Você está na página 1de 33

From Classroom to Voting Booth: The Eect of High School Civic Education on Turnout

Jennifer Bachner This version: September 12, 2010 Original version: May 23, 2010

Abstract A healthy democracy requires a citizenry that participates in political life. While interventions such as removing barriers to registration and mobilizing partisans have received frequent scholarly attention, formal civic education has been largely ignored. Using longitudinal data and a matching analysis, this paper shows that students who complete a year of coursework in American Government/Civics are 3-6 percentage points more likely to vote in an election following high school than those without exposure to civic education. Further, this eect is magnied among students whose parents are not highly politicized. Among students who report not discussing politics with their parents, additional coursework is associated with a 7-11 percentage point increase in the probability of voting. This result suggests that civic education compensates for a relative lack of political socialization at home, and thereby enhances participatory equality.

Many thanks to Claudine Gay, Matthew Platt, Steve Ansolabehere, Sunshine Hillygus, Elena Llaudet and the participants in Harvard Universitys American Politics Research Workshop for helpful comments on earlier versions of this paper.

Introduction

In a March 2010 op-ed piece for the Wall Street Journal, renowned historian and policy analyst Diane Ravitch expressed frustration with current education reform, noting, Most signicantly, we are not producing a generation of students who are more knowledgeable, and better prepared for the responsibilities of citizenship. The conception of citizenship as a learned behavior has existed since ancient times, and it persists as a primary rationale for the public school system in the United States. Yet despite a consensus on the importance of educating for democracy among scholars, policy makers, practitioners and the public, there is a dearth of empirical research documenting the behavioral returns to coursework in American Government and Civics. What political or civic actions, if any, are students prompted to take upon leaving their classrooms? This paper tackles this question by examining the eects of high school civic education on turnout.

Motivating Citizens to Vote

From 1964 to 2008, young citizens voted consistently at lower rates than those over the age of 24. On average, the turnout gap in the presidential elections that occurred during this time period is a stunning 21 percentage points (Current Population Survey, 2010). To account for this disparity, a number of articles argue that a lack of adult experiences, such as marriage, home ownership and employment, depresses youth participation (Squire, Wolnger and Glass, 1987; Strate et al., 1989). More recent research, however, nds the life cycle hypothesis to be of limited value (Highton and Wolnger, 2001; Stoker and Jennings, 1995), and suggests that a lack of political knowledge (Delli Carpini and Keeter, 1996), identity (Kiousis and McDevitt, 2008) and ecacy (Pasek et al., 2008) provide more compelling explanations. By aiding the accumulation of these and other politically relevant resources, civic education can potentially shrink the youth turnout gap. 2

Setting aside the disparity in turnout between age groups, the absolute numbers suggest that there is signicant room for improvement in turnout among all U.S. citizens. To address this broader issue, political scientists have sought to identify interventions that increase turnout. One group of scholars has focused on federal and state government policies, which typically lower the barriers to registration and thereby reduce the costs associated with voting. A second group has devoted attention to get out the vote (GOTV) eorts employed by campaigns and non-partisan organizations. The theory motivating GOTV practices is that potential voters are mobilized by social pressure to participate (Rosenstone and Hansen, 1993; Verba, Schlozman and Brady, 1995). A summary of the main ndings from these types of studies is presented in Table 1.1 The evidence indicates that both public policies and private eorts can boost turnout, with an upper bound of approximately 10 percentage points. The elimination of a registration closing date emerges as the most inuential registration reform law and door-to-door canvassing appears to be the most eective mobilization technique. While this table displays the eects associated with the most commonly studied interventions, numerous others have received scholarly attention, including purging voter registration rolls (Wolnger and Rosenstone, 1980), implementing voter identication laws (Alvarez, Bailey and Katz, 2008) and sending GOTV text messages (Dale and Strauss, 2009). But although there is an overwhelmingly large literature on turnout, studies that treat civic education as a form of intervention are exceedingly rare.

For the government interventions, the eects are calculated by comparing within state (for longitudinal data) and between state (for cross-sectional data) turnout in the presence and absence of voter registration laws. States with Election Day registration allow citizens to register at the polls (or do not require registration). Motor voter laws refer to the requirement that a voter registration form be included with a drivers license application. States with extended registration hours require registration oces to remain open after normal business hours, on weekends or both. The articles cited in the non-governmental interventions section examine non-partisan GOTV practices, though partisan eorts have likewise been found to have positive eects on turnout.

% Point Increase in Turnout Government Interventions Election Day Registration 3.0 - 8.7

Selected Citations

Wolnger and Rosenstone (1980); Teixeira (1992); Fenster (1994); Mitchell and Wlezien (1995); Rhine (1995); Brians and Grofman (2001); Knack (2001) Wolnger and Rosenstone (1980); Teixeira (1992); Mitchell and Wlezien (1995) Knack (1995); Rhine (1995); Franklin and Grier (1997)

Extended Registration Hours

0.7 - 5.5

Motor Voter Program Non-Governmental Interventions Door-to-door canvassing

1.7 - 3.8

7.1 - 10.2

Gerber and Green (2000); Green, Gerber and Nickerson (2003); Gerber, Green and Shachar (2003) Gerber and Green (2000); Gerber, Green and Shachar (2003) Gerber and Green (2000); Gerber and Green (2001); Nickerson (2006)

Direct mailings

0.6 - 2.5

Phone calls

0.0 - 3.8

Table 1: The Inuence of Interventions on Turnout.

Civic Education as Public Policy

Since the late 1980s, the number of states that require civic education to graduate has increased monotonically. Whereas 21 states required a course in American Government/Civics for the graduating class of 1988, 33 states imposed this requirement for the graduating class of 2013.2 The percentage of U.S. graduates who are exposed to classroom-based civic education, however, far exceeds the percentage of students who are required by state laws to take a
I collected these data using newspaper articles, state statutes, state administrative codes and the LBJ School Survey of State Civic Education (Tolo, 1999).
2

course on this subject. Every state allows school districts or the like to supplement the imposed minimum graduation requirements, and many districts take the opportunity to do so.3 Actual course enrollment data provide a more complete picture of the prevalence of civic education in the U.S.
Percentage of U.S. Graduates 2005 2000 1998 1994 1990 1987 Government/Civics (.5 credits) American History (1 credit) World History (1 credit) Economics (.5 credits) 79.2 94.1 76.5 46.6 77.3 90.0 62.1 49.2 81.2 90.5 60.2 45.8 78.0 94.8 59.4 44.5 77.4 91.3 52.0 47.3 71.9 89.9 42.0 33.3

1982 62.2 81.4 36.3 27.4

Table 2: Credits Earned by High School Graduates in Social Studies: 1982-2005. Data are taken from published tables in Roey et al. (2007) and Shettle et al. (2007). Students who completed the minimum indicated credits are included in each estimate.

Table 2 presents the percentage of high school graduates who completed the indicated minimum number of credits in four social studies subjects: Government/Civics, American History, World History and Economics.4 Although American History has continually ranked highest in enrollment, an increasing percentage of students complete coursework in Government/Civics, as enrollment jumped from 62.2% in 1982 to 79.2% in 2005. Figure 1 illustrates the distribution of American Government/Civics course-taking across the U.S. for the graduating class of 1992. The map displays the median number of American Government/Civics credits earned by graduates in each state. Although a substantial portion of graduates were exposed to this type of coursework, there is signicant regional variation. Students in parts of the Northeast and Midwest are most likely to have taken zero courses
Rhode Island, for example, does not require credit in civic education to graduate, yet Ponaganset High School requires students to complete a course in American government and civics beginning with the class of 2007 (Boas, December 17, 2002). In 1993, when Maryland passed a law requiring an American government course beginning with class of 1997, Anne Arundel and Montgomery already had this requirement in place (Le, August 26, 1993). 4 One credit is equal to one Carnegie Unit. A Carnegie Unit is a standard measure of in-class time. One Carnegie Unit is equivalent to 120 in-class hours; most year-long courses are worth 1 unit. If a year-long course were to meet for a total of 150 hours, for example, it would be assigned 1.25 Carnegie Units.
3

Legend: 0 Credits .5 Credits 1 Credit

Figure 1: Median Number of Credits in American Government/Civics. Data are from the NELS:1988. States are coded according to the median number of credits in American Government/Civics earned by high school graduates. in formal civic education.

Previous Research

Taken together, state graduation requirements and course enrollment numbers are evidence of a widespread mandate that high school students receive classroom-based civic education. Why is it, then, that the returns to civic education remain a fairly untilled area of research? The lack of ndings by several well-conducted studies published in the 1960s provides a compelling explanation (Litt, 1963; Langton and Jennings, 1968; Ehman, 1969). Generally considered the seminal work on the topic, Langton and Jennings (1968) use the 1965 wave of the Youth-Parent Socialization Study (YPSS) to investigate whether self-reported amount of civic education is related to political knowledge, interest and attitudes. The null results across all of the dependent variables lead the authors to conclude, Our ndings certainly do not support the thinking of those who look to the civics curriculum in American high 6

schools as even a minor source of political socialization (865). This conclusion persisted for the next three decades. Recently, however, scholars have begun to revisit the relationship between civics coursework and political knowledge. Using the 1988 National Assessment of Educational Progress (NAEP), Niemi and Junn (1998) investigate the relationship between the amount and recency of American Government/Civics coursework and performance on the NAEP Civics Assessment. The authors nd that, in 9 out of the 10 areas of civics covered on the exam, students who indicated that they had completed an American Government/Civics course scored higher than those who had never studied subject and those who had never enrolled in a formal course on the subject.5 The authors conclude, . . . The evidence so far suggests that civics courses do have an eect on student knowledge, an eect that is wide-ranging in terms of content and . . . also appears to raise students capacity for reasoning and exposition about civic matters (70). Others who have tested the coursework-knowledge connection have likewise uncovered positive ndings (Anderson et al., 1990; Gimpel, Lay and Schuknecht, 2003). Noticeably absent among the dependent variables in existing research on civic education are measures of political behavior. In their analysis, Langton and Jennings (1968) construct a variable labeled participative orientation. This measure is based on responses to an open-ended question about what makes a person a good citizen, and it is the only dependent variable in the study that speaks to behavior.6 One concern with this measure is that it lumps together activities such as taking an interest in public aairs, voting and trying to improve the country. It may be that civic education is related to these activities in dierent
When civic knowledge is operationalized as a composite variable (0-100), students who complete an American Government/Civics course in twelfth grade are expected to score two points higher than those who take the course in an earlier grade, and four points higher than those who do not take the course. 6 The survey question reads, People have dierent ideas about what being a good citizen means. Were interested in what you think. Tell me how you would describe a good citizen in this country That is, what things about a person are most important in showing that he is a good citizen (Jennings et al., 1997).
5

ways. Civic education may motivate students to vote but not cause them to stay informed about political developments. And although a question about how citizens in the abstract should act is certainly interesting, it does not elicit information about the political activities in which students actually engage. I address this gap in the literature by treating turnout as the outcome variable of interest.7 A second potential problem with existing research is error in the measurement of civic education. The operational denition of civic education is often derived from a survey question that asks students to report how much American Government/Civics coursework they have completed, with answer choices such as: less than a year, a year, two years or more than two years. Students may inadvertently misreport this number either because they cannot remember correctly or because they are unsure about which courses count as American Government/Civics courses.8 In an analysis of the amount of misreporting of U.S. history course-taking present in the 1994 NAEP, Niemi and Smith (2003) uncover deeply troubling results. The authors nd, for example, that 31.3% of high school students reported enrolling in a U.S. history course during their senior year, whereas the 1994 High School Transcript Study estimates actual enrollment to be 12.2% (18). It seems reasonable to expect that a similarly disturbing amount of misreporting takes place with respect to American Government/Civics classes. Measurement error may have contributed to the null ndings in previous research. If the error is random, meaning students are just as likely to under-report as they are to over-report, then the eect of civic education will be biased toward zero. If students tend
Existing research that investigates the eects of civic education on voting and other forms of participation focus on particular curriculua, such as Kids Voting USA and Project Citizen (Simon and Merrill, 1998; Vontz, Metcalf and Patrick, 2000; Meirick and Wackman, 2004), as opposed to the cumulative amount of classroombased civic education. 8 Many students around the country, for example, take a course titled American Government and Economics. Other ambiguous courses include American Foreign Policy and Contemporary World Aairs. In answering the survey question, some students may include this course in their total count of civics courses while others may not.
7

to systematically over-report, as Niemi and Smiths (2003) ndings suggest, the treatment eect will likely be understated. To ensure accuracy and consistency in my measure of civic education, I rely on student transcript data. These data allow me to control the types of courses that are included in each students count and to tally the amount of actual exposure to this coursework.9 A nal reason to reconsider existing notions about the returns to civic education is the possibility of interaction eects. It may be that the eect of American Government/Civics courses depends upon other agents of political socialization, such as parents, peer groups and the media. Political information can be acquired from each of these sources, and it may be that exposure to one aects the inuence of another. This article examines the eects of civic education from this perspective by testing whether amount of civics coursework interacts with level of parental politicization in its eect on turnout.

5
5.1

From Classroom to Voting Booth


A Model of Participation Applied to Civic Education

The Civic Voluntarism Model (Verba, Schlozman and Brady, 1995) provides a theoretical framework for charting a path from American Government/Civics coursework to turnout. The model states that participation is driven by three factors: psychological engagement, resources and recruitment. Psychological engagement and resources are the two most important of the three factors; whereas participation often occurs in the absence of recruitment, it rarely takes place when the potential participant lacks either resources or psychological engagement (270). High school civic education courses can strengthen students psychological engagement
The transcript data sets list the number of Carnegie Units, a standard measure of in-class time, assigned to each course.
9

with politics by increasing their political knowledge. Political knowledge promotes interest, trust and ecacy in the political system by enabling citizens to judge politicians using more than a few misleading sound bites and images (Verba, Schlozman and Brady, 1995; Delli Carpini and Keeter, 1996; Zaller, 1992). When the media cover a mistake by a politician, knowledgeable citizens will integrate this new piece of information into their existing perception of the featured persons character and policy agenda. An uninformed citizen, in contrast, is likely to form a much more negative opinion of the politician; the uninformed citizen possesses no information to counterbalance the negative news story. As a result, citizens who lack political knowledge may feel mistrustful of the political system. Popkin and Dimock (1999) illustrate how background and contextual knowledge can inuence the perception of a political event: To persons who understand the institutions of politics, a long set of exchanges between, say, Bill Clinton and Robert Dole can be as clear as a sustained volley in tennis; to persons without any knowledge of institutions, their exchange is hard to follow and becomes indistinguishable from a food ght or mud-slinging (134). In short, less knowledgeable citizens develop more negative views on government and elected ocials. And these negative views, whether they take the form of cynicism, mistrust or alienation, have a demobilizing eect on those who posses them. Civic education also has the potential to increase a students store of civic skills. Civic skills are the communications and organizational abilities that allow citizens to use time and money eectively in political life (Verba, Schlozman and Brady, 1995, 304). These skills make participation easier and thus more likely. Even the simple act of voting requires citizens to, at the very least, determine how to register to vote and locate their polling place. Further, the probability that citizens will vote is increased if they are able to discuss politics with friends and persuade others of their political beliefs. 10

A number of studies investigate whether civic education classes in general, and civic education curricula in particular, provide students with politically relevant skills. Gimpel, Lay and Schuknecht (2003) nd that the students who take more than a year of civics coursework, as compared to students with no civics coursework, discuss politics with friends and family ve percent more per week (149). This eect is perhaps attributable to curricula that incorporate political discussion and other interactive forms of learning into lesson plans. Kids Voting USA, for example, is a K-12 curriculum developed in 1987 that encourages students to engage in activities that teach civic skills. These activities include mock elections, debates and role-playing. Evaluations of this curriculum reveal the strong, positive eects the lessons have on students deliberative skills. McDevitt and Kiousis (2004), in a comparison of student survey responses before and after exposure to the curriculum, nd that the students gained both competence and condence in their ability to discuss politics. The authors write, Kids Voting. . . promoted conversational skills and related dispositions such as the willingness to disagree, willingness to listen to opponents, testing out opinions in conversation, and challenging the views of parents (10). A second curriculum that has been found to have signicant eects on students development of civic skills is We the People: Project Citizen. Today, this curriculum is used by teachers in all 50 states. The curriculum asks students to research a policy issue relevant to their community, develop solutions and communicate these solutions to civic and political leaders. An evaluation of the program by Vontz, Metcalf and Patrick (2000) reveals that participation in the program is associated with a signicant increase in a students condence in her ability to identify important public policies, describe public policies to others and formulate opinions on public policies.10
Student Voices is another popular curriculum that has been shown to increase youth civic engagement. Participation in two semesters of the program is associated with an increase in internal ecacy, candidate knowledge and political attentiveness (Pasek et al., 2008).
10

11

5.2

Taking Parents into Account

Parents exert a nontrivial inuence on adolescents political development through explicit lessons and by serving as role models (Hess and Torney, 1967). Parent to child transmission, however, is not inevitable. Although parents have an inherent potential for successful transmission (Tedin, 1974, 1592), this potential is best realized under certain conditions. Political beliefs are transmitted most often from parents to children when perceptual accuracy and issue salience are high. Children, in other words, are likely to adopt their parents attitudes when they develop an accurate understanding of these views, and when these attitudes are important to the parents. This suggests that successful transmission depends upon the level of parental politicization. Highly politicized parents will possess stronger opinions on the importance of voting, their partisan orientation and policy issues than parents who are less interested in or less knowledgeable about politics. As a result, highly-politicized parents are more likely to emit strong and clear cues regarding their political attitudes. But what is the mechanism that drives the transmission process? Political discussion is a key answer to this question (Valentino and Sears, 1998; Verba, Schlozman and Brady, 1995; Jennings, Stoker and Bowers, 2009). Family discussions about politics provide opportunities for children to absorb cues given by their parents. As such, I use the amount of political discussion within the family as an indicator of parental political socialization. I expect that children whose parents are not highly politicized will derive a greater benet from civic education than their peers. This is because low levels of parent politicization should leave the child bereft or relatively open to inuence from other socializing agents (Jennings, Stoker and Bowers, 2009, 13). Parents who are interested in politics, and who share this interest with their children, are likely to crowd out the inuence of schools. Students whose parents are not highly politicized, in contrast, have more to gain from their civics coursework.

12

6
6.1

Data and Methods


National Education Longitudinal Studies

The above discussion yields two hypotheses: (1) additional coursework in civic education will increase turnout and (2) this increase will be largest among students whose parents are not highly politicized. To test these hypotheses, I employ the National Education Longitudinal Studies of 1988 (NELS:1988) and 2002 (NELS:2002). These studies were conducted in multiple waves and contain student surveys, student transcripts and parent surveys. The NELS:1988 was conducted from 1988 to 2000. In 1988, a baseline survey was administered to a nationally representative sample of eighth-graders from 1,052 public and private schools. Subsequent surveys were administered in the spring of 1990, 1992, 1994 and 2000 to a random subset of the original sample. The completion rate for the rst wave is 93%, and the completion rate for all ve waves is 79% (Curtin et al., 2002). The NELS:2002 was rst administered to a nationally-representative sample of high school sophomores in 2002 from 752 public and private schools. The response rate for the rst wave is 87.3%. This sample of students was interviewed again (along with a freshening sample) in 2004 and 2006. The response rates for these waves are 90.8% and 89.1%, respectively (Ingels et al., 2007). In both the NELS:1988 and NELS:2002, transcripts were collected during students senior year of high school. The transcript data sets list every course in which each student enrolled, the year and grade-level during which each course was taken, the length of each course (yearlong, semester or quarter), the grade earned in each course, and the number of Carnegie Units assigned to each course.

13

6.2

Voting Variables

The dependent variables measure students voting behavior in the years following high school. The 1994 and 2000 waves of the NELS:1988 contain questions that ask students whether they voted in the 1992 presidential election, the 1993-94 state/local elections, the 1996 presidential election and any election between 1998 and 2000. In the 2006 wave of the NELS:2002, students were asked whether they voted in the 2004 presidential election and any state/local election between 2004 and 2006. Figure 2 displays the percentage of students who indicated that they voted in each of these six elections.

% of Students Who Voted

0.8

1.0

0.6

0.57 0.51 0.46 0.42 0.31 0.48

0.0
1992

0.2

0.4

199394

1996

199800

2004

200406

Elections

Figure 2: Summary of Voting Variables. This gure displays (1) the percentage of 1992 graduates in NELS:1988 that voted in the 1992, 1993-94, 1996 and 1998-00 elections and (2) the percentage of 2004 graduates in the NELS:2002 that voted in the 2004 and 2004-06 elections.

6.3

Civic Education Variables

The independent variable of interest are the number of credits earned in four categories of civic education.11 There are a variety of courses related to American government, politics
Approximately 70% of students in the NELS:1988 and 73% of students in the NELS:2002 completed some type civic education coursework.
11

14

and citizenship that appear on students transcripts. Guided by previous research and the descriptions of all 44 civic education courses, I grouped these courses into four categories: American Government/Civics, American Problems, American Government and Economics and Other Government. These categories parallel those used by Langton and Jennings (1968), with the exception of the American Government and Economics category, which I have added. Figure 3 displays the distribution of these categories for each data set.
(a) NELS:1988
1.0 1.0

(b) NELS:2002

0.82
0.8 0.8

0.86

% of Students

% of Students 0.12 0.05

0.6

0.4

0.2

0.2

0.4

0.6

0.1 0.02

0.11

0.07

0.0

American Gov/Civics

American Problems

American Gov/Econ

0.0

Other Government

American Gov/Civics

American Problems

American Gov/Econ

Other Government

Civic Education Course Categories

Civic Education Course Categories

Figure 3: Distribution of Civic Education Course Categories. This gure shows the percentage of students who completed coursework in each category among all students who completed civic education coursework. The percentages in each graph sum to more than 100 because some students completed coursework in multiple categories. The American Government/Civics category includes traditional courses, such as American Government, Basic Government and Civics. Of the students who completed civic education coursework, an overwhelming majority (82% in the NELS:1988 and 86% in the NELS:2002) completed this type of course. This is the type of course that is most commonly oered by schools to fulll a state or school district imposed civic education requirement. Although the transcript data sets do not contain information about which courses were required by the school districts for graduation, it is reasonable to expect that a majority of 15

students who complete a traditional American Government/Civics course do so either because it is required or because it is the norm at their school. Further, the inclusion of the other course categories in each model controls for student interest in taking civic education courses beyond a traditional American Government/Civics course. The American Problems category includes courses with titles such as Contemporary American Political Issues. The popularity of this course has declined dramatically since its heyday in the 1960s, and only a small percentage of students in the sample at hand completed this type of course.12 The third category includes courses in which economic issues are woven into the study of government institutions.13 This type of course has grown in popularity over the past few decades. Approximately 12% of students in the NELS:1988 and 11% of students in the NELS:2002 earned credit in this type of course. Lastly, the Other Government category includes courses in specialized areas of American government, comparative politics and international relations. Courses in this category appear to be electives rather than required classes. As a result, it is possible that the students who enrolled in these classes are particularly interested in politics, and it is this interest, rather than the course itself, that drives their future political participation.

6.4

Control Variables

Each model includes four categories of control variables: measures of civic engagement, school and state characteristics, home resources and student demographics. The variables in the civic engagement category are intended to capture a students prior propensity to participate
While 18.5% of students enrolled in this type of course in the 1960-61 academic year, this number had dropped to 8.9% in the 1972-1973 academic year, and continued to decline in the following years (Niemi and Smith, 2001). In Langton and Jenningss (1968) sample, 67% completed an American Government course, 37% completed an American Problems course and 10% completed an Other course (of those students who indicated that they had taken a civic education course) (855). 13 These types of courses are dierent from those in which students spend one semester focused on Government and the second on Economics.
12

16

in civic and political life. These variables include participation in extra-curricular activities, participation in student government and participation in school service clubs.14 In addition, the two data sets contain a few non-overlapping variables that may capture student interest in politics. The regressions that use the NELS:1988 data include a variable that controls for eighth-grade interest in social studies.15 The regressions that use the NELS:2002 data include variables that ask twelfth-grade students how important it is to be (1) patriotic and (2) an active and informed citizen.16 Included in the home resources category is a measure of parental political socialization, which is operationalized as the frequency with which students report discussing current events with their parents. This variable is a three-point scale that ranges from never to often.17 Parents highest level of education, whether the family receives a daily newspaper, and socioeconomic status are also included to control for relevant resources.18 It is reasonable to expect that schools in which students take more civics coursework have other attributes that foster political participation. I control for academic quality by including the percentage of students who attended a four year college in each model. I also control for whether the school is public or private, the urbanicity of the school (urban, rural, suburban) and whether the state imposes a civic education course requirement. Figure A-2 summarizes the state requirement variable. The nal set of independent variables control for student background characteristics that
14 The school activities measure is a count of the number of hours per week spent on extra-curricular activities in tenth and twelfth grade. The student government variable measures participation in eighth, tenth and twelfth grade, and the service club variable measures participation in tenth and twelfth grade. 15 The eighth-grade survey asks students whether they look forward to social studies. This variable is a four-point scale that ranges from strongly disagree to strongly agree. 16 Both of these variables are three-point scales that range from not important to very important. 17 The parent survey also contains a question about the amount of discussion about current events that takes place at home. This variable, however is only weakly correlated with the student variable about political discussion. Because of social desirability, the parent variable seems less reliable than the student variable, although it is included in each model as a control. 18 Socioeconomic status is a composite variable that includes fathers education, mothers education, fathers occupation, mothers occupation and family income.

17

have been shown to inuence turnout. I control for gender, race and parents native language. A binary variable indicating whether the student received a high school diploma and a binary variable indicating whether the student ever transferred schools are also included.

7
7.1

Results
Direct Eect of Civic Education

For each election, I estimate a logit model with clustered standard errors (clustering on school). Included in each model is an interaction between American Government/Civics coursework and the measure of parental politicization.19 The raw output from these six regressions is presented in Table A-1.

200406

90% CI 95% CI

2004

199800 Elections

1996

199394

1992

0.3

0.2

0.1

0.0

0.1

0.2

0.3

Change in Probability of Voting (0 to 1 Course)

Figure 4: Civic Education and Voting. This graph presents the rst dierence estimates associated with taking a year-long course in American Government/Civics versus taking zero courses.
19

The American Government/Civics variable and the measure of parental politicization are centered such that the mean of each of these variables is 0. Alternative specications show that there is no interactive eect between parental politicization and the other categories of civic education.

18

To test the direct eect of civic education, I calculate the change in probability of voting in each election associated with an increase in civics coursework. The only course category that consistently exerts a positive, statistically signicant eect on voting behavior is American Government/Civics. Figure 4 displays the expected changes in probability of voting in each election associated with a one-year increase in American Government/Civics coursework, as well as the 90% and 95% condence intervals associated with these changes.20 Pr(Voting|0 Courses) 1992 1993-94 1996 1998-00 2004 2004-06 .539 .355 .627 .446 .574 .546 Pr(Voting|0 Courses) 1992 1993-94 1996 1998-00 2004 2004-06 .539 .355 .627 .446 .574 .546 Pr(Voting|0.5 Courses) .562 .380 .644 .474 .592 .558 Pr(Voting|1 Course) .584 .407 .660 .501 .609 .571 .023* .026* .016 .027* .017 .013 .045* .052* .032 .055* .034 .025 SE of .011 .010 .011 .010 .009 .009 SE of .021 .019 .021 .020 .019 .018

Table 3: Eects of American Government/Civics Coursework on Voting. The estimates for 1992, 1993-94, 1996 and 1998-00 are based on the NELS:1988 data. The estimates for 2004 and 2004-06 are based on the NELS:2002 data. * p < .05. The precise rst dierence estimates and standard errors associated with an increase in American Government/Civics coursework from zero to one-half and zero to one are displayed in Table 3. For the 1992 presidential election, students who completed a one-year course were 4.5 percentage points more likely to vote than those who took zero courses. The largest eect appears in the 1998-2000 election period, when students who took one year of American Government/Civics were 5.5 percentage points more likely to vote. The smallest
20

Technically the increase is one Carnegie Unit.

19

eect appears in the 2004-2006 election period, where this increase in coursework is associated with a 2.5 percentage point increase in probability of voting.

7.2

Interactive Eect of Civic Education

To test whether parental politicization mediates the relationship between civic education and turnout, I calculate the change in probability of voting associated with an increase in American Government/Civics coursework while holding the political discussion variable xed at never. For all six elections, the estimates are positive, substantively large and statistically signicant.
(a) Interactive Eect of American Gov/Civics (b) Interactive Eect in the 1992 Election
0.4 200406
q

90% CI 95% CI

First Difference Estimates 95% Confidence Band Zero


0.2 Probablity of Voting in 1992

2004

199800 Elections

No Discussion

1996

0.0

Frequent Discussion 0.2 0.4

199394

1992

0.3

0.2

0.1

0.0

0.1

0.2

0.3

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

Change in Probability of Voting (0 to 1 Course)

American Governmet/Civics Coursework (in Carnegie Units)

Figure 5: Civic Education, Parental Politicization and Voting. Graph (a) presents the
rst dierence estimates associated with taking a year-long course in American Government/Civics versus taking zero courses when parental politicization is held xed at never. Graph (b) displays the rst dierence estimates for the 1992 presidential election that are associated with a range of changes in American Government/Civics coursework.

The predicted eects strongly support the argument that civic education interacts with parental politicization in its eect on turnout. Table 4 displays the rst dierence estimates and standard errors associated with increasing American Government/Civics coursework 20

among students who report never discussing current events with their parents. For the 1992 election, a one-year increase in American Government/Civics coursework is expected to increase a students probability of voting by 11.4 percentage points. Across all six elections, the average estimated increase in probability is 8.7 percentage points. Pr Voting|0 Courses, No Discussion 1992 1993-94 1996 1998-00 2004 2004-06 .462 .297 .534 .365 .464 .426 Pr Voting|0 Courses, No Discussion 1992 1993-94 1996 1998-00 2004 2004-06 .462 .297 .534 .365 .464 .426 Pr Voting|0.5 Courses, No Discussion .519 .377 .581 .412 .539 .459 Pr Voting|1 Course, No Discussion .576 .377 .626 .461 .539 .493 .058* .039* .047* .047* .037* .033* .114* .080* .092* .096* .075* .067* SE of .016 .014 .017 .015 .015 .014 SE of .030 .029 .032 .031 .030 .028

Table 4: Eects of Civic Education and Parental Political Socialization on Voting. The estimates for 1992, 1993-94, 1996 and 1998-00 are based on the NELS:1988 data. The estimates for 2004 and 2004-06 are based on the NELS:2002 data. * p < .05.

Robustness of the Results

To test whether the results hold up using alternative statistical methods, I perform a matching analysis in which exposure to one year of American Government/Civics coursework (versus none) is the treatment variable. Matching ensures that the control group resembles the treatment group and, as a result, the estimates derived from the regressions using matched data will be unbiased with respect to the observed covariates (Rosenbaum and Rubin, 1985). 21

(a) NELS:1988: Before Matching

(b) NELS:1988: After Matching

2.5

2.0

1.5

Density

1.0

Density

0.5

0.0

0.0

0.2

0.4 Propensity Score

0.6

0.8

0.0

0.5

1.0

1.5

2.0

2.5

Control Group Treatment Group

Control Group Treatment Group

0.0

0.2

0.4 Propensity Score

0.6

0.8

(c) NELS:2002: Before Matching


3.5

(d) NELS:2002: After Matching

2.5

3.0

2.0

Density

1.5

Density

1.0

0.0

0.5

0.0

0.2

0.4

0.6

0.8

1.0

0.0

0.5

1.0

1.5

2.0

2.5

Control Group Treatment Group

Control Group Treatment Group

0.2

0.0

0.2

0.4

0.6

0.8

1.0

Propensity Score

Propensity Score

Figure 6: Propensity Score Distributions for Matched and Unmatched Data.

22

Observations included in the matched analysis are those students that completed either exactly one year of American Government/Civics coursework or none. This allows the estimated eects to be easily compared to those presented in Figures 4 and 5. Because several of the control variables are continuous, I employ nearest neighbor propensity score matching.21 Figure 6 displays, for each data set, the propensity score distributions for the treatment and control groups before and after matching. The graphs demonstrate that the balance is much improved in the matched data sets. More detailed statistics on the balance are presented in Table A-2.
(a) Direct Eect of American Gov/Civics
.025 200406
Unmatched 95% CI Matched 95% CI
q q

(b) Interactive Eect of American Gov/Civics


.063 200406
q

.030 .034

Unmatched 95% CI Matched 95% CI

.119 .074

2004

q q

2004

q q

.035 .054 199800 Elections


q q

.138 .087 199800 Elections


q q

.060 .032 1996


q q

.082 .091 1996


q q

.049 .052 199394


q q

.126 .073 199394


q q

.061 .045 1992


q q

.095 .111 1992


q q

.057 0.3 0.2 0.1 0.0 0.1 0.2 0.3 0.3 0.2 0.1 0.0 0.1

.149 0.2 0.3

Change in Probability of Voting (0 to 1 Course)

Change in Probability of Voting (0 to 1 Course)

Figure 7: The Eects of Civic Education Before and After Matching. Graphs (a) and (b)
present rst dierence estimates for both the unmatched and matched data sets. The dotted lines are the 95% condence intervals associated with the matched estimates.

Figure 7 displays the rst dierences associated with a one-year increase in American Government/Civics coursework for the both the matched and unmatched data sets. Regarding the direct eects, the matched estimates are very similar to the unmatched estiMore specically, I perform one-to-one matching with a caliper of 0.6. In the matched data sets, the transfer variable is collinear and so it is dropped from the regressions. The regression results are presented in Table A-1.
21

23

mates, although the eects for 2004 and 2004-06 elections lose statistical signicance with the matched data. Turning to the interactive eects, the matched estimates again support the conclusion that civic education strongly encourages students whose parents are not highly politicized to vote. The matched estimates are all signicant at the 10% level (or lower), and the magnitudes of the eects are sizable, with the largest being a 15 percentage point boost in the probability of turnout.

Conclusion

The eects of American Government/Civics coursework on turnout uncovered in this analysis are quite large; they are comparable in size to many of the interventions displayed in Table 1. Further, because the probability that a young citizen votes in a presidential election hovers around 50%, even a three or four percentage point increase in probability could be the dierence between this citizen turning out to vote and staying home on Election Day. Among students whose parents are not highly politicized, the eects of American Government/Civics coursework are even larger. This suggests that civic education compensates for relative lack of political socialization at home. A brief examination of the composition of this group shows that it contains a disproportionate amount of students from the lowest SES quartile, a demographic that is historically under-represented in politics (Verba, Schlozman and Brady, 1995). A graph of this distribution appears in Figure A-1. Civic education, through its ability to ll a politicizing void left by parents, may help level the participatory playing eld among young citizens.

24

Appendix

13.4 19
Often

SES Q1 (lowest) SES Q2 SES Q3 SES Q4

23.5 44 Amount of Discussion

16.9 22.7
Some

26 34.3

29.9 26.7
None

24.2 19.1

10

20

30

40

50

60

70

% of Students

Figure A-1: Socioeconomic Status and Discussion about Current Events. The percentages
for 1992, 1993-94, 1996 and 1998-00 are from the NELS:1988 and the percentages for 2004 and 200406 are from the NELS:2002.

25

Legend: 1992 or Before Between 1993 and 2004 2005 or Later No State Requirement

Figure A-2: State Graduation Requirements in American Government/Civics. This map displays the time period in which each state instituted a civic education requirement. The years refer to graduating classes. The data were collected from newspaper articles, state statutes, state administrative codes and the LBJ School Survey of State Civic Education (Tolo, 1999).

26

Table A-1: Regression Results


1992 1993-94 Unmatched Data 1996 1998-00 2004 2004-06 1992 1993-94 2004 Matched Data 1996 1998-00 2004-06

Civic Education Amer Gov/Civics

Amer Problems

Amer Gov/Econ

Other Gov

.184 (.084) .131 (.251) -.047 (.115) -.191 (.142)

.222 (.084) .118 (.225) .034 (.115) .006 (.13)

.145 (.09) .117 (.219) .072 (.121) .096 (.179)

.216 (.082) -.298 (.186) .283 (.11) .116 (.158)

.146 (.081) -.259 (.342) .285 (.103) .088 (.136)

.114 (.075) -.065 (.333) .095 (.11) .085 (.140)

.234 (.116) -.023 (.344) .047 (.233) -.430 (.227)

.264 (.115) -.488 (.337) -.190 (.178) -.030 (.290)

.222 (.123) .123 (.311) .266 (.259) .270 (.386)

.236 (.120) .157 (.467) .140 (.231) .003 (.263)

.144 (.121) .628 (.508) .337 (.286) -.031 (.299)

.120 (.123) .172 (.542) -.216 (.223) -.106 (.220)

Civic Engagement Extracurricular Hours

Service Club

Student Gov

.017 (.009) .170 (.047) .115 (.053)

.018 (.009) .142 (.044) .151 (.053)

-.003 (.02) .362 (.099) .192 (.103)

.005 (.019) .251 (.094) .167 (.106)

S.S. Look Forward

.038 (.011) .028 (.055) .168 (.043) .034 (.032)

.042 (.011) .038 (.051) .141 (.041) .043 (.032)

.036 (.012) .082 (.059) .275 (.052) .022 (.036)

.036 (.011) .091 (.053) .142 (.043) .127 (.035)

.053 (.021) .030 (.102) .067 (.078) .024 (.065)

.056 (.021) -.018 (.096) .044 (.078) .154 (.066)

.083 (.024) .290 (.118) .227 (.091) .028 (.076)

.057 (.021) .101 (.100) .159 (.088) .117 (.066)

Citizenship - Importance

27
.047 (.046) .077 (.066) .234 (.072) .224 (.045) .284 (.045) .034 (.045) .204 (.071) .004 (.072) .232 (.046) .171 (.049) .01 (.05) .015 (.072) .192 (.08) .317 (.051) .257 (.050) .068 (.046) -.03 (.071) -.025 (.076) .294 (.046) .195 (.049) -.001 (.023) .101 (.058) .309 (.066) .362 (.041) .118 (.041) .002 (.023) .104 (.058) .239 (.064) .392 (.041) .139 (.041) .035 (.058) -.569 (.112) -.597 (.127) -.074 (.125) -.226 (.291) .090 (.057) -.141 (.114) -.315 (.141) .077 (.127) .074 (.286) -.077 (.066) .089 (.128) -.345 (.145) .210 (.136) .104 (.338) -.073 (.059) .179 (.12) -.350 (.154) .088 (.129) -.111 (.382) -.136 (.054) .016 (.105) -.518 (.13) -.280 (.101) -.105 (.301) -.179 (.053) .113 (.097) -.539 (.128) -.280 (.103) -.057 (.312)

Patriotism - Importance

.271 (.047) .096 (.039)

.258 (.045) .110 (.039)

.330 (.098) -.015 (.082)

.281 (.096) .010 (.083)

Home Resources Parent Education

Newspaper

S.E.S. Status

Discuss

Discuss (Parent)

-.010 (.088) -.081 (.13) .435 (.142) .404 (.117) .192 (.093)

.207 (.084) .118 (.136) -.188 (.133) .281 (.123) .198 (.091)

-.163 (.101) -.023 (.141) .300 (.162) .373 (.133) .400 (.101)

.095 (.091) -.214 (.14) -.085 (.145) .34 (.123) .271 (.098)

-.076 (.045) .067 (.110) .405 (.135) .573 (.121) .184 (.089)

-.008 (.047) .174 (.118) .215 (.14) .576 (.117) .209 (.088)

Student Demographics Male

Black

Asian

Hispanic

Amer. Indian

.072 (.111) -.516 (.200) -1.079 (.287) -.352 (.301) -.163 (.530)

-.034 (.111) -.362 (.183) -.412 (.314) -.137 (.291) -.175 (.427)

.191 (.134) -.135 (.23) -.560 (.32) .423 (.346) -.301 (.721)

.051 (.113) .033 (.215) -.370 (.293) .173 (.32) .763 (.8)

-.083 (.121) -.176 (.206) -.481 (.269) .042 (.237) -.292 (.757)

-.193 (.125) -.121 (.196) -.662 (.26) -.463 (.237) -.532 (.841)

Table A-1: Regression Results


1992 1993-94 Unmatched Data 1996 1998-00 2004 2004-06 1992 1993-94 2004 Matched Data 1996 1998-00 2004-06

Parent English

H.S. Diploma

.254 (.258) .336 (.442)

.479 (.252) .472 (.516)

.349 (.314) -.199 (.482)

1.032 (.326) .493 (.535)

.555 (.232) .108 (.356)

.390 (.215) .334 (.376)

Transfer Student

.45 (.114) .38 (.210) -.181 (.122)

.357 (.125) .423 (.238) -.078 (.123)

.468 (.128) .282 (.241) -.361 (.143)

.444 (.129) .222 (.254) -.337 (.134)

.527 (.103) .329 (.212) -.096 (.604)

.426 (.102) .349 (.227) .741 (.559)

School and State Chars. Urban

Rural

Private

% Attending College

State Req. Civ. Ed.

.023 (.087) -.015 (.079) -.116 (.109) .053 (.033) .061 (.068)

-.015 (.084) .055 (.079) -.084 (.11) -.030 (.037) .022 (.067)

.076 (.086) .011 (.077) .106 (.115) .016 (.034) .041 (.069)

.053 (.087) .145 (.077) -.008 (.112) .015 (.035) .185 (.066)

.022 (.082) .024 (.085) -.009 (.092) .035 (.034) -.079 (.071)

-.008 (.08) .024 (.082) .042 (.09) .017 (.033) -.034 (.069)

.202 (.167) -.052 (.144) -.323 (.203) .036 (.063) .337 (.129)

.096 (.169) .081 (.146) -.197 (.22) -.117 (.064) .115 (.131)

.179 (.178) -.059 (.145) .124 (.235) .074 (.068) .099 (.124)

.218 (.159) .249 (.154) -.31 (.212) .105 (.069) .215 (.131)

.074 (.139) .297 (.214) -.104 (.168) .060 (.076) .091 (.132)

-.009 (.139) .336 (.204) -.14 (.169) .058 (.073) .222 (.131)

Interaction Amer Gov/Civics Discuss

28

(Constant)

-.308 (.105) -1.914 (.361) 5775

-.155 (.105) -2.176 (.361) 5769

-.247 (.119) -1.247 (.378) 4985

-.194 (.106) -2.17 (.369) 5010

-.149 (.087) -1.919 (.338) 6565

-.160 (.079) -1.969 (.355) 6575

-.420 (.163) -1.322 (.707) 1560

-.177 (.164) -2.744 (.729) 1560

-.345 (.183) -1.12 (.737) 1346

-.107 (.170) -3.693 (.739) 1340

-.399 (.171) -1.72 (.636) 1542

-.342 (.161) -2.178 (.662) 1544

Estimates are logit coecients. Clustered standard errors are shown in parentheses. Amer Gov/Civics is coded as dummy variable for the matched regressions, where 1 indicates completion of a year-long course.

NELS:1988 Matched Unmatched Mean Mean Mean Mean Treated Control Treated Control 0.59 0.43 0.67 0.52 0.60 0.57 0.01 0.05 0.03 3.81 0.31 0.43 2.57 0.07 0.17 0.09 3.89 0.33 0.41 2.70 0.02 0.06 0.03 3.78 0.32 0.41 2.62 0.03 0.09 0.05 3.86 0.34 0.42 2.60 0.01 0.02 0.03 6.76 0.38 0.27 0.02 0.20 0.09 6.91 0.44 0.26 0.59 0.56 0.61 0.58 0.01 0.03 0.05 6.90 0.43 0.27 0.55 0.38 0.62 0.45 0.59 0.44 0.67 0.51 0.54 0.38 0.61 0.47 0.58 0.56 0.01 0.07 0.05 6.74 0.41 0.28

NELS:2002 Unmatched Matched Mean Mean Mean Mean Treated Control Treated Control

29
3.12 0.77 0.06 1.92 2.36 0.45 0.07 0.04 0.04 0.01 0.94 0.99 0.05 0.24 0.45 0.16 4.35 0.51 3.41 0.79 0.24 1.91 2.38 0.46 0.07 0.08 0.06 0.01 0.88 0.98 0.06 0.26 0.33 0.18 4.41 0.35 3.23 0.78 0.14 1.92 2.36 0.43 0.08 0.05 0.05 0.01 0.92 0.99 0.06 0.27 0.36 0.19 4.40 0.41 3.21 0.77 0.12 1.91 2.36 0.47 0.06 0.05 0.03 0.01 0.94 0.99 0.06 0.24 0.43 0.18 4.39 0.42

Dependent Variables (0,1) Voted 1992 Voted 1993-94 Voted 1996 Voted 1998-00 Voted 2004 Voted 2004-06 Covariates Amer Problems (# of credits) Amer Gov/Econ (# of credits) Other Gov (# of credits) Extracurricular Hours (# per week) Service Club (0,1,2) Student Gov (1,2,3) S.S. Look Forward (1,2,3) Citizenship - Importance (1,2,3) Patriotism - Importance (1,2,3) Parent Edu. (6 categories) Newspaper (0,1) S.E.S. ( -2.25, 2.2) Discuss (0,1,2) Discuss (Parent; 0,1,2) Male (0,1) Black (0,1) Asian (0,1) Hispanic (0,1) Amer. Indian (0,1) Parent English (0,1) H.S. Diploma (0,1) Transfer (0,1) Urban (0,1) Rural (0,1) Private (0,1) % Attending College (6 categories) State Req. Civ. Ed. (0,1) 2.40 2.21 4.77 0.67 0.17 2.07 2.36 0.50 0.10 0.06 0.07 0.01 0.89 0.99 0.00 0.32 0.29 0.24 4.56 0.74 2.37 2.13 4.92 0.69 0.25 2.04 2.38 0.48 0.08 0.08 0.09 0.01 0.86 0.98 0.00 0.28 0.16 0.32 4.93 0.41 2.41 2.14 5.05 0.69 0.30 2.04 2.37 0.49 0.09 0.08 0.10 0.01 0.85 0.98 0.00 0.33 0.16 0.33 4.96 0.59

2.41 2.17 4.81 0.67 0.21 2.07 2.37 0.48 0.08 0.07 0.09 0.01 0.87 0.98 0.00 0.3 0.21 0.28 4.77 0.59

Table A-2: Balance Statistics for Matched and Unmatched Data. Numbers in bold indicate that the dierence between the means is signicant at the 10% level. The treatment is one year of coursework in American Government/Civics.

References
Alvarez, R. Michael, Delia Bailey and Jonathan N. Katz. 2008. The Eect of Voter Identication Laws on Turnout. California Institute of Technology Social Science Working Paper No. 1267R . Anderson, Lee, Lynn B. Jenkins, James Leming, Walter B. McDonald, Ina V.S. Mullis, Mary Jane Turner and Judith S. Wooster. 1990. The Civics Report Card: Trends in Achievement from 1976 to 1988 at Ages 13 and 17; Achievement in 1988 at Grades 4, 8, and 12. National Center for Education Statistics. Boas, Katherine. December 17, 2002. American Civics Class to be Added at Ponanganset. Providence Journal-Bulletin . Brians, Craig Leonard and Bernard Grofman. 2001. Election Day Registrations Eect on U.S. Voter Turnout. Social Science Quarterly 82(1):170182. Current Population Survey. 2010. Voting and Registration Supplement. U.S. Census Bureau. Curtin, Thomas R., Steven J. Ingels, Shiying Wu, Ruth Heuer and Jerey Owings. 2002. Base-Year to Fourth Follow-up Data File Users Manual. National Center for Education Statistics. Dale, Allison and Aaron Strauss. 2009. Dont Forget to Vote: Text Message Reminders as a Mobilization Tool. American Journal of Political Science 53(4):787804. Delli Carpini, Michael and Scott Keeter. 1996. What Americans Know about Politics and Why it Matters. New Haven, CT: Yale University Press. Ehman, Lee H. 1969. An Analysis of the Relationships of Selected Educational Variables with the Political Socialization of High School Students. American Educational Research Journal 6(4):559580. Fenster, Mark J. 1994. The Impact of Allowing Day of Registration Voting on Turnout in U.S. Elections from 1960 to 1992: A Research Note. American Politics Research 22(1):74 87. Franklin, Daniel P. and Eric E. Grier. 1997. Eects of Motor Voter Legislation: Voter Turnout, Registration, and Partisan Advantage in the 1992 Presidential Election. American Politics Research 25(1):104117. Gerber, Alan S. and Donald P. Green. 2000. The Eects of Canvassing, Telephone Calls, and Direct Mail on Voter Turnout: A Field Experiment. American Political Science Review 94(4):653663. Gerber, Alan S. and Donald P. Green. 2001. Do Phone Calls Increase Voter Turnout? A Field Experiment. Public Opinion Quarterly 65:7585. 30

Gerber, Alan S., Donald P. Green and Ron Shachar. 2003. Voting May Be Habit-Forming: Evidence from a Randomized Field Experiment. American Journal of Political Science 47(3):540550. Gimpel, James G., J. Celeste Lay and Jason E. Schuknecht. 2003. Cultivating Democracy: Civic Environments and Political Socialization in America. Washington, DC: Brookings Institution Press. Green, Donald P., Alan S. Gerber and David W. Nickerson. 2003. Getting out the Vote in Local Elections: Results from Six Door-to-Door Canvassing Experiments. Journal of Politics 65(4):10831096. Hess, Robert D. and Judith V. Torney. 1967. The Development of Political Attitudes in Children. Chicago, IL: Aldine Pub. Co. Highton, Benjamin and Raymond E. Wolnger. 2001. The First Seven Years of the Political Life Cycle. American Journal of Political Science 45(1):202209. Ingels, Steven J., Daniel J. Pratt, David Wilson, Laura J. Burns, Douglas Currivan, James E. Rogers and Sherry Hubbard-Bednasz. 2007. Education Longitudinal Study of 2002 BaseYear to Second Follow-up Data File and Documentation. National Center for Education Statistics. Jennings, M. Kent, Gregory B. Markus, Richard G. Niemi and Laura Stoker. 1997. YouthParent Socialization Study, 1965-1997: Four Waves Combined. Ann Arbor, MI: Interuniversity Consortium for Political and Social Research. Jennings, M. Kent, Laura Stoker and Jake Bowers. 2009. Politics Across Generations: Family Transmission Reexamined. Journal of Politics 71(2):118. Kiousis, Spiro and Michael McDevitt. 2008. Agenda Setting in Civic Development: Eects of Curricula and Issue Importance on Youth Voter Turnout. Communication Research pp. 122. Knack, Stephen. 1995. Does Motor Voter Work? Evidence from State-Level Data. Journal of Politics 57(3):796811. Knack, Stephen. 2001. Elecion-day Registration: The Second Wave. American Politics Research 29(1):6578. Langton, Kenneth P. and M. Kent Jennings. 1968. Political Socialization and the High School Civics Curriculum in the United States. American Political Science Review 62(3):852867. Le, Lisa. August 26, 1993. Path to Diploma Has New Wrinkles: Schools Phasing in Tougher Requirements for Students on the Road to Graduation. Washington Post . 31

Litt, Edgar. 1963. Civic Education, Community Norms, and Political Indoctrination. American Sociological Review 28(1):6975. McDevitt, Michael and Spiro Kiousis. 2004. Education for Deliberative Democracy: The Long-term Inuence of Kids Voting USA. Medford, MA: Center for Information and Research on Civic Learning and Engagement. Meirick, Patrick C. and Daniel B. Wackman. 2004. Kids Voting and Political Knowledge: Narrowing Gaps, Informing Votes. Social Science Quarterly 85(5):11611177. Mitchell, Glenn E. and Christopher Wlezien. 1995. The Impact of Legal Constraints on Voter Registration, Turnout, and the Composition of the American Electorate. Political Behavior 17(2):179202. Nickerson, David W. 2006. Volunteer Phone Calls Can Increase Turnout: Evidence from Eight Field Experiments. American Politics Research 34(3):271292. Niemi, Richard G. and Jane Junn. 1998. Civic Education: What Makes Students Learn. New Haven, CT: Yale University Press. Niemi, Richard G. and Julia Smith. 2001. Enrollments in High School Government Classes: Are We Short-Changing Both Citizenship and Political Science Training? PS: Political Science and Politics 34:281287. Niemi, Richard G. and Julia Smith. 2003. The Accuracy of Students Reports of Course Taking in the 1994 National Assessment of Educational Progress. Educational Measurement: Issues and Practice 22(1):1521. Pasek, John, Lauren Feldman, Daniel Romer and Kathleen Hall Jamieson. 2008. Schools as Incubators of Democratic Participation: Building Long-Term Political Ecacy with Civic Education. Applied Developmental Science 12(1):2637. Popkin, Samuel L. and Michael A. Dimock. 1999. Political Knowledge and Citizen Competence. In Citizen Competence and Democratic Institutions, ed. Stephen L. Elkin and Karol Edward Soltan. Pennsylvania State University Press. Ravitch, Diane. March 9, 2010. Why I Changed My Mind about School Reform. Wall Street Journal . Rhine, Staci L. 1995. Registration Reform and Turnout Change in the American States. American Politics Research 23(4):409426. Roey, Stephen, Robert Perkins, Nancy Caldwell, Keith Rust and Janis Brown. 2007. The 2000 High School Transcript Study Tabulations: Comparative Data on Credits Earned and Demographics for 2000, 1998, 1994, 1990, 1987, and 1982 High School Graduates. National Center for Education Statistics. 32

Rosenbaum, Paul R. and Donald B. Rubin. 1985. Constructing a Control Group Using Multivariate Matched Sampling Methods. The American Statistician 39(1):3338. Rosenstone, Steven J. and John Mark Hansen. 1993. Mobilization, Participation, and Democracy in America. New York, NY: Longman. Shettle, Carolyn, Shep Roey, Joy Mordica, Robert Perkins, Christine Nord, Jelena Teodorovic, Janis Brown, Marsha Lyons, Chris Averett and David Kastberg. 2007. Americas High School Graduates: Results from the 2005 NAEP High School Transcript Study. National Center for Education Statistics. Simon, James and Bruce D. Merrill. 1998. Political Socialization in the Classroom Revisted: The Kids Voting Program. Social Science Journal 35(1):2942. Squire, Peverill, Raymond E. Wolnger and David P. Glass. 1987. Residential Mobility and Voter Turnout. American Political Science Review 81(1):4566. Stoker, Laura and M. Kent Jennings. 1995. Life-Cycle Transitions and Political Participation: The Case of Marriage. American Political Science Review 89(2):421433. Strate, John M., Charles J. Parrish, Charles D. Elder and Coit Ford. 1989. Life Span Civic Development and Voting Participation. American Political Science Review 83(2):443464. Tedin, Kent L. 1974. The Inuence of Parents on the Political Attitudes of Adolescents. American Political Science Review 68(4):15791592. Teixeira, Ruy A. 1992. The Disappearing American Voter. Washington, DC: Brookings Institution. Tolo, Kenneth. 1999. School Survey of State Civic Education (K-12) Requirements and Policies. Austin, TX: Lyndon B. Johnson School of Public Aairs. Valentino, Nicholas A. and David O. Sears. 1998. Event-Driven Political Communication and the Preadult Socialization of Partisanship. Political Behavior 20(2):127154. Verba, Sidney, Kay Lehman Schlozman and Henry E. Brady. 1995. Voice and Equality: Civic Voluntarism in American Politics. Cambridge, MA: Harvard University Press. Vontz, Thomas S., Kim K. Metcalf and John J. Patrick. 2000. Project Citizen and the Civic Development of Adolescent Students in Indiana, Latvia and Lithuania. Washington, DC: Oce of Educational Research and Improvement. Wolnger, Raymond E. and Steven J. Rosenstone. 1980. Who Votes? New Haven, CT: Yale University Press. Zaller, John R. 1992. The Nature and Origins of Mass Opinion. New York, NY: Cambridge University Press. 33

Você também pode gostar