In The Good Citizen Michael Schudson describes four interconnected but ultimately distinct eras of American civic life, each characterized by the dominance of a particular model of citizenship. In the first era, roughly corresponding to the 18th and early 19th centuries, citizens deferred to the leadership of political elites — civic responsibility consisted mainly of affirming the legitimacy of this ruling caste. The second era, in place throughout the remainder of the 19th century, was characterized by the dominance of political parties. In this period, citizens played a more central role, though this role was orchestrated by strong local party organizations that mobilized the masses through the tangible incentives of patronage, entertainment and other individual, material rewards rather than through detailed appeals to ideology or issues.

In the late 19th and early 20th centuries, yet another transformation occurred, in large part a reaction to the particular brand of partisan politics that proceeded it. This era, in place until the 1950s, was characterized by two, somewhat competing models. The dominant model, emerging from Progressive reforms, emphasized managerial efficiency, a non-partisan professional press, and government by experts. The second, less dominant model, focused on the importance of the direct input of citizens into the substance of politics and policymaking and was characterized, on the one hand, by local discussion groups, salons, and other forms of civic deliberation, and on the other, by the emerging art and science of public opinion polls. The final era described by Professor Schudson, beginning in the 1950s and characterizing much of today’s non-electoral politics, is dominated by the "rights-conscious" citizen. In this model, individual and collective rights drive the plot lines of politics, and the judicial rather than the executive or legislative branches becomes the center stage on which these dramas unfold.

The description of civic life in The Good Citizen is more nuanced than this brief overview suggests. Professor Schudson is careful in showing that there were alternative models in play in each of these eras; that elements of deference, partisan politics, expertise, direct democracy, and civil rights could be found in each era; and that each new model of citizenship overlaid rather than replaced prior models. He is also careful in pointing out that each of these models of civic life carry with them positive and negative implications for the quality of democracy. But running throughout The Good Citizen is a consistent theme: that the role of the citizen in American civic life has always been more circumscribed in practice than idealized models of "rule by the people" would imply and that, viewed historically, the current state of civic life is at a minimum no less vibrant than in past eras and is arguably a preferable mix of elite and mass democracy.

While there are numerous strands to this argument, it is played out most directly in Professor Schudson’s critique of what he views as the problematic ideal of "the informed citizen" which emerged most explicitly at the end of the 19^th century. In a recent address at the 1999 James K. Batten Awards and Symposium for Excellence in Civic Journalism, he elaborated on his concerns. Professor Schudson argues that this ideal has few roots in the theories and practice of American democracy which proceeded this era; that its ascendancy in the 20^th century has led to a sanitizing of politics, stripping it of the visceral, emotional elements that served as powerful motivating forces in earlier eras; that it sped the rejection of partisan politics and of a politics "intertwined and inextricable from every-day social life and social relations;" and that it has created such impossible intellectual demands on citizens that it serves to weaken rather than strengthen efforts to create a more participatory, democratic civic life.

In the conclusion of The Good Citizen, Professor Schudson writes that "the model of the informed citizen…. still holds a cherished place in our array of political values, as I think it should, but it requires some modification" (p.309). His recommendation is for a more realistic model in which most citizens regularly "monitor" or "scan" the political and social environment, learning enough to be "poised for action if action is required" (p. 311). And while he acknowledges that "[t]here is surely some line of willful ignorance that, once crossed, crosses out democracy itself… [and that the] teaching of democracy and the modeling of democracy should never stop," he also argues that "we should have in view plausible aims that integrate citizenry competence with specialized expert resources" (pp. 311-312). In the end, "[t]here must be some distribution across people and across issues of the cognitive demands of self-government" (p.310).

That the ideal of the informed citizen has been used (consciously and unconsciously) to restrict democracy is incontrovertible: one of the strongest arguments against democracy over the past 2000 years has been the fear that "the masses" lacked the intelligence and knowledge to exercise their power in a responsible way. As Professor Schudson documents, throughout the late 19^th and most of the 20^th centuries the failure of citizens to meet often unrealistic standards was the stated justification for many of the laws and norms designed to limit the franchise to anglo, male, economically well off, and educated citizens. And much of the power behind progressive-era arguments in favor of government-by-experts derived from the belief that average citizens lacked the ability to govern themselves. Certainly it is no coincidence that the rise of the informed citizen model is paralleled by a precipitous decline in voter turnout.

Professor Schudson would find a great deal of support for his argument in much of the political science literature. Public opinion polling since the 1930s has consistently documented low levels of political knowledge among the American public, leading Philip Converse to write that "the most familiar fact to arise from sample surveys is that popular levels of information about public affairs are, from the point of view of an informed observer, astonishingly low" (1975: p. 79). Imbedded, as this research was, in an era in which the model of the informed citizen was dominant, these findings produced a great deal of concern. "It seems remarkable," wrote Bernard Berelson, Paul Lazarsfeld, and William McPhee, "that democracies have survived through the centuries…. That is the paradox. Individual voters today seem unable to satisfy the requirements for a democratic system of government outlined by political theorists" (1954: p.312).

It is not too great a simplification to suggest that most of the public opinion theory and research that has emerged over the last 40 years has been an attempt to resolve this apparent paradox. There is a consensus that most citizens are politically uninformed. There is no consensus, however, on the causes or implications of this state of civic affairs. Many observers, starting from the premise that an informed citizenry is the sine qua non of democracy, conclude that American politics is in crisis: that the tensions inherent in its theory and practice have made it either ungovernable, undemocratic, or both. Robert Entman, in his aptly titled book, Democracy Without Citizens, argues that "people who participate regularly and knowledgeably form a distinct minority," and thus, the U.S. system "represents the general public less well than Americans deserve" (1989: p. 28). Paul Blumberg puts it more starkly:

Much of this research is motivated by a desire to salvage liberal democracy from its critics, to show, as Benjamin Page and Robert Shapiro have put it, that "ordinary citizens are not to be feared" and that "skepticism and disdain [for the civic capacity of the public] are not well founded" (1992: pp. xi, 1). In the remainder of this paper I would like to summarize what the political science literature tells us about what Americans know about politics and why it matters, drawing heavily from my own work (co-authored with Scott Keeter) in this area. In providing this overview I will attempt to do justice to the various points of view that are reflected in the literature. It is my argument, however, that in attempting to rehabilitate the image of ordinary citizens by downplaying the possibility or necessity of an informed public, scholars run an equally great risk of selling both citizens and democracy short. Put more bluntly, I am suggesting that democracy becomes more responsive and responsible the more informed, and the more equitably informed, is its citizenry.

Over fifty years of survey research on Americans’ knowledge of politics leads to several consistent conclusions. The most powerful and influential of these conclusions is that the "average" citizen is woefully uninformed about political institutions and processes, substantive policies and socioeconomic conditions, and important political actors such as elected officials and political parties (Bennett, 1988; Campbell, et al., 1960; Converse, 1964; Ferejohn, 1990; Neuman, 1986). This conclusion has been reinforced, even mythologized, by popular press accounts of public ignorance, as when a 1986 ABC/Washington Post poll reported that, shortly after the widely covered Geneva summit between Ronald Reagan and Mikhail Gorbachev, a majority of Americans could not name the leader of the Soviet Union. A similar, if less scientific, example was given in a 1991 New York Times column:

Unsurprisingly, the average level of knowledge was low — only 4-in-10 of these questions could be answered correctly by over half of those surveyed. But the average alone does not tell the full story. Many of the more commonly known facts included rudimentary but potentially important pieces of information such as details about the separation of powers across branches and levels of government; the definitions of key terms such as veto, inflation, or party platform; civil rights such as the constitutional guarantee to a trial by jury, free speech, and religious freedom; the stands of presidential candidates and political parties on some of the major issues of the day (such as social security, health care, and foreign relations), social and economic conditions (such as the existence of a budget deficit or surplus, or the illiteracy rate), and the like.

Clearly the average American is poorly informed about politics when compared to an idealized citizen. Another, arguably "fairer" way to assess the state of political knowledge among the American public is to compare current levels of knowledge to past levels. While data allowing for a systematic comparison of knowledge levels over the past 50 years is less comprehensive than one would hope, the evidence strongly suggests that Americans are about as informed about politics today as they were fifty years ago (Bennett, 1988; 1989; Delli Carpini and Keeter, 1996: 105-134; Neuman, 1986: 14-17; Smith, 1989: 159-222).

These findings could be seen as good news or bad news, depending on one’s perspective. The good news is that, despite concerns over the quality of education, the decline in newspaper readership, the rise of soundbite journalism, the explosion in national political issues, and the waning commitment to civic engagement, citizens appear no less informed about politics today than they were half a century ago. The bad news is that despite an unprecedented expansion in public education, a communications revolution that has shattered national and international boundaries, and the increasing relevance of national and international events and policies to the daily lives of Americans, citizens appear no more informed about politics today than they were half a century ago.

One could argue, as does Professor Schudson, that these differences simply reflect the fact that "[t]here must be some distribution across people and across issues of the cognitive demands of self-government" (p.310), and ultimately that civic life must "integrate citizenry competence with specialized expert resources" (pp. 311-312). The problem with this view is that differences in levels of knowledge parallel other, more traditional indicators of political, social, and economic power such as race, gender, class, and age.

The extent to which knowledge levels vary across groups of citizens is clearly seen using data from two surveys conducted in the late 1980s. While the size of the knowledge gaps about national politics vary from item to item, the overall pattern is compelling: men are more informed than women; whites are more informed than blacks; those with higher incomes are more informed than those with lower incomes; and older citizens are more informed than younger ones.

The extent of these differences can be summarized in several ways. Of the 68 questions asked across the two surveys, for only five was the percentage correct for women as high or higher than for men, and in no case was the percentage correct for blacks as high as for whites, or was the percentage correct for for low income citizens as high as that for upper income ones. The comparison across age cohorts reveals a somewhat more variable pattern, though 55 of the 68 questions were answered correctly by a greater percentage of "pre-baby boomers" than "post-baby boomers."

The sizes of these gaps in knowledge are substantial. For example, the median percent correct across all the items in the 1989 survey for men was 1.35 times that for women, the median percent correct for pre-baby boomers was 1.38 times that for post-baby boomers, the median percent correct for more affluent citizens was 1.59 times that of relatively poor citizens, and the median percent correct for whites was over twice that for blacks.

The cumulative effect of these question-by-question differences can be gauged by summing across all the items to make a knowledge index. Fully three quarters of the women in the 1989 survey scored below the median for men. Substantially more than three quarters of those from families earning under $20,000 a year scored below the median for those earning over $50,000, as was the case for post-baby boomers when compared to pre-baby boomers. And three quarters of black Americans scored below three quarters of white Americans, a knowledge gap of dramatic proportions. Similar patterns were found in the 1988 data.

As a final demonstration of the extent of group differences in political knowledge, one can compare the average scores on the two knowledge scales (measured as the percent of the questions answered correctly) for members of different segments of the population. The average score for the total 1989 sample was 49 percent while for the 1988 sample it was 50 percent, meaning that the "typical" citizen could answer about half the questions correctly. However, this average masks substantial differences across different segments of the population. These differences are especially dramatic when considered for groups of citizens that combine the advantages and disadvantages associated with age, class, race, and gender. The most informed citizens were older, white males whose family income exceeded $50,000 (65 percent correct on the 1989 scale and 76 percent correct on the 1988 scale). These scores were over two and a half times higher than those achieved by the least informed group in our sample: younger black women whose family income was less than $20,000 a year. More generally, the patterns demonstrated in both samples show the exceptionally close fit between political knowledge and socioeconomic status. Surprisingly, the size of the race, gender, and class knowledge gaps have remained relatively unchanged over the past 40 years, and the size of the generational knowledge gap appears to have increased.

Politics is ultimately about "who gets what" from government, or as David Easton (1965) put it, "the authoritative allocation of goods, services, and values." With this in mind, evidence of systematic differences in political knowledge that are tied to other socioeconomic indicators of political power should give one pause. The political significance of these knowledge gaps depends, however, on whether or not knowledge matters to effective citizenship. While there is some disagreement on this, my own work and my reading of the larger literature strongly suggests that informed citizens are "better" citizens in a number of ways.

Specifically, research has found that more-informed citizens are more accepting of democratic norms such as political tolerance, are more efficacious about politics, are more likely to be interested in, follow and discuss politics, and are more likely to participate in politics in a variety of ways, including voting, working for a political party, and attending local community meetings (Delli Carpini and Keeter, 1996; Junn, 1991; Leighly, 1991; Marcus, et al., 1995; Verba, Schlozman, and Brady, 1995). Research also suggests that more-informed citizens are more likely to have opinions about the pressing issues of the day (Delli Carpini and Keeter, 1996; Krosnick and Milburn, 1990), are more likely to hold stable opinions over time (Delli Carpini and Keeter, 1996; Erikson and Knight, 1993; Feldman, 1989), are more likely to hold opinions that are ideologically consistent with each other (Converse, 1964; Delli Carpini and Keeter, 1996: 235-238; McCloskey and Zaller, 1984: 250-251; Neuman, 1986: 64-67; Nie, Verba, and Petrocik, 1979: 154; Stimson, 1975; Zaller, 1986: 10-11), and are less likely to change their opinions in the face of new but tangential or misleading information (Kinder and Sanders, 1990; Lanoue, 1992) but more likely to change in the face of new relevant or compelling information (Zaller, 1992).

There is also evidence that political knowledge affects the opinions held by different socioeconomic groups (for example, groups based on race, class, gender, and age differences). More-informed citizens within these groups hold opinions that are both significantly different from less-informed citizens with similar demographic characteristics, and that are arguably more consistent with their material circumstances (Delli Carpini and Keeter, 1996: 238-251). For example, informed women are more supportive of government programs designed to protect women’s rights, informed but economically disadvantaged citizens are more supportive of government programs designed to provide jobs and improve their standard of living, and so forth. These group differences are large enough to suggest that aggregate opinion on a number of political issues would be significantly different and more representative of the public interest were citizens more fully and equitably informed about politics (Delli Carpini and Keeter, 1996; Althaus, 1998).

Finally, political knowledge seems to increase citizens’ ability to consistently connect their policy views to their evaluations of public officials and political parties, as well as to their political behavior. For example, more-informed citizens are more likely to identify with the political party, approve of the performance of office holders, and vote for candidates, whose policy stands are most consistent with their own views (Alvarez, 1997; Delli Carpini and Keeter, 1996: 251-258).

While the evidence that knowledge matters, and thus that systematic differences in levels of knowledge are problematic, is compelling, there are a number of arguments which, if correct, could serve to lessen or eliminate these concerns. Two of the most powerful arguments — the "heuristic model" and the "on-line processing model" — focus on the way individuals make political decisions.

One of the major criticisms of the "informed citizen" model is that it expects citizens "to yield an unlimited quantity of public spirit, interest, curiosity, and effort" (Lippmann, 1925:2), thus setting standards so high as to make democracy impossible (Schattschneider, 1960: 134-136). An alternative view is that citizens can make reasonably effective decisions even if they are only moderately-informed (Berent and Krosnick, 1992; Graber, 1988; Hurwitz and Peffley, 1987; Peffley and Hurwitz, 1992; Lau and Sears, 1986; Popkin, 1991; Shapiro, et al., 1991; Stroh, 1992). With only a few exceptions this model accepts the three assumptions that drive the informed citizen model — that beliefs are the mainspring of attitude formation; that beliefs can be based on more or less accurate information; and that attitude formation and expression is an active process. However, citizens are further seen as "cognitive misers" (Hewstone and Macrae, 1994) who attempt to make efficient, rational decisions under circumstances of limited ability to process information, limited incentives to become politically engaged, and limited information (Downs, 1957; Mondak, 1994; Popkin, 1991). Citizens achieve this low-information rationality through the use of information short-cuts or heuristics:

The notion of heuristic decision making is rooted in Anthony Downs’ economic theory of democracy (1957) and research by psychologists Daniel Kahneman and Amos Tversky (1972; 1973; 1982a; 1982b; 1984; Tversky and Kahneman, 1973; 1974; 1980; 1981; 1982a; 1982b). Kahneman and Tversky identified four different simplifying heuristics: representativeness, availability, adjustment and anchoring, and simulation. Representativeness is assigning an item to a particular class and then using what one believes about that class to form opinions about the item in question. For example, I know Bill Clinton is a Democrat, so I use what I believe about Democrats to make judgements about him.

Availability refers to the ease with which an individual can retrieve relevant information from long-term memory. For example, in being asked my opinion about the job Bill Clinton is doing as president, I might easily recall that he recently raised taxes, and so give him an unfavorable rating, since I am opposed to raising taxes. Anchoring and Adjustment is a simplifying process in which individuals form an initial response, and then adjust that response by considering additional information related to that response. For example, I might give Clinton an unfavorable rating based on his raising taxes, but then adjust my opinion in a more favorable direction as I think of ways in which he might have improved economic conditions. Thus, my initial opinion anchors my subsequent reflections.

Finally, simulation "facilitates decision making when information is lacking.... decision makers mentally play out [hypothetical] sequences of events relevant to the judgement under consideration" (Mondak, 1994: 123). For example, in deciding whether to vote for Bill Clinton or Bob Dole, I draw on easily accessible information and beliefs to "predict" how each candidate might address issues of importance to me.

Popkin (1991) uses both representativeness and availability in theorizing about how citizens are able to use heuristics in coming to political judgements, and Ottati and Wyer (1990) and Iyengar (1990) discuss an "accessibility" heuristic that is similar to "availability." Ottati and Wyer (1990) also discuss the use of "stereotypes" in a way that is similar to Kahneman and Tversky’s representativeness heuristic. In addition, political scientists have hypothesized and tested other heuristics. For example, Sniderman, et al., (1986) and Sniderman, Brody, and Telock (1991) refer to a "desert heuristic" in which individuals make political judgements based on whether they believe an individual or group is deserving of the action or policy in question. And Riggle (1992) and Riggle, et al., (1992) distinguish "procedural" heuristics (rules for how information should be processed) from "categorical" heuristics (rules for what kinds of information should be use in different circumstances).

The heuristic model goes a long way towards reconciling evidence of low levels of information with the assumption that citizens can make reasoned decisions that reflect their true preferences. The distinction between the heuristic and informed citizen models is less sharp that often suggested, however, and four related issues make it unclear whether the heuristic model offers a satisfying solution to the paradox of a democracy based on poorly and inequitably informed citizens.

First, both the informed voter and the heuristic models assume citizens come to political judgement with less than full information. Research suggests that even elites such as foreign policy makers make decisions under conditions of imperfect information and use heuristics in making decisions (Jervis, 1976; Khong, 1992; Larson, 1985). The use of short cuts describes a human condition rather than a particular form of decision making. Thus, the issue in both models is not whether people use partial information to make decisions, but the reliability, validity, and relevance of the information used.

Second, the heuristic model is based on low information rationality, not no information rationality. Heuristic models assume that citizens are able to use short cuts precisely because they can draw on relevant information stored in long-term memory. True, the heuristic model suggests that many of the "textbook" facts tapped in quizzes of the public may be unnecessary for making reasoned judgements (Graber, 1994). However, much of the information that is necessary for heuristic decision making — for example, the party affiliations, ideological leanings, past issue stands, and personal characteristics of public figures — is precisely the kind of information that many citizens lack.

Third, while research suggests that many citizens can make reasonably good decisions based on limited information, it also suggests that the process by which such decisions are made, and the quality of the ultimate decisions, is still dependent upon the amount and quality of information citizens have available to them (Sniderman, Brody, and Telock, 1991; Riggle, et al., 1992). This is an especially important consideration, given the systematic group differences in political knowledge documented earlier in this paper.

And fourth, while most of the political science literature focuses on the value of heuristics in reaching decisions that accurately reflect one’s true preferences, much of the psychological literature in this area emphasizes the tendency for such simplified processes to lead to decision errors. At some point the amount or quality of information used for making decisions can become so limited as to be useless or misleading (Kuklinski, et al., 1997). For example, a large percentage of those who voted for George Bush in 1988 did so because they wrongly inferred specific policies to the prior Reagan-Bush administration (Delli Carpini and Keeter, 1996: 263-264). Similarly, heuristic decision making is often at the heart of many of the negative (and inaccurate) stereotypes that drive problematic ethnic and racial attitudes and behaviors (Peffley and Shields, 1996; Smith, 1996).

While the informed citizen and heuristic models differ in their views about how much and what kinds of factual information is necessary for citizens to make political decisions, both see beliefs — or cognitive assumptions about what is true — as the driving force of attitude formation and expression. Alternative approaches, while acknowledging that beliefs matter, place a more central emphasis on the role of affect or emotion.

One well-developed approach to the interaction of emotions and factual information in attitude formation has been developed by Lodge, McGraw, and Stroh (1989; 1990; Also see Anderson and Hubert, 1963; Lodge Steenbergen and Brau, 1995; Sanbonmatsu and Fazio, 1990). According to this model of information processing (known alternately as the "impression driven" or "on-line" model), individuals make political evaluations at the moment information is presented, storing their affective impressions in memory and then "‘forgetting’ the actual pieces of evidence that contributed to the evaluation" (Lodge, McGraw and Stroh, 1989: 401). Affective judgements — rather than factual information — about particular individuals, groups, or issues are mentally stored in a running tally that is updated when new information is encountered. It is these emotional tallies that are retrieved into short-term memory when citizens encounter new information and/or make decisions about the person, group, or issue in question.

The "on-line model" differs from both the informed citizen and heuristic model in two important respects. First, it suggests that findings of generally low recognition and recall of political facts tell us little about people’s exposure to or use of political information. Citizens may have little memory of such facts, yet have used them to develop their attitudes. For example, I may be able to tell you I disapprove of the job the president is doing, and have based that opinion on a wealth of factual information, but be unable to recall what those specific facts are. Second, it suggests that people’s political decisions are driven by affective rather than cognitive schema — citizens come to political judgement about many issues through visceral emotions rather than deliberation and thought. In this model, political sophistication is defined as the speed and efficiency with which citizens can process factual information into affective tallies. At best, tests of factual knowledge are indicators of one’s cognitive processing ability, rather than substantively important pieces of information that are called up for active use in forming and expressing political opinions.

Emotions have also been found to play a role in heuristic decision making. One example is the "likability heursistic" (Brady and Sniderman, 1985; Carmines and Kuklinski, 1990; Sniderman, Brody, and Tetlock, 1991). As the name implies, this model assumes that citizens use short cuts in making political decisions. However, these short cuts are driven by how one feels about the issue, person, or group in question. In the version of this model developed by Brady and Sniderman (1985) and Sniderman, Brody, and Tetlock (1991), citizens infer stands to individuals and groups by attributing their own views to individuals and groups they like, and attributing opposing views to those they dislike. For example, if I am pro-gun control, and like Bill Clinton, then I assume he is pro-gun control as well. Carmines and Kuklinski (1990) also assume that affect (likability) drives decision making, but argue that one’s feelings towards the individual or group, coupled with beliefs about where they stand, cues citizens as to where they themselves stand on the issue in question. For example, if I like Bill Clinton, and I believe he supports gun control, then I decide that I, too, must support gun control. While the direction of causality is important, the point here is that both models see affect, rather than beliefs or knowledge, as the mainspring of attitude formation and change.

Recently, Lodge and Taber (1996) have further developed the "on-line" model, combining it with the concepts of hot cognitions and heuristic decision-making to develop a theory of motivated political reasoning. According to this theory, all social information is affectively charged at the moment the information is encountered, and this "affective tag" is stored directly with the concept in long term memory (p. 2). These hot cognitions (Abelson, 1963) are then updated and revised in the face of new information through the on-line process discussed earlier. Finally, when asked (implicitly or explicitly) to evaluate a political object, people will use the "how-do-I-feel" heuristic (Clore and Isbell, 1996) by moving the affective tally into working memory and using the resulting feelings to guide their response, with negative net tallies producing a negative judgement and positive net tallies producing a positive judgement.

Taken as whole, this research clearly shows that emotions play important, multiple roles in political information processing. They can create moods that affect one’s motivation to attend to or avoid politics, thus affecting the likelihood of learning political facts (Marcus et al., 1996: 52). They can interact with knowledge and beliefs, affecting the way information is perceived, stored, and used (Carmines and Kuklinski, 1990; Lodge, McGraw, and Stroh, 1989; 1990; Lodge and Taber, 1996). And they can substitute for factual information in the formation and expression of political attitudes (Brady and Sniderman, 1985; Lodge and Taber, 1996; Marcus, et al., 1996: 47-51; Sniderman, Brody, and Tetlock, 1991).

What is also clear, is that the specific role played by emotions (and factual knowledge) is context dependent. Lodge, McGraw and Stroh (1989; 1990) found that when experimental conditions encouraged forming immediate impressions (for example, when subjects are told, before being given information about candidates, that they will be asked to evaluate them) political "sophisticates" (significantly, defined as those scoring highest on a test of factual knowledge) are most likely to process new information "on-line." But when the experimental conditions are altered (for example, when subjects are not told they will be asked to make an evaluation until after information is presented) or when the topic being evaluated is relatively complex (for example, a policy issue rather than a candidate), political sophisticates are the most likely to draw on information that is stored in memory. And Lodge and Taber (1996) suggest that the "how-do-I-feel" heuristic is most likely to be employed under certain conditions, including those where affective judgement is called for, where the consequences of being wrong are minor, where objective information is not readily available, where disconfirming evidence is not highlighted, and where one is distracted or under time pressure (p. 3).

More research is needed on the conditions under which various information-processing strategies are employed, and the specific roles of factual information, beliefs, and emotions in these various strategies. In addition, more research is needed on the impact of different information processing strategies on the quality of resulting opinions and behaviors. While research suggests that misinformation and/or certain heuristics can lead to poor decisions, there is little research on the potentially negative consequences of emotion-driven decision making.

In addition to these individual-based theories that potentially mitigate the low and varied levels of knowledge among citizens, there are also several more systemic, collective theories which, if correct, would also suggest that concerns about a poorly and inequitably informed citizenry are misplaced. It is possible that, while there are indisputable gaps in knowledge, these gaps mask a more equitable distribution among the most informed, active citizens. Drawing on the logic of elite theories of democracy, perhaps a focus on the general public, while well-intentioned, is overly idealistic. According to this argument, in the real world of liberal representative democracy, meaningful civic engagement is limited — by choice or necessity — to a relatively small percentage of citizens. These "watchdogs" keep government honest, and, if necessary, sound the periodic alarms that mobilize less engaged citizens to action.

There is evidence in defense of this notion of limited democracy. For example, voters are more informed than non-voters, the former averaging 25.4 correct answers on the 1989 index of knowledge as compared to 19.3 for the latter. Strong partisans are also more informed than true independents, averaging 24.1 and 21.4 respectively. Similar patterns are found using the 1988 data: voters were more informed than non-voters (12.2 correct answers to 7.1) and strong partisans were more informed than independents (11.7 to 7.2). However, these findings conceal important differences among groups whose political views are likely to differ. For example, while partisans are more informed than non-partisans, strong Republicans are significantly more informed (26.1 in the 1989 survey and 13.3 in the 1988 survey) than are strong Democrats (22.7 and 10.3). Further, voters and partisans are not random subsets of the general population: the very groups most likely to be politically uninformed are often the most likely to be underrepresented among these more activist citizens. Finally, politics in the United States extends well beyond partisan politics and voting in periodic elections.

A more direct test of the elite model is to examine the demographic makeup of the most informed segment of society. If women, blacks, the poor, and the young are fairly represented within this "guardian class," then discrepancies in the larger population, while perhaps still a matter of concern, become less serious. A comparison of the demographic makeup of the population to that of the most informed fifth of the population strongly suggests that this is not the case, however. For example, women (who constitute over half the population) make up only 29 percent of this information elite. African Americans, about 12 percent of the population, make up only 3 percent of this more informed group. Low income citizens, well over 30 percent of the population, make up only 16 percent of the "information rich." And so on. The under representation of women, blacks, the poor, the young, and their various combinations, coupled with the overrepresentation of men, whites, the affluent, and older citizens is profound, rivaling the demographic distortions found in comparisons of the general public with elected officials. Thus, to the extent that the real world of politics occurs in the exchanges between elected officials, administrative officials, and a small but informed elite citizenry, this conversation is one that mutes the voices of a large segment of the American public.

Alternative readings of the state civic life have attempted to address the normative and empirical shortcomings of elite democracy. One such approach is to distinguish aggregate from individual public opinion. Benjamin Page and Robert Shapiro, echoing and refining the view first expressed by Aristotle, describe the process by which a polity can move from "individual ignorance to collective wisdom":

A final attempt to reconcile low and inequitable levels of knowledge with democratic politics has been to argue that while most Americans are generally under informed, citizens are "information specialists," knowing more about those issues that matter most to them. In this "pluralist" model, citizens are able to engage the political system effectively on the issues they know and care about, producing a collective politics that is a reasonable approximation of individual and group interests. The evidence belies this vision, however. Despite some evidence of specialization, knowledge about different areas of national politics appears to be highly inter-correlated: put simply, citizens who are more informed about one area of politics (for example, foreign affairs) are generally more likely to be informed about other areas of politics (for example, domestic politics, institutions and processes, and/or political actors) (Delli Carpini and Keeter, 1993; 1996: 138-151; Zaller, 1986).

In sum, there is a good deal of theorizing and research suggesting that effective democracy is possible even if citizens are not fully informed about the details of politics and policy. However, all of these theories still require some non-trivial level of individual and/or collective knowledge, all of them concede (or imply) that the quality of decisions improves as the amount of information increases, and none of them adequately addresses the implications of systematic differences in knowledge among different segments of the population.

The American political system is an enigma. It celebrates the individual while longing for a sense of community. It allows almost unlimited participation while doing little to facilitate it. It combines "thick" civic responsibilities with "thin" civic identities. It has emerged as the world's leading democracy, but is partly designed to limit the impact of the vox populi. And, perhaps most fundamentally, it is built upon both an abiding faith in and a deep-seated suspicion of the public. In his later years, Thomas Jefferson often lamented the lack of trust most of his contemporaries had in the general public. While he agreed that people often fell short of the civic ideal, he argued that the political system, by minimizing what was expected of citizens, guaranteed the nature of their public behavior: "We think one side of this experiment has been long enough tried, and proved not to promote the good of the many; and that the other has not been fairly and sufficiently tried" (Jefferson, 1939: p. 44).

I share Jefferson's concern about the lack of trust in the people themselves, a suspicion that, in many respects, is as prevalent today as it was in the late 18th and early 19th centuries. I also share his beliefs that an informed citizenry is the only true repository of the public will; that, given the incentive, education, and opportunity, the general public is capable of exercising political power in an enlightened way; and that the context in which citizens operate — the social, political, and economic structure — is a critical factor in determining whether or not they are motivated and capable.

Ironically, the belief that most citizens simply cannot or will not acquire sufficient political knowledge drives the arguments of both the proponents and critics of contemporary democracy (this tension is also seen in the period, described in The Good Citizen, during which the informed citizen model emerged). It is my view, however, that attempts to salvage democratic theory and practice by downplaying the responsibilities of citizens or the importance of an informed public do an injustice to the very values they seek to defend. Regardless of what conception of democracy one holds — whether thin or thick, direct or indirect — information is necessary for citizens to function effectively. Further, the real world of American politics makes a surprisingly large number of demands on citizens. While it is impossible to identify the specific pieces of information necessary for assuring good citizenship within this context, clearly some information is important, and all other things being equal, more information is better than less information. Arguments that it is irrational for citizens to become politically informed are based on economic models of rational choice, when civic models are normatively and empirically more appropriate. And while it is certainly true that we all take advantage of heuristics that reduce the amount of information necessary for making political decisions, these short cuts themselves require a nontrivial amount of knowledge to be used effectively.

Throughout this paper I have attempted to provide empirical support for the Jeffersonian vision of the importance and possibility of an informed citizenry, while at the same time confronting those places where the American public falls short of this vision. This overview suggests the following.

First, it is nearly meaningless to talk about how much "the public" as an entity knows about politics. While political knowledge levels are, in many instances, depressingly low, they are high enough among some segments of the population, and on some topics, to foster optimism about democratic possibilities. More than a small fraction of the public is reasonably well informed about politics — informed enough to meet high standards of good citizenship. Many of the basic institutions and procedures of government are known to half or more of the public, as are the relative positions of the parties on many major issues of the day. Further, knowledge levels are too high for us to accept the view, offered by some proponents of the rational choice school, that acquiring and retaining information is fundamentally irrational. Indeed, given their socioeconomic and educational status, the people who are politically well informed are precisely the kind we would expect to engage in rational behavior. None of this discounts the need for increasing the level of public knowledge, nor ignores the fact that large numbers of American citizens are woefully under informed and that overall levels of knowledge are modest at best. Nor is it to downplay the often dramatic disparities in knowledge found between the most and least informed citizens, disparities that rival those found in the distribution of income and wealth in the United States. Rather it demonstrates that enough citizens are able to obtain and retain information in the current political environment — an environment that is only partially supportive of this task — to believe that a more fully and equally informed public is possible.

Second, despite the numerous political, economic, and social changes that have occurred since World War II, overall political knowledge levels in the United States are about the same today as they were 40-to-50 years ago. This stability presents the greatest challenge to the notion that political knowledge levels are strongly affected by structural and contextual conditions. At a minimum it underscores how difficult raising aggregate levels of knowledge may be, and could be construed as evidence of the fundamental intransigence of political ignorance. However, it is important to keep in mind that many of the changes occurring over the past half century seem as likely to depress as increase citizens' civic knowledge. For example, while educational attainment levels in the population have risen, increasing the potential for political knowledge, other changes have clearly depressed levels of public interest and engagement in politics among citizens of all educational levels. Thus, I conclude that the stability in political knowledge is the result of offsetting forces. Further, while long-term trends show little change in what citizens know, short-term patterns suggest that, given the right mix of ability, opportunity, and motivation, citizens are capable of significant political learning. This, coupled with the strong and significant relationship between socioeconomic status, the political and information environment, and political knowledge levels, strongly suggests the potential for improvement.

Third, most citizens are political generalists, rather than specialists, meaning that those who are knowledgeable about one aspect of politics tend to be knowledgeable about others. Several exceptions to this general pattern exist, most notably that women, blacks, and partisans are relatively more informed about gender, race, and party issues, respectively, than they are about other political topics, and that knowledge of local politics is somewhat distinct from knowledge of state or national government. Taken as a whole, these findings suggest that different socioeconomic groups are drawn to politics through a variety of distinct pathways, but that in the long run, differences in the ability, opportunity, and motivation to learn about politics in general outweigh differences in the ability, opportunity, and motivation to learn about specific domains of politics. At the individual level this means that someone who has the resources to learn about one aspect of politics is also likely to have the resources to learn about other aspects of politics. At the aggregate level this means that rather than a pluralist information society in which different groups and classes bring different information to the marketplace of ideas, political information of all kinds tends to be concentrated in the same hands.

A fourth conclusion follows from the third. Inequality in citizen knowledge is not simply an idiosyncratic characteristic of individuals. Groups of citizens vary in knowledge in ways that mirror their standings in the social, political, and economic world, calling into question the fundamental democratic principle of equality among citizens. In particular, women, African-Americans, the poor, and the young tend to be substantially less knowledgeable about politics than are men, whites, the affluent, and older citizens. Much of the knowledge gap between these groups persists even when relevant personal characteristics such as education or occupation are taken into account, pointing to a legacy of the long-term exclusion of socio-economically disadvantaged citizens from many aspects of the public sphere. These systematic differences in political knowledge have serious implications for the ability of some groups to perceive and act on their self interest or their notion of the public interest. If Jefferson is right that the people themselves are the best protectors of their own interests, then many groups are hindered in this effort by their relative lack of political information.

Fifth, being politically informed is the result of many factors. As in most spheres of life, motivation is important in learning about politics. Motivation increases with age, education, social status, a sense of efficacy, and a belief that the political world is directly relevant to the individual. But motivation is only one influence. Individuals with higher levels of cognitive skill and relevant contextual knowledge will tend to learn much more about politics than will others. Cognitive skill and contextual knowledge fall under the rubric of "ability," and are strongly related to one's level of formal education. Indeed, education is the strongest single predictor of political knowledge. Yet while personal factors such as ability or motivation strongly affect knowledge levels, the persistence of, for example, the gender gap in knowledge suggests that these characteristics are themselves greatly influenced by cultural and structural factors, and are not solely the result of autonomous personal choices. Moreover, a key element for political learning is the opportunity to do so, which is neither as constant nor as vast as commonly believed. Where adequate political information is available, in a form easily comprehended, citizens learn more. Indeed, the relationship between availability of information and citizen knowledge levels is so strong in certain situations that the nature of the information environment is the most important predictor of knowledge, surpassing education and interest. Overall, while it is true that individuals with the greatest cognitive skills are likely to learn the most about politics, the type of political information needed to function effectively as a citizen is not especially complex and is well within the reach of individuals with modest cognitive ability — given the motivation and opportunity to do so. With all due respect to Albert Einstein, politics is not harder than physics!

Finally and perhaps most important, informed citizens are demonstrably better citizens, as judged by the standards of democratic theory and practice underpinning the American system. They are more likely to participate in politics, more likely to have meaningful, stable attitudes on issues, better able to link their interests with their attitudes, more likely to choose candidates who are consistent with their own attitudes, and more likely to support democratic norms such as extending basic civil liberties to members of unpopular groups. Differences between the best and least-informed citizens on all of these dimensions are sizable. The impact of political knowledge is independent of, and thus over and above, that of other factors such as interest in politics and political efficacy.

^{1. For examples, see Vera Brodsky Lawrence, Music for Patriots, Politicians, and Presidents: Harmonies and Discords of the First Hundred Years (New York: Macmillan, 1975); The Chestnut Brass Company and Friends, Hail to the Chief! American Political Marches, Songs & Dirges of the 1800s, Sony Classical compact disc SFK 62485; Bernard F. Reilly, American Political Prints, 1766-1876: A Catalog of the Collections in the Library of Congress (Boston: G.K. Hall & Co., 1991); Jean H. Baker, Affairs of Party: The Political Culture of Northern Democrats in the Mid-Nineteenth Century (Ithaca, N.Y.: Cornell University Press, 1983); Roger A. Fischer, Tippecanoe and Trinkets Too: The Material Culture of American Presidential Campaigns, 1828-1984 (Urbana, Ill.: University of Illinois Press, 1988).}

2. Donald B. Cole, Martin Van Buren and the American Political System (Princeton: Princeton University Press, 1984), 150-52; Richard H. Brown, "The Missouri Crisis, Slavery, and the Politics of Jacksonianism," South Atlantic Quarterly 65 (Winter 1966): 55-72.

3. Ronald Satz, American Indian Policy in the Jacksonian Era (Lincoln: University of Nebraska Press, 1975), 9-13. See also, Michael Paul Rogin, Fathers and Children: Andrew Jackson and the Subjugation of the American Indian (New York: Alfred A. Knopf, 1975); David S. Heidler and Jeanne T. Heidler, Old Hickory's War: Andrew Jackson and the Quest for Empire (Mechanicsburg, Pa.: Stackpole Books, 1996).

4. Cole, Van Buren and the Political System, 364-66; Satz, American Indian Policy, 211-45; John K. Mahon, History of the Second Seminole War, 1835-1842, revised ed. (Gainesville: University of Florida Press, 1985); Francis Paul Prucha, The Great Father: The United States Government and the American Indians (1984; reprint, Lincoln: University of Nebraska Press, 1995), Volume I.

5. Quotations from John M. Murrin, "The Great Inversion, or Court Versus Country: A Comparison of the Revolution Settlements in England (1688-1721) and America (1776-1816)," in Three British Revolutions: 1641, 1688, 1776, ed. J.G.A. Pocock (Princeton: Princeton University Press, 1980), 425. For a thorough review and strong critique of the historiography on the American state, see Richard R. John, "Governmental Institutions As Agents of Change: Rethinking American Political Development in the Early Republic, 1787-1835," Studies in American Political Development 11 (1997): 347-380. For the term "politics of affiliation," see Michael Schudson, The Good Citizen: A History of American Civic Life (New York: Free Press, 1998), 6.

6. For just a taste of what the Indians faced in terms of government policies and agencies, see Prucha, Great Father, vol. I; Francis Paul Prucha, The Sword of the Republic: The United States Army on the Frontier, 1783-1846 (Bloomington, Ind.: Indiana University Press, 1969); Michael D. Green, The Politics of Indian Removal: Creek Government and Society in Crisis (Lincoln: University of Nebraska Press, 1982).

7. Schudson, Good Citizen, 313-14.

8. Schudson, Good Citizen, 133-43. For judicious assessments of the debates that find much civic merit in them while steering clear of the mythology debunked by Schudson and Blumin, see David Zarefsky, Lincoln, Douglas, and Slavery: In the Crucible of Public Debate (Chicago: University of Chicago Press, 1993); David M. Potter, The Impending Crisis, 1848-1861 (New York: Harper & Row, 1976), 328-55.

9. Lewis Atherton, Main Street on the Middle Border (1954; reprint, Bloomington: Indiana University Press, 1981), 121-42; Carl Bode, The American Lyceum: Town Meeting of the Mind (New York: Oxford University Press, 1956); Jack Larkin, The Reshaping of Everyday Life, 1790-1840 (New York: HarperPerennial, 1989), 232-44, 258-303.

10. The term "participatory democracy" was popularized by the 1962 "Port Huron Statement" of the Students for a Democratic Society and became a rather vague and flexible catchphrase for the New Left more generally. Tom Hayden first used it denote a supplement to representative democracy, but it quickly came more commonly to mean some "radical alternative to representative institutions." See James Miller, "Democracy is in the Streets": From Port Huron to the Siege of Chicago (New York: Simon & Schuster, 1987), 141-54, quotation on 152-53. Given that the whole concept of "participatory democracy" was designed to point up the inherent shortcomings of representative government and the party system, at a time when citizen participation in the party system was at an all-time high, it seems a standard that no party system, especially one that existed in the nineteenth century, could ever attain. Moreover, even the SDS had leaders who worked hard to organize campuses, mount protests, and generate support for their ideas, despite the group's faith in consensual decision-making and spontaneous mass political expression.

11. Examples of pro-party treatises include Frederick Grimke, The Nature and Tendency of Free Institutions [1856], ed. John William Ward (Cambridge, Mass.: Belknap Press of Harvard University Press, 1968); and Martin Van Buren, Inquiry into the Origin and Course of Political Parties in the United States (1867; reprint, New York: Augustus M. Kelley, 1967). For a convenient sampler illustrating the efflorescence of romantic democratic thought in the antebellum decades, see Joseph L. Blau, ed., Social Theories of Jacksonian Democracy: Representative Writings of the Period (Indianapolis: Bobbs-Merrill, 1964). For broader discussions of nineteenth-century thought on parties, see Michael L. Wallace, "Ideologies of Party in the Ante-Bellum Republic" (Ph.D. diss., Columbia University, 1973); and Richard Hofstadter, The Idea of a Party System: The Rise of Legitimate Opposition in the United States, 1780-1840 (Berkeley: University of California Press, 1969). For examples of nineteenth-century Americans' deep feelings for party, see Baker, Affairs of Party.

12. Charles Grandison Finney, Lectures on Revivals of Religion [1835], ed. William G. McLoughlin (Cambridge, Mass.: Belknap Press of Harvard University Press, 1960), 150, 181, 193, 272.

13. On antiparty movements that started parties, see, Richard H. Sewell, Ballots for Freedom: Antislavery Politics in the United States, 1837-1860 (1976; reprint, New York: Norton, 1980); Paul Goodman, Towards a Christian Republic: Antimasonry and the Great Transition in New England (New York: Oxford University Press, 1988); Tyler Anbinder, Nativism and Slavery: The Northern Know Nothings and the Politics of the 1850s (New York: Oxford University Press, 1992); Mark Voss-Hubbard, "The 'Third Party Tradition' Reconsidered: Third Parties and American Public Life, 1830-1900," Journal of American History 86 (June 1999): 121-50.

14. Michael E. McGerr, The Decline of Popular Politics: The American North, 1865-1928 (New York: Oxford University Press, 1986), 3-41; Schudson, Good Citizen, 228.

15. Works advancing or reflecting this interpretation include: William Nisbet Chambers and Walter Dean Burnham, eds., The American Party Systems: Stages of Political Development (New York: Oxford University Press, 1967); Ronald P. Formisano, "Deferential-Participant Politics: The Early Republic's Political Culture, 1789-1840,"American Political Science Review 68 (1974): 473-487; idem, The Birth of Mass Political Parties: Michigan, 1827-1861 (Princeton: Princeton University Press, 1971); idem, The Transformation of Political Culture: Massachusetts Parties, 1790s-1840s (New York: Oxford University Press, 1983); Paul Kleppner, et al., The Evolution of American Electoral Systems (Westport, Conn.: Greenwood Press, 1981); Richard L. McCormick, The Party Period and Public Policy: American Politics From the Age of Jackson to the Progressive Era (New York: Oxford University Press, 1986); Joel H. Silbey, The American Political Nation, 1838-1893 (Stanford, Calif.: Stanford University Press, 1991).

16. Richard P. McCormick, "New Perspectives on Jacksonian Politics," American Historical Review 65 (Jan. 1960): 288-301.

17. See, for example, Ronald P. Formisano, "Federalists and Republicans: Parties, Yes – System, No," in Kleppner, et al., Evolution of American Electoral Systems, 33-76.

18. Maryland is one of the few states for which a good series of election records have been found and studied, and the results (high voter turnouts and consistently partisan electoral behavior) do not support the dominant interpretation. See David A. Bohmer, "The Maryland Electorate and the Concept of a Party System in the Early National Period," in The History of American Electoral Behavior, ed. Joel H. Silbey, Allan G. Bogue, and William H. Flanigan (Princeton: Princeton University Press, 1978), 146-73.

19. Andrew W. Robertson, telephone conversation with author, 10 November 1999. Some of the First Democratization Project data are utilized in Andrew Robertson and Philip Lampi, "The Election of 1800 Revisited," paper presented at the American Historical Association annual meeting, Chicago, Ill., 9 January 2000.

20. David Waldstreicher, In the Midst of Perpetual Fetes: The Making of American Nationalism, 1776-1820 (Chapel Hill: University of North Carolina Press for the Omohundro Institute of Early American History and Culture, 1997). For other works covering aspects of this festive political culture, see Peter Thompson, Rum Punch and Revolution: Taverngoing and Public Life in Eighteenth-Century Philadelphia (Philadelphia: University of Pennsylvania Press, 1999); Simon P. Newman, Parades and the Politics of the Street: Festive Culture in the Early American Republic (Philadelphia: University of Pennsylvania Press, 1997); Len Travers, Celebrating the Fourth: Independence Day and the Rites of Nationalism in the Early Republic (Amherst, Mass.: University of Massachusetts Press, 1997).

21. Jeffrey L. Pasley, "The Tyranny of Printers": The Rise of Newspaper Politics in the Early American Republic (Charlottesville: University Press of Virginia, forthcoming).

22. Wilson Carey McWilliams, "Parties as Civic Associations," in Gerald M. Pomper, ed., Party Renewal in America: Theory and Practice (New York: Praeger, 1980), 51-68; Cornelius P. Cotter and Bernard C. Hennessey, Politics without Power: The National Party Committees (New York: Atherton Press, 1964), 1-38.

23. Schudson, Good Citizen, 240-93.

24. Ibid., 311.

25. Pauline Maier, From Resistance to Revolution: Colonial Radicals and the Development of American Opposition to Britain, 1765-1776 (1972; reprint, New York: Vintage Books, 1974), 3-26; Paul A. Gilje, The Road to Mobocracy: Popular Disorder in New York City, 1763-1834 (Chapel Hill: University of North Carolina Press for the Institute of Early American History and Culture, 1987); idem, Rioting in America (Bloomington: Indiana University Press, 1999); David P. Szatmary, Shays' Rebellion: The Making of an Agrarian Insurrection (Amherst: University of Massachusetts Press, 1980); Thomas P. Slaughter, The Whiskey Rebellion: Frontier Epilogue to the American Revolution (New York: Oxford University Press, 1986).

HOME