Youth bulges in communities: the effects of age structure on adolescent civic knowledge and civic participation.

Abstract

Youth bulges, cohorts of 16- to 25-year-olds disproportionately large relative to the adult population, are linked with social upheaval in historical research. Limited civic knowledge and heightened civic participation in adolescence, resulting from socialization in communities with large populations of children, are hypothesized to be developmental precursors to the political activism characteristic of youth constituting bulges. In two studies with nationally representative samples, adolescents in communities with disproportionately large populations of children were found to have less civic knowledge than equivalent adolescents in communities without large populations of children. In both studies, civic participation was predicted by the interaction of a community's proportion of children and its poverty level. Similar patterns were identified in a third study using country-level data. Together, the findings demonstrate that the youthfulness of communities and countries influences civic development.

Full text

Page 1

Research Article
Youth Bulges in Communities
The Effects of Age Structure on Adolescent Civic
Knowledge and Civic Participation
Daniel Hart,1Robert Atkins,2Patrick Markey,3and James Youniss4
1Rutgers University,2Temple University,3Villanova University, and4The Catholic University of America
ABSTRACT––Youth bulges, cohorts of 16- to 25-year-olds dispro-
portionately large relative to the adult population, are linked
with social upheaval in historical research. Limited civic
knowledge and heightened civic participation in adolescence,
resulting from socialization in communities with large popula-
tions of children, are hypothesized to be developmental pre-
cursors to the political activism characteristic of youth
constituting bulges. In two studies with nationally representa-
tive samples, adolescents in communities with disproportion-
ately large populations of children were found to have less civic
knowledge than equivalent adolescents in communities without
large populations of children. In both studies, civic participa-
tion was predicted by the interaction of a community’s propor-
tion of children and its poverty level. Similar patterns were
identified in a third study using country-level data. Together, the
findings demonstrate that the youthfulness of communities and
countries influences civic development.
Youth bulge refers to a cohort of youth between the ages of 16 and 25
that is unusually large relative to the adult population in a society.
Historical research has linked youth bulges to revolutions in 17th-
century England, 18th-century France, and 20th-century Indonesia
(Goldstone, 2002; Moller, 1968); to political activism in Western and
Middle Eastern countries (Huntington, 1996); and to the prevalence of
warfare throughout the world (Mesquida & Wiener, 1999; Urdal,
2002). The associations of youth bulges to activism, revolution, and
warfare—the latter two particularly likely in societies experiencing
simultaneous economic difficulties (Huntington, 1996)––leads na-
tional-security analysts (e.g., Helgerson, 2002) and the popular press
(e.g., Zakaria, 2001) to classify countries with disproportionately large
cohorts of youth as at risk for the emergence of political extremism.
Although youth bulges have captured the attention of historians,
political scientists, demographers, and national-security analysts,
they have escaped the scrutiny of psychologists. This is unfortunate
for two reasons. First, a social-psychological analysis can contribute to
the explanation of the origins of youth-bulge phenomena. Second, a
focus on the age structure of societies can broaden psychology’s ap-
preciation for community influences on development. In the three
studies reported in this article, we tested a hypothesis concerning the
relation of a community’s age structure to adolescents’ acquisition of
fundamental civic qualities.
Our hypothesis is that the extraordinary willingness to participate
in political transformation that is characteristic of youth in a youth
bulge is a consequence, in part, of community influences on civic
development. Countries with youth bulges have many communities in
which children and adolescents make up a large fraction of the
population. An adolescent living in a community in which a large
fraction of the population is composed of children and adolescents, a
child-saturated community, will interact more often with peers, and
consequently will be more influenced by them, than will an adolescent
in a community with relatively few children and many adults, or an
adult-saturated community. Our hypothesis is that child saturation
influences adolescents’ acquisition of civic knowledge and civic
participation. The effects of child saturation on adolescents’ civic
knowledge and civic participation lay the foundation for the partici-
pation in political transformation that is characteristic of young adults
in youth bulges.
Civic knowledge is essential to effective citizenship. Citizens
knowledgeable in civics are consistent in political ideology, under-
stand public policy, judge politicians by their leadership rather than
their personal character, trust institutions, and tolerate minority
groups (Delli Carpini & Keeter, 1996; Galston, 2001). Conversely,
citizens low in civic knowledge––and we hypothesize that low civic
knowledge is characteristic of those in child-saturated communities––
have weakly rooted political ideologies, have shallow understanding of
public policy, distrust existing societal institutions, and are intolerant
of minority groups (Delli Carpini & Keeter, 1996; Galston, 2001).
Research demonstrates that revolutions result in part from the suc-
cessful recruitment of large segments of the populace (effective
leadership of the movement and other factors are necessary as well;
see Goldstone, 2001, for a review). We suggest that the distrust of
existing societal institutions and shallow political ideologies that are
characteristic of individuals low in civic knowledge increase the
likelihood that individuals can be recruited into revolutionary
movements.
Adolescents who as young adults will constitute youth bulges are
also hypothesized to be more active civically than other adolescents.
Address correspondence to Daniel Hart, Department of Psychology,
Rutgers University, Camden, NJ 08102; e-mail: hart@camden.rutgers.
edu.
PSYCHOLOGICAL SCIENCE
Volume 15—Number 9
591
Copyright r 2004 American Psychological Society

Page 2

Analysts of youth bulges have focused on the destructive conse-
quences for governments of civic activity when it assumes the form of
protest and revolt (e.g., Goldstone, 2002). Destructive civic activity
may be most likely when a youth bulge is accompanied by depressed
economic conditions (Huntington, 1996; Urdal, 2002), because poor
youth foresee a future without job prospects if the existing social
structures prevail. In the studies that follow, we tested the possibility
of an interaction between youth bulges and poverty in predicting civic
activity.
Although theorists have focused on the relation of youth bulges to
unrest, protest, and revolt, youth bulges can also foster system-sus-
taining civic activity. For example, the civil rights movement in the
United States has been linked to a youth bulge (Huntington, 1996;
Moller, 1968). Because the first two studies in this report focus on
civic activity in the United States in the 1990s––an era largely un-
marked by widespread protest––we focus only on one form of con-
structive civic participation, volunteer community service.
Communities affect civic knowledge and civic participation through
social influence. Knowledge, attitudes, and behaviors are shaped in
daily interactions with other people (Latane ´, Liu, Nowak, Benevento,
& Zheng, 1995). Because adults have more experience in their so-
cieties than youth, they should have more civic knowledge than do
children and adolescents. Research confirms this expectation. For
example, in a survey of a representative national sample, the National
Household Educational Survey of 1996 (NHES:96), adults were asked
five questions about the federal government. The same five questions
were posed to a national sample of adolescents in grades 9 through 12.
Adults answered 55% of the questions correctly, whereas adolescents
knew the answers only 40% of the time (calculated from Collins et al.,
1997).1Compared with other communities, child-saturated commu-
nities have fewer inhabitants (i.e., adults) with high levels of civics
expertise that can be transmitted through informal contact to children
and adolescents. Consequently, we predicted that youth living in
child-saturated communities would know less about the political
system than youth living in adult-saturated communities.
If social influence operates as suggested, then any form of civic
activity more common in adolescents than in adults should be ac-
quired better by adolescents in child-saturated neighborhoods than by
adolescents in adult-saturated neighborhoods. Voluntary community
service is one such activity, as it is more common in American ado-
lescents than among adults (Lopez, 2003). For example, in the
NHES:96, 51% of children and adolescents in grades 6 through 12
answered affirmatively to a question asking if they volunteered in their
schools or communities in the past year. Only 38% of the adults in the
NHES:96 survey answered ‘‘yes’’ to a parallel question (calculated
from Collins et al., 1997).2Child-saturated communities are conse-
quently more likely to offer models of involvement in volunteer ac-
tivities than are adult-saturated ones. Adolescents should be more
likely to volunteer in communities in which many other people vol-
unteer (child-saturated communities) than in communities in which
volunteering is less common (adult-saturated communities). Conse-
quently, we predicted that adolescents in child-saturated communities
would be more likely to volunteer than would adolescents in adult-
saturated neighborhoods.
The child saturation of a community should influence only those
qualities of civic development for which there are substantial differ-
ences in mean level between youth and adults. Such differences exist
for knowledge and volunteering (as we have just documented), but not
for other civic qualities such as tolerance for other individuals’ views
(for the centrality of tolerance in democratic societies, see Sullivan &
Transue, 1999). For example, in the NHES:96, the adult and youth
samples were asked whether people opposed to religion ought to be
allowed to speak in their communities, and whether books with un-
popular messages should be kept in public libraries. Fifty percent of
the youth advocated tolerance in response to both questions, as did
51% of the adults (calculated from Collins et al., 1997).3Because
adult and youth populations do not differ in tolerance, social-influence
theory predicts that there should not be a relation between community
child saturation and tolerance in youth. However, tolerance is clearly
influenced by neighborhood factors and economic distress (Persell,
Green, & Gurevich, 2001). Our theoretical explanation based on so-
cial influence would be strengthened if child saturation’s effects are
(a) limited to civic qualities for which there are substantial mean
differences between youth and adults and (b) distinguishable from the
effects of other neighborhood qualities such as poverty.
In Study 1, we regressed adolescents’ civic knowledge, volunteer-
ing, and tolerance on their communities’ child-saturation levels. We
also included community poverty as a predictor. Finally, the analyses
controlled for individual-level and family-level factors associated with
adolescents’ civic development.
STUDY 1
Method
Participants
The sample was drawn from the National Household Educational
Survey of 1999 (NHES:99). Households representative of the United
States were telephoned, and in households with children, one child or
adolescent served as the participant (for design details, see Nolin
et al., 2000). Children and adolescents in grades 6 through 12 an-
swered questions concerning community service. We included only
those participants age 16 and younger to focus on children and ado-
lescents prior to the youth-bulge age range (16–25), as well as to
maintain roughly equivalent age ranges across all three studies in this
article. Children and adolescents not living with one or more bio-
logical or adoptive parents were excluded from the sample, in order to
permit the assessment of parental educational attainment on civic de-
velopment. These two restrictions resulted in a final sample of 5,616. A
subsample of 2,555, consisting of participants in grades 9 through 12,
was asked questions tapping civic knowledge and tolerance.
Survey Questions
Two parallel sets of five questions were used to assess civic knowledge
(e.g., ‘‘Which of the two major political parties is most conservative at
1The civic-knowledge scores were calculated from the average number of
correct responses for the variables CAVP, CALAW, CAHOUSE, CAVETO, and
CACONSRV for the adult sample in the NHES, with sampling weights used to
estimate scores for the adult population of the United States.
2The variables are SACTY for the youth sample and CASERVC for the adult
sample. National estimates were obtained using the sampling weights.
3The variables for the youth sample are CYBOOK and CYAGNST; the var-
iables for the adult sample are CABOOK and CAAGNST. National estimates
were obtained using the sampling weights.
592
Volume 15—Number 9
Youth Bulges

Page 3

the national level?’’). A given participant responded to one of the two
sets of questions. The number of correct answers was standardized and
indexed civic knowledge.
Participants were asked if they had done voluntary community
service or volunteer work at school or in the community in the pre-
vious year. Those who responded affirmatively to this question, and
who also reported that the community service was neither required by
school nor contributed to a grade, were judged to be involved in
voluntary community service.
Finally, participants were asked whether in their communities a
person ought to be permitted to make a speech opposing religion and
whether books with unpopular messages should be banned from
public libraries. Those who advocated tolerance in both instances
were judged to be tolerant.
Neighborhood Factors
The child saturation of each participant’s community was estimated
from the 1990 U.S. Census (Geolytics, 1999). Community was deter-
mined by the zip (postal) code corresponding to the participant’s home
address. Epidemiological research has demonstrated that zip code
areas correspond to communities of interacting individuals, as re-
flected by patterns of transmission of communicable diseases (e.g.,
Acevedo-Garcia, 2001) and the prevalence of particular kinds of
behavior (e.g., Gruenewald, Johnson, & Treno, 2001). The average
population of zip code areas in 1990 was 8,478 (SD512,330). We
divided the number of children (birth to age 16) by the number of
adults (age 21 and older) in each zip code area to create a child-
saturation quotient (M5.25, SD5.05). In addition, the proportion of
households below the poverty line was calculated for each zip code
(M5.12, SD5.09).
Individual and Family Factors
Participants’ race-ethnicity, age, gender, membership in school-based
and extracurricular teams or clubs, and school grades (reported by the
parents on a 5-point scale, from ‘‘mostly F’s’’ to ‘‘mostly A’s’’) were
available in the survey, as was information on parents’ educational
attainment (mean number of years of education completed by parents
in the household), household income (as reported on an 11-point
scale), parental presence in the home, and whether one or both par-
ents spoke English at home. These characteristics are associated with
volunteering and civic knowledge (Niemi, Hepburn, & Chapman,
2000).
Results and Discussion
Scores for predictors were centered to reduce collinearity and to ease
interpretation of interaction terms. All analyses were conducted using
the weights included in the data set.4
Civic Knowledge
The results in Table 1 indicate that age, gender, educational
achievement, and race-ethnicity were predictive of individual differ-
TABLE 1
Regression of Adolescent Civic Qualities Measured in 1999 on Individual, Family, and Neighborhood Predictors
Predictor
Civic knowledgeVolunteeringTolerance
B SE
b
B SE
Exp(B)
B SE
Exp(B)
(Constant)
Individual
Black
Hispanic
Other non-White
Age
Gender (males50, females51)
School grades
Extracurricular club or team (yes51, 05no)
School club or team (yes51, 05no)
Family
Parental educational attainment
Income
Two-parent vs. one-parent home
Father-only home vs. mother-only home
One parent speaks a non-English language
Two parents speak a non-English language
Neighborhood
Percentage families in poverty
Child saturation
Interaction of percentage in poverty and
child saturation
0.160.05
?1.880.13
?0.040.12
?0.27
?0.28
?0.03
0.12
?0.37
0.25
—
—
0.07
0.08
0.10
0.03
0.04
0.03
—
—
?0.10n
?0.10n
?0.01
0.10n
?0.19n
0.25n
—
—
?0.31
?0.27
?0.10
0.04
0.22
0.23
0.78
0.50
0.13
0.14
0.18
0.02
0.08
0.05
0.10
0.09
0.73
0.76n
0.91n
1.04
1.24
1.26n
2.19n
1.65n
?0.27
0.05
?0.22
0.42
0.05
0.02
—
—
0.17
0.18
0.25
0.07
0.11
0.06
—
—
0.76
1.05
0.80
1.53n
1.05
1.02
—
—
0.06
0.02
0.01
0.06
0.00
?0.03
0.01
0.01
0.07
0.12
0.20
0.11
0.18n
0.05
0.00
0.02
0.00
?0.01
0.04
0.00
0.37
0.57
?0.59
?0.34
0.02
0.02
0.13
0.24
0.36
0.24
1.04n
1.00n
1.45
1.77n
0.55n
0.71
0.05
0.03
0.01
0.02
0.24
?0.43
0.02
0.02
0.16
0.29
0.49
0.28
1.05n
1.03
1.01
1.03
1.28
0.65
?0.44
?0.01
0.30
0.00
?0.04
?0.05n
0.19
0.02
0.54
0.01
1.21
1.02n
?1.66
0.01
0.74
0.01
0.19n
1.01
?0.01 0.040.00
?0.170.08 0.84n
?0.01 0.11 0.99
np < .05.
4The standard errors reported in both Table 1 and Table 2 were corrected for
biases introduced by sampling (the corrected standard errors are 1.4 times
larger than those obtained under standard assumptions; see Nolin et al., 2000).
Volume 15—Number 9
593
D. Hart et al.

Page 4

ences in civic knowledge (the same pattern is observed with adult
samples; e.g., Delli Carpini & Keeter, 1996). Parental educational
attainment was also a predictor of civic knowledge.
The findings relevant to the hypotheses are in the bottom three rows
of Table 1. As predicted, child saturation was negatively associated
with civic knowledge. From the regression equation, it was estimated
that adolescents in a community with a child-saturation quotient of .2
could be expected to be about .2 SD higher in civic knowledge than
equivalent adolescents in a community with an extreme child-satu-
ration quotient of .4 (13 SDs).
Volunteering
Table 1 also presents the results of the logistic regression of volunteer
community service on neighborhood factors and individual- and
family-level variables. Because research has demonstrated that teams
and clubs facilitate entry into community service (Hart, Atkins, &
Ford, 1998), membership in school-based teams and clubs and
membership in extracurricular teams and clubs were entered as
predictors to ensure correct specification of the model. Table 1 indi-
cates that educational achievement, club membership, and parental
educational attainment showed the associations with community
service found in previous studies (e.g., Hart et al.).
The hypothesis-relevant results are in the bottom rows of the
table. The main effect for neighborhood child saturation was positive
and significant, but qualified by an interaction with poverty. This
interaction is illustrated in the top panel of Figure 1. In low-poverty
neighborhoods (in which the poverty rate was 1 SD below the mean, or
approximately 3%), the rate of participation in community service was
nearly twice as high in neighborhoods with a child-saturation quotient
of .4 as in neighborhoods with a child-saturation quotient of .2. In
moderate-poverty neighborhoods (1 SD above the average, or a pov-
erty rate of approximately 21%), child saturation had little effect on
participation in community service. Extremely poor neighborhoods
(13 SD, 40% poverty) with a child-saturation quotient of .4 had
extremely low rates of participation in community service.
Tolerance
Finally, the results in Table 1 suggest that neighborhood poverty level,
but not neighborhood child saturation, was negatively associated with
tolerance.
Discussion
Study 1 demonstrates that adolescents’ civic knowledge and partici-
pation in community service are associated with the extent of
child saturation in an adolescent’s community, a pattern consistent
with the social-influence hypothesis. The results also suggest that
poverty and child saturation interact, with extremes of poverty and
child saturation perhaps exhausting a community’s capacity to direct
adolescents’ civic tendencies into community service. Finally, as
predicted, community poverty, but not child saturation, was related to
tolerance.
STUDY 2
The first goal for Study 2 was to replicate Study 1’s novel findings
concerning child saturation. The second goal for Study 2 was to de-
termine whether the effects reported in Study 1 concerning the as-
sociations of community-level factors with adolescents’ civic
knowledge and participation are attributable to unmeasured parental
civic qualities. Parents do influence the political development of their
children (Jennings & Niemi, 1981), and perhaps it is parents, not
communities, that are responsible for the findings reported in Study 1.
For example, parents in child-saturated neighborhoods might know
less about their governments than do parents in adult-saturated
neighborhoods, and consequently might transmit less information to
their children. Similarly, volunteering among adolescents is influ-
enced by parental modeling (McLellan & Youniss, 2003), and perhaps
parents in affluent, child-saturated neighborhoods volunteer more
often than do parents in affluent, adult-saturated neighborhoods.
In Study 2, we controlled for (a) parental civic knowledge in the
Fig. 1. Volunteering rates estimated from the logistic regression results
for Study 1 (1999 data; top panel) and Study 2 (1996 data; bottom panel)
for neighborhoods of varying levels of poverty and child saturation. To
estimate the means, all predictors in the equation, except for poverty
rate and percentage of children, were set to 0 (the mean for the contin-
uous variables). Sample weights were used, so that the graphs depict
estimated means for the national population (see the text).
594
Volume 15—Number 9
Youth Bulges

Page 5

prediction of adolescents’ civic knowledge and (b) familial volun-
teering in the prediction of adolescents’ volunteer community service.
Method
Participants
Participants were drawn from the sample of the NHES:96. This sample
is independent of the NHES:99, although it was selected in a similar
fashion (for details, see Collins et al., 1997). We used the same in-
clusion criteria concerning age and family status as in Study 1, ob-
taining a sample of 5,776 for the community-service questions and
2,506 for the civic-knowledge and tolerance questions.
Survey Questions
Two sets of five questions were used to assess civic knowledge. In each
household, one set was randomly assigned to the participant, and the
participant’s parent received the other set. The number of correct
responses for each set was standardized for participants and used as
the summary measure of civic knowledge. A summary index for pa-
rental civic knowledge was constructed similarly.
Participants’ voluntary community service and tolerance was
measured in the same fashion as in Study 1. Parents were asked if any
adult in the household was involved in voluntary community service,
with affirmative responses coded as ‘‘1’’ and negative responses coded
as ‘‘0.’’ Parental tolerance was measured with the same questions
posed to participants.
Neighborhood-, Family-, and Individual-Level Predictors
Neighborhood childhood-saturation quotients and poverty rates were
calculated as in Study 1. The family- and individual-level predictors
used in Study 1 were used in this study as well.
Results and Discussion
Scores for predictors were centered, and the analyses were conducted
using weights to correct for biases introduced by the sampling design.
Table 2 presents the results from the regression of civic knowledge,
volunteering, and tolerance on the predictors used in Study 1. Gen-
erally, the results in Table 2 replicate the findings of Study 1.
Civic Knowledge
Age, gender, school grades, parental educational attainment, and,
most important, child saturation had the same associations with ad-
olescents’ civic knowledge as observed in Study 1.
The association of child saturation with individual differences in
civic knowledge in adolescence is not explained by the mediating
influence of parental civic knowledge. When parental civic knowledge
was controlled, the magnitude of the relation between child saturation
and civic knowledge was approximately the same as it was in Study 1,
in which there was no control for parental civic knowledge.
TABLE 2
Regression of Adolescent Civic Qualities Measured in 1996 on Individual, Family, and Neighborhood Predictors
Predictor
Civic knowledge VolunteeringTolerance
B SE
b
B SE
Exp(B)
B SE
Exp(B)
(Constant)
Individual
Black
Hispanic
Other non-White
Age
Gender (males50, females51)
School grades
Extracurricular club or team (yes51, 05no)
School club or team (yes51, 05no)
Family
Parent civic knowledge
Family volunteering
Parental tolerance
Parental educational attainment
Income
Two-parent vs. one-parent home
Father-only home vs. mother-only home
One parent speaks a non-English language
Two parents speak a non-English language
Neighborhood
Percentage families in poverty
Child saturation
Interaction of percentage in poverty and
child saturation
0.145 0.059
?1.74 0.14
?0.240.20
?0.152
?0.103
0.003
0.131
?0.348
0.264
—
—
0.080
0.090
0.109
0.030
0.048
0.028
—
—
?0.05
?0.03
0.00
0.11n
?0.18n
0.25n
—
—
?0.42
?0.15
0.13
0.00
0.12
0.18
0.80
0.36
0.15
0.16
0.19
0.03
0.08
0.05
0.10
0.10
0.66n
0.86
1.14
1.00
1.13
1.20n
2.22n
1.43n
?0.21
0.01
?0.16
0.14
?0.10
0.00
—
—
0.20
0.22
0.27
0.07
0.12
0.07
—
—
0.81
1.01
0.86
1.15
0.91
1.00
—
—
0.194
—
—
0.061
?0.005
0.080
?0.004
0.155
?0.152
0.030
—
—
0.011
0.011
0.080
0.145
0.204
0.117
0.20n
—
—
0.17n
?0.02
0.04
0.00
0.02
?0.04
—
0.27
—
0.05
?0.04
0.16
?0.20
?0.51
?0.80
—
0.09
—
0.02
0.02
0.14
0.25
0.39
0.24
——
—
0.47
0.08
0.00
?0.23
?0.33
?0.42
?0.29
—
—
0.12
0.03
0.03
0.20
0.36
0.57
0.29
—
—
1.31n
—
1.05n
0.96
1.17
0.82
0.60
0.45n
1.60n
1.08n
1.00
0.79
0.72
0.66
0.75
?0.214
?0.014
0.332
0.005
?0.02
?0.07n
0.12
0.02
0.56
0.01
1.12
1.02
?2.25
0.00
0.82
0.01
0.11n
1.00
0.015 0.0430.01
?0.220.08 0.80n
0.05 0.111.05
np < .05.
Volume 15—Number 9
595
D. Hart et al.

Page 6

Volunteering
The mediating effect of familial volunteering on the relation of ado-
lescents’ volunteering to neighborhood poverty and child saturation
was examined with logistic regression. Controlling for familial com-
munity service did not eliminate the association of adolescents’ vol-
untary community service with neighborhood poverty and child
saturation. As in Study 1, there was an interaction between child
saturation and poverty rate in predicting volunteering. This interac-
tion is illustrated in the bottom panel of Figure 1 and resembles the
interaction in Study 1.
Tolerance
Neighborhood poverty, but not neighborhood child saturation, was
negatively associated with tolerance in adolescence.
Discussion
The results of Studies 1 and 2, based on two independent, nationally
representative samples of Americans, converge on three conclusions.
First, child saturation is inversely associated with civic knowledge.
Adolescents and children living in child-saturated communities know
less about their government than do those living in adult-saturated
communities. Second, child saturation is positively associated with
volunteering, although this effect is qualified by poverty. In low-
poverty neighborhoods, children and adolescents are particularly
likely to volunteer to benefit their communities if their neighborhoods
are also child saturated. However, deeply impoverished child-satu-
rated communities have very low rates of volunteering. Third, a com-
munity’s child saturation affects different facets of adolescents’ civic
development than neighborhood poverty does, and the pattern of dif-
ferences is consistent with the social-influence theoretical framework.
STUDY 3
The goal of Study 3 was to extend the analysis of child saturation to
country-level data. Political science and historical research on youth
bulges focuses on differences between countries or generations within
the same country, not on the individual- and community-level data
used in Studies 1 and 2. Consequently, connecting our findings to
other theoretical traditions focusing on youth bulges required an
analysis of child saturation and country-level data. In Study 3, we
examined the relation of country-level demographic information to
national averages for adolescents’ civic knowledge and volunteering.
Method
The units of analysis were country-level indices of civic knowledge
and volunteering from the IEA Civic Education Study (Torney-Purta,
Lehmann, Oswald, & Schulz, 2001). The IEA Civic Education Study
developed measures of civic knowledge and engagement that could be
used across democratic countries. In 1999, representative samples of
14-year-olds in 28 countries (Australia, Belgium, Bulgaria, Chile,
Colombia, Cyprus, Czech Republic, Denmark, England, Estonia,
Finland, Germany, Greece, Hong Kong, Hungary, Italy, Latvia,
Lithuania, Norway, Poland, Portugal, Romania, Russian Federation,
Slovak Republic, Slovenia, Sweden, Switzerland, and the United
States) responded to the measures. The summary report for the study
(Torney-Purta et al., 2001) provided an average for civic knowledge
(overall M5100, SD56.3) for each country’s sample of 14-year-olds,
and also gave the percentage who volunteered for an activity to help
the community (M518.1, SD511.8). We used United Nations data
(Fukuda-Parr, 2002) to determine the percentage of the population
under the age of 15 for each country (M518.8, SD53.9), and the
Central Intelligence Agency’s (n.d.) World Factbook 2002 yielded
each country’s per capita gross domestic product (GDP; in U.S. dol-
lars, M517,960, SD59,034).
Results and Discussion
Table 3 presents the results of regressing countries’ averages for civic
knowledge and volunteering rate on the percentage of the population
younger than 15 and GDP (the interaction of the percentage of a
country’s population under 15 and GDP was not significant in either
equation). The fraction of the population under age 15 was negatively
associated with civic knowledge and positively associated with vol-
unteering rates. For example, a country with 40% children could be
predicted to be 1.8 SD lower in civic knowledge and have a volun-
teering rate approximately 3 SD higher than an equally affluent
country with 20% children. Although there were no countries in the
sample in which 40% of the population consisted of children, there
are more than 40 countries worldwide with fractions this large or
larger (Fukuda-Parr, 2002; Yemen is the highest, with 50.1%). Future
research is needed to determine whether these countries are charac-
terized by the deficits of civic knowledge and high levels of volun-
teering predicted by the regression equation. Poor countries may lack
the institutions to channel the tidal wave of youth activism in child-
saturated societies into constructive activities like volunteering, and
consequently the amplifying effects of child saturation on youth par-
ticipation may be expressed in other, potentially destructive, forms
of political activity.
GENERAL DISCUSSION
Child saturation, measured at the levels of community and country, is
consistently predictive of civic knowledge and volunteering in ado-
lescence. In Studies 1 and 2, community child saturation predicted
individual levels of civic knowledge and voluntary community service.
The associations of child saturation with civic knowledge and vol-
untary community service were statistically significant and practically
important in two separate samples, even when the analyses controlled
TABLE 3
Prediction of Country-Level Civic Knowledge and Volunteering
From the Percentage of the Population 0 Through 15 Years of
Age and Per Capita Gross Domestic Product (GDP)
Predictor
B SE
b
Civic knowledge
107.07
?0.585
Intercept
Percentage population 0–15
Per capita GDP
6.23
0.280
0.0001
?0.36n
0.32w
0.0002
Volunteering
?25.447Intercept
Percentage population 0–15
Per capita GDP
10.283
0.46
0.0001
1.693
0.001
0.56n
0.50n
wp < .10.np < .05.
596
Volume 15—Number 9
Youth Bulges

Page 7

for a variety of individual- and family-level factors. Moreover, the
relation of child saturation to civic knowledge and volunteering was
found using aggregated data from 28 countries on five continents.
Together, the findings suggest that the age structure of communities
and of countries influences the civic development of youth.
The claim that social influence provides the theoretical explanation
for this pattern of findings is strengthened by the analyses in Studies 1
and 2 that demonstrate that other factors cannot account for the
pattern. We showed that the effects of community child saturation are
independent of individual and parental influences, and that the re-
lation of child saturation to civic development differs theoretically and
empirically from the pattern of the relation between neighborhood
poverty and civic development. Moreover, social influence explains
the positive contribution of child saturation (an increase in volun-
teering) on civic development, a finding not easily deduced from other
theories. For example, theoretical accounts of youth bulges that focus
on the poor economic prospects of members of youth bulges (e.g.,
Huntington, 1996) and the predilection of young men in youth bulges
to violence (e.g., Mesquida & Wiener, 1999) cannot readily accom-
modate our findings concerning the benefits of child saturation for
volunteering.
It remains for future research to demonstrate that the relative fre-
quency of contacts with adults versus contacts with children mediates
the relation of child saturation to civic knowledge and volunteering.
At another level, the synthesis of the findings reported in this article
with phenomena studied by historians and political scientists would
benefit from research demonstrating that members of youth bulges in
revolutions were both deficient in civic knowledge and civically active
as adolescents, the combination posited in this article. Nonetheless,
the three studies presented here suggest that the child saturation of a
community––characteristic of neighborhoods previously not explicitly
investigated by psychologists––influences the socialization of youth
residing in that community.
Acknowledgments––We are grateful for the support of the W.T. Grant
Foundation.
REFERENCES
Acevedo-Garcia, D. (2001). Zip code-level risk factors for tuberculosis:
Neighborhood environment and residential segregation in New Jersey,
1985-1992. American Journal of Public Health, 91, 734–741.
Central Intelligence Agency. (n.d.). The world factbook 2002. Retrieved Jan-
uary 15, 2003, from http://www.cia.gov/cia/publications/factbook/
Collins, M.A., Brick, J.M., Nolin, M., Vaden-Kiernan, N., Gilmore, S., Chan-
dler, K., & Chapman, C. (1997). National Household Education Survey of
1996: Data file user’s manual, Volume I (NCES Publication No. 97-425).
Washington, DC: U.S. Department of Education, National Center for
Education Statistics.
Delli Carpini, M.X., & Keeter, S. (1996). What Americans know about politics
and why it matters. New Haven, CT: Yale University Press.
Fukuda-Parr, S. (2002). Human development report, 2002. Retrieved February
15, 2003, from the United Nations Development Programme Web site:
http://www.undp.org/hdr2002/complete.pdf
Galston, W.A. (2001). Political knowledge, political engagement, and civic
education. Annual Review of Political Science, 4, 217–234.
Geolytics. (1999). CensusCD1Maps (Version 2.5). East Brunswick, NJ: Author.
Goldstone, J.A. (2001). Toward a fourth generation of revolutionary theory.
Annual Review of Political Science, 4, 139–187.
Goldstone, J.A. (2002). Population and security: How demographic change can
lead to violent conflict. Journal of International Affairs, 56, 3–21.
Gruenewald, P.J., Johnson, F.W., & Treno, A.J. (2001). Outlets, drinking and
driving: A multilevel analysis of availability. Journal of Studies of Alco-
hol, 63, 460–468.
Hart, D., Atkins, R., & Ford, D. (1998). Urban America as a context for the
development of moral identity in adolescence. Journal of Social Issues,
54, 513–530.
Helgerson, J.L. (2002, April 30). The national security implications of global
demographic change: Proposed remarks to the Denver World Affairs
Council and the Better World Campaign, Denver, CO. Retrieved February
2, 2003, from the Central Intelligence Agency Web site: http://www.cia.
gov/nic/graphics/Denverspeech.pdf
Huntington, S. (1996). The clash of civilizations and the remaking of world
order. New York: Simon & Schuster.
Jennings, M.K., & Niemi, R.G. (1981). Generations and politics: A panel study
of young adults and their parents. Princeton, NJ: Princeton University
Press.
Latane ´, B., Liu, J.H., Nowak, A., Benevento, M., & Zheng, L. (1995). Distance
matters: Physical space and social impact. Personality and Social Psy-
chology Bulletin, 21, 795–805.
Lopez, M.H. (2003). Volunteering among young people. Retrieved Septem-
ber 4, 2003, from the Center for Information and Research on Civic
Learning and Engagement Web site: http://www.civicyouth.org/PopUps/
frequency%20of%20volunteering.pdf
McLellan, J.A., & Youniss, J. (2003). Two systems of youth service: Determi-
nants of voluntary and required youth community service. Journal of
Youth & Adolescence, 32, 47–58.
Mesquida, C.G., & Wiener, N.I. (1999). Male age composition and severity of
conflicts. Politics and the Life Sciences, 18, 181–189.
Moller, H. (1968). Youth as a force in the modern world. Comparative Studies
in Sociology and History, 10, 238–260.
Niemi, R.G., Hepburn, M.A., & Chapman, C. (2000). Community service by
high school students: A cure for civic ills? Political Behavior, 22, 45–69.
Nolin, M.J., Montaquila, J., Lennon, J., Kleiner, B., Kim, K., Chapman, C.,
Chandler, K., Creighton, S., & Bielick, S. (2000). National Household
Education Survey of 1999: Data file user’s manual, Volume I (NCES
Publication No. 2000–076). Washington, DC: U.S. Department of Edu-
cation, National Center for Education Statistics.
Persell, C.H., Green, A., & Gurevich, L. (2001). Civil society, economic
distress, and social tolerance. Sociological Forum, 16, 203–230.
Sullivan, J.L., & Transue, J.E. (1999). The psychological underpinnings
of democracy: A selective review of research on political tolerance,
interpersonal trust, and social capital. Annual Review of Psychology, 50,
625–650.
Torney-Purta, J., Lehmann, J., Oswald, H., & Schulz, W. (2001). Civic knowledge
and engagement at age fourteen. Retrieved January 15, 2003, from The
International Association for the Evaluation of Educational Achievement
Web site: http://www2.rz.hu-berlin.de/empir_bf/reports.html
Urdal, H. (2002, March). The devil in the demographics: How youth bulges
influence the risk of domestic armed conflict. Paper presented at the
annual meeting of the International Studies Association, New Orleans,
LA. Retrieved February 1, 2003, from http://www.prio.no/publications/
papers/YouthBulgesUrdal.pdf
Zakaria, F. (2001, October 15). The politics of rage: Why do they hate us?
Newsweek, p. 22.
(RECEIVED 6/10/03; REVISION ACCEPTED 11/13/03)
Volume 15—Number 9
597
D. Hart et al.