Internet addiction and smartphone addiction are emerging mental health concerns in adolescents. To determine the factors related to these conditions, we investigated their associations with sleep quality, chronotype, sleep patterns, daylight exposure, depression, and anxiety.
Self-rated questionnaires, including the Internet Addiction Proneness Scale-Short Form, Smartphone Addiction Scale–short version, Insomnia Severity Index, Morningness–Eveningness Questionnaire, Epworth Sleepiness Scale, sleep-related questionnaires, and Hospital Anxiety and Depression Scale, were administered in a study population of 1,610 Korean high school students. Multiple linear regression analyses were performed to evaluate the associations with internet addiction and smartphone addiction.
Multiple linear regression showed significant associations of internet addiction with poor sleep quality (β=0.094, p<0.001), eveningness (β=-0.071, p=0.006), less daylight exposure (β=-0.066, p=0.005), anxiety (β=0.252, p<0.001), and older age (β=0.295, p<0.001). Smartphone addiction was significantly associated with poor sleep quality (β=0.076, p=0.005), eveningness (β=-0.164, p<0.001), excessive daytime sleepiness (β=0.109, p<0.001), longer time in bed on weekdays (β=0.057, p=0.024), shorter weekend oversleep (β=-0.081, p=0.005), anxiety (β=0.192, p<0.001), and younger age (β=-0.145, p<0.001).
Our findings suggest that poor sleep quality, eveningness, and anxiety can serve as predictors of both internet addiction and smartphone addiction in adolescents. Internet addiction seems to be negatively associated with daylight exposure, whereas smartphone addiction showed significant associations with sleep patterns. Additional studies regarding the causal relations between these factors are needed.
Internet addiction has been shown to disrupt the daily life of adolescents. Excessive and uncontrollable internet use may lead to serious mental health problems, including sleep problems and emotional dysregulation, and may also have negative effects on academic function and social relationships [
Internet use has increasingly been taking the form of smartphone use, and the number of smartphone devices per person in South Korea reached 0.94 in 2021 [
Internet addiction is dangerous in itself in that individuals may be exposed to excessively provocative content over a long time, but smartphone addiction may have discriminating psychopathology in that smartphones enable easier access to media and allow the user to receive an instant response from others. Therefore, it is necessary to identify the disease-modifying factors associated not only with internet addiction, but also with smartphone addiction. To our knowledge, however, there have been no studies regarding the factors associated with both internet and smartphone addiction in adolescents.
This study was performed to evaluate the associations between sleep quality and internet/smartphone addiction independently in a population of adolescent Korean high school students. In addition, this study also investigated whether morningness-eveningness could predict internet/smartphone addiction, and assessed the associations of daytime sleepiness, daylight exposure, depression, and anxiety with internet/smartphone addiction.
This study was performed as part of a local child/adolescent sleep health survey conducted by Haman-gun Community Mental Health Center, Gyeongsang Province, South Korea. A total of 1,610 students ranging in age from 15 to 19 years were recruited from five high schools in Haman-gun. The participants were asked to complete questionnaires about problematic internet and smartphone use, emotional problems, and sleep patterns. Written informed consent was obtained from all students, parents, and teachers participating in the study. All study protocols were approved by the Institutional Review Board of Gyeongsang National University Hospital (IRB file number: 2015-06-004-004).
The Internet Addiction Proneness Scale–short form (IAPS-SF) is the short version of the Internet Addiction Proneness Scale, developed by the Korean Ministry of Science and ICT to improve its clinical applicability. The IAPS-SF consists of 20 items on seven subfactors: distance of adaptive functions, distance of reality testing, additive automatic thought, withdrawal, virtual interpersonal relationships, deviant behavior, and tolerance. Subjects can be classified according to the standardized total score into high-risk users (score: ≥70), potential-risk users (score: 63–69), or general users (score: <63). In the case of Korean high school students, a total score ≥52 points is classified as indicating high risk and a score ≥47 points is classified as indicating potential risk. The IAPS-SF has been standardized and verified for its validity and reliability in Korea [
The Smartphone Addiction Scale–short version (SAS-SV) is the short version of the original Smartphone Addiction Scale (SAS), which consists of 33 items that evaluate six characteristics of problematic smartphone use: disturbance of adaptive functions, addictive automatic thought, withdrawal, virtual interpersonal relationships, overuse, and tolerance. The SAS-SV has been standardized and verified for its validity and reliability in Korea [
The Hospital Anxiety and Depression Scale (HADS) includes two subscales: Hospital Anxiety and Depression-Anxiety (HAD-A) and Hospital Anxiety and Depression-Depression (HAD-D). A total of 14 items are included with seven items each for depression and anxiety subscales, which allows clinicians to assess the most common emotional symptoms rapidly and simply. Each item is scored from 0 to 3 on a 4-point Likert scale, with higher scores indicating higher levels of anxiety and depression [
All participants completed sleep-related questionnaires assessing information about sleep patterns and sleep problems. The surveys included sleep parameters such as time in bed (TIB) and total sleep time (TST), which were used to evaluate weekend oversleep and social jetlag, calculated by the absolute value of the difference in the midpoint of sleep times between weekdays and weekends. In addition, we collected data on the mean daytime sunlight exposure represented by the duration of any outside activity performed between 10:00 and 15:00 on weekdays and weekends. Such time around midday was considered because early morning and late evening light exposure both lead to circadian phase shifting, and the time interval of 10:00 to 15:00 is generally considered to be the optimal time of sunlight exposure for the effective synthesis of vitamin D [
The Insomnia Severity Index (ISI) assesses the quality of sleep on a seven-item scale. Each item is scored on a 5-point Likert scale (0–4), with a higher total score indicating poorer sleep quality. The ISI has been validated for the clinical measurement of insomnia severity [
The Epworth Sleepiness Scale (ESS), in which participants rate their chances of dozing off or falling asleep while engaged in eight different situations, is used to measure excessive daytime sleepiness. Each item is scored on a 4-point Likert scale (0–3), with a higher total score indicating more severe daytime sleepiness. An ESS score of 0–10 indicates normal daytime sleepiness and a score of 11–24 indicates a high level of daytime sleepiness for which medical attention is recommended. An ESS score ≥16 suggests sleep disorders, including insomnia, obstructive sleep apnea, and narcolepsy. We used the Korean version of the ESS, which has been validated in the Korean language [
The Morningness–Eveningness Questionnaire (MEQ) assesses the alertness and preferred time for sleep and waking in terms of circadian rhythm. The scale can be used to determine chronotype through 19 questions, with higher scores indicating a tendency for higher morningness and lower eveningness [
Correlations between continuous variables were analyzed according to the Pearson’s correlation coefficients. Differences between groups were evaluated by the independent t-test. Multiple linear regression analyses were performed to predict proneness for internet addiction controlling for age, sex, sleep parameters, sleep-related questionnaire scores, and depression/anxiety symptoms. Additional multiple linear regression analyses were performed to predict smartphone addiction. All of the analyses were performed using SPSS Windows Ver. 25 (IBM Corp., Armonk, NY, USA). In all analyses, p<0.05 was taken to indicate statistical significance.
The study population consisted of 1,610 students from five high schools in Haman-gun. The demographic characteristics, sleep parameters, and other clinical information are presented in
To identify predictive factors of internet addiction, a multiple linear regression analysis was conducted with age, sex, sleep parameters, and questionnaire scores as independent variables (
Another multiple linear regression analysis was performed with SAS-SV score as the dependent variable to identify predictive factors of smartphone addiction (
This study identified factors associated with internet and smartphone addiction among adolescents in Korea, including sleep quality, chronotype, other sleep-related factors, and anxiety. The findings suggest that poor sleep quality and eveningness are both associated with a higher risk of smartphone addiction and internet addiction. Higher anxiety was also associated with problems of both internet addiction and smartphone addiction. Students with greater smartphone addiction spent longer times in bed on weekdays and reported higher levels of daytime sleepiness. Students with greater internet addiction had less sunlight exposure during the daytime.
Consistent with several previous reports, the present study showed that poor sleep quality was significantly associated with internet addiction in adolescents. Kawabe et al. [
The present study also demonstrates that students with higher levels of anxiety are more prone to problematic internet use. Gao et al. [
Another important finding of our study is the association between sunlight exposure and internet addiction. Our results show that students with less daytime sunlight exposure during the weekends had higher levels of internet addiction. The role of daylight exposure in the circadian rhythm involving intrinsic photosensitive retinal ganglion cells is widely established to be associated with mood, cognition, and sleep [
The present study shows that poor sleep quality is significantly associated with smartphone addiction in adolescents. Two previous studies of medical school students in India showed a positive correlation between sleep quality and smartphone addiction [
In the present study, eveningness was also shown to predict smartphone addiction. One study of the association between chronotype and smartphone addiction proneness in German adolescents reported that eveningness as evaluated by the Composite Scale of Morningness was an important predictor of smartphone addiction, in which the association was even stronger than that of sleep duration [
In addition, the present study shows that excessive daytime sleepiness could also predict smartphone addiction. Although daytime sleepiness is highly associated with students’ academic performance, and more importantly with depression, a limited number of studies have approached the direct association between daytime sleepiness and smartphone addiction [
In the present study, students who spent more TIB during weekdays and those with shorter weekend oversleep times were more likely to have a smartphone addiction. Smartphone use in bed may contribute to more TIB, and a previous analysis showed that smartphone use before bedtime had both direct and indirect relations with smartphone addiction [
As with the results for internet addiction, higher anxiety was predictive of smartphone addiction in the present study. Whereas depression lacked statistical significance in predicting smartphone addiction in the present study, several previous reports concluded that both depression and anxiety are significantly associated with smartphone addiction. Matar Boumosleh and Jaalouk [
One strength of the present study is that it included sleep patterns, sleep characteristics, and behavioral factors associated with daytime activity. In particular, the present study included an evaluation of sunlight exposure, which was shown to be associated with internet addiction. In addition, most of the questionnaires used in the study were simple and consisted of a relatively small number of items, allowing for excellent accessibility and convenient implementation. Nonetheless, this study had several limitations. First, all variables assessed in the study were acquired using self-reported questionnaires. Complementary parental reports or more objective evaluation methods, such as actigraphy, would have provided higher quality information, but such assessments are limited due to practical constraints in large numbers of subjects in a school setting. Second, due to the relatively rural location of Haman-gun in Korea, it is possible that the results were affected by regional bias. Therefore, future investigations in different cities and different populations are needed. Third, because this was a cross-sectional study, the causal effect of each factor could not be determined.
This study identified poor sleep quality, eveningness, and several other sleep-related patterns or behaviors as factors significantly associated with internet and smartphone addiction in adolescents in Korea. In addition, the results presented here show that anxiety could predict both internet and smartphone addiction. Additional studies should focus on establishing the role of interventions in these factors in preventing or alleviating internet and smartphone addiction in adolescents.
This study was supported by a grant from the Korean Mental Health Technology R&D Project, Ministry of Health & Welfare, Republic of Korea (HM15C1108).
The authors have no potential conflicts of interest to disclose.
The datasets generated or analyzed during the study are available from the corresponding author on reasonable request.
Conceptualization: Nuree Kang, So-Jin Lee, Bong-Jo Kim. Data curation: Nuree Kang, Bong-Jo Kim, Jae-Won Choi, Young-Ji Lee, Eunji Lim. Formal analysis: Nuree Kang, Bong-Jo Kim, Cheol-Soon Lee, Boseok Cha, So-Jin Lee, Dongyun Lee. Funding acquisition: Nuree Kang, Bong-Jo Kim. Investigation: Nuree Kang, So-Jin Lee, Dongyun Lee, Jiyeong Seo, Jae-Won Choi, Young-Ji Lee, Eunji Lim. Methodology: Bong-Jo Kim, Cheol-Soon Lee, Boseok Cha, So-Jin Lee. Project administration: Nuree Kang, Bong-Jo Kim, Cheol-Soon Lee, Boseok Cha, So-Jin Lee, Jiyeong Seo. Resources: Dongyun Lee, Jiyeong Seo, Jae-Won Choi, Young-Ji Lee, Eunji Lim. Software: Dongyun Lee, Jiyeong Seo, Jae-Won Choi, Young-Ji Lee, Eunji Lim. Supervision: Bong-Jo Kim, Cheol-Soon Lee, Boseok Cha, So-Jin Lee, Dongyun Lee. Validation: Bong-Jo Kim, Cheol-Soon Lee, Boseok Cha, So-Jin Lee, Dongyun Lee. Visualization: Dongyun Lee, Jiyeong Seo, Jae-Won Choi, Young-Ji Lee, Eunji Lim. Writing—original draft: Nuree Kang, Bong-Jo Kim, So-Jin Lee. Writing—review & editing: Nuree Kang, Bong-Jo Kim, Cheol-Soon Lee, Boseok Cha, So-Jin Lee, Dongyun Lee.
Sociodemographic and clinical characteristics of the study population (n=1,610)
Characteristics | Value |
---|---|
Age (yr) | 16.91±0.89 |
Sex | |
Male | 739 (45.9) |
Female | 871 (54.1) |
IAPS-SF | 26.793±7.486 |
SAS-SV | 24.772±9.791 |
ISI | 6.824±4.711 |
MEQ | 44.755±7.746 |
ESS | 8.070±3.288 |
Sleep-related parameters | |
Weekend TIB (hr) | 8.640±2.536 |
Weekday TST (hr) | 6.018±1.401 |
Weekend TST (hr) | 8.453±1.954 |
Social jetlag (hr) | 2.840±1.679 |
Weekend oversleep (hr) | 2.436±2.166 |
Duration of sunlight exposure | |
Weekday sunlight exposure (hr) | 0.591±0.765 |
Weekend sunlight exposure (hr) | 1.097±1.232 |
Mean sunlight exposure duration (hr) | 0.735±0.778 |
HADS | |
HAD-A | 5.845±3.689 |
HAD-D | 6.042±3.359 |
Values are presented as mean±SD or number (percentage). IAPS-SF, Internet Addiction Proneness Scale–short form; SAS-SV, Smartphone Addiction Scale–short version; ISI, Insomnia Severity Index; MEQ, Morningness–Eveningness Questionnaire; ESS, Epworth Sleepiness Scale; TIB, time in bed; TST, total sleep time; HADS, Hospital Anxiety and Depression Scale; HAD-A, Hospital Anxiety and Depression-Anxiety; HAD-D, Hospital Anxiety and Depression-Depression
Multiple linear regression model predicting internet addiction
Independent variable | B | Standard error | β | p-value |
---|---|---|---|---|
Age | 4.425 | 0.348 | 0.295 | <0.001 |
Sex | 0.081 | 0.195 | 0.010 | 0.678 |
Weekend TIB | 0.095 | 0.143 | 0.021 | 0.507 |
Social jetlag | -0.140 | 0.096 | -0.041 | 0.145 |
Weekend oversleep | 0.058 | 0.083 | 0.020 | 0.487 |
Weekend sunlight exposure | -0.404 | 0.143 | -0.066 | 0.005 |
ESS | 0.052 | 0.056 | 0.023 | 0.353 |
ISI | 0.149 | 0.042 | 0.094 | <0.001 |
MEQ | -0.069 | 0.025 | -0.071 | 0.006 |
HAD-A | 0.511 | 0.057 | 0.252 | <0.001 |
HAD-D | -0.023 | 0.063 | -0.010 | 0.721 |
R2=0.179, adjusted R2=0.174, F=31.714 (p<0.001), n=1,608. TIB, time in bed; ESS, Epworth Sleepiness Scale; ISI, Insomnia Severity Index; MEQ, Morningness–Eveningness Questionnaire; HAD-A, Hospital Anxiety and Depression-Anxiety; HAD-D, Hospital Anxiety and Depression-Depression
Multiple linear regression model predicting smartphone addiction
Independent variable | B | Standard error | β | p-value |
---|---|---|---|---|
Age | -2.854 | 0.469 | -0.145 | <0.001 |
Sex | 0.152 | 0.261 | 0.014 | 0.560 |
Weekday TIB | 0.480 | 0.213 | 0.057 | 0.024 |
Social jetlag | 0.222 | 0.169 | 0.038 | 0.191 |
Weekend oversleep | -0.365 | 0.130 | -0.081 | 0.005 |
Total sunlight exposure | -0.382 | 0.308 | -0.030 | 0.215 |
ESS | 0.323 | 0.074 | 0.109 | <0.001 |
ISI | 0.158 | 0.056 | 0.076 | 0.005 |
MEQ | -0.208 | 0.034 | -0.164 | <0.001 |
HAD-A | 0.508 | 0.077 | 0.192 | <0.001 |
HAD-D | -0.155 | 0.085 | -0.053 | 0.069 |
R2=0.142, adjusted R2=0.136, F=23.949 (p<0.001), n=1,607. TIB, time in bed; ESS, Epworth Sleepiness Scale; ISI, Insomnia Severity Index; MEQ, Morningness–Eveningness Questionnaire; HAD-A, Hospital Anxiety and Depression-Anxiety; HAD-D, Hospital Anxiety and Depression-Depression