Which form of political participation is the most effective and why? Include a news article from the last four weeks that illustrates this form of participation in action and how effective it is. Also

1 1 Voting Together If you were to ask most American citizens why they voted (or did not vote) in the last election, their explanations of their own choices would almost certainly take the form of a few familiar stories. One voter might speak of his deep and abiding interest in politics; another might speak passionately about her desire to see the right person win. A third might have little interest in the election yet still feel the pull of patriotism or civic duty, and yet another might worry that his family or coworkers would think less of him if he failed to make it to the polls. The stories of nonvoters would be similarly familiar stories. One would likely bemoan the fact that deadlines at work or long lines at the polling station had kept her from participating, another would rattle off a list of objection- able actions taken by politicians, and yet another might wistfully reply that people like him did not really matter anyway. Surprisingly, such stories, writ large, form the basis of most schol- arly explanations of variation in voter turnout. Voters, compared with nonvoters, are more likely to find politics interesting and less likely to find participation prohibitively costly. Strong partisans who care about election outcomes are more likely to vote than weak partisans or inde- pendents who care less about the results. Voters are more likely than nonvoters to have the education and skills needed to register and fig- ure out how to get to the polls and cast a ballot. Finally, citizens are more likely to vote if their family, friends, and housemates are voting as well. Traditional theoretical explanations are cast in terms of individual desires, motives, beliefs, and utilities: People who want to vote and are able to vote are more likely to vote, whereas those lacking the desire or the ability will abstain. Rolfe, M. (2012). Voter turnout : A social theory of political participation. Cambridge University Press.

Created from apus on 2022-03-08 18:36:00.

Voter Turnout 2 This book argues that such explanations are at best partial and at worst biased explanations of variation in political participation because they fail to account for the impact of social context. Individual reasoning does not take place within a social vacuum, but depends in various ways on the actions and opinions of other people. When individual decisions are (even in part) conditional or interdependent, the structure of social ties has the power to shape social outcomes to a much larger degree than previously recognized. Therefore, to avoid bias, explanations of variation in political participation and other social phenomena must go beyond stories about isolated individuals to incorporate the potential impact of social interaction on social outcomes. This is not an idle concern for researchers and policy makers, as the omission of social structure and social interaction limits our ability to understand the fundamental mechanisms through which the demographic correlates of American voter turnout drive participation. Over the years, political scientists have documented almost no change in the empirical predictors of turnout. Immigrants, minorities, young people, the uned- ucated, the poor, and the politically disinterested are systematically less likely to vote than those with higher social status (i.e., wealthy, white, and highly educated citizens), much as Merriam and Gosnell ( 1924 ) estab- lished close to a century ago. (Churchgoers and members of voluntary organizations have since been added to the list of turnout predictors.) Education, the most powerful predictor of turnout, is often described as enhancing benefits or reducing costs (see, e.g., Verba, Schlozman, and Brady 1995 ). However, recent work still acknowledges that the mech- anism connecting education to voting is mysterious (Sondheimer and Green 2010 ), and some scholars even question whether education has a causal effect on turnout at all (Tenn 2007 ; see also Nie, Junn, and Stehlik-Barry 1996 ). The social theory of turnout offered in this book, on the other hand, argues that education is merely a proxy for belonging to a social world whose members have systematically different patterns of social relationships. As a result of incomplete understandings of the causes of turnout, policies designed to increase participation have often failed to achieve their aims. Many recent voting reforms, including absentee voting and Motor Voter legislation, were explicitly designed to lower the costs of participating. Schemes that actively encourage voting at home by mail or the Internet have proven ineffective or even reduced turnout (Berinsky 2005 ), at least after the relaxation of efforts put into marketing such alternative modes of participation. Reducing registration requirements Rolfe, M. (2012). Voter turnout : A social theory of political participation. Cambridge University Press.

Created from apus on 2022-03-08 18:36:00.

Voting Together 3 should raise turnout in the long run, although the impact may take some time to show up. Such cost-reduction policies may backfire, however, if the symbolic virtue of voting is reinforced by the requirement that one stand in lines and follow arcane procedures, or if citizens are less likely to behave in a public-spirited way if their neighbors are not at the polls to greet (and observe) them. This book takes a fresh perspective on this long-standing area of research, one that explicitly accounts for how social and political context drives political participation, and provides new insight into the empiri- cal correlates of voter turnout. The fundamental approach ties together existing strands drawn from scholarship on voting and other social phe- nomena. The ability of social interaction to alter social outcomes will not come as a surprise to game theorists, who have long been aware that most people condition their behavior on the expected choices of those around them. Behavioral economics, inspired by the groundbreak- ing work of Simon on bounded rationality and Kahneman and Tversky on heuristics, provides a firm foundation for the book’s core conditional decision-making model. Key links in the social theory are pulled from long-standing research into the impact of social and political context on mass political behavior. None of these threads contain a full-fledged account of individual decision making and social dynamics akin to con- ditional choice, but they contain important elements that remain present in the social theory of turnout.

the social theory of voter turnout So why do some people turn out to vote whereas others do not? This book proposes and argues for a social theory of voter turnout, grounded in the conditional choice approach. This theory places voters not only in a social context, but also assumes a less familiar logic of decision making. Conditional decision makers rely on conditional decision rules, sometimes termed heuristics or cognitive shortcuts, rather than optimiz- ing payoffs in a forward-looking manner. Thus, it is possible to build a behavioral model of turnout based on a distribution of decision rules in situations like voting without solving the paradox of voter turnout.

The social theory of participation may not speak to individual motives and reasoning, but it can provide a satisfying explanation of documented empirical variation in voter turnout, one in which variation in social net- work structures and citizens’ social locations underlie the well-known demographic patterns. Rolfe, M. (2012). Voter turnout : A social theory of political participation. Cambridge University Press.

Created from apus on 2022-03-08 18:36:00.

Voter Turnout 4 Boiled down to its essence, my argument is that the turnout decision is best represented as a conditionally cooperative response to cooperative decisions made by friends, family, neighbors, and coworkers – members of the social network of the citizen in question. Sometimes a handful of people are willing to cooperate unconditionally, whereas some people will not cooperate under any circumstances. The bulk of the population is willing to cooperate if enough other people will do the same. Having more family members and friends who vote makes one more likely to vote as well. But what of the long-standing empirical correlates of turnout? In the social theory of turnout, these individual-level correlates of voter turnout are reinterpreted as imperfect proxies for social structural variables: var - iation first in the size and structure of the social networks in which indi- viduals are embedded, and, second, in individuals’ social location within a given social network structure. These social structural properties have a primary and direct impact on whether a given individual is likely to vote or to abstain. The general view of turnout as generated by political and social con- texts has several important predecessors in scholarship on voting and other forms of mass political behavior.

conditional choice and conditional decision making This book provides a formal model of the decision to vote, but it is a model built within the theoretical framework of conditional choice (see Rolfe 2009 ). Put simply, conditional choice posits that individual choices are a function of the subjective social meaning of the situation and of the observed and/or expected choices of other people. The conditional choice view is compatible with an assumption of bounded rationality (Simon 1955 ). It is perhaps more persuasive, however, to simply note that most people make decisions that are responsive to the decisions of those around them, at least to some degree. Thus, conditional choice is driven by the reality of observed individ- ual decision making rather than a commitment to a set of assumptions about individual goals and desires. Few people make decisions like clear- headed, forward-looking, goal-directed economists (even if we often wish that we could). Real people continue to vote even when they are not fully informed about the candidates; they will lend their neighbors a cup of sugar, but will also mow the lawn before 8 a.m . on weekends; contrib- ute to charities devoted to world peace, but also yell at their kids; and Rolfe, M. (2012). Voter turnout : A social theory of political participation. Cambridge University Press.

Created from apus on 2022-03-08 18:36:00.

Voting Together 5 iron their shirt each morning for work, but forget to brush dog hair off their jacket. In other words, most people continue to act in ways that are entirely unpredictable if we assume that their behavior is the result a decision-making process that involves the rational, consistent pursuit of personal goals or that can be described by a consistent set of individ- ual traits such as selfishness, altruism, aggression, kindness, laziness, or conscientiousness. How can we understand individual action if individuals do not act con- sistently on the basis of economic rationality, or even predictably on the basis of personal motives, traits, or characteristics (Mischel 1968 )? The answer is to direct attention away from individual decision makers to the social situations in which they find themselves, and to the social interac- tions that take place between them. Nothing of interest to social scientists takes place in a hypothetical social vacuum. 1 All individual action takes place within social situations, and individual action is only intelligible within the social context that gives rise to it. Conditional choice puts social cognition and social interaction – not individual preferences – at the center of individual decision making. How then do individuals make decisions if they are navigating social interactions rather than maximizing payoff functions or minimizing the risk of low payoffs? Conditional actors rely on conditional decision rules, sometimes termed heuristics (Kahneman 2002 ) or rules of thumb (Simon 1955 ).2 People may purposefully condition their actions on those of others, as in a conscious desire to “do one’s fair share.” Alternatively, conditional responses may be automatic and unthinking, operating out- side of the awareness of the individual decision maker. Conditional responsiveness can vary from person to person. Some peo- ple can be largely or even entirely unresponsive to others (“unconditional” actors, whose behavior is a constant when expressed as a mathemati- cal function of others’ decisions). Most people, however, will respond at least somewhat to the actions of others. To account for social interac- tion, conditional decision rules are mathematically modeled as a function of (1) the social meaning of the decision situation, (2) the observed or expected actions of other people, and (3) individual heterogeneity. The 1 Not even laboratory experiments take place in a social vacuum. Subjects enter the lab with a wealth of social knowledge that allows them to interpret all requests and respond accordingly, and leave the lab with that knowledge intact or perhaps even altered by the interactions in the experimental setting. 2 Social interaction is not addressed by the decision rules proposed by the Kahneman and Tvesrsky project. Rolfe, M. (2012). Voter turnout : A social theory of political participation. Cambridge University Press.

Created from apus on 2022-03-08 18:36:00.

Voter Turnout 6 social dynamics of individual decision making in a particular social sit- uation are modeled as a distribution of conditional decision rules – the basis of all conditional choice models. Evidence on the distribution of conditional cooperation comes from behavioral economics, a subfield within economics, which has documented that (seemingly) nonrational “anomalies” in individual decision making are the rule rather than the exception. 3 Considerable empirical evidence supports the conditional interdepen- dence of decision making. Conditional decisions are often consciously made in situations involving public goods or social dilemmas, when individual material benefits come into conflict with what is best for the group. (Voting may be such a social dilemma, or perceived as such by voters – this claim will be discussed in Chapter 3 in the analysis of the social meaning of voting). Conditional decision makers do not seek to maximize their own immediate benefit; rather, they behave in ways that involve cooperation with others but not complete capitulation to them.

Cooperative acts, or contributions to the group, are conditional on the cooperation or contributions of others. The conditional choice approach is neutral on the rationality of deci- sion making, and indeed is silent on the matter of people’s ends or goals; whether or not these ends conform to any external standard is beside the point, for the purposes of this book. What is essential is that people actu- ally do make decisions that reflect social interaction with other people. Thus, a conditional decision model is strictly a mathematical represen- tation of decision outcomes as conditional on the decisions of others, regardless of the rationale, motivation, or mental process of the decision maker (López-Pintado and Watts 2008 ; Young 2009 ). It may well be that the conditional cooperator believes his or her actions to be motivated by a desire to please, or to avoid social disapproval, or to gain the trust of a friend, or even to efficiently gather information in an uncertain world, but such possible motivations are all reducible to a conditional mathe - matical function. Regardless of the motivational story that might be told, the widespread use of conditional decision making has been demon- strated many times in both the real world and in researchers’ artificially constructed lab situations and social dilemmas.

3 Experimental outcomes may seem less anomalous after accounting for what Podolny (2001 ) terms the prismatic aspect of social networks: how social interaction, in particu- lar asymmetries in power and influence in social relationships, shape individual decision making. Rolfe, M. (2012). Voter turnout : A social theory of political participation. Cambridge University Press.

Created from apus on 2022-03-08 18:36:00.

Voting Together 7 With conditional choice as a foundation, the social theory of turn- out will remain silent on the immediate, conscious reasons people invoke to explain why they get out of bed on Election Day and plan a trip to the polls. A motivational story is probably a common expectation for a theory purporting to explain why people vote, but my focus is not on the inner mental states associated with voting, but rather on the mecha- nisms linking political mobilization and social structure to observed vari- ation in turnout rates. I will not attempt to document “social pressure” to vote, or any other such carrier of interpersonal influence. In fact, I would expect that most conditional voters do not experience their decision as one motivated by social pressure, nor are they aware that their decision has been influenced by the choices of those around them. Rather, the level of explanation here focuses on the relationship between social structure and turnout rates. Perhaps this will be a sufficient introduction to the perspective of this book; satisfied readers may wish to move on to the development of con- ditional choice in Chapter 2, or even to the second half of the book, in which conditional choice is translated into specific predictions about voter turnout that are then tested empirically against the predictions of the extant view. For others, the remainder of this chapter presents com- peting views on voter turnout in more detail.

the orthodox view: voters making decisions As citizens in a democracy, Americans are by definition entitled to a voice in the affairs of their country, most commonly in the form of voting in elections. Indeed, Americans are called to the ballot box more often than citizens of most, if not all, other democratic nations, casting ballots in Presidential elections, midterm Congressional races, assorted state and local elections, referenda and initiatives, as well as primaries for the same.

The right to vote is regarded with reverence in American civic culture. The extension of voting rights to women and African Americans constituted a crucial goal for two of the largest, most significant social movements of twentieth-century America. Why then do half or more of all eligible citizens fail to cast a ballot in most elections? This question inspired some of the most innovative and enduring political science research of the early twentieth century (cf. Merriam and Gosnell 1924 ; Gosnell 1927 ), with a book title summing up the basic research question: Non-Voting . Voting was seen as normatively desirable; nonvoters were described in undesir - able and unfavorable terms. (As we will see in Chapter 3, nonvoting is Rolfe, M. (2012). Voter turnout : A social theory of political participation. Cambridge University Press.

Created from apus on 2022-03-08 18:36:00.

Voter Turnout 8 still given quite harsh treatment in the contemporary media.) The ques- tion at the time was: Why would anyone not do the right thing? Downs ( 1957 ) turned the conventional wisdom on its head, proposing that the turnout decision should be modeled as an economic decision, with individuals weighing the costs and benefits of voting before decid- ing whether to vote or abstain. In this view, an individual voter’s decision would consider the relative benefits of having one party (versus the other) in office ( B), multiplied by the probability of casting the deciding vote ( p), and subtracting out the costs of voting ( C) for the basic economic voting model: pB – C. Unfortunately, the probability of casting the deciding vote is quite small, and therefore abstention is always the choice of utility-maximizing citi- zens. Clearly, most citizens do not always abstain from voting, and thus the divergence between actual behavior and the predictions of the cost- benefit model became known as the paradox of voter turnout, sometimes jokingly described as “the paradox that ate rational choice.” Subsequent formal models have offered a number of solutions to the paradox. The most common solution, typically linked to Riker and Ordeshook ( 1968 ), involves including a D-term in the economic voting model: pB – C + D. The additional mathematical term ( D) can be used to signify any addi- tional benefits that an individual receives from the act of voting. These benefits might be linked to an internalized sense of civic duty (Riker and Ordeshook 1968 ), the expressive thrills of participation (Schuessler 2000 ), a sociotropic desire to support democracy or fellow partisans (Edlin, Gelman, and Kaplan 2007 ), and so on. Regardless of the motiva- tional story, inclusion of a D-term reduces electoral participation to an unexplained taste, akin to a taste for chocolate (Tullock 1967 ; Morton 1991 ; Green and Shapiro 1994 ). If someone eats chocolate, we infer that they like it; if someone votes, we infer that they like it, intrinsically. Such modifications essentially reduce the economic voting model to a tautol- ogy that cannot be disproven: Voting is rational for people who vote and not rational for those who do not vote (Green and Shapiro 1994 ).

Rolfe, M. (2012). Voter turnout : A social theory of political participation. Cambridge University Press.

Created from apus on 2022-03-08 18:36:00.

Voting Together 9 Most other formal attempts to solve the paradox of voter turnout also have been heavily criticized, either because they also produced high levels of abstention or relied on unrealistic assumptions. Early parti- san team models that increased the probability of casting decisive votes (Palfrey and Rosenthal 1985 ) produced knife-edge results: models that could not sustain high turnout following minor changes of assumptions (Morton 1991 ). Minimax regret decision making (Ferejohn and Fiorina 1974 ) turns out to sound much less plausible when extended just slightly beyond the narrow decision of whether to vote or not once at the poll- ing station. For example, scholars have argued that a potential voter using the minimax rule would not vote to avoid getting hit by a car en route to the polling place, or would only vote for themselves (Beck 1975 ; Stephens 1975 ; Tullock 1975 ; Aldrich 1993 ). Partisan mobilization mod- els (Uhlaner 1989 ; Morton 1991 ; Bendor et al. 2003 ) fail to make sense of turnout among the huge number of independent and undecided voters in the American electorate, and suffer from second-order collective action problems (Olson 1965 ; Oliver 1980 ) as well as a lack of evidence (Green and Shapiro 1994 ). Thus, it remains difficult to adequately model voter turnout starting with a Downsian logic of costs and benefits.

The Rational Actor Takes a Survey Despite these debates within the formal literature on voter turnout, the basic paradigm of individual costs and benefits was quickly adopted to explain established patterns in the demographic correlates of participa- tion. A significant tradition of survey research into the causes of (non) voting was already well established (e.g., Merriam and Gosnell 1924 ; Lazarsfeld, Berelson, and Gaudet 1944 ; Campbell 1954 ), and was rapidly expanding with the creation of the American National Election Study (see Campbell, Converse, Stokes, and Miller 1960 ). Multiple attempts were made to incorporate a rational calculus of voting directly into the existing survey research literature, rather than relying on a reinterpre- tation of demographic variables, but these attempts did not meet with great success. First, measurement of the three important elements of the utility model (costs, perceived difference between the candidates or benefits, and possible influence over the election outcome) proved difficult. Direct measures – for example, asking survey respondents how likely they were to cast the decisive vote – were usually found to be unrelated to individual-level Rolfe, M. (2012). Voter turnout : A social theory of political participation. Cambridge University Press.

Created from apus on 2022-03-08 18:36:00.

Voter Turnout 10 turnout (Ferejohn and Fiorina 1975 ). Although individual-level measure of civic duty, the D-term, was found to be associated with higher turnout (Campbell et al. 1960 ), this correlation stemmed from different patterns of survey responses among African Americans, and therefore faded after the passage of the 1965 Voting Rights Act (see Chapter 3). Instead of rejecting the cost-benefit logic as a plausible individual mechanism, however, the literature adapted the interpretations of previ- ously established correlates of turnout to fit better with the compelling new logic. Various measures of political interest, already linked to turn- out by Lazarsfeld et al. ( 1944 ), were recast as measuring the benefit term in the Downsian calculus of voting (Katosh and Traugott 1982 ). Sanders (1980 ) argued that rural location depressed turnout because rural resi- dents had to travel longer to get to polling places, thus increasing the costs of turnout. Income and education, long known to predict turnout, were linked to the calculus of voting through multiple pathways: impact- ing the costs of turnout, the benefits of electoral outcomes, and the ability to more cheaply acquire information needed to distinguish between the candidates (cf. Frey 1971 ; Tollison and Willett 1973 ; Niemi 1976 ). Indeed, the most prominent school of thought within survey-based research on voter turnout is still built on the cost-benefit logic, imagining that individual voters’ chances of voting increase or decrease in response to the personal costs and benefits of turnout, even though the cost-benefit logic has such difficulty making sense of anyone voting at all. Earlier mentions of nonrational considerations, including group consciousness (Verba and Nie 1972 ) and socialization and social identity (Campbell, Converse, Stokes, and Miller 1960 ), have been dropped from more recent work in the same tradition (Verba, Schlozman, and Brady 1995 ). The variables originally identified by Merriam and Gosnell are still the pri- mary predictors of individual turnout, with a few additions. What has changed is that demographic characteristics are now described as proxies for civic resources (costs) and political interest (benefits).

contextual explanations of voter turnout The social theory of voter turnout involves a different sort of explana- tion. Rather than asking what it is about individual voters that makes them more or less likely to vote, the social view emphasizes the explana- tory power of voters’ social and political contexts. As noted earlier, this book is not alone in emphasizing contextual rather than individual deter - minants of turnout. But, whereas the social theory belongs to the same Rolfe, M. (2012). Voter turnout : A social theory of political participation. Cambridge University Press.

Created from apus on 2022-03-08 18:36:00.

Voting Together 11 general family of explanation as these alternatives – and draws directly on some of them – this book explicitly pushes the limits of formal, contex- tual explanation further than previous research. The sections that follow review these various strands of research on turnout in more detail, and point out where the social theory picks up on their insights and where it departs from them.

Institutions Beginning with the contextual factors least integral to the social theory of turnout, it is nonetheless important to acknowledge several institutional features that vary across elections and have been associated with varia- tion in aggregate turnout rates. These include (1) institutionally imposed barriers to voter turnout, such as registration requirements, and 2) elec- toral systems (i.e., proportional representation or first-past-the-post, uni- cameralism or bicameralism, other features of polities that affect how votes are translated into representation). We might also regard closeness of elections as something that might vary with electoral institutions (Cox 1999 ), although more commonly it is seen as varying from election to election as candidates and circumstances change. Empirical research on each of these factors has shown a relationship to turnout rates, often con- ceptualized in cost-benefit terms.

Institutional Costs A wealth of evidence confirms that turnout rates vary at the aggregate level across institutional settings. In American elections, turnout is higher in states that allow absentee registration, have no closing date for reg- istration, and have registration offices regularly open during normal office hours and/or during the evenings and weekends (Wolfinger and Rosenstone 1980 ; Rosenstone and Hansen 1993 ). Prior to the passage of the Voting Rights Act, poll taxes, literacy tests, and periodic re-registration requirements were all effectively used to depress turnout among Blacks in the South (Key 1949 ; Kelley, Ayres and Bowen 1967 ). A significant degree of cross-national variation in voter turnout is also associated with institu- tional costs. Turnout is higher in countries where voters are automatically registered, voting is compulsory, or election day is a weekend or holiday (Powell 1986 ; Jackman 1987 ; Lijphart 1997 ; Franklin 2004 ). All of the institutional features described earlier have been theoret- ically linked to individual turnout costs in the Downsian calculus of voting, with the removal of registration barriers or the introduction of Rolfe, M. (2012). Voter turnout : A social theory of political participation. Cambridge University Press.

Created from apus on 2022-03-08 18:36:00.

Voter Turnout 12 compulsory voting assumed to directly decrease the net costs of partici - pation for an individual. Additionally, many of the studies have incor - porated data from extremely large sample surveys, such as the Current Population Study series run by the U.S. Census Bureau, with a sample size of approximately 60,000. However, repeated attempts to show the impact of institutional costs on a more traditional election study with a smaller sample were largely unsuccessful (Campbell et al. 1960 ; Ashenfelter and Kelley 1975 ; Katosh and Traugott 1982 ). Therefore, although it is clear from the evidence that institutional “costs” can and do shift aggregate turnout at the margins, the impact of such costs has only been reliably established at the contextual – not individual – level. 4 As Blais ( 2006 ) points out, political science still lacks “a compelling microfoundation” for most cross-national findings on voter turnout. This pattern – a finding that institutional costs are related to marginal shifts in aggregate level behavior turnout but have almost no perceptible impact on individual-level behavior – fits nicely within the conditional- choice approach. Conditional decision-making models have already been shown to better fit empirical patterns of decision making in quite signif- icant and costly decisions: the shift in patterns of retirement following a change in the legal age of retirement (Axtell and Epstein 1999 ), and adoption of a new form of hybrid corn among farmers (Young 2009 ). I adopt a similar approach, arguing that institutionally imposed costs of decision making are better understood as shifting the distribution of decision rules (in particular the proportion of first movers), rather than as having a direct and consistent impact at the individual level.

Electoral Institutions Evidence on whether other aspects of the institutional context, particu- larly the electoral system, affect voting turnout is inconsistent. Turnout is generally higher in countries with proportional representation (PR) vot- ing systems (Powell 1986 , Jackman 1987 , Franklin 2004 ), as PR systems are associated with competitive national election districts (Powell 1986 , Jackman 1987 ). Subsequent studies outside of Europe have largely failed to replicate the finding that PR systems encourage turnout (Blais 2006 , for a review), however, and turnout can be depressed where PR voting 4 Although no individual mechanism reliably links institutional costs to aggregate turnout, it is worth noting the low rate of agreement with the characterization of voting as a civic duty among African Americans prior to the passage of the Voting Rights Act, followed by an abrupt increase (up to the levels of white American respondents) afterward. Rolfe, M. (2012). Voter turnout : A social theory of political participation. Cambridge University Press.

Created from apus on 2022-03-08 18:36:00.

Voting Together 13 sustains large numbers of parties (Jackman 1987 ). Similarly, the division of power between various branches of the legislature may impact turnout (Jackman 1987 ), although once again subsequent findings on this insti- tutional feature are mixed (Blais 2006 ). Party mobilization is most often mentioned as the individual-level mechanism through which electoral institutions affect turnout, an issue to which we will return shortly. The probability of being decisive in an election, p in the calculus of vot- ing, was recognized as important in early empirical work (Gosnell 1930 ; Key 1949 ) and was later explicitly linked to aggregate turnout variation (Barzel and Silberberg 1973 ; Settle and Abrams 1976 ). However, other scholars who found a relationship between closer elections (measured as the margin of victory for the winning candidate) and increased turnout continued to describe the margin of victory as a contextual, not individ- ual, variable (Patterson and Caldeira 1983 ). As further confirmation of the contextual impact of electoral competition, Ashenfelter and Kelley (1975 ) found that subjective estimates of the closeness of the presiden- tial race did not have a significant impact on individual turnout in the 1960 and 1972 U.S. presidential elections. However, it appeared that the perceived closeness of the race explained almost all of the marginal shift in aggregate turnout levels between the two elections, with 60% of 1972 respondents, but only 10% of 1960 respondents, stating that they believed the election outcome would not be close.

Mobilization and Strategic Politicians The empirical evidence on whether institutions consistently affect voter turnout may be mixed, but one message from the previous section is clear: When institutions impact political participation, they do so only at the contextual (or aggregate) level. Even though some of the empirical results may be consistent with the individual calculus of voting, there is no empirical support for the calculus of voting acting as the mechanism through which institutional and electoral context might (with the sole exception of institutional costs) affect voter turnout. Mobilization by strategic politicians (Morton 1991 Aldrich 1993 ; Rosenstone and Hansen 1993 ; Cox 1999 ) is the leading candidate to fill this void. Strategic politicians (and their supporters) have a strong incentive to invest in elections that they have a chance to win, and there- fore strategically invest both time and energy into mobilizing voters in close races. Available evidence confirms that campaign expenditures do affect turnout (Patterson and Caldeira 1983 ), and reduces the impact of Rolfe, M. (2012). Voter turnout : A social theory of political participation. Cambridge University Press.

Created from apus on 2022-03-08 18:36:00.

Voter Turnout 14 margin of victory to insignificance in properly specified models (Cox and Munger 1989 ). Political actors invest more in mobilization when elec- tions are expected to be close (Rosenstone and Hansen 1993 ; Aldrich 1993 ). The rise in candidate-centered elections decreased the incentives for parties to invest in mobilization, leading to a subsequent decline in turnout in the United States (Rosenstone and Hansen 1993 ). Cox ( 1997 ; 1999 ) extends the strategic politician’s perspective to the puzzling impact of electoral institutions in a sophisticated formal theory linking voters, political parties, and electoral institutions. Cox argues that electoral rules affect voter turnout only indirectly, via the likely margin of victory and resulting mobilization efforts of strategic political elites.

Political parties with an explicit goal of winning elections (and staying in power) put more or less effort into mobilizing voters depending on how various aspects of the electoral system translate “effort-to-votes, votes-to- seats, and seats-to-portfolios” (Cox 1999 ). Although the strategic mobilization perspective makes sense of an impressive range of contextual-level political participation research, it does not focus on decisions of individual voters and thus does not speak directly to the individual-level survey research on turnout discussed earlier. The conventional wisdom holds that mobilization works by sub - sidizing the costs of participation (e.g., a bus to drive voters to the polls) or by offering additional incentives to vote (e.g., social events for union members), but strategic candidates invest far more money in various forms of advertising than they do in hiring buses to drive local resi - dents to the polls. Moreover, empirical support for the direct impact of mobilization on turnout is based largely on personal contact by political elites and activists. I now turn to consider this rapidly growing body of evidence.

Mobilization and Canvassing Citizens who are asked to vote are more likely to do so (Gosnell 1927 ; Rosenstone and Hansen 1993 ; Verba, Schlozman, and Brady 1995 ). Mobilization clearly influences turnout, but the individual-level mecha- nism is left largely unspecified in the current literature. Its impact does not fit readily with the dominant cost-benefit logic – why should a stranger knocking on my door and urging me to vote for a particular candidate (Gerber and Green 2000 ) change the cost-benefit calculus of my turn- out decision? Yet voters who are contacted by parties report not only higher rates of voting, but that they are more likely to urge others to Rolfe, M. (2012). Voter turnout : A social theory of political participation. Cambridge University Press.

Created from apus on 2022-03-08 18:36:00.

Voting Together 15 vote (Rosenstone and Hansen 1993 ) – an indirect mobilization effect also confirmed by recent field experiments (Nickerson 2008 ). I argue that political mobilization does not effect marginal changes in the payoffs of voting, but increases the salience of the election for those exposed to it. As argued by Rosenstone and Hansen ( 1993 ), cit- izen participation takes place in a political environment with multiple actors competing for attention from citizens. Therefore, it is necessary to specify how activities undertaken by the political elite – candidates, organizations, activists, the media – are translated into changes in indi- vidual turnout probability. I conceive of political activity influencing individual decision making in three distinct ways: (1) providing informa- tion about the opportunity to vote in certain low information settings, (2) temporarily increasing the likelihood that someone will vote as a result of face-to-face contact, and (3) most importantly, increasing the salience of politics and the upcoming election among friends, family, neighbors, and coworkers. It is the last of these three that underlies my claim that virtually all turnout is mobilized, either directly or indirectly, by campaigns and related activity. The entire political environment, the campaigns, per - sonal contacting, and other forms of candidate and political activity play a role in the individual turnout decision by encouraging political dis- cussion. Without a campaign, who would vote? Perhaps a few highly informed and highly motivated citizens; but most of us are, to one degree or another, drawn in by the discussion and activity that surrounds an election, whether scattered and highly localized as in a contested city council election or widespread and virtually inescapable as in a presiden- tial race. The theoretical approach of this book fits well with what is known about how canvassing, both face to face and via phone calls and other communication media, affect the probability that mobilized citizens will actually participate. Most strikingly, voters are more likely to vote when told that turnout in the previous election was high rather than low (Gerber and Rogers 2009 ), and when told about their neighbors’ prior voting history (Gerber, Green, and Larimer 2008 ). The proposed conditional decision function is itself nonlinear (and likely to be so for most individuals), and evidence confirms that citizens with a moderate initial propensity to vote are most affected by mobilization (Niven 2004 ; Arceneaux and Nickerson 2009 ). If mobilization also affects the (subjective) framing of the upcom- ing election among potential voters, we would expect the quality of the Rolfe, M. (2012). Voter turnout : A social theory of political participation. Cambridge University Press.

Created from apus on 2022-03-08 18:36:00.

Voter Turnout 16 contact to affect the priming potential of the contact. Again, evidence is in line with this prediction, as volunteer college students making phone calls (Nickerson 2008 ) and local neighborhood residents knocking on doors (Sinclair, Michelson, and Bedolla 2007 ) are more effective at increasing turnout than professionals or volunteers from outside the neighborhood.

Mobilization, in other words, is not a mechanical activity, but depends on the social interaction between those doing the mobilizing and poten- tial voters they wish to mobilize. It is likely that strongly embedded citi- zens who mobilize endogenously, like precinct captains in the Chicago machine or organizational leaders, or, increasingly, ministers in black churches, will be more effective at stimulating both the situational frames and high levels of discussion that promote increased turnout.

Social Context and the Sociological Tradition This book complements existing research documenting the influence of social context on political behavior, including older work from the Columbia School (Lazarsfeld, Berelson, and Gaudet 1948; Berelson, Lazarsfeld, and McPhee 1954 ) and more recent work by other research- ers inspired by the Columbia School tradition (Huckfeldt 1979 , 1980 ; Gimpel, Lay, and Schuknecht 2003 ; McClurg 2004 ; Zuckerman 2005 ; Campbell 2006 ; Klofstad 2007 ). Along with other scholars working inside the social context tradition, I assume that the decisions are not taken by isolated individuals, but individuals embedded in particular social contexts. The impact of social context cannot be reduced to indi- vidual experiences of “peer pressure” or conscious attempts to influence other people. Social context is often working in the background, through the “slow drip” of everyday life (Baybeck and McClurg 2005 ) that shapes social schemas, expectations of others, and understandings of the political world. Traditionally, the impact of social context has been treated as inter - changeable with a simple, linear flow of influence from one person to another. Thus, social context has been conceptualized (and formally expressed) as working at the individual level through alters, or friends, who (presumably consciously) shape the attitudes and behavior of the person whose ego-centered network is under consideration (called “ego”).

The unavoidable stumbling block in this (somewhat reductionist) model of contextual effects on the individual is that people can change either their behavior or their friends (Heider 1944 ). Thus, observational studies Rolfe, M. (2012). Voter turnout : A social theory of political participation. Cambridge University Press.

Created from apus on 2022-03-08 18:36:00.

Voting Together 17 of social influence on political behavior are (perhaps unfairly 5) subject to the same basic criticism: How do we know whether the observed similar - ity between friends is a result of selection or influence (Achen and Shively 1995 )? However, the mathematical approach to social interaction taken in this book is more faithful to the broader understanding of social context that is present in the literature from Lazarsfeld through to the current day. Conditional choice provides a framework for thinking systemati- cally about the impact of social context without reducing it to direct social influence via dyadic relationships. In particular, social structure and social location – not dyadic interaction – step to the forefront as explanations in any situation involving conditional decision making.

Thus, empirical tests of the social theory of turnout, and research within the conditional choice approach more generally, are not subject to the criticisms commonly made of research in the social context tradition. Whereas the underlying notion of conditional cooperation among friends remains indispensable to understanding turnout, it is the way in which these friendships fit into the larger structure of social interaction that offers the possibility of reinterpreting the effects of education and other empirical correlates of turnout. The key insight driving the social theory of voter turnout is that social networks are more than an accumu - lation of the characteristics of the individuals in them (Coleman, Katz, and Menzel 195 7; Blau 197 7; Cook and Whitmeyer 199 2). The actual shape of the network itself can have important implications for the spread of a behavior in a society. In terms of turnout, if Jane the college graduate has more friends who vote than Joe the high school graduate, this may not indicate that Jane’s friends are better able to understand politics, or have developed a stronger sense of civic duty, or possess more of some other characteristic or quality than Joe’s friends. Rather, the difference in turn - out rates may arise simply from the average size and shape of the larger set of personal networks in which these friendship groups are embedded. The importance of this insight cannot be overstated: Social network differences are not reducible to differences in the individuals in those networks. This sociological perspective on the link between education and turn- out bears a strong resemblance to the argument made by Nie, Junn, and Stehlik-Barry ( 1996 ). They resolve the “education puzzle” by claiming 5 McClurg ( 2003 ) correctly points out the flip side of the issue: that any model of political participation that omits informal social interaction is undeniably misspecified. Rolfe, M. (2012). Voter turnout : A social theory of political participation. Cambridge University Press.

Created from apus on 2022-03-08 18:36:00.

Voter Turnout 18 that relative educational attainment, not years of education, is the impor - tant predictor of turnout probability. Relative education attainment is assumed, although never conclusively shown, to be a proxy for social net- work status. Their results do not contradict my claims, and I also argue that education is at best a proxy for something important about social networks. However, I go beyond the claims in that book in two ways.

First, I provide a well-documented and empirically supported model of decision making that includes a mechanism for exactly how social net- work structure impacts individual decisions. Furthermore, I examine in depth their claim that social status is an indicator of access to political power and find that this is only true some of the time. I analyze a large community using both interviews with political candidates and individ- ual and aggregate data to show that political history and institutionally mediated social ties are more important than social status in determining who has social access to political power.

overview of remaining chapters The remainder of the book will unfold in three distinct steps. First, the conditional choice framework and associated research strategy are intro- duced. The second step develops a decision model of conditional coop- eration in situations similar to voting turnout. The third section expands the decision model into a middle-range theory (Hedström 2005 ) of turn- out and tests several implications of the theory. Chapter 2 introduces conditional choice, a formal research frame- work for developing and testing positive propositions about the impact of social context on social outcomes. I present evidence that our decisions and actions are often conditional on the actions of others, even when we are not consciously influenced or even aware of those other actions.

The chapter then moves to a mathematical model of conditional choice, in which any individual action can be expressed as a function of others’ actions, more precisely as a weighted function of several basic decision rules. From these basic decisions rules we can derive a limited number of formal mechanisms that the conditional choice framework can use account for social outcomes, including the “basic dynamics” of a decision situation that make a behavior easy or difficult to diffuse, and aspects of social network structures that further shape and direct this diffusion pro- cess. Although this may sound very mechanistic, subjective perceptions are a keystone of the conditional choice framework: People’s responses Rolfe, M. (2012). Voter turnout : A social theory of political participation. Cambridge University Press.

Created from apus on 2022-03-08 18:36:00.

Voting Together 19 to others’ behavior depend on what sort of situation they think they are in, or the social meaning of the situation. Chapter 3 uses a combination of existing evidence, newspaper articles and editorials, and survey data to establish that the shared social mean- ing of voting is consensually positive and widely shared throughout the citizenry. In other words, there is no socially acceptable justification for nonvoting, even among subgroups in the population. Substantively, vot- ing is understood as the core act defining democratic citizenship in a community of equals. Thus, voting is a social dilemma in which voting is uniformly understood as doing one’s fair share for the community, and American citizens are all equal in terms of citizenship status. Chapter 4 identifies and reviews experimental scenarios with a similar social meaning to voter turnout: low-cost social dilemmas among equals.

Evidence on the conditional (and unconditional) behavior of subjects in experimental social dilemmas is used to estimate a descriptively accurate model of the distribution of conditional decision rules in situations sim- ilar to voting. Chapter 5 describes the general and more specific dynamics of the conditional cooperation model. Turnout is produced when the actions of “first movers,” who are willing to cooperate unconditionally, encourage conditional cooperation among others. The expected level of cooperation (or turnout) responds to changes in two key parameters in the model:

(1) the distribution of decision rules (based on empirical estimates from Chapter 4) and (2) the average size and density of personal networks within an individual’s social circle (estimated using survey data on per - sonal social networks.) The conditional cooperation model successfully produces a wide range of turnout levels within the parameter ranges set by external sources of data, and (as would be hoped) fails to do so out- side those parameters. Chapter 6 incorporates additional empirically based assumptions into the basic decision-making model to produce a complete, midrange social theory of voter turnout. Political mobilization, social location, institu- tional costs, and electoral salience are incorporated into the theory as working through mechanisms associated with either decision rule distri- butions or social structure. A relatively simple demonstration suggests that previously observed aggregate level variation in turnout is in line with predictions of the fully developed social theory. Chapter 7 compares the predictions of the social theory to those of the civic resources view in high-salience presidential elections. Drawing on a Rolfe, M. (2012). Voter turnout : A social theory of political participation. Cambridge University Press.

Created from apus on 2022-03-08 18:36:00.

Voter Turnout 20 relatively unique network component in the 1985 General Social Survey (GSS), I develop a suitable (although rather blunt) proxy for individual location in a larger social network structure. Although the proxy is not perfect, the test does provide support for the claim that social network structure and location provide a better explanation of individual varia- tion in reported voter turnout than a civic resources framework. Chapter 8 provides a more powerful crucial test of the two approaches, comparing the explanatory power of the two perspectives in low-salience elections. A combination of interviews, voter roll data, and geospatial analysis is used to describe the political and social geography of turnout in a large southern county across several elections. Social proximity to candidates is the single greatest driver of turnout in low-salience elec- tions, whereas socioeconomic status has no impact beyond increasing social access to political candidates. Chapter 9 concludes the book. I review the major findings, discuss the practical implications of the project for candidates and political activ- ists, and offer suggestions for future research. I offer thoughts on how to extend the analysis to the individual level, and the applicability of the conditional choice framework to a range of other political behaviors.

Rolfe, M. (2012). Voter turnout : A social theory of political participation. Cambridge University Press.

Created from apus on 2022-03-08 18:36:00.