- Access to this full-text is provided by Springer Nature.
- Learn more
Download available
Content available from International Journal of Behavioral Nutrition and Physical Activity
This content is subject to copyright. Terms and conditions apply.
R E S E A R C H Open Access
Type of screen time moderates effects on
outcomes in 4013 children: evidence from
the Longitudinal Study of Australian
Children
Taren Sanders
1*
, Philip D. Parker
1
, Borja del Pozo-Cruz
1
, Michael Noetel
1,2
and Chris Lonsdale
1
Abstract
Background: Excessive engagement with digital screens is harmful to children’s health. However, new evidence
suggests that exposure at moderate levels may not be harmful and may even provide benefit. Therefore, our
objective was to determine if there are curvilinear relationships between different types of screen time and a
diverse set of outcomes, including health and education.
Methods: We address our objective using a repeated measures design. Children (N= 4013), initially aged 10–11
were assessed every 2 years between 2010 and 2014. Children’s screen time behavior was measured using time-use
diaries, and categorized into five types: social, passive, interactive, educational, or other. We used measures of
children’s physical health, health-related quality of life, socio-emotional outcomes, and school achievement. The
analysis plan was pre-registered. Models were adjusted for gender, socio-economic status, ethnicity, number of
siblings, and housing factors.
Results: There were linear associations between total screen time and all outcomes, such that more screen time
was associated with worse outcomes. However, there was variability when examined by screen time type. Passive
screen time (e.g., TV) was associated with worse outcomes, educational screen time (e.g., computer for homework)
was associated with positive educational outcomes and had no negative relations with other outcomes. Interactive
screen time (e.g., video games) had positive associations with educational outcomes but negative associations with
other outcomes. In all instances, these significant associations were small or very small, with standardised effects <
0.07. We found little evidence of curvilinear relationships.
Conclusions: The small effects of screen time on children’s outcomes appear to be moderated by the type of
screen time. Policy makers, educators, and parents should consider the type of screen time when considering the
benefits and harms of use.
Keywords: Screen time, Children, Health, Education
© The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and
reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to
the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver
(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
* Correspondence: Taren.Sanders@acu.edu.au
1
Institute for Positive Psychology and Education, Australian Catholic
University, North Sydney, NSW, Australia
Full list of author information is available at the end of the article
Sanders et al. International Journal of Behavioral Nutrition and Physical Activity
(2019) 16:117
https://doi.org/10.1186/s12966-019-0881-7
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Background
High levels of engagement with digital screens (i.e.,
‘screen time’) are harmful to children’s physical health
[1]. A body of evidence underpins guidelines that recom-
mend limiting children’s screen time exposure [2,3]. For
example, a recent review found that screen time is dele-
teriously associated with adiposity and cardiorespiratory
fitness [1]. There is also evidence that screen time is as-
sociated with negative psychological and educational
outcomes, such as greater depression [4] and lower aca-
demic achievement [5], respectively. As a result, guide-
lines [3,6] advise that lower levels of screen time are
associated with benefits for children. In our study, we
refer to this as the less-is-better hypothesis.
Evidence that moderate levels of screen time may have
benefits over abstinence or high use contradicts current
guidelines. For example, a review of the literacy develop-
ment literature revealed studies in which moderate
amounts of television was associated with better reading
than low or high amounts of viewing [7]. Curvilinear
relations have also been found with psychosocial out-
comes. For example, in an investigation of more than
120,000 adolescents, Przybylski and Weinstein found
moderate amounts of electronic screen time were associ-
ated with higher mental well-being compared to low or
high levels [8]. Similar curvilinear relationships for screen
time have also emerged in other studies related to chil-
dren’s health and well-being [9–12]. Some researchers
have labeled this the Goldilocks hypothesis [8].
When examining the Goldilocks screen hypothesis, pre-
vious studies have tended to focus on a single outcome, or
a narrow range of variables. For example, Przybylski and
Weinstein (2017) centred their investigation on screen
time’s association with adolescents’well-being, and did
not investigate other important outcomes, such as physical
health or educational achievement. It is possible that the
less-is-better hypothesis and the Goldilocks hypothesis
apply differentially to outcomes. For example, engaging
with moderate amounts of social media may benefit social
functioning, while high levels might displace face-to-face
contact, leading to poorer mental health (i.e., supporting
the Goldilocks hypothesis) [13]. In contrast, passive screen
time (e.g., television) would be unlikely to convey any
form of physical health benefit, and thus lower levels
would be expected to provide health benefits (i.e., support-
ing the less-is-better hypothesis). Studies that examine a
limited range of outcome variables [8–12]cannotexamine
this possibility.
In the current study, we aimed to investigate these two
competing hypotheses across different types of screen
time and different outcomes, including physical health,
psychological outcomes, and educational outcomes. We
further extended the Przybylski and Weinstein (2017)
study of adolescents by examining these hypotheses in a
large sample of children, and by examining if these rela-
tionships are stable as children age. As this is a concep-
tual replication of Przybylski and Weinstein’s work, we
also examine differences by weekday and weekend.
Research Questions
1. Are there linear or curvilinear relations between
screen time and children’s physical health,
psychological outcomes, and educational outcomes?
And, if curvilinear relations exist, at what duration
of screen exposure do they become negative?
2. Are these relationships modified by age, screen time
type (e.g., device or content), and weekday vs.
weekend usage? If so, do these factors shift the
turning point?
Method
Study design and sample
Data were drawn from Growing Up in Australia: The
Longitudinal Study of Australian Children (LSAC), a
population-based study which tracks two cohorts of chil-
dren aged 0/1 years (B-cohort) and 4/5 years (K-cohort)
every 2 years beginning in 2004. We used data from
Waves 4–6 of the K-cohort (2010–2014; ages 10–15).
The overall response rate was 62% in the K-cohort (N=
4013) at baseline, with Wave 6 retention rates of 82%.
Other waves of the K-cohort could not be included be-
cause of significant changes in the design of the time-
use diary instrument used as our exposure measure [14].
We excluded the B-cohort because of the limited avail-
ability of time-use data during the ages of interest.
Further details on the LSAC methodology, including
sampling procedures, are available elsewhere [15].
Exposure variables
Screen time
Time spent engaging with screens was measured using
time-use diaries administered to the child. Children
recorded the activities they participated in during one
randomly allocated day on a paper diary. During a face-
to-face interview on the day following the diary, an
interviewer added additional contextual information
(e.g., where they were and who they were with). Partici-
pants nominated the primary activity they engaged in,
and the time of the activity (the activity ‘window’). The
child could also nominate additional secondary behav-
iors that occurred in parallel during the activity window.
The interviewer applied a coding framework to the chil-
dren’s activities to make the diaries comparable across
children [14]. We divided diary activities which involved
screens into five categories: social screen time (e.g., so-
cial media), passive screen time (e.g., television), inter-
active screen time (e.g., video games), educational screen
Sanders et al. International Journal of Behavioral Nutrition and Physical Activity (2019) 16:117 Page 2 of 10
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
time (e.g., computer use for homework), and other
screen time where the activity did not fit into any of the
categories.
To process the time-use diaries, we calculated the total
length of activities which represented screen time re-
gardless of whether they were primary or secondary
activities. To calculate total screen time, we added all ac-
tivities windows where any of the activities included
screen time. For example, if a child spent 15 min texting
(primary) while also watching TV (secondary), then we
calculated 15 min of both ‘social screen time’and ‘pas-
sive screen time’, but only 15 min of ‘total screen time’
to avoid double-counting. Thus, it should be noted that
the individual categories of screen time variables will not
sum to total screen time. A list of items coded as screen
time is provided in Additional file 1: Table S1. Time-use
diaries have been successfully used in previous studies
investigating health behaviours in children [16–20].
Physical health
Physical outcomes
Weight was measured to the nearest 50 g using glass
bathroom scales (Salter Australia, Springvale, VIC,
Australia; Code 79985) while children were in light
clothing. Height was measured twice, without shoes, to
the nearest 0.1 cm using a stadiometer (Invicta, Leices-
ter, UK; Code IPO955). Waist circumference was also
assessed twice to the nearest 0.1 cm. Body mass index
(BMI) was then calculated as kg/m
2
. The child’s BMI z-
score for age was calculated based on Centre for Disease
Control growth charts [21,22]. All anthropometric mea-
sures were taken by the trained interviewer.
Global health
Parents were asked to report on their perception of their
child’s overall health in a scale ranging from “poor”to “ex-
cellent”[23]. This scale has been previously validated for
Australian children [24]. Because there were fewer than
20 children with “poor”or “fair”health, global health was
dichotomized to “excellent”and “less than excellent”.
Psychological outcomes
Social and emotional functioning
Children’s socio-emotional outcomes were assessed using
the Strengths and Difficulties Questionnaire (SDQ), a vali-
dated, 25-item, parent-reported questionnaire [25]. We
used all five subscales (conduct problems, emotional prob-
lems, hyperactivity, peer problems, and prosocial behavior;
range: 0–10).
Children’s quality of life was assessed via the Paediatric
Quality of Life Inventory (PedsQL), a validated 23-item
parent-reported instrument [26]. We computed two sub-
scale scores (social and emotional functioning), which
ranged from 0 to 100. We chose not to include the
physical functioning subscale as the items were unlikely
to be related to screen time. A higher PedsQL score rep-
resents better quality of life. Parents were the respon-
dents for both the SDQ and PedsQL.
Temperament profile
Children’s temperament was assessed with the School-
Age Temperament Inventory, a 38-item parent-reported
questionnaire with four dimensions: negative reactivity
(intensity and frequency of negative affect), task persist-
ence (the self-direction that a child exhibits in fulfilling
tasks), approach/withdrawal (response to new people and
situations), and activity (moves quickly to get where he/
she wants to go) [27]. In the context of this study, only
negative reactivity and task persistence were included be-
cause of their plausibility as outcomes of screen time.
Higher scores indicate that the child is higher in negative
reactivity and task persistence.
Educational outcomes
School achievement
Estimates of both numeracy and literacy ability were taken
from government administration records of the National
Assessment Program - Literacy and Numeracy (NAPLAN,
https://www.nap.edu.au/naplan). The NAPLAN data are
linked to child data by the LSAC organisers via a unique
identifier. The NAPLAN tests are given to all eligible chil-
dren in Australia in Grades 3 (age 8), 5 (age 10), 7 (age
12), and 9 (age 14). We used scores from Grades 5–9. The
tests are scaled so they are comparable across age cohort
and across grade. Scores have an overall mean of 500 and
a standard deviation of 100. Numeracy was measured
using a single test and literacy was measured using four
tests covering reading, writing, spelling, and grammar. We
conducted principal component analysis on the the four
literacy scores and formed a single factor score to repre-
sent literacy.
Adjustment variables
To provide an all-else-being-equal estimate of the effect of
screen time, we adjusted results for: child gender, Indigen-
ous status, language-other-than-English status, child’s
country of birth (Australia vs. elsewhere), and a composite
measure of family socioeconomic status provided by the
LSAC organizers [28], which is calculated using parent’s
occupational prestige, income and education. We also
used a measure of the average socioeconomic status of the
child’spostcode[29]. To adjust for opportunity to engage
in activities other than screen time, we further adjusted
for home type (detached house vs other), number of sib-
lings of the study child, and a parent-reported index of
neighborhood livability (including parks and safety), as
neighborhood factors have previously been linked to
screen time [17].
Sanders et al. International Journal of Behavioral Nutrition and Physical Activity (2019) 16:117 Page 3 of 10
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Analysis
To minimise potential bias we pre-registered our ana-
lysis plan prior to commencing the study, including
specifying what analyses would be included and our
criteria for including variables in the analysis [30]. Any
deviations from the pre-registered plan are noted below.
Analysis was based on Przybylski and Weinstein’s[8]
study of screen time and well-being that provided sup-
port for the Goldilocks hypothesis. We fitted screen time
as both linear and quadratic effects. If the quadratic ef-
fect was significant, we calculated the turning point (i.e.,
the point at which more screen time moved from having
a beneficial to negative influence) using the equation:
^
xmax ¼
−βscreen:time
2βscreen:time:quadratic . We also calculated the point at
which increases in screen time led to poorer outcomes
than no screen time calculated as twice the turning
point, which we refer to as the ‘zero point’.
The LSAC data comes from a complex sampling de-
sign with postcode as the primary sampling unit. In
addition, we combined data from different waves, mean-
ing that each participant had multiple waves of data. To
account for these factors, we used multilevel models
with observations nested within individuals and individ-
uals nested within postcodes. Our repeated measures
design takes advantage of the multiple waves of data, but
we do not test for longitudinal associations. We accounted
for attrition by using all available information for each
participant and using sample attrition weights provided by
the survey organizers to ensure that the data remained
representative of the population at each wave. We handled
unit non-response missing data using multiple imputa-
tions, combining effects across 10 imputations [31]. We
reverse coded variables such that increases could be con-
sistently interpreted as improvements in these outcomes.
We tested unadjusted models and adjusted models. As
we were interested in whether the effects differed by age
or weekday versus weekend, all models included terms
for age and weekday/weekend. Note that our pre-
registered analysis plan [30] mistakenly included gender
as both an interaction term and a control variable, and
we chose to only include it as a control variable.
Results
Participants
Our analysis included 4013 children in the LSAC study.
Of those analyzed, 51.2% were male, 96.1% were non-
indigenous, 85.7% spoke English as their primary lan-
guage, and 95.9% were born in Australia. Most children
lived in a detached house (88.2%), and the study children
had a mean of 1.7 siblings (SD = 1.2). At age 10 there
were 4013 participants. This declined to 3682 by age 12
and 3276 by age 14. There was a notable increase in
children’s educational and social screen time between
the ages of 12 and 14. This increase may be due to the
participants transitioning from primary to secondary
schooling. Australian children typically begin high
school at age 13, and this transition may increase their
autonomy or change the amount of technology they use
at school. Further description of the sample is found in
Table 1. For unit non-response, the most missing data
was for the time-use diaries (21%) and NAPLAN scores
(16%). All other variables had less than 5% missing data
(see Additional file 4: Figure S1).
Preliminary analysis
Initial analyses showed that the screen time variables
were positively skewed, especially for the less popular
screen time types (e.g., social screen time) where there
were high numbers of participants with zero screen time
(Fig. 1). Therefore, we log transformed the screen time
variables for imputation and translated back to the ori-
ginal scale for analysis. Despite evidence of skew in both
exposure and some outcome variables, assumption
checking revealed few problems in the models.
As per our pre-registered protocol [30], we checked
that outcomes were independent using zero-order corre-
lations and planned to remove variables if any were cor-
related above r= 0.70. The literacy and numeracy
outcomes were correlated at r= 0.71 and we therefore
created a composite score that was the unweighted mean
of the first principal component of each of the two
scores. We refer to this composite score as ‘school
achievement’. We noted that PedsQL emotional subscale
and SDQ emotional subscale were correlated at r=−
0.67 and BMI and waist circumference at r= 0.69 (see
Additional file 2: Table S2 for other correlations). While
these were below our a priori cutpoint, to minimize
spurious associations we chose to keep only the variables
with the least missing data (SDQ emotional subscale and
waist circumference). We also checked for missing data
prior to imputation, and planned to remove variables
where missing data was > 60% [30]. No variables were
removed on this basis.
Linear effects
To examine the less-is-better hypothesis, we first exam-
ined linear models without quadratic terms. These re-
sults are presented in Fig. 2as adjusted linear effects,
standardised for each outcome (β). These effects sizes
are typically interpreted as small effect: β= 0.1; medium
effect: β= 0.3; large effect: β= 0.5. All linear results were
β< 0.07; that is, very small in size.
Total screen time was associated with worse educa-
tional outcomes, but this result was fully attenuated in
adjusted models (see Additional file 3: Table S3). In both
unadjusted and adjusted models, total screen time was
linearly associated with unfavorable temperament
Sanders et al. International Journal of Behavioral Nutrition and Physical Activity (2019) 16:117 Page 4 of 10
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Fig. 1 Density Plots for Components of Total Screen Time
Table 1 Sample descriptive statistics
Wave 4 (Age 10–11) Wave 5 (Age 12–13) Wave 6 (Age 14–15)
Total Screen Time (min/day) 76.6 (35.88) 76.77 (35.85) 97.44 (36.19)
Social Screen Time (min/day) 0.59 (4.03) 0.56 (3.97) 22.3 (34.23)
Educational Screen Time (min/day) 2.56 (10.3) 2.61 (10.36) 11.18 (23.79)
Passive Screen Time (min/day) 61.37 (35.36) 61.35 (35.3) 65.27 (40.76)
Interactive Screen Time (min/day) 26.43 (32.66) 26.8 (32.84) 28.7 (40.19)
Other Screen Time (min/day) 1.24 (6.39) 1.25 (6.41) 12.11 (25.74)
Emotional Functioning (PedsQL) 74.01 (16.49) 75.64 (17.1) 74.97 (18.23)
Social Functioning (PedsQL) 80.24 (18.58) 82.61 (17.86) 80.54 (18.32)
Prosocial Behaviour (SDQ) 8.5 (1.66) 8.3 (1.74) 8.06 (1.86)
Peer Relationship Problems (SDQ) 1.5 (1.71) 1.43 (1.65) 1.54 (1.67)
Emotional Symptoms (SDQ) 1.9 (1.95) 1.94 (1.95) 1.89 (1.98)
Hyperactivity (SDQ) 3.16 (2.36) 2.93 (2.32) 2.66 (2.23)
Conduct Problems (SDQ) 1.33 (1.48) 1.06 (1.42) 0.97 (1.41)
Reactivity (SATI) 2.3 (0.81) 2.38 (0.78) 2.32 (0.81)
Task Persistence 3.5 (0.89) 3.57 (0.85) 3.65 (0.86)
Body Mass Index z-score 0.37 (1.03) 0.35 (1.03) 0.37 (1.13)
Waist Circumference (cm) 66.76 (9.4) 71.89 (10.26) 75.14 (10.19)
Global Health 0.44 (0.5) 0.46 (0.5) 0.48 (0.5)
Literacy 0.02 (1) 0.05 (0.99) 0.05 (0.97)
Numeracy 0.01 (1) 0.04 (1) 0.05 (0.99)
All data presented as means (standard deviations). PedsQL Pediatric Quality of Life Inventory, SDQ Strengths and Difficulties Questionnaire, SATI School-Age
Temperament Inventory
Sanders et al. International Journal of Behavioral Nutrition and Physical Activity (2019) 16:117 Page 5 of 10
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
outcomes, worse socio-emotional outcomes, lower
health-related quality of life, and poorer health out-
comes. However, there was substantial variability when
results were examined by screen time type.
Educational screen time (e.g., homework on electronic
devices) showed the most benefits in unadjusted and ad-
justed models, with positive effects on children’s persist-
ence and educational outcomes, and no significant
Fig. 2 Adjusted Standardised Linear Effects for Each Combination of Screen Time Exposure and Outcome
Sanders et al. International Journal of Behavioral Nutrition and Physical Activity (2019) 16:117 Page 6 of 10
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
effects on psychological or health outcomes. Interactive
screen time (e.g., video games) showed similar trends as
total screen time. However, unlike total screen time,
interactive screen time was associated with positive edu-
cational outcomes. Passive screen time (e.g., TV) was
associated with worse psychological outcomes, poorer
health outcomes, and lower educational outcomes in un-
adjusted and adjusted models. Poorer prosocial behavior
and lower persistence were also associated with higher
passive screen time, but only in unadjusted models. So-
cial screen time was linearly associated with poorer
health-related quality of life, higher reactivity, and worse
socio-emotional outcomes for the prosocial, emotional,
and conduct subscales of the SDQ, with no influence on
the peer or hyperactivity subscales of the SDQ, nor chil-
dren’s persistence, health, or educational outcomes. Fi-
nally, other screen time showed minimal associations
with outcomes, with negative effects on only the pro-
social and peer SDQ subscales.
Quadratic effects
We next examined if any relationships were better rep-
resented by a quadratic function. In unadjusted models,
there were non-linear relationships between total screen
time and the hyperactivity SDQ subscale, social screen
time and the peer SDQ subscale, interactive screen time
and the hyperactivity and prosocial SDQ subscales and
persistence, educational screen time and persistence, and
other screen time and persistence. After adjustment for
covariates, only the total screen time and hyperactivity
SDQ subscale (β
Linear
= 0.028 [0.013–0.043]; β
Quadratic
=
−0.001 [0.002 –−0.000]; turning point: 12.29 [6.44–
18.14] hours; zero point: 24.59 [12.90–36.28] hours), and
the social screen time and peer SDQ subscale (β
Linear
=
−0.096 [−0.159–0.034]; β
Quadratic
= 0.011 [0.003–0.019];
turning point: 4.48 [3.42–5.53] hours, zero point: 8.96
[6.85–11.06] hours) quadratic associations remained sig-
nificant (Additional file 7Table S3). We note that, owing
to the very small quadratic effect, the zero point for the
total screen time and hyperactivity SDQ association is
outside the range of plausible values. Scatterplots of all
associations and the quadratic results are available in
Additional file 5: Figure S2.
Interactions with age and weekday
To determine the extent to which these relationships
changed as the children aged, we tested an interaction
between screen time and sample wave (as an indicator
for age). There were very few significant interactions (6
of 132 for the linear effects and 3 of 132 for the quad-
ratic effects in the adjusted models with p< .05), sug-
gesting that these associations are stable between the
ages of 10 and 15. All interaction results are available in
Additional file 7: Table S3.
For the linear models, all six interactions related to
age. Three interactions were present for waist circumfer-
ence, and one each for prosociality, social PedsQL, and
reactivity. All indicated that increased screen time had a
more detrimental association with these outcomes at
ages 10 and 12 than at age 14. No significant linear in-
teractions were found for weekday vs weekend.
All three of the significant interactions for the adjusted
quadratic relationships related to weekend vs weekday. The
interactions were present for a) conduct problems, b) emo-
tional problems, and c) reactivity as predicted by interactive
screen time. All significant quadratic interactions indicated
a Goldilocks effect for weekends, with turning points at ap-
proximately two to 3 h (see Additional file 6:FiguresS3
and Additional file 7: Figure S4), and no quadratic effects
on weekdays. No significant quadratic interactions were
found for age.
Discussion
In this study, we compared competing hypotheses for
screen time effects on children’s physical health, psycho-
logical outcomes, and educational outcomes. We found
evidence that screen time was associated with children’s
physical health, health-related quality of life, socio-
emotional outcomes, and school achievement, with sub-
stantial variation based on the type of screen time. In
moderation analyses, these results appeared to remain
stable for screen time on weekdays versus weekends.
While there were some significant interactions, none
were meaningful in terms of practical significance. There
was little evidence to support the Goldilocks hypothesis
in our data. Instead, our findings lend qualified support
to the less-is-better hypothesis–qualified because educa-
tional screen time was associated with positive educa-
tional outcomes and higher persistence, with no negative
consequences for other outcomes. Educational screen
time, therefore, appears beneficial and would not fit the
less-is-better or Goldilocks hypotheses. However, the
magnitude of the effects observed in our study were con-
sistently very small, with almost all less than 0.05 of a
standard deviation per hour of additional screen time.
This finding is consistent with meta-analytic results,
where effect sizes for physical health [32] and socio-
emotional and behavioral outcomes [33] have been small
[34]. Yet, screen time has become a major concern that
parents have about their children’s health [35]. Our re-
sults suggest that detrimental effects may be domain-
specific and, as such, some of the concern around screen
time may be unjustified.
Our results also demonstrate a need for future guide-
lines to embrace the complexity of screen time. We
found that interactive screen time can be simultaneously
harmful and beneficial, in that it negatively impacts most
outcomes but is positively associated with educational
Sanders et al. International Journal of Behavioral Nutrition and Physical Activity (2019) 16:117 Page 7 of 10
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
outcomes. Most current guidelines [2,3] focus on redu-
cing harm and largely ignore the potential benefits some
types of screen time can provide. Future evidence-based
guidelines should focus on providing parents, and pro-
fessionals who advise parents and children (e.g., doctors,
teachers), with information that allows them to balance
the risks and benefits of screen time. It is likely useful
for parents to know that duration is not the only screen
exposure variable to consider –content also matters.
For example, our analyses show there are unlikely to be
negative educational consequences, and there may even
be some small benefits, when children engage in educa-
tional types of screen time such as using a computer for
homework.
Our findings are in contrast to previous research that
found non-linear relationships between screen time and
mental well-being [8], socio-emotional outcomes [10], sleep
[11], and other health outcomes [12]. One explanation is
differences in sample sizes. For instance, Przybylski and
Weinstein [8] investigated associations between different
types of screen time and mental well-being in 120,000 ado-
lescents. They found significant results, with standardized
effect sizes for the quadratic terms between 0.03 and 0.13.
It’s possible that even though our data included more than
10,000 data points, it was insufficient to detect these weak
effects. If this is the case, we would question the clinical
significance of such small effects.
Strengths and limitations
We used a nationally-representative, longitudinal data-
set, which provided time-use diary estimates of behavior,
as opposed to simple recall questions. We preregistered
our analysis plan prior to analyzing the data, and used
methods to address the complex survey method and
missing data. Finally, we examined a broad range of
screen time exposures, including educational, interactive
and passive forms of screen time. We also examined di-
verse outcomes, including physical health, psychological,
and educational variables. In addition, we conditioned
on a much broader range of potential covariates than
previous research.
Despite these strengths, our study has several import-
ant limitations. As with the vast majority of screen time
research [36], our study relied on subjectively-reported
screen time. Currently, there are limited options for ob-
jectively measuring screen exposure. More precise meas-
urement devices (e.g., wearable cameras) may yield more
accurate determinations not only of screen exposure
duration, but also the specific content being viewed.
These measurement improvements may have less noise,
and provide a clearer indication of the effects [37]. Des-
pite using longitudinal data we would be reluctant to
draw causal conclusions. The data used covers the
period 2010–2014 and it is plausible that screen time
behaviour has changed since these data were collected.
As such, it is possible that the results presented here are
not generalizable to contemporary children of the target
ages. While we adjusted for important confounders there
is still a risk of unmeasured variable bias influencing the
findings (e.g., parenting style or companion) and we can-
not rule out the possibility of reverse causation.
Conclusion
Previous studies suggested that, compared with very low
or very high amounts of screen time, moderate amounts
of screen media use may benefit children’s mental well-
being. Our findings contradict that research, with little
support for the Goldilocks hypothesis across a wide
range of physical health, psychological and educational
outcomes. Indeed, we observed only very small effect
sizes on the outcomes we measured and across the dif-
ferent types of screen time. We observed that what small
effects do exist seem to be moderated by the type of
screen time, with passive screen time (e.g., TV) having
mostly detrimental effects, while educational screen time
could confer slight benefits in school achievement and
persistence. These results suggest that policymakers,
professionals, and parents should consider the type of
children’s screen time rather than just duration. How-
ever, our overall findings indicate that the high levels of
concern about their children’s screen time exhibited by
parents may be unwarranted.
Supplementary information
Supplementary information accompanies this paper at https://doi.org/10.
1186/s12966-019-0881-7.
Additional file 1: Table S1. Time Use Diary Categories.
Additional file 2: Table S2. Correlations Matrix.
Additional file 3: Table S3. Full Results.
Additional file 4: Figure S1. Missing Data.
Additional file 5: Figure S2. Scatterplots.
Additional file 6: Figure S3. Interactions and Quadratics.
Additional file 7: Figure S4. Example plot of unadjusted linear and
quadratic effects: social quality of life predicted by social screen time with
original scale (left) and modified scale (right).
Abbreviations
BMI: Body Mass Index; LSAC: Longitudinal Study of Australian Children;
NAPLAN: National Assessment Program - Literacy and Numeracy;
PedsQL: Paediatric Quality of Life Inventory; SATI: School-Age Temperament
Inventory; SDQ: Strengths and Difficulties Questionnaire; TV: Television
Acknowledgements
This paper uses unit record data from Growing Up in Australia: the
Longitudinal Study of Australian Children. The study is conducted in
partnership between the Department of Social Services (DSS), the Australian
Institute of Family Studies (AIFS) and the Australian Bureau of Statistics (ABS).
The findings and views reported in this paper are those of the authors and
should not be attributed to DSS, AIFS or the ABS.
Sanders et al. International Journal of Behavioral Nutrition and Physical Activity (2019) 16:117 Page 8 of 10
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Authors’contributions
TS conceptualized the study, processed the time-use diaries, drafted parts of
the initial manuscript, and reviewed and revised the manuscript. PP and BdP-
C conceptualized the study, analyzed the data, drafted parts of the initial
manuscript, and reviewed and revised the manuscript. MN conceptualized
the study, produced the figures, drafted parts of the initial manuscript, and
reviewed and revised the manuscript. CL conceptualized the study, drafted
parts of the initial manuscript, and reviewed and revised the manuscript. All
authors approved the final manuscript as submitted and agree to be
accountable for all aspects of the work.
Funding
No funding was provided for this research. Publication fees were provided
by an Australian Catholic University Faculty of Health Sciences Open Access
Publishing Support Grant.
Availability of data and materials
The LSAC dataset is available from the National Centre for Longitudinal Data
(see https://growingupinaustralia.gov.au). The authors do not have
permission to share this data without endorsement from the Australian
Institute of Family Studies. Materials for this study, including analysis files and
preregistered analysis plans, are available through the Open Science
Framework (https://osf.io/bhzk8/).
Ethics approval and consent to participate
The Australian Institute of Family Studies Ethics Committee provided ethics
approval for the LSAC, and all participants provided written informed
consent.
Consent for publication
Not applicable –no individual data presented.
Competing interests
The authors declare that they have no competing interests.
Author details
1
Institute for Positive Psychology and Education, Australian Catholic
University, North Sydney, NSW, Australia.
2
School of Behavioural and Health
Sciences, Australian Catholic University, Brisbane, QLD, Australia.
Received: 13 May 2019 Accepted: 12 November 2019
References
1. Saunders TJ, Vallance JK. Screen time and health indicators among children
and youth: Current evidence, limitations and future directions. Appl Health
Econ Health Policy. 2017;15:323–31. Available from: https://doi.org/10.1007/
s40258-016-0289-3.
2. Council on Communications and Media. Media use in school-aged children
and adolescents. Pediatrics. 2016;138. Available from: https://doi.org/10.
1542/peds.2016-2592.
3. Department of Health. Australia’s Physical Activity and Sedentary Behaviour
Guidelines: Department of Health; 2014. Available from: http://www.health.
gov.au/internet/main/publishing.nsf/content/health-pubhlth-strateg-phys-
act-guidelines Accessed 10 May 2019
4. Liu M, Wu L, Yao S. Dose-response association of screen time-based
sedentary behaviour in children and adolescents and depression: a meta-
analysis of observational studies. Br J Sports Med. 2016;50:1252–8. Available
from: https://doi.org/10.1136/bjsports-2015-095084.
5. Tremblay MS, LeBlanc AG, Kho ME, Saunders TJ, Larouche R, Colley RC, et al.
Systematic review of sedentary behaviour and health indicators in school-
aged children and youth. Int J Behav Nutr Phys Act. 2011;8:98. Available
from: https://doi.org/10.1186/1479-5868-8-98.
6. Tremblay MS, LeBlanc AG, Janssen I, Kho ME, Hicks A, Murumets K, et al.
Canadian sedentary behaviour guidelines for children and youth. Appl
Physiol Nutr Metab. 2011;36:59–64; 65–71. Available from: https://doi.org/10.
1139/H11-012.
7. Moses AM. Impacts of television viewing on young children’s literacy
development in the USA: A review of the literature. J Early Child Lit. 2008;8:
67–102. Available from: https://doi.org/10.1177/1468798407087162.
8. Przybylski AK, Weinstein N. A large-scale test of the Goldilocks
hypothesis: Quantifying the relations between digital-screen use and
the mental well-being of adolescents. Psychol Sci. 2017;28:204–15.
SAGE Publications Inc. Available from: https://doi.org/10.1177/
0956797616678438.
9. Przybylski AK, Weinstein N. Digital screen time limits and young
children’s psychological well-being: Evidence from a population-based
study. Child Dev. 2017; Available from: https://doi.org/10.1111/cdev.
13007.
10. Parkes A, Sweeting H, Wight D, Henderson M. Do television and
electronic games predict children’s psychosocial adjustment?
Longitudinal research using the UK Millennium Cohort Study. Arch
Dis Child. 2013;98:341–8. Available from: https://doi.org/10.1136/
archdischild-2011-301508.
11. DoYK,ShinE,BautistaMA,FooK. The associations between self-
reported sleep duration and adolescent health outcomes: what is the
role of time spent on Internet use? Sleep Med. 2013;14:195–200.
Available from: https://doi.org/10.1016/j.sleep.2012.09.004.
12. Kim JY. The nonlinear association between Internet using time for
non-educational purposes and adolescent health. J Prev Med Public
Health. 2012;45:37–46. Available from: https://doi.org/10.3961/jpmph.
2012.45.1.37.
13. Song H, Zmyslinski-Seelig A, Kim J, Drent A, Victor A, Omori K, et al.
Does Facebook make you lonely? A meta analysis. Comput Human
Behav. 2014;36:446–52.
14. Mullan K. Longitudinal analysis of LSAC time diary data:
considerations for data users. Australian Institute of Family Studies;
2014. Available from: https://growingupinaustralia.gov.au/sites/default/
files/tp11.pdf Accessed 10 May 2019.
15. Soloff C, Lawrence D, Johnstone R. Sample design (LSAC technical paper
no. 1). Australian Institute of Family Studies, editor. Melbourne: Australian
Institute of Family Studies; 2005. p. 1–30. Available from: https://
growingupinaustralia.gov.au/sites/default/files/tp1.pdf.
16. del Pozo-Cruz B, Perales F, Parker P, Lonsdale C, Noetel M, Hesketh KD, et al.
Joint physical-activity/screen-time trajectories during early childhood: socio-
demographic predictors and consequences on health-related quality-of-life
and socio-emotional outcomes. Int J Behav Nutr Phys Act. 2019;16:55.
BioMed Central. Available from: https://doi.org/10.1186/s12966-019-0816-3.
17. Sanders T, Feng X, Fahey PP, Lonsdale C, Astell-Burt T. The influence of
neighbourhood green space on children’sphysicalactivityandscreen
time: Findings from the longitudinal study of Australian children. Int J
Behav Nutr Phys Act. 2015; Available from: https://doi.org/10.1186/
s12966-015-0288-z.
18. Price AMH, Brown JE, Bittman M, Wake M, Quach J, Hiscock H. Children’s
sleep patterns from 0 to 9 years: Australian population longitudinal study.
Arch Dis Child. 2014;99:119–25. Available from: https://doi.org/10.1136/
archdischild-2013-304150.
19. Price AMH, Quach J, Wake M, Bittman M, Hiscock H. Cross-sectional sleep
thresholds for optimal health and well-being in Australian 4–9-year-olds.
Sleep Med. 2016;22:83–90.
20. Quach J, Price AMH, Bittman M, Hiscock H. Sleep timing and child and
parent outcomes in Australian 4–9-year-olds: a cross-sectional and
longitudinal study. Sleep Med. 2016;22:39–46.
21. Cole TJ, Bellizzi MC, Flegal KM, Dietz WH. Establishing a standard definition
for child overweight and obesity worldwide: International survey. BMJ. 2000;
320:1–6. Available from: https://doi.org/10.1136/bmj.320.7244.1240.
22. Cole TJ, Flegal KM, Nicholls D, Jackson AA. Body mass index cut offs to
define thinness in children and adolescents: International survey. BMJ. 2007;
335:194. Available from: https://doi.org/10.1136/bmj.39238.399444.55.
23. Ware JE Jr, Sherbourne CD. The MOS 36-item short-form health survey (SF-
36). I. Conceptual framework and item selection. Med Care. 1992;30:473–83.
24. Waters E, Salmon L, Wake M. The parent-form Child Health
Questionnaire in Australia: comparison of reliability, validity, structure,
and norms. J Pediatr Psychol. 2000;25:381–91. Available from: https://
doi.org/10.1093/jpepsy/25.6.381.
25. Goodman R. The Strengths and Difficulties Questionnaire: A research note. J
Child Psychol Psychiatry. 1997;38:581–6. Available from: https://doi.org/10.
1111/j.1469-7610.1997.tb01545.x.
26. Varni JW, Seid M, Kurtin PS. PedsQL™4.0: Reliability and validity of the
Pediatric Quality of Life Inventory™Version 4.0 Generic Core Scales in
healthy and patient populations. Med Care. 2001;39:800–12.
Sanders et al. International Journal of Behavioral Nutrition and Physical Activity (2019) 16:117 Page 9 of 10
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
27. McClowry SG. The development of the School-Age Temperament Inventory.
Merrill Palmer Q. 1995;41:271–85.
28. Baker K, Sipthorp M, Edwards B. A Longitudinal Measure of Socioeconomic
Position in LSAC. Australian Institute of Family Studies; 2017. Available from:
https://growingupinaustralia.gov.au/sites/default/files/tp18.pdf
29. Australian Bureau of Statistics. Socio-economic indexes for areas. Canberra:
Australian Bureau of Statistics; 2001.
30. Sanders T, Parker PD, del Pozo-Cruz B, Noetel, M, Lonsdale, C. Screen-time,
health, education and well-being in children: Explorations of non-linear
relationships. Open Sci Framework; 2019. https://osf.io/bhzk8. Accessed 10
May 2019
31. Rubin DB. Multiple Imputation for Nonresponse in Surveys. John Wiley &
Sons; 2004.
32. van Ekris E, Altenburg TM, Singh AS, Proper KI, Heymans MW, Chinapaw
MJM. An evidence-update on the prospective relationship between
childhood sedentary behaviour and biomedical health indicators: a
systematic review and meta-analysis. Obes Rev. 2017;18:712–4. Available
from: https://doi.org/10.1111/obr.12526.
33. Ferguson CJ. Do angry birds make for angry children? A Meta-analysis of
video game influences on children’s and adolescents’aggression, mental
health, prosocial behavior, and academic performance. Perspect Psychol Sci.
2015;10:646–66. Available from: https://doi.org/10.1177/1745691615592234.
34. Orben A, Przybylski AK. The association between adolescent well-being and
digital technology use. Nat Hum Behav. 2019; Available from: https://doi.
org/10.1038/s41562-018-0506-1.
35. Rhodes A. Top ten child health problems: The Royal Children’s Hospital
Melbourne; 2015. Available from: https://www.childhealthpoll.org.au/polls/
top-10-child-health-problems/#detailedreport Accessed 10 May 2019
36. Schmidt ME, Haines J, O’Brien A, McDonald J, Price S, Sherry B, et al.
Systematic review of effective strategies for reducing screen time among
young children. Obesity. 2012;20:1338–54. Available from: https://doi.org/10.
1038/oby.2011.348.
37. Smith C, Galland BC, de Bruin WE, Taylor RW. Feasibility of Automated
Cameras to Measure Screen Use in Adolescents. Am J Prev Med. 2019;57:
417–24.
Publisher’sNote
Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.
Sanders et al. International Journal of Behavioral Nutrition and Physical Activity (2019) 16:117 Page 10 of 10
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
1.
2.
3.
4.
5.
6.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at
onlineservice@springernature.com
Available via license: CC BY
Content may be subject to copyright.
