We used an online research agency (Cross Marketing Inc, Tokyo) to oversee the web-based survey from October 14 to 18, 2021, across Japan. After understanding the purpose of the study and voluntarily agreeing to participate, 600 participants from Japan were recruited through the online research provider.

The participants were parents with children between the ages of 12 and 17 years, and they were asked to fill out an anonymous online questionnaire. Parents were instructed to complete the survey about only 1 child with internet addiction, no matter how many children they had.

We asked the parents, “Please think of your child who is addicted to the internet” and “Please tell us the birth order of that child.”

The survey was divided into two groups: 300 parents who answered “yes” to the question “Do you think your child is dependent on the internet?” and 300 parents who answered “no” to that question. Questionnaires were collected until the number of parents in each group reached 300.

A descriptive analysis (ie, numbers, frequencies, percentages, means, and SDs) of the 600 respondents was conducted. The responses of the 300 respondents in the “yes” group and the 300 respondents in the “no” group were compared for differences in items using a t test. Frequencies of gender, marital status, and birth order were analyzed using the chi-square test or the Fisher exact test. For the characteristics of the participants, P values were considered by applying a 2-tailed significance level of less than .05. For the SDQ, the PSDQ, and the RQ, we used the Bonferroni correction and set the P value threshold of .05/19=.0026 in order to avoid increasing the risk of a type I error by multiple comparisons. All data were analyzed with SPSS (version 22.0; IBM Corp).

This study was approved by the Ethical Review Committee of the Graduate School of Medicine, Chiba University, in September 2021 (M10095).

The characteristics of the participants are shown in Table 1. The mean age of the parents was 49.24 (SD 5.67) years in the “yes” group and 49.07 (SD 5.06) years in the “no” group, with no significant difference between the two groups. Regarding marital status, about 95.1% (571/600) of the respondents in both groups were married; there was no significant difference between groups. There were significant differences in gender between the two groups. Female participants in the “yes” group constituted 45.0% (135/300) of the sample, whereas in the “no” group they constituted 36.0% (108/300) of the sample.

The average age of the participants’ children was 15.01 (SD 1.59) years in the “yes” group and 14.95 (SD 1.58) years in the “no” group, with no significant difference between the groups. In terms of birth order, 58.7% (352/600) of the adolescents were the first child, 31.5% (189/600) were the second child, 8.0% (48/600) were the third child, and 1.8% (11/600) were in a different birth order position, with no significant difference (Table 1).

The total PCIAT score for the group that answered “yes” (mean 55.41, SD 15.78) was significantly higher than that for the group that answered “no” (mean 35.55, SD 11.64). As for the daily time spent on the internet, the children in the group that answered “yes” spent a mean of 4.0 (SD 2.06) hours on the internet, and those in the group that answered “no” spent a mean of 1.7 (SD 1.06) hours on the internet, and there was a statistically significant difference between the two groups (P<.001).

The results of the SDQ, the PSDQ, and the RQ are shown in Table 2. In the SDQ, the mean TDS score for the group that answered “yes” was significantly higher than for the group that answered “no” (P<.001). Regarding the subscale items, mean scores for emotional symptoms, conduct problems, and hyperactivity-inattention for the “yes” group were significantly higher than those for the “no” group (P<.001). There were no significant differences between groups regarding peer problems (P<.049) and prosocial behavior (P<.13).

Regarding the PSDQ, the mean score for authoritarian parenting of the “yes” group was significantly higher than that of the “no” group (P<.001). There were no significant differences between authoritative parenting and permissive parenting. Regarding the RQ, there was no statistically significant difference between the “yes” and “no” groups of parents and children on the whole, whether they had secure, dismissive, preoccupied, or fearful relationships.

From the 300 parents who answered “yes,” we extracted those who scored 50 or higher on the PCIAT (190/300, 63.3%) to examine users who experienced occasional or frequent problems due to internet use. From the 300 parents who answered “no,” 86.0% (258/300) had a PCIAT score of less than 50. The 2 sets were compared to each other. The results of the SDQ, the PSDQ, and the RQ are shown in Table 3.

The SDQ showed statistically significant differences in the TDS and in the subscales of emotional symptoms, conduct problems, hyperactivity-inattention, peer problems, and prosocial behavior (all Ps<.001). The PSDQ showed a significant difference between authoritarian and permissive parents (P<.001) but not authoritative parents (authoritative subscale: P=.24; authoritarian subscale: P<.001; permissive subscale: P<.001).

Regarding the RQ, no statistically significant differences were found for parents under any of the items. Conversely, the children showed a significant difference only in the fearful type (P<.001).

We regrouped participants with a PCIAT cutoff value of 50, ignoring whether parents thought their child was addicted to the internet or not, conducted the analysis, and made a new table (Table 4).

The results showed that the group with a PCIAT score of 50 or higher (n=232) had a PCIAT mean score of 63.58 (SD 10.49). The group with a PCIAT score lower than 50 (n=368) had a PCIAT mean score of 34.07 (SD 8.19). The results of the SDQ, the PSDQ, and the RQ were compared between the two groups and were similar to those in Table 3.

In this study, we administered a questionnaire to investigate the relationship between parenting styles and adolescents’ internet addiction and mental health problems. In recent years, various studies on adolescents have suggested that internet dependence is associated with developmental disorders; in addition, both ADHD and ASD have been found to be associated with the risk of internet dependence. Cakmak and Gul [35] reported that weekly internet usage among children with ADHD aged 12 to 16 years was higher than among children in the control group. Kawabe et al [36] reported that 25 out of 55 participants with ASD were classified as having internet addiction based on results from the IAT.

In this study, we did not take into account the diagnosis of ADHD, ASD, or both, but we did measure SDQ scores and found that the TDS of the SDQ in the group with internet addiction was significantly higher than that in the group without addiction. Baer et al [37] reported that the Computer/Gaming-station Addiction Scale score significantly correlated with the total SDQ score. Akdeniz et al [38] reported that the TDS of the SDQ was higher in the group with internet addition compared to that of the group without internet addiction. Our findings were consistent with previous studies.

Previous research on the parent-child relationship between internet-dependent adolescents and their parents has largely been conducted from the perspective of the adolescents [39-41]. In this study, the perspective of the parents was the focus, and we investigated the parenting styles of those parents with internet-dependent adolescents. In a previous study from the parents’ perspective, Dogan et al [42] investigated the perceptions of internet addiction and parenting styles among adolescents studying in secondary schools between the ages of 14 and 19 years. They used the Parental Attitude Scale by Kuzgun and Eldeleklioğlu [43] to measure parental attitude, and the results showed a negative relationship between internet addiction and a democratic parenting style. Results of that study also showed a negative relationship between a protective-demanding parenting style and an authoritarian parenting style, which was found to have a significant positive relationship with internet addiction. This study used the PSDQ, a parenting style scale created by Robinson et al [28]; the results from Robinson et al’s study were consistent with the findings from our study, showing that parents in the group with internet-dependent children were found to have significantly higher authoritarian parenting tendencies than parents in the group with children who were not dependent on the internet.

Dogan et al [42] also found that a protective-demanding parenting style was a strong predictor of internet dependence, followed by an authoritarian parenting style. Although the 3 subscales of the PSDQ in this study and the 3 subscales of Kuzgun and Eldeleklioğlu’s Parental Attitude Scale in the study by Dogan et al [42] are not comparable, the findings with regard to the relationship of authoritarian parenting style with internet addiction may be common.

Using structural equation modeling analyses of the data from 266 adolescents, Trumello et al [44] suggested that adolescents’ mental health problems measured by the SDQ are an important mediator between parental care and youths’ internet addiction. Our findings were in accordance with their report.

The research implications of this study are that parents who have children with internet addiction may be more aware of their children’s emotional and behavioral problems, and their parenting style is more authoritarian.

Clinicians may encourage parents to stop their authoritarian parenting style, to learn good communication skills, and to reward their children when they choose desirable behaviors. They may also encourage parents to engage children in treatment for internet abuse and emotional and behavioral problems using cognitive behavioral therapy, especially the CRAFT intervention, at the end of their discussion as their recommendation.

As we suggested, although the online survey conducted in this study provided valuable information, it has several limitations. The first limitation was that we did not use the random sampling method. Originally, it would have been ideal to conduct random sampling, in which the probability of being selected for the sample would be equal for all individuals. In the future, an online survey using the random sampling method should be conducted.

The second limitation was that the children in this study were not diagnosed according to the DSM-5 diagnostic criteria for IGD. A structured diagnostic interview based on the online survey about internet addiction will be needed.

The third limitation was that no data were collected from the children in this survey. The parents’ evaluations of their children were based on their assumptions. Considering that there are differences in the understanding of internet addiction between parents and children, future research should focus on collecting data from both parents and children.

ADHD and ASD are known to be related to the risk of internet addiction. Our findings suggest that parents who think their child is addicted to the internet may recognize emotional and behavioral problems in the child measured by the SDQ. In addition, parents with children who suffer from internet addiction may have an authoritarian parenting style. Clinicians may encourage parents to learn good communication skills instead of an authoritarian parenting style.

In the future, studies should conduct additional research on internet addiction in children and their families. Cross-sectional and longitudinal research on families, especially parents, is also needed.