The launch of the iPhone in 2007 has made the use of digital devices (smartphones, tablets, computers, and stream TVs) increasingly popular among young children. A recent review showed that the average percentage of digital overuse among young children reported in previous studies was 48.34% (Wang et al., 2023). However, young children are vulnerable to the negative or side effects of increased screen time caused by using various digital devices with screen; thus, their prolonged use of digital devices, including the long-term or high-frequency use of various digital devices, has become a global concern and public health issue (Dong et al., 2020; Wang et al., 2023). The existing reviews have revealed the association between prolonged digital use in young children and their well-being and development, including behavioral development (Madigan et al., 2019), language development (Zimmerman et al., 2007), sleep quality (Thompson & Christakis, 2005), and executive function performance (Huber et al., 2018). Given the fact that parents play a significant role in early child development, the most important period of behavioral development, there is a need for synthesized evidence to show what parental risk factors contribute to the prolonged use of digital devices in preschoolers and what factors moderate this impact. The findings will provide scientific evidence for parental education, program development, and policy-making. Thus, it will have a far-reaching social impact. To address this urgent need, this study will provide a meta-analysis of the parental factors and related moderators for preschoolers’ prolonged digital use.

1 Parental factors contributing to prolonged digital use

Parents are the primary caregivers of young children and play a crucial role in their early childhood development, including their digital use. Previous studies have investigated the impact of parents on their children’s digital use and identified four major contributors: parents’ digital use, psychological distress, socioeconomic status (SES), and parental behavior. First, several studies have examined the specific role of parents’ digital use in preschoolers’ digital use and reported mixed results (Barr-Anderson et al., 2011; Gao et al., 2022; Hu et al., 2018; Poulain et al., 2019). While some studies have reported a positive impact (Barr-Anderson et al., 2011), others have reported a negative impact (Gao et al., 2022).

Second, many studies have also explored the impact of parental psychological stress, including parents’ anxiety (Ribner & McHarg, 2021), depression (Stienwandt et al., 2022), and parenting stress (Cost et al., 2020). However, the results are inconsistent and the effect sizes differ between studies (Cost et al., 2020; Stienwandt et al., 2022). Therefore, the results of the impact of these two parental factors are still mixed, and their effect sizes remain unclear.

Thirdly, the role of SES has been investigated by numerous studies. For instance, Detnakarintra et al. (2019) examined the association between young children's digital use and parents' education and income and found a medium level of association. In contrast, Yalçın et al. (2021) found a very weak association between young children's digital use and parents' occupation and education. A retrospective look at the research area reveals that few studies have investigated the association between SES and preschoolers' digital use, covering all three items, i.e., occupation, income, and education. Altogether, this area of research has also produced mixed results.

Fourthly, few studies have examined the relationship between preschoolers' digital use and parenting behaviors from an integrated perspective (i.e., covering all sub-items). Some studies have addressed this issue based on measures of parenting style (Lee & Kim, 2022; McArthur et al., 2022), while others have addressed this based on measures of parenting efficacy (Chen et al., 2020; Jago et al., 2015), parental attitude (Goh et al., 2015; Jago et al., 2013), or parent–child interaction (Poulain et al., 2019; Wong et al., 2020). Thus, although many studies have examined the association between preschoolers' digital use and parenting behaviors, it remains unclear what the effect size of the association is and whether the sub-item measures moderate the association.

In summary, the existing studies have identified four main parental factors associated with young children's digital use. However, the effect size of each parental factor remains unclear. Additionally, some studies on parental factors are based on the measurement of different sub-items, and it is also uncertain whether different sub-items have different effect sizes. Therefore, a meta-analysis is needed to fill the research gap and provide a more comprehensive view of the research field.

2 Moderators of the association between parental factors and child digital use

The existing studies have indicated that two factors might be associated with the parental impact on preschoolers' digital use, such as parental roles (mother vs. father/parents) and cultural background (western vs. non-western). First, previous studies have attempted to reveal the effects of different parental roles on children's digital use. For example, Schoeppe et al. (2017) found maternal modeling of outdoor activities was significantly associated with girls' outdoor activities, while paternal modeling of digital use was significantly associated with these screen-based behaviors in boys. Xie et al. (2023) found both parental and maternal ST (i.e., screen time) were associated with children's ST; while, as family income increased, only the association between maternal ST and children's ST was observed to decline. These findings suggested the potential moderating role of parents in children's prolonged digital use. However, this potential moderating effect was not comprehensively tested in preschoolers, although some previous studies focused on the parental role in young children's digital use, especially that of mothers (Corkin et al., 2022; Lee & Kim, 2022). Therefore, this study planned to test whether the effect size of the mother is larger than that of the father in preschoolers' prolonged digital use, as most people would expect.

Secondly, no studies have directly addressed the moderating role of cultural background in the relationship between parental factors and children's digital use. However, there were some clues. For example, Lee et al. (2021) found a relationship between the presence of electronics in children's bedrooms in the early years and their screen time being moderated by cultural background. Given the significant differences in parenting styles between Eastern and Western countries (Chao, 2000), we believe that cultural background must play an important role in the association between parental factors and preschoolers' prolonged digital use. The present study plans to test whether cultural background acts as a moderator of the association, which may be difficult to answer with a small sample of cross-national studies. Together, the moderation analysis in the meta-review can provide some insight into the above research gaps.

3 The current study

To our knowledge, no review or meta-analysis has examined or synthesized the empirical literature on the associations of prolonged digital use with parental factors. Thus, this study seeks to identify the parent-related factors contributing to prolonged digital use in preschoolers via a meta-analysis of the existing studies. Previous research in this field has investigated the relationship between preschoolers' digital use and several parental factors, such as parenting behaviors, SES, digital use, and psychological distress. Accordingly, this study will focus on the influences of these four factors. Additionally, the study will also examine how cultural background (western vs. non-western) and parental role (mother vs. father/parents) may moderate the effects of parental risk factors. Briefly, there were two primary research questions in this study:

  1. 1.

    What are the parental factors significantly associated with prolonged preschoolers’ digital use?

  2. 2.

    Are cultural background and parental role moderators in the association between parental factors and preschoolers’ prolonged digital use?

4 Method

4.1 Search strategies

Systematic literature searches using PubMed and Web of Science were conducted for original research articles published between January 2007 and October 2022. We divided the search terms into three categories, namely, digital devices (specific terms included smartphone, smartphone*, tablet, tablet computer*, tablet device*, tablet tech*, iPhone*, mobile phone*, cell phone*, mobile device*, mobile media, iPad, touch screen*, electronic device*, handheld device*, digital tech*, screen time, screen use, and touchscreen*), participator (specific terms included preschool*, early childhood, kindergarten*, toddler*, infant*, young child, early years, and pre-school*), and parents (specific terms included parent*, father*, and mother*). We used the truncation character (*) to retrieve variants of some search terms. Articles containing matches in at least one of the following three categories were searched on the title, abstract, and Major Topic. In addition, we used operators to combine search terms, such as "AND" between search terms in three categories and "OR" between search terms in each category. The entire search strategy used in the PubMed database can be found in supplementary materials Table S1.

4.2 Inclusion and exclusion criteria

Articles included in the present meta-analysis needed to meet the following six criteria: (1) The article must have been published in English between January 2007 and October 2022, when the final search was conducted. The main reason for choosing 2007 was the release of the early version of the iPhone in 2007 and the beginning of the worldwide popularity of smartphones (Sarwar & Soomro, 2013). (2) participants needed to be the non-clinical population between 0 to 9 years old. (3) digital devices must include several popular digital devices with screens, e.g., computers, tablets, smartphones, handheld gaming devices, and TVs. (4) digital device use or screen time was the dependent variable. (5) the parents’ risk factors could be classified as at least one or more parental SES, parental digital device use, passive parenting behaviors, and parental psychological distress. (6) the article needs to provide an effect size that can be extracted or converted, as detailed in the "Data Extraction" section.

Studies involving the following categories were excluded: (1) if the digital device involved in the study was only a TV or APP. (2) the primary caregiver is a school educator, nannies, grandparent, and other nonparental caregiver. (3) reviews, semi-structured interviews, and commentaries. (4) digital device usage or screen time is the independent variable.

4.3 Study selection

The basic information (e.g., title, authors, year of publication, doi, and PubMed ID) of the1814 articles identified by the search process were imported into Office Excel, and 387 duplicates were marked through the same title and PubMed ID. After removing the duplicates, the first and co-first authors checked the titles and abstracts of the remaining 1427 articles separately. Finally, they identified 89 articles that potentially met the inclusion criteria and progressed to full-text screening. Then, full-text reading, assessment of quality, methodological relevance, and risk of bias were performed by the first author and co-first author on these 89 articles, and a final sample of 52 articles was included in the present meta-analysis. For articles with conflicting opinions, an agreement was reached primarily through critical discussion between the first and the co-first author. If the critical discussion still did not lead to an agreement, articles were determined through consultation with the corresponding author. The two researchers showed 86.21% agreement across all coded items with 89 studies, and an inter-rater agreement rate of Cohen’s κ = 0.72 was maintained during the full-text screening process. Figure 1 provides a detailed PRISMA flow diagram for study inclusion.

Fig. 1
figure 1

PRISMA flow diagram of study selection

Full size image

4.4 Methodological assessment

The methodological quality and risk of bias were assessed with reference to Mallawaarachchi et al. (2022). Specifically, each article was been assessed from 8 aspects, including research question, study design, sample population, eligibility criteria, sample size, measurements, statistical analyses, and confounding variables. Two authors (the first and co-first authors) independently assigned the scores based on 8 criteria. Disagreements between the two author’s judgments were resolved by the decision of the corresponding author. A detailed description of each criterion is provided in supplementary materials Table S2. Each study was classified as (1) low risk of bias with 0 or 1 bias markers, (2) moderate risk of bias with 2 or 3 bias markers, and (3) high risk of bias with 4 or more bias markers. The specific bias scores for each article are shown in Table S3 of the supplementary materials.

4.5 Data extraction

Data was extracted independently by the first author and the co-first author. The following data were extracted from each study: (1) author/s, publication year, and country, (2) study design, (3) sample characteristics, including sample size, age range, mean age and standard deviation, and percentage of male participants, (4) Types of digital devices used, (5) Digital devices use calculation methods or measurement tools, (6) covariant variables, if any. Disagreements between authors’ judgments were resolved through discussion.

The family factors were conceptualized into the four domains: (1) parents’ SES, including family incoming, parents’ education level, and whether parents are working or not, (2) parents’ digital use, including parents' use of common electronic devices such as computers, tablets, handheld gaming devices, TV, (3) passive parenting behaviors, including either positive (love, attention, etc.) or negative (neglect, hostility, etc.), (4) parents’ psychological distress, including parents' perceived stress, depression, and anxiety.

Referring to previous studies (Mallawaarachchi et al., 2022; van Valkengoed & Steg, 2019), the effect size was extracted or transformed according to the following criteria: (1) if studies reported Pearson r then extracted directly, otherwise (e.g., Spearman’s rho, Kendall’s tau, Standardized regression coefficients, and Univariate odds ratios) converted to Pearson r first, (2) if studies used reverse-coded items, the signs of effect sizes were flipped, (3) if studies reported multiple effect sizes per sample, these were combined into one summary effect size per study by averaging the effect sizes, (4) if sample sizes varied across analyses due to missing data, the lowest bound sample size was used, (5) the variance-stabilizing transformation to Fisher’s z (rz) was performed on all correlation coefficients before the meta-analysis was conducted, (6) coefficients were transformed back into r before reporting them using the inverse of Fisher’s z formula. The formulae used for the effect size conversion were given in the “Effect size conversion” section of the Supplementary materials.

4.6 Statistical analyses

We used RStudio® software (version 4.2.1) and the “metafor” package (version 4.0.0) to conduct the meta-analysis with a random-effect model. Pearson’s correlation coefficient (range =  − 1.0 to 1.0) and its 95% confidence interval (CI) were used to calculate the effect sizes for correlations. We reported the estimated effect size, Fisher’s z-value, p-value, and 95% CI. Histograms were used to show the distribution of effect sizes for each parental risk factor (see Supplementary materials Fig. S1). Forest plots were generated to provide a visual representation of Pearson's correlation coefficient for each parental risk factor from the included studies (see Figs. S2-S5). Q statistic and I2 index were calculated to quantify the heterogeneity. The Q statistic is a measure to estimate the between-study variance (i.e., the heterogeneity across all studies), and the I2 index measures the proportion of excess heterogeneity from the total heterogeneity. Egger’s regression test and a visual inspection of the funnel plot (see Supplementary materials Fig. S6) were used to evaluate the publication bias with the cutoff value set at p < 0.05. If bias existed, the trim-and-fill method was used to adjust the publication bias (Duval & Tweedie, 2000).

5 Results

5.1 Study characteristics

Details on the study design, sample characteristics, types of digital devices used, digital devices' use of calculation methods or measurement tools, and the covariates examined in each article are shown in supplementary materials Table S4. The majority of the studies (N = 45, 86.54%) used a cross-sectional study design, and the rest (N = 7, 13.46%) used a longitudinal study design. The 52 samples had a combined sample size of 54,334, which ranged from 36 (Attai et al., 2020) to 10,967 (Gao et al., 2022). The study participants were in the age range of 0–9 years. Most articles (N = 41, 78.85%) have been published since 2019, with the remainder published from 2010 to 2018. The distribution of the 52 studies on each continent was Asian (N = 21, 40.38%), North America (N = 18, 34.62%), European (N = 9, 17.30%), South America (N = 2, 3.85%), and Oceania (N = 2, 3.85%). The most common countries for conducting studies were the United States (N = 12, 23.08%), Canada (N = 6, 11.54%), Turkey (N = 6, 11.54%), China (N = 4, 7.69%), South Korea (N = 3, 5.77%), and the United Kingdom (N = 3, 5.77%). Regarding parental risk factors examined in articles, most examined more than one. Specifically, 25 examined SES factors, 29 examined parental digital use factors, 20 examined passive parenting behavior factors, and 12 examined parental psychological distress factors. Of these, 25 studies examined two or more influencing factors.

5.2 Meta-analysis of four parental risk factors

After a thorough inspection of the available data, a meta-analysis was conducted on each of the four factors. The analysis included associations of young children’s amount of device use (i.e., duration or frequency of use, excluding parental perceptions of problematic use) with (1) parent’s digital use, (2) parents’ psychological distress (including anxiety, depression, and parenting stress), (3) passive parenting behavior (including, parental attitude, parent–child interaction, parenting efficacy, and parenting style), and (4) SES (including, education level, income, and occupation). The overall findings for each factor are illustrated in Fig. 2, while Table 1 summarizes the results. Note that the data distributions for parental risk factors were normal; see Table S5 and Fig. S1 for detailed information. The publication bias analysis and adjusted effects are presented in Table S6, demonstrating the reliability of the results presented below.

Fig. 2
figure 2

Mean meta-analytical effect sizes for four risk factors. Note. Black diamonds show the meta-analytical effect size (r) for each factor. Error bars represent the 95%CI around the effect size. Grey circles represent the effect size for individual studies. The size of the circle indicates the magnitude of the study sample size. SES = socioeconomic status. **p < .01

Full size image
Table 1 Summary of the meta-analyses for each factor
Full size table

5.2.1 Parents’ digital use

A number of researchers have suggested that parents' digital device use is closely connected to the digital device usage of preschool children. Our study results indicated that, overall, parents' digital device use was associated with prolonged screen time of preschool children (r = .24, z = 7.25, p < .01, 95%CI [ 0.17, 0.30]). However, we observed significant variation in effect sizes across studies, ranging from -.18 to .47. There may be also some moderating factors, such as the parental role or cultural background of the parents, which we would test in the moderator analysis section. A forest plot for parents’ digital use factors is presented in Fig. S2.

5.2.2 Parents’ psychological distress

The second factor often linked to preschool children's digital device use is psychological distress in parents. The results showed that parents' psychological distress was associated with prolonged screen time among their children (r = .15, z = 3.77, p < .01, 95%CI [0.07, 0.23]). However, there was significant heterogeneity in effect sizes across studies (Q = 82, df = 11, p < .01). This prompted us to investigate whether sub-items of psychological distress have varying effects on preschool children’s digital device use. We found that the sub-items did not moderate the association above (Q = 2.34, df = 2, p = .31). Nonetheless, as Fig. 3 showed, parents' parenting stress (r = .18, z = 3.57, p < .01) had small-to-medium level effects on digital device usage, while anxiety (r = .10, z = 1.58, p = .12) and depression (r = .07, z = 1.05, p = .29) had non-significant effects (see Fig. S4 for the forest plot).

Fig. 3
figure 3

Mean meta-analytical effect sizes for subset factors of psychological distress, passive parenting behaviors and low SES. Note. Black diamonds show the meta-analytical effect size (r) for each factor. Error bars represent the 95%CI around the effect size. Grey circles represent the effect size for individual studies. The size of the circle indicates the magnitude of the study sample size. SES = socioeconomic status. * p < .05; ** p < .01

Full size image

5.2.3 Passive parenting behaviors

Parenting behavior is another factor that is studied in relation to young children's digital device use. The result indicated that parents' parenting behaviors were associated with their children's screen time (r = .17, z = 4.35, p < .01, 95%CI [0.10, 0.25]; see Fig. S3 for the forest plot). However, given the significant amount of heterogeneity in effect sizes reported (ranging from -.19 to .51), we explored whether the sub-items of parenting behavior have differential effects on children’s screen time. The results showed that none of the sub-items moderated the relationship between parenting behavior and children’s digital device use (Q = 1.76, df = 3, p = .62). Nevertheless, Fig. 3 illustrates that parenting style (r = .24, z = 2.74, p < .01) and parental attitude (r = .18, z = 2.31, p = .02) had small-to-medium level effects on digital device usage, while parent–child interaction (r = .14, z = 1.56, p = .12) and parenting efficacy (r = .09, z = 1.15, p = .25) did not show a significant impact.

5.2.4 Low SES

In most studies, SES is typically considered a control variable rather than a variable of interest. However, the meta results suggested that SES may play a significant role in parents' digital device use and its impact on preschool children's screen time. Specifically, we found that parents' digital device use was positively associated with their children's prolonged screen time (r = .10, z = 7.89, p < .01, 95%CI [0.08, 0.13]). Given the large heterogeneity in effect sizes reported across studies (Q = 106, df = 24, p < .01), we examined the potential moderating effects of different sub-items of SES. Still, no significant moderation was found (Q = 5.05, df = 2, p = .08). The Fig. 3 shows that parents' education level had small-to-medium level effects on children's digital device usage (r = .15, z = 6.08, p < .01), while occupation (r = .09, z = 1.94, p = .05) and income level (r = .07, z = 2.37, p = .02) had smaller effects (see Fig. S5 for the forest plot).

5.3 Moderation analyses for the four parental risk factors

The above results suggest a considerable degree of heterogeneity among the various risk factors. This heterogeneity can be attributed not only to differences in sub-items but also to other moderating factors. In this study, we proposed two potential moderators: the parental role (mother vs. father/parents) and the parents' cultural background (Western vs. non-Western), which have attracted interest from researchers in the field.

5.3.1 Parent’s cultural background

We investigated whether a parent's cultural background (Western vs. non-Western) moderated the impact of each risk factor. As Table 2 showed, culture background did not significantly moderate the effects of low SES (Q = .40, df = 1, p = .53) or psychological distress (Q = .43, df = 1, p = .51). However, we did observe that the impact of parents' digital use was also moderated by culture background (Q = 8.38, df = 1, p < .01), with studies from Western countries reported larger effects (r = .31, z = 7.89, p < .01, 95%CI [0.24, 0.38]) than studies from non-Western countries (r = .15, z = 3.41, p < .01, 95%CI [0.06, 0.23]). Additionally, this moderating effect appeared to be significant in passive parenting behaviors (Q = 3.32, df = 1, p = .06); in contrast to the findings for passive parenting, the studies from non-Western countries reported larger effect sizes (r = .24, z = 4.51, p < .01, 95%CI [0.14, 0.34]) than those from Western countries (r = .11, z = 2.02, p = .04, 95%CI [0.00, 0.21]).

Table 2 Culture background moderation analyses for each parental risk factor
Full size table

5.3.2 Parental role

We also examined whether the parental role moderated each risk factor. As Table 3 showed, the parental role did not significantly moderate any of the four risk factors (passive parenting behavior: Q = 0.96, df = 1.00, p = .33; low SES: Q = 1.50, df = 2, p = .47; psychological distress: Q = 0.72, df = 2, p = .70; parents' digital use: Q = 4.16, df = 2, p = .13). Nevertheless, according to the Table 3, we can find the effect sizes reported by fathers (low SES: r = .08; psychological distress: r = .02; digital use: r = .12) appeared to be smaller than those reported by mothers for both risk factors (low SES: r = .12; psychological distress: r = .15; digital use: r = .16). It is worth noting that there were very few studies that specifically address factors related to fathers (less than 4 articles for each risk factors).

Table 3 Parental role moderation analyses for each parental risk factor
Full size table

6 Discussion

This meta-analysis study has synthesized the existing evidence about the impact of four parental risk factors on preschoolers’ digital use and the associated moderators. This section will discuss these findings and their implications for future studies.

6.1 The four parental risk factors

First, this study has confirmed the four parental risk factors with synthesized evidence from a meta-analysis. In particular, the largest effect size was observed for parents' digital use, followed by parent's psychological distress. This is not surprising as these two factors have been frequently studied, with a previous systematic review demonstrating that maternal distress/depression and maternal television viewing time were associated with young children’s screen media use (Duch et al., 2013), and our meta-analysis further confirmed their important role in preschool children's digital screen time with statistical evidence. Furthermore, in contrast to previous reviews that reported that parents’ SES was not or not clearly associated with their children's digital use (Duch et al., 2013; Paudel et al., 2017), the synthesized findings of this study suggest that this factor is a significant predictor of digital device use in preschool children, although its impact was weaker. In addition, this meta-analysis also revealed that preschool children's digital use was associated with passive parenting behaviors, which is important but not often considered, as suggested by the system review of Paudel et al. (2017). Therefore, the influence of parental behavior and SES should be considered when investigating young children's use of digital devices.

Second, it should be noted that our meta-analysis revealed significant heterogeneity in the effect sizes of all four risk factors. Given that there were sub-items for some of the risk factors, such as parenting behaviors and SES, we investigated whether these subsets may have different levels of contribution to the risk factor, thereby causing heterogeneity. Our findings showed, compared to other sub-items of psychological distress, parenting stress had the largest effect size on young children's digital device usage. This is consistent with the argument that parents would seek out helpers to reduce their parenting stress and digital devices are easily accessible helpers (Radesky et al., 2016). Thus, parenting stress seems to be an important risk factor that requires attention. Regarding to SES, education level had the largest effect size among the three sub-items. This may come as a surprise, but it is reasonable. Compared to income and occupation, a parent's education level is more related to their spending on their children's education and development (Hao & Yeung, 2015). Interestingly, among the four sub-items of parenting behavior, parenting style, and parental attitude were significantly associated with preschool children's digital device use. The results highlight the important role parents' attitudes towards their children's development play in preschool children’s digital screen time, which is consistent with recent systematic research (Chong et al., 2023).

Above all, the findings suggest that we need to pay more attention to specific factors related to parenting practices, such as parenting attitude or parenting stress when attempting to reduce the risk of digital overuse by preschool children.

6.2 The two moderators

This meta-analysis also revealed that the heterogeneity could not be entirely explained by differences in sub-items. This suggests that some moderators may influence the relationship between parental risk factors and young children's digital device usage. The first moderator is the cultural background of the parents. Traditionally, attitudes toward children's education differ between parents from Eastern and Western countries (Chao, 2000). Given that parental attitudes play an important role in young children's screen time (Chong et al., 2023), cultural background may act as a moderator. Our meta-analysis confirmed that cultural background was a significant moderator for parenting behaviors and parents’ digital use. Notably, the parent's digital device usage had a larger effect size in Western countries than in non-Western countries, whereas parenting behaviors had a larger effect size in non-Western rather than Western countries. These findings may suggest that in non-Western, especially Eastern countries, parents regulate their children's activities more strictly, and their parenting behaviors have a greater impact on their children's digital device usage. The view above has also been supported by the study of Lee et al. (2021), which showed the relationship between the presence of digital devices in children's bedrooms and children's screen time was moderated by cultural background, with a stronger relationship in Canada than in Korea. Although parents in Eastern countries may use digital devices frequently, they are more likely to restrict their children's use, whereas in Western countries, parents tend to give their children more freedom. When parents themselves use digital devices frequently, their children are more likely to have access to and use them.

Another moderator we considered is the parental role, specifically the differences between fathers and mothers. Contrary to our expectations, the analyses did not show a moderating effect of the parental role on the effect of each of the four risk factors on young children's digital use. However, the value of effect size suggested that fathers had a smaller effect on preschoolers' digital use than mothers. This is consistent with previous research showing that fathers are less likely to limit their children's screen time use (Chong et al., 2023; Tang et al., 2018). And the lack of research on fathers may have led to the observation of no significant moderating effect. Note that the value of effect size also suggested that the combined moderating effect of fathers and mothers is greater than that of mothers in terms of the influence of low SES status and parental digital use on children's use of electronic devices. This trend may reflect the particular role of fathers in these two areas. Together, the characteristics of mothers appear to be more important for young children's digital device usage, but the influence of fathers' characteristics should not be ignored.

6.3 Limitations and implications

Several limitations should be noted when interpreting this meta-review. First, potentially relevant studies may be missed due to unclear titles or abstracts. Second, a few relevant studies may be missed because only the two most popular databases, PubMed and Web of Science, were used for the search. Third, this review is limited to studies published in English; therefore, some relevant studies written in other languages may be missed. Fourth, when conducting the meta-analysis, approximations of R-values rather than actual values of some studies were used, which may cause slight bias. Fifth, this study focused on normally developing children, so the findings cannot be generalized to clinical populations.

Despite these limitations, the findings of this meta-analysis are reliable given the large sample size (as indicated by fsn index, see Table 1) and provide valuable insights into the risk factors for parental use of young children's digital overuse. First, government agencies and healthcare providers can use the parental risk factors identified here, such as parental digital use and low SES, to identify families with preschoolers of potential digital overuse. Second, cultural background or nationality should be considered when identifying families at high risk of long-term use of digital devices in young children, as cultural background is an important moderator of parental risk factors, including parents' digital use and passive parenting behaviors. Third, parenting practices significantly influence the long-term use of digital devices by preschoolers; thus, there is a need for evidence-based education programs for parents. Fourth, there is a lack of research on father-related factors in the long-term digital use of young children, and future research needs to explore the specific role of fathers in preschoolers' digital use.

7 Conclusion

The meta-analysis confirms that the four parental risk factors play a role in preschool children's digital device usage. However, this influence is moderated by parents' cultural background, and subsets for each risk factor may have different effect sizes. Future research in this area should focus more on the population of developing and least-developed countries and the role of fathers. With a better understanding of the factors, we can develop more effective interventions to reduce the risk of excessive digital device usage among preschool children.