Journal of Psychopathology and Behavioral Assessment

https://doi.org/10.1007/s10862-022-09962-y

A Measurement Model of Media Parenting: Differences Across Parent

and Child Reports and Youth Age and Sex
Erin Corcoran1 · Joy Gabrielli1 · Pamela Wisniewski2 · Todd D. Little3,4 · Jennifer Doty5

Accepted: 7 January 2022

© The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2022

Abstract
For youth raised in the Digital Age, online risks such as cyberbullying and sexting have become increasingly problematic.
Since digital media is primarily consumed at home, parents play an important role in mitigating these risks; parents can
teach children about online dangers, regulate the amount of time spent online, and, to some extent, curate the online content
children see. The present study evaluated the psychometric properties of a four-factor media parenting measurement model
introduced by Livingstone et al. (2011) across self-reports of a U.S. sample of parents (Mage = 38.5) and children (ages 10–14;
Mage = 11.8). To identify meaningful group differences, latent mean comparisons were evaluated across youth age and gender.
Confirmatory Factor Analysis results provided good fit to the data for the four-factor media parenting model based on both
parent [χ2(201, n = 306) = 384.407; ­RMSEA(.046—.063) = .055; CFI = .958; TLI/NNFI = .951; SRMR = .050] and child report
[χ2(203, n = 306) = 378.033; ­RMSEA(.045-.061) = .053; CFI = .942; TLI/NNFI = .934; SRMR = .060]. The final latent parenting
factors included: Active Mediation, Monitoring, Technology Control, and Restrictive Mediation. Latent mean comparisons
revealed that parents of girls reported higher levels of Monitoring than parents of boys, whereas girls reported higher levels
of parental Restriction than boys. Similarly, older children and their parents reported lower Restriction than younger children
and their parents. Overall, latent mean differences identified between media parenting domains may be important for youth
outcomes and provide support for their inclusion as distinct factors in predictive models.

Keywords Media parenting, confirmatory factor analysis · Structural equation modeling · Parent/child reporting ·
Informant discrepancy · Measurement invariance · Sex differences · Developmental differences

Introduction that 95% of American teens either own or have access to a

smartphone and 45% report being online “almost constantly”
America’s youth are growing up in a technology-saturated (Anderson & Jiang, 2018). Like most innovation, the benefits
world where they are constantly connected to each other of advancing technology (e.g., globalization, education, inter-
through the internet on personal digital devices (Anderson personal connection; Lenhart et al., 2010; Wright et al., 2001)
& Jiang, 2018). A 2018 Pew Research Center survey found come with potential disadvantages (e.g., increased access to
mature content and influences, unsupervised social connec-
* Erin Corcoran tion with strangers, and the opportunity to engage in online
[email protected] risk behaviors; Symons et al., 2020). Parents have the primary
responsibility for management of youth media and technologi-
1
Department of Clinical and Health Psychology, University cal risks, yet relatively little is known about how parents man-
of Florida, Gainesville, FL 32611, USA
age youth media access and interaction with media in the home.
2
College of Engineering and Computer Science, University Per socialization theory, parents serve as the primary
of Central Florida, Orlando, FL 32816, USA
means of teaching children how to function in society and,
3
Department of Educational Psychology, Texas Tech increasingly, on the internet (Symons et al., 2020). Included
University, Lubbock, TX 79409, USA
in Symons et al.’s conceptualization of socialization is
4
Optentia Research Focus Area, North-West University, parental mediation theory, whereby parents use a variety
Vanderbijlpark, South Africa
of strategies to mitigate the negative effects of media on
5
Department of Family, Youth and Community Sciences, their children (Clark, 2011; Symons et al., 2020). This idea
University of Florida, Gainesville, FL 32611, USA

13
Vol.:(0123456789)
 Journal of Psychopathology and Behavioral Assessment

became particularly prominent with regards to traditional Padilla-Walker et al., 2020), risky sexual behavior (Bersamin
media use (Clark, 2011; Dorr et al., 1989; Lin & Atkin, et al., 2008), cyberbullying perpetration (Elsaesser et al.,
1989; Valkenburg et al., 1999) and typically includes com- 2017), cyberbullying victimization (Elsaesser et al., 2017), and
ponents of the following five domains: 1) Active Mediation other risk behaviors. For example, a 2019 metanalysis (Chen
(i.e., discussing media content with children); 2) Restric- & Shi, 2019) found that Restrictive Mediation was related to
tive Mediation (i.e., setting rules and restrictions on what decreased screen time, while Active Mediation and Co-viewing
media can be consumed, for how long, and/or where); 3) were most effectively employed for reducing risk behavior.
Co-viewing or Co-using (i.e., joint participation in a media This holds with the idea that children are more able to resist
activity, such as watching TV together); 4) Monitoring (i.e., negative media effects when they have developed skills to criti-
checking on children’s media use, either overtly or covertly, cally engage with the material (Warren et al., 2002).
after use); and 5) Technology Control (i.e., using technology Although most traditional parental media mediation
to limit media time or content; e.g., Valkenburg et al., 1999). behaviors still have obvious application with personalized
While emerging research suggests that media parenting is media devices, the changing media landscape warrants re-
important for youth outcomes, relatively little work has been evaluation of the original media parenting scales that were
done to provide psychometric support for the constructs primarily focused on media such as traditional television
listed above. We also lack consensus on who should be and movie content. For example, parents Co-use/Co-view
reporting on media parenting behaviors; do parents or chil- traditional media, such as television, much more frequently
dren provide more accurate and meaningful interpretation? than video games or individual devices like smartphones
Moreover, evidence suggests media parenting changes as or tablets (Livingstone & Helsper, 2008). Given that youth
youth age, as is developmentally appropriate (Padilla-Walker are spending increasing time with streamed media content
et al., 2012). Research on general parenting has identified (Rideout & Robb, 2019), parenting within internet-based
differences in parenting behaviors based on child gender, contexts seems to be rising in importance. Despite growing
but there is less clarity on whether that applies to media literature on the various media parenting domains, research
parenting (Morawska, 2020). Additional work is needed to is still in its nascent phases for evaluation of the validity and
determine the invariance of these constructs across devel- reliability of measures, particularly with regards to modern
opmental transitions and gender. Subsequently, the present media contexts and technology.
manuscript focuses on assessment of media parenting meas-
urement model fit based on parent and child report as well as Psychometric Properties of Media Parenting Measures
invariance across child age (younger vs older) and gender.
Bybee et al. (1982) first discussed the psychometric proper-
Research Foundations of Media Parenting ties of a media parenting scale in the context of television
use, using Principal Components Analysis to identify three
Alongside the rise of television and personalized media parenting domains: Restrictive Guidance (i.e., restrictions on
devices in the home came the need to distinguish between amount and content of media viewed), Evaluative Guidance
general and media-specific parenting practices. Despite (i.e., parenting behaviors aimed at helping the child “evalu-
conceptual similarities and a strong statistical relationship, ate the meaning, morality and characterization of television
media parenting and general parenting behaviors are distinct programs”), and Unfocused Guidance (i.e., watching televi-
(Eastin et al., 2006; Sanders et al., 2016; Tanski et al., 2010); sion with the child, encouraging certain programs, and talk-
parents who are adept at general adaptive parenting are not ing about the content). In the decades that followed, multiple
always able to translate such practices to technology-based research groups (e.g., Valkenburg et al., 1999; Austin et al.,
parenting. That said, features of the parent–child relationship 1993; Dorr et al., 1989; Van den Voort et al., 1992) across
(e.g., warmth) have been argued to establish the context for fields (e.g., communications, psychology, public health,
media-specific parenting behaviors and may be more impor- and marketing) have undertaken research on the topic. In
tant than the parenting practices themselves in deterring the process, additional/alternative domains such as Active
online risks like cyberbullying (Elsaesser et al., 2017). These Mediation (i.e., active discussion of television content), Co-
findings underscore the idea that general adaptive parenting viewing (i.e., watching media together with children), and
is associated with media parenting but may exert differential Instructive Mediation (i.e., discussing, explaining, and teach-
influence on youth media behaviors. ing around television content) emerged, with often overlap-
Research has established media parenting, separate from ping component behaviors or themes. For this and other rea-
general parenting, as an important intervening factor on sons, a shared understanding of the term “mediation” and
youth risk behavior; parental mediation of media use has the domains within is lacking (Nathanson, 2001). Nathanson
been associated with decreased alcohol and marijuana con- (2001) endeavored to synthesize the extant research, iden-
sumption (Cox et al., 2018), aggression (Nathanson, 1999; tifying Active Mediation, Restriction, and Co-viewing as

13
Journal of Psychopathology and Behavioral Assessment

three primary media parenting domains. Active Mediation Inspection, and Authoritarian Surveillance. Model fit for
encompasses behaviors described in Instructive Mediation these constructs were acceptable and promising for the
and Evaluative Guidance above and is marked by engage- forward progression of a media parenting construct. That
ment with youth about media, with the goal of building their being said, Ho et al.’s (2019) work is focused on social media
ability to assess media independently and decrease suscep- rather than media on the whole, and Nikken and Jansz’s
tibility to mature content. Although Active Mediation often (2014) findings are almost a decade old and focused on a
occurs while parents Co-view content with their children, it is younger population rather than adolescents who use online
not necessary for Co-viewing and is thus distinct. Examples media more heavily. Both are centered on populations in
of Active Mediation might include explaining why a char- their respective countries (Singapore and Denmark). Thus,
acter got sick when drinking alcohol; Restrictive Mediation although suggestive of reliable domains, these findings high-
might include setting a rule in the home that the television light the need for confirmation of a reliable media parenting
cannot be used after 9 in the evening; Co-viewing might be measure that is applicable to the changing media landscape
employed by watching a television show with the child. This in the United States. These findings also underscore the
three-dimensional classification predominates today. question others have asked of whether previously estab-
While providing a valuable framework, limitations exist lished domains are relevant, or if we should be developing
in the use of television-specific domains for new digital/ a new construct rather than modifying an existing construct
social media. Traditional media, or broadcast media, is (e.g., Eastin et al., 2006; Nikken & Jansz, 2003). The pre-
typically created by a production company or otherwise sent study addresses these questions by testing a measure-
external source and consumed passively. Conversely, new ment model of media parenting behaviors, adapted from
digital/social media is both passively consumed and actively those of Livingstone et al. (2011), in a covariance model
engaged with, as individuals can create their own content to assess for relationships across domains. This measure
and consume the content of peers and companies alike has been tested by others utilizing all five original scales
(Chassiakos et al., 2016). If considered in the context of and binary response options (Dürager & Sonck, 2014) and
social learning theory (Bandura, 1986), this has implications has been found to associate significantly with youth out-
for how the content is processed; for example, peer content comes (e.g., parental risk perception, children's online skills,
will likely be interpreted and internalized differently than children's online opportunities, and children's online risks;
content that is industry-derived and then broadcast. Accord- Livingstone et al., 2017). Adaptations made in the present
ingly, Eastin et al. (2006) adapted Valkenburg et al.’s (1999) study address the suggestions of previous researchers to uti-
media parenting domains to be more applicable to modern lize ordinal scales (Dürager & Sonck, 2014) as well as the
technology, adding Technological Mediation to account for need to continually update measures to apply to the current
the use of tracking software to monitor internet use (per media landscape. The present study also extends use of the
Greenberg et al., 2001). Livingstone et al. (2011) also measure, which was developed with a European sample, to
assessed technical mediation in addition to previously a nationally representative United States sample of middle
defined domains (e.g., Active Mediation of child’s internet school-aged youth and their parents. Finally, this study also
use, Active Mediation of child’s internet safety, Monitoring, offers a comparison of parent and child report, which few
Restriction), aiming to capture parenting behaviors specific studies have done with media parenting behaviors.
to modern media and technology. A subset of these domains
and items were used in the present study. Media Parenting Across Age and Gender
Despite laying conceptual groundwork, research on the
topic of parental mediation broadly lacks psychometric sup- Previous research has identified trends in media parenting
port; a 2013 metanalysis of studies on media parenting between younger and older children. For example, the devel-
found that only 20.7% of studies provided information on opmental trajectory of media parenting appears to mirror that
the psychometrics of outcome and exposure measures (Jago of general parenting, whereby parents become less restric-
et al., 2013). In order to address this gap, researchers have tive as their youth age (e.g., Lin & Atkin, 1989; Padilla-
more recently employed structural equation modeling tech- Walker et al., 2012; Rideout et al., 2010). The increased use
niques, thus enabling assessment of model fit and structure. of mobile devices also lends itself to less parental supervi-
Nikken and Jansz (2014) identified a five-factor media sion and monitoring both inside and outside of the home
parenting model with youth aged 2 – 12 years including (Kerr et al., 2010). Similarly, research has found that Active
Co-use, Active Mediation, Restrictive Mediation, Super- Mediation decreases with youth age (Beyens & Valkenburg,
vision, and Technical Safety Guidance, with the latter two 2019; Padilla-Walker et al., 2012; Warren, 2017), although
domains more specific to modern media. Ho et al. (2019) earlier research with traditional forms of media did not find
tested a four-factor model of parental mediation specific to this to be the case (Bybee et al., 1982; Van der Voort et al.,
social media including Active, Restrictive, Non-intrusive 1992). Developmental trends are less clear for Co-viewing;

13
 Journal of Psychopathology and Behavioral Assessment

although some researchers have found evidence of decreased the two-factor structure only accounted for 65% of the
Co-viewing over time (Austin et al., 1999; Warren et al., variance in parent behavior (Livingstone et al., 2017).
2002), others have posited that Co-viewing increases with Furthermore, research has established differential effects
youth age (Dorr et al., 1989), potentially due to converging of the Evaluative factor’s component parts (i.e., Active,
media interests (Gentile et al., 2012). Overall, tapering of Monitoring, and Technology Controls) on outcomes,
mediation with youth age is expected; if parents are adapting calling the efficacy of consolidation of media parenting
to children’s need for monitoring and supervision in balance domains into a single factor into question. For example,
with children’s need to develop skills around autonomy and Campbell and Park (2014) found that parental monitoring
independence, media parenting should show developmental was not associated with decreased youth sexting behaviors
differences in reliable and predictable trends. but frequent family communication was.
Research on gender differences in media parenting Thus, despite demonstrating potential, higher order con-
is relatively scarce and difficult to synthesize due to structs require further exploration. For the purposes of this
contradictory findings and different contexts. A 2020 study, we label the Enabling/Evaluative factor as described
metanalysis of general parenting literature identified gender by Livingstone et al. (2017) as “Proactive” media parent-
differences in parenting (Morawska, 2020), but, to date, ing to better reflect the more active behaviors employed
there is no consensus on whether this occurs with media by parents using Active Mediation, Monitoring, or Tech-
parenting. Early research suggested that there was no nology Control. These behaviors are marked by parent
difference in Active Mediation or Co-viewing based on child engagement with child media use either through discus-
gender (Bybee et al., 1982; Valkenburg et al., 1999; Van der sion, checking, or adaptively implementing technology to
Voort et al., 1992). Eastin (2006) found similar results for monitor use or limit media content/time. This is compared
Interpretive Mediation between males and female children to Restrictive Mediation, which is captured by behaviors
but identified gender differences for Co-viewing, content and intended to prevent exposure to mature content and do
time restrictions. Lin and Atkin (1989) found that parents not assume access to media (e.g., “please specify how
of males were more likely to set rules on VCR-usage (e.g., restrictive your parents are about having your own social
when, how often, with whom, and what they could watch) networking profile,” with options ranging from “never let
than parents of females, although in a literature review, me” to “whenever I want”).
Nathanson (2001) did not identify notable gender differences
in parental Restriction. Given these discrepancies, further
exploration of parental media mediation between genders, Discrepancies by Reporter: Parent Versus Child
particularly in the context of new media, is warranted.
Although most research to this point has utilized parent
Higher‑order Models report (Nathanson, 2001), it remains unknown whether
parents or children are better reporters of media parenting
For the purposes of parsimony, summarizing across behaviors. Comparative studies have found disagreement
domains of media parenting with a higher order construct between parent and child report measures (Fujioka & Austin,
may prove useful for future research. Indeed, mediation 2002); for example, a study of parent and child report of
itself can be considered a higher order construct parental mediation found that reports were correlated but
(Nathanson, 2001), and the manner in which specific significantly different in value (Beyens & Valkenburg, 2019).
media parenting behaviors are grouped within is up for These discrepancies could exist for a variety of reasons,
debate. In recognition of the various behaviors comprising including that children may not be privy to the parental
parental media mediation and the conceptual and mediation strategies being implemented or perceiving them
behavioral overlap implicit between them, research has as strategies at all (Buijzen et al., 2008). They may also
explored the possibility of consolidating media parenting want to overclaim their independence by downplaying rules
behaviors into fewer factors (Livingstone et al., 2017). For placed on them (Greenberg et al., 1972; Lin & Atkin, 1989).
example, Livingstone et al. (2017) utilized factor analysis Conversely, parents are vulnerable to social desirability
with varimax rotation to identify a two-factor structure: bias, thus potentially overstating their use of mediation
Restrictive and Enabling media parenting. This Enabling strategies (Garmendia et al., 2012; Greenberg et al., 1972;
or Evaluative factor included Active media parenting (as Lin & Atkin, 1989). These reporter discrepancies may
previously defined), Technology Controls, and Moni- be reflective of family dynamic and informative for out-
toring, while the Restrictive factor remained the same comes (Des Los Reyes, 2011); for example, discrepancy
and referred to rules and restrictions on media content in parent and child report of parental monitoring sur-
and use. Although conceptually this divide between rounding alcohol behavior was found to be predictive of
more active engagement and passive rule setting is clear, increased youth alcohol use (Abar et al., 2015).

13
Journal of Psychopathology and Behavioral Assessment

Active Monitoring TechControl Restrictive

A1 A2 A3 A4 A5 M1 M2 M3 M4 M5 M6 T1 T2 T3 T4 T5 R1 R2 R3 R4 R5 R6

Fig. 1  Hypothesized four-factor measurement model of media parenting behaviors

Present Study youth gender (male vs female; Aim 2). Once invariance was
established, we examined latent mean differences in the four
Based on prior support for parenting domains, we hypoth- media parenting constructs across groups (Aim 3). Based
esized that a four-factor latent measurement model would on developmental theory suggesting that parents relax rules
provide acceptable fit to the data across both parent and and structure as children age and prior literature suggesting
child reporters (Aim 1; Fig. 1). We adapted the domains that parents may interact with male and female children dif-
proposed by Livingstone et al. (2011; 2017) by combining ferently, we expected to identify latent mean differences in
Active Mediation of internet use with Active Mediation of media parenting approaches across youth age and gender.
internet safety and trimming less relevant items (e.g., digi-
tal safety, which was captured elsewhere in the survey) to
reduce participant burden. We included an exploratory aim Method
to test the possibility of a one-factor (general media parent-
ing; Fig. 2) or two-factor (Restrictive and Proactive media Participants were 306 parent/child dyads representing all
parenting; Fig. 3) higher order model. Further, we hypoth- five regions of the United States enrolled in an online sur-
esized that the final measurement models would be invariant vey of parent and child technology use and online experi-
across youth age (10 – 11 years old vs 12 – 14 years old) and ences. Data included parent report of media management

Media
Parenting

Active Monitoring TechControl Restrictive

A1 A2 A3 A4 A5 M1 M2 M3 M4 M5 M6 T1 T2 T3 T4 T5 R1 R2 R3 R4 R5 R6

Fig. 2  Exploration of single factor higher order measurement model of media parenting behaviors

13
 Journal of Psychopathology and Behavioral Assessment

Proactive

Active Monitoring TechControl Restrictive

A1 A2 A3 A4 A5 M1 M2 M3 M4 M5 M6 T1 T2 T3 T4 T5 R1 R2 R3 R4 R5 R6

Fig. 3  Exploration of two-factor higher order measurement model of media parenting behaviors

behaviors, youth report of the same, and a variety of youth four mediation scales reflected acceptable to excellent internal
report measures regarding online and built-world risk behav- consistency (George & Mallery, 2003): 1) Active Mediation (5
iors. Parents of 10–14-year-old children were identified and items; parent report [PR] α = 0.84; child report [CR] α = 0.77),
recruited by Qualtrics panel specialists. After completing including items such as, “Do you or your child’s other parent/
the initial portion of the survey, parents provided consent for caregiver currently talk to your child about what he/she does on
their 10–14-year-old child to participate and were instructed the internet?” 2) Restrictive Mediation (6 items; PR α = 0.89;
to afford children privacy to complete the survey. Children CR α = 0.76), including items such as, “Please specify the extent
provided assent for participation. to which you or your child’s other parent/caregiver restrict
Of child reporters, 49% were female and 20% minority (pri- your child from giving out personal information to others on
marily Hispanic or Black), with an average age of 11.8 years the internet.” 3) Technology Control (5 items; PR α = 0.91;
(SD = 1.23 years). Of parent reporters, 61% were female, CR α = 0.85), including items such as “How often do you use
20% minority (primarily Hispanic or Black), and 71% mar- parental control technologies to block or filter some types of
ried or living with a partner, with an average age of 38.5 years websites your child visits?” and 4) Monitoring (6 items; PR
(SD = 6.01 years). Parents indicated that 51% of youth received α = 0.99; CR α = 0.81), including items such as “Do either you
free or reduced lunch at school (a descriptive proxy for sample or your child's other parent/caregiver check which websites your
socio-economic status; Table 1). child visited based on his/her internet browsing history?” These
questions were reframed for youth participants; for example,
Measures the first question in the Active domain read “do either of your
parents currently talk to you about what you do on the internet?”
Parents reported on demographic questions about themselves Other survey items that were not included in the measure-
and their children, including age, gender, race/ethnicity, ment model but were included in the imputation include par-
whether the child receives free or reduced lunch (a socio- ent report of youth access to technology and history of school
economic indicator), and parent marital status (included for suspension, as well as child report of digital safety, problem
sample description). Children also provided demographic behavior frequency, cyberbullying perpetration and victimiza-
information on age, gender, and race/ethnicity, which were tion, and social self-efficacy. The parent survey was estimated
compared to parent report for data cleaning purposes. to take 20 minutes and the child survey 30 minutes.
Parents were asked 22 questions about the frequency of
parental mediation strategies that are used in the home based on
previous studies by Livingstone et al. (2011; 2017), with Likert- Procedure
style responses ranging from 1 = Not at All to 5 = Almost All the
Time. Children were asked a random selection of 16 of these Data Analytic Plan
questions (with language adapted to reflect youth perspective)
to reduce burden on child participants. This planned missingness We utilized confirmatory factor analyses (CFA) within a struc-
design is described in more detail below. Cronbach’s α for the tural equation modeling framework using Mplus 8.4 (Muthén &

13
Journal of Psychopathology and Behavioral Assessment

Table 1  Sample Descriptives Variable n % M SD

Youth Age – – 11.8 1.2

10 57 18.6
11 71 23.1
12 81 26.4
13 71 23.1
14 26 8.5
Parent Age – – 38.5 6.0
Youth Gender
Male 155 50.7
Female 151 49.2
Parent Gender
Male 120 39.2
Female 186 60.8
Youth Race/Ethnicity
Black or African American 35 11.4
White or Caucasian 214 69.9
Hispanic/Latino 27 8.8
Other 30 9.9
Parent Race/Ethnicity
Black or African American 39 12.7
White or Caucasian 220 71.9
Hispanic/Latino 21 6.9
Other 47 8.5
Parent Marital Status
Divorced or Separated 40 13.1
Married 189 61.8
Never Married 40 13.1
Widowed 8 2.6
Living Together Not Married 29 9.4

Muthén, 2017) to evaluate the four-factor and higher-order latent the absence of responses decreases power, multiple imputation
measurement models across parent and child reports. Maximum is one of the two modern missing data estimation techniques that
likelihood estimation was utilized. Models were identified via restores missing power without biasing point estimates (Johnson
variance standardization. Indicator loadings were evaluated for & Young, 2011; Little et al., 2013). Regardless of planned miss-
statistical significance at the p < 0.001 level. Model fit was esti- ingness, total missingness for any given item did not exceed
mated using guidelines set forth by Little (2013), with the root 33%. Multiple imputation was used to estimate these values.
mean square error of approximation (RMSEA) and standardized There was no missingness on parent items. Invariance testing
root mean square residual (SRMR) less than 0.08, and compara- was employed to determine measurement equivalence across
tive fit index (CFI) and Tucker–Lewis index or non-normed fit child gender and age (10–11 and 12–14) for both parent and
index (TLI/NNFI) above 0.90. child report. Following the guidelines set forth by Byrne (2013),
In order to reduce test-taking burden on child participants, configural invariance was first established by comparing groups
the child portion of the survey utilized a planned missingness without constraints. Then, factor loadings of each group were
design whereby each child was randomized to receive two thirds constrained to equality to assess metric (or weak) invariance.
of the questions in each parent mediation scale. This approach Finally, intercept means were constrained to equality for scalar
was justified according to the findings of Little and Rhemtulla (or strong) invariance. If model fit of each iteration is signifi-
(2013) as well as Jia et al. (2014). Planned missingness designs cantly worsened from the configural model, as determined by
enable researchers to interpret missing data as MCAR due to a significant Chi Square Difference Test or a decrease in CFI
the completely random nature of assignment, which, in turn, or TLI values of more than 0.01, then invariance is not sup-
mitigates risk of estimation bias (Little et al., 2013). Although ported (Cheung & Rensvold, 2002). Once scalar invariance is

13
 Journal of Psychopathology and Behavioral Assessment

established, latent mean comparisons can be made. Independent CFI = 0.942; TLI/NNFI = 0.934; SRMR = 0.060 (see Fig. 5).
samples T-Tests were conducted to compare parent and child No indicator residuals were freed in the child model.
report of media parenting behaviors. Inter-rater bivariate correlations (between parents and
children; r) ranged from 0.41—0.80, which is interpret-
able as medium to large effect sizes per Cohen (1988; see
Results Table 2) and p < 0.001. Active Mediation items were least
correlated between reporters (0.414—0.632, n = 5), followed
Measurement Model Fit by Monitoring (0.545—0.614, n = 6). Restriction and Tech-
nology Control were most correlated (0.600—0.807, n = 6;
For the parent-report model, all indicator factor loadings for 0.658—0.734, n = 5). Within both parent and child mod-
Active, Monitoring, Technology Control and Restrictive Media- els, Active Mediation, Monitoring and Technology Con-
tion were statistically significant at the p < 0.001 level. Modifica- trol were all significantly positively related to each other,
tion indices were evaluated, and, in conjunction with theoretical while Restrictive Mediation was only significantly related
rational, the following residuals were then freed to correlate: to Active Mediation (see Figs. 4 and 5).
1) Monitoring indicators: “Do either you or your child's other
parent/caregiver check your child’s profile on a social network Higher‑order Models
or online community?” and “Do either you or your child's
other parent/caregiver check which friends or contacts your Two possible high order models were assessed: an overarching
child adds to his/her social networking profile?”; 2) Technol- media-parenting factor and a two-factor model with Restric-
ogy Control indicators: “How often do you use parental control tive and Active Mediation (comprising Active, Technology
technologies to block or filter some types of websites your child Control and Monitoring, similar to the Enabling factor defined
visits?” and “How often do you use parental control technolo- by Livingstone et al., 2017). Both iterations had very simi-
gies to keep track of the websites your child visits?” Follow- lar model fit, which did not provide significant improvement
ing the freeing of these two residuals, CFA using maximum over the four-factor model and, in fact, were slightly worse:
likelihood estimation demonstrated adequate fit in the overall PR χ 2(df = 203) = 397.171; ­RMSEA(0.048—0.064) = 0.056;
sample χ2(df = 201) = 384.407; ­RMSEA(0.046—0.063) = 0.055; CFI = 0.955; TLI/NNFI = 0.949; SRMR = 0.061; CR
CFI = 0.958; TLI/NNFI = 0.951; SRMR = 0.050 (see Fig. 4). χ 2 (df = 205) = 388.537; ­R MSEA (0.046—0.062) = 0.054;
For the child-report model, all factor loadings for CFI = 0.939; TLI/NNFI = 0.931; SRMR = 0.066. While model
Active, Monitoring, Technology Control and Restrictive fit indices were acceptable for the higher order constructs,
Mediation were statistically significant at the p < 0.001 reductions at this level may obscure important cross-scale dif-
level. CFA demonstrated adequate fit in the overall sam- ferences. Thus, the four-factor model was the final model used
ple χ 2(df = 203) = 378.033; ­RMSEA(0.045—0.061) = 0.053; for invariance tests, below.

Fig. 4  Parent Report – Final four-factor measurement model of media parenting behaviors with standardized loadings, factor covariances, cor-
related residuals, and residual error values. ***Indicates significance at p < .001; ns = non-significant

13
Journal of Psychopathology and Behavioral Assessment

Fig. 5  Child Report – Final four-factor measurement model of media parenting behaviors with standardized loadings, factor covariances, and
residual error values. All factor loadings are significant at p < .001. ***p < .001; ns = non-significant

Group Invariance Testing across Child Age children for parent and child report, with both achieving
and Gender good to acceptable model fit per standards outlined above:
PR χ 2(df = 402) = 705.064; ­RMSEA(0.062—0.079) = 0.070;
The first step of invariance, configural, was assessed by sep- CFI = 0.932; TLI/NNFI = 0.922; SRMR = 0.062; CR
arately testing the measurement models of younger and older χ 2 (df = 406) = 823.667; ­R MSEA (0.074—0.090) = 0.082;

Table 2  Correlations Between Correlation n

Parent and Child Report of
Media Parenting Behaviors Active
A1 Talk about internet activity 0.41 243
A2 Explain why some websites good or bad 0.47 245
A3 Suggest ways to use the internet safely 0.51 244
A4 Suggest ways to behave towards other people online 0.63 245
A5 Provide help when something bothers child on the internet 0.46 246
Monitoring
M1 Websites visited based on browsing history 0.59 201
M2 Profile on social media 0.60 204
M3 Friends on social media 0.55 206
M4 Messages on email or instant messaging account 0.61 202
M5 Texts or photo messages on cell phone 0.58 203
M6 Apps installed on cell phone 0.56 202
Technology Control
T1 Block or filter websites 0.73 241
T2 Track websites visited 0.69 245
T3 Limit time on the internet 0.66 243
T4 Monitor text or photo messaging activities from cell phone 0.66 246
T5 Monitor what apps are installed or used on cell phone 0.72 243
Restrictive
R1 Give out personal information to others on the internet 0.60 208
R2 Upload photos, videos or music to share with others 0.73 203
R3 Download music or films on the internet 0.67 203
R4 Have a social media profile 0.78 203
R5 Use instant messaging 0.75 199
R6 Have cell phone 0.81 203

All factor loadings significant at p < .001

13
 Journal of Psychopathology and Behavioral Assessment

CFI = 0.871; TLI/NNFI = 0.854; SRMR = 0.091. A similar (Mdiff = 0.273, p < 0.05). Similarly, older children and their
model fit and pattern of salient and non-salient factor load- parents reported lower Restriction than younger children
ings was observed between groups, allowing us to proceed (Mdiff = 0.429, p < 0.01) and their parents (Mdiff = 0.382,
to test metric invariance. The additional constraints on factor p < 0.01; see Table 4) for a comparison of parent and child
loadings did not result in significantly worse model fit for mean scores). Overall, parents and children reported more
parent and child models across child age and gender based Active Mediation than any other technique (PR M = 4.282;
on the change in χ2, CFI, and TLI (see Table 3), thus allow- CR M = 4.080), followed by Monitoring (PR M = 3.703;
ing us to proceed to scalar invariance. Accordingly, scalar CR M = 3.489), Restriction (PR M = 3.270; CR M = 3.060),
invariance also did not significantly worsen model fit for and Technology Control (PR M = 3.383; CR M = 3.221).
parent or child report (see Table 3). Two-tailed independent sample T-Tests revealed a signifi-
Invariance testing for males and females was conducted cant difference between parent and child report of Active
in a similar fashion to that which was described above. Con- Mediation at p < 0.01, with parents reporting more Active
figural invariance was established across male and female Mediation than their children overall, t(610) = 3.339. Non-
models based on parent and child report, indicating a similar significant differences were found for reports of Monitor-
factor structure. Both achieved good to acceptable model ing t(610) = 2.332, Technology Control t(610) = 2.337, and
fit: PR χ2(df = 402) = 695.870; ­RMSEA(0.060—0.078) = 0.069; Restriction t(610) = 1.527.
CFI = 0.934; TLI/NNFI = 0.924; SRMR = 0.064.; CR
χ 2 (df = 406) = 730.744; ­R MSEA (0.064—0.081) = 0.072;
CFI = 0.897; TLI/NNFI = 0.883; SRMR = 0.080. Metric and Discussion
full scalar invariance were also established for gender and
age models (see Table 3). Parents are in a critical position to intervene on youth online
risk behavior through media mediation. The present study
Media Parenting Factor Means adds to and synthesizes findings of the small body of research
on the psychometric properties of media parenting measures
Latent mean comparisons indicate that parents of females (e.g., Bybee et al., 1982; Ho et al., 2019; Livingstone &
reported higher levels of Monitoring than parents of males Helsper, 2008; Valkenburg et al., 1999, 2013; Van der Voort
(Mean Difference (Mdiff) = 0.297, p < 0.05), whereas female et al., 1992), supporting the presence of four distinct media
youth reported higher levels of Restriction than male youth parenting factors and the efficacy of both parent and child

Table 3  Fit Indices for Invariance Testing Across Youth Gender and Age (10–11 Years Old Versus 12–14 Years Old)
χ2 df p Δχ2 df p CFI Δ CFI TLI/NNFI Tenable?a

Child Report – Gender

Configural 730.744 406 0.000 – – – 0.897 – 0.883 –
Weak 743.829 424 0.000 13.085 18 0.786 0.899 0.002 0.890 Yes
Strong 763.586 442 0.000 19.757 18 0.347 0.898 -0.001 0.894 Yes
Child Report – Age
Configural 823.667 406 0.000 – – – 0.871 – 0.854 –
Weak 836.371 424 0.000 12.704 18 0.809 0.873 0.002 0.862 Yes
Strong 847.893 442 0.000 11.522 18 0.871 0.875 0.002 0.869 Yes
Parent Report – Gender
Configural 695.870 402 0.000 – – – 0.934 – 0.924 –
Weak 713.448 420 0.000 17.578 18 0.484 0.934 0.000 0.927 Yes
Strong 737.978 438 0.000 24.53 18 0.138 0.932 -0.002 0.929 Yes
Parent Report – Age
Configural 705.064 402 0.000 – – – 0.932 – 0.922 –
Weak 718.780 420 0.000 13.716 18 0.747 0.933 0.001 0.926 Yes
Strong 740.662 438 0.000 21.882 18 0.237 0.932 -0.001 0.928 Yes
Weak and Strong Invariance tests evaluated according to Chi Square Difference test and change in CFI. Configural models provide baseline for
metric and scalar tests that follow
a
If Tenable = Yes, invariance is established at that level

13
Journal of Psychopathology and Behavioral Assessment

Table 4  Results of Latent Factor Gender Age

Mean Comparisons of Parent
and Child Reports of Media M Difference p M Difference p
Parenting Behaviors Across
Male (Reference) Versus Parent
Female Children and 10–11 Active 0.152 0.206 -0.090 0.470
Year Olds (Reference) Versus Monitoring 0.257* 0.022 0.033 0.769
12–14 Year Olds
Tech Control -0.024 0.844 0.136 0.268
Restrictive 0.169 0.138 -0.338** 0.004
Child
Active 0.205 0.105 -0.189 0.168
Monitoring 0.194 0.098 0.061 0.602
Tech Control 0.014 0.903 0.114 0.337
Restrictive 0.256* 0.033 -0.378** 0.002
Latent means for males and younger youth, respectively, were used as reference groups and standardized to
have means of zero. Non-equivalence suggests significant latent mean differences
*
p < .05; **p < .01

report. Construct validity, including convergent and discri- and new media, such as the TECH parenting model pro-
minant validity, is well supported by the present study, with posed by Gabrielli et al. (2018) comprising Talk, Educate,
individual items loading onto their respective factors simi- Co-view and House Rules. Included in the Co-view/Co-use
larly to other within-factor items at a significance of p < 0.001. domain is guidance specific to new media. Additionally, a
Results also indicate differences in parenting across child age gap that needs to be addressed is the potential change in par-
and gender, with more Restriction for younger youth, and enting techniques with the emergence of more personalized
more Monitoring (parent report) and Restriction (child report) media devices. For example, monitoring youth television
for females. A second aim of this study was to assess the viewership is considerably different than youth Snapchat
equivalence of measurement models across age and gender for and Instagram usage, which involve private messages and
parent and child report of media parenting behaviors, which time-limited content. It is also different than supervision of
findings confirmed. video games, which is increasingly interactive and immer-
sive (Jiow et al., 2017). At the center of much new media
Implications of Measurement Model is privacy and, without validated measures of youth media
privacy perceptions and behaviors, it is challenging to have
Researchers have questioned whether the parenting domains a comprehensive understanding of youth media use. An
established in the television research are still applicable, or understanding of how media parenting directly relates to
whether the existing domains should be entirely reconsid- youth online behavior, including privacy, will be important
ered in the digital age (e.g., Eastin et al., 2006). Given the in determining whether media parenting is actually effective
present study’s identification of well-fitting models and in the current media environment.
correlated parent–child reports, complete overhaul of the In the present study, Active, Monitoring, and Technology
models developed in earlier research (e.g., Bybee et al., Control factors are positively related across reporters, but
1982; Valkenburg et al., 1999) based on youth TV viewer- only parent report indicated significant associations between
ship appears unwarranted. Rather, these measures might best Restrictive Mediation and any other constructs, in this case a
be updated with question content more specific to today’s positive relationship with Active Mediation. In other words,
technology usage (e.g., social media; Nikken & Jansz, 2014; parents tend to employ these behaviors in tandem, and par-
and Ho et al., 2019), and refining the more active compo- ents who are utilizing Active Mediation techniques are likely
nents of media parenting practices, such as Active, Moni- to be simultaneously Monitoring, Restricting, and using
toring, and Technology Control. The domains set forth by Technology Control. This supports past research which has
Livingstone et al. (2017) make progress towards this goal by argued that the use of multiple forms of mediation in con-
removing Co-viewing and incorporating Technology Con- junction may be most effective for youth outcomes (Padilla-
trols. The present study iterates on such progress via a struc- Walker et al., 2018). More work is needed to assess how
tural equation modeling approach enabling assessment of different strategies can be combined most effectively. The
model fit and comparison of parent and child reports. Future lack of shared covariance between Restrictive Mediation and
research may also explore the re-integration of Co-viewing Technology Control and Monitoring helps explain the lack
with appropriate adaptations for applicability to traditional of significant improvement by adding a single higher order

13
 Journal of Psychopathology and Behavioral Assessment

media parenting construct. By the same logic, the shared however, provide evidence for which reporter is more accurate
variance between Active, Technology Control, and Monitor- or predictive; future research should explore these questions to
ing may appear to lend credence to a two-factor Proactive support understanding of how media parenting behaviors relate
and Restrictive structure. While the two-factor higher order to youth outcomes across parent and child reporters. Also of
constructs did not provide improvements to model fit, they note, the proportion of fathers taking part in this survey is a
did offer acceptable model fit. There are a variety of poten- higher than average (Phares et al., 2005). This is a relative
tial reasons for this finding. Some research has suggested strength of this study as, historically, fathers have been under-
that media parenting behaviors are differentially employed represented in pediatric psychology and clinical child research
in proactive and reactive situations (Wisniewski et al., 2015). and, thus, interventions.
Additionally, there is evidence that parenting behaviors may
have differential effects on outcomes; autonomy-granting Invariance Testing
behaviors (such as Active Mediation) lead to better outcomes
than autonomy-restricting behaviors (which may include Structural equation modeling allows us to assess model fit
Monitoring and Technology Controls; Ghosh et al. 2018b). and invariance between males and females and younger
When considering employment of higher order constructs and older children, which has not been evaluated in media
within measurement of media parenting, future work should parenting behaviors to this point. While developmental
weigh potential benefits (e.g., parsimony of measurement theory and social constructs related to parenting across
model/constructs) and costs (e.g., loss of specificity across youth gender suggests differences in the factor structure
potentially divergent parenting factors). Given the nascent may have arisen, our invariance testing suggests that it did
stage of this literature and emerging work on measurement not, providing good support for global use of these meas-
models, it may be important for researchers to present both ures. Alternatively, latent mean differences indicated that
higher order and lower order models until more consistent both parents and children report different levels of parenting
relations are established. techniques based on child gender and age, with parents of
girls reporting Monitoring more than parents of boys, and
Reporter Discrepancies girls reporting more Restriction than boys. A possible expla-
nation for this difference between parent and child report of
Unique to the present study is a comparison of parent and child Monitoring is that media-specific monitoring is often con-
report of parental Active Mediation, Restrictive Mediation, ducted covertly (Ghosh et al., 2017), so the child may not
Monitoring, and Technology Control through structural equa- be aware of the extent to which their parents are monitoring
tion modeling; historically, researchers will only utilize one their behaviors. Gender differences may reflect differential
reporter depending on their research question or convenience socialization patterns by gender, especially given past find-
(Nathanson, 2001). Beyens and Valkenburg (2019) undertook ings in the built environment that parents monitor girls more
a similar task looking at Restrictive and Active Mediation, than boys (e.g., Doty et al., 2020). Older children and their
albeit through correlations and t-tests, finding that parent and parents reported less Restriction than younger children and
child report on the frequency of media parenting are corre- their parents. This age difference in Restrictive Mediation
lated, with parents reporting significantly more Active and lends credence to study findings, as decreased Restriction
Restrictive behaviors than youth. Nikken and Jansz (2006) also with maturity is developmentally appropriate and empiri-
found that parent and child report of media parenting with cally supported (e.g., Nathanson, 2001).
regards to video gaming were highly aligned. Our findings
were similar, with moderate to high inter-rater correlations for Latent Mean Comparisons
individual media parenting items, providing support for reli-
able response patterns between parents and children. In accord- Scaled scores for each factor reveal that both parents and
ance with prior research, parents report higher levels of media children are reporting Active Mediation more than the
parenting than children, particularly in the domain of Active other forms of mediation, followed by Monitoring, Restric-
Mediation (Beyens & Valkenburg, 2019). On a factor level, tion, and Technology Control. This mirrors results found
Active Mediation items were least correlated between parent by Padilla-Walker et al. (2018) that Active Mediation was
and child reporters and Restrictive and Technology Control employed more frequently than Restriction and Co-use.
were most correlated. This, in conjunction with the signifi- Given Active Mediation of youth media use is most strongly
cant independent samples t-test between reporters on Active related to reducing negative impacts and promoting positive
Mediation, is consistent with studies that have shown that, outcomes (Fujioka & Austin, 2003), this finding is hearten-
despite parents’ perception of active engagement, children are ing. That being said, the potential for social desirability bias
not always interpreting their actions as such (e.g., Valkenburg for the Active Mediation questions may be higher than that
et al., 2013; Warren, 2020). These comparisons do not, of the other categories. Thus, future studies should seek to

13
Journal of Psychopathology and Behavioral Assessment

replicate these findings with different measures in order to used in a reactive or proactive fashion, as has been suggested
determine whether Active Mediation is employed most com- (Wisniewski et al., 2015). Finally, the study reports on the
monly. It has also been argued that the frequency of different four primary domains of media parenting as identified by prior
parenting behaviors is less important than the parenting style research as the most important for youth outcomes. That being
they are employed with, which also remains important for said, there may be other domains that we did not address that
future research to explore (e.g., Valkenburg et al., 2013). are similarly important, such as Co-view and Co-use.
The low endorsement of Technology Control may be due,
in part, to the impact of such methods on the parent–child Conclusion
relationship; some parental control apps involve the covert
monitoring of youth activity, which may compromise trust This study provides evidence for construct validity through
(Ghosh et al., 2017). There is also the possibility that par- well-fitting CFA results for both parent and child measure-
ents do not feel as tech savvy as their children (Livingstone ment models of the media parenting construct. As might
et al., 2019), affecting their ability to confidently utilize have been expected from prior literature, differences in latent
such tools. Accordingly, Ghosh et al. (2018a) found that the mean comparisons across the reporters emerged. Namely,
more youth and their parents used the internet (e.g., screen parents reported higher levels of every parenting behavior
time), the more likely parents were to employ Technology than children, and parents/youth indicated differential paren-
Controls. Further, Eastin et al. (2006) identified a relation- tal employment of mediation strategies for boys and girls
ship between access to technology and screen time. These and younger and older children. These overall latent mean
findings, when taken together, suggest that the less access a differences identified between media parenting domains may
family has to technology, the less likely they will be to use be important for youth outcomes and provide support for
Technology Controls. This may also explain the lack of cor- their inclusion as distinct factors in predictive models. Iden-
relation between Technology Control and Restriction, which tification of a well-fitting measurement model contributes
we may have otherwise expected given the conceptual simi- to a common understanding of media parenting behaviors
larities between them. While access to technology is inher- and enables synchronized assessment of such behaviors’
ent in Technology Control, access is often removed or lim- relationship with youth outcomes. Further, although previ-
ited in Restriction, suggesting that these techniques may be ous research has identified parental mediation as predictive
targeting different populations of technology-using youths. of decreased online and in-person/built world risk (e.g.,
Additionally, there may be an element of parental education Bersamin et al., 2008; Cox et al., 2018; Elsaesser et al.,
involved, whereby parents who have the technological savvy 2017; Livingstone & Helsper, 2008; Nathanson, 1999), few
and resources to utilize Technology Controls may be differ- have used comprehensive validated measures as predictors
ent than parents who rely primarily on Restrictive Media- of youth outcomes, which would be a worthwhile next step.
tion alone. Future research that includes an evaluation of
parental technology knowledge would assist in clarifying
these associations. Declarations

Limitations Conflict of Interest We have no known conflict of interest to disclose.

Limitations are also worth noting. Although representing a

similar racial/ethnic breakdown to that of the United States References
on the whole, the present study had lower-than-average Latinx
Abar, C. C., Jackson, K. M., Colby, S. M., & Barnett, N. P. (2015). Parent–
participation (U.S. Census Bureau, 2019). Given the American
child discrepancies in reports of parental monitoring and their
population, the study cannot generalize to other countries and relationship to adolescent alcohol-related behaviors. Journal of
cultures. Another weakness of the study is that recruitment Youth and Adolescence, 44(9), 1688-1701
occurred through participant self-selection into Qualtrics pan- Anderson, M., & Jiang, J. (2018). Teens, Social Media & Technology
2018. https://​www.​pewre​search.​org/​inter​net/​wp-​conte​nt/​uploa​ds/​
els rather than through community-, hospital-, or school-based sites/9/​2018/​05/​PI_​2018.​05.​31_​Teens​Tech_​TOPLI​NE.​pdf
methods. Thus, given Qualtrics’ employment of internet-based Austin, E. W. (1993). Exploring the effects of active parental media-
recruitment, it may be that the sample represents a portion tion of television content. Journal of Broadcasting & Electronic
of the population with greater technological savvy. The study Media, 37(2), 147–158.
Austin, E. W., Bolls, P., Fujioka, Y., & Engelbertson, J. (1999). How
also utilized observational self-reported data; future research and why parents take on the tube. Journal of Broadcasting &
would benefit from utilization of behavioral observation or Electronic Media, 43(2), 175–192.
triangulation of sources. Furthermore, this research was Bandura, A. (1986). A Social Cognitive Theory. Prentice Hall.
based on cross-sectional data, and longitudinal exploration Bersamin, M., Todd, M., Fisher, D. A., Hill, D. L., Grube, J. W., &
Walker, S. (2008). Parenting practices and adolescent sexual
would also be valuable to assess whether such techniques are

13
 Journal of Psychopathology and Behavioral Assessment

behavior: A longitudinal study. Journal of Marriage and Family, Garmendia, M., Garitaonandia, C., Martinez, G., & Casado del Río,
70(1), 97–112. M. A. (2012). The effectiveness of parental mediation. In S.
Beyens, I., & Valkenburg, P. M. (2019). Parental Media Mediation Livingstone, L. Haddon, & A. Görzig (eds.), Children, risk and
in Adolescence: A Comparative Study of Parent and Adolescent safety online: Research and policy challenges in comparative
Reports. Journal of Broadcasting & Electronic Media, 63(4), perspective (pp. 231–244). The Policy Press.
716–736. Gentile, D. A., Nathanson, A. I., Rasmussen, E. E., Reimer, R. A.,
Buijzen, M., Rozendaal, E., Moorman, M., & Tanis, M. (2008). Parent & Walsh, D. A. (2012). Do you see what I see? Parent and
versus child reports of parental advertising mediation: Exploring child reports of parental monitoring of media. Family Relations,
the meaning of agreement. Journal of Broadcasting & Electronic 61(3), 470–487.
Media, 52(4), 509–525. George, D., & Mallery, M. (2003). Using SPSS for Windows step by
Bybee, C. R., Robinson, D., & Turow, J. (1982). Determinants of step: a simple guide and reference.
parental guidance of children’s television viewing for a special Ghosh, A. K., Badillo-Urquiola, K., Rosson, M. B., Xu, H., Carroll,
subgroup: Mass media scholars. Journal of Broadcasting & Elec- J. M., & Wisniewski, P. J. (2018a). A matter of control or safety?
tronic Media, 26(3), 697–710. Examining parental use of technical monitoring apps on teens'
Byrne, B. M. (2013). Structural equation modeling with Mplus: Basic mobile devices. Proceedings of the 2018 CHI Conference on
concepts, applications, and programming. Routledge. Human Factors in Computing Systems. (pp. 1-14)
Campbell, S. W., & Park, Y. J. (2014). Predictors of mobile sexting Ghosh, A. K., Badillo-Urquiola, K., & Wisniewski, P. (2018b). Exam-
among teens: Toward a new explanatory framework. Mobile ining the Effects of Parenting Styles on Offline and Online Ado-
Media & Communication, 2(1), 20–39. lescent Peer Problems. Proceedings of the 2018 ACM Conference
Chassiakos, Y. L. R., Radesky, J., Christakis, D., Moreno, M. A., & on Supporting Groupwork. (pp-150-153)
Cross, C. (2016). Children and adolescents and digital media. Ghosh, A. K., Badillo-Urquiola, K. A., Xu, H., Rosson, M. B., Carroll,
Pediatrics, 138(5), e20162593. J. M., & Wisniewski, P. (2017). Examining parents' technical
Chen, L., & Shi, J. (2019). Reducing harm from media: A meta-analysis mediation of teens' mobile devices. Companion of the 2017 ACM
of parental mediation. Journalism & Mass Communication Quar- Conference on Computer Supported Cooperative Work and Social
terly, 96(1), 173–193. Computing. (pp. 179-182)
Cheung, G. W., & Rensvold, R. B. (2002). Evaluating goodness-of-fit Greenberg, B. S., Ericson, P., & Vlahos, M. (1972). A comparison of
indexes for testing measurement invariance. Structural Equation parental mediation behaviors for mothers and their children. Journal
Modeling, 9(2), 233–255. of Broadcasting, 565–572.
Clark, L. S. (2011). Parental Mediation Theory for the Digital Age. Greenberg, B. S., Rampoldi-Hnilo, L., & Mastro, D. (Eds.). (2001).
Communication Theory, 21(4), 323–343. https://​doi.​org/​10.​ The alphabet soup of the television ratings. Cresskill, NJ:
1111/j.​1468-​2885.​2011.​01391.x Hampton Press.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences Jago, R., Edwards, M. J., Urbanski, C. R., & Sebire, S. J. (2013). Gen-
(2nd ed.). Hillsdale, NJ, USA: Erbaum Press. eral and specific approaches to media parenting: a systematic
Cox, M. J., Gabrielli, J., Janssen, T., & Jackson, K. M. (2018). Paren- review of current measures, associations with screen-viewing,
tal restriction of movie viewing prospectively predicts adolescent and measurement implications. Childhood Obesity, 9(s1),
alcohol and marijuana initiation: Implications for media literacy S-51-S-72.
programs. Prevention Science, 19(7), 914–926. Jia, F., Moore, E. W. G., Kinai, R., Crowe, K. S., Schoemann, A. M., &
De Los Reyes, A. (2011). Introduction to the special section: More Little, T. D. (2014). Planned missing data designs with small sample
than measurement error: Discovering meaning behind informant sizes: How small is too small? International Journal of Behavioral
discrepancies in clinical assessments of children and adolescents. Development, 38(5), 435–452.
Journal of Clinical Child & Adolescent Psychology, 40(1), 1–9. Jiow, H. J., Lim, S. S., & Lin, J. (2017). Level up! Refreshing Paren-
Dorr, A., Kovaric, P., & Doubleday, C. (1989). Parent-child coview- tal Mediation Theory for Our Digital Media Landscape. Com-
ing of television. Journal of Broadcasting & Electronic Media, munication Theory, 27(3), 309–328. https://​doi.​org/​10.​1111/​
33(1), 35–51. comt.​12109
Doty, J. L., Lynne, S. D., Metz, A. S., Yourell, J. L., & Espelage, D. Johnson, D. R., & Young, R. (2011). Toward best practices in analyzing
L. (2020). Bullying Perpetration and Perceived Parental Monitor- datasets with missing data: Comparisons and recommendations. Jour-
ing: A Random Intercepts Cross-Lagged Panel Model. Youth & nal of Marriage and Family, 73(5), 926–945.
Society, 0044118X20938416. Ho, S., Lwin, M. O., Chen, L., & Chen, M. (2019). Development and
Dürager, A., & Sonck, N. (2014). Testing the reliability of scales validation of a parental social media mediation scale across child
on parental internet mediation. London: EU Kids Online, and parent samples. Internet Research.
LSE. https://​eprin​ts.​lse.​ac.​uk/​60220 Kerr, M., Stattin, H., & Burk, W. J. (2010). A reinterpretation of paren-
Eastin, M. S., Greenberg, B. S., & Hofschire, L. (2006). Parenting the tal monitoring in longitudinal perspective. Journal of Research on
Internet. Journal of Communication, 56(3), 486–504. Adolescence, 20(1), 39–64.
Elsaesser, C., Russell, B., Ohannessian, C. M., & Patton, D. Lenhart, A., Ling, R., Campbell, S., & Purcell, K. (2010). Teens and mobile
(2017). Parenting in a digital age: A review of parents' role in phones: Text messaging explodes as teens embrace it as the centerpiece
preventing adolescent cyberbullying. Aggression and Violent of their communication strategies with friends. Pew Internet & American
Behavior, 35, 62–72. Life Project.
Fujioka, Y., & Austin, E. W. (2002). The relationship of family com- Lin, C. A., & Atkin, D. J. (1989). Parental mediation and rulemaking
munication patterns to parental mediation styles. Communica- for adolescent use of television and VCRs. Journal of Broadcast-
tion Research, 29(6), 642–665. ing & Electronic Media, 33(1), 53–67.
Fujioka, Y., & Austin, E. W. (2003). The Implications of Vantage Little, T. D. (2013). Longitudinal structural equation modeling. Guil-
Point in Parental Mediation of Television and Child’s Attitudes ford Press.
Toward Drinking Alcohol. Journal of Broadcasting & Elec- Little, T. D., Jorgensen, T. D., Lang, K. M., & Moore, E. W. G. (2013).
tronic Media, 47(3), 418–434. On the Joys of Missing Data. Journal of Pediatric Psychology,
Gabrielli, J., Marsch, L., & Tanski, S. (2018). TECH parenting to 39(2), 151–162. https://​doi.​org/​10.​1093/​jpepsy/​jst048
promote effective media management. Pediatrics, 142(1)

13
Journal of Psychopathology and Behavioral Assessment

Little, T. D., & Rhemtulla, M. (2013). Planned missing data designs Rideout, V. J., Foehr, U. G., & Roberts, D. F. (2010). Generation M2:
for developmental researchers. Child Development Perspectives, Media in the lives of 8- to 18-year-olds. Henry J. Kaiser Family
7(4), 199–204. Foundation. https://​www.​kff.​org/​wp-​conte​nt/​uploa​ds/​2013/​04/​
Livingstone, S., Haddon, L., Görzig, A., & Ólafsson, K. (2011). EU 8010.​pdf
Kids Online II: FinalReport 2011. London: EU Kids Online, Lon- Rideout, V. J., & Robb, M. B. (2019). The common sense census:
don School of Economics & PoliticalScience. Available at: http://​ Media use by tweens and teens. Common Sense Media.
eprin​ts.​lse.​ac.​uk/​39351/ Sanders, W., Parent, J., Forehand, R., & Breslend, N. L. (2016). The
Livingstone, S., & Helsper, E. J. (2008). Parental mediation of chil- roles of general and technology-related parenting in managing
dren’s internet use. Journal of Broadcasting & Electronic Media, youth screen time. Journal of Family Psychology, 30(5), 641–646.
52(4), 581–599. Symons, K., Ponnet, K., Vanwesenbeeck, I., Walrave, M., & Van Ouytsel,
Livingstone, S., Stoilova, M., & Nandagiri, R. (2019). Children's J. (2020). Parent-child communication about internet use and
data and privacy online: growing up in a digital age: an evi- acceptance of parental authority. Journal of Broadcasting &
dence review. London School of Economics and Political Science, Electronic Media, 64(1), 1–19. https://d​ oi.o​ rg/1​ 0.1​ 080/0​ 88381​ 51.
Department of Media and Communications. ​2019.​16818​70
Livingstone, S., Ólafsson, K., Helsper, E. J., Lupiáñez-Villanueva, F., Tanski, S. E., Dal Cin, S., Stoolmiller, M., & Sargent, J. D. (2010).
Veltri, G. A., & Folkvord, F. (2017). Maximizing opportunities Parental R-rated movie restriction and early-onset alcohol use.
and minimizing risks for children online: The role of digital skills Journal of Studies on Alcohol and Drugs, 71(3), 452–459.
in emerging strategies of parental mediation. Journal of Commu- U.S. Census Bureau. (2019). Hispanic or Latino. Retrieved from [https://​
nication, 67(1), 82–105. data.​census.​gov/​cedsci/​profi​lemap​widget?​defau​ltVin​tage=​2018&​
Morawska, A. (2020). The effects of gendered parenting on child devel- geoID=​ 0 1000​ 0 0US&​ geoLv=​ 0 10&​ m etri​ c Form​ a t=​ p erce​ n t&​
opment outcomes: A systematic review. Clinical Child and Family metri​cTopic=​Hispa​nic or Latino&heightUS=525px].
Psychology Review, 23(4), 553–576. Valkenburg, P. M., Krcmar, M., Peeters, A. L., & Marseille, N. M.
Muthén, L. K., & Muthén, B. O. (2017). Mplus (pp. 507–518). Chap- (1999). Developing a scale to assess three styles of television
man and Hall/CRC. mediation:“Instructive mediation”,“Restrictive Mediation”, and
Nathanson, A. I. (1999). Identifying and explaining the relationship “social coviewing.” Journal of Broadcasting & Electronic Media,
between parental mediation and children’s aggression. Commu- 43(1), 52–66.
nication Research, 26(2), 124–143. Valkenburg, P. M., Piotrowski, J. T., Hermanns, J., & de Leeuw, R.
Nathanson, A. I. (2001). Mediation of children’s television viewing: (2013). Developing and Validating the Perceived Parental Media
Working toward conceptual clarity and common understanding. Mediation Scale: A Self-Determination Perspective. Human Com-
Annals of the International Communication Association, 25(1), munication Research, 39(4), 445–469. https://​doi.​org/​10.​1111/​
115–151. hcre.​12010
Nikken, P., & Jansz, J. (2003). Parental mediation of children's video Van der Voort, T. H. A., Nikken, P., & Van Lil, J. E. (1992). Replication:
game playing: A similar construct as television mediation. Determinants of parental guidance of children’s television viewing: A
In DiGRA Conference. Dutch replication study. Journal of Broadcasting & Electronic Media,
Nikken, P., & Jansz, J. (2006). Parental mediation of children’s vide- 36(1), 61–74.
ogame playing: A comparison of the reports by parents and chil- Warren, R. (2017). Multi-platform mediation: U.S. mothers’ and
dren. Learning, Media and Technology, 31(2), 181–202. fathers’ mediation of teens’ media use. Journal of Children and
Nikken, P., & Jansz, J. (2014). Developing scales to measure parental Media, 11(4), 485–500.
mediation of young children’s internet use. Learning, Media and Warren, R. (2020). Parental Mediation of Media Use and Effects. The
Technology, 39(2), 250–266. International Encyclopedia of Media Psychology, 1–11.
Padilla-Walker, L. M., Coyne, S. M., Fraser, A. M., Dyer, W. J., & Warren, R., Gerke, P., & Kelly, M. A. (2002). Is there enough time
Yorgason, J. B. (2012). Parents and adolescents growing up in on the clock? Parental involvement and mediation of children’s
the digital age: Latent growth curve analysis of proactive media television viewing. Journal of Broadcasting & Electronic Media,
monitoring. Journal of Adolescence, 35(5), 1153–1165. 46(1), 87–111.
Padilla-Walker, L. M., Coyne, S. M., Kroff, S. L., & Memmott-Elison, Wisniewski, P., Jia, H., Xu, H., Rosson, M. B., & Carroll, J. M. (2015).
M. K. (2018). The protective role of parental media monitoring “Preventative” vs. “Reactive” How parental mediation influences
style from early to late adolescence. Journal of Youth and Ado- teens' social media privacy behaviors. In Proceedings of the 18th
lescence, 47(2), 445–459. ACM conference on computer supported cooperative work &
Padilla-Walker, L. M., Stockdale, L. A., Son, D., Coyne, S. M., & social computing (pp. 302–316).
Stinnett, S. C. (2020). Associations between parental media moni- Wright, J. C., Huston, A. C., Murphy, K. C., & St. Peters, M., Piñon,
toring style, information management, and prosocial and aggres- M., Scantlin, R., & Kotler, J. (2001). The relations of early televi-
sive behaviors. Journal of Social and Personal Relationships, sion viewing to school readiness and vocabulary of children from
37(1), 180–200. low-income families: The early window project. Child Develop-
Phares, V., Lopez, E., Fields, S., Kamboukos, D., & Duhig, A. M. ment, 72(5), 1347–1366.
(2005). Are fathers involved in pediatric psychology research
and treatment? Journal of Pediatric Psychology, 30(8), 631–643. Publisher's Note Springer Nature remains neutral with regard to
https://​doi.​org/​10.​1093/​jpepsy/​jsi050 jurisdictional claims in published maps and institutional affiliations.

13