V. 1%}

7!! . I

I
15-h: .rr
... W4. .

».

   

V V‘u.)-._‘.II. u V
.H..uWrmu.w.w.W.w.Fn ,
Wt‘M‘fiV v . . ‘ I ,
«3.1 . . . ,, ., .
A1 . . . . .V . .
. V V , a. sh . . ,
u! any». V VA . m
. Iv 3.1.121 r1? .. .b
V , .. V .n»..rflr\h.r.h‘un.‘...u.uufi.kq 23.4mm“? ..
. fig 4. h , . . Alfie“.
, _ . V ‘ , .mxmfl: 1%..» . #492?
V . .15ka K: .. ., .
, , . we??? . a.
_ . . wwmtf $
4 fifikwfiw. V V.
~$sfi4w .
V. v.‘
ux
A

um,

.51.. l
. 1W

 

1

 

 

m? ,
u I van.»
1 .. . 53.31:.
.. .% r
. 10‘

h

.\ 2f

 

 

.11... "Rah...” 3

RI ..l \L A

1.5 fiat-Ix...”

 

THEStS

\.

' .

,5: r .. I

. . .-§
[60/515 (7”

This is to certify that the
dissertation entitled

PREDICTORS OF VISUAL MEDIA USE AND
ITS EFFECTS ON IMAGINATIVE PLAYFULNESS
IN PRESCHOOL CHILDREN

presented by
Nary Shin

has been accepted towards fulfillment
of the requirements for the

Ph. D. degree in FamilLand Child Ecologl

 

 

2// ell /

1‘44 4/ M 1 [1.71 1 ll ’
Major Prof :‘ ~ . r’s Signature

Oé/[L/fl%

Date

MSU is an Affirmative Action/Equal Opportunity Institution

' v ‘ ‘v

 

.UBRARY
Michigan State
University

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

ugqu?

 

 

 

 

 

 

 

 

 

 

 

 

 

 

GIOt cJClRC/DatoDuo.p65-p.15

PREDICTORS OF VISUAL MEDIA USE AND THEIR EFFECTS ON
IMAGINATIVE PLAYFULNESS IN PRESCHOOL CHILDREN

By

Nary Shin

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Department of Family and Child Ecology

2004

ABSTRACT

PREDICTORS OF VISUAL MEDIA USE AND THEIR EFFECTS ON
IMAGINATIVE PLAYFULNESS IN PRESCHOOL CHILDREN

By

Nary Shin

The main purpose of this study was an examination of the associations
among preschool children’s visual media use, the quality of their imaginative play,
and the ecological factors of their viewing practices in their family. As a part of
my research, I hypothesized that the levels of preschoolers’ viewing different
programming at home are influenced by ecological variables, which then affect
their imaginative playfulness.

The role of children’s visual media use and its ecological predictors in
imaginative playfulness was researched using Structural Equation Modeling
(SEM) that tested the pathways from ecological factors through children’s visual
media use to their imaginative playfulness. More specifically, this study
examined the relationships among socioeconomic status, the viewing
environment in a family, parental mediation, visual media use of three to five-
year-old children, and their imaginative playfulness.

Children’s visual media use and its ecological factors were collected from
202 children’s primary caregivers, and the head teacher of children in childcare
evaluated the imaginative playfulness component. It was found that
socioeconomic status contributed to the viewing environments in the family, and

in turn, the family viewing environments influenced the children’s viewing

practices. However, the children’s viewing did not affect their imaginative
playfulness. Although no systematic pattern between children’s visual media use
and their imaginative playfulness was found, the final model identified the
pathways among ecological factors, visual media use, and imaginative
playfulness.

Implications for researchers who study media use and children's
development and educators who wish to promote children’s imaginative play are

discussed.

ACKNOWLEDGEMENTS

I have dreamed of being Dr. Shin since I was seven years old. My
inspiration came from my father, Dr. Simon B. Shin. I remember seeing his
nametag on his office door or watching guests who would come to our home and
marvel at my parents’ bedroom filled with books. I wanted to be just like him.
There have been a lot of people since then who have supported my dream and
efforts in earning my doctoral degree.

I would like to express my gratitude to Dr. Whiren, my advisor and
committee chair, for her advice and support throughout my doctoral program as a
mentor, a colleague, and sometimes as a mother. I would also like to thank Dr.
Greenberg for providing basic insights and valuable guidance in the field of
telecommunications with which I was unfamiliar. I would also like to thank Dr.
Soderrnan who challenged me to be a scholar and a researcher, Dr. Luster, who
was a great listener and who guided me in a respectful and encouraging way
whenever I needed his support, and to Dr. von Eye who guided the methodology
I used with extra support throughout my work.

I am very appreciative of the parents and teachers who took the time out
of their busy lives to complete the research questionnaires. Their interest and
participation in this study resulted in my having a much deeper appreciation for
their roles.

I owe my thanks to many friends who provided me with every kind of

support, from encouraging remarks to wonderful meals. I would especially like to

thank my parents-in-law for their thoughtful understanding throughout the
dissertation process, and a special thanks and much love to my mother for her
endless prayers and to my father, the original Dr. Shin, whose special qualities
instilled in me the desire to see this through. Finally, I would like to thank my
husband, Seung-Jin, who is my best and life-long friend, and to my angel, Alexis,
who has been in daycare since she was seven-weeks-old, for the joy that they
have brought and continue to bring to my life. To all of them who have made it

possible to complete my dissertation, I extend my sincere gratitude and love.

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................... viii

LIST OF FIGURES ............................................................................................... ix
CHAPTER 1

INTRODUCTION .................................... 1

Statement of the Problem ........................................................................... 7

Theoretical Framework ............................................................................... 7

Conceptual and Operational Models ........................................................ 11

Conceptual and Operational Definitions ................................................... 15
CHAPTER 2

LITERATURE REVIEW ....................................................................................... 22

Ecological Factors of Children’s Viewing .................................................. 22

Socioeconomic Status ................................................................... 22

Viewing Environment in a Family ................................................... 24

Parental Mediation ......................................................................... 27

Child Characteristics ...................................................................... 29

Playfulness ............................................................................................... 31

Definition ........................................................................................ 31

The Studies of Playfulness ............................................................ 32

Playfulness and Development ....................................................... 36

Effects of Viewing on Imaginative Play ..................................................... 38

Effects of Viewing on the Quantity of Imaginative Play .................. 38

Effects of Viewing on the Quality of Imaginative Play .................... 42

Effects of Viewing on Playfulness ............................................................. 44
CHAPTER 3

METHODOLOGY ................................................................................................ 46

Research Hypotheses .............................................................................. 46

Procedures ............................................................................................... 48

Participants ............................................................................................... 51

Measures .................................................................................................. 54

Socioeconomic Status ................................................................... 54

Viewing Environment in a Family ................................................... 55

Parental Mediation ......................................................................... 56

Children’s Visual Media Use .......................................................... 58

Imaginative playfulness ................................................................. 73

Data Analyses .......................................................................................... 78

Preliminary Analyses ..................................................................... 78

Structural Equation Model ............................................................. 80

Regression Analyses ..................................................................... 83

Data Transformation ...................................................................... 83

vi

CHAPTER 4

RESULTS ........................................................................................................... 85
Relations among the Observed Variables ................................................ 85
Measurement Model ................................................................................. 91

Estimation Procedure .................................................................... 91
Discriminant Validity ...................................................................... 94
Convergent Validity ........................................................................ 94
Measurement Model Modification .................................................. 96
Model Fit ........................................................................................ 98
Structural Equation Model ...................................................................... 100
Hypotheses Testing ..................................................................... 102
Modified Models ........................................................................... 108
Multiple Regression ................................................................................ 119
Hierarchical Regression ......................................................................... 120
Summary ................................................................................................ 122

CHAPTER 5

DISCUSSION .................................................................................................... 124
Socioeconomic Status and Family Viewing Environment ....................... 124
Parental Mediation ................................................................................. 125
Socioeconomic Status and Parental Mediation ...................................... 128
Parental Mediation and Children’s Visual Media Use ............................. 128
Socioeconomic Status and Children's Visual Media Use ....................... 130
Family Viewing Environment and Children’s Visual Media Use .............. 131
Children’s Visual Media Use ................................................................... 132
Child Characteristics and Children’s Visual Media Use .......................... 134
Child Characteristics and Imaginative Playfulness ................................. 135
Socioeconomic Status and Imaginative Playfulness .............................. 136
Children’s visual media use and imaginative playfulness ....................... 137
Limitations .............................................................................................. 144

APPENDICES ................................................................................................... 148
APPENDIX A Rating Instructions ....................................................... 149
APPENDIX B Explanatory Letter ........................................................ 153
APPENDIX C Coversheet of Questionnaire Packet for Parents ......... 159
APPENDIX D Family Background Questionnaire ............................... 160
APPENDIX E Visual Media Use Questionnaire .................................. 163
APPENDIX F Parental Mediation Questionnaire ................................ 164
APPENDIX G Imaginative Playfulness Scale ...................................... 165

BIBLIOGRAPHY ............................................................................................... 167

vii

LIST OF TABLES

Table 1. Summary of Variables and lnstmments ............................................ 21
Table 2. Characteristics of Participants ........................................................... 53
Table 3. Factor Loadings of Parental Mediation Indicators ............................. 59
Table 4. Average Viewing Amount by Dimensions and Visual Media Type ....69
Table 5. Program Titles by Dimension and Visual Media Type ....................... 70
Table 6. Categories and Viewing Amount of Visual Media Uses by Dimensions3
Table 7. Factor Loadings for the Imaginative Playfulness Indicators .............. 79
Table 8. Means, Standard Deviations, and Correlations of Indicators ............. 86
Table 9. Summary of Measurement Model ..................................................... 95
Table 10. Summary of Initial Hypothetical Model ............................................ 103
Table 11. Summary of Modified Model I .......................................................... 111
Table 12. Summary of Modified Model II ......................................................... 115
Table 13. Summary of Modified Model Ill ........................................................ 118
Table 14. Multiple Regression of Indicators of Parental Mediation .................. 120
Table 15. Hierarchical Regression of Indicators of Children Visual Media Use1s21
Table 16. Summary of Hypotheses and Findings ............................................ 123

viii

Figure 1.

Figure 2.

Figure 3.

Figure 4.
Figure 5.
Figure 6.
Figure 7.
Figure 8.

Figure 9.

LIST OF FIGURES
Theoretical Framework of Ecological Factors of Children’s Visual
Media Uses ......................................................................................... 10

Conceptual Model of Predictors of Children’s Visual Media Use and Its
Effects on imaginative Playfulness ...................................................... 12

Operational Model of Predictors of Children’s Visual Media Use and Its

Effects on Imaginative Playfulness ..................................................... 14
Measurement Model ........................................................................... 93
Modified Measurement Model ............................................................. 99
Initial Hypothetical Model .................................................................. 101
Modified Hypothetical Model | ........................................................... 109
Modified Hypothetical Model ll .......................................................... 113
Modified Hypothetical Model III ......................................................... 117

CHAPTER 1

INTRODUCTION

Play is a ubiquitous and key activity in child development. Traditionally,
play has been appreciated as an important vehicle for children’s social,
emotional, and cognitive development, as well as a reflection of their overall
development (Bredekamp & Copple, 1997). Imaginative play, in particular, has
been the focus of much research. According to van der Voort and Valkenburg
(1994), imaginative play refers to the play in which the child transcends the
constraints of reality by acting in an “as if” role.

In imaginative play, children pretend that they are someone or something
else, that objects represent something else, and/or that the participants are in a
different place and time. Imaginative play is sometimes used interchangeably
with such terms as “pretend play,” "fantasy play,” “dramatic play,” and “make-
believe play" (Fein, 1981; Johnson et al., 1999). Imaginative play has been
highly regarded as a process of adaptation to the environment by practicing or
representing events that children experience (Bredekamp & Copple, 1997;
Creasey et al., 1998; Fisher, 1992; Singer, 1996). Accordingly, environments
that children interact with have been acknowledged for playing a key role in the
development of children’s imaginative play (Roopnarine 8 Johnson, 1994).

In modern society, television has become a pervasive ecological factor,
and television viewing has become a main activity for even young children. For

instance, Singer and Singer (1983) reported that television viewing occupies a

larger amount of children’s time than any other single activity, with the exception
of sleeping. Due to the prevalence of television, many scholars studying
children’s imaginative play have been compelled to include the influence of
television.

As stated above, television, which provides a variety of sources of
information for children to use in their imaginative play, has been typically
regarded as a major influence on the content of imaginative play (Huston 8
Wright, 19983). What children watch on television may stimulate imaginative
play by giving children ideas for certain play episodes (Johnson et al., 1999).
Despite the fact that television viewing may have positive effects on the amount
of imaginative play by increasing the repertoire of pretend play, such play has
been speculated to be less creative and imaginative and more repetitive and
simple in terms of scripts and plots (Carlsson-Paige 8 Levin, 1995). In other
words, the “quality" of such play can be questioned, although the effects of
television viewing on the~quantity of imaginative play may be positive.

This presumption is supported by a series of hypotheses, which
presupposes that television has a number of structural attributes that may be
related to poor quality imaginative play. The first structural characteristic of
television is the rapid shift of scene and the constant changes in sound level.
Because this attribute provides children with few opportunities to reflect on what
they are watching, young children, who have limited memory ability, remember
outstanding and repetitive features or sequences, such as action sequences or

aggressive behaviors rather than the whole plot of a show. Accordingly, the plots

and scripts of imaginative play triggered by watching television are simple since
children may mainly retrieve the easily stored action sequences and aggressive
behaviors (Greenfield 8 Beagles-Roos, 1988; Honig et al., 2001 ).

Secondly, television viewing is a cognitively receptive process. Because it
does not provide direct and interactive experience, television offers children little
practice in forming their own images. For this reason, the way that children
watch television is a less meaningful way for them to experience the world. The
repertories and scripts of imaginative play affected by viewing may not be
generated as a result of children’s direct experiences and, therefore, are likely to
be imitative rather than creative (Singer, 1980).

Thirdly, television represents a form of vicarious play or fantasy in a
sense. Namely, children’s television viewing may complement their imaginative
play as part of their development (Singer 8 Singer, 1981 ). Because television
may be providing a major form of alternative consciousness or serving much the
same role as imaginative play, television is thought to increase the “let you
entertain me” attitude that undermines children’s willingness to produce their own
fantasies, and to contribute to less creative imaginative play (Singer et al., 1984a;
Valkenburg 8 van der Voort, 1994).

Besides these structural characteristics of television affecting the quality of
imaginative play, the Iinkup of television programs and toys may also be blamed
for the low quality of play. Because of the marketing of television shows, highly
realistic, single-purpose, and mostly stereotypical or even violent toys based on

program characters are available to young children, which, in turn, may

encourage them to continue imitating what they have seen on television
(Carlsson-Paige 8 Levin, 1990).

In conclusion, many scholars have speculated that imaginative play
generated by viewing television is likely to be imitative and limited in scope. In
addition, the behavior characteristics of such play may easily become aggressive
and disruptive. When cognitive, social, and physical mobility in a spontaneous
and pleasurable manner, is the attribute of high quality play, the imaginative play
influenced by television watching may be considered to be one of lower quality
(Lieberman, 1966; Rubin et al., 1983).

Therefore, efforts to understand the quality of such play and its
contributions to development have taken on a growing sense of importance and
urgency (Carlsson-Paige 8 Levin, 1987, 1990; Meidzian, 1991). Despite the
increasing concern and debate regarding the effects of viewing on the quality of
imaginative play, the preponderance of earlier publications tend to be arguments
at the conceptual level based on developmental theories or societal values,
whereas there is a paucity of empirical research on the effects of television as it
pertains to the quality of imaginative play. One of the reasons for little attention in
the literature may be due to the difficulty of scientifically investigating the quality
of imaginative play.

Playfulness has been used as a way to describe the quality of play since
Lieberman first pioneered and measured the concept. Playfulness is a feature of
children’s play that represents imagination, creativity, and penchant for novelty

seeking, humor, emotional expressiveness, openness, and flexibility (Aguilar,

1985; Barnett, 1990; Boyer, 1997/98; Lieberman, 1966, 1977). Playfulness
provides favorable opportunities to enhance play, not in a disruptive and
impulsive style but in a productive and appropriate way (Aguilar, 1985; Barnett,
1990; Lieberman, 1966, 1977). Playfulness, therefore, is essential to high quality
play, where play is defined as an action that promotes and integrates motor,
cognitive, social, and emotional development and learning in pleasant, voluntary,
child-centered, and supportive contexts (Kieff 8 Casbergue, 2000). In light of
this, playfulness becomes the variable that needs more careful examination,
when the quality of imaginative play influenced by viewing television is being
questioned. Therefore, this study will examine the role that television plays in the
development of children’s imaginative playfulness.

There are several points that were considered in this research in studying
the effects of television on imaginative playfulness based on past studies. First,
according to the ecological theory of Bronfenbrenner (1989), there are
environmental factors that are likely to be associated with child development at
manifold levels. Multiple factors are assumed to contribute to individual
differences in children’s viewing practices at different ecological levels (Anderson
8 Bryant, 1983; Huston 8 Wright, 1998b; Pinon et al., 1989). For preschoolers,
the most important ecology surrounding them is their family. Therefore, this
study examined factors at the familial level, including socioeconomic status, the
family viewing environment, and parental mediation, which are likely to be
predictive of children’s viewing patterns, and may indirectly influence imaginative

playfulness.

Secondly, it has been widely and consistently revealed that the effect of
different programming is an important factor when television influences are
investigated. Earlier studies found that the influences of television may be tied to
the characteristics of television programs such as violent content, educational
components, or intended audiences. It has been found that children played more
imaginatively after they watched educational programs designed for children in
laboratory settings; on the other hand, children who watch a great deal of violent
and action adventure programs are less imaginative and more aggressive
(Singer 8 Singer, 1981; Singer et al., 1984a; Valkenburg et al., 1992; Zuckerman
et al., 1980). Therefore, this study focused on the effects of different
programming in terms of three characteristics: those programs designed primarily
for children, those containing violent episodes, and those produced for an
educational purpose.

Thirdly, although television is the most influential and accessible mass
medium it is not the only mass medium. There are many types of mass media
that young children experience besides television and the use and experience
with television can be affected by other mass media (Funk, 2001). Among a
variety of mass media, watching movies and videos by using VCRs and DVD
players is also essentially a passive activity that may affect children’s play at a
later time by triggering the imagination in some way (Johnson et al., 1999). In
light of this, videos and DVDs are included in this study to examine their

influences on imaginative playfulness.

Statement of the Problem

The main purpose of this study was to examine the associations among
preschool children’s visual media use, the quality of their imaginative play, and
the ecological factors of their viewing practices in their family. For this purpose, I
hypothesized that the levels of preschoolers’ viewing different programming at
home are influenced by ecological variables, which then affect their imaginative
playfulness. Specifically speaking, a child’s viewing of different programs is
expected to be influenced by aspects of the child’s ecology such as
socioeconomic status of the family, the environment of visual media use in the
family, and parental mediation of what the child watches. In turn, in this present
study, the degree to which children view different programs was assumed to

contribute significantly to their imaginative playfulness.

Theoretical Framework

Human ecology, as described by Bronfenbrenner, was used as a
theoretical framework for this study. Ecology is the study of the relationship
between organisms and environments (Griffore 8 Phenice, 2001). According to
the ecological perspective, individuals and environments interacting over time
affect developmental processes. Bronfenbrenner (1979) asserted that research
on human development should include an awareness of the environmental
systems in which people are embedded.

In the ecology of human development, an individual is a psychological
system that is affected by conscious and changing roles, unconscious needs,

and motives. Individuals live simultaneously in multiple nested environments.

Bronfenbrenner (1979, 1989) proposes that there are five levels of environment:
the microsystem, mesosystem, exosystem, macrosystem, and chronosystem.

The microsystem is the child’s immediate environment. The key elements
of the microsystem are activities, role, and interpersonal relations experienced by
the developing person in face-to-face settings. For preschool children, a
microsystem related to visual media use can be the characteristics of their family,
such as family viewing habits or parental mediation.

Children tend to spend time in more than one microsystem. The linkages
and processes that occur between those settings are referred to as the
mesosystem. In other words, the mesosystem is another level of environment
that comprises the linkages and processes that take place between two or more
settings containing the developing person. Accordingly, a mesosystem is a
system of microsystems. For preschool children, the relation between the home
and the child care center is a mesosystem.

The exosystem refers to events that occur in a setting that does not
include the child directly but that can have an effect on the child. Peers’ use of
visual media at their home is an example of the exosystem influencing the visual
media use of the child. The child may never directly interact with exosystems,
but what happens there may affect the child through their microsystems.

The macrosystem is the broadest level of the environment where the child
and his/her family are living. The macrosystem consists of the overarching
pattern of microsystem, mesosystem, and exosystem. Characteristics of a given

culture, subculture, or other broader social context that the family belongs to are

examples of the microsystem influencing visual media use of the child.

Bronfenbrenner (1989) later proposed another concept for his ecological
system, which was the chronosystem. The chronosystem refers to an ecological
transition that occurs whenever a person’s position in the ecological environment
is altered as a result of a change in role, setting, or both. For example, the birth
of a baby, the beginning of school, graduating from school, moving to a new
community, getting a job, changing careers, getting married, and retirement are
some ecological changes that individuals and families can experience.

These multiple environments, as a series of embedded and concentric
circles, start with the closest environment, such as family and friends, and extend
to encompass societal level factors, such as nation, social class, and ethnic
group. Based on the Bronfenbrenner ecological theory of human development, a
theoretical framework of factors that are likely to be associated with children’s
visual media use is illustrated in Figure 1. As depicted, potential ecological
systems that are likely to influence visual media use are at the level of the
individual and a series of contexts that are important for children, are at the level

of the family, peer group, community, and other broader level of environments.

Figure 1. Theoretical Framework of Ecological Factors of Children’s Visual
Media Uses (Adapted from Luster 8 Oh, 2001, p. 718)

  

Time Context

Constancy
Change over time

  

Macro Context
Laws regarding visual media
Cultural values and beliefs

 
 

    
   
 
  

Extrafamilial Context
Peers’ use of visual media
Neighbors, Community

Visual media industry _‘

 
    

   
 

x//
f,
/ \
/
/ i‘.‘
\
Familial Context \
, Family viewing habits \
/ Maternal employment \ i
/' Availability of visual media equipments .\
Parental mediation \ /'

 

Gender
‘ Cognitive development ./ / / ,,
\ \ Interests / ,/

\\
k.

\\ \\

  
 
 

 

 
       
 

1O

    
 
 
 
  
 

Popularity of programs
.\

 

 

 

/
Age / /
[/I/ / ' /
(1’, 1/

Conceptual and Operational Models

According to Bronfenbrenner (1986), the family, which is the most
proximal environment, is one of the most significant influences in the child’s
ecology. Among the numerous ecological predictors affecting children’s visual
media use displayed in Figure 1, several factors at the individual level and family
level were selected in this study. Figure 2 illustrates a conceptual model for this
study showing the pathways of ecological predictors to playfulness through
children’s visual media use.

In the conceptual model, the family’s socioeconomic status is expected to
affect the viewing environment in a family and parental mediation, which then
would be the predictors of children’s viewing of television, videos, and DVDs.
Besides the family viewing environment and parental mediation, the variables of
child characteristics at the individual level are likely to influence children’s
viewing practices. Finally, children’s imaginative playfulness is expected to be
determined by the children’s characteristics and their visual media use.

In the conceptual model, therefore, the familial factors are expected to
indirectly influence children’s imaginative playfulness through their visual media
use. Also, child characteristics are anticipated to affect imaginative playfulness

directly and indirectly through their visual media use.

11

 

 

mozmtmaomcmzo
EEO

 

 

 

>m_a
o>_umc_mmE_
ho mmmciSmi

 

 

 

 

mmocSSmE o>=mc_mmE_ co mwootw m: cam 33 Eco—2 .955 $65.30 Co 99235 Co Eco—2 .maamocoo .N 6.5mm

 

 

mm:
£09: .933
$9230

 

 

 

\

 

cozmfimE
.mEoLmn.

 

 

 

 

 

éEfimE
EoEcocscm
@5265

 

macaw
o_Eocooo
-o_oow

J
1
_
_
.

 

 

 

12

Variables, which are operationally defined and measured to indicate the

constructs in the conceptual model, are listed in the operational model (see
Figure 3. At the family level, socioeconomic status, which is likely to influence
viewing environment in a family and parental mediation, was indicated by gross
family income, family head’s occupation, and family head’s education. Viewing
environment in a family is expected to be indicated by three observed variables,
including amount of visual media use, number of visual media devices in the
family, and number of visual media devices that children own. Parental
mediation of visual media use is indicated by instructive mediation and restricted

mediation.

At the individual level, child characteristics are likely to influence children’s
visual media use and their imaginative playfulness. In this study, child
characteristics are indicated by age and gender.

Finally, children’s imaginative playfulness is indicated by five observed
variables, including physical spontaneity, social spontaneity, cognitive

spontaneity, manifest joy, and sense of humor.

13

 

 

5950 .N
8< .F BEBE 82.25% .N
8556955 EEO cosmfimE 62826:. .F
cozmfimE .mucmema

 

 

 

 

 

 

 

 

 

 

 

 

 

/
o 83 .o ,
58:: Co omcow .m m on? .m 950:. >__Em mmoL .
>ormmzcms. .v v 69:. .v 53838 mbmmcszmw .M
£29.83 o>=Emoo m m on: m cosmazooo m com; >__Emn_ é
$29,.on Eoom .N N 83 .N .39... 25.88868
35:9:on .mofiEn. ._. F on? .F . . 4
mmm£2>ma 93269:. cm: £va 1
.953 95.220

 

 

 

 

 

 

mm: mBoE _m:w_> co anE< .m
3033 SEE
, .963 $5.22". v.o LonEzz .N
/ woo_>ou
Enos _m:m_> Co 59:52 .F
éESmE
EmEcocSco @5365

 

 

 

“$chng 026588. co maomtm m: new $3 Enos. .m:m_> MEEEEO Co 2905de Lo .805. $8530 .m «taut

Conceptual and Operational Definitions

Socioeconomic Status

Conceptual: Social and economic factors that distinguish people in a
society (Reiss, 1961).

Operational: The reports of primary caregivers on items of gross annual
income of their family and years of education of the family head on the Family
Background Questionnaire (see Appendix D) and the scores of parental
occupation level as coded using the Socioeconomic Index (Duncan, 1961 ).

Viewing Environment in a Family

Conceptual: The physical traits or attributes that influence what children
watch and how much they watch on television, video, and DVD within the
household.

Operational: The reports of primary caregivers regarding total minutes of
visual media use at home during the week previous to the reporting day, the
number of visual media devices in their family, and the number of their child’s
visual media devices as measured by items in the Family Background
Questionnaire.

Parental Mediation

Conceptual: Parent-child interaction intended to influence the effects of

watching television, videos, and DVDs on their children’s development (Weaver

8 Barbour, 1992).

15

Operational: The scores on the Parental Mediation Questionnaire (see
Appendix F) adapted from the inventory of Parents’ Television Mediation Styles
(Valkenburg et al., 1999).

Instructive Mediation

Conceptual: Parents’ active efforts to discuss and criticize what their child
watches on television, video, and DVD (Valkenburg et al., 1999).

Operational: The scores reported by primary caregivers on items 1-5 on
the Parental Mediation Questionnaire.

Restrictive Mediation

Conceptual: Parents’ active efforts to set rules on what and how much
their child watches on television, video, and DVD (Valkenburg et al., 1999).

Operational: The scores reported by primary caregivers on items 6-10 on
the Parental Mediation Questionnaire.

Child Characteristics

Conceptual: The individual traits or attributes of the child that can affect
his/her visual media use.

Operational: The reports of the primary caregivers on their child’s age and
gender on the Family Background Questionnaire.

Children ’3 Visual Media Use

Conceptual: Children’s actual viewing of television, video, and DVD at
home.

Operational: Children’s time spent viewing Type 1, Type 2, Type 3, Type

4, Type 5, and Type 6 programs during the week previous to the reporting day at

16

home, as reported by primary caregivers on the Visual Media Use Questionnaire
(see Appendix E).
Type 1

Conceptual: The amount of children’s viewing at home of nonviolent
informative programs designed for children.

Operational: The reports of the primary caregivers on the Visual Media
Use Questionnaire stating the titles and frequencies of viewing of nonviolent
programs designed primarily for viewers 12 or younger that included informative
or educational messages.
Type 2

Conceptual: The amount of children’s viewing at home of nonviolent and
non-inforrnative programs designed for children.

Operational: The reports of the primary caregivers on the Visual Media
Use Questionnaire stating the titles and frequencies of viewing of nonviolent
programs designed primarily for viewers 12 or younger that do not include
informative or educational messages.
Type 3

Conceptual: The amount of children’s viewing at home of violent and non-
informative programs designed for children.

Operational: The reports of the primary caregivers on the Visual Media
Use. Questionnaire stating the titles and frequencies of viewing of violent
programs designed primarily for viewers 12 or younger that do not include

informative or educational messages.

17

Type 4

Conceptual: The amount of children’s viewing at home of nonviolent
informative programs for general audiences.

Operational: The reports of the primary caregivers on the Visual Media
Use Questionnaire stating the titles and frequencies of viewing of nonviolent
programs with informative or educational messages not designed primarily for
viewers 12 or younger.
Type 5

Conceptual: The amount of children’s viewing at home of nonviolent and
non-informative programs for general audiences.

Operational: The reports of the primary caregivers on the Visual Media
Use Questionnaire stating the titles and frequencies of viewing nonviolent
programs without informative or educational messages, not designed primarily
for viewers 12 or younger.
Type 6

Conceptual: The amount of viewing at home of violent and non-inforrnative
programs for general audiences.

Operational: The reports of the primary caregivers on the Visual Media
Use Questionnaire stating the titles and frequencies of viewing of violent
programs without informative or educational messages not designed primarily for

viewers 12 or younger.

18

Imaginative Playfulness

Conceptual: Children’s disposition during imaginative play that is
characterized by physical spontaneity, social spontaneity, cognitive spontaneity,
manifest joy, and sense of humor (Lieberman, 1966).

Operational: Children’s score on the adapted Children's Playfulness Scale
(CPS; Barnett, 1990, 19913), which deals with imaginative playfulness as
reported by their head teacher on the Imaginative Playfulness Scale (see
Appendix G).
Physical Spontaneity

Conceptual: The coordination and motor activity level during imaginative
play (Barnett, 1990; Lieberman, 1966).

Operational: Children’s score on items 1-4 on the Imaginative Playfulness
Scale.
Social Spontaneity

Conceptual: Ability to get along with others and to move in and out of
groups during imaginative play (Barnett, 1990; Lieberman, 1966).

Operational: Children’s score on items 5-9 on the Imaginative Playfulness
Scale.
Cognitive spontaneity

Conceptual: Imaginative quality including uniqueness, unconventionality,
flexibility and variety of thoughts during the child’s imaginative play (Barnett,

1990; Lieberman, 1966).

19

Operational: Children’s score on items 10-13 on the Imaginative
Playfulness Scale.
Manifest joy

Conceptual: The extent to which the child expresses enthusiasm,
exuberance, enjoyment, lack of restraint, and vocalization during imaginative play
(Barnett, 1990; Lieberman, 1966).

Operational: Children’s score on items 14-18 on the Imaginative
Playfulness Scale.
Sense of humor

Conceptual: Appreciation of comical events, recognition of funny
situations, and gentle teaching during imaginative play (Barnett, 1990;
Lieberman, 1966).

Operational: Children’s score on items 19-23 on the Imaginative
Playfulness Scale.

A summary of variables and instruments is presented in Table 1.

20

Table 1.

Summary of Variables and Instruments

 

Variable

Instrument

 

Exogenous variables

 

Socioeconomic stiatus
Family head’s occupation
Family head’s education
Gross family income

Child characteristics
Age
Gender

Socioeconomic Index
Panel Study of Income Dynamics
Panel Study of Income Dynamics

Family Background Questionnaire
Family Background Questionnaire

 

Mediating variables

 

Viewing environment in a familv
Number of visual media devices
Number of children’s of visual media devices
Amount of visual media use

Parental mediation
Instructive mediation
Restrictive mediation

Children’s visutal media use
Type 1 programs
Type 2 programs
Type 3 programs
Type 4 programs
Type 5 programs
Type 6 programs

Panel Study of Income Dynamics
Panel Study of Income Dynamics
Panel Study of Income Dynamics

Parental Mediation Questionnaire
Parental Mediation Questionnaire

Visual Media Use Questionnaire
Visual Media Use Questionnaire
Visual Media Use Questionnaire
Visual Media Use Questionnaire
Visual Media Use Questionnaire
Visual Media Use Questionnaire

 

Dependent variables

 

Imaginative glaflulness
Physical spontaneity
Social spontaneity
Cognitive spontaneity
Manifest joy
Sense of humor

Imaginative Playfulness Scale
Imaginative Playfulness Scale
Imaginative Playfulness Scale
Imaginative Playfulness Scale
Imaginative Playfulness Scale

 

21

CHAPTER 2

LITERATURE REVIEW

In this chapter, a comprehensive review of research literature was
organized around four topics: ecological factors of children’s viewing;
playfulness; effects of viewing on imaginative play; and effects of viewing on

playfulness.

Ecological Factors of Children’s Viewing
Socioeconomic Status

The three most commonly used measures of socioeconomic status are
income, education, and occupation. The educational levels of parents have been
found to have a direct impact on predicting the amount of children’s viewing.
There is a correlation between parents with advanced degrees and relatively low
television use by their children from preschoolers to adolescents. In contrast, it is
shown that the low educational level of parents has a direct influence on the
higher amounts of children’s viewing (Anderson et al., 2001; Bianchi 8 Robinson,
1997; Comstock et al., 1978; Condry, 1989; Huston et al., 1992; Huston 8
Wright, 1996; Pinon et al., 1989; Tangney 8 Feshbach, 1988). Of course, the
influence of the parents’ education should be discussed more specifically in
terms of types of television programming. In general, preschoolers and
elementary school children of better—educated parents view an equal or greater

amount of informative programs, such as Sesame Street (Ball 8 Bogatz, 1970;

22

Cook et al., 1975; Huston et al., 1990; Truglio et al., 1996). This implies that
parental educational status influences what, as well as how much, their children
watch.

Similarly, occupational levels of parents also influence their children’s
viewing practices; the high occupational status of parents is associated with
relatively low television use by adolescents, as well as preschoolers and
elementary school children (Anderson et al., 2001; Bianchi 8 Robinson, 1997;
Comstock et al., 1978; Condry, 1989; Huston et al., 1992; Huston 8 Wright,
1996; Tangney 8 Feshbach, 1988).

In regard to income level, Atkin and his colleagues (1991) found that
family income levels account for the amount of television viewing of fifth-graders
and tenth-graders (B = -.09, R2 = .16, p = .04). Considering the types of
programs, preschoolers and adolescents in low-income families watch an equal
amount of non-animated entertainment shows and more animated entertainment
shows (Anderson et al., 2001; Huston 8 Wright, 1996). Huston and Wright
(1996), however, showed family income does not predict viewing of children
ranging in age from two years to seven years when parental education is
statistically controlled. Therefore, it can be concluded that the differences in the
amount of children’s television viewing, in accordance with income differences,
are explained by differences in parents’ educational levels.

As Huston and Wright (1998b) argued, these descriptive influences
provide little insight into the effects of socioeconomic status on children’s viewing

practices. Therefore, it is important to investigate variables that are involved in

23

pathways from socioeconomic status to viewing practices in families.
Viewing Environment in a Family

Television Viewing Practices in a Family

Television sets are located in the common living areas of most homes, so,
young children are often present and exposed to programs watched by their
family members (Carpenter et al., 1989). Because much of television viewing
occurs in the company of their family, the viewing practices of family members
are influential on children’s viewing. Huston and Wright (1996) found, for
example, that a parent was present during the majority of time when their three-
to seven-year—olds watched general audience programs; by contrast, parents
were present for about 25% of the time that children viewed child audience
programs. Consequently, television viewing habits and preferences of family
members are a powerful source of early exposure to television for young
children. For example, the choices of programs were guided by the parents’
preferences and habits and not those of the children between three and seven
years old (Desmond et al., 1985; St. Peters et al., 1991 ). For this reason, the
presence of parents has been mainly studied as a factor influencing young
children’s viewing practices.

The availability of parents is determined by family structure and size.
F ifth-graders and tenth-graders in single-parent homes often view without
supervision, and children view programs more in homes where the father is
absent (Atkin et al., 1991). Children in larger families are also likely to view more

television and their parents are less likely to mediate viewing of preschool and

24

elementary school children (Gross 8 Walsh, 1980).

Maternal employment is another factor that is related to the presence of
parents affecting children’s television viewing. It is widely assumed that children
of employed mothers view more television because the children may be less
closely supervised and can access television more easily (Huston 8 Wright,
1998b). However, maternal employment was negatively related to viewing
practices of children ranging in ages from three to seven years (Huston 8 Wright,
1996; Huston et al., 1999; Pinon et al., 1989). In other words, children of
employed mothers watch less television than children whose mother are full-time
homemakers. One of the reasons could be that many of those children are in
organized childcare settings where little or no television is shown during the day.

In addition to parents, children also frequently watch television with their
sisters and brothers. The presence of older siblings is likely to reduce viewing;
by contrast, the presence of younger siblings increases it (Pinon et al., 1989). In
particular, older children play a more powerful role in program selection than
younger children (Lull, 1978) and they serve as a model of viewing for their
younger siblings (Dunn, 1983). Only children, or children with younger siblings,
watch more child-inforrnative programs; however, young children view more
entertainment programs with older siblings (Huston 8 Wright, 1996).

In summary, family structure, the number of older and young siblings, and
mother’s employment in a family appear to set the boundaries for children’s
opportunities to watch television, because they might impose constraints on

viewing practices at home. Therefore, television viewing practices in a family can

25

be a factor that indicates the familial characteristics affecting children’s television
viewing.
Availability of Television Sets at Home

Time use by young children is likely to depend on opportunities and
restrictions provided by the home environment or other child rearing
environments (Huston et al., 1999). Accordingly, the number and the location of
television devices for viewing are likely to affect the viewing patterns of children.

The effect of the number of television sets has been thoroughly
investigated. Not surprisingly, children tend to watch television more when
children have their own television sets than children in families with a single
television set. When a family has many television sets, the parents are likely to
view television with preschoolers as well as school age children (Atkin et al.,
1991; Bower, 1973; Chaffee 8 Tims, 1976; Lull, 1978; McDonald, 1986). Gross
and Walsh (1980), however, found that parents in households with fewer
television sets have more chances to control viewing practices of their children
ranging in age between three and 12 years and tend to discuss television
programs because they usually view television together. This implies that the
opportunity of parental control of children’s television viewing decreases when
children possess their own television sets. Therefore, the number of television
sets and their location, and children’s ownership of television sets can predict the
practices of children’s television viewing at home. Not surprisingly, the number
and location of television sets in a household tends to increase with social class.

Dorr and Rabin (1995), therefore, speculated that the physical environment

26

factors could be related to socioeconomic status of the family. Interestingly, this
result and interpretation contradict the findings reported above that high
socioeconomic status of parents is associated with relatively low television use of
their children. Parental mediation of what their children watch may be the reason
for inconsistent findings.
Parental Mediation

Parents have been consistently named as the most influential family
members. It is commonly recommended that parents take an active role in
mediating their children’s experiences with television. Broadly speaking, parental
mediation is defined as a group of behaviors that are relevant to interactions
between parents and their children (Desmond et al., 1990). In a narrow sense,
Desmond and his colleagues (1985) define parental mediation for young children
as a form of active effort by parents to help children understand their physical
and social world. Based on this definition, parental mediation of television’s
effects means an active and purposeful process in which parents interpret, filter,
or augment some specific television message (Rothschild 8 Morgan, 1987). The
concept of parental mediation of television viewing assumes that television is a
powerful source of modeling and learning but the ability of young children to
understand the messages on television is limited. Therefore, mediating roles
enacted by parents are critical for intervention in television viewing of young
children who are possibly in greater need of guidance.

Operational definitions of parental mediation differ across previous studies

because the actual practices of parental mediation are quite diverse. Accordingly,

27

the categories of parental mediation that have been used in past research are
also inconsistent and diverse. Weaver and Barbour (1992), who reviewed the
literature on parental mediation categorized three kinds of parental mediation:
restrictive mediation, evaluative mediation, and unfocused mediation that refers
to coviewing and non-evaluative discussion. Recently, Valkenburg and her
colleagues (1999) tried to provide a reliable instrument of measuring parental
mediation with statistical scales. They found three styles of television mediation
by Dutch parents and the categories are instructive, restrictive, and social
coviewing.

Past studies on the effects of parental mediation on children’s television
viewing mainly focused on restrictive mediation rather than instructive mediation.
In general, the degree of restrictive mediation was negatively correlated with the
amount of children’s viewing television. Atkin and his colleagues (1991) found
that fifth—graders and tenth-graders were likely to watch less television when their
parents set rules for viewing or prohibited some types of viewing, although recent
studies found no relationship between restrictive mediation and viewing practices
of preschool and elementary school children (Warren, 2003; Warren et al., 2002).
If parents provided recommendations to their children of which programs to
watch, and there were explanations or discussions of programs with their
children, then there were few relationships between the children’s nonrestrictive
mediation and their frequency of television viewing (Warren, 2003; Warren et al.,
2002). Rather, nonrestrictive parental mediation enhanced what children learned

from television (Desmond et al., 1985, Desmond et al., 1987; Messaris 8 Sarett,

28

1981).

The influences of socioeconomic status on parental mediation have not
been consistently found to be at variance. While Valkenburg et al. (1999) found
that the more highly educated the parents were, the more rules they had
regarding elementary school children’s viewing. It was also revealed that
parents’ education was not a predictor of parental mediations for adolescents,
(Lin 8 Atkin, 1989), with elementary children (Atkin et al., 1991; Gross 8 Walsh,
1980, Warren et al., 2002), nor for preschool children (Warren, 2003). The
findings regarding income level are also inconsistent; parental income did not
predict parental mediation, such as rule making in several studies (Gross 8
Walsh, 1980; Lin 8 Atkin, 1989), although Atkin and his colleagues (1991)
reported parents’ income was a positive predictor of nonrestrictive mediation.
Accordingly, it is appropriate to examine the affects of socioeconomic status on
parental mediation.

Child Characteristics

Individuals watch television to satisfy a variety of needs or to gratify a wide
range of interests. Interests, goals, and motivations of viewing are formed by the
sets of individual characteristics guiding their viewing.

Age

Total viewing time increases rapidly with age from infancy to the first few
years of life, and the amount of viewing stagnates or declines when they enter
school (Anderson 8 Field, 1983; Comstock, 1991; Dennison et al., 2002; Huston

8 Wright, 1996; Paik, 2001, Warren, 2003). Because schools require children’s

29

attendance and participation during certain time periods, children’s schooling
sets boundaries for opportunities to watch television. Later on, in middle school,
it was found that the time children spend watching television peaks and then
declines when they become teenagers (Atkin et al., 1991; Bianchi 8 Robinson,
1997; Huston 8 Wright, 1998b; Liebert et al., 1982). This is likely a function of
teenagers’ social alternatives that replace television viewing.

In addition to the developmental changes that occur as a function of age-
related life events, such as schooling, the types of programs that children watch
change with age because of the age appropriateness of programs. For example,
the viewing of informative programs for children peaked at age four (3-4.5 hours
per week) and declined with age among children from ages five to seven (Pinon
et al., 1989); viewing cartoons and comedies increased at age five (3-5 hours per
week), then leveled off (Huston et al., 1990). No age changes occurred on other
types of entertainment programs.

Gender

Gender differences in the use of time are of theoretical interest because
participation in gender-stereotyped activities may be one basis for sex-
differentiated behavior patterns (Huston et al., 1999). Gender differences in
viewing presumably reflect the influences of the individuals’ gender-typed
interests and motivations. Accordingly, children’s gender becomes an individual
characteristic that contributes to the amount of programming they watch. On
average, boys watch more television than girls in various age groups from

preschoolers to adolescents (Bianchi 8 Robinson, 1997; Carpenter et al., 1989;

30

Rosengren, 1994; Singer 8 Singer, 1981), although some research found no
gender difference (Condry, 1989).

The gender difference particularly occurs in viewing such programs as
cartoons and action adventures (Huesmann 8 Eron, 1986; Huston et al., 1990).
According to Alvarez et al. (1988), during observations in laboratories, boys were
more attentive than girls to cartoons and to commercials with high rates of action,
violence, and visual special effects. They interpreted the gender difference of
total viewing to the greater number of masculine gender-typed programs, such as
cartoons, action adventure shows, and sports, than the gender-typed programs
for girls on television. That difference in choice of programs typically appears
around age four or five, and usually continues into late childhood (McKenzie et

al., 1992; Timmer et al., 1985) or even increases with age (Huston et al., 1999).

Playfulness
Definition
Even though the question of “what is play” has been of interest for many

years, the definition of play has not been clearly stated (Garvey, 1977).
According to Barnett (1990), one of the reasons why attempts to define play
failed may have stemmed from the failure to provide empirical evidence
regarding the definition and corresponding theory. Therefore, it has been a
major issue in presenting direct empirical data on children’s play in scientific
ways. Matthews and Matthews (1982), who classified a variety of approaches to

defining play, categorized the approach to playfulness as an attempt to define

31

play in terms of a psychological predisposition.

According to Aguilar (1985), playfulness is the perception or attitude that
allows an individual to behave spontaneously in play. The studies on playfulness
basically originate from interest in the way children play as opposed to why
children play. Playfulness is the orientation of one’s personality to be flexible,
active, imaginative, creative, and find pleasure in what one chooses. The
interest in and studying of playfulness is an attempt to focus on play as a
characteristic of the individual, rather than as what the child “does” (Barnett,
1990). If being playful means that free expression, creativity, and imagination are
released, then perhaps it can provide favorable opportunities for enhanced play
(Aguilar, 1985).

It implies that playfulness can be viewed as a positive characteristic or
individual resource for play. In other words, playfulness is a dispositional aspect
of play that implies the “quality” of play, by focusing on the playful feature of the
child rather than focusing on specific behavioral elements within certain time
frames and physical contexts (Barnett, 1991a). Therefore, it is more useful in
studying the concept of playfulness to focus on the qualities that children bring to
their environment rather than on activity, as traditional scales have done.

The Studies of Playfulness

There are several groups of scholars who have developed a definitive
description of playfulness, an explanation for its importance, and a method of
measuring it. Lieberman (1965, 1966), as a pioneering researcher, proposed

that playfulness was a stable disposition. She clearly postulated the existence of

32

the playful trait in children and developed a measurement for the construct that
would provide direct empirical evidence of its underlying nature. In her work,
Lieberman (1965, 1977) operationally defined playfulness using sub-dimensions
and developed the Playfulness Scale. The items of a 5-point rating scale, which
was designed to be completed by teachers, includes “While playing, how often
does the child show flexibility in his interaction with the surrounding group
structure” or “What degree of imagination does the child show in his/her
expressive dramatic play?” According to her, playfulness is comprised of five
constituent dimensions labeled as physical spontaneity, social spontaneity,
cognitive spontaneity, manifest joy, and sense of humor. Her studies indicated
that playfulness in children is a unitary behavioral dimension across the physical,
social, and cognitive domains, although playfulness is conceptualized in terms of
subcomponents, and goes beyond the childhood years to exist as a trait in
adolescents (Lieberman, 1977).

Barnett, who was a strict follower of Lieberman, elaborated on
Lieberman’s early work. She thought the identification and description of
playfulness in young children theoretically and conceptually contributed to the
documented attempts to explain the nature, antecedents, and consequences of
play behaviors. Her major work on playfulness developed an empirically sound
method to assess playfulness. In order to do this, she replicated Lieberman’s
concepts and developed the Children’s Playfulness Scale (CPS), which is a
methodologically reliable and valid measurement of playfulness (Barnett, 1990,

1991a). The examples of questionnaire items, which are 5-point rating scale, are

33

“The child invents his/her own games to play” or “The child is restrained in
expressing emotion during play.” The items were developed to be evaluated by
teachers, because Barnett believed playfulness was an internal personality
construct that is not limited to certain time frames and physical context. The
results obtained by factor analysis confirmed the existence of the general
playfulness factor and its five component dimensions that are consistent with the
five component traits of playfulness posed by Lieberman (Barnett, 1990, 1991b).
The third school of thought is composed of Rogers and her colleagues,
who proposed playfulness as a psychological construct involving individual
differences in the disposition to play (Rogers et al., 1998). They criticized
Lieberman because she focused on behaviors during play, such as facial
expression, laughing, or joking, rather than on the personality disposition.
Therefore, Rogers and her colleagues did not focus on Lieberman’s work but
relied on existing definitional criteria based on other literature (Krasnor 8 Pepler,
1980; Rubin et al., 1983; Smith 8 Vollstedt, 1985), which is useful in
distinguishing the playful disposition during play. They generated a pool of items
representing six criteria for the disposition of play from the definitions and
developed a trait-rating instrument, named the Child Behaviors Inventory of
Playfulness (CBI) to measure individual differences in playfulness as a
personality disposition. The items with a 5-point rating scale included, “The child
uses toy/objects in unusual ways” or “the child starts activities for his/her own
enjoyment.” Similar to the CPS developed by Barnett, the CBI was expected to

be completed by persons who have knowledge of a child’s behaviors in multiple

34

contexts. Accordingly, rating by teachers in school settings were used instead of
observation by researchers in a specific setting. Although the item pool was
constructed to represent six criteria for the disposition of play, the results
obtained by factor analysis reveal that children’s playfulness is a uni-dimensional
trait.

The last group of scholars dealing with playfulness is composed of Bundy
and her colleagues who are occupational therapists. Occupational therapists
have regarded play as the child’s primary occupation (Bundy, 1992; Knox, 1996;
Okimoto et al., 1999). They appreciate the therapeutic importance of play to help
children improve functional performance (Case-Smith, 1996). Bundy’s writing
(1993) supports the position that playfulness is the one aspect that differentiates
play from work, because without playfulness, all activities become work.

According to Bundy and her colleagues, playfulness can be determined
within any transaction by evaluating the presence of four elements: intrinsic
motivation, internal control, freedom to suspend reality, and framing (Bundy,
1993, 1998). Intrinsic motivation suggests motivation by some aspect of the
activity itself rather than for an external reward. Internal control refers to the fact
that an individual is in charge of his or her actions. Freedom to suspend reality
means the individual chooses how close to objective reality the transaction will
be. Framing refers to the effects that the individual creates by adjusting
behaviors and the surrounding environment so that the playful act is clearty
distinguishable from a serious act.

Based on the concept of playfulness as a reflection of the combined

35

presence of the four components, the Test of Playfulness (ToP) has been
developed and revised (Bundy, 1998; Metzger, 1993). The examples of items on
a 4-point rating scale are “The child is actively engaged” or “T he child
incorporates objects or other people into play in novel, imaginative, creative,
unconventional, or variable ways.” Contrary to other instruments of playfulness,
ToP was designed to be scored from videotapes of children engaged in free play
by trained raters, who are mostly therapists. Therefore, ToP was useful in
measuring the playfulness of very young children and children with disabilities;
however, uses of the measurement are limited to trained therapist observers.
Playfulness and Development

Lieberman, who postulated the concept of playfulness, argued that
children’s playfulness is related to their development. Among a variety of
developmental characteristics, she was interested in imagination and creativity
which are the important elements of children’s play. According to Lieberman
(1965, 1967), children’s playfulness is associated with imagination and creativity.
As she speculated, the relationship among them has been proposed by a
number of scholars; more playful children tend to be more creative and
imaginative (lverson, 1982; Kogan, 1983).

In addition to Lieberman’s early exploration of the connections between
playfulness, imagination, and creativity, a number of studies that investigated the
relationship between playfulness and other developmental characteristics have
been implemented. Bundy (1998) examined playfulness of children ranging in

age from 22 to 118 months and found that a child who has more internal control,

36

is more intrinsically motivated, is freer from some constraints of reality, and is
better able to give and receive interactional cues than a child who is less playful.
Rogers and his colleagues (1998) found that more playful children from
preschool age to fourth grade appear to be more approachable, adaptable, and
persistent. Arabasz (2000) found that playful children between the ages of 41
and 131 months typically show flexibility, which leads to more creative problem
solving, and increased adaptability.

Similarly, the temperament of playful children is characterized by a
positive mood, persistence, adaptability, and approachability (Rogers et al.,
1998). Children between ages of three to six years who are more playful are
also more likely to think divergently as well as creatively (Barnett 8 Kleiber,
1982), and show elaboration in play (Barnett 8 Kleiber, 1984). In addition, it has
been revealed that playfulness predicts prosocial behavior, exploratory behavior,
variation-seeking, competence, and coping. Also, the playful children showed
flexibility, leadership, and elaboration in play (Barnett 8 Kleiber, 1984; Knox,
1996; Singer 8 Rummo, 1973; Singer et al., 1980). Interestingly, the playfulness
trait and its relationship with other developmental features has been supported at
later age levels, as Lieberman proposed. In conclusion, children’s playfulness
can be regarded as a function of a generally positive disposition that the child
should possess in order to adapt to and have an impact on her environment
(Singer 8 Singer, 1976).

Because of the relationships of playfulness to positive developmental

outcomes, many studies focused on the usefulness of playfulness among

37

children with special needs. In general, playfulness in children with disabilities
such as developmental delay or cerebral palsy is scored lower than playfulness
in children who are normally developing. Play training for the children with
disabilities has been widely implemented in the occupational therapy field.
Intervention programs, such as play training or environmental support programs
for play have been effective in increasing playfulness scores of infants and
toddlers with disabilities (Okimoto et al., 1999; Lane 8 Mistrett, 2000)

To summarize, the research on playfulness has mostly focused on
conceptualizing what is playfulness and developing instruments based on the
concept of playfulness. In addition, the studies, which provided the evidence
needed to validate the developed measures and reveal the important roles of
playfulness in child development, were implemented. However, there is little
research about the ecological factors influencing children’s playfulness at
multiple levels of the environment. Therefore, in addition to the awareness of the
benefits of playfulness, which is the initial step of studying playfulness, it is also
important to recognize the impact of ecological factors, such as that of family and
peers, community institutions and organizations, such as schools, and inventions
and creations, such as television, as environmental elements which influence

playfulness (Aguilar, 1985).

Effects of Viewing on Imaginative Play
Effects of Viewing on the Quantity of Imaginative Play

Previous research on the association between children’s television

38

viewing and their imaginative play examined the roles of television. The research
that investigated influences of television on such play followed the tradition of
communication research that focuses on the "effects approach” of television on
development (Anderson et al., 2001). Past studies on how children’s imaginative
play is affected by television viewing mainly examine whether television viewing
increases or decreases the amount of their imaginative play. There are two
contradictory hypotheses speculating the influences of television on the
frequency of imaginative play. The two perspectives are the stimulation
hypothesis and the reduction hypothesis (van der Voort 8 Valkenburg, 1994).

According to the stimulation hypothesis that assumes the positive effects
of television watching, it enables children to get involved in imaginative play more
frequently by providing a variety of information sources for children to use in
imaginative play. Young children have difficulty playing in creative forms
because they are not able to mentally represent something without adequate
information (Smilansky, 1968). For this reason, they imitate the entity and
behaviors of what they have seen or experienced previously during their play.
Television, which is an engrossing, ready-made story-telling device, provides
models and storylines that can be easily imitated in children’s imaginative play.
Consequently, television is assumed to promote imaginative play by providing a
rich source of ideas from which children can draw when engaged in imaginative
play (Greenfield et al., 1990; Huston 8 Wright, 1998b).

The other hypothesis, the reduction hypothesis, proposes that television

viewing decreases the amount of imaginative play. According to van der Voort

39

and Valkenburg (1994), there are several assumptions for explaining this
hypothesis. The displacement assumption insists that television viewing may
become an easy alternative to children’s imaginative play because it provides
self-entertainment for children, and eventually displaces the time to play (Singer
8 Singer, 1986; Winn, 1985). Structural characteristics of television as a medium
also provide several possible reasons for the reduction of imaginative play.
According to the passivity assumption, frequent television viewing may lead the
young viewers to a passive attitude that undermines willingness or ability to
engage in imaginative play, because television, an easy medium requires little
mental effort to produce the children’s own imaginations (Singer et al., 1984a).
The rapid pacing assumption suggests that the rapid procession of television
programs allows little time for children to process what they watch and leads to
cognitive overload and impulsive thinking (Anderson et al., 2001). Consequently,
children are less likely to engage in imaginative play. The arousal assumption
also attributes the reduction of imaginative play to hyperactivity and impulsive
behaviors that may be caused by watching television because watching
television with the format of rapid and sophisticated visual pacing techniques has
arousing effects on the viewers (Singer et al., 1984b; Valkenburg 8 van der
Voort, 1994). To conclude, the frequency of imaginative play episodes may be
reduced by frequent consumption of television for one or more of these reasons.
According to several studies on the effects of television viewing on
imaginative play, few relationships between the amount of time spent in viewing

television and the frequency of imaginative play of preschoolers were found

40

(Gadberry, 1980; Singer 8 Singer, 1976, 1981). The reasons for insignificant
associations can be that only the overall amount of viewing time was considered
as a predictor of imaginative play in the early research, since recent publications
indicate that there are significant relationships between watching specific types of
programs and children’s imaginative play. In general, a decrease of children’s
imaginative play was associated with television viewing of programs with a high
level of violence and action. That is, children who saw television with low
violence or who saw no television, showed an increase in imaginative play; those
who saw programs with high action and violence declined in imaginative play in
various age groups preschoolers to school age children (Huston-Stein et al.,
1981; Singer 8 Singer, 1976, 1981; van der Voort 8 Valkenburg, 1994), although
several experimental studies have shown insignificant effects in children ranging
in age between 3 to 6 years (Potts et al., 1986). On the other hand, the
programs that were nonviolent or informative, or programs with low action and
fantasy did not affect the frequency of children’s imaginative play (Greer et al.,
1982; Huston et al., 1999; Huston-Stein et al., 1981). Interestingly, a program for
children specifically designed to stimulate imagination, such as Mr. Rogers’
Neighborhood, increased imaginative play of preschoolers (Friedrich-Cofer et al.,
1979; Singer 8 Singer, 1976; Tower et al., 1979).

According to the conclusion of van der Voort and Valkenburg (1994) who
reviewed the studies on television’s impacts on imaginative play, there is little
evidence that television viewing in general stimulates or reduces children’s

imaginative play. However, there are strong indications that the influence of

41

television on the frequency of imaginative play is primarily dependent on the
specific programs such as well-designed educational programs for children. In
other words, it is the kind of programs watched, not the presence or absence of
television that is crucial to imaginative play.

Effects of Viewing on the Quality of Imaginative Play

Although television supplies a vast array of stories and characters that
may stimulate fantasy and imagination, it has been argued that much of what is
generated is with low quality represented by stereotypes and constrained by the
formulaic content of entertainment television (Huston 8 Wright, 1998b). When
high quality play is supposed to provide opportunities for children to develop their
understanding of how the world works (Kieff 8 Casbergue, 2000; Rubin et al.,
1983), more imaginative and creative play script that has children concentrate on
ideas and make connections between ideas is an indicator of good quality of play
(Fromberg, 2002). In light of this issue, the quality of the imaginative play
triggered by watching television does not seem to be high (Johnson et al, 1999).
While a number of quantitative studies have been implemented focusing on the
extent of imaginative play affected by television viewing, there is little qualitative
research on the effects of television on the quality of imaginative play.

The ethnographic studies were likely to focus on the content rather than
the frequency of the play; in general, the studies provide a number of anecdotal
examples that indicate that children’s imaginative play is associated with what
they watch. James and McCain (1982) observed preschoolers’ play episodes in

order to identify patterns, themes and uses of television in children’s play. They

42

found that children’s play that is related to television involved frequent uses of
fantasy superheroes. French and Pena (1991) implemented a retrospective
study using interviews with adults and revealed that television generated more
heroic adventure play and more fantasy heroes instead of real-life heroes, as
compared to the pre-television generation. Guddemi (1985) also found that low
viewers at age four did play more overall themes and more non-television related
themes; and also initiated more overall themes. In particular, non-television
themes had a much wider range of flexibility as to what plans could occur and as
to what role or identity the character could take. Television themes were much
less flexible in terms of players, roles of characters, and scripts. Van der Voort
and Valkenburg (1994), however, concluded that television-related play cannot
be determined to be more or less imaginative and creative than play that is not
derived from television after reviewing of a series of ethnographic studies.
Therefore, they suggested that scholars examine the “quality” of imaginative play
that is affected by television viewing.

There are limited numbers of studies that focus on the quality of
imaginative play influenced by television viewing. Singer and Singer’s (1981)
classic research on the systemic pattern between imaginative play and patterns
of television viewing among preschoolers was the exemplary study that implies
the quality of such play. After observation of preschoolers over one year, they
found that children who play imaginatively in the nursery school were somewhat
more likely to be watching what might by termed more “benign” programming and

they did not appear to be watchers of the more active shows or the action-

43

detective shows.

Effects of Viewing on Playfulness

Hogue and Johnson (1999) completed a study comparing the playfulness
of three- to twelve-year olds watching less than 10 hours of television a week
with children watching more than 20 hours per week, using t-test analyses. They
found a significant difference in the two groups” playfulness; the heavy viewers’
playfulness scores on Bundy’s Test of Playfulness (ToP) were significantly lower
than that the scores in the low television group. It substantiates the concern that
television has a deleterious effect on the playfulness of children.

In contrast, Arabasz (2000) who replicated the study of Hogue and
Johnson with thirty children between the ages of 41 and 131 months, using the
ToP to measure the children’s playfulness, compared ToP scores of children
watching 10 hours or less of television each week to the scores of children
watching 20 hours or more of television each week. Interestingly, her results
indicated that children watching more television scored significantly higher on the
playfulness. Because results that were not consistent with Hogue and Johnson’s
study, the groups were compared for age and gender. Neither the means for age
nor gender were statistically different between groups. She concluded that
television might hold some benefit in respect to children’s playfulness by
promoting social skills such as reading and giving appropriate cues, persisting
through difficulties and the improving the ability to skillfully clown and joke.

As reviewed, very little of the literature on the effects of television pertains

44

to playfulness. Only two pieces of literature exist that describe the impact of
television on playfulness and the findings of the two studies were not consistent
with each other. In addition, they had several limitations. First, ecological factors
were not taken into account; even the demographic information of the subjects,
such as race and socioeconomic status, was not considered or reported. The
studies also considered only the overall amount of viewing as an influencer on
playfulness without investigating the effects of different television programming
on playfulness. Therefore, it is necessary to examine the effects of viewing
different programming on children’s playfulness based on an ecological

perspective.

45

CHAPTER 3

METHODOLOGY

This chapter describes the methodology of this study. Five sections are
included: research hypotheses; procedures; participants; measures; and data

analyses.

Research Hypotheses
This study tested the model shown in Figure 3 to determine if the data
were consistent with the model. For this purpose, hypotheses regarding the
proposed relationships among the selected variables were generated based on
previous studies described in the first two chapters. The first set of hypotheses
examines relationships between variables that are proposed in the model.
H 1. The higher the socioeconomic status of a family is, the lower the level of
the family viewing environment.
H 2. The higher the socioeconomic status of a family is, the more the parents of
the family mediate what their child views.
H 3. The higher the level of the family viewing environment is, the more visual
media a child in the family uses.
H 4. Parental mediation has mixed effects on children’s visual media use. That
is, children’s visual media use varies by parental mediation.
H 4.1. The more the parents of the family mediate what their child view,

the more Type 1 programs their child views.

46

H5.

H6.

H7.

H 4.2. The more the parents of the family mediate what their child view,
the less Type 2, Type 3, Type 4, Type 5, and Type 6 programs their child
views.

Child characteristics have mixed effects on children’s visual media use.
That is, children’s visual media use varies by child characteristics.

H 5.1. Younger children spend more time viewing Type 1 programs than
older children; older children spend more time viewing Type 2, Type 3,
Type 4, Type 5, and Type 6 programs than younger children.

H 5.2. Girls spend more time viewing Type 1, Type 2, Type 4, and Type 5
programs than boys; boys spend more time viewing Type 3 and Type 6
programs than girls.

Children’s visual media use has mixed effects on their imaginative
playfulness. That is, children’s imaginative playfulness varies by program
types that children view.

H 6.1. The more Type 1 programs children watch, the more playful they
are.

H 6.2. The more Type 2, Type 3, Type 4, Type 5 and Type 6 programs
children watch, the less playful they are.

Child characteristics have mixed effects on imaginative playfulness. That
is, children’s imaginative playfulness use varies by child characteristics.
H 7.1. Older children are rated higher than younger children in
playfulness including physical spontaneity, social spontaneity, cognitive

spontaneity, manifest joy, and sense of humor in imaginative play.

47

H 7.2. Boys are rated higher than girls in physical spontaneity, manifest
joy, and sense of humor in imaginative play; gins are rated higher than
boys in social and cognitive spontaneity in imaginative play.

In addition to direct effects among the variables, a following hypothesis
was generated to examine the overall explanatory power of the structural model
for children’s visual media use and their imaginative playfulness, integrating all
the variables selected for the present study. It is stated as follows:

H 8. The hypothesized model of the relationship of socioeconomic status,
children’s characteristics, and three mediating variables, including the
viewing environment in a family, parental mediation, and children’s visual
media use, to the level of imaginative playfulness in this study fits the

data.

Procedures

Data were collected from 202 primary caregivers, including parents, foster
parents, stepparents, and/or grandparents of 202 preschool children, and 43
head teachers of the children involved.

The recruiting process consisted of a series of stages. First, local centers
providing programs for three to five-year-olds, were selected in three adjacent
counties in Michigan. The procedure of center selection was accomplished using
the child care center list of the State Department of Consumer and Industry
Services (CIS). In the list, licensed centers for young children that were

regulated by the state government were listed with the information about the age

48

and number of children that they served and the types of program that they
provided (full-day or half-day). Because of the large sample size of this study,
the more children the programs served, the more participants could be recruited.
Accordingly, it was decided to select the larger programs for three, four, or five-
year olds. For this study, therefore, 25 centers, which provided programs for
more than 100 children, were selected as the first tier to be contacted.

Secondly, directors of the selected centers were contacted. This was
completed during the period January 17 through April 3 in 2003. The initial
meeting with directors was intended for gathering the center information that the
CIS list did not include. There were several criteria that the participating centers
should meet. In order for a valid assessment of children’s imaginative
playfulness, the participating children were expected to have free choice time
with an assortment of play options, both indoors and outdoors, based on the
philosophy of their program appreciating children’s imaginative play. For this
reason, the directors were asked if their center provided free choice time in a
daily routine, and if their center was equipped with interest areas, especially a
pretend play area and a block area in the classrooms where imaginative play
episodes often occur. Also, the participating centers were expected not to use
visual media on a regular basis at their centers during a day to control for the
children’s regular exposure to visual media. This was because the children’s
amount of visual media use was collected only at home.

After the initial meetings with the 25 directors and visits to their centers,

there were three centers that were excluded by the author: a Montessori center

49

that did not value imaginative play as a main activity in their philosophy; 3 center
where children watched video on a regular basis during the day; and a center
that does not have a playground for outdoor play.

The directors of the other 22 centers, which were judged to be appropriate
for the research, were asked about their willingness to recruit the families and
head teachers who will be invited to participate in this study. After the meeting
with the directors, seven directors declined to participate in the research.
Ultimately, the directors of 15 centers in two counties granted permission to
recruit research participants at their centers.

Head teachers of the 15 centers whose classes were for three, four and/or
five year-old children were contacted in order to obtain their consent to
participate in this study. The head teachers sent survey packets, which included
parental questionnaires and an envelope for return, to every family with three,
four, or five-year olds.

With the head teachers’ agreement, primary caregivers, including parents
or guardians of all children whose head teacher agreed to participate, were
contacted. The packets of the parental questionnaires, including an overview
and general introduction to the purpose of this research with the set of
questionnaires for primary caregivers, were distributed (see Appendix B, C, D, E,
and F). The primary caregivers who voluntarily completed and returned the
questionnaires in a sealed envelope to their head teacher were considered as
participating families. A total of 1,021 packets were sent to families and 239

families returned their packet to the head teachers. Thus, the response rate of

50

the research was 22.63%. The questionnaires for parents were completed by 10
fathers and one grandmother; the others were done by mothers of participating
preschoolers.

Finally, 43 head teachers whose children participated in this research
collected the returned packets and completed the Imaginative Playfulness Scale

for each child of the participating families (see Appendix G).

Participants

For this study, the participants were 202 primary caregivers of
preschoolers and 43 head teachers. Although 239 packets were originally
returned, 37 of them were not included in this research for one of four reasons.
First, some programs that the participating children watched as reported by
primary caregivers, were either not identified or not available during the data
coding period. Second, some packets had missing items. Because structural
equation modeling (SEM) results can be adversely affected by the presence of
missing values, these had to be discarded. Third, the families whose parents
were full-time students were dropped because students were not coded either in
the employment or unemployment categories in the occupational prestige scale.
Fourth, one response with an outlier was excluded because gross income was
too high ($500,000).

The deletion from the analysis of the outlier was judged based on their
contribution to kurtosis (Bentler, 1989; Raykov et al., 1991). The kurtosis statistic,

which represents the tendency for observations to cluster both tightly around the

51

mean and broadly in the tails (Harris, 1985), was 17.06 with the sample,
including the income outlier. After deleting the case, the kurtosis statistic that
had been skewed to a significant degree became .89. Accordingly, it was
decided to exclude the case with the income outlier.

Table 2 presents a summary of the demographic and background
characteristics of the sample of this study. The mean age of the 202
preschoolers sampled was 52.24 months, with a range from 36 months to 72
months. The sample children consisted of 104 boys (51.5%) and 98 girts
(48.5%). There were 171 European-Americans (84.7%), three African-
Americans (1.5%), six Hispanics (3.0%), 13 Asian Americans (6.4%), and nine
others (4.4%). The sample included 119 first-born children (58.9%), 50 second-
bom children (24.8%), 22 third-born children (10.9%), and 11 fourth-born or later-
bom children (5.5%). Most of the children (92.6%) attended only one early
childhood program; 15 children attended another program such as Kindergarten
or Montessori school. Because visual media uses are controlled at the programs,
the children who attended two programs were included in this study.

The primary caregivers of this study were composed of 202 mothers
between the ages of 20 and 52, with a mean of 35 years. The number of years
of mothers’ education in the sample ranged from 9 to 24 years, with a mean of
16.50 years. One hundred forty-seven of the 202 mothers (72.6%) were
employed full-time or part-time. Their mean income was $27,407.92 (SD =
$26,506.73), and the average occupational prestige was 61.28 (SD = 21.51)

whose possible points were from one to 96.

52

Table 2.

Characteristics of Participants

 

 

 

 

 

 

 

 

 

Characteristics Percentage Mean SD
Child (N= 202)
Age (in months) 52.22 8.03
Sex (Male) 51.5
Ethnicity (European-American) 84.6
(African-American) 1 .6
(Hispanic) 3.0
(Asian) 6.4
(Others) 4.4
Birth order (First born) 58.9
(Second born) 24.8
(Third born) 10.9
(Fourth born or greater) 5.4
First language (English) 93.2
Attending programs (None) 92.7
Mother (N= 202)
Age (in years) 35.00 5.89
Educational level (in years) 16.50 2.94
Employed mother (including part-time) 72.6
Ea_ther (N= 181)
Age (in years) 37.07 5.97
Educational level (in years) 16.55 3.46
Employed father (including part-time) 98.3
Family Characteristics (N= 202)
Annual income (Dollars) 78,719.31 45,540.35
Occupational level 64.95 22.76
Married status (Married) 82.7
(Divorced) 4.5
(Separated) 1 .0
(Never married) 7.4
(Others) 4.3
Number of adults in the family 1.97 .37
Number of children in the family 2.08 1.06

 

53

The demographic information for 181 fathers out of 202 participating
families was reported (89.6%), because mothers of the separated and divorced
families did not report the information. The fathers were between 24 and 55
years of age, with a mean of 37.07 years. On average, they achieved 16.55
years of formal education, ranging from 8 to 26 years. Their average income
was $61,790.12 (SD = $39,786.60), and the mean occupational prestige was
65.71 (SD = 23.48).

About 167 participating families (82.7%) included married couples; the
sample contained only nine divorced families (4.5%), two separated families
(1.0%), 15 families whose mother never married (7.4%), and nine families that

did not belong to anyone of them (4.4%) such as engaged couples.

Measures
Socioeconomic Status
The child’s family background information, including socioeconomic
status, was obtained by the Family Background Questionnaire. The
questionnaire included questions about the level of income, occupation, and
education of participating parents. Primary caregivers of the participating
children reported their annual income, years of education, and the occupation of
themselves and their spouses. The items regarding income and education were
adapted from the questionnaires of the Panel Study of Income Dynamics (PSID)

implemented by the Institute for Social Research at the University of Michigan

(Hofferth et al., 1998).

The sum of the annual incomes of the parents was coded as family
income. The years of education of the family head were coded as educational
level of families. The occupational status of a family head was coded using the
Socioeconomic Index (Duncan, 1961), which assesses the general standing of
occupations to evaluate occupational prestige. Although Duncan’s
Socioeconomic Index is somewhat outdated, it has been used as a
comprehensive scale to quantify occupational status until recent times (Lemelle,
2002). The range of the occupation scale is from one to 96.

As for the family head issue, the employed person was regarded as the
head of the family. In the case of dual income families, a parent who made more
money was regarded as the head of the family. When the income level of
parents was equal, occupational levels of parents were considered for deciding
the head in a family.

Viewing Environment in a Family

In the Family Background Questionnaire, three items regarding viewing
environment in a family were included. Primary caregivers responded to the
items regarding total amount of visual media use at home yesterday, the number
of visual media devices including television sets, videos, and DVD players in their
family, and the number of visual media devices that children own. The items
from the questionnaires of the study of 1997 Child Development Supplement
(CDS) to the Panel Study of Income Dynamics (Hofferth et al., 1998) were

adapted for this study.

55

Parental Mediation
Measure Modification I

Parental Mediation was obtained by the Parental Mediation Questionnaire
designed by Valkenburg and her colleagues (1999). The original Parental
Mediation Questionnaire consisted of 15 items of instructive mediation, restrictive
mediation, and social co-viewing with five questions regarding each mediation
style. The items were scored on a 5-point scale (never = 1, rarely = 2,
sometimes = 3, often = 4, always = 5).

The Valkenburg’s questionnaire was modified by the investigator. As
stated above, the original questionnaires were designed to report parental
mediation of television only. For this study, the items of the measure, which do
not include video and DVD, were adapted to report average viewing of video and
DVD as well as television. For example, the original item of “How often do you
and/or your spouse try to help the child understand what s/he sees on television
was changed to “How often do you and/or your spouse try to help the child
understand what s/he sees on television, video, or DVD?”

The modified instrument for this study was reviewed by a focus group in a
pilot study prior to data collection. The pilot study was done with a convenience
sample of 17 primary caregivers of 3-5 year olds. Most of the parents barely
answered the items of social co-viewing and commented that social co-viewing
did not occur with their preschooler. The examples of social co-viewing
questions were: “How often do you and/or your spouse watch together because

you both like the same television, videos, or DVDs? How often do you and/or

56

your spouse watch together because of a common interest in the television
programs, videos, or DVDs?” These results could be due to the fact that the
original measure was developed for children between five and 12 years old.
Social co-viewing that implies children as partners of parents’ viewing might not
be plausible for preschoolers who are much younger. Accordingly, another
modification made was to exclude the social co-viewing items. Therefore, only
the 10 items of instructive mediation and restrictive mediation were rated by the
children’s primary caregiver in this study.

Measure Modification II

Exploratory Factor Analysis (EFA) using Principal Component Analysis
was performed to evaluate the parent mediation measure by providing explicit
measurement errors through uncovering the underlying structure of the sets of
variables. Because the original measurement of Valkenburg and her colleague
(1999) was modified to assess only two constructs out of three, EFA was
necessary prior to the main analysis of the hypothesized model.

As expected, two components were extracted; there were two eigen
values greater than one (F1= 3.90, F2 = 1.84). After rotation using the Varimax
with Kaiser Normalization to maximize the variance of the loadings, it was found
that item six of the measure had weak loadings on both factors (F1 = .36, F2 =
.30), while the other items primarily loaded on one factor with high loadings from
.68 to .82 (See Table 3). Although the meaning of the factor loading magnitudes
varies by the research context, factor loadings of the item were marginally on the

minimum cut-off of loadings, so they could be termed as weak (Norman 8

57

Streiner, 1994). In fact, item six “How often do you and/or your spouse ask your
child to turn off television, video, or DVD when s/he is watching an unsuitable
program?” was not answered by nine parents; it was the item that was most
frequently left blank. Additional comments stated that they never turned off any
visual media device because of inappropriate content. Therefore, item six was
not included for the later analyses.

Cronbach’s alpha coefficients for the nine items of the modified measure
for this sample was .80; Cronbach’s alpha coefficient for the five instructive
mediation items was .86 and Cronbach’s alpha coefficient for the four restrictive
mediation items was .76. In this study, the sum of the indicators of each factor
was calculated and used in subsequent analyses.

Children ’3 Visual Media Use
Procedure

Primary caregivers reported television shows, videos, and DVDs watched
by their children at home during the week previous to the reporting day, on the
Visual Media Use Questionnaire. They tracked the name and frequency of each
individual program being watched in the report in order to consider content as

well as frequency of a program that children watched.

58

Table 3.

Factor Loadings of Parental Mediation Indicators

 

 

 

Theoretical Factor Loading
Item
Component
F1 F2

Instructive How often do you and/or your spouse point .82
Mediation out why some things actors do on

television, video, or DVD are good? (2)
Instructive How often do you and/or your spouse point .81
Mediation out why some things actors do on

television, video, or DVD are bad? (3)
Instructive How often do you and/or your spouse try to .78
Mediation help the child understand what s/he sees

on television, video, or DVD? (1)
Instructive How often do you and/or your spouse .78
Mediation explain what something on television,

video, or DVD really means? (5)
Instructive How often do you and/or your spouse .73
Mediation explain the motives of the characters on

television, video, or DVD? (4)
Restrictive How often do you and/or your spouse set .79
Mediation specific viewing hours for your child? (7)
Restrictive How often do you and/or your spouse .78
Mediation restrict the amount of child viewing? (9)
Restrictive How often do you and/or your spouse .77
Mediation specify in advance the television programs,

videos, or DVDs that may be watched? (10)
Restrictive How often do you and/or your spouse forbid .68
Mediation your child to watch certain programs? (8)

 

Note. 1)The number of parentheses is the item number on the Parental

Mediation Questionnaire. 2) F1: Instructive Mediation, F2: Restrictive Mediation.

59

A pilot study was also done with the convenience sample of 17 primary
caregivers to test the measure. To measure the total amount of children’s
viewing per program, the participants who completed the pilot Visual Media Use
Questionnaire commented that reporting frequency of viewing per program would
be more convenient and accurate than reporting actual time of viewing per
program. Also, they suggested that ten rows to report programs on the
questionnaire that their child watch during a week would be enough.

After the revision based on the comments of the pilot sample, the Visual
Media Use Questionnaire was distributed and returned. One hundred and sixty
two television shows and 340 videos/DVDs were reported by the participants of
this study. The viewing reports were tabulated initially according to the number
of minutes each child watched each program; the total minutes of watching the
individual programs were computed by multiplying the frequency and the running
time of each program. At the same time, each program was also coded by
dimensions of the coding schemes (see 4.2 in Chapter 4). Eventually the
minutes that children watched during a week were obtained per program
category.

Coding Schemes

There were two rating scales that were adapted to classify the reported
programs in this research. First, a coding system by Anderson and his
colleagues (2001), which has been used for their earlier studies (Center for
Research on Influences of Television on Children, 1983) was adapted. In their

research, there were three dimensions used to categorize all the television

60

programs: (a) Intended Audiences, (b) Animation Components, and (c)
lnforrnative Purposes. Among the three elements, only rating scales of the
Intended Audience and the Informative Purpose were adapted from the
Anderson’s study; the scale of Animation Components was not used in this study.
According to the researchers, they included the animation components in their
research to identify whether a program includes aggression or not, which is a
major factor in this research. As they admitted, however, not every animated
program contained violence and other program types sometimes show violence.
For this reason, it was concluded that animation cannot be the component used
to rate the prevalence and salience of violence.

A second rating scale, named Television and Movie Violence Rating
developed by Huesmann and his colleagues (2003), was adapted as a scale
measuring violence components. Their coding instructions were specifically
developed to rank the degree of violence components on television shows and
movies.

As shown in the Rating Instructions (see Appendix A), all the programs
reported by participants were coded in terms of three dimensions in the present
study as follows: (a) Intended Audiences (child audiences or general audiences),
(b) Violence components (yes or no), and (c) lnforrnative Purposes (yes or no).

Intended Audiences. The operational definition for the Intended
Audiences was whether a program is intended primarily for viewers under 13
years old or not. When a program is determined to be primarily designed for

children 12 years old or younger, the program was classified as a children’s

61

program; other programs were coded as programs for general audiences.

One problem which was found at the first coder training session was rating
so-called family movies, which were produced for all the family members
together. They were appropriate for young children but were not judged to be
designed only for children. Accordingly, the coders were provided with specific
instructions that could be referred to in order to make a decision: the
characteristics of programs, including age of a main actor/actress,
appropriateness of theme, complexity of story, and rating of the Motion Picture
Association of America (MAPP) such as G, PG, or PG13. If the age of a main
actor/actress was under 13, a movie does not include an inappropriate theme for
children such as sexual content, a storyline was not too complex to be
understood by children, and/or a movie is rated G, then the movie was
considered as a program intended for child-audiences.

Violence Components. Analyses of the prevalence and salience of
violence as a feature of programming through assessing content directly has
been done based on a content-analytic perspective (Gunter, 1981). Instead of
rating schemes employed in content analysis studies, a violent rating on a scale
based on the decision of raters was necessary for this study in order to judge
whether a program is a violent one or not. For this reason, the Television and
Movie Violence Rating Instructions of Huesmann and his colleagues (2003) were
adapted.

In the rating scheme, a violent act was conceptually defined as “a physical

act intended to harm another person” (Eron, 1963, 1993; Huesmann 8 Eron,

62

1986). Their rating instruction was based on the established theory that visible
physical violence has much more influence than invisible violence, while all
violence in shows has some influence on the viewer (Huesmann et al., 2003). In
their rating scheme, a violent act was operationally defined as “an explicit act that
was committed intentionally and interpersonally to harm others”. Based on the
lengths and frequencies of episodes with violence in a show, the coders rated the
degree of aggression in a program.

Although their rating scheme was useful to judge the degree of violence in
each program to categorize a show into a violent show or nonviolent show, it was
found that more details of operational definitions were needed after peer
reviewing and coder training sessions. Because some acts were not clear such
as contact sports, theft, chasing, and scaring, the rating scheme of Greenberg
(1980) was used to supplement the Huesmann’s measure. Contact sports in
dramatic plots, which are not basically intended to harm others despite physical
contact in a common and tolerated way, was not coded as a violent act; theft was
also not coded as an aggressive one, because it is not a violent act but an
antisocial behavior that harms others. Chasing and soaring, which are visible,
intentional, and interpersonal behaviors, were rated as violent acts.

Another adaptation of Huesmann’s measure was his response categories.
The original scale was a 5-point scale (0 = No violence, visible or invisible, or
perhaps a very little invisible violence, 1 = A very little visible violence or a little to
medium invisible violence, 2 = A lot of invisible or implied violence with only a

little visible violence, 3 = Medium to high visible violence, 4 = High visible

63

violence). Based on feedback from coders that the term ‘invisible’ violence in the
scales was confusing and inconsistent with the operational definition of violence
in the present study, the original scales were simplified. Therefore, the violence
components measure in this study included a 4-point scale (1 = no violence, 2 =
a very little violence, 3 = medium violence, 4 = high violence).

Informative Purposes. Based on operational definitions of Anderson and
his colleagues (2001), programs primarily designed to be educational or
informative were rated as informative programs. The decision was based on the
apparent intention of the producer.

At the first coder training and reliability session, it was found that
educational components in children’s programs were not clearly differentiated
from social morality messages delivering a message, which encourages good
and punishes evils typically in fairy tales. For this reason, specific coding
schemes were added based on components of educational shows that Singer
and Singer (1998) presented. The content of children’s informative programs
should be designed to provide information, knowledge, or cognitive skills; to help
children to develop constructive prosocial attitudes or actions such as sharing,
turn taking, and self-esteem; to expose children to the varieties of culturally and
physically different types of children and adults and fostering a sense of mutual
respect for these cultural and physical differences like ethnic diversity, disability,
and single parent family; or to encourage awareness of one’s own and others’
emotional responses.

In addition to coding instructions based on Anderson and his colleagues

64

and Singers’ schemes, another direction about messages regarding religious
beliefs was added in the coding scheme of this study. This direction was
included because there were several reported programs which were based on
Christianity designed in an informative manner. Religious programs were
regarded as informative only if they included non-religious messages that
promote children’s socially constructive attitudes. For example, a message that
Christmas is not all about presents but about love for others was coded as
informative. On the other hand, a message that God has a plan for you was
regarded as non-infonnative.
Reliability

Coding was done by two graduate students who majored in natural
science and telecommunication. The coders had a one-hour training session to
be introduced to the research and the coding instructions. After the instruction
session, ten programs were randomly picked and used for the first intercoder
reliability training. The two coders and the author discussed the difficulties in
coding family movies according to their intended audiences, and scary episodes,
antisocial episodes, and chasing episodes according to their violence
components. Accordingly, there was further clarification of the coding schemes.

The second training session and a pilot reliability test with the revised
coding schemes took place with another 17 programs. At that point, satisfactory
intercoder reliability was not produced (Intended Audiences = .71, Violence
Components = .62, and Informative Purposes = .79). After the testing, religious

movies according to their informative purposes and shows of contact sports

65

according to their violence components were discussed and further clarification
was added in the coding schemes.

The main intercoder reliability test was performed with 49 shows randomly
selected from the program list reported by participating parents, because
approximately 10 percent of the population is recommended for the reliability test
(Riffe et al., 1998). In order to exclude programs for the reliability test from the
shows for the main analyses, an episode of the selected programs was used for
the reliability test in the case of television shows; for the test of videos and DVDs,
a sequel of either movies or shows from the same series, or a movie made by
the same production company of the selected programs, were randomly selected
as long as it was not included on the program list.

The researcher conducted the intercoder reliability tests, using the
Program for Reliability Assessment with Multiple Coders (PRAM), a Windows —
based application for the PC, designed to simplify the calculation of intercoder
reliability coefficients (Neuendorf, 2002). The reliability per dimension was
calculated respectively. For Intended Audiences and lnforrnative Purposes
dimensions, which were dichotomous, percent agreements were obtained; for
Violence Components dimension, on the other hand, Pearson correlation was
calculated because the scores were done with 4-point scale. The final intercoder
reliability on each dimension were obtained as follows: Intended Audiences, .93,

Violence Components, .91, and lnforrnative Purposes, .96 for the 49 programs.

66

Coding

After a satisfactory interrater reliability was obtained, 502 programs that
participants reported were reviewed and coded by the two coders. Each coder
analyzed 251 programs. In the case of television shows, a randomly selected
episode from May 18 to June 21, 2003 was analyzed. When a show had more
than one independent episode in one segment, only the first episode was coded
for this study. Title rolls and commercials during a show were not included in the
analysis. The running time of television programs were obtained from TV Guide.
in the case of videos and DVDs, the reported programs were reviewed and the
official running time on each tape or DVD was used. The content and amount of
the previews at the beginning of tapes and DVDs were not considered in this
research.

After each of the reported programs was categorized, the total number of
minutes was calculated by categories in order to estimate the amount of
children’s visual media use by summing the running minutes of programs by
categories. Based on these three dimensions of the coding schemes, all
programs reported were originally classified into 16 different categories.

Table 4 summarizes the number of minutes viewed by program
dimensions, as reported on Visual Media Use Questionnaire by primary
caregivers during the one-week sampling period. Total time spent on viewing per
child was 951 minutes (SD = 622.40) during the sampling week watching
television and videos/DVDs; the children spent 500.03 minutes (SD = 349.59) on

viewing television, and total time spent on viewing videos or DVDs per child was

67

449.68 minutes (SD = 408.57).

Child audience/No violence/Informative programs were viewed most
frequently (387.52 minutes). Most programs of this category were television
shows (television = 80.8%, video/DVD = 19.2%); virtually, all Public Broadcasting
System (PBS) shows for children were coded here (see Table 5). Children spent
42.4% of their total viewing time watching programs of this type on television,
videos and DVDs.

The second most frequently viewed type was child audience/medium
violence/ non-informative programs. Children watched 227.27 minutes and spent
22.6% of their total viewing minutes with these programs; in actual minutes,
children watched 215.17 minutes of videos/DVDs and 12.10 minutes of television
in this category of programs. Because Disney movies which were the favorite
programs of young children were categorized as child audience/medium
violence/non-informative programs (see Table 5), videos and DVDs were the
majority of this category’s programs.

Following in descending hours viewed was child audience/little
violence/non-informative programs (122.80 minutes). This type included a
variety of children’s television shows and movies. Children watched almost the
same amount of television shows (61.20 minutes) as they watched movies of this

type (61.60 minutes).

68

Table 4.

Average Viewing Amount by Dimensions and Visual Media Type (N = 202)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Visual Child Audiences General Audiences

Media

Type Informative Entertaining lnforrnative Entertaining
No Violence

Television 313.3 (267.5) 29.1 (74.5) 26.1 (82.9) 22.9 (63.3)

Video/DVD 74.2 (157.5) 35.4 (79.1) 2.0 (20.3) 9.1 (38.3)

Total 387.5 (327.6) 64.5 (112.3) 28.2 (85.1) 32.0 (81.0)
Little Violence

Television 8.8 (26.6) 61.2 (129.4) .0 (.0) 2.1 (15.1)

Video/DVD .7 (5.6) 61.6 (96.7) .0 (.0) 5.4 (37.2)

Total 9.4 (28.2) 122.8 (166.6) .0 (.0) 7.4 (42.5)

Medium Violence

Television .0 (.0) 12.1 (48.1) .0 (.0) .0 (.0)

Video/DVD 1.9 (17.5) 215.2 (260.5) .0 (.0) 12.1 (49.7)

Total 1.9 (17.5) 227.3 (268.6) .0 (.0) 12.1 (49.7)
High Violence

Television .0 (.0) 24.5 (61.7) .0 (.0) .9 (9.4)

Video/DVD .0 (.0) 21.1 (63.8) .0 (.0) 11.0 (48.5)

Total .0 (.0) 45.6 (101.8) .0 (.0) 11.9 (49.2)

 

Note. 1) The unit of the time spent on the viewing activity is minute, 2) Values
enclosed in parentheses represent standard deviation.

69

Table 5.

Program Titles by Dimension and Visual Media Type

 

Program

 

DimenSIons Type Televusuon Video/DVD
Child audience/ 1 Angelina Ballerina, Berenstain Baby Bach, Kids Yoga,
No violence/ Bears, Barney and Friends, Book of Schoolhouse Rock, There Goes
lnfonnative Pooh, Dragon Tales, Madeline, Out a Bulldozer, Arthur Series,
of the Box, Stanely. Teletubbies, Blue's Clue Story Time,
Wiggles Berenstain Bears Series
Child audience/ 1 Liberty’s Kids, Anne of Green American Legends, Peter and
Little violence/ Gables, Seven Little Monsters the Wolf
Non-informative
Child audience/ N/A Care Bears Movie
Medium violence/
Non-informative
Child audience/ 2 As told by Ginger, Cousin Skeeter, Country Bears, Fantasia, Tigger
No violence/ Hamtaro, Lizzie McGuire, Recess, movie, Willy Wonka and the
Non-informative Rugrats Chocolate Factory
Child audience/ 2 Amanda Show, Doug, Hey Arnold! Barbie as Rapunzel, Beethoven,
Little violence/ Jimmy Neutron, Johnny Bravo, E.T., Fly Away Home, Free
Non-informative Scooby-doo, Sponge Bob Square Willy, Muppet Show, Mary
Pants, Yu Gi Oh Poppins, Stuart Little, Toy Story,
Wizard of Oz
Child audience/ 3 Acme Hour, Aladdin, Bugs Bunny. 101 Dalmatians, Aladdin,
Medium violence! Kim Possible, Transformers Atlantis, Barbie in the
Non-informative Armada Nutcracker, Beauty & the Beast,
Harry Potter Series, Spy Kids
Child audience/ 3 Buzz Lightyears, Digimon, Dragon Batman Beyond, Peter Pan ll,
High violence/ Ball 2, Justice League, Pokemon, Power Rangers, Powerpuff Gins
Non-informative Power Rangers, Powerpuff Girls
General audience/ 4 Antiques Roadshow, Emeril Live Crocodile Hunter
No violence/ Changing Rooms, Jack Hannah’s
lnforrnative Animal Adventures, Jeopardy
General audience/ 5 Brady Bunch, Cosby Show, Fear Billy Elliot, My Big Fat Greek
No violence/ Factor, King of the Hill, Wheel of Weeding, Shallow Hal
Non-informative Fortune, Will & Grace, Star Treck

General audience/ 5 Boy Meets World, Cops, Diagnosis Knight's tale, Sound of Music
Little violence/ Murder
Non—infon'native

General audience/ 6 Back to the Future, Cats and

Medium violence! dogs, Chicken Run, Ghost

Non-informative Busters, Home Alone, Master of
Disguise, Matilda

General audience/ 6 Malcolm in the Middle, WWE Charlie's Angels, Jurassic Park

High violence/ Wrestling series, Men in Black, Mouse

Non-informative Hunt, Mummy, Star Wars series

 

70

As shown in Table 4, children watched a relatively small amount of the
other types of programs. In particular, there was no program which was coded
as the following types: child audience/high violence/informative programs,
general audience/little violence/informative programs, general audience/medium
violence/informative programs, and general audience/high violence/informative
programs. Additionally, child audience/medium violence/informative television
shows and general audience/medium violence/non-informative television shows
were not viewed by the participating children. In Table 5, examples of each
program type are reported.

Categorization

For most analyses in previous studies on children’s viewing practices, the
coding dimensions were used to generate mutually exclusive content categories
(Anderson et al., 2001; Huston et al., 1990, 1999; Pinon et al., 1989; St. Peters et
al., 1991; Wright et al., 2001). In the present study, the large number of initial
categories in Table 4 was abridged to generate simpler groups of programs that
the children watched.

One grouping decision involved the placement of programs into one of
only two categories: violent or nonviolent. In this study, the programs rated as
‘no violence’ and ‘little violence’ were combined as ‘nonviolent programs’
although the programs with little aggressive episodes were conceptually
programs containing violence; the programs rated as ‘medium violence’ and ‘high
violence’ were coded as ‘violent programs’.

There were two main reasons to combine categories of the violence

71

component. First, the episodes on the programs with little violence were too
short and infrequent to be retrieved in children’s imaginative play. Although both
programs without violent episodes and programs with little violence were
conceptually different, it did not seem to be reasonable to differentiate the two
types of programs. Second, the decision about whether a program is “nonviolent”
or ‘little violent” was dependent on the episodes that were chosen. In other
words, violent segments in the coded programs with little violence seemed to be
included by chance in the case of television series.

As reported above, children had no viewing in children’s informative
programs with high violence, general informative programs with little violence,
general informative programs with medium violence, and general informative
programs with high violence. Accordingly, these four types were not categorized
for analysis in this study.

Besides these four types, child audience/medium violence/informative
programs were excluded for categorization also. The Care Bear Movie was the
only program that was intended for child audiences programs and contained a
medium level of violent episodes. Only three children viewed this program; the
average amount of watching that type of program was very little (1.86 minutes)
with a large standard deviation (17.45 minutes). Because of the small number of
programs that were reported and the small number of children who viewed the
program, the effects of viewing the movie was unlikely to be powerful and
representative of a single category of visual programs. Therefore, child

informative programs with medium violence were not analyzed for hypothetical

72

models using SEM.
Thus, a total six mutually exclusive categories of programs were identified
for further analyses in this study. Table 6 summarizes the types of programs and

the numbers of minutes viewed by program types.

Table 6.

Categories and Viewing Amount of Visual Media Uses by Dimensions

 

 

 

 

Violence Child Audiences General Audiences
Level Informative Non-informative Informative Non-informative
_ Type 1 Type 2 Type 4 Type 5
Nonvrolent _ _ _ _
Programs (M - 396.94, (M - 187.28, (M — 28.18, (M - 39.44,
SD = 335.60) SD = 223.91) SD = 85.07) SD = 101.42)
. Type 3 Type 6
Violent _ _
Programs (M - 272.88, (M — 24.02,
SD = 304.37) SD = 76.85)

 

Note. The unit of the time spent on the viewing activity is minutes.

Imaginative playfulness
Original Measure: Children ’s Playfulness Scale
The Children’s Playfulness Scale (CPS) developed by Barnett (1990,
1991a) was modified to measure imaginative playfulness in this study. The CPS
consisted of five dimensions as Lieberman originally found: physical spontaneity,
social spontaneity, cognitive spontaneity, manifest joy, and sense of humor.

Because the CPS was designed to measure the overall characteristics of

73

playfulness as a whole and each dimension of playfulness, it is useful to identify
the specific influences of television on different dimensions of playfulness as well
as playfulness itself. There are 23 items which were scored on a 5-point scale
(doesn’t sound at all like the child = 1, sounds a little like the child = 2, sounds
somewhat like the child = 3, sounds a lot like the child = 4, sounds exactly like
the child = 5) by the children’s head teacher.

Reliability and validity for the original CPS have been determined (Barnett,
1990). The reliability of CPS was acceptably high with Cronbach's alpha
coefficients representing internal consistency ranging from .70 to .89, depending
on the child care settings that the participating child attended, and test-retest
reliabilities ranged from .89 to .95 depending on test sessions. Construct validity
was assessed by examining correlations with other play measures including
Lieberman's Playfulness Scale (Lieberman, 1965, 1977), Child Behaviors
Inventory of Playfulness (CBI; Rogers et al., 1987), and Rubin’s Play Observation
Scale (Rubin, 1977, 1982). All comparisons of CBI with other play measures
were positive and statistically significant; correlations of CPS with Lieberman’s
Playfulness Scale ranged from .51 to 82; correlations with the CBI ranged from
.29 to .55, and correlations with the Play Observation Scale ranged from .38 to
.81, depending on sub-dimensions of the CPS.
Measure Modification I

There were several modifications made on the original CPS by the
investigator for the purpose of this study. First, the CPS was designed to

evaluate children’s playfulness in general, so adaptation of the measure to

74

indicate imaginative playfulness was necessary. For this reason, the original
items were specified to evaluate the playfulness of imaginative play. Pretend
play, which is an interchangeable term with imaginative play, was used on the
imaginative playfulness scale of this study, because pretend play was more
easily understood in the pilot study. For example, the item of “The child is
physically active during play” was changed to “The child is physically active
during pretend play.”

Second, an item regarding children’s playfulness in general was also
added to get information about the general tendency of children to be playful (see
item 24 of the questionnaire in Appendix G). Therefore, there were 24 items with
scores on a 5-point scale (doesn’t describe the child at all = 1, describes the child
a little = 2, describes the child somewhat = 3, describes the child a lot = 4,
describes the child exactly = 5), which were rated by the children’s head
teachers.

Because the original CPS was modified for this study, a pilot study was
done to refine the instrument packet prior to data collection. A focus group,
composed of student teachers majoring in child development (n = 27), provided
written and verbal feedback regarding the modified CPS. They also reported that
pretend play is an easier term to understand. In addition, directions for the
questionnaire were modified based on their comments.

Measure Modification II
Exploratory Factor Analysis (EFA) using Principal Component Analysis

was performed to determine the number of underlying dimensions. Like EFA of

75

the parental mediation measurement above, this analysis was necessary for this
present study because the original measurement, that is, CPS assessing
children’s playfulness in general, was modified for this study to measure
imaginative playfulness.

The results found in the original CPS identified the same five underlying
variables but sets of items that were correlated with each other were a little bit
different from the original CPS.

As expected, the results of EFA of imaginative playfulness revealed that
the number of clusters of indicators was the same as the original CPS: five
groups were found out of 23 items of imaginative playfulness. According to
Garson (n.d.), a rule-of-thumb for factor loadings to be considered weak is less
than .4, strong as more than .6, and otherwise as moderate, although the
meaning of the factor loading magnitudes can vary by research context. Based
on his proposition, there were four items that did not load high on any factor,
although they loaded moderately on at least one. Those items were items 1, 6,
13, and 18. Item 1 “The child’s movements are generally well-coordinated and
playful during pretend play" which was an indicator of physical spontaneity, was
moderately cross-loaded on the social spontaneity (.51) and cognitive
spontaneity (.47) factors. Item 6 “The child initiates pretend play with others”
which was an indicator of social spontaneity, was moderately cross-loaded on
social spontaneity (.48) and manifest joy (.42). Item 13 “The child stays with one
activity rather than changing activities during pretend play” which was an

indicator of cognitive spontaneity, showed a moderate load on physical

76

spontaneity (.51). Item 18 “The child sings and talks during pretend play” which
was an indicator of manifest joy, was moderately cross-loaded on cognitive
spontaneity (.41) as well as manifest joy (.48). The results for indicators with
moderate and weak loadings show that the items do not measure the factors that
they are designed to indicate. Consequently, it was decided to remove the items
that did not have a high factor loading for further analyses.

Additionally, there was an indicator that loaded on another factor that it
was not expected to represent. Item 9 “T he child assumes a leadership role
when playing with others during pretend play” which was an indicator of social
spontaneity, was highly loaded on cognitive spontaneity (.80). It was
conceptually understandable that children’s leadership can be related to cognitive
spontaneity as well as social playfulness in the context of imaginative play,
because leadership can be derived from creative and unconventional play
themes as well as social ability to lead during imaginative play. Accordingly, it
was decided to consider item 9 as an indicator of cognitive spontaneity.

In addition to the empirical criteria, the conceptual clarity of the modified
model was also supportive of a five-factor model. Thus, the measure of
imaginative playfulness for this study included 19 items that indicate five factors:
three items for physical spontaneity, three items for social spontaneity, four items
for cognitive spontaneity, four items for manifest joy, and five items for sense of
humor. Table 7 summarizes the components and their loadings of the 19 items.
Cronbach’s alpha coefficient representing the internal consistency of 23 items

was .93; Cronbach’s alpha coefficient for the physical spontaneity items was .90;

77

Cronbach’s alpha coefficient for the social spontaneity items was .89; Cronbach’s
alpha coefficient for the cognitive spontaneity items was .89; Cronbach’s alpha
coefficient for the manifest joy items was .86; and Cronbach’s alpha coefficient
for the sense of humor items was .90. This is generally considered a very high
level. In this study, the sum of indicators of each factor was calculated and used

for further analyses.

Data Analyses
Preliminary Analyses

The data in this study were analyzed in two steps. First, preliminary
statistics were obtained using the Statistical Package of the Social Science
(SPSS). Descriptive statistics including the mean, standard deviation, skewness,
and range were calculated for each of the 21 variables. Correlation matrices
were also computed to examine the extent of associations among the indicators.

Principle component analyses were also performed to test the
appropriateness of two measurement models containing a latent variable, using:
(a) the measurement model of parental mediation indicated by instructive
mediation and restrictive mediation, and (b) the measurement model of the

imaginative playfulness indicated by its five sub-dimensions. These analyses

were to ensure that the constructs and their indicators are acceptable.

 

Table 7.

Factor Loadings for the Imaginative Playfulness Indicators

 

 

 

Theoretical Factor Loading
Item
Component F1 F2 F3 F4 F5
Sense of Humor The child tells funny stories during pretend play (21) .82
Sense of Humor The child laughs at humorous stories during pretend 77
play (2)
Sense of Humor The child gently teases others during pretend play (20) 75
Sense of Humor The child enjoys joking with other children during 75
pretend play (19)
Sense of Humor The child likes to “clown around" during pretend play 74
(23)
Cognitive Spontaneity The child invents his/her own games during pretend .81
play (10)
Social Spontaneity The child assumes a leadership role when playing with .80
others during pretend play (9)*
Cognitive Spontaneity The child assumes various characters’ roles during .70
pretend play (12)
Cognitive Spontaneity The child uses unconventional objects during pretend .68
play (1 1)
Social Spontaneity The child responds easily to others' approach during .90
pretend play (6)
Social Spontaneity The child plays cooperatively with other children during .88
pretend play (7)
Social Spontaneity The child responds easily to others’ approach during .82
pretend play (8)
Manifest Joy The child is restrained in expressing emotion during .82
pretend play (17)
Manifest Joy The child shows enthusiasm during pretend play (16) .72
Manifest Joy The child demonstrates exuberance during pretend .69
play (15)
Manifest Joy The child expresses enjoyment during pretend play .62
(14)
Physical Spontaneity The child prefers to be active rather than quiet during .81
pretend play (3)
Physical Spontaneity The child is physiwlly active during pretend play (2) .74
Physical Spontaneity The child runs (skips, hops, jumps) a lot during pretend .73

play (4)

 

Note. 1) The number of parentheses is the item number on the Imaginative Playfulness Scale.

2) F1: Sense of Humor, F2: Cognitive Spontaneity, F3: Social Spontaneity, F4: Manifest Joy, F5:

Physical Spontaneity.

* Reclassified as Cognitive Spontaneity.

79

Structural Equation Model

After these statistics were examined, the second step was an examination
of the Structural Equation Modeling (SEM) using the Analysis of Moment
Structure (Amos) in order to evaluate the hypothetical model.

What is SEM?

SEM is a path analysis with latent variables, so it allows testing the explicit
representation of direct and indirect effects among latent variables as well as
observed variables (Kline, 1998). That is, SEM is useful to evaluate the
hypothetical model of this study that includes latent variables and observed
variables. In SEM analysis, relationships among the variables are examined
simultaneously via analysis of covariance. When a model fits the data, the
reasonableness of the model can be confirmed. In this study, Generalized Least
Square (GLS) estimation was used to estimate model parameters and to test the
fit of the equations of the model to the data.

Fit lndices

There are several fit lndices to check model fit. Because each fit index
tells goodness-of-fit in a particular way, sets of them are generally used to
examine the reasonableness of the model. For this study, X2. the Goodness of
Fit Index (GFI), the Adjusted Goodness of Fit Index (AGF I), the Comparative Fit
Index (CFI) and the Root Mean Square Error of Approximation (RMSEA; Root
mean square residual) were obtained. The stmctural models were considered to
fit if the X2 is not significant, the x2/df ratio is less than 3, if GFI, AGFI, and CF I

are each greater than .90, and if RMSEA is less than .1 (Bagozzi & Yi, 1988;

80

Kline, 1998).
Procedures

When a SEM model consists of latent variables and their indicators, a two-
stage process should be followed: testing measurement models and analyzing a
hypothetical model with the tested measurement models. The first stage is to
test a measurement model and the second stage is to test an initial hypothetical
model.

Measurement Model Testing. Prior to the estimation of the initial
hypothetical model, measurement model testing was done as a first step of SEM.
Measurement model testing is a type of Confirrnatory Factor Analysis (CFA)
conducted to corroborate its underlying constructs. Traditionally, CFA is
performed to give evidence for convergent and discriminant validity prior to the
main analysis (Rogers, 1987).

According to Garson (n.d.), high loadings of indicators on the predicted
factors indicate convergent validity. In order to examine convergent validity,
estimates of regression weights, that is, factor loadings, the critical ratio (CR),
and significance level were obtained to test null hypotheses.

Discriminant validity demonstrates that the correlation between factors is
not so high as to lead one to think the factors overlap conceptually. Correlations
among latent variables of the measurement model were obtained and examined
in the present study.

Depending on statistical evidence of convergent and discriminant validity

and a good fit, the researcher did further evaluation of the adequacy of the

81

measurement model and modified it based on conceptual reasonableness.

Hypothetical Model Testing. The measurement model testing is followed
by data analyses to test hypothetical models to examine the research
hypotheses. That is, the subsequent analyses of measurement models are done
to test the data for the degree to which they fit the proposed patterns (Bryne,
1994). In this study, the second stage of modeling, including the structural paths
between the latent constructs and the observed variables that were exogenous
variables, was tested. Fit lndices, including )8, GFI, AGFI, CFI, and RMSEA,
were calculated to test the fits of the hypothetical models.

Model Trimming and Model Building. In addition to testing the overall fit of
an initial model, efforts to improve the initial model can be done using SEM.
Model improvement was done by constraining some of the insignificant paths
(model trimming) or by adding some of the significant paths (model building).
Modification lndices, which are calculated for every parameter that is fixed to
zero, would guide possible changes of X2 test statistics in the initial model.

After the initial model was modified in order to obtain a better fitting model,
12 difference statistics, which refer to the difference between the overall fit of the
initial model and revised models, were used to compare the fit of revised models
with the fit of the initial model to test model improvement. In addition to statistical
evidence, respecification and reestimation of these values were guided by theory
in order for conceptual plausibleness (Benishek, 1992).

After modification, 78 difference statistics were obtained to examine

significant improvement of the modified models. Modification of the initial model

82

was finally decided when it was conceptually reasonable and statistically
supported.
Regression Analyses

Multiple regression and hierarchical regression analyses were performed
as supplementary analyses to investigate the contributions surrounding the two
indicators of parental mediation, that is, instructive mediation and restrictive
mediation.

Data Transformation

Data transformation of some of the variables was done. First, the range of
the raw scores for income and the amount of family visual media use was too
large a range with large standard deviations; family income ranged from zero to
$250,000 with an average of $78,719.31 (SD = 45,540.35) and family visual
media use varied from zero to 2,700 minutes with an average of 360.74 (SD =
290.63). Therefore, they were transformed as standardized values, whose mean
is zero and standard deviation is one.

Another data transformation was completed for data whose distributions
were not normal. The first transformed variable was restrictive mediation, which
was an indicator of parental mediation. When the researcher conducted an
“eyeball check” of the distribution of each variable to identify normality
assumption violations, the distribution of restrictive mediation was found to be
extremely negatively skewed. Restrictive mediation ranged from eight to 25 with
an average of 16.40 (SD = 3.21 ). The median of restrictive mediation was 17

with a skewness statistic of -.72. Because the possibility of a skewed distribution

83

increases as the skewness statistic departs further from zero (Brown, 1997), the
distribution of restrictive mediation was ascertained as a skewed one. Therefore,
raw data of restrictive mediation measure was transformed using the square root
transformation, which is a general approach to handling non-normal data, was
done in this study. In addition to restrictive mediation, instructive mediation,
which indicates parental mediation construct with restrictive mediation (M = 18.88,
SD = 2.89), was also transformed using the square root transformation in order to
be consistent with the other measure, although its normality of distribution was
acceptable (Skewness = -.29).

The six variables of children’s visual media use were also not normally
distributed. All the distributions of the variables were zero-inflated ones, so they
were extremely positively skewed. In addition to the visual inspections of the
distributions of the variables, the statistics of skewness were also very large
(Type 1 programs = 1.31, Type 2 programs = 2.16, Type 3 programs = 1.67,
Type 4 programs = 4.28, Type 5 programs = 3.79, Type 6 programs = 4.08). For
this reason, the six variables were also converted using square root

transformation due their skewed distribution.

84

CHAPTER 4

RESULTS

This chapter describes the findings of this study. Six sections are
included: the relations among the observed variables; the measurement model;
the structural equation model; multiple regression; hierarchical regression; and a

summary of the findings.

Relations among the Observed Variables

Correlational analyses were performed to determine the extent of the
associations among the 21 observed variables. Those at the familial level
included family head’s occupation (OCP), gross income (INC), family head’s
education (EDU), number of visual media devices (FVMD), number of children’s
devices (CVMD), amount of family visual media use (FMVU), instructive
mediation (IPM), and restrictive mediation (RPM). At the individual level, the
variables included age (AGE) and gender (GNDR). As mediators, there were the
amount of viewing Type 1 programs (TY PE1 ), amount of viewing Type 2
programs (TYPE2), amount of viewing Type 3 programs (TYPES), amount of
viewing Type 4 programs (TYPE4), amount of viewing Type 5 programs
(TYPE5), and amount of viewing Type 6 programs (TY PE6). The outcome
variables were physical spontaneity (PSIP), social spontaneity (SSIP), cognitive
spontaneity (CSIP), manifest joy (MJIP), and sense of humor (SHIP). Table 8

presents the zero-order correlations among the variables.

85

68:: 8 emcew .FN $2 88:82 .8

88:88:88 88:80 .2 8888888 .888 .2 588.88 .8888 .2 .8880 .2 .82. .2 88.895 8 83 888.) 8 .895 .2 88.8.8

nab-9858083228.. .mEEmBavéhm:§888E<.~2 .Qtfimoammafimcghofiaog: .mEEmBamg-rmcfiés

8 8.8.5 .2 88.828 2 8&2 856$ 8 822. .8 88885 28588 .8 88885 9838. .2 .88: 885 .883 8 .885 .8 .8838
8.88.26 8 8.82 .8 .8838 985 88> .8 .8282 .8 .8585 8890 .8 888:8 8.888; 288”. .8 88888 8.888; 2.88”. .2 .8888

 

 

8. 8. 8.- 8.- 8.- 8. 8.- 8. 8.- ...2.- 2.- 8.18 .8
8. 8. 8.- 2.- 1.2.- 8. 8.- 8. .2.- :2.- .2.- n_=.s_ .8
8. 8. 8.- 8.- ...2.- 8. 8.- 8. .2.- ..2.- 8.- 8.80 .2
8. 8. 8.- 8.- ...2..- 8.- .2.- 8.- 8.- 8. 8.- 8.88 .2
a2. .2. 8.- 8.- 2.- 8. 8. 8. .2.- ..2.- .2.- 8.8.". .2
8.- 8.- 8.- 8.- 8.- 8.- 2.- a2.- 2.- 8. 8. amaze .2
8.- 8. .2.- 8. 8.- 8. 8. 8. .2.- .2.- ..2.- 22 .2
.2. 8. 8.- 8.- 8. 2. 8. 8. t2.- 2.- ..2.- 88882 .2
:8. t8. 8.- 2.- 8. .2. 8. .2. 12.- t2.- 188.. 88882 .2
.2. .2.- 8.- 8.- 8. 8.- 8.- 8.- 8. 8. 8.- 8882 .2
8.2 :8. 8. :8.- 8. ..8. .2. :88. 8.- 8.- 2.- 88882 .2
8.2 8.- :2.- 8. .88. :2. .2. :8.- ..N8.- :8... 88882 .2

8.2 8.- 2.- ...8. 8.- 2. 8. 8. 8. 82882 .8

8.2 :8. .2.- t2.- :8.- 8. 8. ...2. 8.288 .8

8.2 8. 8.- 8.- 2. 8. 2. 8.28. .2

8.2 :8. :88. 8.- :8.- ..88: 2:22“. .8

8.2 :8. :2.- Lu.- ....8.- 92>o .8

8.2 :2. .2.- 88.. 82>... .8

8.2 :8. :88. :oz. .8

8.2 L8. :88 .N

8.2 800 .2

2 2 8 8 2 8 8 v 8 N 2 8.88%

 

ANON H ZV whom—@065 ho mcozflmtoo Ucm .wcozmsmo 0.50:9” 5:50: 0 90m...

86

._.o. v Q... .mo. v at .o... v Q.

.2" EO .o u >om 8 62882888: woo. 88:08 9.8: 228250 AN .8888 N 2 .882

 

87

 

56 3.3. 2:. tom. the. iVN. tum. 9..- imN. iNN. to F. No. _.N
Dav 2.3 2:. 1N0. imm. 1mm. mo. ¥MN. an... 5.. mo. ON
m _..v mmdr 00... tom. tom. .NF. 1mm. mo. mo. mo. 9.
EN mmdw 00.? in. 19.. top. No. 2.. S. e
Nod Nwdw 00... tm F... ..t. .12. .m _.. mo. 5
om. av. 2.: mo. 5... mo. 3.- or
mod NN.Nm ooé SN. 3. or. B
ems“~ m5? ooé 1 FN. imN. 3
05m nod oo._. «9.. m?
56 VON on: NF
mod Nvdw :-
nod so. 3 or
no.0 m5: m
NV. 86 m
cm. and N
oo._. co. m
wk. on. m
3N vmé v
co... co. m
Ed mnbr N
onNN mméo F

Om memo—2 PN ON 9. Q. t. 0—. m2 3 m_. N _. mm_nmtm>

 

8.88 8 8.882

Most of all, the variables indicating the same construct were found to be
correlated with each other. In the measurement model, there were four latent
variables: socioeconomic status (SES), viewing environment in a family (FVE),
parental mediation (PM), and imaginative playfulness (IP). Family head’s
occupation, family head’s education, and gross income indicating socioeconomic
status, were significantly related to each other in a positive direction (p < .001 ).
Correlations between the number of visual media devices, the number of
children’s devices, and amount of visual media use, which indicated a construct,
that is, viewing environment in a family, were also positively significant (p < .001).
A significant correlation between instructive mediation and restrictive mediation
that were indicators of parental mediation was found in a positive direction
(p < .001). The indicators of imaginative playfulness, which were physical
spontaneity, social spontaneity, cognitive spontaneity, manifest joy, and sense of
humor, were also positively related to each other (p < .001).

Several significant correlations between the variables were found as
hypothesized, mostly in the expected directions. For example, socioeconomic
status indicators were negatively related to the indicators of the family viewing
environment. Among nine correlations between three socioeconomic status
indicators and three family viewing environment indicators, seven of them were
negative and significant (p s .005); only two correlations between occupation
and the number of visual media devices (p = .69) and between income and
amount of family visual media use were not significant (p = .22).

On the contrary, the relationships between indicators of socioeconomic

88

status and parental mediation were expected to be positive. Although
associations between three socioeconomic status indicators and the two parental
mediation indicators were positive, only the correlation between income and
restrictive mediation was significant (p = .04).

Children whose family viewing environment was rated with high scores
tended to watch more television, videos, or DVDs. Three family viewing
environment indicators and six program types were mostly positive as expected
but they were small to moderate in magnitude and sometimes were not
significant (.07 s p s .51). There were four correlations between the family
viewing environment indicators and the six visual media use variables which
were negative, but they were very small and not significant (.26 s p s .87).

However, parental mediation of what their children watched was likely to
be related to the children’s visual media use. Interestingly, the relationship
between parental mediation and visual media use depended on the type of
parental mediation. Instructive mediation was related to visual media use
variables positively (.20 s p s .77) except for the amount of viewing of Type 1
programs (p = .10). On the other hand, restrictive mediation was negatively
related with visual media use variables. Among the correlations among parental
mediation indicators and visual media use variables, only the relations between
restrictive mediation and Type 2 and 3 programs were significant (p = .01 );
children whose parents restricted what children viewed were unlikely to watch
Type 2 and 3 programs.

As for the relationships of child characteristics with visual media use

89

variables, children’s gender was not associated with all types of programs
(.12 s p s .82). On the other hand, there were two significant correlations
between age and visual media use variables. That is, older children tended to
watch more Type 6 programs (p = .002) and less Type 1 programs (p = .05).

The correlations between child characteristics and imaginative playfulness
indicators were mostly compatible with the hypotheses. Older children tended to
be more playful in terms of their physical, social, and cognitive spontaneity,
manifest joy, and sense of humor during imaginative play (p s .02). Boys tended
to have higher scores on physical spontaneity (p = .04), and girls were likely to
have higher scores on social spontaneity (p = .01).

The relationships between the five imaginative playfulness indicators and
the six observed variables of visual media use were contradictory to the
hypotheses. The amount of viewing Type 1 programs was negatively related to
imaginative playfulness indicators (.21 s p s .66), while correlations between
other types of programs and imaginative playfulness indicators were mostly
positive but were not significant.

Several relationships which were not hypothesized were found. For
example, indicators of the family viewing environment were significantly
correlated with restrictive mediation (p s .003), while instructive mediation was
not statistically correlated with any indicator of family viewing environment
(.79 s p s .85). Additionally, indicators of socioeconomic status were also found
to be related to viewing amounts of several programs; children of a higher

socioeconomic status family tended to watch fewer Type 2, Type 5, and Type 6

90

programs (p s .03). The socioeconomic status of a family was also found to be
related to the imaginative playfulness of their child. In general, socioeconomic
status indicators were negatively correlated with imaginative playfulness
indicators; children of a higher socioeconomic status family were less likely to be
playful. In particular, the level of physical spontaneity, cognitive spontaneity, and
manifest joy tended to be low when socioeconomic status was high

(.03 Sp 5.07).

Measurement Model
A measurement model refers to a model with a single latent variable
indicated by observed variables. Prior to the main analysis to test hypotheses
using SEM, measurement model testing is conventionally required for measure
validation. Convergent validity and discriminant validity, which together give
evidence for construct validity, were tested using Confirrnative Factor Analysis
(CFA). In CFA, a finding that indicators have high loadings on the predicted
factors indicates convergent validity. In an oblique rotation, discriminant validity
is demonstrated if the correlation between factors is not so high as to lead one to
think the two factors overlap conceptually (Garson, n.d.).
Estimation Procedure
As stated in Chapter 3, socioeconomic status, parental mediation, family
viewing environment, and imaginative playfulness measures were the latent
variables indicated by observed variables. On the other hand, age, gender, and

visual media use were treated as one-item scales. Therefore, these variables

91

were not included in the measurement model testing in the CFA prior to the
analyses of the hypothetical models (see Figure 4).

Although the square root transformation was performed to normalize the
distributions of parental mediation indicators, particularly the restrictive mediation
variable, such change was not powerful enough to transform the original
distribution into an acceptable normal distribution. This presents a problem for
the hypothesis testing in SEM using Maximum likelihood (ML) estimation that has
been conventionally used in most studies. Because ML assumes a normal
distribution in the population from which you drew your sample data, non-
normally distributed data are more likely to cause problems for CFA. A violation
of normality may lead to a rejection of models that may not be false and decide
that particular parameter estimates are different from zero when in fact this is not
the case (type 1 error) because the chi-square test statistic of overall model fit
will be inflated and the standard errors used to test the significance of individual
parameter estimates will be deflated (Information Technology Services, 2002).

Generalized Least Squares (GLS), which is another estimation to compute
goodness-of-fit lndices, parameter estimates, and standard errors, was used to
deal with the non-normal data. GLS is widely recommended to test data with
non-normal distribution because it is more robust, nevertheless it under-perfonns
relative to ML in the testing of overall model fit and the significance of individual
parameter estimate values relative to ML (Osborne, 2002). Therefore, GLS
estimators were obtained to test the measurement models instead of ML

estimation, as is conventionally used in most studies.

92

.Loto EeEosmmm—z m. .LoEsz 20 mmcem “dim So.- fiomcmi 8:22 Sgccficoaw

82:88 8.80 8889888 .888 ”8.88 8888888 .8228 888 888685 88588 Sam 88888:.
82828:. ”in: .88: 868E _m:w_> ho anE< ”02>... .8833 8.55.28 v6 LmnEsz H.550 .8838 86.2: :8wa

ho LmnEsz “92>u .cozmosuc mbmmc 22:8“. 50m .0885 8890 ”02_ 60:88:80 8.888; >__Emn_ ”moo 88859383
0228:5825 i. 60:86.88 .95me ”SE 5:82 8 :_ EoEcoLSco m:_>>m_> MS“. .8388 o_E0coomo_oow ”mmm .8288me

n=Iww $228 1508 n=mww n=mnt 21%. SEE 22>“? D§>Ow 055% Doww oz_w QOOw
L , , , L

L L L L L
ILIIIL . < L L Lu. IILIL lH L L P Lu‘. e LLI L LI KIIIIL IL.) L L RI , III.LLI
L L . L L L L L
L nzIw L n_=-._>_ L nsz fizwm L n:w.n_ L SEE L L STE L LD§>mL LBS—>0 0L LD§>HLL LIv Dow IILL L 02_ L L 100
L L L L L
Liv I L y I (I L L\.&III. L L1|L L «I L .IIJ1, IL IN IIL LI -. ‘IyL I L L x
. , . I I \LLLL ..I/... LLL \\\. LLLL .
IIIFIIIIII I . . IWILI - IL.I
IL A Iwwm ,8. I
/ l/ I x I I\ II.\ / /I. I \ I\
I/ III I I / IIIIIIIII I I II III) \

.2805. 888058885. .28 859m

93

Discriminant Validity

Discriminant validity, referring to the distinctiveness of the factors
measured by different sets of indicators, can be determined by examining
estimated correlations between the factors. In this study, therefore, discriminant
validity was tested by examining relationships between each pair of constructs.
Conventionally, the correlations which are not excessively high (e.g., >.85) can
be evidence of discriminant validity (Kline, 1998).

As shown in Table 9, there were no high correlations among the
constructs. The correlation between the family viewing environment and parental
mediation was the highest one (r = -.43, p = .10); the family viewing environment
and imaginative playfulness was the lowest correlation (r = .10, p = .37). The
results show that an adequate level of discriminant validity among the constructs
was achieved in the measurement model. Therefore, it was concluded that
socioeconomic status, family viewing environment, parental mediation and
imaginative playfulness measured different concepts.

Convergent Validity

Table 9 summarizes key parameter estimates, and indices from the

confirmatory factor models, showing directions, significance, and effect sizes of

indicators.

94

Table 9.

Summary of Measurement Model

 

Parameter estimates

Initial measurement model

Modified measurement model

 

 

 

 

 

 

 

 

 

Parameters Estimates” C.R. Significance Estimates C.R. Significance
SES -—> OCP .83 7.37 .00 .81 8.88 .00
see —> INC .62 —2’ — .68 —2’ —
SES —+ EDD .82 7.87 .00 .85 9.16 .00
FVE —> FVMD .72 4.74 .00 .78 4.24 .00
FVE —> CVMD .62 4.72 .00 .59 4.71 .00
FVE _. FVMU .51 —2’ — .48 —2’ —
PM —> IPM .35 1.77 .07

PM —+ RPM .84 —2’ —

IP —» PSIP .70 9.58 .00 .70 9.67 .00
IP —+ SSIP .38 4.65 .00 .41 5.03 .00
IP —> CSIP .77 10.65 .00 .78 10.85 .00
IF —> MJIP .78 10.37 .00 .79 10.55 .00
IF _. SHIP .81 —2’ — .80 —2’ —
SES 9 FVE -.39 -3.01 .00 -.44 -3.33 .00
SE3 2. PM .17 1.24 .22

SES 8 IF -.26 -2.67 .01 -.23 -2.60 .01
FVE <—> PM -.43 -1.65 .10

FVE <—> IF .10 .90 .37 .08 .85 .40
PM 8 IF -.12 -1.01 .31

INC —-> FVMD .35 4.39 .00
Goodness of fit sgmmary

Model 16 df ledf GFI AGFI CFI RMSEA
Initial model 68.70 59 1.16 .95 .92 .94 .03
Modified model 34.63 40 .86 .97 .95 1.00 .00
Modifica_tion indices 3’

Path Modification indices

INC —+ FVMD 10.49

elNC4) H eFVMD4) 7.62

eFVMo”) ... ses 7.37

FVMD —+ INC 5.99

ses —+ FVMD 4.46

 

Note. 1) Standardized regression weights, 2) Constrained parameters to 1.0, 3)
ML of the initial measurement model, 4) Measurement error of the variables.

95

With convergent validity, all constructs except parent mediation showed
significant coefficients; the regression coefficient of instructive mediation
indicating parental mediation was 1.77 (p = .07) although parental mediation was
found as a latent variable significantly indicated by instructive mediation and
restrictive mediation when the measurement model was tested using EFA. On
the other hand, the remaining measures showed sufficiently significant scores,
indicating measures of the respective constructs are internally consistent. That
is, the results of CFA showed that all loading parameters except PM were large
and loaded significantly on the latent variables which they were intended to

represent.

Indicators having non-significant factor loadings are conventionally
dropped unless there is a compelling reason to keep them. As an example,
parental mediation has been shown to be as an important mediator of children’s
viewing practices in past studies. However, convergent validity of the parental
mediation measurement model was revealed as poor, because instructive
mediation and restrictive mediation were not loading on parental mediation which
they were intended to represent. In other words, the measure of parental
mediation in the present research was statistically not appropriate despite the
theoretical significance. Thus, it was decided to eliminate the parental mediation

measurement model from the hypothetical model.
Measurement Model Modification
There were two major modifications to improve the model fit of the
measurement model. The first modification was to remove the parental

mediation measurement model as stated above. Another modification was to

96

add a significant path to the measurement model.

The researcher inspected the modification indices (M.|.) to find paths to be
added. As showed in Table 9, six paths were found that could improve the
overall fit of the measurement model. Because the fit of the initial measurement
model was already good, it was decided that a path from income to the number
of visual media devices in a family, which had a ML of 10.49, was added in the
model because the ML of the path was higher than the other paths.

Besides the statistical significance, the addition of the path from income to
the number of visual media devices in a family was also conceptually important
because the direction of the path was opposite to the direction of the path
between the indicators’ latent variables. That is, the path from income to the
number of visual media devices in a family is the relationship between indicators
representing their latent variables, which are socioeconomic status and family
viewing environment. Accordingly, the effects of income on the number of visual
media devices in a family were covered by the influences of socioeconomic
status on family viewing environment. While the coefficient of the path from
socioeconomic status to the family viewing environment was negative, the path
from income to the number of visual media devices in a family was positive. That
is, the relationship between income and the number of visual media devices
could not be covered by the relationships between the two latent variables. In
fact, it is reasonable that families with low income cannot afford to purchase
multiple visual media devices. Based on this rationale, income, which has

unique effects on the number of visual media devices, was decided to be added

97

to the measurement model.

Therefore, it was hypothesized that the indicator, income, statistically
influenced another indicator, the number of visual media devices, separately from
the path of its effect through socioeconomic status and the family viewing
environment in the measurement model. The modified measurement model
appears on Figure 5.

Model Fit

The fit indices of the overall measurement model were obtained, and the
results indicated the confirmatory measurement model was a good
representation of the data (see Table 9). The ledf and RMSEA statistics were
small enough to indicate a good fit; the values of GFI, AGFI, and CFI were also
sufficient to accept the model as a good fit (ledf = 1.16, GFI = .95, AGFI = .92,
CFI = .94, RMSEA = .03). Examination of the modified measurement model,
after allowing for the path from income to the number of visual media devices
and removing the parental mediation measurement model, provided a better fit;
ledf and RMSEA decreased, and GFI, AGFI, and CFI increased (ledf = .86,
GFI = .97, AGFI = .95, CFI = 1.00, RMSEA = .00). After modifying the
measurement model, the significance of improvement was examined using 12
difference tests. According to the result, the difference in x2 was significant dx2
(1) = 34.07 (p < .001). The X2 difference test results led to the conclusion that the
modified measurement model was more adequate than the original

measurement model.

98

. f0. V Que.

cote EoEoSmmos. ”w .583; he omcow Hn=Iw $0.. 63:22 No.22 56:95am o>=_cmoo ”Emu same—Scone

_m_oow nn=mm 55:3:on Reina ”Ema .3: £86 .m:m_> ho anE< ”92>“. 683% 3.9220 8 59:52 ”0550
583% Even. 353 B 89:32 ”92>“. .cozmoauo wbmo; >__Emn_ ”Dow .0885 890 ”02. 5038308 mbmoc 2E5”.
”moo .mmocSSma 92.9.39... ”n: .>__Emu_ m s EoEcBSco @5265 um>u_ .mBSw o_Eocoooo_oom ”mum .mocmcmkom

.oé 2 999893 8593200 3 .302

n=Iww n=22w n=mOw Emmw $wa 355%.. 05506.. 92>“...o Dom—w 02.6.. QOOw

H a a H a H a Hijs a

n=Iw n:22 n:wo Emm n:mn_ 22>“. n=2>O 92>“. Dow 02. Q00

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

:. 3. ton.

   

eon/4.2.

 

.1..me

.2322 265658.82 858.2 .m 65mm

99

From the overall evidence on discriminant and convergent reliability, and
fit indices of the original and modified measurement models, the researcher
concluded that adequate levels of the measurements were achieved. Thus,
three constructs, socioeconomic status, family viewing environment, and
imaginative playfulness, were utilized in the initial hypothetical model to be

tested.

Structural Equation Model

Figure 6 presents the initial hypothetical model that incorporated the
hypotheses for the present study. This model was a revised model of the original
hypothetical model shown in Figure 4, adjusted for the results of the measures
validation.

In the initial structural model, there were 19 observed variables and three
latent variables, namely: occupation, education and income indicating
socioeconomic status; the number of visual media devices, the number of
children’s devices, and amount of family visual media use indicating viewing
environment in a family; physical spontaneity, social spontaneity, cognitive
spontaneity, manifest joy, and sense of humor indicating imaginative playfulness;
age and gender which were child characteristics variables treated as exogenous
variables; and six exogenous variables assessing the amount of viewing of Type

1 programs through Type 6 programs.

100

uoEzc ho omcom dim $0.. 7.02:0: No.22 .Eocficoqm

052600 un=mo .Eocflcoam _0_00w unzwm .roflcoqm .mofiEn. unzmn. .._00:00 #520 .0m< ”mod. .mESmoa

m on: 9:53 *0 E30E< no mat. .mEEmoa m 09? @5303 *0 E:0E< ”m mm>h .mEEmoa v 09:. @5303 *0 E30E<
”v wntrr .mESmoa m 09:. @5303 *0 E30E< ”m mat. .mESmoa N 09? @5303 *0 EzoE< ”N mm: 680505

w 0%... @5303 F0 E30E< ”F un_>._. .00: 060E .053 *0 E30E< ”92>“. £00300 0.56:5 *0 anEzz ”0550

600300 m_00E .0303 *0 59:32 ”02>“. 50:00:00 9000; >__Emn_ ”Dom .oEooE 30.0 “oz. .cozmasooo 0.000: 2E0“.
No.00 .mmocsgma 0>=m:_mmE_ ”a. .>__E& m c_ E0Ec0._>co @5305 um>m 63.3w 0_E0c0000_00m nwmm .mocogokom

Io Tl
3>2d Hm 02>0
L. z

I lIIIIWI

/ 7,
« xx/r

 

 

\

_ .9 :9 _

 

duos. 62652;: .625 .0 9:94

101

In the initial model, visual media use was assumed to be influenced by
socioeconomic status through the family viewing environment and eventually
affecting imaginative playfulness. Age and gender were hypothesized to
influence the six program types and imaginative playfulness respectively. In
addition, a path from income to the number of visual media devices was added.

Hypotheses Testing

The key parameters obtained for testing the hypotheses is reported in
Table 10. The hypotheses, based on theoretical grounds and previous empirical
findings in Chapter 3, were examined. Among a large number of hypotheses, 2
and 4 were not examined using SEM because the parental mediation
measurement model was not included in the structural model testing.

H 1. The higher the socioeconomic status of a family is, the lower their level of
the family viewing environment.

The hypothesis predicted socioeconomic status to have a negative effect
on family viewing environment. For the data of the present study, the
hypothesized effect of socioeconomic status on the family viewing environment
was supported. That is, higher levels of socioeconomic status were associated
with lower levels of family viewing environment.

H 3. The higher the level of the family viewing environment is, the more visual

media a child in the family uses.

102

Table 10.

Summary of Initial Hypothetical Model

 

 

 

flrameter estimates

Parameters Estimates" C.R. Significance
$53 —> OCP .85 — 2’ —
SES —> INC .62 7.39 .00
SES —> EDU .77 8.33 .00
FVE —> FVMD .75 5.69 .00
FVE —> CVMD .55 — 2) —
FVE —9 FVMU .63 5.1 1 .00
IP -+ PSIP .70 9.34 .00
IP —> SSIP .42 5.06 .00
IP —> CSIP .82 10.77 .00
IP —> MJIP .77 10.27 .00
IF —> SHIP .82 — 2’ —
INC —> FVMD .32 4.50 .00
SES -> FVE -.47 -3.87 .00
FVE —> TYPE1 .21 2.08 .04
FVE —> TYPE2 .44 3.97 .00
FVE —> TYPE3 .54 4.41 .00
FVE —> TYPE4 -.03 -0.32 .75
FVE —> TYPE5 .37 3.45 .00
FVE —> TYPE6 .19 2.02 .04
AGE —> TYPE1 -.11 -1.42 .16
AGE —> TYPE2 -.02 -.26 .79
AGE _, TYPE3 -.04 -.58 .57
AGE —> TYPE4 .16 1.75 .05
AGE —> TYPE5 .05 .67 .50
AGE —> TYPE6 .26 3.22 .00
GNDR —> TYPE1 -.04 -0.50 .61
GNDR —+ TYPE2 -.03 -0.39 .69
GNDR —> TYPE3 .04 .53 .60
GNDR —9 TYPE4 -.11 -1.21 .23
GNDR -+ TYPE5 .10 1.36 .18
GNDR —* TYPE6 -.04 -.53 .60
AGE —> IP .28 3.26 .00
GNDR —> IP -.01 -.01 .99

 

103

Table 10 (Cont’d)

 

 

 

 

 

 

Parameter estimates
Parameters Estimates C.R. Significance
TYPE1 —> IP -.04 -.43 .67
TYPE2 —* IP -.07 -.76 .44
TYPE3 -> IP .07 .79 .43
TYPE4 —> IP -.04 -.43 .67
TYPE5 —-> IP .05 .59 .55
TYPE6 —> IP .10 1.09 .28
@dness of fit swmary

x2 df ledf GFI AGFI RMSEA

192.43 132 1 .46 .90 .86 .05

 

Modification indices
Path

SES
eTYPE4
€TYPE4
eTYPE3
eTYPE3
eSHW’
6PSW’
EEDU
eOCP
eFMVU
eCVMD
TYPE4
TYPE4
TYPE2
TYPE6
TYPE4
TYPE4
GNDR
TYPE2
FVMD
TYPE1

llllllllllllllliiilil

AGE
eTYPE2
eTYPE6
eTYPE2
eTYPE4
eTYPE4
GNDR
6TYPE2
EFVMD
eTYPE1
eTYPE1
TYPE2
TYPE6
TYPE4
TYPE4
TYPE3
SHIP
PSIP
EDU
OCP

F MVU

Modification index

6.18
8.75
6.98
4.03
6.24
5.26
4.89
4.84
6.20
9.08
4.01
8.38
6.67
6.58
6.24
5.95
4.72
4.89
5.56
4.90
8.84

 

Note. 1) Standardized regression weights, 2) Constrained parameters to 1.0.

104

In this hypothesis, the family viewing environment was expected to have a
positive effect on visual media use regardless of the type of programs. This
hypothesis was mostly supported. In particular, the family viewing environment
contributed to children’s visual media use. Following in descending magnitude in
the family viewing environment effects, were Type 5, 2, 3, 1, and 6 programs.
Only Type 4 programs did not show a significant pathway from family viewing
environment. That is, the children whose family rated higher in the level of family
viewing environment tended to watch more Type 1, 2, 3, 5, and 6 programs, but
Type 4 programs were not related to that level.

H 5. Children’s visual media use varies by child characteristics.

H 5.1. Younger children spend more time viewing Type 1 programs than

older children; older children spend more time viewing Type 2, Type 3,

Type 4, Type 5 and Type 6 programs than younger children.

As expected, children’s age had mixed effects on children’s visual media
use by program types. In addition to Type 1 programs that were hypothesized to
have negative effects, children’s time spent watching Type 2 and Type 3
programs also decreased with age. On the other hand, older children spend
more time viewing Type 4, Type 5 and Type 6 programs than younger children
as anticipated. In short, programs designed for children tended to be viewed by
younger children, and general audience programs were likely to be watched by
older children regardless of the presence of violence components and
educational purposes.

Among these six paths from family viewing practice to each program type,

105

only two programs were influenced by family viewing practice; family viewing
practice had significant positive effects on Type 4 and Type 6 programs but did
not contribute to the other programs. That is, older children spent more time
viewing Type 4 and Type 6 programs than younger children.

H 5. 2. Girls spend more time viewing Type 1, Type 2, Type 4, and Type 5

programs than boys; boys spend more time viewing Type 3 and Type 6

programs than girls.

The results indicated that the hypothesis was not supported for all the
types of programs; all coefficients were insignificant. That means children’s
viewing practices did not vary by children’s gender.

H 6. Children’s imaginative playfulness varies by their visual media use.

H 6.1. The more Type 1 programs children watch, the more playful they

are.

H 6. 2. The more Type 2, Type 3, Type 4, Type 5 and Type 6 programs

children watch, the less playful they are.

Both sub-hypotheses were not supported for all the types of programs.
There were no significant relationship for the hypotheses even at a marginal
level. The results were consistent across all six program types. Thus, children’s
viewing practices were not related to their imaginative playfulness.

H 7. Children ’s imaginative playfulness use varies by child characteristics.

H 7.1. Older children are rated higher than younger children in

playfulness including physical spontaneity, social spontaneity, cognitive

spontaneity, manifest joy, and sense of humor in imaginative play.

106

In the hypothesis, children’s age was expected to have a positive, direct
effect on their imaginative playfulness regardless of its dimension. For the data
of the present study, the level of imaginative playfulness increased by age as
expected. The results were consistent for all five dimensions. That is, children’s
age contributed to a higher level of imaginative playfulness.

H 7. 2. Boys are rated higher than girls in physical spontaneity, manifest

joy, and sense of humor in imaginative play; girls are rated higher than

boys in social spontaneity and cognitive spontaneity in imaginative play.

The results indicated that the hypothesis on the influence of gender on
imaginative playfulness was not supported. Regardless of the dimension of
imaginative playfulness, the level of children’s imaginative playfulness did not
vary by their gender.

H 8. The hypothesized model of the relationship of socioeconomic status,
children’s characteristics, and three mediating variables, including the
viewing environment in a family, parental mediation, and children’s visual
media use, to the level of imaginative playfulness in this study fits the
data.

The ledf ratio was less than 3, the value of GFI was greater than .90, and
RMSEA was less than .10. On the other hand, AGFI and CF I were each less
than .90. That is, only three fit indices of the initial hypothetical model succeeded
to show the model was a good representation of the data. Despite the results of
fit measures representing fairly adequate fit of the model, the model did not fit the

data well.

107

Modified Models

In addition to the fit measures indicating a quite unsatisfactory fit, there
were too many constraints in the initial model. Based on the coefficients and
modification indices (M.l.), the model was modified in order to create a more
parsimonious model with better fit in several steps.

Modified Model I

In the first modification, the model was obtained by constraining all the
insignificant paths, p-values less than .10, to be zero (see Figure 7). The first
modified model which removed statistically insignificant paths became a more
parsimonious model. This is more desirable because of its relative simplicity.

A total of 17 paths of the initial model were constrained in the modified
model I; the path from the family viewing environment to Type 4, from Type 1 to
imaginative playfulness, from Type 2 to imaginative playfulness, from Type 3 to
imaginative playfulness, from Type 4 to imaginative playfulness, from Type 5 to
imaginative playfulness, from Type 6 to imaginative playfulness, from age to
Type 1, from age to Type 2, from age to Type 3, from age to Type 5, from gender
to Type 1, from gender to Type 2, from gender to Type 3, from gender to Type 4,
from gender to Type 5, and from gender to Type 6 were excluded. Because
there was no significant path involving gender, the variable gender was removed
from the model. Therefore, there were 18 observed variables and three latent

variables in this model with 20 paths.

108

.5. v 0: .00. v 0..

..0:.:: .0 00:90. ”0.1m. $0.. .00._:0_2

”0.22 $._0:0.:000 02::000 ”050 $._0:0.:000 0.00m unzmw $._0:0.:000 _0o_0$:n. ”060 .09. ”m9. .0E0.00.0

0 00$ ._. 05303 .0 .::0E< no w0>... 0.00.090 m 00$ ._. 05303 .0 .::0E< ”m m0>... .0:.0.00.0 v 00$ .. 05303 .0 .::0:.<
”v m.n.>._. 08050.0 0 00$ .. m:_>>0_> .0 .::0E< ”m mat. .0E0.00.0 N 00$... 05.503 .0 .::0E< ”N m0>._. 08050.0

. 00$ ._. 05303 .0 .::0:.< ”_. m0>._. .00: 0.00E .0003 .0 .::0:.< ”02>“. 000300 000.220 .0 .0060: ”0550

000300 0.00E .0003 .0 .00532 “02>... 60:00:00 0.000... $__E0n. 50m. .3805 000.0 ”02. 00:00:80 0.000: $__:.0...
”000 0005200 0>_.0:_00E_ ”n: $._E0. 0 :_ .:0E:0..>:0 052,05 MS“. 0300 0.E0:0000_00w ”mum .00:0.0.0~.

0.. 0. 0.908000 00:.0..0:00 C 0.02

 

 

 
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

109

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. 0 00>. :22. 0530 02>.
..NN.
00.4 0 Bi
0.10 e.
a 00?.
0.22
mmo>F
0.00
N 00¢
20
. 00¢ :00 oz_ 000
E00
2082 8:269... 8582 N 6.50.“.

The 18 difference test, which evaluates the significance of the change in
the model, was not available. The X2 difference test is available only for the
relative fit of two hierarchically related models. The modified model that does not
contain an exogenous variable, gender, is not hierarchically related to the initial
model. Although significance of improvement of the revised model could not be
statistically tested, the fit indices in Table 11 indicate that the revised model is
still not satisfactory although the model was improved a bit (ledf = 1.35, GFI =
.90, AGFI = .87, CFI = .77, RMSEA = .04).

Modified Model II

The next modification to enhance model fit was to add pathways. The M.I.
estimates of the modified model I were used for the additional paths. As for the
modifications, some paths which were not empirically and theoretically supported
were not included in the second modified model even if the MI. indicated that
they would produce a numerically acceptable fit (Arbuckle & Wothke, 1999). For
example, the paths between program types were not included because causal
relationships among uses of visual media types could not be inferred in this study.
Also, the paths whose estimate of MI was marginally greater than 4, which is
the threshold in the Amos statement, were also excluded, because there were
too many paths displayed on the output. The paths among indicators of latent
variables were also not included because their relationships were already

explained by the relationship among latent variables.

110

Table 11.

Summary of Modified Model I

 

 

 

 

 

 

 

 

 

Parameter estiflates
Parameters Estimates” C.R. Significance
$53 —+ OCP .85 — 2’ —
SES —* INC .62 7.25 .00
SES —> EDU .77 8.18 .00
FVE —> FVMD .75 5.75 .00
FVE —> CVMD .55 — 2’ —
FVE —> FVMU .62 5.19 .00
IP —> PSIP .67 8.92 .00
IP —+ SSIP .43 5.02 .00
IP . CSIP .82 10.44 .00
IP —* MJIP .77 9.88 .00
IF —> SHIP .79 — 2) —
INC —> FVMD .33 4.80 .00
SES - FVE -.45 -3.87 .00
FVE - TYPE1 .23 2.26 .02
FVE —* TYPE2 .43 3.94 .00
FVE - TYPE3 .56 4.56 .00
FVE -+ TYPE5 .36 3.41 .00
FVE —> TYPE6 .18 2.05 .04
AGE - TYPE4 .16 2.76 .05
AGE H TYPE6 .22 1.84 .01
AGE —-> IP .30 3.71 .00
Goodness of fit summary
x2 df ledf GFI AGFI CFI RMSEA

178.72 132 1.35 .90 .87 .77 .04
Mocmcation indices
Path Modification index
eTYPE4 H €SHIP 6,14
eTYPE4 H eTYPE2 8.75
eTYPE4 H eTYPE6 7.72
eTYPE3 H eTYPE4 5.65
eFVMD H AGE 4.35
eEDU H eTYPE2 5.73

 

111

Table 11 (Cont’d)

 

Modification indfices

 

Path Modification index
eFMVU H eTYPEl 8.57
ECVMD H eTYPE1 4.94
TYPE4 H SHIP 6.12
TYPE4 H TYPE2 8.77
EDU H TYPE2 5.31
FMVU H TYPE1 4.64
TYPE4 H TYPE6 7.49
TYPE2 H TYPE4 7.10
TYPE6 H TYPE4 6.95
TYPE4 H TYPE3 5.00
EDU H TYPE3 4.05
AGE H FVMD 4.35
TYPE2 H EDU 6.50
FVMD H OCP 4.38
TYPE1 H FMVU 8.16
TYPE1 H CVMD 4.46

 

Note. 1) Standardized regression weights, 2) Constrained parameters to 1.0.

Accordingly, four paths were added that could be conceptually justified
among 25 paths identified as relationships that could improve the goodness of fit.
The four paths were from Type 4 programs to sense of humor, and between
measurement error (e) of Type 2 programs and e of Type 4 programs, between a
of Type 3 programs and e of Type 4 programs, and between 6 of Type 4
programs and Type 6 programs. The modified model II derived empirically and

theoretically is depicted on Figure 8.

112

”6:22 0.296% 02.280 5:0

.5. v 0: .8. v 0..

..0..0 .:0:.0.:000_2 m. ..0:.:: .0 00:00 ”0.10 $0.. .00._:0_2

$._0:0.:000 _0_0om ”n._wm $._0:0.:000 .0202... “0.00 .09. ”m0... .0E0.00.0

o 00$... m:..s0.> .0 .::0E< no wn.>... 080. 0.0 m 00$ H 05303 .0 .::0E< ”m m0>._. 0E0.00.0 v 00$ F 05303 .0 .::0E<
”v m0>... .0E0.00.0 m 00$ ._. 05.503 .0 .::0E< ”m m.n.>... 08050.0 N 00$ ._. 05303 .0 .::0:.< ”N m0>._. .0E0.00.0

F 00$ ... 05303 .0 .::0E< ”.. mat. .00: 0.00E .0303 .0 .::0E< ”02>“. 000300 0:00.20 .0 00:52 ”0550
000300 0.00E .0003 .0 .00Eaz ”02>”. 00:00:00 0.000: $__E0u. ”Dow .0E00:_ 000.0 ”02. 00:00:80 0.000: $__E0...
.000 00050300 020500.... ”0. $__:.0. 0 :_ .:0E:0._>:0 05305 nm>0 .0300 250580.80 .wmw 00:00.00.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

own.>._.0 lv
:8.
m0< :40.
0..
0.10 y, 5.- \vamn.>.$ ll
.0 t
052 /v 000.61.
.1 FN.1
Emu Nm.n.>._.0 Iv
n._.wm
0.9”.

 

 

.o. .. 0. 0.0.0E0.00 00:.0..0:00 3 0.02

 

 

 

o wn.>...

 

 

32).“. 0530 92>“.

 

 

 

 

 

 

 

 

m mat

 

 

.4 wn.>._.

 

 

 

m mat

 

 

 

 

N mat

 

 

 

 

w uni

 

 

 

/ \\
k :00. \.\.0...

v :00.
/ x

/ _\ :00.

 

 

 

 

 

 

 

 

 

 

__ 08.2 8.058.: 80.32 .0 6.50.0

113

The fit indices of the modified model II indicated that the model fit the data
better. GFI and RMSEA showed the model fit the data, although AGFI and CFI
were marginally unsatisfactory for good-fit. The results of the X2 difference test,
which used the modified model I as the baseline model to be compared with,
indicated the modified model II was improved significantly after the modification
(see Table 12). Therefore, the results indicated that the model was a substantial
improvement over the modified model I and is adequate in terms of the relative
fit.

Modified Model lll

Although the modified model II was a fairly good representation of the
data, the model was modified again. This modification was done because the
regression weights of two paths became insignificant in the modification model ll.
Therefore, insignificant paths were removed in the modified model III to create the
desired more parsimonious model. One path was from Type 4 to sense of humor,
which was found to improve the model in the earlier analysis. The other
excluded path was from age to Type 4 programs. This had been a significant
path in the earlier models but not significant after new paths were added. The
other one was the path from Type 4 programs to sense of humor, which
was newly added in the modified model II. Although ML of the modified model ll
indicated that the path from Type 4 programs to sense of humor would improve

the model, the coefficient of the path was not significant. Therefore, the two

paths dropped from the modified model lll (see Figure 9).

114

Table 12.

Summary of Modified Model ll

 

 

 

 

 

Parameter estimates
Parameters Estimates” C.R. Significance
ses _. OCP .85 — 2’ -—
SES H EDU .61 8.12 .00
SES H INC .77 7.19 .00
FVE H FVMD .75 5.76 .00
FVE —> CVMD .55 — 2’ —
FVE H FVMU .62 5.10 .00
IP H PSIP .67 8.91 .00
IP H SSIP .43 5.05 .00
IP H CSIP .82 10.46 .00
IP H MJIP .76 9.88 .00
IF —> SHIP .79 — 2’ —
FVE H TYPE1 .24 2.01 .02
FVE H TYPE2 .39 3.18 .00
FVE H TYPE3 .53 4.58 .00
FVE H TYPE5 .36 3.40 .00
FVE H TYPE6 .17 2.12 .04
AGE H TYPE4 .12 1.61 .11
AGE H TYPE6 .20 2.67 .01
AGE H IP .31 3.77 .00
INC H FVMD .33 4.79 .00
SES H FVE -.45 -3.75 > .00
TYPE4 H SHIP -.07 -1.26 .21
eTYPE2 H eTYPE4 -.21 -2.80 .00
eTYPE3 H eTYPE4 .18 2.30 .02
eTYPE4 H eTYPE6 .24 3.13 .00
Goodness of fit summarv

Contrast with

Baseline model
x2 df ledf may... dim.ference GFI AGFI CFI RMSEA

Modell 178.72 132 1.35 — —- .90 .87 .77 .04
Model II 150.13 128 1.17 28.59“ 4 .92 .89 .89 .03

 

 

 

115

Table 12 (Cont’d)

 

Modification indices

 

Path Modification index
eFVMD H AGE 14.53
eFVMU H eTYPE1 8.52
FVMU H TYPE1 8.09
eOCP H eFVMD 5.83
TYPE2 H EDU 5.64
EDU H TYPE1 5.35
eEDU H eTYPE2 5.28

 

Note. 1) Standardized regression weights, 2) Constrained parameters to 1.0.
** p < .01.

The results in Table 13 showed that the model fit was slightly improved
after modification but the achievement in increasing the fit was trivial; the
improvement was not significant according to the X2 difference test. In the
modified model III, however, there were no insignificant paths, so the hypothetical
model became simpler. Because a more parsimonious model is a better model,
the modified model III was selected as the final model, as it has the most
adequate and satisfactory fit representing the pathways of the effects among

latent and observed variables.

116

.5. v 0.... .00. v 0.

..0..0 .:0E0.00005_ .0

..0E0: .0 00:00 0.10 $0.. 00.205. ”0.05. $..0:0.:000 02.2000 ”0.00 .$._0:0.:000 .0000 “0.00 $200.50 .00.0$:0
”0.00 .00< M02. .0E0.00.0 o 00$ P 05303 .0 .:00E< .0 m0>._. 0.00090 0 00$ 0. 05303 .0 .:00:.< .0 mat 0:.0.00.0
v 00$ .. 05303 .0 .:00E< ”.4 m0>k 080.090 0 00$ .. 05303 .0 .:00:.< .m m0>... 0:.0.00.0 N 00$ 0. 05.503 .0 .:00E<
.N 00>... .0E0.00.0 . 00$ ._. 05303 .0 .:00E< H. 00>... .000 0.00:. .0003 .0 .:00:.< .055“. 000300 000.03... .0 .00E0z
.n.5.>0 000300 0.00:. .0003 .0 00:52 ”92>“. 05.00000 0.000.. $__E0n. ”Dom. 0.0005 000.0 .02. 05.000000 0.000:
2.80". .000 00050300 02.05005 .0. .230. 0 5 .:0:.:0.3:0 05.505 m.>n. 00.0.0 0_E0:0000_0ow .wmw 00:00.02
0.. 0. 0.0.0E0.00 0050..0:00 C 0.02

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

«NF.
m0<
\ 00 0 > El 0 00>. .350 0530 053.
4/ \
Q_Iw /xi~.N®. 3mm. kmfi.
7 for”. oimN. m mm: 05F. // \\
/ / .\
/ ..\ :00.
00$”. .\
0:22
\MvaQtw Iv .v wn_>._i
0.1mm. w>l .5me
150 ,/v mutt... Iv m wn_>._.
. 3mm.
«CNN 1
/ .
0.00 7,000.6. I. N 00>. .0. \
n:wn_ _. mat Dow 02— 100

 

 

 

 

 

 

 

 

 

 

 

 

E 08.2 8.3582. 8.0.8.2 .0 0.00.0

117

Table 13.

Summary of Modified Model Ill

 

 

 

 

 

 

 

Parameter estimates
Parameters Estimates” C.R. Significance
SES —> OCP .85 —2’ —
SES —* INC .61 7.18 .00
SES -> EDU .77 8.11 .00
FVE —> FVMD .75 5.71 .00
FVE —+ CVMD .55 —2’ —
FVE —> FVMU .62 5.10 .00
IP ——> PSIP .67 8.90 .00
IP —> SSIP .43 5.06 .00
IP —-> CSIP .82 10.45 .00
IP -—> MJIP .76 9.87 .00
IP —> SHIP .79 —2’ —
FVE —> TYPE1 .24 2.30 .02
FVE —> TYPE2 .38 3.77 .00
FVE -+ TYPE3 .53 4.51 .00
FVE —> TYPE5 .36 3.42 .00
FVE —> TYPE6 .17 2.90 .04
AGE —> TYPE6 .17 2.68 .01
AGE —> IP .31 3.71 .00
INC —’ FVMD .33 4.79 .00
SES —> FVE -.45 -3.73 .00
eTYPE2 H eTYPE4 -.23 -3.06 .00
eTYPE3 H eTYPE4 .19 2.30 .01
eTYPE4 H eTYPE6 .25 3.39 .00
goodness of fit sumearv

Contrast with

Baseline model

x2 df ledf fidmm dfdmeme GFI AGFI CFI RMSEA

Model II 150.13 128 1.17 — — .92 .89 .89 .03
Model Ill 153.97 130 1.18 3.84 2 .92 .89 .89 .03

 

118

Table 13 (Cont’d)

 

Modification irfiices

 

Path Modification index
eFVMD H AGE 14.79
EFVMU H eTYPE1 8.63
FVMU H TYPE1 8.20
eOCP H eFVMD 5.82
TYPE2 H EDU 5.54
EDU —> TYPE1 5.28
eEDU H eTYPE2 5.16

 

Note. 1) Standardized regression weights, 2) Constrained parameters to 1.0.

Multiple Regression

As stated above, parental mediation was dropped from the hypothetical
model and the effects of the construct were not examined using SEM.
Accordingly, multiple regression analyses were performed to test the hypothesis
2 as a supplementary analysis. As shown in Table 14, no statistically significant
variable of socioeconomic status influencing either instructive mediation or
restrictive mediation was found. The levels of income, occupation and education
of parents were unlikely to affect their mediation while their child watched
television, video, or DVD. Therefore, hypothesis 2 was not supported with the

data.

119

Table 14.

Multiple Regression of Indicators of Parental Mediation

 

 

 

Predictor Variables Instructive Mediation Restrictive Mediation
OCP .1 1 .19
EDU -.08 -.08
INC" .09 .01
F 1.20 1.62
R Square .00 .01

 

Note. 1) Z scores.
Reference. OCP: Family head’s occupation, EDU: Family head’s education, INC:
Gross income.

*p < .05, **p < .01.

Hierarchical Regression

Hierarchical regression analyses were used to examine the contributions
of instructive mediation and restrictive mediation on six different types of
programs after the effects of other predictors, including indicators of
socioeconomic status indicators and the family viewing environment, were
controlled. The results of the supplementary analyses provide partial support for
hypothesis 4. As shown in Table 15, there were only two significant relationships
between the two indicators of parental mediation and program types when the

effects of other indicators were controlled.

120

Table 15.

Hierarchical Regression of Indicators of Children Visual Media Uses

 

 

 

 

 

 

 

 

predictor TYPE 1 TYPE 2 TYPE 3
Variables M0de|1 Mode|2 Model1 Model2 Mode|1 Model2
OCP .05 .06 .01 .02 -.15 -.14
EDU .08 .08 -.20* -.21 .12 .11
INC" -.05 -.04 -.15 -.15 -.02 -.02
FVMD .05 .05 .13 .1 1 .34" .32"
CVMD -.12 -.11 .01 -.00 -.07 -.08
FMVU” .35" .36* .09 06* .19" .18“
IPMZ’ -.14* .10 .08
RPM” .05 -.16* -.13
F 4.40" 3.64" 5.31" 4.74" 6.94“ 5.72“
RSquare .09 .10 .11 .13 .15 .16
R Square Change .01 .02 .01
predictor TYPE 4 TYPE 5 TYPE 6
Variables Model1 Model2 Mode|1 Mode|2 Mode|1 Model2
OCP -.08 -.07 -.26** -.26** -.08 -.09
EDU .03 .02 .08 .08 .06 .07
INC" .01 .00 -.1o -.11 -.18* -.19*
FVMD .04 .03 .16“ .16" .10 .11
CVMD -.10 -.11 -.12 -.13 -.01 -.01
FMVU" -.04 -.05 .09 .09 .06 .06
IPMZ’ .13 .09 .09
RPM” -.10 -.06 -.02
F .35 .71 4.66" 3.71“ 1.84 1.57
R Square -.02 -.01 .10 .10 .02 .02
R Square Change .01 .00 .00

 

Note. 1) 2 scores, 2) Converted using square root transformation.
*p < .05, “p < .01.

121

First, instructive mediation was found to be negatively predictive of
viewing Type 1 programs. In other words, the more the parents of the family
instructively mediated what their child viewed, the less Type 1 programs their
child viewed. This result was contrary to the hypothesis 4.1 that assumed
positive effects of parental mediation on the uses of Type 1 programs (r = -.14,
p = .05).

Second, restrictive mediation negatively predicted uses of Type 2
programs. That is, children whose parents restrictively mediated were likely to
watch less Type 2 programs. This result partially supported the hypothesis 4.1
that assumed negative effects of parental mediation on the uses of Type 2, Type
3, Type 4, Type 5, and Type 6 programs (r = -.16, p = .03).

Despite the significant relationships from parental mediation variables, R
Square Change after the two predictors added were very trivial even in the
models of Type 1 and 2 programs. It implies that instructive mediation and
restrictive mediation are the predictor variables are not good predictors of the

criterion variable.

Summary
A summary of the research hypotheses tested and the findings of the

study is presented in Table 16.

122

Table 16.

Summary of Hypotheses and Findings

 

 

 

 

 

 

 

 

 

 

 

 

 

Hypotheses Findings

H 1. The higher the socioeconomic status of a family is, the lower Supported
the level of the family viewing environment.

H 2. The higher the socioeconomic status of a family is, the more Not supported
the parents of the family mediate what their child view.

H 3. The higher the level of the family viewing environment is, the Partially
more visual media a child in the family uses. supported

H 4.1. The more the parents of the family mediate what their child Not supported
view, the more Type 1 programs their child views.

H 4.2. The more the parents of the family mediate what their child Partially
view, the more Type 2, Type 3, Type 4, Type 5, and Type 6 supported
programs their child views.

H 5.1. Younger children spend more time viewing Type 1 programs Partially
than older children; older children spend more time viewing supported
Type 2, Type 3, Type 4, Type 5, and Type 6 programs than
younger children.

H 5.2. Girls spend more time viewing Type 1, Type 2, Type 4, and Not supported
Type 5 programs than boys; boys spend more time viewing
Type 3 and Type 6 programs than girls.

H 6.1. The more Type 1 programs children watch, the more playful Not supported
they are.

H 6.2. The more Type 2, Type 3, Type 4, Type 5 and Type 6 Not supported
programs children watch, the less playful they are.

H 7.1. Older children are rated higher than younger children in Supported
playfulness including physical spontaneity, social
spontaneity, cognitive spontaneity, manifest joy, and sense
of humor in imaginative play.

H 7.2. Boys are rated higher than girls in physical spontaneity, Not supported
manifest joy, sense of humor in imaginative play; girls are
rated higher than boys in social spontaneity and cognitive
spontaneity in imaginative play.

H 8. The hypothesized model of the relationship of socioeconomic Partially
status, children’s characteristics, and three mediating supported

variables including the viewing environment in a family,
parental mediation, and children’s visual media use to the
level of imaginative playfulness in this study fits the data.

123

CHAPTER 5

DISCUSSION

The present study investigated the factors affecting children’s visual media
use and their effects on their imaginative playfulness. Overall, this research
found that children’s visual media use was influenced by its predictors, including
socioeconomic status and the family viewing environment. Socioeconomic
status contributed to the viewing environments in the family, and in turn, the
family viewing environments influenced the children’s viewing practices.
However, the children’s viewing did not affect their imaginative playfulness.
Because no association between children’s viewing practices and their
imaginative playfulness was found, this study failed to find the pathways from
children’s visual media use and imaginative playfulness, even though the fit
indices of the final model, including all the variables, were acceptable. In this
chapter, the results of each hypothesis are given and suggestions for future

research are made.

Socioeconomic Status and Family Viewing Environment

Socioeconomic status, which was indicated by income, education, and
occupation, negatively affected the family viewing environment, supporting
Hypothesis 1. Therefore, this study confirmed numerous past studies
demonstrating that socioeconomic status is an important predictor of the viewing

environment in a family (Anderson et al., 2001; Bianchi & Robinson, 1997;

124

Tangney & Feshbach, 1988).

An interesting predictor was family income, which was an indicator of
socioeconomic status, because it positively predicted the number of visual media
devices in a family. That is, the higher the income level of a family, the more
visual media devices the family owned. The pathway between two indicators
was exclusive because the finding was contradictory to the relationship between
the latent variables that each indicator represented. That is, income positively
predicted the number of visual media devices in a family, while socioeconomic
status negatively contributed to family viewing environment. This distinctive
pathway between the two indicators, which might be imbedded in the effects of
socioeconomic status on family viewing environment, was identified using SEM
in this study. It is reasonable that affordability is an important factor in the
possession of visual media devices in a family. Therefore, family income should
be carefully considered as a unique factor when the environmental

characteristics in a family are studied.

Parental Mediation
Parental mediation has been a key variable predicting children's viewing in
previous literature. Despite the importance, parental mediation was excluded in
the hypothetical models for several reasons after the measurement model testing.
The measurement model test found the indicators of parental mediation were not
adequate for analyses using SEM; the coefficients of the indicators were not

significant. Therefore, hypothesis 2 and 4 were not tested with the SEM models.

125

The main rationale of insignificant coefficients can be said to be a problem
with the measure. Even after the exploratory factor analysis showed the
existence of two variables, instructive mediation and restrictive mediation did not
represent parental mediation significantly in the SEM models. In other words, the
results of the measurement model testing showed the coefficients for instructive
mediation and restrictive mediation were not significant.

There are several reasons for the inadequate measure that can be
considered. Statistically speaking, first, a measurement model with a latent
variable indicated by only two manifest variables is not recommended for SEM
because too many constraints are needed to identify the measurement model
(Kline, 1998). Parental mediation in this study was a construct represented by
only two indicators. Therefore, it might result in an inadequate measure.

Another possible reason why the parental mediation measure was not a
good one can be explained by difficulties in adapting the measure of Valkenburg
and her colleagues (1999). First, the original measure was designed to assess
parental mediation of elementary school children, whose viewing practices and
whose parents’ interaction might be different from preschoolers. According to the
parents who did not provide all the answers in the Parental Mediation
Questionnaire, for example, interaction related to items 6 and 10 barely occurred
in their family (see the Appendix F). The parents commented that preschoolers'
viewing practices were likely to be controlled by their parents before viewing
activities occur, because the parents mediated before their child’s viewing begins,

not while their child is viewing. Although the social coviewing items in the original

126

measure were excluded for an appropriate measure for parental mediation of
preschoolers, some of the other items might not be sufficiently applicable to the
preschool children.

Another characteristic of the Valkenburg’s measure is that it was designed
to assess parental mediation of children's viewing of television, while this
research uniquely included their mediation of children’s viewing of videos and
DVDs. Because young children needed parental approval and technical
supports in order to view videos and DVDs, the parents’ restrictions on an
unsuitable program that their child was watching (item 6) and their prior-to-
specification of the program that their child could view (item 10) seems to be
inappropriate. That is, such mediation may differ when their child is viewing
television from when their child is viewing videos and DVDs. Accordingly, the
adaptation of the original measure to assess parental mediation of videos and
DVDs uses might have created unanticipated measurement problem.

These reasons for unsuccessful adaptation of the measure imply that
parental mediation may be beyond interpersonal and verbal actions that are
traditionally regarded as mediating strategies; In addition to such mediation,
parental mediation may include structuring the physical and psychological
environments related to children’s viewing. Therefore, it would be valuable to
develop a measure to investigate different practices of parental mediation of
young children during the viewing of videos and DVDs with questions considering

this matter.

127

Socioeconomic Status and Parental Mediation

In order not to lose a chance to examine the relationships related to
parental mediation, multiple regression analyses were used as supplementary
analyses. There was no relationship between socioeconomic status and parental
mediation, so hypothesis 2 was not supported with the data. Accordingly, this
result of multiple regression implies that retention of the parental mediation
measurement model would also be insignificant and would not improve fits of the
SEM models.

As reviewed in Chapter 2, the influences of socioeconomic status on
parental mediation have not been consistent. The findings of multiple regression,
which indicate that parents’ socioeconomic status did not contribute to parental
mediation of children’s viewing, provide evidence that supports no influence of
socioeconomic status on parental mediation. Thus, the data strongly suggest

that other factors should be taken into account in any study of parental mediation.

Parental Mediation and Children's Visual Media Use
The results of hierarchical regression showed that the effects of parental
mediation were trivial. There were only two significant relationships between the
two indicators of parental mediation and six different program types; instructive
mediation was found to be negatively predictive of viewing Type 1 programs and

restrictive mediation also negatively predicted uses of Type 2 programs.
It is an interesting finding that children whose parents mediated in an

instructive way tended to watch fewer Type 1 programs, because it was contrary

128

to the hypothesis 4.1. Instructive mediation is the parents’ efforts to enhance
their child’s learning from visual media through providing comments to maximize
positive effects and minimize negative effects of viewing. Instructive mediation is
the active, deliberate parental behaviors to enhance children’s knowledge and
behaviors. The parents try to make sense of and evaluate content of what they
watch (Austin et al., 1990), and to increase children's critical viewing skills,
thereby to make children more capable of managing their own interactions with
visual media (Door & Rabin, 1995). Therefore, parents who mediate in an
instructive way may be more concerned about cognitive functioning and mental
skills. Nevertheless, children’s visual media use is basically a passive and
receptive process that does not involve direct hands-on experiences. Perhaps,
such parents are more interested in activities that supplant viewing. Accordingly,
an alternative reason might be that parents who instructively mediate engage
children in other activities instead of watching even Type 1 programs that are

well designed programs for children.

Another significant path was from restrictive mediation to Type 2
programs; children whose parents mediated what they watch in an instructive
way tended to watch fewer Type 2 programs. Restrictive mediation involves the
implementation and enforcement of parental rules regarding what their children
watch and how much their children view. Parental restrictive mediation tends to
prohibit children from being exposed to television programs with violent, sexual,
or frightening content (Jordan, 2001). Based on this explanation, it can be
speculated that parents tried to mediate the negative influences of programming

that contains violent episodes like Type 3 or 6 programs and programming that is

129

not designed for young children like Type 4, 5, or 6 programs. Accordingly, the
finding of a significant relationship between restrictive mediation and Type 2
programs is an unexpected result. This unexpected finding can be interpreted,
considering the added paths in the SEM models. In the modified model II and III,
systemic patterns between the programs were found because paths between
measurement errors of Type 2, 3, 4, and 6 were found as significant
relationships. Hence, the data indicate that more careful examination of the
systemic patterns between measurement errors would be very helpful to interpret

the effects of restrictive mediation on children’s visual media use.

Socioeconomic Status and Children’s Visual Media Use

The relationships between socioeconomic status and children’s visual
media use were not assumed, although many studies found that the levels of
income, education, and occupation of a family could predict viewing practices of
children from preschoolers to adolescents. In this study, socioeconomic status
did not directly affect children’ viewing practices. Instead, it appears to have had
impact on the more general family viewing environment, which in turn, had an
effect on the children's visual media use. This implies that the association
between socioeconomic status and viewing practices is more complex then
previously understood. Therefore, it will be useful to investigate additional
mediating functions that may exist as well.

As hypothesized, SEM found no path from socioeconomic status to

children's viewing variables, although they were partially correlated with each

130

other. In other words, socioeconomic status contributed to other factors such as
the family viewing environment as originally hypothesized; in turn, the ecological
factors such as the family viewing environment are likely to predict what children
view. Thus, further research focusing on investigating mediators between

socioeconomic status and children’s viewing would be meaningful.

Family Viewing Environment and Children’s Visual Media Use

Hypothesis 3 was partially supported in this study. The number of visual
media devices in a family, the number of children’s devices, and the amount of
visual media use in a family indicated the family viewing environment as a latent
variable as hypothesized. It implies that the family viewing environment can be
understood as a construct that is adequately indicated by a series of
characteristics related to family viewing practices. Accordingly, the results
showed that the viewing environment of a family had positive effects on
children’s media use in the family. Therefore, the findings of this study confirm
many studies showing that features of the family viewing environment are
important predictors of children’s viewing behaviors (Atkin et al., 1991; Bower,
1973; Chaffee & Tims, 1976; Dorr & Rabin, 1995; Huston et al., 1999; Lull, 1978;
McDonald, 1986).

Considering the result supporting Hypothesis 1, it can be said that
socioeconomic status was not a factor that directly predicted children’s viewing
practices; rather, socioeconomic status was a variable that influenced indirectly

the children’s visual media use mediated by the family viewing environment. In

131

particular, the effects of socioeconomic status on visual media use were not
substantial after the effect of the family viewing environment was controlled.
Therefore, the family viewing environment should be interpreted as an important
factor which mediates the effect of socioeconomic status.

Among the six program types, there was one program type that was not
influenced by family viewing environment. The coefficient from the family viewing
environment to Type 4 programs was not significant in the initial model.
Therefore, it can be said that the family viewing environment mainly influenced
the children’s viewing of non-informative programs, including Type 1, 2, 3, 5, and
6 programs. On the other hand, there was no systematic pattern between
viewing Type 4 programs and family viewing environment.

Type 4 programs, which are designed to deliver information without
aggression for general audiences, are likely to include content that may be of
less interest to preschool-aged children and are unlikely to contain the
visual/auditory effects with a rapid pace to catch the eyes of children.
Consequently, the programs that are not attractive to children may not be
watched, even though their family viewing environment is more susceptible for

the children to be exposed to such programs.

Children’s Visual Media Use
Children's visual media use, which was the major variable of this study,
had several characteristics. The descriptive data of visual media use provide

insights into the viewing practices of preschoolers.

132

First, individual differences of visual media use were very large; in
particular, the distributions of visual media use were zero-inflated. Because of
this attribute of the presence of excess zeros, statistics assuming linear
relationships among variables seem to be less appropriate. Accordingly, it was
thought that investigating group differences would be a valid alternative method
for analyses. In order to investigate associations between children’s viewing
practice and other related variables, it will be valid to compare the characteristics
of heavy viewing groups with light-viewing groups in further studies.

Second, the final model indicated that there are many relationships among
measurement errors of program types: a negative relationship between errors of
Type 2 and 4 programs, positive relationships between errors of Type 3 and 4
programs, and errors of Type 4 and 6 programs. These findings are the most
powerful aspects of SEM assuming that observed variables are measured with
errors. This error component is an attempt by SEM to model the imperfection of
measuring the true scores of each variable. In the context of SEM, the
measurement error term is the unique factor associated with each measure.
First, indicators of the same construct might share variance because they were
measured by a common method. Second, the associations among
measurement errors may imply that there were systematic variances that are not
explained by viewing different types of programs. That is, there might be other
variables affecting children’s viewing in addition to the ecological factors included
in this study. In further research, therefore, the efforts to identify the additional

variables will be important.

133

Child Characteristics and Children’s Visual Media Use

Earlier studies showed that children’s gender contributed to their viewing.
Typically, it has been known that boys tend to view more than girls. Despite
these eanier findings, no pathway from gender to visual media use was found in
this study. In particular, the most interesting finding was that even uses of Type
3 and 6 programs, which contained violent episodes, were not different by
gender, although previous research found that programs with violence including
cartoons and action adventure programs are more often watched by boys than
girls among preschool-aged children (Huston et al., 1990; Huston 8. Wright,
1998b; Singer & Singer, 1981).

One of the possible reasons for no gender difference in consuming violent
programs is the inclusion of videos and DVDs. The majority of violent programs
that the preschoolers watched were videos and DVDs for child audiences.
Typical programs that were categorized in this type were the Disney movies,
which are mostly intended for moralizing and didactic themes but include violent
episodes. In addition, female characters as a heroine became more available in
action adventure programs which were likely to be very violent. For example,
Powerpuff Gids and Kim Possible, which were rated as highly violent programs,
were frequently reported in the questionnaire. For these reasons, girls” might
also become frequent viewers of violent programs when their uses of videos and
DVDs are included.

During the preschool years, children are frequent viewers of media. In this

research, preschoolers’ viewing patterns changed with age depending on the

134

types of programs. In the final model, older children viewed more Type 6
programs, which contain violence designed for general audiences with
entertaining purposes, than younger children; on the other hand, uses of the
other programs did not change by age. A large number of reported programs
were movies mostly with action-adventure content such as the Star Wars series,
Jurassic Park series, Mummy, and Back to the Future. According to Cantor
(1994), the viewer’s age is an extremely important determinant of emotional
reactions to media that may cause fright. Based on her statement, it would
appear that older children can endure stress produced by exposure to aggressive
movies and find enjoyment in watching such programs which younger children

cannot.

Child Characteristics and Imaginative Playfulness

Children’s imaginative playfulness was expected to change by age and
gender. However, only children’s age contributed to their imaginative playfulness.
The findings are consistent with Bamett’s (1991b) research investigating
correlates between playfulness and individual attributes: children’s social
spontaneity, cognitive spontaneity, and sense of humor scores increased from
age two to five. The results of this study using SEM, in particular, provide useful
evidence that young children’s imaginative playfulness as a construct as well as
its individual sub-dimensions increases by age. It implies that children become
more playful as they grow through developing their social, cognitive, and physical

skills. Thus, the data strongly suggest that age be taken into account in any

135

study of imaginative playfulness.

No relationship between gender and imaginative playfulness was found.
In addition, no individual path from gender to the individual playfulness sub-
dimensions was found. The result showing that children’s gender did not
contribute to their imaginative playfulness was inconsistent with the findings of
previous studies. Barnett and Kleiber (1982) implemented eany research on the
relationship between playfulness and cognitive functioning including divergent
thinking and intelligence between the ages of three and six years. They found
gender differences between males and females for the mediating effect of
intelligence on the relationship between playfulness and divergent thinking.
Later, Barnett (1991b) found that a significant gender effect was obtained for the
physical spontaneity and manifest joy dimensions in children ranging in age from
21 to 61 months. On the other hand, girls were rated higher than boys in
cognitive spontaneity and no gender differences were detected for either social
spontaneity or sense of humor. Using SEM, however, this research was able to
discover no contribution of children’s gender to any sub-dimension of imaginative
playfulness. As a result, imaginative playfulness may not differ by gender, even

though there are small differences across its dimensions.

Socioeconomic Status and Imaginative Playfulness
It has been widely known that children’s imaginative play varies from child
to child, not only seen from their unique personalities, but also from differences

stemming from their socioeconomic status. The famous work of Smilansky

136

(1968), in discussing sociodramatic play of children from economically
disadvantaged children, sparked this interest. In particular, Smilansky and other
researchers, influenced by her work, have characterized the imaginative play of
disadvantaged children, compared to that of middle-class children as of lower
quality (Weiner & Weiner, 1974). Typically, their imaginative play was described
as “unimaginative”, "repetitive”, “simplistic”, “desultory”, and “concrete” (Johnson
et al., 1999, p. 131). Accordingly, children’s imaginative playfulness indicating its
quality of imaginative play was also assumed to be negatively related to their
socioeconomic status.

As demonstrated in Chapter 4, however, neither direct nor indirect
pathway from socioeconomic status to imaginative playfulness in the hypothetical
models was identified. It implies that there may be other mediators between
socioeconomic status and imaginative playfulness. Thus, the effort to identify
other variables that mediate the influences of socioeconomic status on
imaginative playfulness will be useful for further understanding of these

dimensions.

Children’s Visual Media Use and Imaginative Playfulness
The key pathway of this study was from children’s visual media use to
their playfulness. Despite the wealth of literature speculating the link between
children’s viewing practices and their quality of imaginative play, children’s
viewing practices did not make a significant contribution to their imaginative

playfulness regardless of the types of programs. This finding is consistent with

137

the conclusion of van der Voort and Valkenburg (1994) who reviewed the
empirical research on this relationship. The results of this study will be discussed
with regard to the program types that have been the main focus of previous
literature.
Children’s Use of Educational Programs for Child Audiences and
Imaginative Playfulness

The first program type that has been studied is well-designed programs
with educational purposes for young children (Type 1 programs). In this study,
Type 1 programs were assumed to promote children’s playfulness during their
imaginative play. Type 1 programs were deliberately designed to enhance
children’s physical, language, social, and logical knowledge. Accordingly, the
children who watched more Type 1 programs were assumed to show enhanced
imaginative play. Contrary to the hypothesis, no relationship between children’s
visual media use and imaginative playfulness was found. In addition to this
result, SEM also failed to find any individual path from the types of visual media
programs to the indicators of imaginative playfulness; even cognitive spontaneity
and social spontaneity in imaginative playfulness, which might be more directly
related to the information that Type 1 programs deliver, were not influenced by
viewing Type 1 programs.

There are many studies showing that children learn and enhance their
cognitive ability and social skills by watching and learning from children’s
educational programs; these programs such as Sesame street, Bamey and

friends, and Mr. Rogers’neighborhood were effective in promoting particular

138

attributes, including cognitive skills, language, school readiness, and prosocial
interactions (Comstock, 1991; Huston & Wright, 1998b; Singer 8 Signer, 1998;
Zill et al., 1994). However, the results of this research showed that the effects of
learning cognitive and social skills through watching these programs were not
enough to be incorporated in children’s imaginative play and ultimately enhance
the quality of such play. As pointed out eartier, children’s imaginative play
shows the development of each child comprehensively. If watching Type 1
programs does not enhance playfulness in imaginative play, it implies that
knowledge and skills that children acquire from Type 1 programs are not fully
assimilated. lnforrnation delivered in Type 1 programs tends to be fairly easily
acquired and readily forgotten or not sufficiently assimilated into youngsters'
performance. Thus, it would be very beneficial to develop educational programs
that are designed to promote children’s dispositions such as imagination ability
and self-control that can encourage positive development and quality play.
Children ’3 Use of Violent Programs and Imaginative Playfulness

The second kind of programs that have been debated were those
containing aggressive episodes. In this study, it was thought viewing violent
programs would result in a lower quality of imaginative play. Contrary to the
hypothesis, neither Type 3 nor Type 6 programs, which were programs
containing violent segments, predicted children’s imaginative playfulness in this
study.

Preschoolers’ imaginative play with aggression involves frequent uses of

fantasy superheroes, play-fighting, and/or war play, and has been blamed due to

139

its low quality (James & McCain, 1982; Johnson et al., 1999). Mass media,
particularly television with violent content, has been blamed as the source of
such play. Although the aggressive imaginative play has been recently coupled
with mass media, children’s pretend play involving violence has existed in
various themes and forms before the mass media was prevalent; it has been
common with popular toys, such as bows and arrows, spears, or daggers,
usually made of sticks, reeds, or grasses, and guns that shoot corn, peas, or
thorns in children’s imaginative play (Power, 2000). Children also engage in
symbolic play-fighting in other cultures. For example, children in New Guinea
pretend to be sharks in their play-fighting. Sutton-Smith (1988) pointed out that
mass media may influence, but is largely subordinate to, preexisting themes in
children’s play. That is, violence components of mass media are used and
transformed by children because it contributes to their desired play scripts. The
interpretation of Sutton-Smith is consistent with the major play theorists of our
time, Garvey (1977), Schwartzman (1978), and Fein (1975), who consider
imaginative play to be largely a transformational, rather than an imitative activity.
If the themes, scripts, characters, and other attributes of aggressive programs
are to serve as instruments or vehicles to satisfy certain developmental needs in
imaginative play as Sutton-Smith pointed out, it is understandable that children’s
viewing of violent programs does not harm the children’s imaginative playfulness
because imaginative play with aggression dealing with children’s developmental
needs would occur among children who do not watch the violent programs as

well as among heavy viewers.

140

Play materials have been regarded as an important factor in children’s
imaginative play. The effects of commercialized toys typically tied in to mass
media have been considered when the influences of mass media on children’s
play is discussed. In particular, the toys related to the characters of action-
adventure programs and war toys have been the center of such concern because
the toys do provoke the behaviors, scripts, and themes relevant to their character
during imaginative play (Carlson-Paige & Levin, 1988; Pott et al., 1986). Even
Sutton-Smith (1986), who appreciated children’s imaginative play with
aggression, was concerned that the range of what children will think about has
been increasingly influenced by commercialized toys, which are tied to mass
media. Consequently, Line and Pentecost (1990) insisted that the link that is
being forged between television fiction and toys may have significant implications
for the styles and content of children’s imaginative play. One contextual
characteristic of commercialized toys tied to mass media that should be
considered is that such play materials are rarely available in school settings
(Shin, 2001). On the other hand, Davie et al. (1984), who observed children’s
play at home, found how important these young people’s heroes still are for their
pretending. Allowing for the above, further research investigating the effects of
viewing programs with violence on playfulness in imaginative play should
consider pretend play ’at home’ where commercialized toys are more available.

Contextual factor is another ecological factor that should be discussed in
interpreting the findings of no relationship between viewing programs with

violence and imaginative playfulness. According to King (1982), there are three

141

types of play at school: instrumental play, real play, and illicit play. Instrumental
play includes activities that are required by, controlled by, and evaluated by the
teachers. Real play is recreation that includes voluntary and self—directed
activities. Illicit play, which is unauthorized and surreptitious interaction, is
unlikely to be permitted at school. It includes such actions as making faces,
goofing off, or giggling. It has been widely accepted that children’s play with
pretend action-adventure characters containing violence is discouraged. That is,
imaginative play incorporating the content in Type 3 or Type 6 programs is
regarded as illicit play and is likely to be restrained at school where instrumental
play is preferred. Due to this reason, school settings might not be the best
context where children act out what they would like to pretend. Consequently,
the imaginative playfulness that teachers assessed in this study might be found
not to be influenced by viewing Type 3 or Type 6 programs. Therefore, further
studies investigating the effects of viewing programs with violence on playfulness
in imaginative play should consider pretend play ‘at home’ where children are
able to play out what they see on television and video with fewer restrictions.
Although imaginative play influenced by watching aggressive programs
may contribute to child development and tend to be unavailable even at school
settings (Shin, 2001), there has been increasing and wide-spread vocal
opposition and concerns of educators about perceived violence involved in
children’s imaginative play. The opponents blame violent programs as the
sources of play with aggressions in various position papers (Carlsson-Paige 8.

Levin, 1987, 1990). Therefore, strict opponents of children’s imaginative play

142

with violence insist children should be discouraged from engaging in such kinds
of play (Nilsson, 1995); even if scholars believe that such behavior serves
important developmental functions, they propose that it should be channeled into
more constructive activities (Carlsson-Paige & Levin, 1987, 1990). Considering
the results of this study, the reason for debates on the effects of viewing violent
programs on the quality of children’s imaginative play can be questioned
(Carlsson-Paige & Levin, 1990; Crosser, 1995; Levin, 1998; Levin & Carlsson-
Paige, 1989).

There are several reasons why many teachers as well as researchers
have become aware of pretend play with violence. First, children’s playful
aggression is easily misinterpreted as real aggression in school settings.
According to Connor (1989) who had various adults interpret children’s play,
people use their own subjective criteria in classifying a particular behavior as
aggressive. Observer characteristics, such as gender and occupation, were an
important determinant of what is considered as aggressive behavior even though
most children can differentiate between play-fighting and real aggression (Smith
& Lewis, 1985; Snow, 1974). Second, according to Boyd (1997), what teachers
are sincerely concerned with was not about the developmental outcomes of
violent play but the safety of children when children engaged in such play at
school. Third, the contemporary trend toward controlling real school violence is
likely to lead to restrictions on any episode involving perceived aggression in
school settings (Achenbach et al., 2002; Lickel et al., 2003; Slater, 2003).

Therefore, it can be said that the arguments on the negative aspects of pretend

143

play influenced by viewing Type 3 and 6 programs may not be based on the
concerns about the quality of children’s imaginative play but have gone beyond
an academic debate about play to concerns about general aspects of child
behavior related to aggression.

To conclude, the findings of this research show that the use of visual
media on imaginative playfulness cannot be justified as beneficial or harmful,
despite the increasing concerns and debates at the conceptual level on the
negative effects of viewing on the quality of imaginative play. As Lull (1980)
argued, however, the nature of the relationship between imaginative play and
viewing practices would be more implicit and subtle than explicit, if the influences
of visual media use are embedded in the children’s daily activities and are
unintentional and inconspicuous. Children’s visual media use does not affect
imaginative playfulness, but it would significantly contribute to marketing media
tied in to toys that leads certain play themes, producing children’s shared
experiences that can be sources of their imaginative play, and triggering peer
pressures as to what to play. Therefore, a more sophisticated model related to
the variables should be hypothesized to explore the pathways from visual media

use to imaginative playfulness.

Limitations
Although this research provided empirical data examining the relationships
between different program types and imaginative playfulness, there are several

limitations that should be considered when the findings are interpreted.

144

The measure of visual media use can be an issue. In this study, the title
roll and the advertisements during television shows were not included for
categorization. In the case of serial television shows, only one randomly
selected episode was represented. In the case of videos and DVDs, previews
arranged before each show or movie were not included either. Because the
content of programs is inconsistent with the characteristics of their title roll,
advertisements or previews, the effects of visual media use might be
contaminated due to the exclusion of children’s experiences that they were
exposed to but not analyzed.

Among the three dimensions of the program categorization, the violence
components should be also discussed. Verbal violence, which was traditionally
overlooked as compared to physical violence, has been the focus of scholars of
mass media (Glascock, 2003; Greenberg et al., 1980) as well as
developmentalists studying young children (Halperin et al., 2003; Kostelnik et al.,
2002). Despite the significance and prevalence of verbal violence in
programming, it was not included in this study due to several reasons, which
follow. When children’s imaginative play was debated in terms of the effects of
viewing aggressive programming, mostly action adventure programs were
reproached (Carlsson-Paige 8. Levin, 1990; Crosser, 1995; Frontis-Diaz, 1997;
Levin, 1998; Levin & Carlsson-Paige, 1989) because it is more easily imitated in
imaginative play by young children than verbal aggression. In particular, most of
programs were likely to be categorized as violent programs when verbal

aggression was included because verbal aggression is very prevalent in

145

programming. For these reasons, it was decided that the inclusion of verbal
aggression would not be helpful for the valid categorization of violent/nonviolent
programs. Therefore, the programs with violence in this research should be
carefully interpreted considering the scope of the meaning of violence.

Low response rate can be another limitation of this study. Only 22.6
percent of the parents who were invited to participated in this research completed
and returned their survey packet. The parents who did not participate in this
study might have felt it an imposition to complete three different questionnaires
that were fairly long. Also, they might have some items that they did not feel
comfortable in reporting. The most common missing item that participating
parents were reluctant to report was family income, although their confidentiality
was guaranteed. Because certain characteristics of parents who refused to
participate in this research is assumed, it is important to this point to be careful in
generalizing the results of this study despite the large sample size.

There is another problem of generalization in terms of the final model.
SEM was very useful in modifying the initial model to find the final model
explaining the relationships between visual media use, imaginative playfulness,
and their ecological factors well. Although it is statistically acceptable to modify
the initial model based on the results of modification indices to improve the
overall fit, the final model has many weights in the model. Accordingly, the error
of the final model may be estimated when the same coefficients and model fit are
computed in another sample. Because of the potential cross validation problem,

the prediction errors of the final model can be estimated. Therefore,

146

generalization of the final model of this research should be done carefully, even

though there is a large sample size.

147

APPENDICES

148

APPENDIXA Rating Instructions

Dimension 1. Intended audiences

1 = Programs intended primarily for viewers 12 years old or younger: Programs for child

audiences.

2 = Programs not intended primarily for child audiences: Programs for general audiences.

- Decision is based on producer’s intentions.

- Focus on the primary target audiences of each program. General programming can

entertain children under 13-year-olds; general audiences also can enjoy children’s

program.

- When intended audiences are not clear, consider the following components

1)

2)

3)

4)

5)

Age of a main actor/actress: A program whose main actor/actress’s age is under
13 can be children’s programming.

Theme: A program dealing with the issues or interests of young children can be
children’s programming.

Complexity of story: A program with a simple storyline can be children’s
programming. Programs for children do not necessarily satisfy all three
components. For example, Harry Potter whose main actor is under 13 and whose
theme deals with children’s interest can be classified as children’s programming
despite the complex story. These components can be referred when target
viewers are both children and general audiences, mostly family movies.

Ratings of which TV Guide: Television shows, which TV guide classified as
children’s programs, mostly indicate child programming.

Rating of the Motion Picture Association of America (MAPP): Movies, which are
rated G, mostly indicate child programming; PG13 movies are mostly programs
for general audiences; however, PG movies cannot be automatically categorized

as programs for child audiences (e. g. Star Wars).

149

Dimension 2. Violence components

1 = No violence:

2:

3:

4:

There is no episode with violence. 6. g. Cosby Show, Barney and Friends, Fantasia,
Winnie the Pooh

A very little violence:

There are only several short episodes with violence. e. g. Muppets, Snow White, Cops .
Medium violence:

There are several long episodes or somewhat frequent short episodes with violence.

6. g. Bugs Bunny, Lady & the Tramp, Harry Potter, Little Mermaid

High violence:

There are long episodes with violence appearing very frequently through programs.

e.g. Mummy, Powerpuff Girls, Power Ranger

Decision is based on an aggressor, not a victim.

For the purposes of rating programs, a violent act is a physical act intended to harm

someone. In other words, it is visible, intentional, and interpersonal.

What to include: Acts that include all the following three components are regarded as

violence.

1) Visible: Violent acts include visual portrayals of physical performances of
aggressors. Physical aggression includes assaulting without an object (6. g. hitting,
kicking, pushing, pinching, strangling, scratching, grabbing, jerking), assaulting
with an object (e. g. bombing, burning, shooting, stabbing), detention, chasing,
physical scaring, and physical threatening (e. g. pointing a gun, physically
menacing others, holding knife against others). Invisible violence episodes in
which explicit acts of aggression are not shown do not count as violence, although
commitment of violent acts is implied. For example, a detective’s investigation
after a crime does not count as a violent act. Also, verbal aggression that involves
sending noxious symbolic messages (0. g. messages with criticism, insult, cursing,
a negative affective reaction, verbal threat, yelling, screaming, shouting) generally
does not count as violence, unless the verbal act results in a physical act.

2) Intentional: Violent acts include antisocial acts with the potential to do harm, but

150

3)

from which the victim escapes uninjured. Accidental aggression is not included.
For example, if someone tries to run down another individual with an automobile,
it counts as a violent act whether or not he/she is successful; on the other hand, if
someone accidentally crashes into another automobile in which a person is injured,
this is not a violent act by our definition.

Interpersonal: Violent acts must be performed by a human or a character (e. g.
Stuart Little or King Kong) against another human or character. Episodes of
property damage or injuring/killing uncharacterized animals and beings, which

are not intended to harm others would not be rated as violent acts. For example, if
an angry throws plates on the ground and breaks them does not count as a violent
act. Episodes of property damage or injuring/killing uncharacterized animals,
which are intended to harm others would be rated as violent acts. For example, if

a couple who are angry to each other breaks other’s items are rated as violence.

- What not to include

1)

2)

3)

Slapstick episodes or adventure episodes driven by catastrophes (e. g. fire) and
acts of nature (0. g. volcanoes, tomados) are not regarded as violence although
people and personified character are harmed.

Common and tolerated physically aggressive behaviors (e. g. contact sports,
hunting and fishing for game animals, hunting or butchering of animals in real
world events, and legal demolition of property) are not considered as aggression.
Make sure that aggression is one behavioral category of antisocial behaviors.
Theft and deceit, which are antisocial behaviors, would not be regarded as

violence.

151

Dimension 3. Informative Purposes

1 = Program primarily designed to be educational or informative.

2 = Program not primarily designed to be contain educational or informative.

- This distinction refers to apparent intention of the producer.

- It does not refer to the entertainment value. A program may be both educational and
entertaining. In particular, educational programs designed for children can be highly
entertaining.

- Programs for general audience: Programs about real world events are usually
informative. These include news, public affairs, sports events, religious services,
cultural events, discussions and analysis.

- Programs for child audience: Programs are primarily designed to contribute the
following items.

1) Providing information on knowledge or cognitive skills.

2) Helping children to develop constructive prosocial attitudes or actions (e.g.,
sharing, turn taking, self-esteem, self-restraint, cooperation, manner).

3) Exposing children to the varieties of culturally and physically different types of
children and adults and fostering a sense of mutual respect for these cultural and
physical differences (e. g., ethnic diversity, disability, single parent family).

4) Encouraging awareness of one’s own and others’ emotional responses.

- The messages should be specific. Do not include just because a social message or

religious beliefs could be extracted.

152

APPENDIX B Explanatory Letter

(for primary caregivers)

Dear Parents:

I am a doctoral student at Michigan State University in the Department of Family and
Child Ecology. I am conducting a study about the relationship between visual media use
practices and the play of young children. This research is being conducted under the

supervision of Dr. Alice Whiren.

Attached are three sets of questionnaires that you are being asked to complete, if you
choose to participate. It is hoped that you will take time to complete these materials. It
should take no more than 30 minutes of your time. Also, your participation will include
granting permission to the researcher to gather research information from your child’s
head teacher. The teacher will be asked to share information about your child’s playful

behavior, e.g., tells funny stories during imaginative play.

Because your child’s name will be written on this survey packet and on the questionnaire
that your head teachers will complete, your reports may be identified. Your privacy will
be protected to the maximum extent allowable by law and your responses will be kept
confidential. To protect against a breach of confidentiality, all forms I collect will be
stored in a secure place that is accessible only to project staff, and no names will be used
in any report of the research findings. After your questionnaires are coded, they will be
destroyed. Your responses will not be discussed with school staff, and no private

information from school records will be used.
Your participation is voluntary and you have the right to refuse to participate in this

study. Further, you may discontinue your participation at any time. Also, your non-

participation will not in any way be used to penalize you.

153

If you have any question about this study, please feel free to contact me by phone: (517)
272-2980, e-mail: shinnary@msu.edu, or regular mail: 3015 Staten Ave. #5, Lansing, MI
48910, or Dr. Alice Whiren, by phone: (517) 432-0893, e-mail: whiren@msu.edu, or
regular mail: 3 Human Ecology, East Lansing, MI 48824. If you have any questions or
concerns regarding your right as a research participant, or are dissatisfied at any time
with any aspect of this study, you may contact - anonymously, if you wish - Dr. Ashir
Kumar, Chair of the University Committee on Research Involving Human Subjects
(UCRIHS) by phone: (517) 355-2180, fax: (517) 432-4503, e-mail: uchris@msu.edu, or
regular mail: 202 Olds Hall, East Lansing, MI 48824.

I appreciate your time and cooperation in advance. In appreciation of your participation
in this research, you will receive two children’s books upon receipt of your completed
questionnaires. Also, your head teacher will receive a gift certificate for the classroom
because of your participation and completion of their questionnaire. You and your head
teachers will not receive books if you do not complete questionnaires for this study. You
indicate your voluntary agreement to participate by signing below. After you sign and
complete the questionnaires, please seal and return the packet in the enclosed envelope to

your head teacher without indicating the name of you or your child.

 

 

 

 

Sincerely,

Nary Shin

Child’s name (Please print) Name of the parent/legal guardian
Signature of the parent/legal guardian Date

154

(For head teachers)

Ecological factors as predictors of visual media uses and its effects on
imaginative playfulness in preschool children

Letter and Consent Form

Dear Head Teachers:

I am a doctoral student at Michigan State University in the Department of Family and
Child Ecology. I am conducting a study about the relationship between visual media use
and the play of young children. The purpose of the study is to examine the associations
among children’s practices of visual media use, environmental factors that can affect the
practices, and the quality of their imaginative play. This research is being conducted

under the supervision of Dr. Alice Whiren.

Your participation will involve distributing a survey packet to every family with three,
four, or five-year olds, collecting returned packets to your director, and completing
questionnaires for each pair of participating children. The questionnaire is regarding
children’s playful behaviors. An estimate of time to complete a questionnaire is no more
than 15 minutes. Because you are expected to complete one copy per participating child,
your time that will take for participation will vary depending on the number of
participating children in you class. While you may experience inconvenience for

participation in this study, there is no personal risk involved with this research.

Your participation is voluntary and you have the right to refiise to participate in this
study. Further, you may discontinue your participation at any time. Also, your non-

participation will not in any way be used to penalize you.

Because the child’s name whom you rate will be written on the questionnaire, your
responses may be identified. Your privacy will be protected to the maximum extent
allowable by law and your responses will be kept confidential. To protect against a
breach of confidentiality, all forms I collect will be stored in a secure place that is

accessible only to project staff, and no names will be used in any report of the research

155

findings. After your questionnaires are coded, they will be destroyed. Your responses

will not be discussed with parents and other school staff.

While specific benefits for participating in this research cannot be guaranteed, the results
will be valuable to those who are interested in the relation between children’s visual

media use and their play.

If you have any question about this study, please feel free to contact me by phone: (517)
272-2980, e-mail: shinnary@msu.edu, or regular mail: 3015 Staten Ave #5, Lansing, MI
48910, or Dr. Alice Whiren, by phone: (517) 432-0893, e—mail: whiren@msu.edu, or
regular mail: 3 Human Ecology, East Lansing, MI 48824. If you have any questions or
concerns regarding your right as a research participant, or are dissatisfied at any time
with any aspect of this study, you may contact - anonymously, if you wish - Dr. Ashir
Kumar, Chair of the University Committee on Research Involving Human Subjects
(UCRIHS) by phone: (517) 355-2180, fax: (517) 432-4503, e-mail: uchris@msu.edu, or
regular mail: 202 Olds Hall, East Lansing, MI 48824.

I sincerely appreciate your participation in this project. In appreciation of your
participation in this research, upon receipt of your completed questionnaires, you will
receive a $5 gift certificate of the Teachers ’ Store in Lansing for your evaluating of each
pair of participating children. You will not receive the gift certificate if you do not
complete questionnaires for this study. By signing below, you indicate that you have

read the information in this form and that you voluntarily agree to participate.

Sincerely,

Nary Shin

 

 

 

Name (Please print) Signature Date

I would be interested in receiving a summary of the results of this study. Yes
No

156

(For directors)

Ecological factors as predictors of visual media uses and its effects on
imaginative playfulness in preschool children

Letter and Consent Form

Dear Directors:

I am a doctoral student at Michigan State University in the Department of Family and
Child Ecology. I am conducting a study about the relationship between visual media use
and the play of young children. The purpose of the study is to examine the associations
among children’s practices of visual media use, environmental factors that can affect the
practices, and the quality of their imaginative play. This research is being conducted

under the supervision of Dr. Alice Whiren.

Data will be collected from primary caregivers and head teachers of three, four or five
olds who will complete questionnaires. Your participation will involve allowing me to
recruit head teachers voluntarily participating in this study and collecting survey packets
completed by the participating families and head teachers. Participating families will
receive two children’s books and head teachers will receive a $5 gift certificate of the
Teachers ’ Store in Lansing for your evaluating of each pair of participating children,
upon receipt of their completed questionnaires. While you may experience
inconvenience for participation in this study, there is no personal risk involved with this

research.

Your participation is voluntary and you have the right to refuse to participate in this
study. Further, you may discontinue your participation at any time. Also, your non-

participation will not in any way be used to penalize you.
Because the name of participating children will be written on questionnaires rated by

their primary caregiver and head teacher, their responses may be identified. The privacy

of participating families and head teachers will be protected to the maximum extent

157

allowable by law and their responses will be kept confidential. To protect against a
breach of confidentiality, all forms I collect will be stored in a secure place that is
accessible only to project staff, and no names will be used in any report of the research

findings. After the questionnaires are coded, they will be destroyed.

While specific benefits for participating in this research cannot be guaranteed, the results
will be valuable to those who are interested in the relation between children’s visual

media use and their play.

If you have any question about this study, please feel free to contact me by phone: (517)
272-2980, e-mail: shinnary@msu.edu, or regular mail: 3015 Staten Ave #5, Lansing, MI
48910, or Dr. Alice Whiren, by phone: (517) 432-0893, e-mail: whiren@msu.edu, or
regular mail: 3 Human Ecology, East Lansing, MI 48824. If you have any questions or
concerns regarding your right as a research participant, or are dissatisfied at any time
with any aspect of this study, you may contact - anonymously, if you wish - Dr. Ashir
Kumar, Chair of the University Committee on Research Involving Human Subjects
(UCRIHS) by phone: (517) 355-2180, fax: (517) 432-4503, e-mail: uchris@msu.edu, or
regular mail: 202 Olds Hall, East Lansing, MI 48824.

By signing below, you indicate that you have read the information in this form and that

you voluntarily agree to participate.

Sincerely,

Nary Shin

 

 

 

Name (Please print) Signature Date

I would be interested in receiving a summary of the results of this study. Yes

No

158

APPENDIX C Coversheet of Questionnaire Packet for Parents

Ecological factors as predictors of visual media uses and its effects on
imaginative playfulness in preschool children

Thank you for your participation.

Please seal your completed questionnaire in the enclosed envelope
and return it to your head teacher.

In appreciation of your participation, please select two books.
The list below includes Caldecott honored books.
They are the most distinguished American children’s books
awarded by the American library Association.

The books will be delivered to your family
after your questionnaires are received

 

BOOK LIST

Click, clack, moo cows that type by Doreen Cronin

 

In the small small pond by Denise Fleming

 

Inch by inch by Leo Lionni

 

Madeline by Ludwig Bemehnans

 

Make way for ducklings by Robert McCloskey

 

No David by David Shannon

 

Oflicer Buckle and Gloria by Peggy Rathmann

 

Strega Nona by Tomie De Paola

 

Swimmy by Leo Lionni

 

The paperboy by Dav Pilkey

 

159

APPENDIX D Family Background Questionnaire

Ecological factors as predictors of visual media uses and its effects on
imaginative playfulness in preschool children
Questionnaire 1
Please fill in the blanks or check the appropriate responses.

Name of the participating child

 

 

 

 

Age of the child years months
Sex of the child Male Female
Position of the child
First born Second born Third born
Fourth born F ifih born Sixth born or greater
The first language of the child
Does the child attend more than one early childhood program? Yes No
If yes, please name the programs 1)
2)
3)

 

Number of adults living in the household
Number of children living in the household
Number of total family members living in the household

Your age years
Your spouse’s age (if applicable) years
Sex of person completing the questionnaire Male Female

Relationship to the child

 

Your current marital status

 

 

 

 

Married Divorced
Widowed Separated
Never married Other

Your ethnic origin
White Afi‘ican American Hispanic
Asian Others

Your spouse’s ethnic origin (if applicable)
White African American Hispanic
Asian Others

Your child’s ethnic origin
White African American Hispanic
Asian Others

 

160

Your education

Some middle school Middle school degree
Some high school High school degree
Some college/no degree Associate degree (2 years)
4-year college degree Master’s degree
Doctoral degree
Total years of your education (beginning with the lowest elementary grade with the
exception of kindergarten) years

Your spouse’s education (if applicable)

Some middle school Middle school degree

Some high school High school degree

Some college/no degree Associate degree (2 years)

4-year college degree Master’s degree

Doctoral degree
Total years of your spouse’s education (beginning with the lowest elementary grade with
the exception of kindergarten) years

Your current occupation (please specify. ex: Homemaker, Owner of factory that employs
about 100 people; public school teacher for second grade children; official of an
international labor union; farm owner and operator)

 

 

Are you a full-time student? Yes No
If yes, please specify your former occupation and current part-time job (if
applicable)

 

 

If no, please check the category of your occupation.
Unemployed Homemaker
Workers requiring no training or skill Skilled worker
Owner of small business, clerical, sales, technician
Administrative personnel, small independent business, minor professional
Business managers, proprietors of medium sized business
Executive and proprietors of large concern, major professionals

Your annual income dollars

Your spouse’s current occupation (if applicable)

 

 

Is your spouse a full-time student? Yes No
If yes, please specify your spouse’s former occupation and current part-time job
(if applicable)

 

 

161

 

If no, please check the category of your spouse’s occupation.
Unemployed Homemaker
Workers requiring no training or skill Skilled worker
Owner of small business, clerical, sales, technician
Administrative personnel, small independent business, minor professional
Business managers, proprietors of medium sized business
Executive and proprietors of large concern, major professionals

Your spouses’ annual income dollars

Gross annual income of your family dollars

Number of television set(s) in your home
Please list the location of each TV set in your home in the first column below.
Location Hours watched
TV Set 1
TV Set 2
TV Set 3
TV Set 4
Now in the second column, please indicate how many total hours it was watched
yesterday by any family members.

 

 

 

 

 

 

 

 

Does the participating child have a television set in his/her room? Yes No
Do you have cable? Yes No
Please specify the type of cable Basic cable

Premium cable
Cable plus movie channe1(s)
Number of VCR(s) in your home
Please list the location of each VCR in your home in the first column below.
Location Hours watched
VCR 1
VCR 2
VCR 3
VCR 4
Now in the second column, please indicate how many total hours it was used yesterday
by any family members.
Does the participating child have a VCR in his/her room? Yes No
Number of DVD player(s) in your home
Please list the location of each DVD PLAYER in your home in the first column below.
Location Hours watched

 

 

 

 

 

 

 

 

DVD player 1
DVD player 2
DVD player 3
DVD player 4
Now in the second column, please indicate how many total hours it was used yesterday
by any family members.
Does the participating child have a DVD player in his/her room? Yes No

 

 

 

 

 

 

 

 

162

 

APPENDIX E Visual Media Use Questionnaire

Questionnaire 2

Please reflect back on the LAST WEEK and base the answers to these
questions on the amount and type your participating child viewed television
shows, videos, or DVDs during that week.

Please list up to ten television shows that your child watches regularly, including
children’s shows and shows for general audiences, and report how many times your child
watched those shows in the last week. Please report the full name of the shows as clearly
as you can.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Program Frequency
Program Frequency
Program Frequency
Pro gram Frequency
Program Frequency
Pro gram Frequency
Program Frequency
Program Frequency
Program Frequency
Program Frequency

 

 

Please recall videos or DVDs that your family members owned, borrowed, or rented.
Please list up to ten videos and DVDs that your child watched in the last week, including
children’s programs and programs for general audiences, and report how many times
your child watched them. Please report the full name of the programs as clearly as you
can. (e. g. Veggie Tale-Jonah, Spiderrnan-the movie)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Name Frequency
Name Frequency
Name Frequency
Name Frequency
Name Frequency
Name Frequency
Name Frequency
Name Frequency
Name Frequency
Name Frequency

 

 

163

 

APPENDIX F Parental Mediation Questionnaire

Questionnaire 3

Please read each statement and circle the number next to the statement that
BEST describes your participating child and your family.

 

How often do you and/or your spouse Never Rarely Sometimes Often Always

 

1. try to help the child understand what 1 2 3 4 5
s/he sees on television, video, or
DVD?

2. point out why some things actors do 1 2 3 4 5
on television, video, or DVD are
good?

3. point out why some things actors do 1 2 3 4 5
on television, video, or DVD are bad?

4. explain the motives of the characters 1 2 3 4 5
on television, video, or DVD?

5. explain what something on television, 1 2 3 4 5
video, or DVD really means?

6. say to your child to turn off television, 1 2 3 4 5
video, or DVD when s/he is watching
an unsuitable program?

7. set specific viewing hours for your 1 2 3 4 5
child?

8. forbid your child to watch certain 1 2 3 4 5
programs?

9. restrict the amount of child viewing? 1 2 3 4 5

10. specify in advance the television 1 2 3 4 5
programs, videos, or DVDs that may
be watched?

 

Note 1. Original measure sub-dimensions: 1-5 = lnstnrctive mediation, 6-10 =
Restrictive mediation.

Note 2. Modified measure sub-dimensions: 1-5 = Instructive mediation items, 7-
10 = Restrictive mediation items.

164

 

APPENDIX G Imaginative Playfulness Scale

Ecological factors as predictors of visual media uses and its effects on
imaginative playfulness in preschool children

This questionnaire is regarding children’s playful behaviors during their
IMAGINATIVE PLAY. Imaginative play refers PRETEND PLAY that children
make believe that they are someone or something else, that they are in a
different place and time, and/or that playthings represent something else.
Please read each statement and circle the number next to the statement that
best describes the characteristics of imaginative play INDOORS and
OUTDOORS.

 

 

 

Name of child:
Doesn’t Describes Describes Describes Describes
describe the the child the child the child the child
child at all a little somewhat a lot exactly

1. The child’s movements are 1 2 3 4 5
generally well-coordinated and
playful during pretend play.

2. The child is physically active during 1 2 3 4 5
pretend play.

3. The child prefers to be active rather I 2 3 4 5
than quiet during pretend play.

4. The child nms (skips, hops, jumps) 1 2 3 4 5
a lot during pretend play.

5. The child responds easily to others’ 1 2 3 4 5
approach during pretend play.

6. The child initiates pretend play with 1 2 3 4 5
others.

7. The child plays cooperatively with 1 2 3 4 5
other children during pretend play.

8. The child is willing to share 1 2 3 4 5
playthings during pretend play.

9. The child assumes a leadership role 1 2 3 4 5
when playing with others during
pretend play.

10. The child invents his/her own 1 2 3 4 5
games during pretend play.

 

165

 

 

Doesn’t Describes Describes Describes Describes
describe the the child the child the child the child

 

child at all a little somewhat a lot exactly

11. The child uses unconventional 1 2 3 4 5
objects during pretend play.

12. The child assumes various 1 2 3 4 5
characters’ roles during pretend
play.

13. The child stays with one activity 1 2 3 4 5
rather than changing activities
during pretend play.*

14. The child expresses enjoyment 1 2 3 4 5
during pretend play.

15. The child demonstrates exuberance 1 2 3 4 5
during pretend play.

16. The child shows enthusiasm during 1 2 3 4 5
pretend play.

17. The child is restrained in expressing 1 2 3 4 5
emotion during pretend play.*

18. The child sings and talks during 1 2 3 4 5
pretend play.

19. The child enjoys joking with other 1 2 3 4 5
children during pretend play.

20. The child gently teases others 1 2 3 4 5
during pretend play.

21. The child tells funny stories during 1 2 3 4 5
pretend play.

22. The child laughs at humorous 1 2 3 4 5
stories during pretend play.

23. The child likes to “clown around” 1 2 3 4 5
during pretend play.

24. The child is playful during general 1 2 3 4 5
play as well as pretend play. **

 

* indicates inverted coding.

** not included in the playfulness score.

Note 1. Original measure sub-dimensions: 1-4 = Physical Spontaneity, 5-9 =
Social Spontaneity, 10—13 = Cognitive Spontaneity, 14-18 = Manifest Joy, 19-23
= Sense of Humor.

Note 2. Modified measure sub-dimensions: 2-4 = Physical Spontaneity, 6-8 =
Social Spontaneity, 9-12 = Cognitive Spontaneity, 14-17 = Manifest Joy, 19-23 =
Sense of Humor.

166

BIBLIOGRAPHY

Achenbach, T. M., Dumenci, L., & Rescorla, L. A. (2002). Ten-year comparisons
of problems and competencies for national samples of youth: Self, parent,
and teacher reports. Journal of Emotional and Behavioral Disorders, 10
(4), 194-203.

Aguilar, T. E. (1985). Social and environmental barriers to playfulness. In J. L.
Frost, & S. Sunderlin (Eds.), When children play (pp. 73-76). Wheaton,
MD: Association for Childhood Education lntemational.

Alvarez, M., Huston, A. C., Wright, J. C., & Kerkman, D. (1988). Gender
differences in visual attention to television form and content. Journal of
Applied Developmental Psychology. 9, 459-475.

Anderson, D. R., & Bryant, J. (1983). Research on children’s visual media uses:
The state of the art. In J. Bryant & D. R. Anderson (Eds.), Children’s
understanding of TV: Research on attention and comprehension (pp. 331 -
353). New York: Academic Press.

Anderson, D. R., & Field, D. E. (1983). Children’s attention to television:
Implications for production. In M. Meyer (Ed.), Children and the formal
features of television (pp. 56-96). New York: Saur.

Anderson, D. R., Huston, A. C., Schmitt, K. L., Linebarger, D. L., & Wright, J. C.
(2001). Early childhood visual media uses and adolescent behavior.
Monographs of the Society for Research in Child Development, 66 (No. 1,
Serial No. 264). Boston, MA: Blackwell.

Arabasz, J. (2000). The effects of visual media uses on the playfulness of
children. Unpublished master’s thesis, Colorado State University, Fort
Collins.

Arbuckle, J. L., & Wothke, W. (1999). Amos 4.0 user’s guide. Chicago, IL: Small
Waters.

Atkin, D. J., Greenberg, B. S., & Baldwin, T. F. (1991). The home ecology of
children's visual media uses: Parental mediation and the new video
environment. Journal of Communication, 41 (3), 40-52.

Austin, E. W., Roberts, D. F., & Nass, C. I. (1990). Influences of family

communication on children’s television-interpretation processes.
Communication Research, 17, 545-564.

167

Bagozzi, R. P., 8 Yi, Y. (1988). On the evaluation of structural models. Journal of
Academy of Marketing Science, 16, 74-94.

Ball, S. 8 Bogatz, G. A. (1970). The first year of "Sesame Street”: An evaluation.
Princeton, NJ: Educational Testing Service.

Barnett, L. A. (1990). Playfulness: Definition, design, and measurement. Play 8
Culture, 3, 319-336.

Barnett, L. A. (1991a). The playful child: Measurement of a disposition to play.
Play 8 Culture, 4, 51-74.

Barnett, L. A. (1991b). Characterizing playfulness: Correlates with individual
attributes and personality traits. Play 8 Culture, 4, 371-393.

Barnett, L. A., 8 Kleiber, D. A. (1982). Concomitants of playfulness in early
childhood: Cognitive abilities and gender. Journal of Genetic Psychology,
141,115-127.

Barnett, L. A., 8 Kleiber, D. A. (1984). Playfulness and the early play
environment. Joumal of Genetic Psychology, 144, 153-164.

Benishek, L. A. (1992). Hardiness, life stress, and neuroticism: A structural
equation model of self-reported illness. Unpublished Doctoral Dissertation,
Michigan State University, East Lansing.

Bentler, P. M. (1989). E08: A structural equations program manual. Los
Angeles: BMDP Statistical Software Inc.

Bianchi, S. M., 8 Robinson, J. (1997). What did you do today? Children’s uses of
time, family composition, and acquisition of social capital. Journal of
Marriage and the Family, 59, 332-344.

Bower, R. (1973). The people look at television. New York: John Wiley.

Boyd, B. J. (1997). Teacher response to superhero play: To ban or not to ban?
Childhood Education, 74 (1 ), 23-28.

Boyer, W.A.R. (1997/98). Playfulness enhancement through classroom
intervention for the 21st century. Childhood Education, Winter, 90-96.

Bredekamp, S., 8 Copple, C. (Eds). (1997). Developmentally appropriate

practice in early childhood programs. Washington, DC: National
Association for the Education of Young Children.

168

Bronfenbrenner, U. (1979). The ecology of human development. Cambridge, MA:
Harvard University Press.

Bronfenbrenner, U. (1986). Ecology of the family as a context for human
development: Research perspectives. Developmental Psychology, 22,
723-742.

Bronfenbrenner, U. (1989). Ecological system theory. In R. Vasta (Ed.), Annals of
child development (Vol. 6, pp. 187-249). Greenwich, CT: JAI Press.

Brown, J. A. (2001). Media literacy and critical television viewing in education. In
D. G. Singer, 8 J. L., Singer (Eds), Handbook of children and the media
(pp. 681-697). Thousand Oaks, CA: Sage Publications, Inc.

Brown, J. D. (1997). Skewness and Kurtosis. Testing 8 Evaluation SIG
Newsletter, 1 (1), 16-18.

Bryne, B. M. (1994). Structural equation modeling with E08 and EQS/Windows.
Thousand Oaks, CA: Sage.

Bundy, A. C. (1992). Play: The most important occupation of children. Sensory
Integration Special Interest Section Newsletter, 15(2), 1.

Bundy, A. C. (1993). Assessment of play and leisure: Delineation of the problem.
American Journal of Occupational Therapy, 47, 217-222.

Bundy, A. C. (1998). Test of Playfulness (ToP) manual, Version 3.5. Ft. Vollins,
CO: Colorado State University.

Cantor, J. (1994). Fright reactions to mass media. In J. Bryant 8 D. Zillmann
(Eds.), Media effects: Advances in theory and research (pp. 213-246).
Hillsdale, NJ: Lawrence Erlbaum.

Carlsson-Paige, N., 8 Levin, D. E. (1987). The war play dilemma: Balancing
needs and values in the early in the earty childhood classroom. New York:
Teachers College Press.

Carlsson-Paige, N., 8 Levin, D. E. (1990). Who’s calling the shots?: How to
respond effectively to children’s fascination with war play and war toys.
Philadelphia, PA: New Society Publishers.

Carlsson-Paige, N., 8 Levin, D. E. (1995). Can teachers resolve the war-play
dilemma? Young Children, 50(5), 62-63.

169

Carpenter, C. J., Huston, A. C., 8 Spare, L. (1989). Children’s use of time in their
everyday activities during middle childhood. In M. N. Bloch 8 A. D.
Pellegrini (Eds.), The ecological context of children ’3 play (pp. 165-190).
Norwood, NJ: Ablex.

Case-Smith, J. (Ed.). (1996). Occupational therapy for children. St. Louis, MO:
Mosby.

Center for Research on Influences of Television on Children. (1983). CRITC
program categorization system coding manual. Lawrence, KS: University
of Kansas.

Chaffee, S. H. 8 Tims, A. R. (1976). Interpersonal factors in adolescent television
use. Journal of Social Issues, 32, 98-115.

Christopher, F. S., Fabe, R. A., 8 Wilson, P. M. (1989). Family visual media
uses: Implications for family life education. Family Relations, 38, 210-214.

Comstock, G (1991). Television and the American child. San Diego: Academic
Press.

Comstock, G., Chaffee, S., Katzman, N., McCombs, M., 8 Roberts, D. (1978).
Television and human behavior. New York: Columbia University Press.

Condry, J. (1989). The psychology of television. Hillsdale, NJ: Lawrence
Erlbaum.

Connor, K. (1989). Aggression: Is it in the eye of the beholder? Play 8 Culture, 2,
213-217.

Cook, T. D., Appleton, H., Conner, R., Shaffer, A. Tamkin, G., 8 Weber, S. J.
(1975). “Sesame Street” revisited. New York: Ressell Sage.

Creasey, G. L., Jarvis, P. A., 8 Berk, L. E. (1998). Play and social competence.
In O. N. Saracho 8 B. Spodek (Eds.), Multiple perspectives on play in
early childhood education. SUNY series, eariy childhood education:
Inquiries and insights (pp. 116—143). Albany, NY: State University of New
York Press.

Crosser, S. (1995). Mighty Morphine Power Ranger play: Research and reality.
Eariy Childhood News, 7(3), 25-27.

Davie, C. E., Hutt, S. J., Vincent, E., 8 Mason, M. (1984). The young child at
home. London: NFFR-Nelson.

170

Dennison, B. A., Erb, T., 8 Jenkins, P. L. (2002). Television viewing and
television in bedroom associated with ovenrveight risk among low-income
preschool children. Pediatrics, 109(6), 1028-1035.

Desmond, R. J., Hirsch, 8., Singer, J. L., 8 Singer, D. G. (1987). Gender
differences, mediations, and disciplinary styles in children’s responses to
television. Sex Roles, 16, 75-389.

Desmond, R. J., Singer, J. L., Singer, D. G., Calam, R., 8 Colimore, K. (1985).
Family mediation patterns and visual media uses: Young children’s use

and grasp of the medium. Human Communication Research, 11 (4), 461-
480.

Desmond, R. J., Singer, J. L., 8 Singer, D. G. (1990). Family mediation: Parental
communication patterns and the influences of television on children. In J.
Bryant (Ed.), Television and the American family (pp. 293-310). Hillsdale;
NJ: Lawrence Erlbaum.

Dorr, A., Kovaric, P., 8 Doubleday, C. (1989). Parent-child coviewing of television.
Journal of Broadcasting 8 Electronic media, 33 (1 ), 35-51.

Dorr, A., 8 Rabin, B. E. (1995). Parents, children, and television. In M. Bomstein
(Ed.), Handbook of parenting (Vol. 4, pp. 323-351 ). Mahwah, NJ: Erlbaum.

Duncan, 0. D. (1961). A socioeconomic index for all occupations. In Reiss, A. J.
(Ed.), Occupations and social status (pp. 109-138). New York: The Free
Press of Glencoe.

Dunn, J. (1983). Sibling relationship in early childhood. Child Development, 54,
787-81 1.

Eron, L. D. (1963). Relationship of television viewing habits and aggressive
behavior in children. Joumal of Abnormal and Social Psychology, 67, 193-
196.

Eron, L. D. (1993). No doubt about it, media violence affects behavior. Media and
Values, 64, 14.

Fein, G. (1975). A transformational analysis of pretending. Developmental
Psychology, 1 1, 291 -296.

Fein, G. (1981). Pretend play in childhood: An integrative review. Child
Development, 52, 1095-1 1 18.

Fisher, E. P. (1992). The impact of play on development: A meta-analysis. Play 8
Culture, 5, 159—1 81 .

171

 

French, J., 8 Pena, S. (1991). Children’s hero play of the 20th century: Changes
resulting from television’s influence. Child Study Journal, 21 (2), 79-94.

Friedrich-Cofer, L., K., Huston-Stein, A., McBride K. D., Susman, E. J., 8 Clewett,
A. S. (1979). Environmental enhancement of prosocial television content:
Effects on interpersonal behavior, imaginative play, and self-regulation in
a natural setting. Developmental Psychology, 15, 637-646.

Fromberg, D. R. (2002). Play and meaning in early childhood education. Boston,
MA: Allyn 8 Bacon.

Funk, J. (2001). The effects of video games on Youth. Paper presented at the
Annual Meeting of the Society for Research in Child Development,
Minneapolis, MN.

Gadberry, S. (1980). Effects of restricting first granters’ TV-viewing on leisure
time use, IQ change, and cognitive style. Journal of Applied
Developmental Psychology, 1, 45-57.

Garson, G. D. (n.d.). Factor analysis. Retrieved September 20, 2003, from
http://www2.chass.ncsu.edu/garson/pa765/factor.htm.

Garvey, C. (1977). Play. Cambridge, MA: Harvard University Press.

Glascock, J. (2003). Gender, race, and aggression in newer TV networks’
primetime programming. Communication Quarteriy, 51 (1 ), 90.

Greenberg, B. 8., Edison, N., Femabdez-Colladp, C., 8 Atkin, C. K. (1980)
Antisocial and prosocial behaviors on television. In B. S. Greenberg (Ed),
Life on Television (pp. 99-128). Norwood, NJ: Ablex.

Greenfield, P., 8 Beagles-Roos, J. (1988). Radio vs. television: Their cognitive
impact on children of different socioeconomic and ethnic groups. Joumal
of Communication, 38 (2), 71-92.

Greenfield, P. M., Yut, E., Chung, M., Land, 0., Kreider, H., Pantoja, M., 8
Horsley, K. (1990). The program length commercial: A study of the effects

of television toy tie-ins on imaginative play. Psychology 8 Marketing, 7(4),
237-255.

Greer, D., Potts, R., Wright, J. C., 8 Huston, A. C. (1982). The effect of television
commercial form and commercial placement on children’s social behavior
and attention. Child Development, 53, 611-619.

Griffore, R. J., 8 Phenice, L. P. (2001). The language of human ecology: A
general systems perspective. Dubuque, Iowa: Kendall/Hunt.

172

Gross, L. S. 8 Walsh, R. P. (1980). Factors affecting parental control over
children’s visual media uses: A pilot study. Joumal of Broadcasting, 24,
411-419.

Guddemi, M. R. (1985). The effects of television on children’s play themes.
Unpublished Doctoral Dissertation, The University of Texas, Austin.

Gunter, B. (1981). Measuring television violence: A review and suggestions for a
new analytical perspective. Current Psychological Review, 1, 91-112.

Halperin, J. M., McKay, K. E., Grayson, R. H., 8 Newcom, J. H. (2003).
Reliability, validity, and preliminary normative data for the children’s
aggression scale-teacher version. Journal of the American Academy of
Child and Adolescent Psychiatry, 42 (8), 965.

Harris, R. J. (1985). A primer of multivariate statistics (2nd ed.). Orlando, FL:
Academic Press.

Hofferth, S. L., Davis-Kean, R, Davis, J., 8 Finkelstein, J. (1998). Child
development supplement to the Panel Study of Income Dynamics, 199 7:
Users guide. Retrieved September 17, 2000, from
http://www.isr.umich.edu/src/child-development/home.htlm

Hogue, L. 8 Johnson, C. (1999). Does watching television affect playfulness in
children? Journal of Occupational Therapy Students. Retrieved December
12, 2001, from http://www.aota.org/nonmembers/area7/Iinks/LlNK04D.asp

Honig, A. S., Fitzgerald, H. E., 8 Brophy-Herb, H. E (eds.). (2001). Infancy in
America: Encyclopedia. Santa Barbara, CA: ABC-CLIO, Inc.

Huesmann, L. R., 8 Eron, L. D. (Eds.). (1986). Television and the aggressive
child: A cross-national comparison. Hillsdale, NJ: Erlbaum.

Huesmann, L. R., Eron, L. D., Lefkowitz, M., 8 Walder, L. (1986). Stability of

aggression over time and generation. Developmental Psychology, 20(6),
1 120-1 134.

Huesmann, L. R., Moise-Titys, J., Podolski, C, 8 Eron, L. D. (2003). Longitudinal
relations between children’s exposure to TV violence and their aggressive
and violent behavior in young adulthood: 1977-1992. Developmental
Psychology, 39 (2), 201-221.

Huston, A. C., Donnerstein, E., Fairchild, H., Feshbach, N., Katz, P., Murray, J.,
Rubinstein, E., Wilcox, B., 8 Zuckerman, D. (1992). Big world, small
screen: The role of television in American society. Lincoln, NE: University
of Nebraska Press.

173

Huston, A. C., 8 Wright, J. C. (1996). Television and socialization of young
children. In MacBeth, T. M. (Ed.), Tuning in to young viewers: Social
science perspectives on television (pp. 37-60). Thousand Oaks, CA: Sage.

Huston, A. C., 8 Wright, J. C. (19983). Television and informational and
educational needs of children. The Annals of the American Academy, 557,
9-21.

Huston, A. C., 8 Wright, J. C. (1998b). Mass media and children’s development.
In I. Sigel 8 K. A. Renninger (Eds.), Handbook of child psychology: Child
psychology in practices (Vol. 4, pp. 999-1058). New York: Wiley.

Huston, A. C., Wright, J. C., Marquis, J. M., 8 Green, S. B. (1999). How young
children spend their time: Television and other activities. Developmental
Psychology, 35(4), 912-923.

Huston, A. C., Wright, J. 0., Rice, M. L., Kerrnan, D., 8 St. Peters, M. (1990). The
developmental of visual media uses pattern in early childhood: A
longitudinal investigation. Developmental Psychology, 26, 409-420.

Huston-Stein, A., Fox, 8., Greer, D., Watkins, B. A., 8 Whitaker, J. (1981). The
effects of TV action and violence on children’s social behavior. The
Journal of Genetic Psychology, 138, 183-191.

Information Technology Services. (2002). Handling non-normal data in structural
equation modeling. Retrieved October 18, 2003, from
http://www.utexas.edu/cc/faqs/stat/general/gen33.html.

Iverson, K. (1982). Play, creativity, and schools today. Phi-DeIta-Kappa, 63 (10),
693-694.

James, N. C., 8 McCain, T. A. (1982). Television games preschool children play:
Patterns, themes, and uses. Journal of Broadcasting, 26(4), 783-800.

Johnson, J. E., Christie, J. F., 8 Yawkey, T. D. (1999). Play and early childhood
development (2nd ed.). New York: Longman.

Jordan, A. B. (2001). Public policy and private practice: Government regulations
and parental control of children’s television use in the home. In D. G.
Singer 8 J. L. Singer (Eds.), Handbook of children and the media (pp.
651-662). Thousand Oaks, CA: Sage Publications, Inc.

Kieff, J. E., 8 Casbergue, R. M. (2000). Playful Ieaming and teaching: Integrating
play into preschool and primary programs. Needham Heights, MA: Allyn 8
Bacon.

174

King, N. R. (1982). Children’s play as a form of resistance in the classroom.
Journal of Education, 164 (3), 320-329.

Kline, R. B. (1998). Principle and practices of structural equation modeling. New
York: The Guilford Press.

Kogan, N. (1983). Stylistic variation in childhood and adolescence: Creativity,
metaphor and cognitive styles. In J. H. Flavell 8 E. Markman (Eds),
Handbook of child psychology: Cognitive development (Vol. 3, pp. 630-
706). New York: Wiley.

Kostelnik, M. J., Whiren, A. P., 8 Stein, L. C. (1986). Living with He-Man:
Managing superhero fantasy play. Young Children, 41 (4), 3-9.

Kostelnik, M. J., Whiren, A. P., Soderman, A. K., Gregory, K., 8 Stein, L. C.
(2002). Guiding children’s social development: Theory to practice (4“ ed.).
Albany, NY: Delmar.

Knox, S. H. (1996). Play and playfulness in preschool children. In R. Zemke 8 F.
Clarke (Eds.), Occupational science: The evolving discipline (pp. 81-88).
Philadelphia, PA: F.A. Davis.

Krasnor, L. R., 8 Pepler, D. J. (1980). The study of children’s play: Some
suggested future directions. In K. H. Rubin (Ed.), New directions in child
development: Children ’3 play (pp. 85-95). San Francisco, CA: Jossey-
Bass.

Lane, S. J., 8 Mistrett, S. G. (2000). Play of children with disabilities: The impact
of intervention on playfulness. Paper presented at the annual meeting of
the Association for the Study of Play, Baltimore, MD.

Lemelle, A. (2002). The effects of the intersection of race, gender and
educational class on occupational prestige. Western Joumal of Black
Studies, 26(2), 89-98.

Levin, D. E. (1998). Play with violence: Understanding and responding
effectively. In D. P. Fromberg 8 D. Bergen (Eds.), Play with birth to twelve
and beyond: Context, perspectives, and meanings (pp.348-356). New
York, NY: Garland Publishing.

Levin, D. E., 8 Carlsson-Paige, N. (1989). Piaget, war play and peace. The
Genetic Epistemologist, 17(3), 11-15.

Lickel, L., Schmader, T., 8 Hamilton, D. L. (2003). A case of collective

responsibility: Who else was to blame for the Columbine High School
shootings? Personality and Social Psychology Bulletin, 29(2), 194-204.

175

Lieberman, J. N. (1965). Playfulness and divergent thinking: An investigation of
their relationship at the kindergarten level. Joumal of Genetic Psychology,
107, 219-224.

Lieberman, J. N. (1966). Playfulness: An attempt to conceptualize a quality of
play and of the player. Psychological Report, 19, 1278.

Lieberman, J. N. (1967). A developmental analysis of playfulness as a clue to
cognitive style. Journal of Creative Behavior, 1, 391-397.

Lieberman, J. N. (1977). Playfulness: Its relationship to imagination and
creativity. New York: Academic Press.

Liebert, R. M., Sprafkin, J. N., 8 Davidson, E. S. (1982). The eariy window. New
York: Pergamon.

Lin, C. A., 8 Atkin, D. J. (1989). Parental mediation and rulemaking for
adolescent use of television and VCRs. Journal of Broadcasting 8
Electronic Media, 32 (4), 403-413.

Line 8., 8 Pentecost, D. (1990). The characterization of play: Marketing
children’s toy. Play 8 Culture, 3, 235-255.

Lull, J. (1978). How failures select television programs: A mass-observational
study. Journal of Broadcasting, 26, 801 -81 1.

Lull, J. (1980). The social uses of television. Human Communication Research,
6, 197-209.

Luster, T., 8 Oh, S. M. (2001). Correlates of male adolescents carrying handguns
among their peers. Journal of Marriage and Family, 63, 714-726.

Matthews, R. J., 8 Matthews, W. S. (1982). A paradigm case approach to the
study of fantasy. In D. J. Pepler 8 K. H. Rubin (Eds), The play of children:
Current theory and research (pp. 21 -29). New York: Karger.

McDonald, D. G. (1986). Generational aspect of television coviewing. Journal of
Broadcasting 8 Electronic Media, 30, 75-85.

McKenzie, T. L., Sallis, J. F., Nader, P. R., 8 Broyles, S. L. (1992). Anglo- and
Mexican-American preschoolers at home and at recess: Activity patterns

and environmental influences. Journal of Developmental and Behavioral
Pediatrics, 13, 173-180.

Meidzian, M. (1991). Boys will be boys: Breaking the link between masculinity
and violence. New York: Doubleday.

176

Messaris, P., 8 Sarett, C. (1981). On the consequences of television-related
parent-child interaction. Human Communication Research, 7, 226-244.

Metzger, P. (1993). Validity and reliability of test of playfulness. Unpublished
master’s thesis, University of Illinois, Chicago, IL.

Neuendorf, K. A. (2002). The content analysis guidebook. Thousand Oaks, CA:
Sage.

Nilsson, N. (1995). Children and war toys: Who had the idea to introduce children
to violence? lntemational Play Theory, 3, 20-33.

Norman, G. R., and D. L. Streiner (1994). Biostatistics: The bare essentials. St.
Louis, MO: Mosby.

Okimoto, A. M., Bundy, A., 8 Hanzik, J. (1999). Playfulness in children with or
without disability: Measurement and intervention. The American Journal of
Occupational Therapy, 54(1), 73-82.

Osborne, J. W. (2002). Normalizing data transformations. (ERIC Documentation
Reproduction Service No. ED470204).

Paik, H. (2001). The history of children’s use of electronic media. In J. L. Singer
8 D. G. Singer (Eds), Handbook of children and the media (pp. 7-28).
Thousand Oaks, CA: Sage.

Pinon, M. F., Huston, A. C., 8 Wright, J. C. (1989). Family ecology and child
characteristics that predict young children’s educational visual media uses.
Child Development, 60, 846-856.

Potts, R., Huston, A. C., 8 Wright, J. C. (1986). The effects of television form and
violent content on boys’ attention and social behaviors. Journal of
Experimental Child Psychology, 41, 1-17.

Power, T. G. (2000). Play and exploration in children and animals. Mahwah, NJ:
Lawrence Erlbaum.

Raykov, T, Tomer, A., 8 Nesselroade, J. R. (1991). Reporting Structural Equation
Modeling results in psychology and aging: Some proposed guidelines.
Psychology and Aging, 6 (4), 499—503.

Reiss, A. J. (1961). Occupations and social status. Glencoe. IL: Free Press.

Riffe, D., Lacy, S., 8 Fico, F. G. (1998). Analyzing media messages: Using

quantitative content analysis in research. Mahwah, NJ: Lawrence Erlbaum
Association.

177

 

Rogers, C. S., Impara, J. C., Frary, R. 8., Harris, T, Meeks, A., Semanic-Lauth,
S., 8 Reynolds, M. R. (1998). Measuring playfulness: Development of the
Child Behaviors Inventory of Playfulness. Play 8 Culture Studies, 1, 121-
135.

Rogers, C. S., Meeks, A. M., Impara, J. C., 8 Frary, R. (1987). Measuring
playfulness: Development of the child behaviors inventory of playfulness.

Paper presented at the Southwest Conference on Human Development,
New Orleans, LA.

Rogers, J. C. (1987). Selection of evaluation instruments. In L. King-Thomas 8 B.
J. Hacker (Eds), A therapists guide to pediatric assessment (pp. 3-33).
Boston: Little Brown.

Roopnarine, J. L., 8 Johnson, J. E. (1994). The need to look at play in diverse
cultural settings. In J. L. Roopnarine, J. E. Johnson, 8 F. H. Hooper (Eds),
Children’s play in diverse culture (pp. 1-8). Albony, NY: State University of
New York Press.

Rosengren, K. E. (1994). Media effects and beyond: Culture, socialization, and
lifestyles. London: Routledge.

Rothschild, N., 8 Morgan, M. (1987). Cohesion and control: Adolescents’

relationships with parents as mediators of television. Journal of Early
Adolescence, 7(3), 299-314.

Rubin, K. H. (1977). Play behaviors of young children. Young Children, 32(6),
16-24.

Rubin, K. H. (1982). Nonsocial play in preschoolers: Necessarily evil? Child
Development, 53, 651-657.

Rubin, K. H., Fein, G. G., 8 Vanderberg, B. (1983). Play. In E. M. Hetherington
(Ed.), Handbook of child psychology: Socialization, personality, and social
development (Vol. 4, pp. 693-774). New York: Wiley.

Schwartzman, H. (1978). Transformations. New York: Plenum.

Shin, N. (2001). The effects of television on children ’3 play. Paper presented at
the Association of Study Play. San Diego, CA.

Singer, D. G., 8 Rummo, J. (1973). Ideational creativity and behavioral style in
kindergarten aged children. Developmental Psychology. 8, 154-161.

178

Singer, J. L. (1980). The power and limitations of television: A cognitive-affective
analysis. In P. H. Tannenbaum (Ed.), The entertainment functions of
television (pp. 31-65). New York: Lawrence Erlbaum.

Singer, J. L. (1996). Cognitive and affective implications of imaginative play in
childhood. In M. Lewis (Ed.), Child and adolescent psychiatry: A
comprehensive textbook (pp. 202-212). Baltimore, MD: Williams 8 Wilkins.

Singer, J. L. 8 Singer, D. G. (1976). Visual media uses and imaginative play in
preschoolers: A developmental and parent-intervention study. Progress
Report, National Science Foundation Rann Grant No. Apt 76-20772, Yale
University.

Singer, J. L., 8 Singer, D. G. (1981). Television, imagination, and aggression: A
study of preschoolers. Hillsdale, NJ: Lawrence Erlbaum.

Singer, J. L., 8 Singer, D. G. (1983). Psychologists look at television. American
Psychologist, July, 826-834.

Singer, J. L., 8 Singer, D. G. (1986). Family experiences and visual media uses
as predictors of children’s imagination, restless and aggression. Journal of
Social Issues, 42 (3), 107-124.

Singer, J. L., 8 Singer, D. G. (1998). Barney 8 Friends as entertainment and
education: Evaluationg the quality and effectiveness of a television series
for preschool children. In J. K. Asamen 8 G. L. Berry (Eds), Research
paradigms, television, and social behavior (pp. 305-367). Thousand Oaks,
CA: Sage.

Singer, J. L., Singer, D. G., 8 Rapaczynski, W. (1984a). Children’s imagination
as predicted by family patterns and visual media uses: A longitudinal
study. Genetic Psychology Monographs, 110, 43-69.

Singer, J. L., Singer, D. G., 8 Rapaczynski, W. S. (1984b). Family patterns and
visual media uses as predictors of children’s belief and aggression.
Journal of Communication, 34 (2). 73-89.

Singer, J. L., Singer, D. G., 8 Sherrod, L. R. (1980). A factor analytic study of
preschooler’s play behavior. American Psychology Bulletin, 2, 143-155.

Slater, M. D. (2003). Alienation, aggression, and sensation seeking as predictors
of Adolescent use of violent film, computer, and website content. Joumal
of Communication, 53(1), 105-121.

Smilansky, S. (1968). The effects of sociodramatic play on disadvantaged
preschool children. New York: Wiley.

179

Smith, P., 8 Lewis, K. (1985). Rough-and tumble play, fighting, and chasing in
nursery school children. Ethology and Sociobiology, 6, 175-181.

Smith, P., 8 Vollstedt, R. (1985). On defining play: An empirical study of the
relationship between play and various play criteria. Child Development,
56, 1042-1050.

Snow, R. P. (1974). How children interpret television violence in a play context.
Journalism Quarterly, 51, 13-21.

St. Peters, M., Fitch, M., Huston, A. C., Wright, J. C., 8 Eakins, D. (1991).
Television and families: What do young children watch with their parents?
Child Development, 62, 1409-1423.

Sutton-Smith, B. (1988). War toys and childhood aggression. Play and Culture, 1,
57-69.

Tangney, J. P., 8 Feshbach, S. (1988). Children’s television-viewing frequency:
Individual differences and demographic correlates. Personality and Social
Psychology Bulletin, 14 (1), 145-158.

Timmer, S. G., Eccles, J., 8 O’Brien, K (1985). How children use time. In F. T.
Juster 8 F. P. Stafford (Eds), Time, Goals, and well-being (pp. 353-382).
Ann Arbor, MI: Survey of Research Center, Institute for Social Research.

Tower, R. 8., Singer, D. 6., Singer, J. L., 8 Biggs, A. (1979). Differential effects
of television programming on preschoolers’ cognitions, imagination, and
social play. American Journal of Orthopsychiatry, 49, 265-281.

Truglio, R. T., Murphy, K. C., Oppenheimer, S., Huston, A. C., 8 Wright, J. C.
(1996). Predictors of children’s entertainment visual media uses: Why are
they turning in? Journal of Applied Developmental Psycho/091’. 17, 475-
494.

Valkenburg, P. M., Krcmar, M., Peeters, A. L., 8 Marseille, N. M. (1999).
Developing a scale to assess three styles of television mediation:
“Instructive mediation”, “restrictive mediation”, and “social coviewing".
Journal of Broadcasting 8 Electronic Media, 43 (1 ), 52-66.

Valkenburg, P. M., 8 van der Voort, T. H. A. (1994). Influence of TV on
daydreaming and creative imagination: A review of research.
Psychological Bulletin, 1 16 (2), 316-339.

Valkenburg, P. M., Vooijs, M. W., 8 van der Voort, T. H. A (1992). The influence

of television on children’s fantasy styles: A secondary analysis.
Imagination, Cognition 8 Personality, 12(1), 55-67.

180

van der Voort, T. H. A., 8 Valkenburg, P. M. (1994). Television’s impact on
fantasy play: A review of research. Developmental Review, 14, 27-51.

Warren, R. (2003). Parental mediation of preschool children’s television viewing.
Journal of Broadcasting 8 Electronic Media, 47(3), 394-417

Warren, R., 8 Gerke, P., 8 Kelly, M. A. (2002). Is there enough time on the
clock? Parental involvement and mediation of children’s television viewing.
Journal of Broadcasting 8 Electronic Media, 46 (1 ), 87-101.

Weaver, 3., 8 Barbour, N. G. (1992). Mediation of children’s televiewing. Journal
of Contemporary Human Services, 73, 236-242.

Weiner, E. A. 8 Weiner, B. J. (1974). Differentiation of retarded and normal
children through toy-play analysis. Multivariate Behavioral Research, 9,
245-252.

Winn, M. (1985). The plug-in drug: Television, children and the family. New York:
Viking Penguin.

Wright, J. C., Huston, A. C., Murphy, K. C., S. Peters, M., Pinon, M., Scantlin, R.,
8 Kotler, J. (2001). The relations of early television viewing to school

readiness and vocabulary of children from low-income families: The early
window project. Child Development, 72(5), 1347-1366.

Zill, N., Davies, E., 8 Daly, M. (1994). Viewing of Sesame Street by preschool
children in the United States and its relationship to school readiness.
Rockville, MD: Westat.

Zuckerman, 0., Singer, D. G., Singer, J. L. (1980). Visual media uses, children’s
reading and related classroom behavior. Journal of Communication, 30(1),
166-175.

181

 

   

IIIIinIIIIIIjjIIIIIiII