Вы находитесь на странице: 1из 16

American Association for Public Opinion Research

The News Media in Children's Political Socialization


Author(s): M. Margaret Conway, Mikel L. Wyckoff, Eleanor Feldbaum and David Ahern
Source: The Public Opinion Quarterly, Vol. 45, No. 2 (Summer, 1981), pp. 164-178
Published by: Oxford University Press on behalf of the American Association for Public Opinion
Research

Stable URL: http://www.jstor.org/stable/2748789 .


Accessed: 24/09/2014 00:46
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

American Association for Public Opinion Research and Oxford University Press are collaborating with JSTOR
to digitize, preserve and extend access to The Public Opinion Quarterly.

http://www.jstor.org

This content downloaded from 202.75.66.124 on Wed, 24 Sep 2014 00:46:57 AM


All use subject to JSTOR Terms and Conditions

The News Media in


Children's Political Socialization
M. MARGARETCONWAY, MIKEL L. WYCKOFF,
ELEANOR FELDBAUM, AND DAVID AHERN
previous research on the development of children's political orientations has emphasized interpersonal sources of influence
such as families, peer groups, and schools. Central to this literature
are the active learningconcepts of motivation, practice, and achievement. But children also learn throughpassive experiences, which are
"typicallyeffortless, responsive to animated stimuli amenable to artificial aid to relaxation, and characterizedby an absence of resistance
to what is learned, thus opening up possibilities that, dependingupon
one's view, one may welcome or deplore" (Krugman and Hartley,
1970:184).Many of these passive experiences are providedto children
MUCH

Abstract The researchreportedexaminesthe impactof the news mediaas an agent of


political socializationon children'spolitical knowledge, their attitudestoward political
parties and electoral systems, and their activities relative to politics. Children'snews
mediause and level of politicalknowledgefunctionas primarysources of causal effects
for each other. They are consistently more importantin their effects on the attitudes
and types of participationstudied in this researchthan are the parentaland educational
system socializationindicatorsor the structuralvariables of grade and gender.
M. MargaretConway is an Associate Professorof Governmentand Politics, University of Maryland.Mikel L. Wyckoffis an Associate Study Directorwith MarketFacts,
Inc.-Washington. Eleanor Feldbaumis a Senior Scientist with The Analytic Sciences
Corporation.David Ahern is an Assistant Professor of Political Science, University
of Dayton. The authors wish to express their appreciationto the school system
which cooperated in collecting the data, and to acknowledge the support for data
analysis providedby the ComputerScience Centerand the Departmentof Government
and Politics, University of Maryland,the assistance of Anne Marso, MarkWeinberg,
and GlennGardnerin collectingthe data, and the help of MaryLynn Tischerand Joyce
Kaufman in coding the data. The contributionsof our research colleague Marsha
Grayson in the design of the research and the data collection are acknowledgedand
appreciated.The GeneralResearch Board of the GraduateSchool of the University of
Marylandprovidedresearchsupportto the senior author. The comments of an anonymous reviewerwere very helpfulin revisingan earlierdraftand are gratefullyacknowledged.
Public Opinion Quarterly Vol. 45:164-178
Published by Elsevier North-Holland, Inc.

1981 by The Trustees of Columbia University


0033-362X81l/0045-164/$2.50

This content downloaded from 202.75.66.124 on Wed, 24 Sep 2014 00:46:57 AM


All use subject to JSTOR Terms and Conditions

NEWS MEDIA N CHLDREN'S SOCIALIZATION

165

through the mass media. Childrenlearn when they read newspapers


or magzines, when they watch television or films, or when they listen
to the radio just as surely as when they read textbooks or listen to
lectures. And they learn not only informationbut also attitudes and
opinions.

While considerable research has examined the influence of the


media, and especially television violence, in stimulating aggressive
behavior in children, the impact of news media use on children's
political information,political participation,and attitudes toward the
political regime has received less attention. Such relationshipsare the
concern of the research reported here.
Researchers disagree about the importance of the mass media in
political socialization. In their earlierworks Klapper(196O}and Gerbner (1960) treat the media as secondary socializing agents which
reinforceexisting views throughselective exposure. Others argue that
children have few political views to be reinforced; hence, direct
learning of political attitudes and beliefs can occur (Chaffee et al.,
1970).
A numberof studieshave foundthat the news mediaare an important
informationsource for children (Alper and Leidy, 1970; Atkin, 1977;
Atkin et al., 1976; Byrne, 1969; Connell, 1971; Conway et al., 1975;
Hawkins et al., 1975; Hirsch, 1971; Hollander, 1971; Rubin, 1976;
Tolley, 1973; Torney et al., 1975). Research on the impact of news
media use on children's political attitudes has been more limited.
However, while the evidence is less plentiful, several studies have
suggested that the news media can influence affective as well as
cognitive orientations (Conway et al., 1975; Chaffee et al., 1970;
Jackson-Beeck and Chaffee, 1975; Kraus and Davis, 1976).
Certain generalizations concerning children's mass media consumption patterns have been established, especially the matter of
gender differences in viewing and reading. For example, girls like
music and nonconflictualthemes, whereas boys prefer adventureand
action fantasy content (Roberts, 1973).Girls are more likely than boys
to acquire informationthroughincidentallearning(Hale et al., 1968),
and they seek out the reality content of the mass media at an earlier
age than do boys (Schramm et al., 1959). However, boys report
viewing more news than do girls (Atkin, 1978; Rubin, 1976).
Age or grade differences also exist in media use. Since radio and
television are usually more accessible and understandable,children
generally use the electronic media first. Use of print media increases
steadily as children progress throughthe early school years (Roberts,
1973). Olderchildrenalso reportmore frequentviewing of televisions
public affairs content (Atkin, 1978; Rubin, 1978).

This content downloaded from 202.75.66.124 on Wed, 24 Sep 2014 00:46:57 AM


All use subject to JSTOR Terms and Conditions

166

CONWAY, WYCKOFF, FELDBAUM, AND AHERN

Older studies suggest that parents have a direct impact on their


children's media use patterns (Himmelweit et al., 1958; Schrammet
al., 1961). More recent analyses have questioned the validity of this
direct effects model. One alternative is that of reverse modeling, in
which children influence parentaltelevision viewing patterns (Clarke,
1963). However, children generally report spending more time
watching television than do parents (Chaffee et al., 1971). An opportunity explanationfor children'snews viewing, for example, children
watch television news because that is what their parentsare watching,
also cannot be discounted, although some research reports low correlations between parentaltelevision news viewing and the children's
level of news viewing (Chaffee et al., 1971; Atkin, 1978). Furthermore, patterns of television news viewing by children may be influenced by family communicationspatterns (Chaffee et al., 1971).
Those children who attend more to the news media may be more
cognizant of the political system's strengths and weaknesses. News
programsprovide extensive coverage of the electoral system and the
role of political parties in nominating and electing public officials.
They also discuss goals and outcomes of public policy programs,
accomplishmentsand misdeeds of public officials, levels of government expenditures, and on occasion the misuse of public funds. Thus,
exposure to attitude-formingrole models and to information about
complex political institutions and processes may result in the development of favorableattitudestoward some aspects of governmentand
less favorable attitudes toward others.
Methods,Models, and Data
Research dealing with the role of the news media in children's
political socialization suggests a developmental model in which certain antecedent variables influence news media use and political
knowledge, with both antecedent and interveningvariables having an
impact on additionalpolitical attitudes and behaviors. Therefore, the
analysis proceeds in two stages. First, linkages among four antecedent variables, children'snews media use, and political knowledge are
examined. Then, relationshipsbetween elements of the initial model
and children's party system support, electoral system support, and
political participationare examined.
The analysis is based on data collected in seven elementaryschools
located in a Middle Atlantic metropolitanarea. A paper and pencil
questionnairewas administeredto 760 fifth and sixth grade students
early in 1973. The schools were located in three distinct types of
socioeconomic neighborhoods:two were in workingclass areas, three

This content downloaded from 202.75.66.124 on Wed, 24 Sep 2014 00:46:57 AM


All use subject to JSTOR Terms and Conditions

NEWS MEDIA IN CHILDREN'S SOCIALIZATION

167

in middle-classareas, and two in upper-middle-classareas. Classes in


five schools used election simulations, while classes in the two remainingschools did not.1 Since the type of civics instructionto which
children are exposed may have considerable impact on their political
knowledge, attitudes, and participationpatterns, an indicator of instructionalmethod is included in the analysis.
An initial model for analysis is presented in Figure 1. Four antecedent variables have been specified: parentalnews media use (xl); sex
(x2); grade in school(x3);and type of instructionalmethod used in the
classroom (X4). Parentalnews media use is measuredusing a summed
score which reflects use of both print and electronic news media.
Two causally prior variables which previous research suggests
might be significant are grade in school and sex. The subject matter
considered in the fifth and sixth grade social studies curriculums
differs substantiallybetween the grades in all the schools studied, but
by and large, sixth grade students are expected to be more knowledgeable than fifth graders. The research reviewed also suggests that
boys are more likely to watch television news, but girls are more
likely to acquire informationfrom the media throughpassive learning
processes.
A fourth antecedent variable, instructionalmethod, is coded from
zero to four, with zero indicatingthat the traditionalteaching method
Parents' News
MediaUse
(_
\

.2

GRADE

P51

51

Y. Child'sNews
Use

P_
<Media

SEX

25~~~~
c,

Instructional

Method

Knowledge

P
X4

In64

ndex

Figure 1. Initial Model to be Estimated


Equations:
1. Y5 = P51X1 + P52X2 + P53X3 + P56Y6 + e5
2. Y6 = P62X2 + P63X3 + P64X4 + P65Y6 + e6
lThere is no relationship between use of simulation or the elaborateness of a
simulationand the social class characteristicsof the schools' clienteles; for example,
the most elaborate simulationwas conducted in a working class area school.

This content downloaded from 202.75.66.124 on Wed, 24 Sep 2014 00:46:57 AM


All use subject to JSTOR Terms and Conditions

168

CONWAY, WYCKOFF, FELDBAUM, AND AHERN

of lecture, discussion, and readingwas used. A non-zero score means


a simulationwas used to instruct children in Americangovernmental
processes, with a score of one assigned to the most simple simulation
(discussion and voting) and higher scores assigned to more complex
simulations. A score of four indicates a very complex simulation,
involving use of voter registrationsystems, presidentialprimariesand
nominatingconventions, debates on issues, campaigns, and an election.
The knowledge index is a summed score of children's correct responses to 17 questions about political institutions, includingpolitical
parties and the electoral process, with correct answers to these questions being assigned a score of one and incorrect answers scored
zero.2 Child's news media use is measuredby a summed score based
on the child's answers to three questions about the frequency of
consuming three different news media: television news, political and
public affairs news in newspapers, and news magazines such as Time
or Newsweek.3
Does media use precede knowledge, or does knowledge precede
2 The questions focus on the functions and processes of the three branches of
government, the presidentialand general electoral systems, and the political party
system. The index scores range from zero to 14, and the mean score for the total
sample is 5.8.
3 It is theoreticallyreasonableto combinechildren'selectronic and printmediause in
a single indicatorof news media consumptionor, alternatively,to separatethem into
distinctelectronic and printmediause variables.Both approachesare supportedby the
outcomes of the following principalcomponents analyses:
Loadings for Single
Component Solution

Loadings for Rotated


2-Component Solution

1
.70
.79
.66

News magazinereading
Newspaperreading
Television news watching
h2 =

1
.91
.66
.09

52.0%

2
-.01
.44
.95
72.2%

In an effortto treat electronicand printmediause as separatedimensions,we sought


to estimate the following model:
Parental X2
Print News 2
Media Use

Y Child'sPrint News
5
MediaUse

Child's TV

NewsUsey
X
Parental TV
X
News Media Use

Y Knowledge
6

InstructionalA
3
Method

Unfortunately,multi-collinearityamong system variables (a problem frequently encounteredwhen using two-stageleast squares;see Kritzer, 1976)preventedthe generation of meaningfulpath estimates.

This content downloaded from 202.75.66.124 on Wed, 24 Sep 2014 00:46:57 AM


All use subject to JSTOR Terms and Conditions

NEWS MEDI

IN CHLDEN'S

SOCILIZATION

169

media use? It can be argued that knowledge must precede media use;
in this view, one does not consume the media'snews content until one
has the capacity to understandthe meaningof what is viewed or read.
An alternative argumentholds that children consume the news content because it is available to them, and thus increased knowledge
comes from this pattern of either incidental or deliberately sought
exposure to the news.
Discarding the notion that these two positions must be mutually
exclusive, a developmental process can be suggested in which both
endogenous variables have causal effects. Therefore, rather than rejecting one of these positions on a priori grounds, we propose an
alternative nonrecursive model in which political knowledge and
media use are conceived as mutually dependent variables. In Figure
1, this situation of reciprocal influence is denoted by paths P65 and
P56.

The estimation of nonrecursive paths presents a number of methodological problems not encountered with recursive systems. First,
because a situation of reciprocal causation is specified, it cannot be
assumed that explanatory variables are uncorrelatedwith model disturbanceterms. As a result, ordinaryleast squares estimators are not
appropriate.The method of two-stage least squares provides a solution to this problem.
A more serious problem encountered with nonrecursivesystems is
the problem of identification.In order to generate unique path coefficients, certain conditions must be met. The order condition for identification requires that the numberof model paths to be estimated be
limited so that the numberof variableseffectively excluded from each
of the model's equations "be at least as great as the number of
equationsin the model less one" (Johnston, 1972:348).4This condition
is met for the model specified in Figure 1 by eliminatingcausal paths
P61 and Pm. That is, the effect exerted on child's political knowledge
(Y6) by parental news media use (X1) is assumed to be exclusively
indirect and mediated by child's news media use (Y5). Similarly,
instructional method (X4) is assumed to have no direct effect on
child's news media use (Y5).5
4 A necessary and sufficientconditionfor identification,the rank condition, is more
difficultto establish. In general, establishmentof the order condition is consideredan
adequate basis for assuming identification.
S While the validity of this specificationmust be determinedon a prioriratherthan
empiricalgrounds, some readers may wish to note the following partialcorrelations:
r16

.27,

r16.5

.14, p

- .001

.11, r45.6 = .01, n.s.


The correlationbetweenX4 and Y4 does drop out, but the relationshipbetweenX1 and
Y6remainssignificant.However, with a partialcorrelationof only .14, the amountof
variance explained attributablesolely to X1 is a mere 2 percent.
r45

This content downloaded from 202.75.66.124 on Wed, 24 Sep 2014 00:46:57 AM


All use subject to JSTOR Terms and Conditions

170

CONWAY, WYCKOFF, FELDBAUM, AND AHERN

The paths specified in Figure 1 have been estimated through the


method of two-stage least squares. Paths which are insignificantat p
S .05 have been discarded and the remainingpaths reestimated. The
outcomes of this procedure are presented in Figure 2.
ResearchResults
Figure 2 presents the initial model with insignificant paths eliminated. The relationshipsbetween parent'sand child's news media use
and between instructionalmethod and level of political knowledge are
positive and significant.Grade in school is marginallyrelated to level
of political knowledge, but unrelated to news media use. Gender is
not significantlyrelated to either news media use or level of political
knowledge.
Most important, the hypothesized nonrecursive relationship between child's news media use and level of political knowledge is
confirmed. The influence of child's news media use on political
knowledge is slightly larger than the effect of knowledge on news
media use.
To test further the conclusion that differences in gender have no
significantimpact on relationshipswithin the system, the estimation
of unstandardizedcoefficients has been done separately for boys and
girls.6The unstandardizedpath coefficients for the two subgroupsare
quite similar, as Table 1 indicates, except for the path from level of
Parents' News
Media Use
MediaUse

.2
.25

VYChild's News
j Media Use

~~~~~~~~~~~~.62
.46Ine

Method

Knowledge

Figure 2. Estimated Model with Insignificant Paths Eliminated


(Standardized Coefficients)
6 Unstandardizedpath coefficients must be used for this comparison;if standardized
path coefficients were compared,the differencesin magnitudesof the path coefficients
might be due to differences in sample variances rather than differences in effects
(Schoenberg, 1972).

This content downloaded from 202.75.66.124 on Wed, 24 Sep 2014 00:46:57 AM


All use subject to JSTOR Terms and Conditions

NEWS MEDIA IN CHILDREN'S SOCIALIZATION

171

Table 1. Estimated Path Coefficientsa (Unstandardized)

P51
P64

P63
P65
P56
a

Total Sample

Boys

Girls

.23
.31
.64
.69
.41

.22
.32
.61
.70
.46

.27
.31
.69
.63
.27 (n.s.)

Significantat p =.05 or less, unless otherwise indicatedby n.s.

knowledgeto news media use (P56); for girls this is small and insignificant. Thus, knowledge may be a more importantinfluence on news
media use for boys than for girls.
Until now we have examined only the process involved in the
developmentof predispositionstoward news media use and the acquisition of political knowledge. The second stage of our analysis treats
these variables as interveningfactors in the development of electoral
system support, political party support, and political participation.7
The four independent variables of parent's media use score, instructional method used in the classroom, grade, and sex are again
included in the analysis.
All possible recursive paths between the six variables in the initial
model and the three additional dependent variables were estimated
initially. Paths which are insignificantat p S .05 have been discarded
and the remainingpaths re-estimated.In order to accomplishthis task
while adheringto the order condition for identification,the new models are treated as bloc-recursive systems.8
The direct, indirect, and total effects of antecedent and intervening
variables on electoral system support, party system support, and
politicalparticipationare presented in Table 2 and Figures 3, 4, and 5.
Direct effects are simply the standardizedcoefficients associated with
each independent-dependentvariable path. Indirect and total effects
are computed using the method explicated by Lewis-Beck and Mohr
(1976).
The data in Table 2 and Figure 3 indicate that political knowledge
has a much greater direct impact on electoral system support than
does news media use. However, because the indirect effects of news
7 A political participationindex was created from the students' responses to
questions about participationin various kinds of political activity. Questions used to measure party system supportand electoral system supportare based on items reportedin
other studies. The two attitudinalscales used as dependentvariableswere developed
throughprincipalcomponentsanalyses with varimaxrotation.(See the Appendixfor a
list of items used in the party system supportand electoral system supportscales and
the political participationindex.)
8 For a discussion of bloc-recursivemodels see Duncan
(1975:85).

This content downloaded from 202.75.66.124 on Wed, 24 Sep 2014 00:46:57 AM


All use subject to JSTOR Terms and Conditions

172

CONWAY, WYCKOFF, FELDBAUM, AND AHERN

Table 2. Summary of Effects on Three Dependent Variables


Parent's
Media
Use
Electoral system support
Direct
Indirect
Total
Party system support
Direct
Indirect
Total
Political participation
Direct
Indirect
Total

.12
.12

Instructional
Grade Method Sex

.06
.06
-.09
-.04
-.13

-.05
-.05
.08
.16
.24

.09
.06
.15

.11
.10
.21
-.09
-.06
-.15

Child's
Media
Use Knowledge R2

.20

.14
.33
.47

.20
-.12
-.12

.31
.21
.52

-.22
-.22

-.25
-.10
-.35

.33
.30
.63

.19
.09
.28

.20
.29
.49

.20

.10

.31

media use are substantial, the total effects associated with these
intervening variables are very similar and they are clearly the most
important explanatory variables in the analysis. The total effects
attributableto method of instructionare considerablysmaller and are
equally the result of-irect and indirect influence. On the other hand,
the effects of gender on electoral system support are entirely direct.
This suggests that boys tend to evaluate electoral politics more favorably than girls regardlessof news media use patternsand levels of
political knowledge. Finally, parental news media use and grade in
school have only weak and indirect effects on electoral system support.
Turningto the analysis of party system support (Figure 4), a quite
different pattern is found. First, the signs of all paths leading to the
dependent variable are negative. Child's media use has only an indi-

Parents'News
MediaUse

.ir
0

.C

3.
SEX

ChildIsNews
MediaUse

Sts

Insrcinl

RD
GRADE

56 .62 .46
Knwldg
31

Insrut2oalR
Method
-Y4

Y6

P
2
InexY7 ELECTORAL

~SYSTEM
SUPPORT
2
.20

Index
~~~~~~Knowledge

Figure 3. Electoral System Support Model (Standardized Coefficients)

This content downloaded from 202.75.66.124 on Wed, 24 Sep 2014 00:46:57 AM


All use subject to JSTOR Terms and Conditions

NEWS MEDIA IN CHILDREN'S SOCIALIZATION

173
e95

Child'sNews
MediaUse

Parents'News
MediaUse
1

SE

~~~~5

A2

6.4

~~~~
le7
195

U'~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~7
09
PARTY
o
~~~~~~~~Y7
21
.2<GADE
SEUPPORT
0~~~~~~~~~~~~~~~~~~~R
oXX
_Y

SYSTEM
1

KnowledgeIndex
e06

Figure 4. Party System Support Model (Standardized Coefficients)


4t(Method

rect effect on party system support, but both grade and sex have a
small direct effect. Once again, the knowledge variablehas the largest
direct effect, and children who are more knowledgeable about the
political party system are more negative about it. Instructional
method effects, both direct and indirect, are slight but negative. Parental news media use has only an indirecteffect throughchild's news
media use and level of political knowledge.
That the signs from the knowledge and instructionalmethods variables to party system support are negative is not surprising.Political
parties receive low levels of support from adults (Feigert and Conway, 1976),civics instructionin the schools tends to emphasize voting
for the person and not for the party (Hess and Torney, 1967; Merelman, 1971), and news media content tends to present a negative view
of the political party system (Hofstetter, 1976). The results also indi,

Child'sNews
MediaUse

Parents'News
MediaUse

Xs

e7

o
c

.09

kGRADE

.X
cn

.83

46

.? /SEX

.31
~~~~~~~R

.62

Instruction ai
Method
)'4

Y POLITICAL
PARTICIPATION

6 KnowledgeIndex

F6

Figure 5. Political Participation Model (Standardized CoefficientS)

This content downloaded from 202.75.66.124 on Wed, 24 Sep 2014 00:46:57 AM


All use subject to JSTOR Terms and Conditions

174

CONWAY, WYCKOFF, FELDBAUM, AND AHERN

cate that girls are less supportiveof the political parties and that sixth
grade children are less supportive than fifth graders.
In the political participationanalysis (Figure 5 and Table 2), yet
another pattern is found. In contrast to the previous two models,
gender is not a significantvariable. However, all other variables have
both direct and indirect effects. Child's media use has the largest
impact through both direct and indirect effects, with level of knowledge also having substantialtotal effects. Comparedto instructional
method and parentalnews media use, teaching method has a stronger
direct effect on children's political participation.

Conclusionsand Discussion
To explore the potential impact of the news media vis-a-vis other
agents of socialization on preadolescents' political orientations, a
system of six variables thought to contribute to such orientations,
includingindicatorsof mass media, school, and family influences, has
been analyzed using causal modeling techniques. Grade in school and
instructional method have statistically significant, but moderate to
marginal effects on children's political knowledge, while parental
news media use has moderate and statistically significant effects on
children'snews media use. On the other hand, gender differences are
not found to have significant additive effects on either political
knowledge or news media use.
More important, the analysis indicates that children's news media
use and level of political knowledge function as primary sources of
causal effects for one another. However, upon dividingthe sample on
the basis of sex and reestimating the model for the resulting subsamples, knowledge appears to be a more important influence on
news media use for boys than for girls.
A plausible interpretationof these findings is that knowledge of
politics and attention to political affairs through news media use are
related in a developmental process which may be initially stimulated
or partially reinforced by the primary aspects of socialization but
which also seems capable of independentlybuilding and maintaining
its own momentumonce engaged. In the absence of longitudinaldata
this interpretationaccounts for our results in a theoretically reasonable fashion.
Furthermore,our findings indicate that children's news media use
and level of knowledge about American political processes are consistently more importantin determininglevel of political participation,
supportfor the political parties, and supportfor the electoral process

This content downloaded from 202.75.66.124 on Wed, 24 Sep 2014 00:46:57 AM


All use subject to JSTOR Terms and Conditions

NEWS MEDIA IN CHILDREN'S SOCIALIZATION

175

than are instructional method, parental news media use, or the


structural variables of grade and gender. In general, only marginal
effects on the dependent yariables can be credited to grade in school
and to gender differences. An exception occurs with electoral system
support, where gender differences are moderatelyassociated with the
dependent variable despite the fact that male-female differences in
levels of political knowledge and news media use are not discernible.
Among the socialization agent variables, parental news media use
displays largely indirect effects on the attitudeand behavior variables
and emerges as a relatively importantinfluence-only with regard to
politicalparticipation.Instructionalmethod operates both directly and
indirectly to produce moderate effects on all three dependent variables. The use of increasinglyelaborateelection campaignsimulations
seems to be particularly influential in encouraging high levels of
political participation.
But contrasting the overall effects produced by the antecedent
variableson each of the dependentvariables, the superiorexplanatory
power of children'snews media use and political knowledge is apparent. In most instances, the total effects associated with the latter two
are nearly twice as large as those attributableto the remainingvariables. In short, the analysis indicates that news media use, alone and
in conjunctionwith knowledge of the Americanpolitical process, is a
significantdeterminantof children'spolitical attitudes and patternsof
political participation.
As a final note, the value of the relatively elaborate statistical
models used in the analysis must be emphasized. Importantcomponents of the total effects attributedto many of the variables(and most
notably child's news media use) derive indirectly from complex developmental sequences and nonrecursive relationships. Given the
likelihood that such complex relationships do exist in the "real
world,"using less elaboratemodels, and particularlythe simple linear
model which underlies basic regression analysis, would almost certainly produce misleading results.
As a case in point, had we simply regressed party system support
on the six antecedent variables, the resulting standardizedregression
coefficient derived for child's media use would be zero. Obviously,
that outcome would lead to the conclusion that news media have little
influence on children's orientations to the party system.
Alternatively, the analysis presented here suggests that the news
media function as primary sources of children's political knowledge
and informationand, as a result, exert an importantindirectinfluence
on attitudes and behaviors even though direct linkages may be weak
or nonexistent. Thus, in our analysis of children's support for the

This content downloaded from 202.75.66.124 on Wed, 24 Sep 2014 00:46:57 AM


All use subject to JSTOR Terms and Conditions

176

CONWAY,WYCKOFF,FELDBAUM,ANDAHERN

party system, news media use emerged as an important independent


variable.
In conclusion, in order to treat adequately the processes involved in
political socialization, socialization theory and research should develop models which reflect the complexities of those processes.
Based on a model which provides for developmental sequences and
nonrecursive relationships among system variables, the analyses presented here support the argument that the role of the mass media in
political socialization has not received adequate attention. While
media effects have frequently been perceived as reinforcing rather
than creating individual attitudes and beliefs, our analysis suggests that
the explanation of within- and between-generation differences in
political orientations requires greater attention to the creative effects
of observation and vicarious experience made possible through the
news media.
Appendix: Dependent Variable Scale and Index Items Used
(Agree /Disagree /Don't Know Questions)

Loadings

THE ELECTORAL SYSTEM SUPPORT SCALE

Most election campaignsare silly or ridiculous.


Many primaryelections aren'timportantenough to vote in.
If the Democrats and Republicanscan't agree on important
things, it would be bad for the country.
It isn't so importantto vote when you know your party
doesn't have a chance to win anyway.
My family has no effect upon what the governmentdoes.
So many other people vote in nationalelections that it does
not matter whether a person votes or not.

.460
.458
-.417
.641
.410
.551

THE PARTY SYSTEM SUPPORT SCALE

It is good to stick with your party throughthick and thin.


.495
A congressman should follow his party leaders even if he
doesn't want to.
.546
Lower taxes should be allowed for people who give money
to the political party of their choice.
.525
The most importantthing to consider when voting is the
political party to which each person belongs.
.478
A total of 12 componentswere extractedfrom a set of 43 attitudeitems, all
with eigenvaluesof 1.0 or more;the total varianceexplainedby the 12 components was 46 percent.
PARTICIPATION

INDEX ITEMS

1.
2.
3.
4.

Did you work at the voting places on election day?


Did you wear a button for a real national or local candidate?
Did you talk to your friends about Nixon and McGovern?
Did you help othermembersof yourfamilywith work they were doingin the
real election?
5. Did you watch any of the national political conventions on TV?

This content downloaded from 202.75.66.124 on Wed, 24 Sep 2014 00:46:57 AM


All use subject to JSTOR Terms and Conditions

NEWS MEDIA IN CHILDREN'S SOCIALIZATION

177

References
Alper, W. S., and T. R. Leidy
1969- "The impact of informationtransmissionthroughtelevision." Public
70 Opinion Quarterly33:556-62.
Atkin, Charles K.
1977 "The effects of campaign advertising and newscasts on children."
JournalismQuarterly54:503-08.
1978 "Broadcastnews programmingand the child audience." Journal of
Broadcasting22 (Winter):47-61.
Atkin, C. K., J. Galloway, and 0. B. Nayman
1976 "News media exposure, politicalknowledge, and campaigninterest."
JournalismQuarterly53:231-37.
Byrne, G. C.
1969 "Mass media and political socialization of children and pre-adults."
JournalismQuarterly47:647-59.
Chaffee, S. H., J. M. McLeod, and C. K. Atkin
1971 "Parentalinfluenceson adolescent media use." AmericanBehavioral
Scientist 14:323-40.
Chaffee, S. H., L. S. Ward, and L. P. Tipton
1970 "Mass communications and political socialization." Journalism
Quarterly47:647-59, 666.
Clarke, P.
1963 "An experimentto increase the audience for educationaltelevision."
Ph.D. dissertation, University of Minnesota.
Connell, R. W.
1971 The Child'sConstructionof Politics. Melbourne:MelbourneUniversity Press.
Conway, M. M., A. J. Stevens, and R. G. Smith
1975 "The relation between children's media use and children's civic
awareness."JournalismQuarterly52:531-38.
Dawson, R. E., and K. Prewitt
1969 Political Socialization. Boston: Little, Brown.
Duncan, Otis Dudley
1975 Introductionto StructuralEquation Models. New York: Academic
Press.
Easton, D.
1965 A Systems Analysis of Political Life. New York: Wiley.
Feigert, F. B., and M. M. Conway
1976 Parties and Politics in America. Boston: Allyn and Bacon.
Gerbner, George
1960 "Mass communicationsand the citizenshipof secondary school children."Pp. 179-204 in F. K. Patterson(ed.), The Adolescent Citizen.
New York: Free Press.
Hale, G. A., L. K. Miller, and H. W. Stevenson
1968 "Incidentallearningof film content." Child Development 39:69-77.
Hawkins, R. P., S. Pingree, and D. F. Roberts
1975 "Watergateand political socialization."AmericanPolitics Quarterly
3:406-22.
Hess, R. D., and J. Torney
1967 The Developmentof PoliticalAttitudesin Children.Chicago:Aldine.

This content downloaded from 202.75.66.124 on Wed, 24 Sep 2014 00:46:57 AM


All use subject to JSTOR Terms and Conditions

178

CONWAY, WYCKOlF?, FELDBAUM, AND AHERN

Himmelweit, Hilda T., A. N. Oppenheim, and Pamela Vince


1958 Television and the Child. London: Oxford University Press.
Hirsch, Herbert
1971 Poverty and Politicization. New York: Free Press.
Hofstetter, R.
1971 Bias in the News. Columbus:Ohio State University Press.
Hollander, N.
1971 "Adolescents and the war: the sources of socialization."Journalism
Quarterly48:472-79.
Jackson-Beeck, M., and S. Chaffee
1975 "Family communication, mass communication, and differential
political socialization."Paper presented at the InternationalCommunicationAssociation, Chicago, April.
Johnston, J.
1972 Econometric Methods (2nd ed.). New York: McGraw-Hill.
Klapper, Joseph T.
1960 The Effects of Mass Communications.New York: Free Press.
Kritzer, Herbert M.
1976 "Problemsin the use of two-stage least squares: standardizationof
coefficients and multicollinearity."Political Methodology 3:71-97.
Krugman,Herbert E., and Eugene L. Hartley
1970 "Passive learning from television." Public Opinion Quarterly
34:184-90.
Lewis-Beck, M., and L. Mohr
1976 "Evaluatingthe effects of independentvariables."Political Methodology 3:27-48.
Merelman, R. M.
1971 Political Socialization and EducationalClimates. New York: Holt,
Rinehart, and Winston.
Prisuta, Robert H.
1979 "The adolescent and television news: a viewer profile." Journalism
Quarterly56:277-82.
Roberts, Donald F.
1973 "Communicationand children: a developmental approach." Pp.
174-215 in I. Pool et al. (eds.), Handbook of Communication.
Chicago: Rand McNally.
Rubin, Alan M.
1976 "Television in children's political socialization."Journal of Broadcasting 20 (Winter):51-60.
Schoenberg, R.
1972 "Strategies for meaningfulcomparisons."Pp. 1-35 in Sociological
Methodology 1972. San Francisco: Jossey-Bass.
Schramm, W., J. Lyle, and B. Parker
1961 Television in the Lives of Our Children.Stanford:StanfordUniversity Press.
Tolley, H., Jr.
1973 Childrenand War. New York: Teachers College Press.
Torney, J. S., A. N. Oppenheim, and R. F. Farnen
1975 Civic Education in Ten Countries. New York: Halstead Press.

This content downloaded from 202.75.66.124 on Wed, 24 Sep 2014 00:46:57 AM


All use subject to JSTOR Terms and Conditions

Вам также может понравиться