*Result*: The influence of children's emotional regulation on internet addiction during the COVID-19 pandemic: the mediating role of depression.

*Internet addiction (IA) and depression present significant public health challenges, especially during a pandemic. Previous research conducted outside of pandemic contexts highlighted the importance of emotional regulation (ER) for depression, with specific strategies such as cognitive reappraisal (CR) and expressive suppression (ES) showing effectiveness in predicting the internalization of problems. However, knowledge regarding ER strategies for depression and IA during the pandemic remains limited, thus hindering our implementation of effective strategies. This study aimed to examine the relationships between IA, ER strategies, and depressive symptoms, and to identify protective factors against depression and IA. Additionally, it sought to investigate the mediating role of depressive symptoms. Data were collected from 824 students across the pre-pandemic and pandemic periods. Results highlight the risks associated with ES for both depressive symptoms and IA, while CR demonstrates potential in reducing depressive symptoms and IA. Interventions that promote the development of CR and discourage reliance on ES can effectively mitigate depressive symptoms and IA.*

AN0191136548;0ux01feb.26;2026Jan29.02:21;v2.2.500

The influence of children's emotional regulation on internet addiction during the COVID-19 pandemic: the mediating role of depression 

Internet addiction (IA) and depression present significant public health challenges, especially during a pandemic. Previous research conducted outside of pandemic contexts highlighted the importance of emotional regulation (ER) for depression, with specific strategies such as cognitive reappraisal (CR) and expressive suppression (ES) showing effectiveness in predicting the internalization of problems. However, knowledge regarding ER strategies for depression and IA during the pandemic remains limited, thus hindering our implementation of effective strategies. This study aimed to examine the relationships between IA, ER strategies, and depressive symptoms, and to identify protective factors against depression and IA. Additionally, it sought to investigate the mediating role of depressive symptoms. Data were collected from 824 students across the pre-pandemic and pandemic periods. Results highlight the risks associated with ES for both depressive symptoms and IA, while CR demonstrates potential in reducing depressive symptoms and IA. Interventions that promote the development of CR and discourage reliance on ES can effectively mitigate depressive symptoms and IA.

Keywords: Cognitive reappraisal; COVID-19; expressive suppression; negative emotion; problematic internet use; elementary school students

Introduction

Internet usage has become an integral part of life, especially due to social distancing and lockdown strategies during the COVID-19 pandemic. While the internet offers convenience, uncontrolled internet use may lead to internet addiction (IA), which is defined as individuals' excessive and compulsive internet use, and involves a persistent and uncontrollable urge to engage in online activities to the extent that it disrupts other areas of life and impairs functioning (B. Dong et al., [12]; Giordano et al., [17]). For example, adolescents who experience IA often experience negative mental health (Han et al., [21]; Ho et al., [24]; Huang et al., [26]). Additionally, excessive internet use can lead to sleep disturbances and can have a significant impact on students' academic performance, potentially resulting in poor school attendance, decreased concentration, and even contributing to the risk of school failure (M. P. Lin, [34]; Xin et al., [52]).

Recent meta-analyses reported a global pooled prevalence of IA at 8.23% (Meng et al., [39]), which increased to 10.6% during the COVID-19 pandemic (Alimoradi et al., [1]). IA is a serious health problem worldwide, and particularly in Asian countries (H. Dong et al., [13]; Lia et al., [30]). Studies have indicated an IA prevalence of 10.7% in South Korea (Park et al., [41]), 10.4% to 26.5% in China (B. Dong et al., [12]; Tang et al., [48]; Xin et al., [52]), and 11.0% to 14.3% in Taiwan among adolescents (Department of Information and Technology Education, [10]). The prevalence of problematic internet use and IA significantly increased during the pandemic, especially for adolescents (H. Dong et al., [13]; Huang et al., [26]; M. P. Lin, [34]).

During the pandemic, children and adolescents faced significant public health challenges, not only related to IA but also to depression (Alimoradi et al., [1]; Duan et al., [15]; Masaeli & Farhadi, [37]). Most studies exploring the relationship between depression and IA have focused on adolescents (e.g. Bai et al., [2]; Cheung et al., [5]; Houghton et al., [25]; Raudsepp & Kais, [42]; Zhao et al., [57], [58]; Zhen et al., [59]) or adults (e.g. S. K. Chen & Lin, [4]; Jasso-Medrano & López-Rosales, [27]; Lim et al., [33]), although some studies have included samples spanning multiple developmental stages. For example, Seo et al. ([45]) examined children and adolescents (aged 9–18), while Worsley et al. ([50]) focused on adolescents and young adults (aged 17–25). However, few studies have specifically targeted children (e.g. Chiang et al., [6]). Considering that early IA can have significant effects on subsequent development, this study sought to investigate the underlying mechanisms for elementary school children, addressing a gap in the existing literature.

Moreover, existing research, mostly conducted outside of pandemic contexts, suggests that emotion regulation (ER) plays a crucial role in addressing depression, and specific ER strategies are effective for managing the internalization of problems (Dryman & Heimberg, [14]; L. Liang et al., [31]). However, limited knowledge exists regarding specific ER strategies for depression and IA, and much of the research has relied on cross-sectional designs. To address these unique public health challenges, it is essential to gain a comprehensive understanding and to implement effective strategies. Therefore, this study aimed to investigate the relationships between IA, ER strategies, and depressive symptoms among children by collecting data at two different time points. Additionally, the study sought to identify protective factors against depression and IA. Although there is no consensus on the causal relationship between depressive symptoms and IA, increasing evidence suggests that depressive symptoms can predict IA in adverse contexts (Seo et al., [45]; Worsley et al., [50]). Longitudinal research indicates that an increase in depressive symptoms raises the likelihood of problematic internet use (S. K. Chen & Lin, [4]). Therefore, this study also explored the mediating role of depressive symptoms in the relationship between ER and IA within the context of the pandemic.

Internet addiction and depression

IA has been found to be associated with increased stress and internalizing problems, including depressive symptoms (Han et al., [21]; Huang et al., [26]). Some individuals use the internet persistently as a coping mechanism to distract themselves from negative emotions (DiBlasi et al., [11]; Hernández et al., [23]). Conversely, others may engage in excessive internet use to avoid real-life responsibilities, leading to difficulties in controlling their internet usage, and procrastinating to perform important tasks, which in turn exacerbates their stress and depression (DiBlasi et al., [11]; Hernández et al., [23]). The comorbidity between IA and depression has been established (Tang et al., [48]; L. Yang, [55]). Research has shown that the pandemic made children and adolescents more susceptible to mental health issues and behavioral addiction problems (M. P. Lin, [34]). Reviews have indicated increased IA during the pandemic, linked to mental health conditions (Alimoradi et al., [1]; Masaeli & Farhadi, [37]). However, most empirical studies examining IA and depression during this period employed cross-sectional designs, collecting data at only one-time point (e.g. H. Dong et al., [13]).

Research has also suggested that adverse childhood experiences, stress, and trauma are associated with IA and depressive symptoms, with depressive symptoms potentially mediating the relationship between stress and IA. For instance, Seo et al. ([45]) identified depressive symptoms as a mediator between IA and adverse childhood experiences, while Worsley et al. ([50]) demonstrated that depression mediates the relationship between childhood maltreatment and IA. Moreover, a developmental perspective and longitudinal studies provide evidence that changes in depressive symptoms influence the likelihood of problematic internet use. For example, S. K. Chen and Lin ([4]) found that trajectories of depressive symptoms significantly impacted problematic internet use among college students, with faster declines in depression associated with lower levels of problematic internet use. Similarly, using a representative sample of 2,155 Taiwanese children, Chiang et al. ([6]) demonstrated that depressive symptoms in fifth grade were a significant predictor of smartphone addiction in sixth grade. Given this existing evidence, our research model assumes that depressive symptoms impact IA.

Emotion regulation strategies and depression

ER is a key predictor of behavioral addiction and emotional outcomes, allowing us to modify the intensity, timing, and expression of our emotions (Giordano et al., [17]; Gross & Jazaieri, [18]). Cognitive reappraisal (CR) and expressive suppression (ES) are central ER strategies (Dryman & Heimberg, [14]). CR involves viewing circumstances in a different, less stressful way, while ES involves denying the importance of emotions and attempting to hide them from others (Gross & Jazaieri, [18]; L. Liang et al., [31]). Systematic reviews have shown a positive correlation between emotion dysregulation and problematic technology use among adolescents, and a reliable association between ES, CR, and internalizing symptoms (Dryman & Heimberg, [14]; H. Yang et al., [54]). CR is often linked to lower levels of internalizing problems, while ES is associated with maladaptive outcomes including internalizing problems (Dryman & Heimberg, [14]; L. Liang et al., [31]).

Although the association between maladaptive outcomes and ER strategies has been supported as mentioned above, research findings regarding the impact of CR and ES on depression have been inconsistent. For instance, a study of 1,343 adolescents found that only ES, but not CR, was associated with depression, suggesting that ES may be an outcome rather than a cause of depression (De France et al., [9]). Studies conducted during the COVID-19 pandemic have also provided mixed results regarding the relationships between CR, ES, and mental health outcomes (S. Liang et al., [32]; Sun et al., [47]). Sun et al. ([47]) suggested that both CR and ES reduced negative emotions among Chinese frontline medical staff. In another study, CR was viewed as a protective factor for worry, while ES was viewed as a risk which increased worry during the pandemic in Italy (Sebri et al., [43]). Because adults and children are different in terms of how they perceive stress, and possess different abilities and resources to cope with pandemics, the above information calls for further understanding of the role of CR and ES on emotional outcomes for adolescents during stressful times.

Emotion regulation strategies and internet addiction

Dysfunctional emotion regulation has been consistently associated with problematic internet use including IA (P. Y. Lin et al., [35]; Uçur & Dönmez, [49]; Yan et al., [53]; Zhao et al., [58]). For example, a study of 262 high school students found that external and internal dysfunctional emotion regulation, as well as internal functional emotion regulation, significantly predicted both IA and smartphone addiction (Yildiz, [56]). Wu et al. ([51]) observed that childhood maltreatment increased dysfunctional regulation, which in turn heightened internet gaming addiction among middle school students. Additionally, a panel study indicated that childhood emotion regulation difficulties predicted increased IA in adolescence (Cimino & Cerniglia, [8]).

Despite the established link between ER and IA, limited research has explored the specific roles of cognitive reappraisal (CR) and expressive suppression (ES) in relation to IA in children, as well as the underlying mechanisms involved. L. Liang et al. ([31]) found that negative emotions mediate the relationship between ES and IA, as well as the relationship between CR and IA. Their study also highlighted the protective role of reappraisal in mitigating negative emotions and its predictive association with IA. While these findings are insightful, further empirical investigations with longitudinal data and diverse samples are necessary to validate these results. A deeper understanding of the relationships between specific ER strategies, IA, and depression can inform evidence-driven prevention and intervention strategies to effectively address these issues.

Based on the literature review, we proposed the first two hypotheses:

H1:

There is a positive association between IA and depressive symptoms (Tang et al., [48]; L. Yang, [55]).

H2:

ES is positively associated with depressive symptoms, while CR is negatively associated with depression (Dryman & Heimberg, [14]; L. Liang et al., [31]).

Although little research has examined the relationship between specific ER strategies (CR, ES) and IA, research (Yildiz, [56]) has found that not every ER strategy significantly predicts IA. Moreover, previous studies suggested that the effects and mechanisms of CR and ER on mental health outcomes are more likely to be different (De France et al., [9]; Dryman & Heimberg, [14]; Sun et al., [47]). We, therefore, proposed the hypothesis that the paths whereby CR and ES impact IA could differ (H3).

Materials and methods

Participants and procedure

This study utilized a two-wave follow-up design. The Municipal Education Bureau of a northern city in Taiwan facilitated the recruitment of 18 elementary schools, comprising 47 classes, to participate in the project. Students in each class completed the online questionnaires in their schools' computer classrooms. Only students who clicked the 'agree' icon on the consent form displayed on the first page of the online questionnaire could participate in the survey. The consent form clearly outlined the purposes, procedure, confidentiality, and ethical considerations in language understandable to elementary school students. During the data collection process, teachers were instructed to neither interrupt nor guide the students while they completed the questionnaire. The procedure, instrument, consent forms, confidentiality, and ethics of participants' safety were approved by the Institutional Review Board of the University of Taipei (UT-IRB No. IRB-2022-067) (Omitted during anonymous review).

There were 957 fifth graders (aged 11) who participated in the first-wave survey (T1 October 2020), and 899 who participated in the second wave (T2 June 2021). The duration between the two waves was 8 months. The second wave was conducted during the school closure period regulated by the government to combat the COVID-19 pandemic. Those who participated in both waves (n = 824(824/957 = 86%), 52.55% females) were included in the following analysis.

Measures

Emotional regulation

The emotional regulation scale (ERS) consisted of two subscales: cognitive reappraisal and expressive suppression (Gross & John, [19]). Six items measured the concept of cognitive reappraisal, such as 'I control my emotions by changing the way I think about the situation I'm in' and 'When I want to feel less negative emotion, I change the way I'm thinking'. Four items captured the content of expressive suppression. Sample items include 'I control my emotions by not expressing them' and 'I keep my emotions to myself'. The options ranged from 0 ('strongly disagree') to 6 ('strongly agree'). The standardized Cronbach's alpha of the Chinese version ERS in the first wave of this sample was.84 for cognitive reappraisal and.66 for expressive suppression. The averaged item score represented the levels of cognitive reappraisal and expressive suppression and ranged from 0 to 6.

Depressive symptoms

The study adopted the Center for Epidemiologic Studies Depression Scale for Children (CES-DC) to measure depressive symptoms (Shahid et al., [46]). CESD measured four dimensions of depressive symptoms, including depressive affect, somatic complaints, interpersonal problems, and positive affect, using 20 items. The options ranged from 0 ('not at all') to 3 ('a lot'). The averaged item score representing the level of depressive symptoms ranged from 0 to 3. The standardized Cronbach's alpha of the Chinese version of CES-DC (Chien & Cheng, [7]) adopted in this study was.87 in the first wave.

Internet addiction

The IA measurement was adopted from the Chinese Internet Addiction Scale (CIAS) which included five dimensions: withdrawal symptoms, compulsive internet use, tolerance, time management, and interpersonal relationships (S. H. Chen et al., [3]). Each dimension comprised two items. The options ranged from 1 (very like me) to 4 (very unlike me). Cronbach's alpha was.86 in the second wave of this study. The averaged item score represented the level of IA and ranged from 1 to 4.

Control variables

Sex and socioeconomic status are frequently mentioned factors related to IA (Lee & McKenzie, [29]; Mari et al., [36]). Thus, this study included sex and socioeconomic status as control variables in the model. Sex was measured by one item, where 1 = boys and 2 = girls. Socioeconomic status was measured by providing a graph showing a 10-step social ladder and asking participants: 'Please imagine a ladder, the 10<sups>th</sups> floor represents the best life in Taiwan, the 0<sups>th</sups> floor represents the worst life in Taiwan, which level of the ladder do you think your current life is on?' A higher rung on the ladder indicated a better life in Taiwan.

Statistical analysis

Sample characteristics were demonstrated by means and standard deviations for continuous variables and numbers and percentages for categorical variables. The correlation matrix showed the correlations between all variables in this study. Mediation hypotheses were examined using a series of multiple regression models conducted using PROCESS Model 4 (Hayes, [22]). The model was used to investigate the direct association and indirect association among emotional regulation, depressive symptoms, and internet addiction. SAS 9.4 was used to conduct the analysis. The study framework is shown in Figure 1.

Graph: Figure 1. Mediation model of internet addiction, emotional regulation, and depressive symptoms.

Result

Sample characteristics

Table 1 shows the descriptive statistics of the study sample. The participants, 47.45% boys, had a relatively high socioeconomic status (M = 7.4, SD = 2.13), moderate levels of CR (M = 3.82, SD = 1.41), ES (M = 3.10, SD = 1.43), and IA (M = 2.14, SD = 0.63). In addition, they had low levels of depressive symptoms (M = 0.64, SD = 0.44).

Table 1. Sample characteristics (N = 824).

<table><thead><tr><td /><td>Range</td><td>Mean</td><td><italic>SD</italic></td><td>Correlation Coefficients</td></tr><tr><td>X1</td><td>X2</td><td>M</td><td>Y</td><td>C1</td><td>C2</td></tr></thead><tbody><tr><td>X1: Cognitive Reappraisal (T1)</td><td>0&#8211;6</td><td>3.82</td><td>1.41</td><td>1.00</td><td /><td /><td /><td /><td /></tr><tr><td>X2: Expressive Suppression (T1)</td><td>0&#8211;6</td><td>3.10</td><td>1.43</td><td><bold>0.30</bold></td><td>1.00</td><td /><td /><td /><td /></tr><tr><td>M: Depressive Symptoms (T1)</td><td>0&#8211;3</td><td>0.64</td><td>0.44</td><td><bold>&#8722;0.27</bold></td><td><bold>0.14</bold></td><td>1.00</td><td /><td /><td /></tr><tr><td>Y: Internet Addiction (T2)</td><td>1&#8211;4</td><td>2.14</td><td>0.63</td><td>&#8722;0.01</td><td><bold>0.11</bold></td><td><bold>0.14</bold></td><td>1.00</td><td /><td /></tr><tr><td>C1: Sex (T1) &#8211; Boys (%)</td><td>-</td><td>47.45%</td><td>-</td><td>0.05</td><td>0.04</td><td>0.00</td><td>0.04</td><td>1.00</td><td /></tr><tr><td>C2: Socioeconomic Status (T1)</td><td>1&#8211;10</td><td>7.40</td><td>2.13</td><td><bold>0.22</bold></td><td>&#8722;0.01</td><td><bold>&#8722;0.28</bold></td><td><bold>&#8722;0.11</bold></td><td><bold>0.09</bold></td><td>1.00</td></tr></tbody></table>

1 T1 = Time 1; T2 = Time 2. Time 1 variables were measured at the first wave, and Time 2 variables were measured at the second wave.

2 Missing data: 23 participants did not report their socioeconomic status. SD = standard deviation. Bold numbers indicate p <.01.

Table 1 also demonstrates the bivariate associations among study variables. The results revealed positive associations between CR and ES (r =.30), as well as between ES and depressive symptoms (r =.14). On the other hand, CR was negatively associated with depressive symptoms (r = −.27). Additionally, IA showed positive associations with both ES (r =.11) and depressive symptoms (r =.14). Socioeconomic status was positively related to CR (r =.22) and negatively associated with depressive symptoms (r = −.28) and IA (r = −.11). Participants' sex did not exhibit significant associations with other factors (p >.05).

Results of the mediation analysis

The findings revealed the total effects of first-wave CR and ES on second-wave internet addiction, taking sex and socioeconomic status into account (see Table 2 and Figure 1). First-wave CR showed no significant association with second-wave IA (c<subs>1</subs> = −.015, SE =.0169, p =.386, 95% CI [−.0477,.0185]). However, participants with higher levels of ES in the first wave were found to have significantly higher levels of IA in the second wave (c<subs>2</subs> =.049, SE =.0162, p =.002, 95% CI [.0176,.0813]).

Table 2. Results of the mediation analysis.

<table><thead><tr><td /><td>M: Depressive symptoms</td><td>Y: Internet Addiction</td><td>Total and Indirect effect</td></tr><tr><td>B</td><td>SE</td><td><italic>p</italic></td><td>B</td><td>SE</td><td><italic>p</italic></td><td>Effect</td><td>Boot SE</td><td>LLCI</td><td>ULCI</td></tr></thead><tbody><tr><td>Constant</td><td>1.032</td><td>0.072</td><td>&#60;.001&#42;&#42;&#42;</td><td>2.034</td><td>0.128</td><td>&#60;.001&#42;&#42;&#42;</td><td /><td /><td /><td /></tr><tr><td>Cognitive Reappraisal (CR)</td><td>a₁: &#8722;0.088</td><td>0.890</td><td>&#60;.001&#42;&#42;&#42;</td><td>c₁': 0.001</td><td>0.018</td><td>.964</td><td /><td /><td /><td /></tr><tr><td>Expressive Suppression (ES)</td><td>a₂: 0.069</td><td>0.010</td><td>&#60;.001&#42;&#42;&#42;</td><td>c₂': 0.037</td><td>0.017</td><td>.025&#42;</td><td /><td /><td /><td /></tr><tr><td>Depressive symptoms (M)</td><td>&#8211;</td><td>&#8211;</td><td>&#8211;</td><td>b: 0.175</td><td>0.056</td><td>.002&#42;&#42;&#42;</td><td /><td /><td /><td /></tr><tr><td>Sex</td><td>0.036</td><td>0.028</td><td>.194</td><td>0.038</td><td>0.044</td><td>.388</td><td /><td /><td /><td /></tr><tr><td>Socioeconomic status</td><td>&#8722;0.045</td><td>0.007</td><td>&#60;.001&#42;&#42;&#42;</td><td>&#8722;0.024</td><td>0.011</td><td>.027&#42;</td><td /><td /><td /><td /></tr><tr><td>R²</td><td>0.175</td><td /><td>&#60;.001&#42;&#42;&#42;</td><td>0.037</td><td /><td>&#60;.001&#42;&#42;&#42;</td><td /><td /><td /><td /></tr><tr><td>Total Effect of CR on Y</td><td /><td /><td /><td /><td /><td /><td>c₁: &#8722;0.015</td><td>0.0169</td><td>&#8722;0.0477</td><td>0.0185</td></tr><tr><td>Total Effect of ES on Y</td><td /><td /><td /><td /><td /><td /><td>c₂: 0.049</td><td>0.0162</td><td>0.0176</td><td>0.0813</td></tr><tr><td>Indirect Effect (CR &#8594; M &#8594; Y) (a1&#42;b)</td><td /><td /><td /><td /><td /><td /><td>&#8722;0.015</td><td>0.0055</td><td>&#8722;0.0266</td><td>&#8722;0.0051</td></tr><tr><td>Indirect Effect (ES &#8594; M &#8594; Y) (a2&#42;b)</td><td /><td /><td /><td /><td /><td /><td>0.012</td><td>0.0045</td><td>0.0039</td><td>0.0217</td></tr></tbody></table>

CR in the first wave was negatively associated with depressive symptoms, which were positively linked to IA in the second wave. The indirect path from CR to IA through depressive symptoms was significantly negative (a<subs>1</subs>b = −.015, Percentile bootstrap 95% CI [−.0266, −.0051]), as the confidence interval did not include zero. By contrast, ES in the first wave increased depressive symptoms, which subsequently raised the risk of IA in the second wave. The analysis identified a significantly positive indirect effect from ES to IA through depressive symptoms (a<subs>2</subs>b =.012, Percentile bootstrap 95% CI [.0039,.0217]).

Overall, depressive symptoms significantly mediated the association between ES and IA during quarantine. CR had a negative indirect effect on IA through reduced depressive symptoms, while ES had a positive indirect effect on IA through increased depressive symptoms.

Discussion

The present study aimed to investigate the association between ER strategies (i.e. CR and ES), depression, and IA, while also testing the mediating role of depressive symptoms. As shown in Table 2 and Figure 1, the findings supported hypothesis 1, demonstrating a positive association between IA and depressive symptoms. These results suggest that depressive symptoms may be a risk factor for IA, consistent with previous studies (Seki et al., [44]; Tang et al., [48]; L. Yang, [55]). Moreover, the findings supported hypothesis 2, indicating that ES is positively associated with depressive symptoms, whereas CR is negatively associated with depressive symptoms. These results align with prior research, highlighting that maladaptive ER strategies, such as ES, contribute to internalizing problems, while adaptive ER strategies, such as CR, mitigate such issues (Klein et al., [28]; Sebri et al., [43]).

One possible explanation for the negative consequences of expressive suppression is that the act of suppressing emotions may paradoxically intensify the severity and frequency of negative emotions, while also contributing to unwanted thoughts and behaviors (Sebri et al., [43]). Conversely, cognitive reappraisal, which involves reframing a stressful event in a less emotional manner, is associated with adaptive emotional responses and well-being (McRae et al., [38]; Sebri et al., [43]). Another possibility is that CR enables students to reframe and reassess emotional events, enhancing their adaptability in negative situations and providing a greater sense of control. In contrast, ES may contribute to heightened negative emotional experiences, leading to feelings of reduced control and maladaptive behavior problems (Gross & John, [19]).

In this study, we also explored the mediating role of depressive symptoms and proposed hypothesis 3, suggesting that the pathways through which CR and ES influence IA may differ. Previous studies investigating the association between problematic internet use or IA with emotional dysregulation consistently supported their relationship (Mo et al., [40]; H. Yang et al., [54]; Yildiz, [56]). In spite of the relevance, there is a scarcity of research specifically investigating the relationship between specific ER strategies and IA. Yildiz ([56]) found that not all of the four tested ER strategies predicted IA. The results of our study are consistent with those of Yildiz. However, measurements and methods of this study differed from those utilized in Yildiz's study. Furthermore, the current study focused on elementary students in an urban city sample, and specifically examined the expressive suppression and cognitive reappraisal aspects of emotion regulation. In this regard, the present study is the first in the literature. Furthermore, there is a lack of studies that have utilized a sample consisting of elementary school students and collected data at two different time points spanning the pandemic period, thus improving the reliability of the findings on the relationship between emotion regulation, depressive symptoms, and internet addiction compared to cross-sectional studies. This is despite the fact that differing time measurements may underestimate the predictive ability of emotion regulation strategies on IA due to potential interfering factors.

Due to the lack of comprehensive information regarding the underlying mechanisms of these findings, we are unable to clearly articulate why there was no direct effect from CR to IA after controlling for sex and economic status. While it may not be directly relevant to children's ER strategies and IA, a study on ER and IA among adolescents and young adults could potentially offer some insights. Yan et al. ([53]) found that individuals with IA have lower frontal alpha asymmetry and a positive relationship with decreasing reappraisal use. Their study suggested that IA causes ER difficulties. However, due to the divergent directions of our model and the unavailability of complete data for all variables across both waves, we were unable to explore this possibility. The variations in sample characteristics, methodology, and measurement employed in our study and Yan et al.'s study render direct comparisons impossible.

There was a partial mediating effect of depression on the relationship between ES and IA. An indirect effect was found for the relationships of CR and IA through depression. Nevertheless, depression did not act as a mediator for the relationship between CR and IA because this relationship was insignificant. Higher levels of CR corresponded to lower levels of depression, which in turn lessened IA. The results demonstrated that instead of influencing IA directly, CR might affect IA by reducing depression. This means that CR's positive impact on adolescents may not be limited to the mental aspect, but may extend to addictive behavior.

The findings were somewhat different from those of a previous study. L. Liang et al. ([31]) found that CR directly negatively predicted IA, while ES could not predict IA directly for junior high school students; negative emotion (measured by depression and anxiety) was the partial mediating factor for the relationship between CR and IA, and was the complete mediating factor between ES and IA.

One possible explanation for the difference is the disparity in age groups, resulting in significant variations in cognitive and emotional development. Additionally, the utilization of different measurement techniques, particularly for assessing IA and depression, further complicates the comparison of these two studies. Despite differences, the findings of both this study and Liang's research team partially supported the idea that negative emotions increase the risk of IA, while CR, not ES, can help mitigate IA. Students who rely more on ES struggle to regulate their emotions against daily or pandemic-related stress, leading to increased internet use as a way to cope with negative emotions (Elhai et al., [16]; Gross & Thompson, [20]). Conversely, students who employed more CR tended to experience reduced stress levels and fewer negative emotions. As a result, they were more capable of maintaining focus on their daily life activities rather than seeking distractions in the form of internet usage.

Limitations and future directions

The present study has several limitations. First, the reliance on self-report scales instead of formal diagnoses for depression and internet addiction may introduce reporting biases of the findings. Second, the sample was restricted to a single city, and data collection occurred during the COVID-19 pandemic, which may limit the generalizability of the results to individuals of different ages, from different regions, or in different circumstances. Third, factors such as adverse childhood experiences and environmental stressors, which can influence ER, depression, and IA (Lim et al., [33]; Worsley et al., [50]), were not included in the analysis. Future research should address these limitations by incorporating diverse samples, formal diagnostic tools, and additional contextual variables.

Despite its limitations, this study provides valuable insights into the relationships between IA, ER strategies, and depressive symptoms, particularly in the context of the COVID-19 pandemic. By analyzing data collected from 824 students across pre-pandemic and pandemic periods, the findings underscore the critical role of ER in influencing both depression and IA. Specifically, CR emerged as a protective factor, demonstrating its potential to reduce depression and IA, whereas ES was associated with increased risks for both conditions.

These findings highlight the importance of incorporating depressive symptoms and ER strategies into interventions targeting adolescent internet-related problems, particularly during periods of heightened stress. Promoting adaptive ER strategies such as CR and discouraging maladaptive strategies like ES could play a pivotal role in mitigating the psychological impacts of IA and depression. Routine screening for depression and ER strategies by professionals working with adolescents may facilitate the early identification of high-risk individuals and enable timely, targeted support. Future research should investigate the longitudinal effects of ER on IA and depression across diverse populations and contexts, further advancing the development of effective prevention and treatment strategies.

Acknowledgments

We sincerely recognize the support from Professor Yuwen Chang (National Taipei University of Education) and Professor April Chiung-Tao Shen (National Taiwan University). Special thanks to Hung-Chien Chang, Meng-Jung Lee, and Wen-Shan Lin for their valuable discussions. We also thank the National Taiwan University Children and Family Research Center, which sponsored the project and the data-gathering process.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Author contributions

Dr. Chen and Dr. Wu jointly conceived the research idea. Dr. Chen manages the manuscript organization, literature review, hypothesis formulation, and sections on reviews, discussion, and conclusion. Dr. Wu leads the research methodology, data analysis, table/figure creation, result writing, manuscript formatting, journal submission, communication with the journal, and handling of editorial revisions. Dr. Lee provides manuscript draft feedback, and all three authors review and approve the final version.

Data statement

The data that support the findings of this study are available upon reasonable request.

Correction Statement

This article has been corrected with minor changes. These changes do not impact the academic content of the article.

By Hsing-Jung Chen; Tony Szu-Hsien Lee and Wen-Chi Wu

Reported by Author; Author; Author