Human Communication Research ISSN 0360-3989

ORIGINAL ARTICLE

The Impact of “Scary” TV and Film

on Children’s Internalizing Emotions:
A Meta-Analysis
Laura J. Pearce & Andy P. Field
School of Psychology, University of Sussex, Brighton BN1 9QH, UK

Despite a general perception that violent or scary television creates anxiety in children, the
research literature is small and disparate. A meta-analysis quantifies the impact of scary
television and film on children’s internalizing emotions (fear, anxiety, sadness, and sleep
problems). Scary television has a relatively small impact on children’s internalizing emo-
tions (r = .18) overall, and this association was not significantly moderated by whether the
televised material was factual (e.g., news) or fictional or contained violence. Children under
10 were more susceptible to scary TV. The overall result contrasts with the dramatic effects
found for individual children within studies, suggesting that research is needed to unpick
the factors that moderate the effect that scary television has on children.

Keywords: Television, Children, Anxiety.

doi:10.1111/hcre.12069

Children spend a large proportion of their waking hours watching television: in the
Unites Kingdom, 4–9 year olds watch an average of 17 hours and 34 minutes per week,
and 10–15 year olds 16 hours and 31 minutes (Ofcom, 2012a). In the United States
these figures are even higher with 2–5 year olds watching, on average, 32 hours of
television per week, and 6- to 11-year-olds watching 28 hours (McDonough, 2009).
Public opinion is that children should be protected from unsuitable material on TV,
including material that might scare them (Cantor, 2012; Ofcom, 2012b). For example,
the Federal Communications Commission (FCC) require that all television sets with
screens larger than 33 centimeters are fitted with the “V-Chip,” a tool enabling parents
to block content based on FCC program ratings (Federal Communications Com-
mission, 1998a). These TV ratings indicate the suitability of TV content for different
audiences. These initiatives are consistent with advice from expert psychologists that
parents closely monitor and restrict their children’s TV viewing (Cantor, 2012).
Fear and anxiety in children and adolescents is on the rise. From 1952 to 1993,
levels of anxiety in American children increased by around one standard deviation

Corresponding author: Andy P. Field; e-mail: [email protected]

98 Human Communication Research 42 (2016) 98–121 © 2015 International Communication Association

L. J. Pearce & A. P. Field “Scary” TV and Children’s Internalizing

and children in the 1980s reported higher levels of anxiety than psychiatric patients
in the 1950s (Twenge, 2000). These increased levels persist to the current day (Twenge,
2011). This generational increase in anxiety is believed to be explained partly by chil-
dren’s exposure to scary or threatening television and media (Twenge, 2000).

What is scary TV?

Most of the research has focused on the consequences of viewing violent TV on
children’s externalizing behaviors, such as conduct problems and antisocial behav-
ior (Browne & Hamilton-Giachritsis, 2005). The impact of scary or violent TV
on children’s internalized emotional responses, such as fear and anxiety, remains
under-researched. This lack of research is surprising given public policy action plans
to improve child mental health from both the Surgeon General and U.K. government
(Department for Children Schools and Families, 2010; U.S. Public Health Service,
2000). Given the relatively greater body of work looking at the effects of violent
TV on children, we are not surprised that TV parental guidelines tend to focus
on violent rather than scary content. The FCC approved TV ratings1 have content
labels for violence, fantasy violence, suggestive dialog, course or crude language, and
sexual situations but not nonviolent scary content (TV Parental Guidelines, 2012).
Similarly, although one distinction between the TVY and TVY7 ratings is whether
a program might frighten young children, the definitions of TVY7FV, TVG, TVPG,
TV14, and TVMA do not mention scary content at all; instead they reflect escalating
degrees of violence, crude language, sexual activity, or suggestive dialog (TV Parental
Guidelines, 2012). The focus on violence in TV ratings might not seem important
because violent content will often be scary; however, it could be an oversight because
scary content need not be violent.
The question of what constitutes scary televised content has been answered in a
bottom-up way by asking people what TV shows have scared them (e.g., Cantor &
Sparks, 1984; Harrison & Cantor, 1999; Valkenburg, Cantor, & Peeters, 2000). As such,
“scary” TV is defined as anything that has evoked a fear response. This approach allows
too much subjectivity: someone with a butterfly phobia would find a documentary
about butterflies Frightening, but that does not make it a scary program in general.
Others adopt a top-down approach by assuming that TV depicting events that
have threatening implications for the world, and by implication the viewer, has the
potential to scare. These studies typically focus on the impact of news reports of ter-
rorism, war, and natural disasters (e.g., Otto et al., 2007; Pfefferbaum et al., 2001; Pfef-
ferbaum et al., 2008; Smith & Moyer-Gusé, 2006; Weems, Scott, Banks, & Graham,
2012). By defining scary media in terms of local or global threats, we overlook the
countless TV shows and movies that depict fictional and fantastical events that pose
no real threat (e.g., zombies).
A definition of scary TV needs to encompass both fact and fiction, and the poten-
tial, but not certainty, of eliciting fear. We, therefore, define scary TV and movies as
those depicting situations and events in which there is social, interpersonal, physi-
cal, or psychological threat that is real, imagined, or implied (for example, through

Human Communication Research 42 (2016) 98–121 © 2015 International Communication Association 99

“Scary” TV and Children’s Internalizing L. J. Pearce & A. P. Field

contravening commonly accepted physical and natural laws). This definition allows
for TV and films to be scary because of depicting real threat (e.g., news), fictional
threat (e.g., aliens), and through direct acts (e.g., violence), psychological states (e.g.,
depicting hallucinations or psychological stress) and deviations from normal experi-
ence (e.g., zombies, your hand turning into a spider). The inclusion of psychological
and implied threat acknowledges that although scary TV might depict violence it
need not.

Does scary TV affect children’s emotional health?

At face value, there are reasons to believe that scary television can have a large impact
on children’s emotional health. Fear and intense emotional reactions are common
after media exposure (Cantor & Nathanson, 1996; Valkenburg, Krcmar, Peeters, &
Marseille, 1999; Valkenburg et al., 2000). Over 75% of 5–6 year olds report television
induced fear, and over 25% report that they experienced “great fear” (Korhonen &
Lahikainen, 2008). Although in Korhonen and Lahikainen’s study children quantified
their fear only with a self-report visual scale, other studies suggest that TV-induced
fears can be extreme enough to warrant hospitalization (Mathai, 1983) and persist for
months and even years in a substantial portion of children (Cantor & Reilly, 1982).
Perhaps more telling is that up to 90% of adults recall an incident from their childhood
when they were “really frightened” by a TV program or movie; 26.1% of these adults
still experienced residual fear or avoidance in relation to the event (Harrison & Cantor,
1999; Hoekstra, Harris, & Helmick, 1999).
Nevertheless, for the 25% of children who experienced great fear there were 75%
who did not. Consequently, what the overall impact of scary TV is on children as
a whole is unclear. In addition, there is considerable variability across studies in the
percentage of children reporting fear to TV: Estimates vary from around 30% up to
about 80% (Cantor, 1998; Korhonen & Lahikainen, 2008). This variability raises the
question of what individual or contextual variables moderate the impact that scary TV
has on an individual child. Several candidate moderators emerge from the literature,
falling under the broad headings of study characteristics, individual characteristics
and media characteristics.

Study characteristics
Although the least theoretically interesting of the three categories of moderators, we
would predict that differences in the methods adopted to study the impact of scary TV
and films on children’s emotional health will influence the size of the observed effect.
For example, experimental studies are likely to yield stronger effects because of the
tighter control of potential confounding variables. In addition, self-report measures
are more likely to be influenced by social desirability than more indirect measures of
fear, such as physiological measures, which are difficult to consciously influence.
Another measurement issue relates to how “emotional health” is defined. The log-
ical emotion to experience having watched something frightening is fear or anxiety.
The American Psychiatric Association’s Diagnostic and Statistical Manual (DSM-5)

100 Human Communication Research 42 (2016) 98–121 © 2015 International Communication Association
L. J. Pearce & A. P. Field “Scary” TV and Children’s Internalizing

is a widely used guide for diagnosing psychological disorders (American Psychiatric

Association, 2013). Within this manual, internalizing problems are those for which
the symptoms are kept internal by the person experiencing them.
There are several categories of problems including anxiety disorders, depressive
disorders, and trauma-related anxiety (such as post traumatic stress disorder, PTSD).
Sleep disturbance is common across these disorders. Although some researchers
believe that because of the comorbidity between anxiety and depression they should
share a common label (Craske et al., 2009; Watson, 2005), they are still distinguished
in DSM-5.
Therefore, theoretically we predict different effects of scary TV and film on differ-
ent measures of internalizing. In particular, we predict stronger effects on measures
directly tapping fear and anxiety (e.g., fear, PTSD) than depression. Research on the
impact of scary TV has used outcome measures that align to these internalizing prob-
lems (fear/anxiety, depression, PTSD, and sleep disturbance) making this hypothesis
testable.
Finally, parental reports of anxiety often do not converge with other indices of
children’s emotional responses: Typically parents underestimate their child’s anxi-
ety (Durbin & Wilson, 2012). Given that work on the impact of TV on children’s
emotional health sometimes relies on parental report, we predict that estimates of
anxiety will be lower when parents report on behalf of their child than when the child
self-reports.

Individual characteristics
Early research quite rightly acknowledged that child development would have a role
to play in moderating the emotional impact of scary TV. Theoretically, there are good
reasons to predict that the effect of TV should be moderated by children’s cognitive
development (Cantor & Sparks, 1984). Cognitive abilities will affect how children per-
ceive, comprehend, and interpret TV. However, the research to date has produced
conflicting results about the impact of TV on children at different stages of develop-
ment. For example, children under 10 years were most affected by TV news about
the 9/11 attacks (Otto et al., 2007), but older children were most affected by TV news
about the Iraq war (Smith & Moyer-Gusé, 2006). The reason for these opposite find-
ings is unknown. Three studies suggest that fears caused by fantasy characters reduce
as children get older but fears in relation to news or fiction depicting reality increase
(Cantor & Nathanson, 1996; Cantor & Sparks, 1984; Valkenburg et al., 2000). How-
ever, all of these studies looked at development based on age rather than specific
markers of development.
Looking at what markers of development might be important, because individ-
ual differences in processing ability should mediate emotional responses to scary
TV. Peri-traumatic perceptual, sensory-based processing rather than conceptual,
meaning-based processing, tends to lead to the intrusions and subsequent distress
associated with PTSD (Brewin & Holmes, 2003; Ehlers & Clark, 2000), and predicts
more frequent intrusions and disorganized memory (Ehlers, Mayou, & Bryant, 2003;

Human Communication Research 42 (2016) 98–121 © 2015 International Communication Association 101
“Scary” TV and Children’s Internalizing L. J. Pearce & A. P. Field

Halligan, Clark, & Ehlers, 2002; McKinnon, Nixon, & Brewer, 2008). In contrast,
greater conceptual processing during therapy for PTSD predicts a better treatment
outcome (Kindt, Buck, Arntz, & Soeter, 2007). Consequently, we predict greater
negative emotional reactions to scary media in younger children who process infor-
mation more perceptually (especially for media in which threats are depicted very
visually).

Media characteristics
We have already touched on the possibility that whether a program depicts reality
(e.g., the news) or fantasy should moderate the effect it has; this too will be influenced
by the child’s cognitive development. Children before the age of 6 are more likely to
be absorbed by fantasy material and engage with it as though it were real (Harris,
2000). Older children are more able to use conscious appraisal strategies (e.g., “There
is nothing to be afraid of, because the monster is not real”) to control their emotional
reactions to fantasy material (Harris, 2000). Therefore, we predict that because the
control of emotional involvement increases with age, scary fantasy TV will have less
impact in older children. Conversely, as children get older and start to be capable of
understanding the implicational meaning of TV depicting real events, this type of
content will have an increased effect.
Related to this point, TV and film content has changed over that time: a recent
study showed that violence in movies has increased linearly from the 1950s to 2006
(Bleakley, Jamieson, & Romer, 2012). In addition, technological advances in program
production, such as rapid improvements in computer generated imagery, is ever
increasing the realism of fantasy TV. There is evidence from videogames that this
technology will make the viewing experience more immersive than in years gone
by (Barnett & Coulson, 2010). Therefore, we predict stronger effect sizes from more
recent studies, which reflect the raised threshold of threat and violent content in
children’s TV and film.
Finally, when defining what constitutes scary TV, we acknowledged that violence
may or may not be present; given the known effect of violence on conduct prob-
lems and antisocial behavior (Browne & Hamilton-Giachritsis, 2005), we predict that
violent and scary content will have a greater impact on emotional health than scary
content alone.

Aims of the current study

Given that anxiety in youth is increasing and scary television has been posited as an
explanation for this trend, having a precise estimate of the impact that scary TV gen-
erates is important. Also, given the focus on TV violence (rather than threat content
more generally) in TV ratings (Federal Communications Commission, 1998b) poli-
cymakers need to be better informed about the emotional impact of scary television
on internalizing problems such as fear, anxiety, and sleep problems. Finally, program
makers, parents, and policy makers need to be better informed about the possible

102 Human Communication Research 42 (2016) 98–121 © 2015 International Communication Association
L. J. Pearce & A. P. Field “Scary” TV and Children’s Internalizing

moderators (e.g., age of the child) that might protect or exacerbate the impact of scary
TV on children’s emotional health.
To this end, we report a meta-analysis of the existing literature. The primary aim
was to estimate the population effect that “scary” television has on children’s emo-
tional responses associated with internalizing disorders (fear, anxiety, PTSD, sadness,
or sleep problems). The secondary aim was to see whether methodological details,
characteristics of the media, or attributes of the child moderated this effect.

Method
Conceptual rationale
Research synthesis should include studies that address an identical conceptual
hypothesis (Cooper, 2010). Our conceptual hypothesis was that “scary” television
affects the internalizing symptoms of children. Within this context, we restricted that
analysis to studies looking at the effects of television or related media (e.g., films) and
to exclude studies that tested adult samples or included adults within the sample.
We decided not to look at media more generally (e.g., newspapers, internet)
because although there is good evidence that verbal threat information can affect
fears (see Muris & Field, 2010, for a review,) these experiences are quite different
from films and TV, which will contain visual images. Similarly, although studies have
looked at the link between internet use and fear (e.g., Comer, Furr, Beidas, Babyar,
& Kendall, 2008) ‘internet use’ could entail a diffuse set of experiences, from reading
a written report to watching TV-type footage. In addition, we could find no studies
that focussed exclusively on online video media. In the interests of ensuring that
studies investigated relatively comparable childhood experiences, we focussed on
exposure to TV and films.
The focus on childhood was driven by evidence that TV might have a greater
impact under the age of 10 (Otto et al., 2007), and that pervasive internalizing prob-
lems have their roots in childhood and adolescence (Muris & Field, 2011; Silverman
& Field, 2011). As such, although internalizing problems can develop at any time of
life, preadulthood is a sensitive period during which the seeds of pervasive anxiety
are sewn (Muris & Field, 2011). Consequently we decided not only to focus on the
impact that TV has on youths 18 and under, but also to examine whether those under
10 are particularly at risk. Doing so is a first step in determining the extent to which
scary TV might lead to clinical levels of anxiety, what factors moderate this link (if it
exists), and what might be done to reduce the impact of scary TV in this particularly
vulnerable population.

Inclusion criteria
Given that other reviews have looked at the impact of violent TV on externalizing
behaviours (Browne & Hamilton-Giachritsis, 2005), this review was restricted to stud-
ies measuring children’s internalizing behaviors. We considered any program or film
containing threat content (e.g., news reports about a terrorist attack) or intending to

Human Communication Research 42 (2016) 98–121 © 2015 International Communication Association 103
“Scary” TV and Children’s Internalizing L. J. Pearce & A. P. Field

scare (e.g., a televised adaptation of a horror novel) to be “scary TV” (or TV/film that
might be expected to influence internalizing symptoms). A study was included if it
met these criteria:

1. Outcomes were reported only for children aged 18 years and under, including
self-report by the child, parent reports of child behavior, and physiological mea-
sures in an experimental laboratory setting. If a study included an age range with
an upper limit beyond 18 years, it was excluded.
2. The primary outcomes included measures of internalizing behaviors such as fear,
anxiety, worry, sadness, and depression. Studies which focused specifically on
externalizing behaviors such as aggression and violence were excluded because
the focus of this study was internalized emotions.
3. Initially, studies were selected for inclusion if the outcome measures were clinically
validated scales or physiological measures. However, due to the limited number
of studies including such measures, inclusion criteria were broadened to include
nonvalidated self-report measures such as Likert and visual analogue measures of
fear, anxiety, etc.
4. Experimental studies must include a control group or condition, either a baseline
measurement before exposure to television, or a group with no, or limited exposure
to television. Correlational studies must have measured the quantity of exposure
to warrant inclusion.
5. There needed to be sufficient information to compute effect sizes.

Figure 1 shows our search strategy. An initial search was run in January 2012
using electronic databases, reference lists of reviews and relevant studies, and a hand
search of books. Searches were conducted using PsycARTICLES/PsycINFO, PubMed,
Web of Knowledge, and Google Scholar with the following keywords used as search
terms alone and in combination: Child, Television, Scary, Emotion, Negative, Positive,
Fear, Anxiety, Depression, Sad. The reference lists of relevant reviews and studies were
examined to identify any further studies for inclusion. The library catalogue at our
university was searched using the same terms as the electronic database to identify
books to be hand searched. In addition, books which included chapters on children’s
emotional reactions to the media in general were also hand searched. This search
identified 101 studies in which children’s reactions to scary television were reported;
Eighty of these did not meet all inclusion criteria.
After this initial search, 60 experts in the field and lead researchers for the studies
identified by the initial search were emailed with the list of studies produced by the
initial database search and were asked to add any further studies of which they were
aware. They were also asked for any relevant unpublished data that fit the inclusion
criteria. This effort yielded no additional studies.
A second database search was conducted in February 2013 to identify any studies
published since the first search. The same methods as the first search were used. This
search identified a further five studies, one of which did not have sufficient data to

104 Human Communication Research 42 (2016) 98–121 © 2015 International Communication Association
L. J. Pearce & A. P. Field “Scary” TV and Children’s Internalizing

Figure 1 The search strategy for articles included in the meta-analysis.

compute effect sizes. Where effect sizes could not be computed authors were contacted
for additional information but none provided it.
A final round of searches was conducted just prior to submission. This included
another database search but adding the search terms TV, Film, Movie, and Happy in
isolation and in combination with the search terms on our original list. This yielded 25

Human Communication Research 42 (2016) 98–121 © 2015 International Communication Association 105
“Scary” TV and Children’s Internalizing L. J. Pearce & A. P. Field

studies that were not included in our January 2012 and February 2013 searches. Unfor-
tunately only one of these met the inclusion criteria (see Figure 1). We also e-mailed 66
researchers (the 19 nonresponders from our January 2012 e-mail, and 47 new names).
This yielded two published studies that were included (Pfefferbaum et al., 1999, 2008),
two unpublished studies that did not meet the inclusion criteria, and three datasets
from published studies that had not reported data on fear outcomes but included mea-
sures that could be used in the meta-analysis (Davies & Gentile, 2012; Ferguson, San
Miguel, & Hartley, 2009; Gentile & Walsh, 2002). In total, 31 studies met our inclu-
sion criteria and were included in the analysis. Across studies the analysis is based on
12,454 individuals.

Analysis strategy
For each study, Pearson’s r was computed as the effect size estimate of television view-
ing on the outcome variables, and converted to Zr. In correlational studies, r was taken
directly from the articles, otherwise standard and widely used formulae were used to
convert descriptive information (e.g., means and standard deviations, frequencies)
into r (see, e.g., Borenstein, Hedges, Higgins, & Rothstein, 2009; Rosenthal, 1991).
Some studies produced several effect sizes, which violates the assumption of inde-
pendence of effect sizes and means that studies with multiple effect sizes are double
(or more) counted. This is a common issue in meta-analysis (Field, in press). Three
of the leading texts in meta-analysis (Borenstein et al., 2009; Cooper, 2010; Hunter
& Schmidt, 2004) suggest two (broadly) acceptable approaches. The first is to use the
study as the unit of analysis and aggregate effect sizes within studies. This approach
potentially misses the nuances of what happens in subgroups within a study. However,
provided those groups have a common control it is still meaningful to aggregate them
because the nuances can be teased apart later using moderator analysis. In short, con-
duct an overall analysis that retains the independence of effect sizes (as the statistical
model assumes) followed by moderator analyses that forgo that assumption to tease
apart nuances (Hunter & Schmidt, 2004).
The second approach is to treat effect sizes as the unit of analysis; that is “pretend”
that effect sizes within a study are independent. This method, at best, results in giving
studies with multiple effect sizes more weight than those yielding a single effect sizes,
and at worst biases the parameter estimates from the analysis (Cheung, in press).
A preferable approach is to treat the meta-analysis as a multilevel model (Che-
ung, in press; Field, in press; Goldstein, Yang, Omar, Turner, & Thompson, 2000;
Konstantopoulos, 2011; Raudenbush & Bryk, 2002) in which effect sizes (level 1) are
nested within studies (level 2). In doing so, dependence between effect sizes within
studies is explicitly modeled. For the overall analysis, effect sizes are predicted from
the model intercept (i.e., the average effect size weighted by sampling precision),
which is treated as a random effect, but moderators can also be added as fixed effect
predictors.
With these concerns in mind, the overall analysis and publication bias anal-
ysis was conducted with effect sizes aggregated within studies. A random-effects

106 Human Communication Research 42 (2016) 98–121 © 2015 International Communication Association
L. J. Pearce & A. P. Field “Scary” TV and Children’s Internalizing

meta-analysis (Hedges & Vevea, 1998) using the DerSimonian-Laird estimator of

between study variability (Dersimonian & Laird, 1986) was used for three rea-
sons: first, in real social science data effect size variability is the norm (Field, 2005;
Hunter & Schmidt, 2000); second, the random effects model reduces to the fixed
effect model if effect sizes do not vary but not vice versa (Field, 2003); and finally
random-effects meta-analyses enable the results to be generalized beyond the studies
included in the meta-analysis (Hedges & Vevea, 1998). The result is an average
effect size (and confidence interval) that is weighted by sampling precision (that
is, studies with large samples are weighted more heavily than ones based on small
samples).
To quantify the likely effect of publication bias, a sensitivity analysis based on
Vevea and Woods (2005) was conducted. This analysis adjusts the population effect
size estimate for moderate and severe one- and two-tailed selection bias using.
For moderator analysis, a multilevel model was fitted with effect sizes nested
within studies. The intercept was a random effect (varying across studies) and mod-
erators treated as fixed effects. Using a multilevel approach enabled us to maximize
the power of moderator analyses by including effect sizes within studies connected
to different levels of the moderator variable of interest.
All analyses were conducted using R 3.1.1. (R Core Team, 2014). The overall analy-
sis was conducted using the functions rma() function in the metafor package (Viecht-
bauer, 2010). The moderator analysis multilevel models were fitted with the rma.mv()
function in the metafor package. The publication bias analysis used Vevea and Woods’
scripts for S-plus adapted for R.

Moderator variables
Ten variables were examined as possible moderators of the relationship between
scary TV and children’s internalizing responses. These reflect the three categories
of study characteristics (1–4), individual characteristics (5–6), and media charac-
teristics (7–10): (1) methodology (experiment vs. correlational); (2) measure type
(self-report vs. physiological measures); (3) responder (did the study use parent or
child report or the child’s emotional state?); (4) outcome measure (what internalizing
response was measured: fear, PTSD, physiological, sadness, sleep problems?); (5) The
mean age of participants; (6) age under 10 (were participants aged under 10 or not?);
(7) age of the study (years since publication); (8) media type (did the study measure
purely exposure to scary TV of did their measure of exposure allow for other forms of
scary media too?); (9) media content (was the media content in the study factual or fic-
tional?); and (10) violent content. For violent content, we could not always determine
the exact content of the TV program to which children were exposed; therefore, stud-
ies were split into three categories: studies where we could be sure that the content was
not violent (k = 4), studies containing general measures of TV exposure, which would
include both violent and non-violent programs (k = 8), and studies that specifically
looked at violent events such as news reports of terrorist attacks, wars, assassinations,
etc. (k = 19).

Human Communication Research 42 (2016) 98–121 © 2015 International Communication Association 107
“Scary” TV and Children’s Internalizing L. J. Pearce & A. P. Field

Becker−Blease et al. (2008) 0.21 [ 0.16 , 0.25 ]

Braun. Lewensohn et al. (2009) 0.05 [ . 0.02 , 0.11 ]
Buijzen et al. (2007) 0.28 [ 0.18 , 0.37 ]
Comer et al. (2008) 0.26 [ −0.08 , 0.60 ]
Davies & Gentile (2012) 0.15 [ 0.10 , 0.20 ]
Fairbrother et al. (2003) 0.14 [ 0.05 , 0.24 ]
Ferguson et al. (1999) 0.06 [ . 0.02 , 0.14 ]
Gentile & Walsh (2002) 0.15 [ 0.06 , 0.24 ]
Gilissen et al. (2007) 0.25 [ 0.02 , 0.48 ]
Gilissen et al. (2008) 0.44 [ 0.22 , 0.66 ]
Lengua et al.(2005) 0.11 [ −0.05 , 0.28 ]
Ortiz et al. (2011) 0.18 [ 0.06 , 0.31 ]
Otto et al.(2007) 0.12 [ . 0.04 , 0.27 ]
Paavonen et al. (2006) 0.15 [ 0.03 , 0.26 ]
Paavonen et al. (2009) 0.28 [ 0.17 , 0.40 ]
Pfefferbaum et al. (1999) 0.07 [ . 0.02 , 0.17 ]
Pfefferbaum et al. (2000) 0.32 [ . 0.11 , 0.74 ]
Pfefferbaum et al. (2001) 0.24 [ 0.20 , 0.29 ]
Pfefferbaum et al. (2003) 0.34 [ 0.10 , 0.58 ]
Pfefferbaum et al. (2008) 0.12 [ . 0.06 , 0.30 ]
Phillips et al. (2004) 0.33 [ 0.03 , 0.62 ]
Saylor et al. (2003) 0.22 [ 0.07 , 0.37 ]
Schredl et al. (2008) 0.10 [ . 0.02 , 0.23 ]
Schuster et al. (2001) 0.28 [ 0.12 , 0.43 ]
Smith & Moyer–Guse (2006) 0.21 [ 0.05 , 0.36 ]
Smith & Wilson (2002) . 0.09 [ . 0.27 , 0.09 ]
Sparks & Cantor (1986) 0.00 [ −0.28 , 0.29 ]
Unz et al. (2008) 0.56 [ 0.39 , 0.73 ]
van der Molen & Bushman (2008) 0.14 [ 0.03 , 0.25 ]
Weems et al. (2012) 0.19 [ 0.03 , 0.36 ]
Wilson et al. (2005) 0.16 [ 0.01 , 0.30 ]

RE Model

. 0.40 0.00 0.40

Figure 2 Forest plot of the studies in the meta-analysis.

Results
Overall analysis
Figure 2 shows a forest plot (Lewis & Clarke, 2001) with the effect of scary TV on chil-
dren’s internalizing emotions (Zr) for each study and its associated 95% confidence
interval. The hashed vertical line indicates no effect. There was consistent evidence
that the effect of scary TV on children’s internalizing responses is greater than zero
in the population: 20 of the 31 studies had confidence intervals that did not cross the
hashed line.
There was significant heterogeneity in effect sizes, Q(30) = 93.38, I 2 = 67.87%,
p < .001, indicating variability in study findings that might be explained by mod-
erators. Overall, there was a modest but consistent effect of exposure to scary TV
on children’s internalizing responses, Ẑ 𝜌 = .18, 95% CI (.14, .22), p < .001.2 When
converted back to r this is an overall effect of r = .178.

108 Human Communication Research 42 (2016) 98–121 © 2015 International Communication Association
L. J. Pearce & A. P. Field “Scary” TV and Children’s Internalizing

0.000

0.054
Standard Error

0.109

0.163

0.217

. 0.20 0.00 0.20 0.40 0.60

Observed Outcome

Figure 3 Funnel plot of the studies in the meta-analysis.

Publication bias
Figure 3 shows a funnel plot (Egger, Smith, Schneider, & Minder, 1997) of the studies
in the analysis. This plot shows evidence of publication bias (the so-called “file draw-
er” problem) in that the lower left corner, which represents small studies showing
small effects, is relatively empty. This implies that small studies showing small effects
are missing because they have failed to be published (they remain in the author’s file
drawer). To quantify the likely effect of publication bias, a sensitivity analysis based on
Vevea and Woods (2005) was conducted which adjusts the population effect size esti-
mate for moderate and severe one- and two-tailed selection bias. In effect, this analysis
answers the question of ‘what would the overall effect size from the meta-analysis
have been had we been able to access all of the unpublished studies?’ The degree to
which the adjusted population effect size estimate differs from the unadjusted value
indicates the extent to which publication bias has influenced the original analysis.
The original analysis yielded a population estimate of Ẑ 𝜌 = .18, and the most different
adjusted estimate was for severe one-tailed selection bias, Ẑ 𝜌 = .15, which is a rela-
tively trivial reduction in the population effect size estimate and would not change
our interpretation of it. These findings give us confidence that the estimated pop-
ulation effect size has not been severely inflated by unpublished studies not in the
meta-analysis.

Human Communication Research 42 (2016) 98–121 © 2015 International Communication Association 109
“Scary” TV and Children’s Internalizing L. J. Pearce & A. P. Field

Table 1 Effects of Moderators on the Relationship Between Exposure to Scary TV and

Internalizing Responses

95% CI
Moderator k Estimate Lower Upper Test p

Study characteristics
Experimental or correlational 129 −0.159 −0.269 −0.049 χ2 (1) = 8.08 .005
Self-report of physiological outcome 129 −0.177 −0.359 0.005 χ2 (1) = 3.64 .056
Responder 129 χ2 (2) = 0.14 .932
Child vs. Parent −0.010 −0.067 0.047 z = −0.34 .736
Child vs. Both 0.016 −0.215 0.246 z = 0.13 .895
Outcome measure 129 χ2 (5) = 22.39 < .001
General vs. Fear 0.213 0.111 0.315 z = 4.11 < .001
General vs. Physiology 0.356 0.150 0.561 z = 3.39 < .001
General vs. PTSD 0.126 0.050 0.201 z = 3.25 .001
General vs. Sadness/Depression 0.244 0.132 0.356 z = 4.28 < .001
General vs. Sleep 0.208 0.061 0.355 z = 2.77 .006
Individual characteristics
Mean age 129 −0.006 −0.022 0.009 χ2 (1) = 0.62 .432
Age under 10 129 0.123 0.001 0.246 χ2 (1) = 3.88 .049
Media characteristics
Study age 129 −0.007 −0.017 0.002 χ2 (1) = 2.31 .128
Media type (TV only or mixed media) 117 0.035 −0.076 0.145 χ2 (1) = 0.38 .539
Media Content (fact vs. fantasy) 129 −0.017 −0.067 0.033 χ2 (1) = 0.45 .501
Violent content 129 χ2 (2) = 2.50 .287
No violence vs. General −0.102 −0.256 0.052 z = −1.30 .193
No violence vs. Violent −0.034 −0.175 0.108 z = −0.47 .640

Moderators of the impact of scary TV

Table 1 summarizes the multilevel models (see earlier) in which each moderator
predicts effect sizes. In terms of study characteristics, whether the child or par-
ent responded did not significantly influence effect sizes. However, the following
moderators were significant: (a) the methodology used, with effect sizes signifi-
cantly stronger in experimental designs (k = 17), Ẑ 𝜌 = .30 [.13, .47], p < .001, than
correlational ones (k = 112), Ẑ 𝜌 = .17 [.13, .20], p < .001; (b) there was borderline
significance that effect sizes measured using self-report (k = 123), Ẑ 𝜌 = .18 [.14, .22],
p < .001, were weaker than those measured physiologically (k = 6), Ẑ 𝜌 = .37 [.24, .50],
p < .001; and (c) compared to general indicators of internalizing (k = 3), Ẑ 𝜌 = .10
[−.10, .30], p = .325, effect sizes were significantly stronger for specific measures
such as fear (k = 68), Ẑ 𝜌 = .19 [.13, .25], p < .001, physiology (k = 6), Ẑ 𝜌 = .37 [.24,
.50], p < .001, PTSD, (k = 32), Ẑ 𝜌 = .17 [.10, .24], p < .001, sadness or depression
(k = 11), Ẑ𝜌 = .27 [.06, .47], p = .010, and sleep disorders (k = 9), Ẑ 𝜌 = .18 [.10, .27],
p < .001.3

110 Human Communication Research 42 (2016) 98–121 © 2015 International Communication Association
L. J. Pearce & A. P. Field “Scary” TV and Children’s Internalizing

For individual characteristics, the mean age of participants did not significantly
predict effect sizes, however, effect sizes from participants who were under 10 years
old (k = 16), Ẑ𝜌 = .28 [.17, .38], p < .001, were significantly stronger than those from
above the age of 10 (k = 113), Ẑ 𝜌 = .18 [.13, .22], p < .001.
For media characteristics, all moderators had a small, nonsignificant, influence on
study effect sizes: media type (purely TV or other media too?), media content (factual
vs. fictional content), and the presence of violent content.

Discussion
Our research supports the long-held belief that “scary” TV can have an impact on
children’s own fears and anxiety but also provides a better estimate of the true pop-
ulation effect of scary TV than in any individual study. Although other studies have
assimilated studies examining the impact of TV viewing on externalizing behaviors
(Paik & Comstock, 1994; Wood, Wong, & Chachere, 1991), ours is the first to quantify
its impact on internalizing. The results showed that exposure to scary TV was consis-
tently associated with internalizing responses such as fear, anxiety, PTSD, depression,
and sleep problems. Interestingly, the overall effect of scary TV on internalizing symp-
toms (r = .18) was comparable to the impact that violent TV has on externalizing
behaviors: Meta-analyses focusing on violent TV suggest that it has fairly small effect
(d = 0.27, which equates to r = .13) on aggression (Wood et al., 1991) and a larger
effect (r = .31) on antisocial behavior (Paik & Comstock, 1994). The effect of scary
TV on internalizing fell between these values.

Study characteristics
Several predictions were made about the impact that study characteristics would
have on effect sizes. Contrary to expectation, effect sizes were not stronger from
child report compared to parent report, however, the rest of our predictions were
broadly on the mark. Experimental research seemed to yield larger effects than corre-
lational research, and physiological measures yielded stronger effects than self-report
(although this effect was borderline significant). This finding could reflect the greater
ability of experimental studies to limit the background noise against which effects
are measured; however, it could also be due to the relatively short follow-up period
typically used in experimental studies compared to correlation studies, which often
assess internalizing months after exposure to scary TV.
The most theoretically interesting finding was that specific measures of internal-
izing yielded stronger effects than general ones. In addition, PTSD measures yielded
significantly lower effect sizes than measures of fear. Two caveats to these findings are
that the effect of general measures is based on only three effect sizes. Although this
means that the effect is likely to be big because our power to detect it would have been
low, small samples are inherently more variable than large ones and the significance
could reflect a fluke finding that will not replicate. Also, despite the significant differ-
ence, effect sizes from fear measures (Ẑ 𝜌 = .19) and PTSD measures (Ẑ 𝜌 = .17) were

Human Communication Research 42 (2016) 98–121 © 2015 International Communication Association 111
“Scary” TV and Children’s Internalizing L. J. Pearce & A. P. Field

comparable from a practical perspective. However, the key messages for researchers
are that (a) we need additional well-controlled experimental research that capitalizes
on the increased sensitivity that these designs afford; (b) we should not rely solely
on self-report; and (c) we should target specific measures of internalizing rather than
general ones.
The theoretically interesting message is that differences were not found with
respect to fear, depression, and sleep problems. This finding suggests that scary
TV has a nonspecific effect on internalizing: It does not create fear only, it is also
associated with depressive symptoms and sleep disturbance. Notwithstanding the
debate about whether anxiety and depression should represent separate diagnostic
categories (Craske et al., 2009; Watson, 2005), the impact of scary TV on these
different manifestations of internalizing symptoms is comparable.

Individual characteristics
We predicted that scary TV would have a greater impact on younger children because
they process more perceptually and are less able to regulate their emotional responses.
There is some empirical evidence that age 10 years might be a reasonable threshold for
detecting these developmental differences (Cantor & Sparks, 1984; Otto et al., 2007;
Valkenburg et al., 2000). Although the mean age of the sample did not significantly
predict effect sizes, we did find support for the theory that children under 10 might
be particularly vulnerable to scary TV. This finding makes sense given that, compared
to older children, younger children seem to process more perceptually and are, there-
fore, affected more by “visual” news stories, for example, those about natural disasters
(Cantor & Nathanson, 1996). However, our analysis used age as a proxy for markers of
cognitive development (such as processing style). Age is not important, but rather the
psychological correlates of age that affects how the child processes the media that they
watch. More experimental and longitudinal studies are needed that track theoretically
driven cognitive and emotional skills (e.g., emotion regulation, understanding of fact
and fiction, and processing styles), TV viewing, and internalizing symptoms.

Media characteristics
Although our focus was on scary TV rather than violent TV, scary TV often portrays
violence. We noted earlier that the overall effect of scary TV on internalizing in the
current study was comparable to similar analyses looking at the effect of violent TV
on externalizing. This correspondence could be explained by the violent content in
the scary TV used in the studies in our meta-analysis.
For example, 19 of the 31 studies in our analysis investigated the impact of vio-
lent world events such as wars or terrorist attacks. In only four studies could we be
relatively certain that the programs used were scary but not violent. When compar-
ing these studies to others that definitely contained violence and a category of studies
for which the violent content was uncertain, we found no significant differences: The
level of violence (coded per study) had virtually no impact in moderating the effect

112 Human Communication Research 42 (2016) 98–121 © 2015 International Communication Association
L. J. Pearce & A. P. Field “Scary” TV and Children’s Internalizing

sizes. This result suggests that violent content is not driving the overall effect of scary
TV on internalizing.
In addition, whether the content was fact or fantasy had no significant impact,
which contradicts the theoretical expectation that TV depicting real threats will have
a greater impact than those depicting fictional ones. However, this effect is likely to be
further moderated by cognitive development. Several of studies suggested that emo-
tional reactions to scary fantasy TV declined with cognitive development (Cantor &
Sparks, 1984), whereas emotional reactions to factual TV increased as children better
comprehend it (Smith & Wilson, 2002). The finding that fact or fantasy content did
not moderate effect sizes could be because the effect will vary as a function of cogni-
tive development, and there were too few studies to test this hypothesis. As a result, we
still know relatively little about the interplay between developmental processes and
emotional reactions to TV. Controlled experimental studies that manipulate age and
media type, or longitudinal studies that track the impact of scary TV across age
and media content are needed before we can draw firm conclusions.
We would also note that comparing factual and fictional TV might be like com-
paring apples and pears: Controlling for the scariness of the content (that is, tightly
control the factual vs. fantasy elements while keeping other aspects of the program
comparable) would be extremely difficult. Therefore, although the nonsignificance of
our findings suggested that the effect on internalising of factual and fantasy TV was
comparable,4 we think it is hard to know what this comparison actually represents
because content other than the factual or fantastical nature of the TV will not have
been tightly controlled in this comparison.

Future directions
The overall effect of scary TV on internalizing behaviors was modest compared to
the levels of fear experienced by an important minority of children in individual
studies. For example, studies have reported PTSD symptoms on or around the
clinical threshold for subgroups of children exposed to scary TV (Holmes, Creswell,
& O’Connor, 2007; Otto et al., 2007). Although these results might at first seem
contradictory, a meta-analysis invariably glosses over these individual differences by
focusing on group-level effects. The difference between the group-level effect (quan-
tified by the meta-analysis) and the extremity of anxiety experienced by individual
youths in particular studies suggests that there are individual differences in children’s
processing of or susceptibility to scary TV that have yet to be explored within this
small literature.
For example, certain personality characteristics and coping styles are known to
correlate with resilience in adolescents: adolescents with high scores on disagreeable-
ness and emotional instability and low scores on extroversion, agreeableness, and
openness to new experiences were considered to have a “low protective” personal-
ity cluster (Davey, Eaker, & Walters, 2003). We believe it makes theoretical sense to
hypothesize that children with this personality profile might be particularly vulnera-
ble to negative televisual emotional experiences. The studies within the meta-analysis

Human Communication Research 42 (2016) 98–121 © 2015 International Communication Association 113
“Scary” TV and Children’s Internalizing L. J. Pearce & A. P. Field

did not permit this fine-grained analysis of the interaction between personality clus-
ters and TV, Therefore, future research needs to unpack the individual differences in
susceptibility to scary televised content.
Related to personality is child temperament. As with personality more generally,
little is known about how the temperament of the child affects both what they watch
and their reactions to it. For example, based on theory we might expect the effects
of scary TV to be greater in children with risk factors for anxiety such as behavioral
inhibition or trait anxiety (Lonigan, Philips, Wilson, & Allan, 2011). The only available
evidence has suggested that child and family anxiety histories do not exacerbate the
effects of scary TV (Otto et al., 2007); however, this study used quite broad brush
measures of TV exposure to a specific event (the 9/11 attacks) and more systematic
research is needed.

Policy implications
Based on the first 25 years or so of research there is little to suggest that, at the group
level, scary TV has a severe impact on children’s mental health (in terms of internaliz-
ing problems). However, such programming does have a consistent effect that is com-
parable to effects found for externalizing behaviours. Given the ubiquity of violence
within many of the programs used in the studies included in the meta-analysis, our
findings suggested that the current FCC TV parental rating guidelines (TV Parental
Guidelines, 2012) are wise to highlight violent content: it affected children’s internal-
ized emotions as much as it did their externalized ones. However, given that the level
of violence did not moderate effect sizes, there was a suggestion that the scariness of
the program, irrespective of the violent content, might also be an important criterion
for TV parental ratings. This conclusion is somewhat speculative given the inherent
difficulty in classifying the studies in the meta-analysis by the amount of violence
within the scary TV.
Furthermore, given that some individual studies do show extreme reactions in a
substantial minority of children, the important question for policy makers is what
characteristics of the TV and of the child predict the strength of effect that scary TV
has on an individual. Similarly, although the advice that all parents monitor what their
children watch (Cantor, 2012) is generally supported by our analysis, parents of cer-
tain types of children should be more mindful of what television their children view
than others. Until we know more about what characteristics of children make them
particularly vulnerable, or which specific characteristics of the TV program drive the
effect giving parents specific advice that best suits their child is difficult. The point to
take forward is that scary TV certainly does affect children’s internalized emotions,
but more needs to be known about the parameters affecting when and why it has this
effect.

Studies in the meta-analysis

(Becker-Blease, Finkelhor, & Turner, 2008; Braun-Lewensohn, Celestin-Westreich,
Celestin, Verte, & Ponjaert-Kristoffersen, 2009; Buijzen, van der Molen, & Sondij,

114 Human Communication Research 42 (2016) 98–121 © 2015 International Communication Association
L. J. Pearce & A. P. Field “Scary” TV and Children’s Internalizing

2007; Comer et al., 2008; Davies & Gentile, 2012; Fairbrother, Stuber, Galea, Fleis-
chman, & Pfefferbaum, 2003; Ferguson et al., 2009; Gentile & Walsh, 2002; Gilissen,
Bakermans-Kranenburg, van Ijzendoorn, & van der Veer, 2008; Gilissen, Koolstra,
van Ijzendoorn, Bakermans-Kranenburg, & van der Veer, 2007; Lengua, Long, Smith,
& Meltzoff, 2005; Ortiz, Silverman, Jaccard, & La Greca, 2011; Otto et al., 2007; Paavo-
nen, Pennonen, Roine, Valkonen, & Lahikainen, 2006; Paavonen, Roine, Pennonen, &
Lahikainen, 2009; Pfefferbaum et al., 1999, 2000, 2001, 2003, 2008; Phillips, Prince, &
Schiebelhut, 2004; Saylor, Cowart, Lipovsky, Jackson, & Finch, 2003; Schredl, Anders,
Hellriegel, & Rehm, 2008; Schuster et al., 2001; Smith & Moyer-Gusé, 2006; Smith &
Wilson, 2002; Sparks & Cantor, 1986; Unz, Schwab, & Winterhoff-Spurk, 2008; van
der Molen & Bushman, 2008; Weems et al., 2012; Wilson, Martins, & Marske, 2005)

Acknowledgments
Andy P. Field wrote this article based on work by Laura Pearce for her Masters disser-
tation, which he supervised.

Notes
1 The FCC rating categories are, in broadly ascending order, suitable for all children (TVY)
and general audience (TVG), children over 7 years old only (TVY7 or TVY7FV if the
program included fantasy violence), parental guidance needed to determine suitability
(TVPG), suitable for children over 14 only (TV14), and suitable for audiences over 17
years old only (TVMA).
2 Running this overall analysis as a multilevel model in which studies could contribute
multiple effect sizes yielded comparable parameter estimate, Ẑ 𝜌 = .19, 95% CI (.15, .24),
p < .001.
3 In a different analysis, we coded fear as the baseline category rather than ‘general’. Of
course, as in the other analysis, general measures yield significantly smaller effect sizes
than fear, but this analysis showed that PTSD measures yield significantly lower estimates
than fear, b = −.088 [−.16, −.01], p = .02. Physiological, depression and sleep disturbance
did not yield effect sizes significantly different from fear.
4 Ignoring statistical power as an explanation.

References
American Psychiatric Association (2013). Diagnostic and statistical manual of mental
disorders (Vol., 5th ed.). Arlington, VA: American Psychiatric Publishing.
Barnett, J., & Coulson, M. (2010). Virtually real: A psychological perspective on massively
multiplayer online games. Review of General Psychology, 14, 167–170.
Becker-Blease, K. A., Finkelhor, D., & Turner, H. (2008). Media exposure predicts children’s
reactions to crime and terrorism. Journal of Trauma & Dissociation, 9(2), 225–248.
doi:10.1080/15299730802048652
Bleakley, A., Jamieson, P. E., & Romer, D. (2012). Trends of sexual and violent content by
gender in top-grossing U.S. films, 1950–2006. Journal of Adolescent Health, 51(1), 73–79.
doi:10.1016/j.jadohealth.2012.02.006

Human Communication Research 42 (2016) 98–121 © 2015 International Communication Association 115
“Scary” TV and Children’s Internalizing L. J. Pearce & A. P. Field

Borenstein, M., Hedges, L. V., Higgins, J. P. T., & Rothstein, H. R. (2009). Introduction to meta
analysis. Chichester, England: Wiley-Blackwell.
Braun-Lewensohn, O., Celestin-Westreich, S., Celestin, L.-P., Verte, D., &
Ponjaert-Kristoffersen, I. (2009). Adolescents’ mental health outcomes according to
different types of exposure to ongoing terror attacks. Journal of Youth and Adolescence,
38(6), 850–862. doi:10.1007/s10964-008-9305-8
Brewin, C. R., & Holmes, E. A. (2003). Psychological theories of posttraumatic stress disorder.
Clinical Psychology Review, 23(3), 339–376. doi:10.1016/s0272-7358(03)00033-3
Browne, K. D., & Hamilton-Giachritsis, C. (2005). The influence of violent media on children
and adolescents: A public-health approach. Lancet, 365(9460), 702–710.
doi:10.1016/s0140-6736(05)70938-7
Buijzen, M., van der Molen, J. H. W., & Sondij, P. (2007). Parental mediation of children’s
emotional responses to a violent news event. Communication Research, 34(2), 212–230.
doi:10.1177/0093650206298070
Cantor, J. (1998). Ratings for program content: The role of research findings. Annals of the
American Academy of Political and Social Science, 557, 54–69.
Cantor, J. (2012). The media and children’s fears, anxieties, and perceptions of danger. In D.
G. Singer & J. L. Singer (Eds.), Handbook of children and the media (Vol., 2nd ed., pp.
215–229). Thousand Oaks, CA: Sage.
Cantor, J., & Nathanson, A. I. (1996). Children’s fright reactions to television news. Journal of
Communication, 46(4), 139–152.
Cantor, J., & Reilly, S. (1982). Adolescents fright reactions to television and films. Journal of
Communication, 32(1), 87–99.
Cantor, J., & Sparks, G. G. (1984). Children’s fear responses to mass-media – Testing some
Piagetian predictions. Journal of Communication, 34(2), 90–103.
Cheung, M. W.-L. (in press). Modeling dependent effect sizes with three-level meta-analyses:
A structural equation modeling approach. Psychological Methods. doi:10.1037/a0032968
Comer, J. S., Furr, J. M., Beidas, R. S., Babyar, H. M., & Kendall, P. C. (2008). Media use and
children’s perceptions of societal threat and personal vulnerability. Journal of Clinical
Child and Adolescent Psychology, 37(3), 622–630. doi:10.1080/15374410802148145
Cooper, H. M. (2010). Research synthesis and meta-analysis: A step-by-step approach (Vol., 4th
ed.). Thousand Oaks, CA: Sage.
Craske, M. G., Rauch, S. L., Ursano, R., Prenoveau, J., Pine, D. S., & Zinbarg, R. E. (2009).
What is an anxiety disorder? Depression and Anxiety, 26(12), 1066–1085.
doi:10.1002/da.20633
Davey, M., Eaker, D. G., & Walters, L. H. (2003). Resilience processes in adolescents:
Personality profiles, self-worth, and coping. Journal of Adolescent Research, 18(4),
347–362. doi:10.1177/074355840323810
Davies, J. J., & Gentile, D. A. (2012). Responses to children’s media use in families with and
without siblings: A family development perspective. Family Relations, 61(3), 410–425.
doi:10.1111/j.1741-3729.2012.00703.x
Department for Children Schools and Families. (2010). Promoting the emotional health of
children and young people: Guidance for Children’s Trusts partnerships, including how
to deliver NI50.
Dersimonian, R., & Laird, N. (1986). Meta-analysis in clinical-trials. Controlled Clinical
Trials, 7(3), 177–188.

116 Human Communication Research 42 (2016) 98–121 © 2015 International Communication Association
L. J. Pearce & A. P. Field “Scary” TV and Children’s Internalizing

Durbin, C. E., & Wilson, S. (2012). Convergent validity of and bias in maternal reports of
child emotion. Psychological Assessment, 24(3), 647–660. doi:10.1037/a0026607
Egger, M., Smith, G. D., Schneider, M., & Minder, C. (1997). Bias in meta-analysis detected
by a simple, graphical test. British Medical Journal, 315(7109), 629–634.
Ehlers, A., & Clark, D. M. (2000). A cognitive model of posttraumatic stress disorder.
Behaviour Research and Therapy, 38(4), 319–345.
Ehlers, A., Mayou, R. A., & Bryant, B. (2003). Cognitive predictors of posttraumatic stress
disorder in children: Results of a prospective longitudinal study. Behaviour Research and
Therapy, 41(1), 1–10.
Fairbrother, G., Stuber, J., Galea, S., Fleischman, A. R., & Pfefferbaum, B. (2003).
Posttraumatic stress reactions in New York City children after the September 11, 2001,
terrorist attacks. Ambulatory Pediatrics, 3(6), 304–311.
doi:10.1367/1539-4409(2003)003<0304:psriny>2.0.co;2
Federal Communications Commission. (1998a). Report and order: In the matter of technical
requirements to enable blocking of video programming based on program ratings;
implementation of sections 551(c), (d), and (e) of the telecommunications act of 1996.
Washington, DC. Federal Communications Commission.
http://transition.fcc.gov/Bureaus/Engineering_Technology/Orders/1998/fcc98036.html
Federal Communications Commission. (1998b). Report and order: In the matter of the
implementation of section 551 of the telecommunications act of 1998 video programming
ratings. Washington, DC. Federal Communications Commission.
http://transition.fcc.gov/Bureaus/Cable/Orders/1998/fcc98035.html
Ferguson, C. J., San Miguel, C., & Hartley, R. D. (2009). A multivariate analysis of youth
violence and aggression: The influence of family, peers, depression, and media violence.
The Journal of Pediatrics, 155(6), 904–908.e903. doi:10.1016/j.jpeds.2009.06.021
Field, A. P. (2003). The problems in using Fixed-effects models of meta-analysis on real-world
data. Understanding Statistics, 2, 77–96.
Field, A. P. (2005). Is the meta-analysis of correlation coefficients accurate when population
correlations vary? Psychological Methods, 10(4), 444–467.
doi:10.1037/1082-989x.10.4.444
Field, A. P. (in press). Dread returns to Mega-Silly One. Health Psychology Review.
doi:10.1080/17437199.2013.879198
Gentile, D. A., & Walsh, D. A. (2002). A normative study of family media habits. Journal of
Applied Developmental Psychology, 23(2), 157–178. doi:10.1016/s0193-3973(02)00102-8
Gilissen, R., Bakermans-Kranenburg, M. J., van Ijzendoorn, M. H., & van der Veer, R. (2008).
Parent–child relationship, temperament, and physiological reactions to fear-inducing
film clips: Further evidence for differential susceptibility. Journal of Experimental Child
Psychology, 99(3), 182–195. doi:10.1016/j.jecp.2007.06.004
Gilissen, R., Koolstra, C. A., van Ijzendoorn, M. H., Bakermans-Kranenburg, M. J., & van der
Veer, R. (2007). Physiological reactions of preschoolers to fear-inducing film clips: Effects
of temperamental fearfulness and quality of the parent–child relationship. Developmental
Psychobiology, 49(2), 187–195. doi:10.1002/dev.20188
Goldstein, H., Yang, M., Omar, R., Turner, R., & Thompson, S. (2000). Meta-analysis using
multilevel models with an application to the study of class size effects. Journal of the Royal
Statistical Society: Series C: Applied Statistics, 49, 399–412. doi:10.1111/1467-9876.00200

Human Communication Research 42 (2016) 98–121 © 2015 International Communication Association 117
“Scary” TV and Children’s Internalizing L. J. Pearce & A. P. Field

Halligan, S. L., Clark, D. M., & Ehlers, A. (2002). Cognitive processing, memory, and the
development of PTSD symptoms: Two experimental analogue studies. Journal of Behavior
Therapy and Experimental Psychiatry, 33(2), 73–89.
Harris, P. L. (2000). The work of the imagination. Oxford, England: Blackwell.
Harrison, K., & Cantor, J. (1999). Tales from the screen: Enduring fright reactions to scary
media. Media Psychology, 1(2), 97–116. doi:10.1207/s1532785xmep0102_1
Hedges, L. V., & Vevea, J. L. (1998). Fixed- and random-effects models in meta-analysis.
Psychological Methods, 3(4), 486–504.
Hoekstra, S. J., Harris, R. J., & Helmick, A. L. (1999). Autobiographical memories about the
experience of seeing frightening movies in childhood. Media Psychology, 1(2), 117–140.
doi:10.1207/s1532785xmep0102_2
Holmes, E. A., Creswell, C., & O’Connor, T. G. (2007). Posttraumatic stress symptoms in
London school children following September 11, 2001: An exploratory investigation of
peri-traumatic reactions and intrusive imagery. Journal of Behavior Therapy and
Experimental Psychiatry, 38(4), 474–490. doi:10.1016/j.jbtep.2007.10.003
Hunter, J. E., & Schmidt, F. L. (2000). Fixed effects vs. random effects meta-analysis models:
Implications for cumulative research knowledge. International Journal of Selection and
Assessment, 8(4), 275–292.
Hunter, J. E., & Schmidt, F. L. (2004). Methods of meta-analysis: Correcting error and bias in
research findings (Vol., 2nd ed.). Newbury Park, CA: Sage.
Kindt, M., Buck, N., Arntz, A., & Soeter, M. (2007). Perceptual and conceptual processing as
predictors of treatment outcome in PTSD. Journal of Behavior Therapy and Experimental
Psychiatry, 38(4), 491–506. doi:10.1016/j.jbtep.2007.10.002
Konstantopoulos, S. (2011). Fixed effects and variance components estimation in three-level
meta-analysis. Research Synthesis Methods, 2, 61–76. doi:10.1002/jrsm.35
Korhonen, P., & Lahikainen, A. R. (2008). Recent trends in young children’s
television-induced fears in Finland. Journal of Children and Media, 2(2), 147–162.
doi:10.1080/17482790802078664
Lengua, L. J., Long, A. C., Smith, K. I., & Meltzoff, A. N. (2005). Pre-attack symptomatology
and temperament as predictors of children’s responses to the September 11 terrorist
attacks. Journal of Child Psychology and Psychiatry, 46(6), 631–645.
doi:10.1111/j.1469-7610.2004.00378.x
Lewis, S., & Clarke, M. (2001). Forest plots: Trying to see the wood and the trees. British
Medical Journal, 322(7300), 1479–1480. doi:10.1136/bmj.322.7300.1479
Lonigan, C. J., Philips, B. M., Wilson, S. B., & Allan, N. P. (2011). Temperament and anxiety
in children and adolescents. In W. K. Silverman & A. P. Field (Eds.), Anxiety disorders in
children and adolescents: Research, assessment and intervention (Vol., 2nd ed., pp.
198–224). Cambridge, MA: Cambridge University Press.
Mathai, J. (1983). An acute anxiety state in an adolescent precipitated by viewing a horror
movie. Journal of Adolescence, 6(2), 197–200. doi:10.1016/s0140-1971(83)80027-x
McDonough, P. (2009). TV viewing among kids at an eight-year high: Nielsenwire.
McKinnon, A. C., Nixon, R. D. V., & Brewer, N. (2008). The influence of data-driven
processing on perceptions of memory quality and intrusive symptoms in children
following traumatic events. Behaviour Research and Therapy, 46(6), 766–775.
doi:10.1016/j.brat.2008.02.008

118 Human Communication Research 42 (2016) 98–121 © 2015 International Communication Association
L. J. Pearce & A. P. Field “Scary” TV and Children’s Internalizing

Muris, P., & Field, A. P. (2010). The role of verbal threat information in the development of
childhood fear. "Beware the Jabberwock!". Clinical Child and Family Psychology Review,
13(2), 129–150. doi:10.1007/s10567-010-0064-1
Muris, P., & Field, A. P. (2011). The normal development of fear in children and adolescents.
In W. K. Silverman & A. P. Field (Eds.), Anxiety disorders in children and adolescents:
Research, assessment and intervention (Vol., 2nd ed., pp. 76–89). Cambridge, MA:
Cambridge University Press.
Ofcom. (2012a). Children’s TV viewing: BARB analysis.
Ofcom. (2012b). Protecting audiences in a converging world: Deliberative research report.
Ortiz, C. D., Silverman, W. K., Jaccard, J., & La Greca, A. M. (2011). Children’s state anxiety in
reaction to disaster media cues: A preliminary test of a multivariate model. Psychological
Trauma-Theory Research Practice and Policy, 3(2), 157–164. doi:10.1037/a0020098
Otto, M. W., Henin, A., Hirshfeld-Becker, D. R., Pollack, M. H., Biederman, J., & Rosenbaum,
J. F. (2007). Posttraumatic stress disorder symptoms following media exposure to tragic
events: Impact of 9/11 on children at risk for anxiety disorders. Journal of Anxiety
Disorders, 21(7), 888–902.
Paavonen, E. J., Pennonen, M., Roine, M., Valkonen, S., & Lahikainen, A. R. (2006). TV
exposure associated with sleep disturbances in 5-to 6-year-old children. Journal of Sleep
Research, 15(2), 154–161.
Paavonen, E. J., Roine, M., Pennonen, M., & Lahikainen, A. R. (2009). Do parental
co-viewing and discussions mitigate TV-induced fears in young children? Child: Care,
Health and Development, 35(6), 773–780.
Paik, H., & Comstock, G. (1994). The effects of television violence on antisocial-behavior: A
meta-analysis. Communication Research, 21(4), 516–546.
doi:10.1177/009365094021004004
Pfefferbaum, B., Jeon-Slaughter, H., Pfefferbaum, R. L., Houston, J. B., Rainwater, S. M., &
Regens, J. L. (2008). Media coverage and children’s reactions to disaster with implications
for primary care and public health. The Journal of the Oklahoma State Medical
Association, 101(12), 312–317.
Pfefferbaum, B., Nixon, S. J., Krug, R. S., Tivis, R. D., Moore, V. L., Brown, J. M., … Gurwitch,
R. H. (1999). Clinical needs assessment of middle and high school students following the
1995 Oklahoma City bombing. American Journal of Psychiatry, 156(7), 1069–1074.
Pfefferbaum, B., Nixon, S. J., Tivis, R. D., Doughty, D. E., Pynoos, R. S., Gurwitch, R. H., &
Foy, D. W. (2001). Television exposure in children after a terrorist incident.
Psychiatry-Interpersonal and Biological Processes, 64(3), 202–211.
doi:10.1521/psyc.64.3.202.18462
Pfefferbaum, B., Seale, T. W., Brandt, E. N., Jr., Pfefferbaum, R. L., Doughty, D. E., &
Rainwater, S. M. (2003). Media exposure in children one hundred miles from a terrorist
bombing. Annals of Clinical Psychiatry, 15(1), 1–8. doi:10.1023/a:1023293824492
Pfefferbaum, B., Seale, T. W., McDonald, N. B., Brandt, E. N., Rainwater, S. M., Maynard, B.
T., … Miller, P. D. (2000). Posttraumatic stress two years after the Oklahoma City
bombing in youths geographically distant from the explosion. Psychiatry-Interpersonal
and Biological Processes, 63(4), 358–370.
Phillips, D., Prince, S., & Schiebelhut, L. (2004). Elementary school children’s responses 3
months after the September 11 terrorist attacks: A study in Washington, DC. American
Journal of Orthopsychiatry, 74(4), 509–528. doi:10.1037/0002-9432.74.4.509

Human Communication Research 42 (2016) 98–121 © 2015 International Communication Association 119
“Scary” TV and Children’s Internalizing L. J. Pearce & A. P. Field

R Core Team. (2014). R: A language and environment for statistical computing. Vienna,
Austria: R Foundation for Statistical Computing. Retrieved from
http://www.R-project.org
Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical linear models: Applications and data
analysis methods. Thousand Oaks, CA: Sage.
Rosenthal, R. (1991). Meta-analytic procedures for social research (Vol. ,2nd ed.). Newbury
Park, CA: Sage.
Saylor, C. F., Cowart, B. L., Lipovsky, J. A., Jackson, C., & Finch, A. J. (2003). Media exposure
to September 11 – Elementary school students’ experiences and posttraumatic
symptoms. American Behavioral Scientist, 46(12), 1622–1642.
doi:10.1177/0002764203254619
Schredl, M., Anders, A., Hellriegel, S., & Rehm, A. (2008). TV viewing, computer game
playing and nightmares in school children. Dreaming, 18(2), 69–76.
doi:10.1037/1053-0797.18.2.69
Schuster, M. A., Stein, B. D., Jaycox, L. H., Collins, R. L., Marshall, G. N., Elliott, M. N., …
Berry, S. H. (2001). A national survey of stress reactions after the September 11, 2001,
terrorist attacks. New England Journal of Medicine, 345(20), 1507–1512.
Silverman, W. K., & Field, A. P. (Eds.) (2011). Anxiety disorders in children and adolescents:
Research, assessment and intervention (Vol., 2nd ed.). Cambridge, MA: Cambridge
University Press.
Smith, S. L., & Moyer-Gusé, E. (2006). Children and the War on Iraq: Developmental
differences in fear responses to television news coverage. Media Psychology, 8(3),
213–237.
Smith, S. L., & Wilson, B. J. (2002). Children’s comprehension of and fear reactions to
television news. Media Psychology, 4(1), 1–26.
Sparks, G. G., & Cantor, J. (1986). Developmental differences in fright responses to a
television program depicting a character transformation. Journal of Broadcasting &
Electronic Media, 30(3), 309–323.
TV Parental Guidelines. (2012). Understanding the TV ratings and parental controls. Retrieved
from http://www.tvguidelines.org/resources/TV_Parental_Guidelines_Brochure.pdf
Twenge, J. M. (2000). The age of anxiety? Birth cohort change in anxiety and neuroticism,
1952–1993. Journal of Personality and Social Psychology, 79(6), 1007–1021.
Twenge, J. M. (2011). Generational differences in mental health: Are children and adolescents
suffering more, or less? American Journal of Orthopsychiatry, 81(4), 469–472.
doi:10.1111/j.1939-0025.2011.01115.x
U.S. Public Health Service (2000). Report of the surgeon general’s conference on children’s
mental health: A national action agenda. Washington, DC: U.S. Department of Health and
Human Services.
Unz, D., Schwab, F., & Winterhoff-Spurk, P. (2008). TV news – The daily horror? Journal of
Media Psychology: Theories, Methods, and Applications, 20, 141–155.
doi:10.1027/1864-1105.20.4.141
Valkenburg, P. M., Cantor, J., & Peeters, A. L. (2000). Fright reactions to television - A child
survey. Communication Research, 27(1), 82–99.
Valkenburg, P. M., Krcmar, M., Peeters, A. L., & Marseille, N. M. (1999). Developing a scale
to assess three styles of television mediation: “Instructive mediation,” “restrictive
mediation,” and “social coviewing”. Journal of Broadcasting & Electronic Media, 43(1),
52–66. doi:10.1080/08838159909364474

120 Human Communication Research 42 (2016) 98–121 © 2015 International Communication Association
L. J. Pearce & A. P. Field “Scary” TV and Children’s Internalizing

van der Molen, J. H. W., & Bushman, B. J. (2008). Children’s direct fright and worry reactions
to violence in fiction and news television programs. Journal of Pediatrics, 153(3),
420–424. doi:10.1016/j.jpeds.2008.03.036
Vevea, J. L., & Woods, C. M. (2005). Publication bias in research synthesis: Sensitivity analysis
using a priori weight functions. Psychological Methods, 10(4), 428–443.
Viechtbauer, W. (2010). Conducting meta-analyses in R with the metafor Package. Journal of
Statistical Software, 36(3), 1–48.
Watson, D. (2005). Rethinking the mood and anxiety disorders: A quantitative hierarchical
model for DSM-V. Journal of Abnormal Psychology, 114(4), 522–536.
doi:10.1037/0021-843x.114.4.522
Weems, C. F., Scott, B. G., Banks, D. M., & Graham, R. A. (2012). Is TV traumatic for all
youths? The role of preexisting posttraumatic-stress symptoms in the link between
disaster coverage and stress. Psychological Science, 23(11), 1293–1297.
doi:10.1177/0956797612446952
Wilson, B. J., Martins, N., & Marske, A. L. (2005). Children’s and parents’ fright reactions to
kidnapping stories in the news. Communication Monographs, 72(1), 46–70.
doi:10.1080/0363775052000342526
Wood, W., Wong, F. Y., & Chachere, J. G. (1991). Effects of media violence on viewers
aggression in unconstrained social-interaction. Psychological Bulletin, 109(3), 371–383.
doi:10.1037/0033-2909.109.3.371

Human Communication Research 42 (2016) 98–121 © 2015 International Communication Association 121