Abstract
Background/Objectives: Internet Gaming Disorder (IGD)-related symptoms are a concern for children’s digital mental health, yet evidence on digital parenting as a family-context correlate of these symptoms is limited. This study examined the concurrent associations of total digital parenting and its two operational dimensions, Digital Knowledge and Technical Competence (DKTC) and Digital Guidance, Protection, and Engagement (DGPE), with parent-reported IGD-related symptom severity. Methods: This cross-sectional survey included 285 parent reports from Riyadh, Saudi Arabia, each concerning one child aged 6–13 years (M = 8.76, SD = 1.32; 55.4% boys). Parents completed an adapted parent-report IGDS9-SF and a 15-item digital-parenting scale. Correlations and hierarchical regression with HC3 standard errors and bootstrap intervals were used. Results: Symptom severity was inversely associated with total digital parenting (r = −0.747), DKTC (r = −0.658), and DGPE (r = −0.710), all p < 0.001. The model containing the two dimensions accounted for 56.0% of the variance, nearly identical to the model containing the total score (55.9%); both dimensions showed unique inverse associations (β = −0.333 and −0.481, p < 0.001). Conclusions: Higher digital parenting was associated with lower parent-reported IGD-related symptom severity. Because the data were cross-sectional and reported by the same parent, causal or protective effects cannot be inferred.
1. Introduction
Internet gaming disorder (IGD) represents an important clinical and psychosocial concern in childhood and adolescence because it may be associated with significant impairment in personal and social functioning [1]. In the ICD-11, gaming disorder is characterized by impaired control over gaming, increasing priority given to gaming over other activities, and continuation or escalation of gaming despite negative consequences [2].
Children’s gaming behavior develops within a broader family context. A recent systematic review found that family factors associated with IGD among children and adolescents encompassed parental mediation, positive and poor parenting, family disharmony, and socioeconomic conditions. Parental knowledge and positive parenting emerged as potential protective factors, whereas findings for restrictive mediation were inconsistent [3]. Research on parental mediation of adolescents’ Internet use further indicates that parents may combine active and restrictive strategies and that their associations with risk awareness, autonomy, and self-regulation vary according to the form of mediation used [4]. Accordingly, parental knowledge, communication, supervision, and boundary setting provide a relevant framework for examining variation in children’s gaming-related symptoms. Restrictive practices, however, should not be assumed to be uniformly protective.
The broader family literature provides a consistent starting point. Reviews link gaming-related difficulties with lower family cohesion, conflict, weak communication, reduced warmth and support, ineffective monitoring, poorer attachment, and negative parental attitudes toward Internet use [5,6,7,8]. Individual studies similarly associate gaming problems with poorer family functioning and less favorable parent–child relationships [9,10].
More favorable parent–child relationships and appropriate supervision have been associated with fewer gaming problems, whereas harsh or inconsistent discipline has been associated with more problems [11]. Family processes are also relevant before late adolescence: maternal parenting style has been associated with IGD among primary-school children, and marital conflict has prospectively predicted subsequent IGD features partly through father–child attachment and self-esteem [12,13].
Taken together, this evidence situates gaming symptoms within a relational ecology, but it does not establish that broad family functioning and technology-specific parenting are interchangeable constructs [5,6,8].
Digital parenting refers to a multidimensional set of technology-related parental resources and practices used to understand, support, and regulate children’s activities in digital environments. It encompasses parents’ readiness to manage these environments, including their knowledge, attitudes, and self-efficacy, as well as mediation practices used to guide digital engagement, discuss online risks and opportunities, establish boundaries, and supervise children’s online activities [14,15,16].
Modecki et al. [17] documented substantial heterogeneity in the measurement of digital parenting and showed that the literature has emphasized co-use and close supervision more than positive uses of technology, bidirectional influence, differences among families, and support beyond parents.
Recent studies have examined digital-parenting awareness, attitudes, and digital self-efficacy as separately measurable constructs. Among parents of children with special needs, Altındağ Kumaş and Sardohan Yildirim [18] found significant positive associations among digital-parenting awareness, attitudes, and perceived self-efficacy. Similarly, Kalkim et al. [19] assessed digital-parenting awareness and digital self-efficacy as distinct constructs among parents of early adolescents and reported significant associations between awareness and selected self-efficacy dimensions, including digital literacy and digital communication. These findings indicate that the constructs are empirically related but do not establish that they constitute a single construct.
Digital competence may shape mediation without guaranteeing uniformly favorable outcomes. Livingstone et al. [20] found that enabling mediation was more common when parents or children had stronger digital skills, whereas restrictive mediation was more common when skills were lower. Enabling mediation coincided with more online opportunities but also greater exposure to risk; restriction coincided with fewer risks and fewer opportunities. The relevant issue is therefore not only whether parents possess digital skills, but how those skills are translated into context-sensitive decisions.
Associations of digital-parenting awareness with self-efficacy, digital literacy, communication, parental well-being, and sociodemographic characteristics support the conceptual relevance of parental digital competence [18,19,21]. However, these studies did not test whether technical competence had an independent association with children’s IGD symptoms. Its association with lower IGD-related symptom severity should therefore be tested rather than assumed.
The practice dimension draws on parental-mediation research. Parents commonly combine active discussion and co-use with rules, technical restrictions, and monitoring rather than relying on a single strategy [22]. Jiow et al. [23] argued that mediation theory must be adapted to interactive media and described restrictive, active, and co-use strategies as operating through gatekeeping, discussion, diversion, and investigation. Digital guidance may therefore extend beyond rule enforcement to include explanation, joint engagement, information seeking, safety responses, and direction toward alternatives.
Protection-related awareness may be modifiable: a parent-training intervention produced maintained improvement in protection-from-risk scores [24]. The broader evidence, however, cautions against ranking practices without context. Some restrictive strategies have been associated with less problematic gaming, whereas active mediation has not been consistently protective [25]. Restriction may also reduce beneficial opportunities or arise reactively after problems become visible.
Longitudinal findings reinforce this caution. Higher parental monitoring predicted lower subsequent IGD symptoms, but IGD symptoms also predicted later reductions in monitoring; father–child relationship quality participated in these reciprocal pathways [26]. Koning et al. [27] also reported selected bidirectional associations, but not a uniformly protective effect of Internet-specific parenting. More frequent parent–adolescent communication about Internet use predicted higher subsequent IGD symptoms among boys, whereas earlier IGD symptoms predicted subsequent changes in reactive rules and in the frequency or quality of communication. Other prospective evidence is more differentiated. After adjustment for background variables, baseline parental support predicted fewer subsequent IGD symptoms, whereas parental abuse predicted more symptoms. Parental psychological control and a positive parent–adolescent relationship did not retain significant prospective effects, and baseline IGD symptoms did not predict parental factors at follow-up [28]. By contrast, Su et al. [26] and Koning et al. [27] reported reciprocal longitudinal associations between parenting-related factors and gaming-related difficulties. These findings caution against a universal one-way account: parenting practices may precede changes in gaming-related symptoms in some contexts, whereas existing symptoms may shape later parental responses in others; relationship quality and unmeasured conditions may influence both processes.
Distinguishing behavioral from psychological control is more informative than treating parental control as unitary. Turan et al. [29] described behavioral control as supervision and rule setting, whereas psychological control involves intrusive practices, including manipulation of the parent–child relationship, guilt induction, criticism, and excessive personal control. In a two-year longitudinal study, parental psychological control was indirectly associated with subsequent IGD through adolescents’ affiliation with deviant peers [30]. In a cross-sectional study, core self-evaluation partially mediated the association between parental psychological control and IGD tendency [31]. Turan et al. [29] also reported associations between perceived psychological-control indicators and IGD risk among Turkish adolescents, although the pattern differed by parent and family characteristics.
Parental knowledge may function as a protective condition within a specific peer-victimization context: the positive association between peer victimization and IGD was stronger among adolescents reporting lower parental knowledge, and this moderated association involved basic psychological-needs satisfaction [32]. Longitudinally, parental involvement was indirectly associated with fewer subsequent IGD symptoms through consideration of future consequences, with peer victimization moderating this association [33]. Cross-sectionally, emotional warmth was associated with IGD through time-management tendency [34].
Taken together, these findings suggest that parenting-related factors may be connected with gaming outcomes through relational, peer-related, and intrapersonal processes. They also caution against conflating parental knowledge, behavioral supervision, or guidance with coercive psychological control, because these constructs differ conceptually and have been examined using distinct measures and explanatory pathways [29,30,31,32,33,34].
Evidence focused specifically on digital parenting and gaming is more directly relevant but remains heterogeneous in its constructs, developmental periods, and outcomes. Among young children, active, supportive, and explanatory mediation, together with some structured restrictive practices, was associated with a lower tendency toward digital-game addiction, whereas restrictive-blocker role did not show the same inverse association [35]. In preschool children, dimensions of maternal digital-parenting awareness were associated with children’s digital-game-addiction tendency, whereas maternal digital literacy was associated with several awareness dimensions rather than directly with the overall addiction-tendency score [36]. In early childhood, digital-parenting awareness was inversely associated with digital-game addiction and positively associated with parent–child relationship quality and children’s social competence [37].
These studies strengthen the case for examining technology-specific parenting in childhood, but their outcomes are not equivalent to DSM-based IGD symptom severity. They also leave unresolved whether competence is directly associated with gaming symptoms, operates indirectly through enacted practices, or is relevant mainly under particular family conditions [35,36,37].
Broader parenting studies offer convergent rather than construct-equivalent evidence. Neglectful parental attitudes have been associated with the highest digital-game-addiction levels, whereas authoritarian attitudes have been more strongly related to social anxiety; gaming duration, child sex, and parental control have also contributed to statistical prediction [38]. Positive parent–child interaction and support have predicted changes in gaming-related problems [10], but family cohesion and conflict do not specify the technology-related decisions captured by digital parenting.
Accordingly, an association of a warm family climate or positive parent–child interaction with fewer gaming-related problems cannot directly validate parents’ digital knowledge or technical competence [10,38]. Evidence concerning psychological control likewise should not be generalized to supportive digital guidance, because psychological control and behavioral guidance are conceptually distinct forms of parental influence [29].
A transactional perspective is more directly relevant to the parent–child processes examined here. Longitudinal findings indicate that associations between parenting and gaming-related difficulties may operate in both directions: parental monitoring or Internet-specific practices may precede changes in gaming-related symptoms, whereas emerging symptoms may be followed by changes in parental monitoring or restriction [26,27].
Three gaps motivate the present study. First, much of the available evidence concerns adolescents, broad family functioning, general parenting styles, psychological control, or isolated mediation strategies rather than multidimensional, technology-specific parenting among children [3,5,8]. Second, the literature employs heterogeneous constructs and non-equivalent outcomes, moving among IGD, problematic gaming, digital-game addiction, and addiction tendency; this limits direct comparison. Third, evidence concerning guidance, monitoring, and parent–child relationship processes is more extensive than evidence testing whether parents’ digital knowledge and technical competence account for unique variation in DSM-based, parent-reported IGD symptom severity. Studies of digital-parenting awareness and self-efficacy clarify relationships among parental digital resources but do not establish their unique associations with IGD symptoms [17,18,19,20].
The literature therefore supports examining both parental digital resources and enacted practices, but it does not justify assuming that the dimensions have the same direction or magnitude of association with parent-reported IGD symptom severity.
The present cross-sectional study addresses these gaps by examining the concurrent associations of total Digital Parenting and its two operational dimensions—Digital Knowledge and Technical Competence and Digital Guidance, Protection, and Engagement—with parent-reported IGD-related symptom severity. It further examines whether each dimension is uniquely associated with symptom severity after child age and sex are considered, and whether separating the dimensions adds explanatory value beyond the total score.
Research Questions
- What are the descriptive statistics for total Digital Parenting, its two dimensions, and parent-reported IGD-related symptom severity?
- How are digital parenting (total and dimension scores) associated with parent-reported IGD-related symptom severity?
- Do the two digital-parenting dimensions account for additional variance in symptom severity beyond child age and sex?
- What are the unique associations of each dimension, child age, and child sex with symptom severity?
2. Materials and Methods
2.1. Study Design
This study employed a cross-sectional correlational design to examine the concurrent associations between digital parenting and parent-reported Internet Gaming Disorder (IGD)-related symptom severity among children. It also examined the unique statistical contribution of two digital-parenting dimensions—Digital Knowledge and Technical Competence and Digital Guidance, Protection, and Engagement—after accounting for child age and sex. Because all variables were measured at a single time point, the findings do not establish temporal precedence or causality.
2.2. Participants and Statistical Power
Two separate and non-overlapping samples were collected during the period covered by the institutional ethical approval. The first was an independent model-development sample (N = 310), comprising 174 reports completed by fathers and 136 completed by mothers, and was used exclusively for exploratory factor analyses of the two study measures. The second was a separate cross-validation sample (N = 304), used for the initial ordinal confirmatory factor analyses and as the source from which the age-restricted analytic sample was derived. No parent or child records overlapped between the two samples.
The N = 304 cross-validation sample was a non-probability convenience sample of parent reports from Riyadh, Saudi Arabia. It comprised reports concerning 167 boys and 137 girls aged 4–13 years. The electronic questionnaire initially permitted reports concerning children aged 4–13 years; no IGD-specific lower-age boundary had been specified a priori. All 304 reports satisfied the original data-collection eligibility criteria.
After data collection and completion of the confirmatory factor analyses in the full cross-validation sample (N = 304), but before the focal bivariate associations or regression models were examined, the analytic lower age limit was revised from four to six years. This decision restricted hypothesis testing to school-aged children because several IGD criteria—including deception about gaming, loss of interest in previous activities, and gaming to escape negative moods—presuppose behavioral autonomy and self-regulatory capacities that are less applicable to, and less readily observed in, preschool-aged children. The decision was not based on the CFA findings, scale scores, observed associations, or statistical significance. Consequently, 19 reports concerning children aged 4 or 5 years (6.3% of the full cross-validation sample) were excluded. These reports concerned nine boys and ten girls; therefore, the exclusion did not disproportionately remove children of either sex.
In an exploratory comparison, the excluded reports had higher IGD-related symptom scores than the retained reports (M = 27.68, SD = 7.52 versus M = 22.02, SD = 9.86; Welch’s t(22.34) = 3.11, p = 0.005; Hedges’ g = 0.58). Digital-parenting scores and child sex did not differ between the excluded and retained reports, but all 19 excluded questionnaires were completed by fathers. Because of this respondent-role imbalance, respondent role was examined in supplementary sensitivity analyses. When respondent role (father versus mother) was added as a covariate to the final regression model, it was not statistically significant (p = 0.22), and both digital-parenting dimensions retained significant inverse associations with parent-reported IGD-related symptom severity. The inverse association between total Digital Parenting and symptom severity was also similar for reports completed by mothers (r = −0.760) and fathers (r = −0.735).
The questionnaire link was distributed to parents through participating schools and parent networks without deliberately targeting fathers of children aged 4–5 years. One possible explanation for the exclusive representation of fathers among the 19 excluded reports is that the survey link may have circulated through parent communication groups comprising predominantly or exclusively fathers. However, the specific distribution pathways could not be verified, and this explanation remains speculative. Accordingly, the observed pattern is regarded as an unexplained recruitment imbalance rather than evidence of an intentional sampling strategy.
The final age-restricted analytic sample comprised 285 parent reports, each concerning one child aged 6–13 years (M = 8.76, SD = 1.32), including 158 boys (55.4%) and 127 girls (44.6%). The respondents comprised 150 mothers (52.6%) and 135 fathers (47.4%). Accordingly, the initial confirmatory factor analyses were conducted in the full cross-validation sample (N = 304), and the same measurement models were subsequently re-estimated in the age-restricted analytic sample (N = 285) to examine the stability of the measurement findings after application of the revised age criterion. The descriptive, correlational, and regression analyses addressing the research questions were also based on the age-restricted analytic sample (N = 285). The N = 285 sample was not an independent third sample; rather, it was an age-restricted subset of the N = 304 cross-validation sample.
The age distribution in the final analytic sample was uneven: 16 children were aged 6 years, 12 were aged 7 years, 122 were aged 8 years, 30 were aged 9 years, 96 were aged 10 years, none were aged 11 years, six were aged 12 years, and three were aged 13 years. This uneven distribution reflects the non-probability convenience recruitment through participating schools and parent networks rather than any age-based selection within the 6–13-year range.
For the initial data collection, respondents were eligible if they: (a) resided in Riyadh; (b) were the parent or primary caregiver of a child aged 4–13 years; (c) had sufficient knowledge of the child’s digital activities and gaming behavior to complete the study measures; (d) reported that the child had engaged in Internet or digital gaming during the preceding 12 months; (e) reported that the child had no diagnosed developmental disability; and (f) provided electronic informed consent. For inclusion in the present hypothesis-testing analyses, the child was additionally required to be aged 6–13 years. Responses were excluded if the respondent was not the child’s parent or primary caregiver, the child did not meet the applicable eligibility criteria, electronic informed consent was not provided, or the questionnaire contained substantial missing or invalid data. No observations were excluded on the basis of high or low scale scores.
IGD was examined as parent-reported symptom severity rather than as a clinical diagnosis. Accordingly, the scores represent the extent of IGD-related symptoms reported by parents and should not be interpreted as establishing a formal diagnosis.
A sensitivity power analysis based on the noncentral F distribution was conducted for the overall four-predictor regression model using the achieved analytic sample of 285 participants and α = 0.05. The analysis indicated that the sample provided 80% power to detect an effect of Cohen’s f2 = 0.043, 90% power to detect f2 = 0.055, and 95% power to detect f2 = 0.066. Thus, the achieved sample was adequately powered to detect small effects of approximately f2 = 0.043 or greater with 80% power, but it was not sufficiently powered to detect an effect as small as f2 = 0.02 with 80% power.
2.3. Data Collection
Data were collected electronically in Riyadh, Saudi Arabia, from late June through July 2026, following institutional ethical approval. The questionnaire link was distributed to prospective participants through participating schools and parent networks. Before accessing the questionnaire, parents received information about the study’s purpose, initial eligibility requirements, voluntary nature, confidentiality protections, and their right to discontinue participation without penalty. Electronic informed consent was obtained before respondents proceeded to the questionnaire. Each parent was instructed to report on one child who met the initial data-collection eligibility criteria, and no directly identifying information was collected.
Following data collection, submitted responses were screened for completeness, eligibility under the initial data-collection criteria, and allowable response ranges. The revised analytic age criterion described in the Participants and Statistical Power section was then applied before the focal bivariate associations or regression models were examined. No observations were excluded on the basis of scale scores or subsequently observed statistical results. Each row in the analytic dataset represented one submitted questionnaire. The final analytic dataset contained no missing values for the variables included in the analyses, and all item responses fell within the intended range of 1–5.
2.4. Measures
2.4.1. Parent-Report Internet Gaming Disorder Scale–Short Form
Internet Gaming Disorder-related symptom severity was assessed using an Arabic parent-report adaptation of the Internet Gaming Disorder Scale–Short Form (IGDS9-SF) [39]. The original English-language instrument assesses the severity of the nine DSM-5 IGD criteria. Translation and cultural adaptation followed a back-translation procedure. Two specialists in special education and psychology independently translated the items into Arabic, and the research team reconciled the two versions. A bilingual native English speaker with an advanced qualification in English, who had not seen the original instrument, then back-translated the Arabic version. Comparison with the original identified semantic discrepancies, after which minor culturally and educationally appropriate wording adjustments were made. Twelve specialists reviewed the final Arabic version for content relevance and linguistic clarity. Because some self-report items concern internal experiences requiring introspection, the wording was adapted so that parents reported on the child’s gaming-related behaviors and experiences while preserving the substantive criterion content. The adaptation targeted children aged 6–13 years. No diagnostic cutoff or clinical classification was applied because the reporting source had changed and no clinically validated threshold was available for this Arabic parent-report adaptation.
A parent-report format was selected because the study focused on children aged 6–13 years, whose reading comprehension, introspective capacity, and ability to estimate the frequency and consequences of their gaming behavior may vary developmentally. Parents were therefore asked to evaluate observable gaming-related behaviors and consequences based on their knowledge of the child’s daily activities. Nevertheless, parental reports may be less sensitive to internal experiences or gaming activities occurring outside parental observation. Accordingly, the resulting scores were consistently interpreted as indicators of parent-reported IGD-related symptom severity rather than as clinical diagnoses.
Each item was rated for the preceding 12 months on a five-point response scale: 1 (never), 2 (rarely), 3 (sometimes), 4 (often), and 5 (always). The nine item scores were summed to produce a total score ranging from 9 to 45, with higher scores indicating greater parent-reported IGD-related symptom severity. The total score was treated as a continuous variable, and no diagnostic cutoff or clinical classification was applied.
In an independent model-development sample (N = 310), the item matrix demonstrated good suitability for factor analysis, KMO = 0.883, and Bartlett’s test of sphericity was statistically significant, χ2(36) = 1214.08, p < 0.001. Principal-axis exploratory factor analysis with parallel analysis supported the retention of one factor: the first observed eigenvalue (4.64) exceeded the corresponding random eigenvalue (1.34), whereas the second observed eigenvalue (0.90) was lower than its corresponding random eigenvalue (1.23). The first eigenvalue from the unreduced correlation matrix corresponded to 51.60% of the total item variance, whereas the factor extracted using principal-axis factoring accounted for 45.76% of the total item variance. Factor loadings ranged from 0.557 to 0.780.
The unidimensional structure was subsequently examined using ordinal confirmatory factor analysis with the WLSMV estimator (diagonally weighted least squares with mean- and variance-adjusted test statistics) in a second, independent cross-validation sample (N = 304), from which the present age-restricted analytic sample was drawn. The one-factor model showed acceptable comparative and residual-based fit, scaled χ2(27) = 246.29, p < 0.001, scaled CFI = 0.972, scaled TLI = 0.963, and SRMR = 0.061, whereas scaled RMSEA was elevated (0.164, 90% CI [0.145, 0.183]). Because fit-index cutoffs developed for continuous data under maximum likelihood do not transfer directly to ordinal data estimated with WLSMV, and the indices can lead to different conclusions in this setting [40], fit was judged across indices rather than against a single cutoff. Standardized factor loadings ranged from 0.767 to 0.916, and all nine items were retained. Composite reliability and average variance extracted were satisfactory (CR = 0.952; AVE = 0.689), providing favorable reliability and extracted-variance estimates for the specified model in the cross-validation sample. The same model was re-estimated in the age-restricted analytic sample (N = 285) with essentially unchanged results: scaled χ2(27) = 245.52, scaled CFI = 0.971, scaled TLI = 0.962, SRMR = 0.063, and scaled RMSEA = 0.169; loadings ranged from 0.781 to 0.909 (CR = 0.955; AVE = 0.701). Given the elevated RMSEA, these results were interpreted as supporting, rather than definitively confirming, the unidimensional structure.
The scale demonstrated excellent internal consistency (Cronbach’s α = 0.938; McDonald’s ω = 0.939).
2.4.2. Digital Parenting Skills
Digital parenting was assessed using a 15-item Arabic parent-report scale developed for this study. Item development was guided by the study’s operational definition of digital parenting, which distinguishes parents’ digital knowledge and technical competence from the practices through which they guide, protect, and engage with their children’s digital activities. The initial items were formulated to represent these content domains and were reviewed for relevance, clarity, and correspondence with the intended construct before administration. The content-based allocation of items to the two proposed dimensions is a working specification of this newly developed measure and warrants replication and external content-validity evaluation.
In an independent model-development sample (N = 310), the item matrix demonstrated excellent suitability for factor analysis, KMO = 0.946, and Bartlett’s test of sphericity was statistically significant, χ2(105) = 3716.33, p < 0.001. Principal-axis exploratory factor analysis with parallel analysis supported the retention of one dominant general factor. The first eigenvalue from the unreduced correlation matrix corresponded to 60.52% of the total item variance, whereas the factor extracted using principal-axis factoring accounted for 57.72% of the total item variance. Factor loadings ranged from 0.683 to 0.817. Thus, the exploratory analysis supported a broad digital-parenting construct but did not, by itself, establish sharply separated dimensions.
A theoretically specified correlated two-factor model was subsequently examined using ordinal confirmatory factor analysis with the same WLSMV estimator in the same independent cross-validation sample (N = 304), from which the present age-restricted analytic sample was drawn. The two-factor model showed acceptable comparative and residual-based fit, scaled χ2(89) = 662.95, p < 0.001, scaled CFI = 0.949, scaled TLI = 0.940, and SRMR = 0.071, with an elevated scaled RMSEA (0.146, 90% CI [0.136, 0.156]), and the scaled chi-square difference test favored it over the one-factor alternative, scaled Δχ2(1) = 110.27, p < 0.001. This comparison indicates relative improvement only; the elevated scaled RMSEA continues to qualify the model’s absolute fit. Standardized loadings ranged from 0.520 to 0.871 for Digital Knowledge and Technical Competence and from 0.772 to 0.923 for Digital Guidance, Protection, and Engagement. Composite reliability and average variance extracted were satisfactory for Digital Knowledge and Technical Competence (CR = 0.880; AVE = 0.601) and Digital Guidance, Protection, and Engagement (CR = 0.957; AVE = 0.690). The latent factor correlation (r = 0.816) was compatible with substantial overlap between the theoretically specified dimensions. A bifactor model with one general and two specific factors fitted better, scaled χ2(75) = 444.43, scaled CFI = 0.967, scaled TLI = 0.954, SRMR = 0.049, and scaled RMSEA = 0.128, and indicated a strong general factor (ωH = 0.884; explained common variance = 0.792), supporting the total Digital Parenting score as the primary indicator. Re-estimation in the age-restricted analytic sample (N = 285) yielded essentially unchanged results: the two-factor model showed scaled CFI = 0.947, scaled TLI = 0.937, SRMR = 0.073, and scaled RMSEA = 0.151, again fitted better than the one-factor model (scaled Δχ2(1) = 111.37, p < 0.001), and the latent correlation was 0.807; the bifactor model again indicated a strong general factor (ωH = 0.892). These labels describe the operationalization and factor structure of the measure developed for this study and are not proposed as a universally established taxonomy of digital parenting.
The Digital Knowledge and Technical Competence dimension comprised five items assessing parents’ knowledge of digital tools and environments, awareness of digital opportunities and risks, and perceived capacity to respond to technology-related demands. The five items were summed to produce a score ranging from 5 to 25. The Digital Guidance, Protection, and Engagement dimension comprised 10 items assessing enacted parental practices, including discussing and guiding digital activities, establishing boundaries, protecting children from digital risks, supervising use, and remaining constructively involved in children’s digital activities. The 10 items were summed to produce a score ranging from 10 to 50.
All items were rated on a five-point Likert-type agreement scale: 1 (strongly disagree), 2 (disagree), 3 (neutral), 4 (agree), and 5 (strongly agree). A total Digital Parenting score was calculated by summing responses across all 15 items, producing a possible range of 15–75. Higher dimension and total scores indicated higher reported levels of the corresponding digital-parenting resources and practices. The total score was used in the descriptive and bivariate analyses because the exploratory findings supported a broad general construct. The two dimension scores were entered separately into the primary regression model to estimate their unique concurrent associations with parent-reported IGD-related symptom severity. The total score was not entered alongside its component dimensions because doing so would introduce mathematical redundancy and multicollinearity.
Internal consistency was satisfactory for Digital Knowledge and Technical Competence (Cronbach’s α = 0.844; McDonald’s ω = 0.846); excellent for Digital Guidance, Protection, and Engagement (α = 0.928; ω = 0.930); and excellent for the total Digital Parenting score (α = 0.934; ω = 0.936).
2.4.3. Demographic and Control Variables
Parents reported the child’s age and sex. Child age was recorded in completed years and entered as a continuous developmental covariate. Child sex was coded as 1 = male and 2 = female in the source dataset and recoded as 0 = male and 1 = female for regression analysis. Age and sex were included as control variables to estimate the associations of the two digital-parenting dimensions with parent-reported IGD-related symptom severity after accounting for potential demographic variation. Parents also reported their educational attainment and their level of electronic-gaming use; these variables were included only in the additional robustness analysis described below.
2.4.4. Data Analysis
Primary analyses were conducted in Python 3.12.14 (Python Software Foundation, Beaverton, OR, USA) using pandas 2.2.3 (NumFOCUS, Inc., Austin, TX, USA), NumPy 2.3.5 (NumFOCUS, Inc., Austin, TX, USA), SciPy 1.17.0 (NumFOCUS, Inc., Austin, TX, USA), and statsmodels 0.15.0 (NumFOCUS, Inc., Austin, TX, USA). Ordinal confirmatory factor analyses were estimated in R 4.3.3 (R Core Team, R Foundation for Statistical Computing, Vienna, Austria) using lavaan 0.6-17 (Yves Rosseel, Ghent University, Ghent, Belgium) with the WLSMV estimator. Data management, score computation, statistical estimation, and bootstrap resampling were implemented using reproducible analysis syntax. Frequencies and percentages were calculated for categorical variables, and means, standard deviations, and observed ranges were calculated for continuous variables. Internal consistency was evaluated using Cronbach’s α and McDonald’s ω. All code was executed on the original study data, and the resulting outputs and interpretations were reviewed and verified by the authors.
Pearson product–moment correlations were used for associations among continuous variables. Because child sex was coded as 0 = male and 1 = female, its associations with continuous variables were estimated using point-biserial correlations. The correlation analysis included child age and sex, parent-reported IGD-related symptom severity, the two digital-parenting dimensions, and the total Digital Parenting score.
Hierarchical ordinary least-squares regression was used to examine the unique concurrent associations of the two digital-parenting dimensions with parent-reported IGD-related symptom severity. Child age and sex were entered at Step 1, followed simultaneously by Digital Knowledge and Technical Competence and Digital Guidance, Protection, and Engagement at Step 2. The total Digital Parenting score was excluded from the regression because it was mathematically composed of the two dimension scores. Incremental concurrent variance was summarized using ΔR2 and the conventional F-change statistic. Because the latter assumes homoscedasticity, it was treated as a model-comparison summary rather than as the primary basis for coefficient-level inference.
Coefficient-level inferences were based on HC3 heteroscedasticity-consistent standard errors and percentile confidence intervals obtained from 5000 case-resampled bootstrap samples. Unstandardized coefficients (B), standardized coefficients (β), HC3 standard errors, p values, and 95% confidence intervals were reported. All tests were two-tailed, with α = 0.05.
Regression diagnostics included variance inflation factors and tolerance values, the Durbin–Watson statistic, the Breusch–Pagan test, the Shapiro–Wilk test, the Ramsey RESET test, and visual inspection of residual-versus-fitted and normal Q–Q plots. Standardized residuals, leverage values, and Cook’s distances were also examined. Diagnostic findings informed the use of robust inferential procedures but were not applied as automatic criteria for excluding observations.
To evaluate whether separating the two digital-parenting dimensions added explanatory information beyond the total score, an additional model containing child age, child sex, and the total Digital Parenting score was compared with the primary model containing age, sex, and both dimension scores. Because the total score is the sum of the two dimensions, this comparison evaluated whether allowing separate dimension slopes improved model fit. The change in R2 and conventional F-change were reported as model summaries, and an HC3-robust Wald test evaluated the equality constraint on the two unstandardized dimension slopes imposed by the total-score model. Squared semipartial correlations were calculated from the decrease in R2 after removing each dimension from the final model. A separate HC3-robust contrast was used to examine whether the two standardized dimension coefficients differed statistically.
An additional robustness model was estimated to examine whether the focal associations remained after adjustment for parental educational attainment and parental electronic-gaming-use level, which were the relevant parental covariates available in the dataset. Both variables were represented categorically, with bachelor’s degree and moderate electronic-gaming use serving as the reference categories. Child age and sex were retained in the model, and coefficient-level inferences continued to use HC3 heteroscedasticity-consistent standard errors and 5000-sample case-bootstrap confidence intervals.
Because all 19 questionnaires excluded under the revised age criterion had been completed by fathers, a supplementary sensitivity analysis added respondent role (coded 0 = mother and 1 = father) as a categorical covariate to the final regression model. Correlations between total Digital Parenting and parent-reported IGD-related symptom severity were also estimated separately for reports completed by fathers and mothers. Coefficient-level inference for the respondent-role-adjusted sensitivity model used HC3 heteroscedasticity-consistent standard errors.
To examine whether the focal associations varied with child age without imposing an arbitrary developmental cutoff, an exploratory moderation model added mean-centered age-by-Digital Knowledge and Technical Competence and age-by-Digital Guidance, Protection, and Engagement interaction terms to the primary model. The incremental change in R2 was treated as a model summary, and joint inference for the two interaction terms was based on an HC3-robust Wald test.
Two supplementary diagnostics were conducted in response to concerns about common-method bias and item-content proximity. First, a Harman-type unrotated principal-components diagnostic was applied to all 24 study items; this diagnostic was treated as descriptive and not as evidence capable of ruling out common-method bias. Second, a content-based sensitivity model recalculated the digital-parenting dimensions after removing DP2, DP3, DP10, and DP15, the items most explicitly referring to monitoring applications or games, time limits, discussion of Internet activities, and participation in children’s digital activities. The sensitivity model retained child age and sex and used HC3-robust coefficient inference.
3. Results
3.1. Descriptive Statistics and Score Reliability
The first research question examined the descriptive characteristics of overall digital parenting, its two dimensions, and parent-reported Internet Gaming Disorder (IGD)-related symptom severity. Table 1 presents the descriptive statistics and internal-consistency estimates for the study measures.
Table 1.
Descriptive Statistics and Reliability Estimates.
Among the 285 children aged 6–13 years, the mean parent-reported IGD-related symptom-severity score was 22.02 (SD = 9.86), within a possible range of 9–45. The observed scores ranged from 9 to 44, and the corresponding mean item score was 2.45 on the five-point response scale.
For digital parenting, the mean Digital Knowledge and Technical Competence score was 19.22 (SD = 4.02; observed range = 5–25), corresponding to a mean item score of 3.84. The mean Digital Guidance, Protection, and Engagement score was 40.35 (SD = 7.09; observed range = 20–50), corresponding to a mean item score of 4.04. The mean total Digital Parenting score was 59.57 (SD = 10.28; observed range = 29–75), with a mean item score of 3.97. Because no validated categorical thresholds were specified for these measures, the scores were interpreted continuously rather than classified as low, moderate, or high.
All study scores demonstrated good-to-excellent internal consistency. Cronbach’s α and McDonald’s ω were 0.938 and 0.939, respectively, for IGD-related symptom severity; 0.844 and 0.846 for Digital Knowledge and Technical Competence; 0.928 and 0.930 for Digital Guidance, Protection, and Engagement; and 0.934 and 0.936 for total Digital Parenting.
3.2. Bivariate Associations
The second research question examined the bivariate associations of overall digital parenting and its two dimensions with parent-reported IGD-related symptom severity. As shown in Table 2, IGD-related symptom severity was strongly and inversely associated with Digital Knowledge and Technical Competence, r = −0.658, p < 0.001, and Digital Guidance, Protection, and Engagement, r = −0.710, p < 0.001. Total Digital Parenting was also strongly and inversely associated with IGD-related symptom severity, r = −0.747, p < 0.001.
Table 2.
Correlations Among the Study Variables.
The two digital-parenting dimensions were positively correlated, r = 0.690, p < 0.001, indicating that they were related but not identical components of digital parenting. Correlations involving the total Digital Parenting score and its component dimensions were not interpreted as evidence of discriminant validity because these scores overlap mathematically.
A small inverse point-biserial association was also observed between female sex and Digital Knowledge and Technical Competence, rpb = −0.156, p = 0.008. This association was not a primary study outcome and is reported descriptively.
Neither child age, r = 0.014, p = 0.814, nor child sex, rpb = 0.012, p = 0.840, was significantly associated with IGD-related symptom severity. Because sex was binary-coded, correlations between sex and continuous variables are reported as point-biserial correlations, although they are computationally equivalent to Pearson correlations.
3.3. Incremental Variance Explained by Digital Parenting
The third research question examined whether the two digital-parenting dimensions accounted for additional variance in parent-reported IGD-related symptom severity after controlling for child age and sex. The hierarchical regression results are presented in Table 3.
Table 3.
Hierarchical Regression Model Summary.
At Step 1, child age and sex collectively explained less than 0.1% of the variance in IGD-related symptom severity, R2 < 0.001, adjusted R2 = −0.007. The control model was not statistically significant, F(2, 282) = 0.04, p = 0.957.
Adding Digital Knowledge and Technical Competence and Digital Guidance, Protection, and Engagement at Step 2 increased the proportion of concurrently observed variance accounted for by 56.0 percentage points, ΔR2 = 0.560. The conventional model-comparison test indicated a statistically significant improvement, ΔF(2, 280) = 178.22, p < 0.001. The final four-predictor model accounted for 56.0% of the concurrent variance in parent-reported IGD-related symptom severity, R2 = 0.560, adjusted R2 = 0.554, conventional F(4, 280) = 89.16, p < 0.001.
Because heteroscedasticity was detected, the conventional overall and change-in-F statistics were treated as model-summary indices. Primary coefficient-level conclusions were based on HC3 heteroscedasticity-consistent standard errors and bootstrap confidence intervals.
For comparison, the model containing age, sex, and the total Digital Parenting score accounted for 55.9% of the variance, R2 = 0.559, adjusted R2 = 0.555. Allowing separate slopes for the two dimensions increased R2 by less than 0.1 percentage point, ΔR2 = 0.001, and did not significantly improve conventional model fit, ΔF(1, 280) = 0.55, p = 0.459. The corresponding HC3-robust Wald test of the equal-unstandardized-slope constraint was also nonsignificant, χ2(1) = 0.32, p = 0.572. Thus, separating the dimensions did not materially increase overall explained variance relative to the total score; its value lay in estimating the dimensions’ unique associations.
3.4. Unique Associations with IGD-Related Symptom Severity
The fourth research question examined the unique associations of the two digital-parenting dimensions, child age, and child sex with parent-reported IGD-related symptom severity. The HC3-robust regression coefficients are reported in Table 4.
Table 4.
Final Regression Model for Parent-Reported IGD-Related Symptom Severity.
After adjustment for child age, sex, and the other digital-parenting dimension, Digital Knowledge and Technical Competence had a statistically significant inverse association with IGD-related symptom severity, B = −0.817, HC3 SE = 0.176, β = −0.333, p < 0.001, 95% CI [−1.164, −0.470]. Its percentile bootstrap interval was also entirely below zero, 95% bootstrap CI [−1.192, −0.514].
Removing Digital Knowledge and Technical Competence from the final model reduced R2 by 0.055, corresponding to sr2 = 0.055.
Digital Guidance, Protection, and Engagement similarly demonstrated a statistically significant inverse association with IGD-related symptom severity, B = −0.669, HC3 SE = 0.092, β = −0.481, p < 0.001, 95% CI [−0.851, −0.488]. The bootstrap result supported the same conclusion, 95% bootstrap CI [−0.834, −0.477]. Removing this dimension from the final model reduced R2 by 0.118, corresponding to sr2 = 0.118. Although its standardized coefficient was numerically larger, an HC3-robust Wald test did not show a statistically significant difference between the two standardized coefficients, χ2(1) = 1.22, p = 0.270. The results therefore do not establish that one dimension was consistently stronger than the other.
Neither child age, B = −0.191, β = −0.026, p = 0.465, nor child sex, B = −0.737, β = −0.037, p = 0.346, made a statistically significant unique contribution. The HC3 and bootstrap confidence intervals for both demographic covariates included zero.
3.5. Regression Diagnostics and Robustness
Multicollinearity was not problematic: variance inflation factors ranged from 1.006 to 2.004, and tolerance values ranged from 0.499 to 0.994. The Durbin–Watson statistic was 1.962, providing no indication of problematic residual autocorrelation. The Ramsey RESET test was nonsignificant, F = 0.55, p = 0.577, providing no evidence of major functional-form misspecification.
The Breusch–Pagan test was significant, LM = 19.47, p < 0.001, and the residuals-versus-fitted-values plot showed nonconstant residual dispersion. These findings supported the use of HC3 standard errors and bootstrap confidence intervals. The Shapiro–Wilk test was also significant, W = 0.984, p = 0.003, while the Q–Q plot indicated limited departures from normality, particularly in the distributional tails. No standardized residual exceeded an absolute value of 3. Nine observations exceeded the conservative Cook’s distance screening benchmark of 4/N = 0.014; however, the maximum Cook’s distance was 0.078, which remained well below 1. These observations were therefore treated as potentially influential screening cases rather than automatically excluded. Because the HC3 and bootstrap results were concordant and the study protocol retained all eligible cases, all 285 observations remained in the final analysis.
3.6. Additional Covariate-Adjusted Robustness Analysis
After additional adjustment for parental educational attainment and electronic-gaming-use level, the model accounted for 62.6% of the variance (R2 = 0.626), and the two dimensions added 40.3% beyond the covariates (ΔR2 = 0.403, p < 0.001). Both Digital Knowledge and Technical Competence (B = −0.875, HC3 SE = 0.162, β = −0.357, p < 0.001) and Digital Guidance, Protection, and Engagement (B = −0.538, HC3 SE = 0.086, β = −0.387, p < 0.001) retained significant inverse associations.
3.7. Developmental and Method-Related Sensitivity Analyses
Because all 19 excluded questionnaires had been completed by fathers, respondent role was examined in a supplementary sensitivity analysis. When respondent role (father versus mother) was added to the final regression model, it was not statistically significant, B = 1.00, p = 0.22, and both digital-parenting dimensions retained significant inverse associations with parent-reported IGD-related symptom severity. The inverse correlations between total Digital Parenting and symptom severity were descriptively similar for reports completed by mothers (r = −0.760) and fathers (r = −0.735). Thus, the focal associations were not materially altered after accounting for respondent role.
Adding the two age-interaction terms increased R2 from 0.560 to 0.571 (HC3-robust joint Wald χ2(2) = 15.14, p < 0.001). Only the age-by-Digital Guidance, Protection, and Engagement interaction was significant (B = 0.158, HC3 SE = 0.052, p = 0.003), indicating a weaker inverse association at older ages; given the uneven age distribution, this finding is exploratory.
In the Harman-type diagnostic, the first unrotated component accounted for 51.2% of the variance across the 24 items, exceeding the 50% heuristic; common-method bias therefore cannot be ruled out. After excluding DP2, DP3, DP10, and DP15, both reduced dimension scores remained inversely associated with symptom severity (DKTC: B = −1.520, HC3 SE = 0.254, β = −0.365, p < 0.001; DGPE: B = −0.800, HC3 SE = 0.099, β = −0.464, p < 0.001; R2 = 0.577), indicating that the focal associations did not depend on these items.
4. Discussion
This study examined the concurrent associations of digital parenting, assessed as a total score and as two dimensions, with parent-reported Internet Gaming Disorder (IGD)-related symptom severity among children aged 6–13 years in Riyadh. Overall digital parenting, Digital Knowledge and Technical Competence (DKTC), and Digital Guidance, Protection, and Engagement (DGPE) were strongly and inversely associated with symptom severity. After child age and sex were considered, DKTC and DGPE accounted for an additional 56.0 percentage points of the concurrently observed variance, and each dimension retained a significant unique inverse association. Thus, both parental digital resources and enacted digital-parenting practices were statistically relevant to symptom severity within the present sample. These findings represent concurrent associations and do not establish temporal precedence or causality.
For comparison, a model containing age, sex, and the total Digital Parenting score accounted for 55.9% of the variance, nearly identical to the 56.0% accounted for by the model containing separate dimension scores. Most of the explanatory information was therefore shared across the dimensions, and their separation did not materially improve overall model performance. The dimensional model remains informative for estimating unique concurrent associations, but it should not be presented as substantially more explanatory than the total score.
The measurement evidence qualifies the dimension-specific findings. The correlated two-factor digital-parenting model fitted better than the one-factor alternative in the cross-validation sample and in the analytic sample drawn from it, but the high latent correlation (r = 0.816) and the near-equivalence of the total-score and dimension-score regression models indicate substantial shared variance. The two dimension scores are therefore interpreted as theoretically specified, strongly overlapping operational components rather than definitively established distinct constructs; their value in the present analyses lies in estimating unique concurrent associations. A bifactor model nonetheless indicated a strong general factor, supporting the total score as the primary indicator. The limitations of the factor-analytic evidence are considered in the Limitations section.
The mean item score for parent-reported IGD-related symptom severity was 2.45. Because no validated diagnostic cutoff was applied to the adapted parent-report version, this value should not be interpreted as a prevalence estimate or assigned a categorical severity level. The IGDS9-SF was developed as a brief nine-item self-report measure based on the nine DSM-5 criteria [39]. In the present study, scores from the adapted parent-report version were interpreted as indicators of symptom severity rather than clinical diagnoses. The mean item scores for DKTC (3.84) and DGPE (4.04) likewise describe the positions of the observed scores on the five-point response scale rather than validated categorical levels. The relatively high DGPE mean may be compatible with favorable parental self-presentation or restricted variability near the upper end of the scale, but social desirability and ceiling effects were not formally assessed.
The direction of the inverse associations is broadly consistent with longitudinal evidence concerning related parental practices. Su et al. [26] found that greater parental monitoring at the first assessment predicted lower subsequent IGD symptoms, whereas higher symptoms at the second assessment predicted lower parental monitoring at the third. She et al. [28] reported differentiated longitudinal associations between specific parental factors and subsequent IGD symptoms. Dou et al. [33] found that greater parental involvement was prospectively associated with fewer IGD symptoms through consideration of future consequences, with the pathway differing according to adolescents’ levels of peer victimization. These studies examined parental constructs related to, but not equivalent to, DKTC and DGPE and primarily involved adolescents rather than children with a mean age of 8.76 years. They therefore provide directional context but cannot be generalized directly to the present sample.
Evidence involving younger children is also consistent in direction, although it has generally examined digital-game-addiction tendency rather than DSM-based IGD symptom severity. Among children aged 3–6 years, Çalhan and Göksu [35] found inverse associations between digital-game-addiction tendency and five parental media-mediation roles, whereas the restrictive-blocker role was not significantly associated with the outcome. Türen and Bağçeli Kahraman [36] reported significant relationships between maternal digital-parenting awareness and preschool children’s digital-game-addiction tendencies, as well as between maternal digital literacy and digital-parenting awareness. Kay and Sağlam [37] reported a significant inverse path between digital-parenting awareness and early-childhood digital-game addiction in their structural model. Differences in developmental period, construct definition, measurement, and outcome terminology nevertheless limit direct comparison with the present findings.
The findings also address the identified gap concerning the unique contribution of parental digital knowledge and competence. DKTC retained a significant inverse association after DGPE, child age, and sex were considered, and DGPE likewise retained a significant unique inverse association. Within the present operationalization, DKTC represents parental digital resources, whereas DGPE represents reported enacted practices; however, the cross-sectional analysis did not test whether knowledge and competence were translated into guidance, protection, or engagement. Livingstone et al. [20] linked parental and child digital skills to enabling and restrictive mediation strategies, whereas Jiow et al. [23] identified gatekeeping, discussion, diversion, and investigation as activities underlying parental mediation of video-game play. Together, these studies are compatible with a conceptual distinction between digital resources and mediation practices, but they do not validate the specific two-factor structure used here. DGPE had a numerically larger standardized coefficient than DKTC, but the HC3-robust Wald test did not indicate a statistically significant difference between the coefficients, χ2(1) = 1.22, p = 0.270. The result therefore does not establish that DGPE was statistically more important than DKTC.
Parental involvement should not be assumed to be uniformly associated with fewer gaming-related symptoms. Koning et al. [27] reported selected bidirectional associations between Internet-specific parenting practices and IGD symptoms that varied according to the parenting practice and the adolescent’s sex. Benrazavi et al. [25] reported that associations differed across parental-mediation strategies; some restrictive strategies were associated with lower problematic online gaming, whereas active mediation was not consistently supported. Because Benrazavi et al. studied older participants and problematic online gaming rather than parent-reported IGD symptom severity, their findings are not directly comparable with those of the present study.
The longitudinal evidence is compatible with a transactional interpretation in which parenting may precede changes in gaming behavior while children’s symptoms may also modify some parental responses. Su et al. [26] and Koning et al. [27] reported selected pathways in both directions. Reciprocal relationships are therefore plausible but not uniformly supported, and the present data cannot determine whether digital-parenting resources and practices preceded symptoms, developed in response to them, or operated reciprocally. The Interaction of Person–Affect–Cognition–Execution model describes addictive behaviors as emerging through interactions among personal predispositions, affective and cognitive responses, coping, reward expectancies, and executive control [41]. Digital parenting may be considered part of the broader environmental context surrounding children’s gaming, but the present study did not assess the mechanisms specified by the model and therefore does not test them.
The magnitude of the observed associations requires particular caution. The correlation between overall digital parenting and symptom severity was strong (r = −0.747). Because both principal constructs were reported by the same parent at the same time, their association may have been influenced by common-source perceptions, shared response tendencies, or conceptual proximity between parental monitoring practices and parents’ perceptions of gaming-related difficulties. Consequently, the 56.0% value represents concurrent statistical overlap in parent-reported continuous symptom scores rather than the proportion of clinically diagnosed IGD or symptoms caused or prevented by digital parenting. The Harman-type diagnostic indicated that the first unrotated component accounted for 51.2% of item variance, which exceeded the conventional 50% heuristic threshold and therefore did not alleviate the common-method concern. Although the inverse associations remained significant after removing four digital-parenting items selected for their potential content proximity to monitoring and management of digital use, this sensitivity analysis could not separate substantive covariance from shared-source inflation.
Child sex was not statistically significant in the adjusted model. Previous studies reported greater gaming-related problems among boys [35,38], but this pattern was not reproduced in the present sample. Child age also had no significant linear main effect; however, the exploratory moderation analysis indicated a significant joint age interaction. The inverse association of Digital Guidance, Protection, and Engagement became weaker as age increased, whereas the age interaction involving Digital Knowledge and Technical Competence did not reach individual statistical significance. Because the age distribution was concentrated at ages 8 and 10 and included very few children aged 12–13 years, these moderation findings should be treated as provisional and should not be interpreted as establishing uniform developmental-stage differences.
The study provides locally situated evidence that parental digital resources and enacted practices were associated with parents’ reports of children’s gaming-related symptoms in Riyadh. This pattern is directionally consistent with Saudi evidence showing that lower family harmony and the absence of parental restriction or supervision were associated with higher problematic Internet use among adolescents [42]. Similarly, stronger parent–adolescent relationships were associated with lower problematic Internet use, whereas authoritarian fathering was associated with higher problematic Internet use among Saudi adolescents [43]. These comparisons should be interpreted cautiously because problematic Internet use is conceptually broader than IGD-related symptom severity, and the cited studies examined adolescents rather than predominantly younger children. Moreover, cultural values were not measured and no cross-cultural comparison was conducted; therefore, the present associations cannot be attributed specifically to Saudi culture. The findings identify parental digital competence and enacted guidance as potentially relevant domains for future parent-oriented research and program development. However, intervention studies are required before concluding that strengthening either domain reduces children’s gaming-related symptoms. Supportive digital guidance should also be distinguished from coercive psychological control, which has been associated with IGD symptoms, risk, or tendency, depending on the study [29,30,31]. The healthcare relevance of the present findings lies in identifying family-context correlates for future research on child digital mental health and family-centered assessment, rather than in establishing diagnostic, preventive, or treatment effects.
Limitations
Several limitations define the scope of the findings. First, the cross-sectional, single-informant, single-method design prevents determination of temporal direction and may be affected by shared-method variance. No marker variable, child report, independent behavioral indicator, or additional caregiver report was available to separate substantive from shared-method covariance, and the Harman-type diagnostic cannot quantify or rule out common-method bias. Although respondent role was not incorporated into the primary analysis, supplementary sensitivity analyses adjusted for father-versus-mother respondent role and estimated the total-score correlation separately by respondent role; the focal associations were not materially altered. These analyses were supplementary and descriptive, however, and reports from other caregivers were unavailable.
Second, the IGDS9-SF was developed as a self-report measure [39] and was translated into Arabic and reformulated for parent report; changes in language, developmental target, and reporting source may affect item interpretation, and parents may have limited access to children’s internal experiences or to gaming outside their observation. The reported factor structure was estimated in the full cross-validation sample (N = 304) and re-estimated, as a descriptive sensitivity check, in its nested age-restricted analytic subset (N = 285); although comparative and residual-based fit indices were acceptable, RMSEA values were elevated, indicating some local misfit. The total score should therefore be interpreted as an operational composite indicator rather than as evidence that unidimensionality was confirmed. Criterion-related validity against children’s self-reports, clinical assessment, or independently observed gaming behavior was not examined.
Third, digital parenting was self-reported by parents. The relatively high DGPE mean may partly reflect social desirability or ceiling effects, neither of which was formally assessed. Conceptual proximity between reported monitoring practices and parental perceptions of children’s gaming difficulties may also have contributed to the strong associations; the associations remained significant in the content-based sensitivity model, but this does not exclude shared-source inflation. Future studies should combine parent reports with child reports, behavioral indicators, gaming records, or reports from additional caregivers.
Fourth, the non-probability convenience sample from Riyadh limits generalizability. The lower age limit was revised from four to six years after data collection, and this decision was not preregistered. This post-hoc modification of the analytic eligibility criteria was not specified a priori and represents a methodological deviation from the original study plan. Although theoretically motivated and implemented before the focal association and regression analyses, this decision may have introduced selection bias. The 19 excluded reports had higher IGD-related symptom scores than the retained reports and were all completed by fathers. Although respondent role was not statistically significant in the supplementary adjusted analysis and the total-score correlations were descriptively similar for reports completed by mothers and fathers, selection effects related to the age restriction cannot be excluded. Within the retained sample, the age distribution was highly uneven, with most children concentrated at ages 8 and 10, no children aged 11, and very few children aged 12–13 years. This imbalance limits the generalizability of the findings across the full 6–13-year age range and constrains the interpretation of the exploratory age-moderation analysis. The findings should therefore not be generalized to preschool-aged children.
Finally, several potentially relevant variables were unavailable, including child gaming duration and frequency, parental age, socioeconomic status, parents’ general digital-media use, family functioning, and peer-related factors; residual confounding therefore remains possible despite adjustment for parental education and electronic-gaming-use level. The high latent correlation between the digital-parenting dimensions, and the content of DP2 and DP3, which resemble parental management practices, indicate substantial overlap between the component scores; their factor structure and measurement invariance across developmental groups require evaluation in independent samples with more evenly distributed ages.
5. Conclusions
Total digital parenting and both of its operational dimensions were inversely associated with parent-reported IGD-related symptom severity among children aged 6–13 years. After child age and sex were considered, DKTC and DGPE were each uniquely associated with symptom severity, and both associations remained statistically significant after additional adjustment for parental education and electronic-gaming-use level. However, the model containing the total Digital Parenting score accounted for nearly the same concurrent variance as the model containing the two dimensions (55.9% versus 56.0%), and the dimensions were highly correlated; the total score therefore captures most of the explanatory information, whereas the dimensions add component-specific information, and neither dimension was shown to be stronger than the other. Because the associations were cross-sectional and based on reports from the same parent, they do not establish that digital parenting prevents or reduces IGD-related symptoms. Longitudinal, multi-informant, and developmentally stratified research is required to clarify temporal direction, reduce shared-method bias, and determine whether changes in digital-parenting resources and practices precede changes in children’s gaming-related symptoms. These conclusions pertain to observed parent-report composite scores.
Author Contributions
Conceptualization, S.R.S. and M.M.E.; methodology, S.R.S. and M.M.E.; formal analysis, A.N.A.; investigation, A.N.A.; resources, A.N.A.; data curation, S.R.S. and A.N.A.; writing—original draft, M.M.E.; writing—review & editing, M.M.E.; supervision, M.M.E.; funding acquisition, S.R.S. All authors have read and agreed to the published version of the manuscript.
Funding
This research project was funded by the Deanship of Scientific Research and Libraries, Princess Nourah bint Abdulrahman University, through the Pioneer Researcher Funding Initiative, Grant No. (PRFI-2026).
Institutional Review Board Statement
The study was conducted in accordance with the Declaration of Helsinki, and approved by the Institutional Review Board of Princess Nourah bint Abdulrahman University, Riyadh, Saudi Arabia (IRB Log Number: 26-0647; Category of Approval: EXPEDITED-1; IRB Registration Number with KACST: HAP-01-R-059; date of approval: 22 June 2026).
Informed Consent Statement
Electronic informed consent was obtained from all participating parents or primary caregivers before they proceeded to the questionnaire.
Data Availability Statement
The data supporting the findings of this study are available from the corresponding author upon reasonable request and subject to ethical and institutional approval. The data are not publicly available because they contain information concerning minors and are subject to privacy and ethical restrictions.
Acknowledgments
During the preparation and revision of this manuscript, the authors used ChatGPT website version (OpenAI, San Francisco, CA, USA; accessed September–November 2026) to assist with English-language refinement, manuscript organization, and the generation and checking of portions of the analysis syntax. All analysis syntax, statistical outputs, and interpretations were reviewed and verified by the authors against the original study data and reproducible analysis files. The tool was not used to collect, generate, fabricate, or alter the study data. The authors take full responsibility for the accuracy, integrity, and final content of the manuscript.
Conflicts of Interest
The authors declare no conflicts of interest. The funder had no role in the study design; data collection, analysis, or interpretation; preparation of the manuscript; or the decision to submit the manuscript for publication.
References
- Paulus, F.W.; Ohmann, S.; von Gontard, A.; Popow, C. Internet gaming disorder in children and adolescents: A systematic review. Dev. Med. Child Neurol. 2018, 60, 645–659. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- World Health Organization. Addictive Behaviours: Gaming Disorder. Available online: https://www.who.int/news-room/questions-and-answers/item/addictive-behaviours-gaming-disorder (accessed on 25 September 2026).
- Petrescu, T.C.; Błachnio, A.; Enea, V. Internet gaming disorder in children and adolescents: A systematic review of familial protective and risk factors. Addict. Behav. 2025, 167, 108345. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Steinfeld, N. Parental mediation of adolescent Internet use: Combining strategies to promote awareness, autonomy and self-regulation in preparing youth for life on the web. Educ. Inf. Technol. 2021, 26, 1897–1920. [Google Scholar] [CrossRef] [Scilit]
- Atalan Ergin, D.; Essau, C.A. Family factors and internet gaming disorder among adolescents: A systematic review. Int. J. Dev. Sci. 2025, 19, 122–135. [Google Scholar] [CrossRef] [Scilit]
- Nielsen, P.; Favez, N.; Rigter, H. Parental and family factors associated with problematic gaming and problematic Internet use in adolescents: A systematic literature review. Curr. Addict. Rep. 2020, 7, 365–386. [Google Scholar] [CrossRef] [Scilit]
- Rosales-Navarro, C.; Torres Pérez, I. Parental behaviors and video game addiction in adolescents: A systematic review. Int. J. Ment. Health Addict. 2026, 24, 1660–1681. [Google Scholar] [CrossRef] [Scilit]
- Schneider, L.A.; King, D.L.; Delfabbro, P.H. Family factors in adolescent problematic Internet gaming: A systematic review. J. Behav. Addict. 2017, 6, 321–333. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Bonnaire, C.; Phan, O. Relationships between parental attitudes, family functioning and Internet gaming disorder in adolescents attending school. Psychiatry Res. 2017, 255, 104–110. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Li, A.Y.L.; Lo, B.C.-Y.; Cheng, C. It is the family context that matters: Concurrent and predictive effects of aspects of parent–child interaction on video gaming-related problems. Cyberpsychol. Behav. Soc. Netw. 2018, 21, 374–380. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Cuong, V.M.; Assanangkornchai, S.; Wichaidit, W.; Minh Hanh, V.T.; My Hanh, H.T. Associations between gaming disorder, parent–child relationship, parental supervision, and discipline styles: Findings from a school-based survey during the COVID-19 pandemic in Vietnam. J. Behav. Addict. 2021, 10, 722–730. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Areshtanab, H.N.; Fathollahpour, F.; Bostanabad, M.A.; Ebrahimi, H.; Hosseinzadeh, M.; Fooladi, M.M. Internet gaming disorder and its relationship with behavioral disorder and mother’s parenting styles in primary school students according to gender in Iran. BMC Psychol. 2021, 9, 110. [Google Scholar] [CrossRef] [Scilit]
- Jeong, H.; Yim, H.W.; Lee, S.Y.; Lee, H.K.; Potenza, M.N.; Jo, S.J.; Son, H.J. A partial mediation effect of father–child attachment and self-esteem between parental marital conflict and subsequent features of Internet gaming disorder in children: A 12-month follow-up study. BMC Public Health 2020, 20, 484. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Fidan, N.K.; Olur, B. Examining the relationship between parents’ digital parenting self-efficacy and digital parenting attitudes. Educ. Inf. Technol. 2023, 28, 15189–15204. [Google Scholar] [CrossRef] [Scilit]
- Tan, C.Y.; Pan, Q.; Tao, S.; Liang, Q.; Lan, M.; Feng, S.; Cheung, H.S.; Liu, D. Conceptualization, measurement, predictors, outcomes, and interventions in digital parenting research: A comprehensive umbrella review. Educ. Res. Rev. 2024, 45, 100647. [Google Scholar] [CrossRef] [Scilit]
- Zhao, P.; Bazarova, N.N.; Valle, N. Digital parenting divides: The role of parental capital and digital parenting readiness in parental digital mediation. J. Comput.-Mediat. Commun. 2023, 28, zmad032. [Google Scholar] [CrossRef] [Scilit]
- Modecki, K.L.; Goldberg, R.E.; Wisniewski, P.; Orben, A. What is digital parenting? A systematic review of past measurement and blueprint for the future. Perspect. Psychol. Sci. 2022, 17, 1673–1691. [Google Scholar] [CrossRef] [Scilit]
- Altındağ Kumaş, Ö.; Sardohan Yildirim, A.E. Exploring digital parenting awareness, self-efficacy and attitudes in families with special needs children. Br. J. Educ. Technol. 2024, 55, 2403–2418. [Google Scholar] [CrossRef] [Scilit]
- Kalkim, A.; Konal Korkmaz, E.; Uysal Toraman, A. Examining the relationship between digital parenting self-efficacy and digital parenting awareness of early adolescents’ parents. J. Pediatr. Nurs. 2024, 78, 1–6. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Livingstone, S.; Ólafsson, K.; Helsper, E.J.; Lupiáñez-Villanueva, F.; Veltri, G.A.; Folkvord, F. Maximizing opportunities and minimizing risks for children online: The role of digital skills in emerging strategies of parental mediation. J. Commun. 2017, 67, 82–105. [Google Scholar] [CrossRef] [Scilit]
- Pazarcıkcı, F.; Karataş, N.; Kaya, A. The relationships of parents’ mental well-being and sociodemographic characteristics with digital parenting awareness: Structural equation model analysis. J. Pediatr. Nurs. 2024, 75, e102–e111. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Livingstone, S.; Helsper, E.J. Parental mediation of children’s Internet use. J. Broadcast. Electron. Media 2008, 52, 581–599. [Google Scholar] [CrossRef] [Scilit]
- Jiow, H.J.; Lim, S.S.; Lin, J. Level up! Refreshing parental mediation theory for our digital media landscape. Commun. Theory 2017, 27, 309–328. [Google Scholar] [CrossRef] [Scilit]
- Canpolat, M.; Karadaş, C. A mixed method research on increasing digital parenting awareness of parents. Educ. Inf. Technol. 2024, 29, 6683–6704. [Google Scholar] [CrossRef] [Scilit]
- Benrazavi, R.; Teimouri, M.; Griffiths, M.D. Utility of parental mediation model on youth’s problematic online gaming. Int. J. Ment. Health Addict. 2015, 13, 712–727. [Google Scholar] [CrossRef] [Scilit]
- Su, B.; Yu, C.; Zhang, W.; Su, Q.; Zhu, J.; Jiang, Y. Father–child longitudinal relationship: Parental monitoring and Internet gaming disorder in Chinese adolescents. Front. Psychol. 2018, 9, 95. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Koning, I.M.; Peeters, M.; Finkenauer, C.; van den Eijnden, R.J.J.M. Bidirectional effects of Internet-specific parenting practices and compulsive social media and Internet game use. J. Behav. Addict. 2018, 7, 624–632. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- She, R.; Zhang, Y.; Yang, X. Parental factors associated with Internet gaming disorder among first-year high school students: Longitudinal study. JMIR Serious Games 2022, 10, e33806. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Turan, N.; Alkan, A.; Çekiç, Y. Perceived parental psychological control and the risk of internet gaming disorder in adolescents: A cross-sectional study. Curr. Psychol. 2024, 43, 21735–21744. [Google Scholar] [CrossRef] [Scilit]
- Lin, S.; Yu, C.; Chen, J.; Sheng, J.; Hu, Y.; Zhong, L. The association between parental psychological control, deviant peer affiliation, and Internet gaming disorder among Chinese adolescents: A two-year longitudinal study. Int. J. Environ. Res. Public Health 2020, 17, 8197. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Ji, Z.; Wei, S.; Ding, H. Parental psychological control and internet gaming disorder tendency: A moderated mediation model of core self-evaluation and intentional self-regulation. Int. J. Ment. Health Promot. 2024, 26, 547–558. [Google Scholar] [CrossRef] [Scilit]
- Liang, Q.; Yu, C.; Xing, Q.; Liu, Q.; Chen, P. The influence of parental knowledge and basic psychological needs satisfaction on peer victimization and Internet gaming disorder among Chinese adolescents: A mediated moderation model. Int. J. Environ. Res. Public Health 2021, 18, 2397. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Dou, K.; Feng, X.-K.; Wang, L.-X.; Li, J.-B. Longitudinal association between parental involvement and Internet gaming disorder among Chinese adolescents: Consideration of future consequences as a mediator and peer victimization as a moderator. J. Behav. Addict. 2022, 11, 820–830. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Chen, I.-H.; Lee, Z.-H.; Dong, X.-Y.; Gamble, J.H.; Feng, H.-W. The influence of parenting style and time management tendency on internet gaming disorder among adolescents. Int. J. Environ. Res. Public Health 2020, 17, 9120. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Çalhan, C.; Göksu, İ. An effort to understand parents’ media mediation roles and early childhood children’s digital game addiction tendency: A descriptive correlational survey study. Educ. Inf. Technol. 2024, 29, 17825–17865. [Google Scholar] [CrossRef] [Scilit]
- Türen, Ş.; Bağçeli Kahraman, P. The predictive relationships between children’s digital game addiction tendencies and mothers’ digital parenting awareness and digital literacy levels. Educ. Inf. Technol. 2025, 30, 3115–3144. [Google Scholar] [CrossRef] [Scilit]
- Kay, M.A.; Sağlam, M. As predictives of digital game addiction in early childhood: Digital parenting, family relations and social competence. Ann. Med.-Psychol. 2025, 183, 429–438. [Google Scholar] [CrossRef] [Scilit]
- Geniş, Ç.; Ayaz-Alkaya, S. Digital game addiction, social anxiety, and parental attitudes in adolescents: A cross-sectional study. Child. Youth Serv. Rev. 2023, 149, 106931. [Google Scholar] [CrossRef] [Scilit]
- Pontes, H.M.; Griffiths, M.D. Measuring DSM-5 Internet gaming disorder: Development and validation of a short psychometric scale. Comput. Hum. Behav. 2015, 45, 137–143. [Google Scholar] [CrossRef] [Scilit]
- Xia, Y.; Yang, Y. RMSEA, CFI, and TLI in structural equation modeling with ordered categorical data: The story they tell depends on the estimation methods. Behav. Res. Methods 2019, 51, 409–428. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Brand, M.; Wegmann, E.; Stark, R.; Müller, A.; Wölfling, K.; Robbins, T.W.; Potenza, M.N. The Interaction of Person–Affect–Cognition–Execution (I-PACE) model for addictive behaviors: Update, generalization to addictive behaviors beyond Internet-use disorders, and specification of the process character of addictive behaviors. Neurosci. Biobehav. Rev. 2019, 104, 1–10. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Saquib, N.; Saquib, J.; AlSalhi, A.; Colder Carras, M.; Labrique, A.B.; Al-Khani, A.M.; Chamsi Basha, A.; Almazrou, A. The associations between family characteristics and problematic Internet use among adolescents in Saudi Arabia. Int. J. Adolesc. Youth 2023, 28, 2256826. [Google Scholar] [CrossRef] [Scilit]
- Abusaleh, H.; Güloğlu, B. Problematic Internet use among adolescents in Saudi: Parenting style and parent-adolescent relationship. Turk. Psychol. Couns. Guid. J. 2025, 15, 584–609. [Google Scholar] [CrossRef] [Scilit]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2026 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license.