This file is also available in Adobe Acrobat PDF format
ISSUE IMPORTANCE, POLITICAL CONTEXT, AND DEMOCRATIC RESPONSIVENESS
ARGUABLY the most important issue facing the country in 1968 was the war in Vietnam. By the beginning of that year, almost 30 percent of Americans had friends or relatives among the approximately half a million troops stationed in Southeast Asia (Lau, Brown, and Sears 1978). Additionally, the war was costing taxpayers well over twenty billion dollars each year, and roughly one hundred U.S. soldiers--and an untold number of Vietnamese--were being killed per week (Page and Brody 1972). Americans had become accustomed to, as well as disgusted with, the nightly images of the conflict that were displayed in their homes via the network news. Public opinion was sharply divided between supporters of the war and an increasingly vocal peace movement. This movement did not yet represent majority sentiment in the country, but the prospect that it might eventually influence American foreign policy seemed very real. Still, when the leader of the antiwar movement in Congress, Senator Eugene McCarthy, emerged to challenge President Lyndon Johnson for the leadership of the Democratic Party, many observers thought his chances for success were remote. Nevertheless, McCarthy engaged in a determined effort to unseat the president, beginning with an intensive campaign in the New Hampshire primary. McCarthy garnered a surprising 42 percent of the vote in this contest, compared to 48 percent for President Johnson. Although technically a victory for the president, many interpreted this election as a moral victory for the antiwar movement and a resounding defeat for the administration's policies in Vietnam. Shortly after the New Hampshire primary, Johnson withdrew from the Democratic primary. His presidency was effectively over.
On its face, the preceding account appears to be an excellent example of the significant role that voters play in our representative form of democracy. However, scholars of public opinion recognize that there is one important point missing from this story. Public opinion polls would later reveal that most of McCarthy's supporters in the New Hampshire primary actually favored an escalation of hostilities in the war rather than withdrawal (Converse 1975). Moreover, most of these voters mistakenly thought that McCarthy also favored a more "hawkish" stance. In short, one of the most glaring examples of the influence of public opinion in recent American history highlights the deficiencies of the electorate more so than its strengths.
Decades of public opinion research have provided voluminous support for this conclusion. For example, less than half of respondents in national surveys know both the name and party affiliation of their representative in Congress (Jacobson 1992). Further, barely a majority can provide this information for the senator seeking reelection in their home state. Typically, these percentages are even lower for congressional challengers. A skeptic might counter that this information is relatively unimportant as long as voters have a fairly accurate sense about where politicians stand on the issues. Unfortunately, even this information is unknown to much of the public. For example, over 40 percent of Americans either had no idea, or an inaccurate perception, of how their senators voted on the high-profile decision to go to war in the Persian Gulf or the confirmation of Clarence Thomas to the Supreme Court. When one considers that roughly half would accurately identify their senators' vote based purely on chance, this figure seems particularly low. Moreover, at least this many citizens are also unaware of the traditional differences between the Republican and Democratic parties on issues such as social welfare spending, social security, defense spending, and taxes (Bennett 1995).
All of this might suggest that politicians can safely ignore the opinions of their constituents. Strangely enough, this is not the case. An equally large literature indicates that politicians are often quite concerned with how their constituents will react to their policy positions. For instance, as explored in more detail in chapter 3, many senators publicly agonized over their vote on the Thomas confirmation because of concerns about constituent reaction. More than a few senators also declared that it was the toughest vote they ever had to cast.
The well-known political ignorance of the American voter juxtaposed with the genuine concern politicians express about faithfully representing their constituents begs the following question. Why do politicians worry about their voting record if voters are only dimly aware of this information? The aim of this book is to provide an answer to this question through an examination of survey data drawn from both Senate and gubernatorial elections. In brief, I argue that the perception of the American public as generally uninformed on political matters, although strictly accurate, is also misleading. In fact, under the right circumstances, voters are surprisingly well informed on the issues that they care about.
It is true that most citizens are often only vaguely aware of the issue positions of major political candidates (Bennett 1995; Smith 1989). This ignorance becomes more pronounced as one moves from presidential elections down to the less visible campaigns in Congress (Jacobson 1992; Mann and Wolfinger 1980; Stokes and Miller 1962). There are a number of reasons why this is so. Chief among these, ironically, is the generally high level of policy responsiveness. For example, presidential candidates, particularly the successful ones, generally agree with most voters on the issues (Page 1978). Additionally, members of Congress frequently share the same views and values as their constituents and thus need only follow their own preferences in order to successfully represent their state or district (Miller and Stokes 1963). In summary, voters often do not pay attention to politics because politicians see to it that they do not have to.
This is not to say that politicians never vote against their constituents' interests. I will argue, however, that this is not as common as many expect in large part because of the specter of constituent vigilance. This general tendency toward responsiveness makes it difficult for challengers to exploit an incumbent's record, and, consequently, voters are not regularly confronted with issue-laden campaigns. Only infrequently do incumbents fail to anticipate what some scholars refer to as the "potential preferences" of voters. I will show that when politicians do misread the public, however, interested voters learn about it, provided the media or political challengers convey this information to them. In short, voters are generally as informed about their incumbent's performance in office as they ought to be, given the relatively high levels of responsiveness, and they are about as informed as they can be, given the information made available.
This study provides a broad description and theoretical assessment of how voters observe and evaluate political actors. As indicated above, this book focuses entirely on state-level contests, but there is no reason why its conclusions cannot also be applied to national or local elections. The latter part of the book explores the factors that influence the prospective judgments that voters make about the likely actions of political candidates. The bulk of this study, however, focuses on the retrospective evaluations that voters make of their incumbent's actual performance in office. Specifically, I examine the process by which citizens acquire information about the performance of Senate incumbents and how they subsequently use that information to hold them accountable at election time. I refer to this process as "monitoring." The concept of voter monitoring is centrally important to my argument, but it cannot be examined independent of the context in which it occurs. It must be assessed in light of the political information that is readily available and the motivation of voters to pay attention to this information. As we shall see, previous works have not always considered each of these factors; consequently, they have underestimated the attentiveness and influence of the American voter.
To assess whether citizens are indeed capable of fulfilling their democratic obligations, I address four interrelated questions: (1) How much information on issues of public policy is made available by the mass media, incumbents, and political challengers? (2) How much of the information that is made available do voters actually receive? (3) Are citizens who are interested in particular issues also more informed about the candidate's position on those issues? (4) Do interested voters rely more heavily upon the information they receive when evaluating incumbents at election time?
In summary, the argument of this book is that the generally high levels of political responsiveness can be explained, in part, by incumbents' anticipation of constituents' likely reaction to public policy initiatives. This, and the ideology that legislators and constituents often share by default, generally lead them to come down on the "right," or at least the popular, side of an issue, thereby preventing challengers from exploiting their record. As indicated in the next section, other scholars have also noted the importance of the voter's latent political preferences (Arnold 1990; Key 1961; Miller and Stokes 1963). What has not been fully appreciated, however, is that the influence of latent public opinion hinges on how easily it can become activated. Indeed, if the voters' potential preferences cannot be easily activated--and if politicians come to recognize this--then incumbents will have little incentive to consider latent attitudes at all. I argue in this book that politicians are wise to consider the attitudes of their constituents, even when their constituents are effectively "asleep" or not actively engaged in policy debates. When incumbents do not correctly anticipate and respond to voter preferences, the mass media, interest group leaders, and potential challengers become alerted, thereby increasing the likelihood that interested voters will become informed. Moreover, once they are informed, interested voters will defend their interests and values at the ballot box.
This view of the mass public suggests that the electorate might best be described as a loose collection of "sleeping giants." These giants are not routinely vigilant, and in any case they do not all share the same political priorities. When the interests or values of one or more of these giants are at stake in a political contest, however, they can become surprisingly alert. Of course, this attentiveness is contingent on the presence of favorable contextual conditions, such as the availability of sufficient political information. When political elites provide this information, interested members of the public pay attention. Conversely, when elites neglect their responsibility, the sleeping giants continue their slumber.
Theories of the Link between Public Opinion and Public Policy
When the framers of the Constitution designed our system of government, they envisioned the legislative branch, especially the House of Representatives, as the most important and the most responsive to the people (Hamilton, Madison, and Jay 1961). Indeed, contrary to popular perceptions, there is now ample evidence showing that the voting behavior of members of Congress adhere closely to the preferences of their constituents (Bartels 1991; Jackson and King 1989; Jacobson 1992; Kingdon 1989; Miller and Stokes 1963; Overby et al. 1992; Powell 1982; Shapiro et al. 1990; Whitby and Gilliam 1991). While at first glance this evidence suggests the existence of a vigilant and attentive citizenry, decades of research in public opinion offer little support for this view. Researchers in this field have consistently shown that the electorate is generally uninformed and uninterested in issues of public policy (Campbell et al. 1960; Conover and Feldman 1981; Converse 1964, 1975, 1990; Smith 1989). Moreover, some scholars go a step further, arguing that, in spite of steady increases in education, levels of political information are not likely to increase substantially in the foreseeable future (Bennett 1995; Knight 1990; Nie, Junn, and Stehlik-Barry 1996; Smith 1989).
In short, previous research shows that legislators are generally responsive, but that this is not due to active monitoring by the public. This conclusion is puzzling. The framers of the Constitution expected the electorate to assess the performance of their representatives and periodically remove nonresponsive legislators from office. Thus, a vigilant citizenry was regarded as one of the chief mechanisms for ensuring democratic accountability (Dahl 1989; Hamilton, Madison, and Jay 1961). But if the electorate is indeed uninformed, they may be ill prepared to review the performance of their representatives, as the framers anticipated. What reasons, then, are there for lawmakers, or any other group of elected officials, to be responsive? More importantly, what penalties (if any) can legislators realistically expect to suffer if they do not faithfully represent their constituents?
Resolving this apparent contradiction is of more than academic importance. If voters are politically inattentive, then elected officials can ignore constituent opinion with impunity. Other scholars have also recognized this paradox and have developed several models that may account for this prima facie inconsistency (Erikson and Luttbeg 1973; Fiorina 1974; Jacobson and Kernell 1981; Miller and Stokes 1963). These models, although not mutually exclusive, invoke one of four main explanations: the general partisan or ideological orientation of the district; the ability of the electorate as a whole to cancel out the weaknesses of individual voters; the mediating role of organized interest groups; and elite attentiveness to, or anticipation of, constituency preferences.
District Political Orientation
According to the first explanation, policy agreement between constituents and legislators occurs primarily because of the manner in which many political jurisdictions are drawn. Specifically, most congressional districts are typically quite homogenous, both politically and socially. Even when there is significant ideological heterogeneity, many states and districts tend to have a dominant political orientation. Thus, the kind of representative most likely to be elected from such districts will tend to share this orientation. As a result, all that these legislators need do is to follow the dictates of their own conscience in order to represent their constituents effectively as well. This seems especially plausible in more homogeneous states or districts (Erikson 1978; Fiorina, 1974; Key 1961; Miller and Stokes 1963; Powell 1982).
In an elaboration on this model, representation may come about indirectly through voter reliance on party identification. According to this explanation, many political jurisdictions have a dominant partisan character just as they have a dominant ideological character. Because voting decisions rely heavily on partisan identification, members of the dominant party are far more likely to attain political office. Consequently, because partisan identification generally corresponds closely with issue preferences among both the elite and the mass public, legislators typically end up representing their districts even though voters do not consciously consider issues at election time (Erikson 1978; Franklin 1984; Pomper 1972; Popkin 1991; Repass 1971; Stokes and Miller 1962).
A second manner through which the deficiencies of the electorate can be reconciled with widespread evidence of congressional responsiveness is preference aggregation. This view does not dispute the fact that most citizens are uninterested in politics and uninformed about policy debates at the elite level. However, proponents of this argument maintain that collective public opinion is often remarkably informed and influential (Abramowitz 1988; Condorcet 1785; Converse 1990; Page and Shapiro 1992).
Page and Shapiro (1992) offer perhaps the most detailed exposition of this account. They argue that, contrary to the view of many experts, collective public opinion is frequently stable (although not fixed), is coherently organized, takes into account the best available information, and changes in predictable and reasonable ways. In short, collective public opinion is rational. This can occur because most citizens are not terribly interested in politics and are not often critical of the political views they are exposed to and subsequently accept. This tendency to internalize politically inconsistent messages can result in seemingly random fluctuations in individual policy preferences. However, Page and Shapiro argue that the high levels of variance apparent in individual issue positions often mask a "true" underlying attitude.1 When examined at the aggregate level, these fluctuations cancel out to reveal a public opinion responsive to elite behavior and one in which policy questions have a nontrivial impact on election outcomes (Erikson and Wright 1989; Kahn and Kenny 1999; Wright and Berkman 1986). According to Erikson and Wright, the "electorate is much more capable in the aggregate than as individual voters. It is as though all our individual ignorance and misinformed judgments cancel out, so that average perceptions and judgments are responsive to what candidates say and do. The result is perhaps a more representative Congress than the electorate sometimes seems to deserve" (114).
Interest Group Pressure
A third explanation for policy responsiveness relies less on the abilities of the average citizen and more on the presence of organized pressure groups (Blumer 1948; Erikson and Lutbegg 1973; Truman 1971). These groups can often effectively influence Congress because of their ability to identify and mobilize the inchoate interests of like-minded citizens through television advertising campaigns, petitions drives, and organized rallies (Kollman 1998). More recently, interest groups have also begun to rely on Internet-based techniques, or "cyber-lobbying," to organize potential followers and bring pressure to bear on Congress (Davidson and Oleszek 1998). This strategy of "outside lobbying" is often accompanied by more traditional efforts to influence legislation, such as providing policy information and technical expertise, or campaign contributions (Davidson and Oleszek 1998; Hall and Wayman 1990; Kingdon 1989). Thus, to the extent that interest groups are representative of constituent opinion, ordinary citizens may still exert an influence on their political representatives in spite of their inattentiveness. This is because it is interest groups that are actively monitoring politicians and not the average voter.
Finally, as indicated above, some scholars maintain that even in the face of voter ignorance, politicians have an incentive to be responsive because of their perceptions of the latent or potential preferences of inattentive citizens (Arnold 1990; Erikson and Lutbeg 1973; Fenno 1978; Fiorina 1974; Key 1961; Kingdon 1989; Miller and Stokes 1963). This explanation will be referred to as the potential preferences model. According to this model, legislators will tend to be responsive as long as they are convinced that someone (particularly a would-be challenger) is paying attention and might inform their constituents at election time.
Miller and Stokes explain the influence of the voter's potential preferences on legislative behavior in their 1963 article, "Constituency Influence in Congress."
By voting correctly on [the issues, House members] are unlikely to increase their visibility to constituents. Nevertheless, the fact of constituency influence, backed by potential sanctions at the polls, is real enough. That these potential sanctions are all too real is best illustrated in the election of 1958 by the reprisal against Representative Brooks Hays in Arkansas' Fifth District. Although the perception of Hays as too moderate on civil rights resulted more from his service as intermediary between the White House and Governor Faubus in the Little Rock school crisis than from his record in the House, the victory of Dale Alford as a write-in candidate was a striking reminder of what can happen to a Congressman who gives his foes a powerful issue to use against him. (55; italics added)
Generally speaking, then, the models outlined above view the mechanism of representation as deriving either from lack of serious divisions within the political jurisdiction resulting in a shared ideological orientation between legislator (or political representative more generally) and constituent, the effects of aggregation, the influence of interest groups, or the potential attentiveness of the electorate. Although each of the models offers a persuasive explanation for legislative responsiveness, none of them is entirely satisfactory. Moreover, they generally provide only a limited role for the electorate. For example, the first model neither expects nor requires any direct monitoring on the part of most voters. At most, this model allows voters to exercise their influence on policy through their selection of partisan representatives. Beyond that, the average voter should demonstrate little active engagement in the political process.
Clearly, partisanship is an important component in ensuring legislative accountability, but relying solely on a candidate's party as an indication of their issue positions can also be misleading. Although there is considerable variance between the two major parties on numerous issues, there are also many issues on which there is considerable variance within the parties. For example, Democrats tend to be the more liberal party on the abortion question, yet in Louisiana all four Democratic House members along with the two Democratic senators scored a perfect 0 on the National Abortion Rights Action League (NARAL) index in 1990.2 There were also few partisan differences among House members on abortion this year in states such as Connecticut, Idaho, Kentucky, Maine, Mississippi, Washington, and elsewhere in the country.
Differences on important policy questions within the major parties are at times even greater in the Senate. In 1990 alone, senators of the same party representing the same state differed by at least 30 points on the NARAL index in Alabama, Alaska, Illinois, Kansas, Missouri, Nebraska, Nevada, New Hampshire, North Dakota, Oregon, and Washington. Moreover, Republican senators scored uncharacteristically high (i.e., 60 or above) on this measure in seven states in 1990. In the same year, six of the seventeen Democratic senators facing reelection also received scores of 60 or higher on the conservative National Security Index of congressional votes (NSI). Clearly, knowing which party a politician belongs to reveals a great deal about his or her issue positions, but it does not reveal everything. Therefore, it seems likely that the generally high levels of congressional responsiveness do not occur simply because of the partisan or ideological orientation of the district.
Preference aggregation models find that an informed consideration of the issues does influence the collective voting decisions of the electorate. Other scholars, however, have found that the aggregation process does not quite live up to its billing. Althaus (1998), for example, reports that the collective opinion represented in surveys often misrepresents the distribution of attitudes in society. In short, aggregate support for or opposition to some policies would look considerably different if the average voter were more informed about politics (also see Bartels 1996). Converse (1990) also points out that preference aggregation models often overlook the fact that some citizens are vastly more informed than are others. He concludes that the "rationality" attributed to aggregate public opinion derives disproportionately from the relative few who are especially well informed.
There is some controversy in the literature as to the identity of this informed group of voters. Some argue that there is a small group of citizens who are typically attentive to a broad range of issues whereas most voters have little or no interest in politics (Price and Zaller 1993; Zaller 1992). If this is so, then the influence of aggregate public opinion can more accurately be described as the influence of the relatively few Americans who are unusually attentive to politics. According to another view, one adopted in this book and explored more fully in the next section, the more informed sentiment at the core of aggregate public opinion differs from issue to issue. Converse (1964) referred to these groups as "issue publics." Subsequent chapters will demonstrate that aggregate public opinion is indeed attentive to the activities of incumbents. However, issue publics are as likely as citizens more generally informed about politics to represent the driving force behind this attentiveness.
Pressure group models also expect little active monitoring by most citizens. Instead, interest group leaders watch over public officials in order to safeguard the welfare of their members. There is at least one reason why this explanation cannot fully account for the relationship between public opinion and public policy. Sometimes group leaders and rank-and-file members disagree on major policy questions. A conspicuous example is provided in chapter 3's discussion of the Clarence Thomas confirmation battle. Mainstream black political leaders opposed this unusually controversial nomination, but opinion polls showed that most African Americans supported Thomas. In the end, many Democratic senators also sided with Thomas. Presumably, these senators were more concerned with the opinion of their black constituents than that of African American group leaders. This concern could only be justified if the average black voter, and not merely black opinion leaders, were more informed about this vote than other Americans.
Even if one assumes that, on the majority of issues, interest group leaders do most of the heavy lifting in a representative democracy, there is still an important role for ordinary citizens. If the threats of an interest group regarding voter mobilization are to have any credibility among politicians, then group leaders must demonstrate an ability to alert and mobilize its membership. In other words, the "sleeping giant" must stir, at least occasionally, if governmental leaders are to pay it any heed.
Unlike district orientation models, potential preference models do not explicitly preclude issue-based monitoring, but they do suggest that legislators try to anticipate potential voter concerns and defuse them before they arise. Hence, as with the pressure group models, actual voter attentiveness is unnecessary and unusual. Still, representatives rarely have complete information about their constituents' current or future preferences, and so they may still unwittingly provoke voters and thereby encourage would-be challengers. Moreover, the heterogeneity of their state or district may also preclude a cost-free vote on some issues (Hutchings, McClerking, and Charles 2000). Of course this is only problematic if the electorate is informed or can easily become so. If the likelihood of voters becoming informed is indeed low, then legislators will typically suffer no costs for incorrectly anticipating voter preferences.
None of the general models described above fully explains the observed correlation between legislators' roll call votes and district/state opinion. Shared partisanship and ideology undoubtedly play an important role, but there is sufficient variation within and across congressional parties to undermine confidence in this explanation. The aggregation model seems to require at least some voters to have informed opinions, but it does not specify which voters are informed and under what circumstances they become so. Finally, the pressure group and potential preferences models suggest a more indirect role for voters in ensuring congressional responsiveness. Still, both would have little influence over policy makers unless there are at least occasional instances when voters demonstrate significant interest in, and knowledge about, matters of public policy.
None of these broader explanations is the only way to account for the relationship between constituent attitudes and legislative outcomes. Some scholars simply reject the premise that substantial levels of political knowledge are required in order for the mass public to effectively monitor their elected representatives. Some of these researchers argue that many citizens rely on information "short cuts," or heuristics, to direct their political judgments (Grofman and Norrander 1990; Lupia 1994; Lupia and McCubbins 1998; Popkin 1991). Others argue that voter impressions of political candidates and parties are constantly updated with new information even though much of that information is quickly forgotten (Fiorina 1981; Lodge, McGraw, and Stroh 1989; Lodge, Steenbergen, and Brau 1995). I am generally persuaded by both sets of explanations. However, they do not negate the basic importance of voter monitoring. If politicians behave in a manner inconsistent with the preferences of their constituents, then the only way for voters to hold them accountable is if they learn about this action and vote accordingly at election time. Whether this information is gained via shortcuts or more traditional routes ultimately involves the process of attentiveness and not the product. This book is more concerned with whether citizens gain the necessary political information than with precisely how they gain it. Further, whether or not this information influences political impressions but is quickly forgotten should not affect whether the voting record of members of Congress influence citizens' vote choice. Under either scenario, voters must reward their "friends" and punish their "enemies" at the ballot box in order for them to play any direct role in ensuring legislative responsiveness.
An Alternative Framework for Understanding the Relationship between Public Opinion and Policy Responsiveness
This study argues that voter monitoring does not play merely a passive, or indirect, role in helping to guarantee legislative responsiveness. Along with the other theories discussed above, voter monitoring (or at least the specter of it) plays a critical role in the relationship between constituent attitudes and political representation. To understand this relationship, we must turn our attention to two interrelated sets of concepts: individual motivation and contextual factors that promote heightened political attentiveness.
The motivation to learn about politics is central to the thesis of this book. A growing body of literature suggests that this motivation can sometimes compensate for traditional deficiencies in political information. More specifically, many scholars argue that shifting groups of citizens are remarkably informed about, and more likely to vote on the basis of, issues that they perceive as important (Campbell 2003; Hutchings 2001; Iyengar 1990; Krosnick 1988, 1990b; Krosnick, Berent, and Boninger 1994; McGraw and Pinney 1990; Popkin 1991). Previous research indicates that these perceptions generally originate from at least one of three sources: self-interest, group interests, or core values (Boninger, Krosnick, and Berent 1995; Campbell, et al. 1960; Conover 1984, 1985; Popkin 1991).3 This helps to explain why Iyengar (1990) found that Jewish Americans were more informed than other citizens about Middle East politics, blacks were more informed than whites about civil rights issues, and blue-collar workers were more informed than non-group members about the economy.4 Philip Converse, in "The Nature of Belief Systems in Mass Publics"(1964), was among the first to recognize the importance of individual motivation (see also Key, 1961; Truman 1971). He noted that
different controversies excite different people to the point of real opinion formation. One man takes an interest in policies bearing on the Negro and is relatively indifferent to or ignorant about controversies in other areas. His neighbor may have few crystallized opinions on the race issue, but he may find the subject of foreign aid very important. Such sharp divisions of interest are part of what the term "issue public" is intended to convey. (245; italics added)
There is an additional reason to believe that voters tend to focus on a few, mostly group-relevant issues. The costs of becoming well informed on more than a few issues are too great for most people (Downs 1957; Fiske and Taylor 1991; Lau and Sears 1986). Concentrating on a relatively narrow range of issues is one way of minimizing these costs. Citizens also rely on partisan cues, friends and co-workers, or formal groups such as unions and churches as sources of political information (Dawson 1994; Huckfeldt 1986; Tate 1993). Previous studies on the role of issues in congressional elections have often overlooked or understated the importance of issue salience (Abramowitz 1988; Erikson and Wright 1989; Mann and Wolfinger 1980; Stokes and Miller 1962; Wright and Berkman 1986).5 Consequently, they have understated the influence that policy concerns play in voters' political judgments and the role that constituent preferences play in affecting political outcomes.
Yet another reason for believing that the electorate is primarily made up of citizens who are most attentive to issues they view as important is that legislators themselves view the public in this way (Fiorina 1974). For example, Fenno (1978) reports a House member stating that
'[t]here isn't one voter in 20,000 who knows my voting record . . . except on that one thing that affects him.' And another said,' 'Only a few discerning people know my voting record: labor, the environmentalists, and the Chamber of Commerce.' But it is, of course, the voter dissatisfied with 'that one thing that affects him' or the 'few discerning people' who will press for explanations. (142; italics added)
Kingdon (1989) reaches similar conclusions. He notes that many representatives feel that while their constituencies were generally ignorant of their votes in Washington, subgroups within the district could be extremely aware, depending on the issue. One member of Congress interviewed by Kingdon was particularly concerned about the attentiveness of African Americans and organized labor.
[M]ost of my constituents don't care [about the Adam Clayton Powell vote] . . . But there is one group that will notice--the black community. They'll take account of what you do, and hold it against you if you go wrong. This is often the way it is. Take the compulsory arbitration matter last year. Most of the people don't have the vaguest notion about this, but the labor groups will notice and take account of it. (32)
This research is important because it supports the view that voters need not be generally attentive to political issues in order to ensure accountability. As long as voters are informed about the issues they care about, they are likely to hold members of Congress accountable when they are not responsive. Moreover, as long as members of Congress recognize this, they will have an incentive to anticipate such reactions and act to defuse them by voting in line with the preferences of attentive--and politically significant--publics.
The notion that when citizens are interested in politics they will also learn more about politics is a simple and compelling idea. Unfortunately, support for this thesis has not always been overwhelming. Although many scholars do find support for this hypothesis (Iyengar 1990; Hutchings 2001; Krosnick 1988, 1990b; Krosnick et al. 1994; Popkin 1991; Repass 1971), others do not (Margolis 1977; Niemi and Bartels 1985; Nueman 1986; Price and Zaller 1993; Rabinowitz, Prothro, and Jacoby 1982; Wolpert and Gimpel 1997).6 Why does evidence of issue salience emerge under some circumstances but not others? One explanation for the mixed findings is that previous efforts to study issue publics have typically not considered contextual factors.
Forces outside of the individual also play a critical role in the process of voter monitoring. More specifically, contextual factors such as the level of media coverage devoted to a legislator's position on various issues and the manner in which those issues are covered can also influence citizen engagement in politics. These considerations are important because they help to determine when heightened attentiveness should occur among issue publics. For example, Iyengar and Kinder (1987), along with others, have found that particular issues are more apt to influence political judgments when those issues are prominently covered in the media. This effect, referred to as "priming," is extremely robust and has been supported with both experimental and survey methodologies (Iyengar and Kinder 1987; Krosnick and Brannon 1993; Krosnick and Kinder 1990; Mendelberg 2001; Valentino, Hutchings, and White 2002).
The impact of priming on citizen engagement is considerable. Nevertheless, I believe that previous work may have underestimated the influence of these contextual factors. For one thing, the priming hypothesis was designed to identify media effects, and few subsequent elaborations have explicitly applied the theory to campaign settings.7 This is significant because priming effects should also occur when candidates emphasize issues (albeit through the mass media). More importantly, however, scholars have not paid sufficient attention to the joint effects of contextual variables and measures of issue salience.8 Both factors should have independent effects on voter monitoring, but their combined effects should be especially powerful.
The hypothesized relationship between context, motivation, voter monitoring, and political outcomes is summarized in figure 1.1. Perceptions of issue importance and contextual factors should have both direct (as indicated by the dotted lines) and interactive (as indicated by the solid line) effects on levels of political engagement. Further, the latter effects are likely to be stronger than the former effects. When these factors are present, interest in the political process should increase along with information levels, issue voting, and participation rates. These forces should in turn influence policy decisions both because of politicians' anticipation of these effects (i.e., the impact of potential preferences) and through their direct influence on election outcomes.
This issue salience perspective on voting behavior fits comfortably with most of the models described above. For example, the aggregation model implicitly requires some small set of voters to actively monitor the political process. The issue salience theory identifies which voters are most likely to engage in this activity. The pressure group and potential preferences models undoubtedly provide a partial explanation for the relatively high levels of political responsiveness, but they are also incomplete. An issue salience perspective supplements these accounts by explaining why politicians are motivated to anticipate voter preferences or respond to the requests of interest group leaders. In both cases, successful politicians recognize that sometimes voters can be exceptionally attentive to politics, especially when their values or self-interests or group interests are threatened. In short, an emphasis on voter perceptions of issue importance clarifies our understanding of the relationship between public opinion and public policy. As the works of Fenno, Kingdon, and others have shown, legislators are generally convinced that some members of their constituencies do monitor their roll call votes. This perception is at odds with the bulk of public opinion research, unless one relies on the issue salience hypothesis.
The research cited above indicates that, while citizens may not monitor their representatives on every issue, voters are attentive to legislative activity on issues they perceive as important. This kind of voter oversight may not be routine, given the expectations of the potential preferences model, but it should predictably occur when certain conditions are present. To explore these conditions, this book will focus on the following research questions:
Do the mass media provide sufficient information for voters to monitor their political representatives?
When this information is provided, are the people most interested in the issue also more likely to learn about it?
Do voters rely more on the issue they care about when casting their ballots? Under what circumstances does this occur?
participation rates increase when issues voters care about are raised in campaigns?
The public opinion literature has clearly shown that, by most standards, the American electorate is generally uninformed about issues of public policy. At the same time, however, scholars have shown that politicians in general, and members of Congress in particular, adhere quite closely to the preferences of their constituents. Legislators are encouraged to be broadly responsive at least in part because they are convinced that the "wrong" vote (or set of votes) could prematurely end their career. On its face, this belief does not seem well founded, given what we know about the attentiveness of the average voter. I maintain that the gap in these findings can be bridged, in part, by paying greater attention to the issues voters care about. Such an emphasis takes into account the relative lack of interest among most citizens with regard to many issues of public policy yet still helps to explain why politicians are responsive.
That the public is often asleep does not mean that representatives can simply disregard their constituents' interests. On the contrary, the theme running throughout this book is that legislators are responsive to their constituents because of a realistic fear that interested voters can become informed and hold discrepant votes against them. It is true that part of the reason that citizen monitoring does not regularly occur is that voters do not often possess sufficient interest. However, the public is also inattentive to politics because their politicians do not typically behave in a way that departs from the expected. The ensuing chapters will show that when voters are unusually interested, and particularly when a politician's voting records or issue positions receive unusual coverage, voter monitoring can be considerable. Thus, the sleeping giants that make up public opinion may not often stir, but when they do politicians would be foolish to disregard them. More importantly, a wise representative will see to it that his or her constituents are not motivated to closely monitor legislative activities. The best way to accomplish this is by being responsive--at least on the issues significant voting blocs find important and are likely to find out about.
Finally, it should be noted that this book focuses entirely on the ways in which citizens react to the policy decisions of their elected, or prospective, representatives. Although the model summarized in figure 1.1 links these reactions, or the anticipation of them, to political outcomes, this step in the representation process is not directly explored here. However, two previous studies have confronted this issue. In the first (Hutchings 1998), I showed that, at least under some circumstances, legislators are concerned with the potential preferences of their constituents. Examining support among southern Democrats for the highly publicized 1990 Civil Rights Act, and an equally important yet obscure amendment to this legislation, I found that the size of the black constituency was a much more significant determinant of support for the 1990 act than for the amendment. This suggests that these typically moderate legislators are more likely to support expansive civil rights legislation when their African American constituents might learn of their vote.
In a follow-up piece (Hutchings, McClerking, and Charles 2000), the range of bills was broadened to include both civil rights-related and social welfare votes across three separate congresses. Consistent with the earlier work, we found that white southern Democrats with significant black constituencies were more likely to support "black interests" on votes that received more media coverage. These results provide some support for the contention that concern with voter monitoring can significantly affect legislative outcomes. In the following chapters, I explain why this concern is justified.
Return to Book Description
File created: 8/7/2007
Questions and comments to: firstname.lastname@example.org
Princeton University Press