the results below, we control for a wide variety of demographic variables that are also likely to be related to preferences for issue representation (e.g., age, gender, income, and so forth) to avoid spuriously attributing effects to sorting or education that are really due to other demographic factors. Table 3 presents the results.⁹

Table 3.

The Voters Who Prioritize Issue Position Representation.

	(1)	(2)	(3)	(4)	(5)	(6)
	Constituency service ranking	Issue position ranking	Issues ranked above constituency service	Constituency service ranking	Issue position ranking	Issues ranked above constituency service
Sorted partisan	−0.255*** (0.0414)	0.153*** (0.0395)	0.525*** (0.0871)	−0.205*** (0.0516)	0.0928* (0.0507)	0.410*** (0.109)
College graduate	−0.114*** (0.0427)	0.143*** (0.0386)	0.274*** (0.0895)	−0.0244 (0.0594)	0.0338 (0.0550)	0.0694 (0.126)
Sorted Partisan × College Graduate				−0.166** (0.0764)	0.200*** (0.0700)	0.398** (0.163)
Tea Party identifier	−0.0972* (0.0556)	−0.0200 (0.0533)	0.0438 (0.116)	−0.0992* (0.0557)	−0.0182 (0.0532)	0.0480 (0.117)
Female	0.115*** (0.0389)	0.0567 (0.0387)	−0.162* (0.0848)	0.116*** (0.0388)	0.0547 (0.0386)	−0.166* (0.0849)
Caucasian	−0.0755 (0.0845)	−0.0782 (0.0805)	−0.00258 (0.188)	−0.0784 (0.0845)	−0.0744 (0.0805)	0.00408 (0.188)
African-American	0.118 (0.102)	−0.226** (0.0981)	−0.499** (0.224)	0.110 (0.102)	−0.217** (0.0980)	−0.482** (0.224)
Asian-American	0.217 (0.155)	−0.260* (0.143)	−0.652* (0.343)	0.200 (0.155)	−0.240* (0.142)	−0.616* (0.344)
Latino	0.111 (0.0767)	−0.0712 (0.0747)	−0.361** (0.169)	0.107 (0.0768)	−0.0662 (0.0744)	−0.350** (0.169)
Middle income	−0.0229 (0.0461)	0.0782* (0.0446)	0.0948 (0.100)	−0.0265 (0.0461)	0.0821* (0.0445)	0.102 (0.100)
High income	−0.228*** (0.0568)	0.171*** (0.0522)	0.407*** (0.120)	−0.227*** (0.0568)	0.169*** (0.0521)	0.405*** (0.120)
Age 30-44	−0.0122 (0.0637)	−0.0907 (0.0613)	−0.0578 (0.141)	−0.0144 (0.0636)	−0.0874 (0.0611)	−0.0517 (0.141)
Age 45-64	0.0495 (0.0560)	−0.191*** (0.0553)	−0.259** (0.126)	0.0479 (0.0559)	−0.188*** (0.0554)	−0.254** (0.126)
65 or older	0.0226 (0.0646)	−0.354*** (0.0615)	−0.449*** (0.136)	0.0217 (0.0645)	−0.352*** (0.0616)	−0.446*** (0.136)
Political interest	−0.0120 (0.00758)	0.00763 (0.00738)	0.0175 (0.0161)	−0.0122 (0.00759)	0.00784 (0.00739)	0.0179 (0.0161)
Constant	3.100*** (0.123)	3.075*** (0.117)	−0.0400 (0.266)	3.084*** (0.124)	3.093*** (0.118)	−0.00396 (0.267)
Observations	3,841	3,855	3,793	3,841	3,855	3,793
R ²	.060	.039	—	.061	.041	—

Source. Annenberg Institutions of Democracy data.

The cell entries are ordinary least squares regression coefficients (columns 1 and 2) or logistic regression coefficients (column 3), with associated standard errors in parentheses. Survey weights are used, and week fixed effects are included.

Coefficients that can be statistically distinguished from 0 are given as follows: *p < .1. **p < .05. ***p < .01.

Column 1 predicts rankings of constituency service (with higher rankings indicating greater importance), column 2 does the same for issue positions, and column 3 predicts which respondents will assign a higher ranking to a member’s issue positions than to his or her constituency service. All three sets of results handsomely confirm our expectations: sorted partisans, and the better educated, have an especially strong preference for issue-based representation.¹⁰ For example, taking model 3 (ranking issue representation above constituency service), sorted partisans are 13 percent more likely to prefer issue representation, and those who are better educated (college graduates) are 6 percent more likely to do so. While voters generally prefer issue-based representation, these groups have an even stronger tendency to do so.¹¹

Columns 4 to 6 of Table 3 give an even further wrinkle to the above. They show that there is a positive and significant interaction effect between sorted partisans and education. This shows that even among the sorted, higher education has a large effect, presumably because it helps even those voters grasp why they value issue-based representation. For example, looking at column 6 (ranking issues above constituency service), the marginal effect of sorting among those without a college degree is approximately 10 percentage points (a change in predicted probability from 49% to 59%). But among those with a college degree, the marginal effect nearly doubles to 19 points (from 50% to 69%). While both sorted partisans and the better educated prioritize issue-based representation, those who are both sorted and well educated are especially likely to do so.

These findings become even more significant when put in the context of who actually votes. We know that those who are sorted partisans, better educated, wealthier, and so forth are more likely to turn out to vote in American elections (see, among others, Levendusky 2009; Verba, Schlozman, and Brady 1995). Indeed, the elderly are the only high-participation group that prioritizes constituency service over issue positions in a member’s performance.

Such patterns almost certainly influence congressional behavior. Obviously, voters still want members to provide constituency services, and members undoubtedly are highly sensitive to such requests (Butler, Karpowitz, and Pope 2012), in part because they know that voters will reward them electorally for providing services (Cain, Ferejohn, and Fiorina 1987; Serra and Moon 1994). That said, however, the importance of issue-based representation presents a real challenge to members. Members’ behavior is constrained by high participation, high-interest voters’ prioritization of substantive ideological (and therefore partisan) representation, and the focus of these groups on issues drives member behavior in particular ways (Harbridge and Malhotra 2011). Furthermore, as we saw in Table 3, those who prioritize constituency service over issue-based representation are less informed, less engaged, less partisan, and so forth—and hence are far less likely to turn out and vote. Our data confirm this pattern: respondents who prioritized issues over constituency service were significantly more likely to report that they would turn out and vote (5.2% more likely, p < .01) and reported more interest in the election (0.33 points on a 10-point scale, p < .01; both analyses are difference-of-means tests).

Do These Patterns Shape Evaluations and Vote Choice?

So far, we have demonstrated that in today’s political environment, citizens expect their member of Congress to represent them on the salient national issues of the day and that this tendency is especially pronounced among sorted partisans and the better educated. But we have not shown the consequences of these patterns. Do they shape how citizens evaluate their members, or their eventual vote choice?

Our argument suggests that they should. In particular, we would expect those who prioritize issue representation to put more weight on issue and partisan congruence when evaluating members. This implies that voters who weight issues more heavily should evaluate their same-party members more positively (and opposite-party members more negatively), and be more likely to vote for candidates from their own party.

The TAPS data give us the necessary data we would need to evaluate whether this holds for approval of one’s member of Congress. Using the TAPS data, we estimate the following model: Approve_i = β₀ + β₁SameParty_i + β₂IssueAgree_i + β₃National_i + β₄National_i × IssueAgree_i +u_i and Approve_i = β₀ + β₁SameParty_i + β₂IssueAgree_i + β₃National_i + β₄National_i × Same Party_i +u_i

Here, Approve is a binary variable (1 = approve of your member, 0 = do not approve), SameParty is a binary variable indicating whether the member and the respondent share the same partisanship, IssueAgree measures the respondents’ perceived level of issue agreement with his member,¹² National indicates whether the respondent thinks working on national issues is a very important part of a member’s job (1 = thinks working on national issues is a very important part of a member’s job, 0 otherwise), and u is a stochastic disturbance term. Our core hypothesis is that the interaction term (β₄) should be significant in both models, as this implies that when a respondent thinks national issues are more important that they give greater weight to issue/partisan congruence when evaluating their own member of Congress. Table 4 gives the estimates of our model.

Table 4.

Effects of Issue Prioritization on Evaluations and Vote Choice.

	(1) Member approval (TAPS)	(1) Member approval (TAPS)	(2) Vote choice (AIOD)
Same-party incumbent	0.35 (0.43)	−0.29 (0.59)	0.88*** (0.21)
Issue agreement	1.83*** (0.47)	1.91*** (0.27)
National issues very important	−0.65 (1.84)	−0.49 (0.50)
National Issues Very Important × Issue Agreement	0.27 (0.55)
National Issues Very Important × Same-Party Incumbent		1.44** (0.73)
Strong partisan			1.53*** (0.30)
Prioritize issues over service			0.32* (0.19)
Constant	−5.85 (1.57)	−5.80 (0.96)	1.73 (0.24)
N	440	440	3,163

Source. TAPS (Column 1-2), AIOD (Column 3).

Cell entries are logistic regression coefficients with associated standard errors in parentheses. Survey weights are used for both datasets, and week fixed effects are included for the AIOD data. TAPS = The American Panel Survey; AIOD = Annenberg Institutions of Democracy.

Coefficients that can be statistically distinguished from 0 are given as follows: *p < .1. **p < .05. ***p < .01.

Column 1 of Table 4 is not consistent with our expectations: unfortunately, there is no significant interaction between issue agreement and thinking working on national issues is a very important part of a member’s job. But, as column 2 illustrates, there is a significant and positive interaction between partisan congruence and thinking national issues are a very important part of a member’s job—those who put more emphasis on issues view an incumbent from their own party more positively.¹³

Because interactions can be difficult to understand, it is helpful to consider some simple average differences. Of those who do not think working on national issues is a very important part of a member’s job, 35 percent approve of their member if they are from opposite parties, and that figure rises 57 percent if they are from the same party. But among those who think that issues are a very important part of a member’s job, 31 percent approve of a member from the opposing party, but 83 percent approve of a member from their own party. So among those who do not view working on national issues as a very important part of a member’s job, the gap in approval between an opposite-party and same-party member is only 22 percentage points. But that figure rises to 52 percentage points among those who view working on national issues as a very important part of a member’s job. There is much greater partisan polarization in approval among voters who see working on national issues as a very important part of a member’s job. The member’s party is a much more important determinant of approval among those who prioritize issue representation from their member of Congress.

We also expect such preferences to shape vote choice as well. Here, our underlying theoretical expectation would be the same: those who attach more importance to national issues would place more weight on issue/partisan congruence when casting their ballot. This implies that those who attach more weight to issues should be more likely to vote for their same-party candidate, or a candidate who they think represents them more fully on the issues. We can examine vote choice in the AIOD data, which ask respondents (prior to the election) whether they plan to vote, and if so, which party’s candidate they will support in the 2014 election. Because the AIOD data do not ask about issue congruence, we cannot look at its effects here, but we can look at party voting as a close substitute. We can then estimate a simple model of the following form:

V o t e P a r t y_{i} = β_{0} + β_{1} S a m e P a r t y_{i} + β_{2} S t r o n g P a r t i s a n_{i} + β_{3} I s s u e s O v e r S e r v i c e_{i} + δ_{w (i)} + u_{i} .

Here, SameParty is dummy variable equal to 1 if the respondent’s party matches that of the incumbent, 0 otherwise; StrongPartisan is a dummy equal to 1 if the respondent is a strong partisan, 0 otherwise; IssuesOverService is a dummy variable equal to 1 if the respondent rates issue representation as being more important than constituency service when evaluating members (see Table 3); $δ_{w (i)}$ is a week-of-interview fixed effect; u is a stochastic disturbance term. Here, β₁ captures the effect of incumbency, β₂ captures the effect of partisan strength, and β₃ which is our key coefficient, captures the effect of issue prioritization. β₃ tells us whether, controlling for incumbency and party, prioritizing issues above constituency service makes a respondent more likely to cast a ballot for their own party. Column 3 of Table 4 gives our results.

Column 3 again confirms our theoretical expectations. Unsurprisingly, there is a strong effect of incumbency: people are more likely to cast a ballot for their party when the incumbent House member is from their party. Similarly, being a strong partisan makes respondents more likely to cast a ballot for their party. But even controlling for these two factors—the two most crucial determinants of congressional vote choice, party and incumbency—there is still an effect of issue prioritization. Those who think issue representation is more important than constituency service are more likely to vote for their party’s candidate. This is particularly striking because 2014 was such a deeply polarized election at the ballot box: in our data, 93 percent of respondents report intending to vote for their party’s candidate, so there is almost no room for other factors to matter. But nevertheless, we still find an effect for our core argument.

Taken together, these results nicely complement our earlier findings. In particular, they highlight that our findings in Tables 1 to 3 are not just mere “cheap talk” from respondents. Respondents say that national issues matter to them, and that in turn shapes how they evaluate members of Congress.

Have These Patterns Changed over Time?

Because our data from TAPS and the AIOD come from the past few years, we cannot really make strong claims about how the patterns we have uncovered here may have changed over time. But given our argument above, one could imagine that in an earlier, less polarized era, voters would have evaluated members differently, with more emphasis on constituency service or district projects, and less on issues.

While we cannot test such claims rigorously, we do have some data—albeit of a limited nature—that can give us a suggestive glance at how voters evaluated members in an earlier era. This data comes from the American National Election Study, which has, in some years, asked questions about how voters evaluate members of Congress. We present three pieces of evidence that suggest that our findings above represent a departure from earlier periods. First, the NES asked in 1978, 1980, 1986, and 1990 whether individuals had contacted their member of Congress, and if so, what they had requested from them (the options were to express an opinion, request information, or request help with a problem). This is a slightly different list than in the TAPS data discussed above, but similar enough that we can make some comparisons. Here, much like TAPS, contact with a member is relatively rare (approximately 14 percent of respondents report such contact), and approximately ½ of respondents contacted a member to express an opinion. But unlike TAPS, an equal or larger number contacted a member with a constituency service request: in every year the item was asked, at least 45 percent of respondents who had contacted a member did so to request help with some problem (respondents could give more than one response, so these categories do not sum to 100 percent). In contrast, in our more recent TAPS data, the figure was 19 percent, suggesting a decline in such service-driven contact in the contemporary era.

Second, in the 1988–1990–1992 Senate Election Study, respondents were asked whether, on average, they agreed with their House incumbent’s voting decisions. Overall, 54 percent agree with their member. Not surprisingly, there is a large gap between those who have a same-party incumbent and those represented by an incumbent from the other party. For respondents with a same-party incumbent, 68 percent agree with their member, and of those with an opposite-party incumbent, the figure is 40 percent, for a 28-point gap. We can compute a similar effect for our TAPS data. There we find a 50-point gap: 68 percent of respondents with a same-party incumbent agree (on average) with their member, but that figure plummets to 18 percent for those with an incumbent from the other party. Such findings are also consistent with the fact that voters now recognize that the parties have sharply different policy platforms (Levendusky 2009), and hence agreement with a representative from the other party is unlikely. Twenty-five years ago, with less polarized parties, it was easier to agree with the ideological positions of a member of the opposing party, but not so today. Agreement with a member’s decisions today is much more sharply separated by party.

Third, in that same Senate Election Study, respondents were asked to rank order the importance of three activities for their representative: securing federal monies for the district, working on issues, and providing constituency service. As in our TAPS data, most people ranked working on issues and securing money about equally. But more interestingly, in the Senate Election Study, these rankings do not relate to member approval and voting. In the previous section, we saw that, in both the TAPS and AIOD data, voters who placed more importance on the issues were more likely to approve of their same-party member and to vote for their party’s candidate. But here, we find no such effects. In the contemporary highly polarized era, prioritizing issues leads one to same-party candidates—these are the candidates who will represent you on the issues. But twenty-five years ago, in a less divided era, party was a less reliable cue of issue congruence (see point 2 above), and hence it had less of an effect on evaluations and vote choice.

Together, these pieces of evidence suggest that in an earlier era, contact with Congress was more service driven, and evaluations were less tightly tied to partisanship and issue congruence. The key word there is suggest, as the data we have cannot definitively demonstrate this, but they do indicate interesting opportunities for future over-time data collection projects.

Broader Implications

In this paper, we ask what voters want from their members of Congress. We show that, in the contemporary environment, voters value both constituency service and issue representation, but that issues are somewhat more important for most voters. Using evidence from two nationally representative surveys, we have shown that ordinary voters do indeed weigh partisan concerns more heavily than service-based concerns. Most voters assume that any member will perform the constituency service functions of a member of Congress, but feel that only one party will actually represent them on the issues. Thus, partisan and ideological considerations drive what they want from a member, and this pattern is particularly pronounced for sorted partisans and the well educated. Finally, we provide evidence that these preferences shape how voters evaluate their members of Congress, and how they vote. These views are not simply cheap talk, but actually shape how citizens behave in the voting booth.

Like all analyses, ours suffers from weaknesses. In particular, we only have data from survey self-reports. This limits us in two ways. First, we might suffer from some social desirability bias, in that people think it is “good” to select issue representation, as it makes them seem serious and policy minded. This is, of course, a possibility. While we cannot rule out this possibility completely, our results about the effects of such preferences suggest this is not the primary driver of our results. If social desirability bias primarily caused our results, we would not see those who emphasize issues to more strongly approve of their same-party members, or to be more likely to vote for them. Because these preferences actually shape behavior, this suggests that these preferences are not simply the product of people wanting to seem high-minded. Second, while we have some data on the implications of these attitudes, there is certainly more to be done to link voter responses to member behavior more broadly (for such an analysis, see Sulkin, Testa, and Usry 2015). While these are limits, our work nevertheless establishes an important baseline finding about how contemporary voters weigh the relative importance of constituency service and issue representation.

Such findings have a number of important implications for our understanding of Congress and representation. First, voters’ prioritization of ideological/partisan considerations effectively nationalizes elections, making them much more about national issues and concerns than about local issues. We saw evidence of this in 2014, when issues like Obamacare, the spread of the Islamic State, and the Ebola virus drove voter decision making more than local issues (Jacobson 2015b). Instead of all politics being local, all politics—at least congressional politics—is now trending in a national direction (though local factors still, of course, matter). While we are still in the early days of 2016, all signs point to this year being a similarly nationalized election (Confessore 2016). It is possible that recent cycles have been unique; our data do not allow us to say how such considerations would play out in other years. But given the broad structural factors we identify here, this seems like more of a systematic pattern than a onetime aberration (see also Abramson et al. 2015, 287).

Such changes are also consistent with Jacobson’s (2015a) finding that partisanship has become more important, and incumbency less so, in recent congressional elections (see also Abramowitz and Webster 2016). Our findings help to explain why. Ordinary voters today see no difference between the parties on issues such as constituency service, bringing federal monies to the district, and standing up to special interests, but they see an enormous difference between the parties when it comes to representation on the issues. This being the case, a same-party challenger becomes a much more attractive candidate than an opposite-party incumbent: voters think that both candidates will deliver the services but only one—their party’s nominee—will represent them on the issues. The way voters prioritize the components of a member’s job drives the partisan factor in voting.¹⁴

The tendency we observe here also likely exacerbates congressional polarization. When voters value issue representation, that makes a member’s position-taking activity all the more important. It gives members an incentive to engage in polarizing behavior, and it makes it more difficult to engage in the nonpartisan cultivation of support that was at the heart of the personal vote. Congressional politics increasingly becomes about issue considerations, especially on divisive, national issues, and as a result, there is more fuel for the fire of polarization.

Finally, these patterns will likely continue to strengthen the hand of party leaders. While we cannot say what has happened over time (given that we have a cross-section), our findings are consistent with an argument that issues have become increasingly important over time. If this is so, it might prompt us to rethink one of the consequences of the personal vote. One of the most important consequences of the rise of the personal vote was the ensuing weakening of party leaders (Cain, Ferejohn, and Fiorina 1987)—if a member’s support was nonpartisan, then he or she could more easily resist efforts from the party leaders to control that member. But now, with support being much more partisan, members have more incentives to cede control to party leaders to coordinate policy (Cox and McCubbins 1993). Of course, this tendency is not absolute: for example, the Tea Party helped lead to John Boehner stepping down as speaker. But overall, as voters have shifted to wanting more partisan/ideological representation from their members in Congress, Congress has delivered. Whether that will continue is unclear, but at least at the moment, the importance of issues to voters has changed the contemporary incentives of members of Congress.

Footnotes

Acknowledgements

We thank the Annenberg Public Policy Center at the University of Pennsylvania and the Weidenbaum Center at Washington University, St. Louis for providing us with access to their data. We also appreciate the helpful comments provided by the referees and editors, as well as those by E. Scott Adler, Josh Clinton, Charles Finocchiaro, David Parker, Patrick Tucker, and seminar participants at the Annenberg Public Policy Center and APSA 2015.

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) received no financial support for the research, authorship, and/or publication of this article.

Notes

References

Abramowitz

Alan

Webster

Steven

. 2016. “The Rise of Negative Partisanship and the Nationalization of U.S. Elections in the 21st Century.” Electoral Studies 41 (1): 12–22.

Abramson

Paul

Aldrich

John

Gomez

Brad

Rohde

David

. 2015. Change and Continuity in the 2012 and 2014 Elections. Washington, DC: CQ Press.

Bartels

Larry

. 2000. “Partisanship and Voting Behavior, 1952–1996.” American Journal of Political Science 44 (1): 35–50.

Butler

Daniel

Karpowitz

Christopher F.

Pope

Jeremy C.

2012. “A Field Experiment on Legislators’ Home Styles: Service versus Policy.” Journal of Politics 74 (2): 474–86.

Cain

Bruce

Ferejohn

John

Fiorina

Morris

. 1987. The Personal Vote: Constituency Service and Electoral Independence. Cambridge: Harvard University Press.

Confessore

Nicholas

. 2016. “Candidates in ’16 Race Showing Less Interest in Regional Concerns.” The New York Times, 17 January, http://tinyurl.com/hzjqgba.

Cox

Gary

McCubbins

Mathew

. 1993. Legislative Leviathan. Berkeley: University of California Press.

Doherty

David

. 2013. “To Whom Do People Think Representatives Should Respond: Their District or the Country?” Public Opinion Quarterly 77(1): 237-55.

Eulau

Heinz

Karps

Paul

. 1977. “The Puzzle of Representation: Specifying Components of Representation.” Legislative Studies Quarterly 2 (3): 233–54.

10.

Fenno

Richard

. 1975. “If, as Ralph Nader Says, Congress Is ‘The Broken Branch,’ How Come We Love Our Congressmen So Much?” In Congress in Change: Evolution and Reform (pp.277-87), edited by Ornstein

Norman

. New York: Praeger.

11.

Fenno

Richard

. 1978. Home Style. Boston: Little, Brown.

12.

Griffin

John

Flavin

Patrick

. 2011. “How Citizens and Their Legislators Prioritize Spheres of Representation.” Political Research Quarterly 64 (3): 520–33.

13.

Grimmer

Justin

Westwood

Sean

Messing

Solomon

. 2014. The Impression of Influence: Legislator Communication, Representation, and Democratic Accountability. Princeton: Princeton University Press.

14.

Harbridge

Laurel

Malhotra

Neil

. 2011. “Electoral Incentives and Partisan Conflict in Congress: Evidence from Survey Experiments.” American Journal of Political Science 55 (3): 494–510.

15.

Harden

Jeffrey

. 2013. “Multidimensional Responsiveness: The Determinants of Legislators’ Representational Priorities.” Legislative Studies Quarterly 38 (2): 155–84.

16.

Harden

Jeffrey

. 2016. Multidimensional Democracy: A Supply and Demand Theory of Representation in American Legislatures. New York: Cambridge University Press.

17.

Jacobson

Gary

. 2015a. “It’s Nothing Personal: The Decline of the Incumbency Advantage in U.S. House Elections.” Journal of Politics 77 (3): 861–73.

18.

Jacobson

Gary

. 2015b. “Obama and Nationalized Electoral Politics in the 2014 Midterm.” Political Science Quarterly 130 (1): 1–25.

19.

Jewitt

Caitlin

Goren

Paul

. 2016. “Ideological Structure and Consistency in the Age of Polarization.” American Politics Research 44 (1): 81–105.

20.

Joachim

David

. 2014. “With Stealth, Congressional Spending on Pet Projects Persists, Report Says.” The New York Times, May 7, http://tinyurl.com/zo9saes.

21.

Keith

Bruce

Magelby

David

Nelson

Candice

Orr

Elizabeth

Westlye

Mark

Wolfinger

Raymond

. 1992. The Myth of the Independent Voter. Berkeley: University of California Press.

22.

Lee

David S.

Moretti

Enrico

Butler

Matthew J.

2004. “Do Voters Affect or Elect Policies? Evidence from the U.S. House.” Quarterly Journal of Economics 119 (3): 807–59.

23.

Levendusky

Matthew

. 2009. The Partisan Sort: How Liberals Became Democrats and Conservatives Became Republicans. Chicago: University of Chicago Press.

24.

Lockerbie

Brad

. 1999. “The Partisan Component to the Incumbency Advantage, 1956-1996.” Political Research Quarterly 52 (3): 631–46.

25.

Mayhew

David

. 1974. Congress: The Electoral Connection. New Haven: Yale University Press.

26.

McCarty

Nolan

Poole

Keith

Rosenthal

Howard

. 2006. Polarized America: The Dance of Ideology and Unequal Riches. Cambridge: MIT Press.

27.

Miller

Warren

Stokes

Donald

. 1963. “Constituency Influence in Congress.” American Political Science Review 57 (1): 45–56.

28.

Nie

Norman

Junn

Jane

Stehlik-Berry

Kenneth

. 1996. Education and Democratic Citizenship in America. Chicago: University of Chicago Press.

29.

Parker

David

Goodman

Craig

. 2009. “Making a Good Impression: Resource Allocation, Home Styles, and Washington Work.” Legislative Studies Quarterly 34 (4): 493–524.

30.

Serra

George

Moon

David

. 1994. “Casework, Issue Positions, and Voting in Congressional Elections: A District Analysis.” Journal of Politics 56 (1): 200–13.

31.

Sulkin

Tracy

Testa

Paul

Usry

Kaye

. 2015. “What Gets Rewarded? Legislative Activity and Constituency Approval.” Political Research Quarterly 68 (4): 690–702.

32.

Verba

Sidney

Schlozman

Kay

Brady

Henry

. 1995. Voice and Equality: Civic Volunteerism in American Politics. Cambridge: Harvard University Press.

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

0.14 MB