Journal of Adolescence 62 (2018) 38–46

Contents lists available at ScienceDirect

Journal of Adolescence
journal homepage: www.elsevier.com/locate/adolescence

Prevalence of internet addiction and its risk and protective factors

T
in a representative sample of senior high school students in Taiwan
Min-Pei Lina,∗, Jo Yung-Wei Wub, Jianing Youc, Wei-Hsuan Hub, Cheng-Fang Yend,e
a
Department of Educational Psychology and Counseling, National Taiwan Normal University, No.162, Sec. 1, Heping E. Rd., Da-an District, Taipei
City 106, Taiwan
b
Department of Counseling and Guidance, National University of Tainan, No.33, Sec. 2, Shu-Lin St., Tainan 700, Taiwan
c
Center for Studies of Psychological Application, Guangdong Key Laboratory of Mental Health and Cognitive Science, Research Center for Crisis
Intervention and Psychological Service of Guangdong Province, and School of Psychology, South China Normal University, No. 155 Zhongshan W. Rd,
Tianhe District, Guangzhou, China
d
Department of Psychiatry, Kaohsiung Medical University Hospital, No.100, Tzyou 1st Road, Kaohsiung 807, Taiwan
e
Department of Psychiatry, School of Medicine, and Graduate Institute of Medicine, College of Medicine, Kaohsiung Medical University, No.100,
Tzyou 1st Road, Kaohsiung 807, Taiwan

AR TI CLE I NF O AB S T R A CT

Keywords: The aim of this study investigated the prevalence of Internet addiction (IA) in a large re-
Adolescents presentative sample of secondary school students and identified the risk and protective factors.
Internet addiction Using a crosssectional design, 2170 participants were recruited from senior high schools
Prevalence throughout Taiwan using both stratified and cluster sampling. The prevalence of IA was 17.4%
Risk and protective factors
(95% confidence interval, 15.8%–19.0%). High impulsivity, low refusal self-efficacy of Internet
use, high positive outcome expectancy of Internet use, high disapproving attitude of Internet use
by others, depressive symptoms, low subjective well-being, high frequency of others' invitation to
Internet use, and high virtual social support was all independently predictive in the logistic re-
gression analysis. The prevalence of IA among secondary school students in Taiwan was high.
Results from this study can be used to help educational agencies and mental health organizations
create policies and design programs that will help in the prevention of IA in adolescents.

1. Introduction

The last decade has witnessed a dramatic proliferation of research on the newly emerging mental health problem of Internet
addiction (IA) (Kuss, Griffiths, & Binder, 2013; Lin, Ko, & Wu, 2011). IA is characterized by a maladaptive pattern of Internet use
leading to clinically significant impairment or distress (Weinstein & Lejoyeux, 2010). Frequently, Internet use is portrayed as an
inclusive medium without differentiating the types of online activities and applications that might be engaged by the users (Király,
Nagygyörgy, Koronczai, Griffiths, & Demetrovics, 2015). In accordance with the reviews conducted by Kuss, Griffiths, Karila, and
Billieux (2014), with the development in IA research, many researchers have adapted the diagnostic criteria for substance depen-
dence or/and pathologic gambling in the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV/DSM-IV-TR; American
Psychiatric Association, 1994, 2000), and modified the IA diagnostic criteria according to diagnostic interview results; suggesting an
IA cut-off point. Worldwide prevalence varied significantly; 0.8% of Italian high school students were considered to be seriously
addicted (Poli & Agrimi, 2012), and the prevalence of IA in Hong Kong adolescents ranged from 17% to 26.8% during the high school

∗
Corresponding author.
E-mail address: lmmpp@ntnu.edu.tw (M.-P. Lin).

https://doi.org/10.1016/j.adolescence.2017.11.004
Received 15 May 2017; Received in revised form 3 October 2017; Accepted 10 November 2017
0140-1971/ © 2017 The Foundation for Professionals in Services for Adolescents. Published by Elsevier Ltd. All rights reserved.
M.-P. Lin et al. Journal of Adolescence 62 (2018) 38–46

years (Shek & Yu, 2016). Although studies have yielded inconsistent prevalence, there is without a doubt that IA has emerged as a
rapidly growing problem in young people and has attracted worldwide attention (Wu et al., 2016). When compared to adults,
adolescence is a critical period for addiction vulnerability (Pallanti, Bernardi, & Quercioli, 2006). Adolescents have a natural ten-
dency towards the Internet and are more likely to adopt patterns of IA (Tsitsika et al., 2009). Their noticeable Internet literacy has
been linked to IA (Leung & Lee, 2012; Veen & Vrakking, 2006). Furthermore, adolescents typically have free and unlimited Internet
access, flexible living schedules, and freedom from parental interference (Kuss, Griffiths, et al., 2013). Generally speaking, IA is
common among adolescents, and associated risk and protective factors are found both at home and at school (Wu et al., 2016).
Therefore, it is important to assist adolescents from becoming addicted to Internet use, and close attention need to be given by
parents, teachers, and health care professionals regarding these risk and protective factors (Tang et al., 2014; Wu et al., 2016).
A large number of studies have examined the associations between IA and psychosocial risk and protective factors in adolescents.
Comprehensively, we incorporated the theory of triadic influence (TTI) (Flay & Petraitis, 1994; Petraitis, Flay, & Miller, 1995) as a
conceptual framework for understanding psychosocial risk and protective factors for IA. TTI integrates variables and processes from
many psychological and social theories of behavior onset and change, and it provides a unified theoretical framework that in-
corporates influential factors, or the causes of the behavior (Flay, Petraitis, & Hu, 1999). More recently, the TTI has been used as a
framework in research to examine adolescents' addictive behaviors (Chun, 2015; Defoe, Semon Dubas, Somerville, Lugtig, & van
Aken, 2016; Grigsby, Forster, Unger, & Sussman, 2016). The TTI organizes the influences on addictive behaviors into three sub-
stantive domains of influence (interpersonal stream, cultural/attitudinal stream, and intrapersonal stream) and discusses the extent to
which different factors influence behavior as ranging from very proximal to distal and ultimate (Grigsby et al., 2016).
According to the reviews conducted by Chun (2015) and Klein Velderman et al. (2015), proximal determinants are conceptualized
as rather addictive behavior-specific, being highly predictive for one addictive behavior, which include attitudes (an adolescent's
expectations and evaluations of the benefits of addictive behaviors, such as positive outcome expectancies), poor coping skills, and
lower refusal self-efficacy. Distal determinants of addictive behaviors are causes of addictive behaviors that are intermediate between
proximal and ultimate, and these include substance-specific behaviors of influential role models (such as knowing and observing the
smoking and/or drinking behavior of parents), general values that affect the attitudes of adolescents towards substance use, and
affective states related to internal motivation to substance uses, which include low self-esteem and emotional factors. Ultimate
determinants of addictive behavior are believed to also affect addictive behaviors, but to be almost unchangeable (i.e. more deeply
rooted), and these include an adolescent's immediate surroundings, such as neighbors, schools, and culture, including weak public
policies on substance use; characteristics of parents, family members, and other influential role models, as well as personal traits and
biological dispositions that may encourage an internal motivation for substance use (such as genetic susceptibility to substance
dependence, impulsivity, and emotional instability) (Flay et al., 1999; Petraitis et al., 1995).
On the other hand, previous studies have also indicated that behavioral factors (greater impulsivity) (Zhang et al., 2015), cog-
nitive factors (higher positive outcome expectancy of Internet use and lower refusal self-efficacy of Internet use) (Lin, Ko, & Wu,
2008; Lin et al., 2011; Wu, Ko, Tung, & Li, 2016a; Wu, Ko, Wong, Wu, & Oei, 2016b), emotional factors (more depressive symptoms)
(Lee, Shin, Cho, & Shin, 2014; Nie, Zhang, & Liu, 2017), self-esteem (Nie et al., 2017; Younes et al., 2016), alexithymia (Dalbudak
et al., 2013; Scimeca et al., 2014), and subjective well-being (Lai et al., 2015; Mei, Yau, Chai, Guo, & Potenza, 2016), are all
correlated with IA. Moreover, studies also discovered that social factors, including actual social support (Wegmann & Brand, 2016;
Wu et al., 2016), virtual social support (Yeh, Ko, Wu, & Cheng, 2008), social influence for Internet use (Wu et al., 2016b), and
accessibility (Ak, Koruklu, & Yilmaz, 2013; Lin et al., 2011; Wu et al., 2016), were associated with IA.
Although a number of psychosocial risk and protective factors have been associated with IA, few studies have surveyed a large
representative sample of senior high school students through a comprehensive investigation that incorporated psychological and
social factors targeting this population. Most studies, however, have been limited by relatively small or unrepresentative samples, or
have only examined a minimal number of psychosocial factors, which restricts the comprehensiveness to understand or compare the
related factors associated in the relationship. In educational practices, agencies and mental health organizations have limited re-
sources and time, thus, can only choose two to three important psychosocial factors at a time to focus on preventive education. For
this reason, it is important to integrate related psychosocial factors in IA, and probe at two to three critical psychosocial factors.
Therefore, the aim of this study was to examine the prevalence of IA and the associations between IA, and psychological and social
factors in a large representative sample of senior high school students in Taiwan. These findings can be used to help educational
agencies and mental health organizations create policies and design programs that will help in the prevention of IA in adolescents.

2. Methods

2.1. Participants and procedure

A cross-sectional design was applied in this study. Participants were recruited from senior high schools throughout Taiwan using
both stratified (by school type (i.e. regular high school or vocational high school) and clustered (by class) sampling. The 2011 school
year data provided by the Department of Statistics, Ministry of Education (https://stats.moe.gov.tw/qframe.aspx?qno=MQA1AA2)
showed a total of 401,958 regular high school students and 366,449 vocational high school students in Taiwan (ratio of 52: 48). In
accordance to the ratio, the present study invited two regular high schools and two vocational high schools during the time period of
October through December of 2013, and selected 2253 students (1152 regular high school students, yielding 51.13% of the total
sample size, which was close to the ratio of the population group). 2170 students (Mage = 15.83, SD = 0.38) participated in the final
data analysis, resulting in a response rate of 96.32%.

39
M.-P. Lin et al. Journal of Adolescence 62 (2018) 38–46

This study collected data through in-class survey administration. Full consent from the principals, and the counseling and gui-
dance team of each senior high school was obtained prior to conducting the study. In addition, teachers of the participating classes
gave consent before administrating the survey battery. Students were also fully informed of the purpose of the study and voluntary
participation was emphasized. Prior to administering the questionnaires, confidentiality of the surveys were highlighted, and written
consent forms were collected from the participants.

2.2. Measurement

2.2.1. The Chen Internet Addiction Scale

The Chen Internet Addiction Scale (CIAS; Chen, Weng, Su, Wu, & Yang, 2003) is a 26-item self-reported questionnaire with a 4-
point Likert scale, which has a Cronbach's alpha of 0.93 in this study. Containing five subscales, the CIAS includes “Compulsive Use of
Internet” (5 items), “Withdrawal Symptoms of IA” (5 items), “Tolerance Symptoms of IA” (4 items), “Interpersonal and Health-
Related Problems of IA” (7 items), and “Time Management Problems” (5 items). The internal consistency coefficients of each subscale
ranged from 0.78 to 0.81, and the split-half reliability over two weeks is 0.83 (Chen et al., 2003). Correlation analyses yielded
significantly positive correlations among the total scale and subscale scores of CIAS with weekly hours spent on Internet activity
(Chen et al., 2003).
Ko, Yen, Chen, Chen, and Yen (2005a) modified diagnostic criteria for pathologic gambling and substance dependence in DSM-IV-
TR (American Psychiatric Association, 2000) and developed the Diagnostic Criteria of Internet Addiction (DC-IA), which contained
nine characteristic symptoms of IA. The cut-off point of six among the nine characteristic symptoms yielded the best diagnostic
accuracy (95.4%) (Ko et al., 2005a). Furthermore, Ko et al. (2005b) recruited 468 high school students to complete both the CIAS and
diagnostic interview, and found that a 63/64 cut-off point of the CIAS was the optimal diagnostic cutoff point among high school
students in Taiwan, which gave the best diagnostic accuracy, Cohen Kappa, diagnostic odds ratio, and 87.6% of cases were correctly
classified. This discriminative potential makes the scale appropriate for use as a reliable diagnostic tool in a massive epidemiologic
survey, as it can provide the estimated prevalence and identify the high-risk case group (Ko et al., 2005b). As a result, those with CIAS
scores of 64 or more were classified as being Internet addicted in the current study.

2.2.2. The short-form of the impulsivity scale

The 15-item impulsivity scale used in this study was adapted from the Barratt Impulsivity Scale (BIS-10). Confirmatory factor
analysis revealed two subscales: motor impulsiveness and non-planning (Li, Ko, Weng, Liau, & Lu, 2002). This scale also demon-
strated satisfactory construct validity (Fu, Ko, Wu, Cherng, & Cheng, 2007). The internal consistency coefficient for this scale was
0.78 in this study.

2.2.3. Positive outcome expectancy of internet use questionnaire

This questionnaire consisted of 26 items and was measured with a 6-point Likert scale (Lin, Hu, Lin, & Wu, 2013). The total scores
ranged from 26 to 156 with a higher total score representing a greater degree in positive outcome expectancy of Internet use among
senior high school students. The internal consistency coefficient was 0.91 (Lin et al., 2013). The Cronbach's alpha coefficient was 0.94
in the present study.

2.2.4. Refusal self-efficacy of internet use questionnaire

This scale is a 19-item self-reported questionnaire, measured with a 6-point Likert scale (Lin et al., 2013). The total scores ranged
from 19 to 114 with lower total scores representing a lesser level in refusing self-efficacy of Internet use among senior high school
students. Previously, the internal consistency coefficient was found in be 0.92 (Lin et al., 2013), while the present study yielded a
consistency coefficient of 0.91.

2.2.5. Ko's depression inventory

This inventory was constructed in reference to various common depression scales (Chien, Ko, & Wu, 2007). It is a self-reported
scale consisting of 26 items, and the Cronbach's alpha coefficient of this inventory was 0.89 in the present study. Chiu, Ko, and Wu
(2007) also found this inventory to be positively correlated with the depression scale in the Symptom Check List-90-R (0.73) among
adolescents in Taiwan.

2.2.6. The Rosenberg Self-Esteem scale

This scale comprises of 10 items that evaluates global self-esteem. The scale yields a single overall score for self-esteem, and has
good reliability and construct validity (Rosenberg, 1965). Successfully used among adolescents in Taiwan (Chang, 2016), the Ro-
senberg Self-Esteem scale is rated on a 6-point scale, ranging from 1 = strongly disagree to 6 = strongly agree, and the internal
consistency coefficient was 0.91 in this study.

2.2.7. The Toronto Alexithymia Scale-20

The Toronto Alexithymia Scale-20 is a self-report questionnaire containing 20 items and rated on a 5-point Likert scale, ranging
from 1 = strongly disagree to 5 = strongly agree. This scale has good internal consistency, test-retest reliability, and factorial validity
(Bagby, Parker, & Taylor, 1994; Parker, Taylor, & Bagby, 2003). In the current study, the Cronbach's alpha coefficient for this scale
was 0.84.

40
M.-P. Lin et al. Journal of Adolescence 62 (2018) 38–46

2.2.8. Chinese Happiness Inventory

Subjective well-being was assessed using the Chinese Happiness Inventory, which is a 10 item self-report questionnaire rated on a
4-point scale. This inventory has been shown to have good reliability and validity among adolescents in Taiwan (Lu, 2006). In the
present study, this inventory demonstrated a good internal consistency (α = 0.90).

2.2.9. Social support scale

This scale comprises of 16 items that assesses perceived social support from parents and friends. Yeh et al. (2008) revealed a good
reliability and validity of this scale among adolescents in Taiwan. It was rated on a 4-point scale, ranging from 1 = do not apply to me
at all to 4 = apply to me very much or most of the time, and the internal consistency coefficient was 0.90 in the current study.

2.2.10. Virtual social support scale

This scale was revised from the Social Support Scale. It assesses perceived social support from those who are acquainted only
through the Internet as distinguished from those known in real life (Yeh et al., 2008). In addition, Yeh et al. (2008) added two
questions to the scale in accordance with the online environment. The final virtual social support scale has 10 items. The scale was
rated on a 4-point scale, ranging from 1 = do not apply to me at all to 4 = apply to me very much or most of the time, and the internal
consistency coefficient was 0.94 in the present study.

2.2.11. Social influence for internet use scale

This scale was revised from the Parental influence for IGA Scale and Peer influence for IGA Scale (Wu et al., 2016b). This scale
contained three subscales, including “Disapproving Attitude of Internet Use by Others,” “Perceived Frequency of Others' Internet
Use,” and “Frequency of Others' Invitation to Use the Internet.” 15 items were constructed to assess the social influence for Internet
use, including perceived use from parents, surrounding relatives, siblings, peers and best friend regarding the disapproving attitudes
toward Internet use (5 items); perceived use from parents, surrounding relatives, siblings, peers and best friend regarding Internet use
frequencies (5 items); and the frequency of parents, surrounding relatives, siblings, peers and best friend regarding invitations to use
the Internet (5 items). The three subscales were all rated on a 4-point scale, and higher scores indicated a stronger disapproving
attitude of Internet use, perceived higher frequencies of Internet use by others, and higher frequencies of invitations to use the
Internet by others, respectively. In this study, the Cronbach's alpha values of the three subscales were 0.76, 0.67, and 0.76, re-
spectively.

2.2.12. Internet use accessibility

Four items were constructed to assess Internet use accessibility, including “Can you use the Internet at home?” “Can you use the
Internet at school?” “Is there an Internet café near your home or school that you can go online?” and “Do you often come in contact
with media/advertisement that encourages you to use the Internet? The questionnaire is rated on a 4-point scale, ranging from
1 = never to 4 = always, and higher scores represented an easier accessibility to Internet use. Access to the Internet included using
desktops, notebooks, tablets, smart phones, and other mobile devices.

2.2.13. Statistical analysis

SPSS version 18.0 for Windows was used to analyze the data, and the significance level was set at 0.05. The prevalence for IA was
calculated. The association between IA in the groups and psychosocial factors were examined using t-test and Pearson correlation.
Significant factors were further selected and included in the forward logistic regression analysis to examine their associations with IA
among senior high school students, and to identify two to three critical psychosocial factors.

3. Results

IA was fulfilled based on the cutoff criteria in the CIAS total scores. Participants with CIAS total scores of 64 or more were
considered Internet addicts in the high school student population (Ko et al., 2005b). The diagnostic cutoff point has the optimal
diagnostic accuracy (87.6%), specificity (92.6%), Cohen Kappa value (0.61), and diagnostic odds ratio (26.17) to discriminate high
school students with IA from non-cases (Ko et al., 2005b). The discriminative potential of the CIAS makes it appropriate to serve as a
reliable diagnostic criterion for a one-stage massive epidemiologic research (Ko et al., 2005b). Therefore, CIAS provided the esti-
mated prevalence of IA and identified the Internet addicted group in this study. The prevalence of IA in the large representative
sample of senior high school students in Taiwan was 17.4% (95% confidence interval, 15.8%–19.0%).
The results in Table 1 showed that all psychological and social characteristics were significantly different between the Internet
addicted and non-Internet addicted groups. Moreover, the correlations displayed in Table 2 also revealed that IA diagnosis
(CIAS > 63) was significantly and positively correlated with impulsivity, positive outcome expectancy of Internet use, depressive
symptoms, alexithymia, virtual social support, disapproving attitude of Internet use by others, frequency of others' Internet use,
frequency of others' invitation to use Internet, and internet use accessibility, but were negatively related to refusal self-efficacy of
Internet use, self-esteem, subjective well-being, and actual social support, respectively.
Table 3 revealed that the following characteristics significantly increased the risk of IA: high impulsivity, high positive outcome
expectancy of Internet use, high disapproving attitude of Internet use by others, depressive symptoms, high frequency of others'
invitation to use the Internet, and high virtual social support. On the other hand, high refusal self-efficacy of Internet use and high
subjective well-being decreased the risk of IA. Impulsivity was the first variable to be entered in the regression model, followed by

41
M.-P. Lin et al. Journal of Adolescence 62 (2018) 38–46

Table 1
Psychosocial risk and protective factors among the Internet addicted and non-Internet addicted groups.

Risk and protective factor Internet addiction

Yes No t Effect size

(n = 377) (n = 1787)
Mean (SD) Mean (SD)

Psychological factor
Impulsivity 34.90 (6.43) 29.59 (5.98) −15.44∗∗∗ 0.83
Positive outcome expectancy of Internet use 124.19 (16.61) 115.98 (16.02) −8.96∗∗∗ 0.50
Refusal self-efficacy of Internet use 54.58 (20.78) 67.65 (18.11) 11.30∗∗∗ 0.68
Depressive symptoms 17.34 (10.23) 11.12 (8.72) −10.67∗∗∗ 0.67
Self-esteem 35.57 (8.37) 39.78 (8.98) 8.67∗∗∗ 0.47
Alexithymia 58.28 (9.84) 52.23 (10.72) −10.52∗∗∗ 0.56
Subjective well-being 22.13 (5.19) 25.15 (5.78) 9.43∗∗∗ 0.52
Social factor
Actual social support 45.62 (8.43) 47.15 (8.45) 3.17∗∗ 0.18
Virtual social support 18.59 (7.77) 16.23 (6.72) −5.42∗∗∗ 0.34
Disapproving attitude of Internet use by others 11.47 (2.01) 10.98 (2.28) −3.82∗∗∗ 0.22
Frequency of others' Internet use 15.45 (2.35) 14.95 (2.33) −3.72∗∗∗ 0.21
Frequency of others' invitation to use Internet 12.13 (2.71) 11.37 (2.93) −4.84∗∗∗ 0.26
Internet use accessibility 11.21 (2.39) 10.22 (2.46) −7.09∗∗∗ 0.40

∗∗
p < 0.01; ∗∗∗p < 0.001.
Note: The effect size is adopted from Cohen's, 1988 effect sizes46.

refusal self-efficacy and positive outcome expectancy of Internet use.

However, self-esteem, alexithymia, actual social support, frequency of others' Internet use, and Internet use accessibility were not
associated with IA in the forward logistic regression analysis. We added each factor separately into the logistic regression analysis,
and found that alexithymia (Wald. coeff. = 89.940, p < 0.001), frequency of others' Internet use (Wald. coeff. = 13.672,
p < 0.001), Internet use accessibility (Wald. coeff. = 47.922, p < 0.001), self-esteem (Wald. coeff. = 63.948, p < 0.001), and
actual social support (Wald. coeff. = 9.948, p < 0.01) were all able to independently and significantly predict IA, but was unable to
significantly predict IA when other psychosocial factors were added into the regression analyses.

4. Discussion

This study sought to uncover and examine the prevalence, and psychological and social predictors of IA in a large representative
sample of secondary school students in Taiwan. A CIAS cutoff score of 63/64 (Ko et al., 2005b) indicated that the prevalence of IA in
the sample examined in this study was 17.4% (95% confidence interval, 15.8%–19.0%). The prevalence appears to be near or even
lower than that found among a high school sample in Hong Kong (Shek & Yu, 2016) (17.0%–26.8% during the high school year), but
higher than the findings in a majority of researches targeting high school student samples, including Kilic, Avci, & Uzuncakmak's
(2016) high school student sample in Turkey (1.3%), Johansson & Götestam's (2004) Norwegian youth population (aged 12–18 years;
1.98%), Stavropoulos, Alexandraki, & Motti-Stefanidi's (2013) Greek high school student sample (3.1%), Kuss, Van Rooij, Shorter,
Griffiths, & van de Mheenb's (2013) adolescents in the Netherlands (aged 11–19 years; 3.7%), Pallanti et al.'s (2006) high school
student sample in Italy (5.4%), Chen et al.'s (2016) high school student sample in China (8.7%), Park, Kim, & Cho's (2008) high
school student sample in South Korean (10.7%), and Yang & Tung's (2007) high school student sample (13.8%) in Taiwan. According
to previous researches, prevalence varied widely depending on the different social and cultural contexts, as well as varying definitions
and criteria for measuring IA. In addition, the cut-off criteria for most studies were set without clinically established diagnostic
validity, which make it difficult to effectively compare findings (Lin et al., 2011). However, the present study used the CIAS to
establish a good diagnostic validity for the cut-off point, and yielded a high prevalence in IA. Such a result indicates that IA pre-
vention deserves to be given a greater attention among senior high school students in Taiwan.
Through forward logistic regression analyses, behavioral and cognitive factors, among other psychosocial categories, were found
to be the greatest contributing psychosocial factor in the development of IA. In educational practices, due to the limited resources and
time available among educational agencies and mental health organizations, only two to three psychosocial factors are incorporated
and focused in preventive education. Thus, the present study suggests that IA prevention among high school students should
prioritize in focusing on the behavioral and cognitive factors of IA.
Impulsivity was found to be associated with IA (Zhang et al., 2015). Our results showed that behavioral impulsivity made the
greatest contribution to the prediction of IA above and beyond the effects of other psychosocial factors. One explanation for this may
be that long-term benefits are often overlooked when faced with the immediate short-term gratifications of Internet use, and in turn,
increases the possibility of IA. Using delayed discounting tasks, Li et al. (2016) examined such a hypothesis and discovered that
Internet addicts discounted delayed amounts more steeply than the non-Internet addicted group, regardless of the reward sign and
monetary amount. As a result, when adolescents with high impulsivity come in contact with the Internet, their insensitivity to
delayed outcomes increases their risk of IA. Therefore, we suggest that campus prevention programs focus on educating students in

42
 M.-P. Lin et al.

Table 2
The correlations among variables.

Variables 1 2 3 4 5 6 7 8 9 10 11 12 13 14

1. Internet addiction Diagnosis (CIAS > 63) –

2. Impulsivity 0.32∗∗∗ –
3. Positive outcome expectancy of Internet use 0.19∗∗∗ 0.17∗∗∗ –
4. Refusal self-efficacy of Internet use –0.26∗∗∗ –0.18∗∗∗ –0.19∗∗∗ –
5. Depressive symptoms 0.25∗∗∗ 0.36∗∗∗ 0.01 –0.09∗∗∗ –
6. Self-esteem –0.18∗∗∗ –0.22∗∗∗ 0.09∗∗∗ 0.11∗∗∗ –0.58∗∗∗ –
7. Alexithymia 0.21∗∗∗ 0.35∗∗∗ 0.02 –0.17∗∗∗ 0.48∗∗∗ –0.53∗∗∗ –

43
8. Subjective well-being –0.20∗∗∗ –0.20∗∗∗ 0.09∗∗∗ 0.17∗∗∗ –0.50∗∗∗ 0.63∗∗∗ –0.47∗∗∗ –
9. Actual social support –0.07∗∗ –0.02 0.19∗∗∗ 0.11∗∗∗ –0.17∗∗∗ 0.32∗∗∗ –0.27∗∗∗ 0.42∗∗∗ –
10. Virtual social support 0.13∗∗∗ 0.11∗∗∗ 0.29∗∗∗ –0.11∗∗∗ 0.09∗∗∗ –0.03 0.06∗∗ 0.02 0.11∗∗∗ –
11. Disapproving attitude of Internet use by others 0.08∗∗∗ 0.04 –0.12∗∗∗ –0.06∗∗ 0.07∗∗ –0.13∗∗∗ 0.11∗∗∗ –0.16∗∗∗ –0.15∗∗∗ –0.04 –
12. Frequency of others' Internet use 0.08∗∗∗ 0.18∗∗∗ 0.24∗∗∗ –0.10∗∗∗ 0.06∗∗ 0.04 0.04 0.03 0.13∗∗∗ 0.08∗∗∗ –0.26∗∗∗ –
13. Frequency of others' invitation to use Internet 0.10∗∗∗ 0.11∗∗∗ 0.25∗∗∗ –0.12∗∗∗ 0.02 0.10∗∗∗ 0.02 0.12∗∗∗ 0.19∗∗∗ 0.11∗∗∗ –0.20∗∗∗ 0.36∗∗∗ –
14. Internet use accessibility 0.15∗∗∗ 0.22∗∗∗ 0.24∗∗∗ –0.13∗∗∗ 0.08∗∗ 0.05∗ 0.01 0.04 0.17∗∗∗ 0.12∗∗∗ –0.11∗∗∗ 0.24∗∗∗ 0.32∗∗∗ –
M 0.17 30.52 117.42 65.37 12.18 39.06 53.27 24.64 46.89 16.64 11.06 15.04 11.51 10.40
SD 0.38 6.38 16.43 19.24 9.29 9.02 10.81 5.80 8.46 6.97 2.24 2.34 2.91 2.47

∗ ∗∗ ∗∗∗
p < 0.05; p < 0.01; p < 0.001.
Journal of Adolescence 62 (2018) 38–46
M.-P. Lin et al. Journal of Adolescence 62 (2018) 38–46

Table 3
Forward logistic regression analyses in predicting Internet addiction from psychosocial risk and protective factors.

Risk and protective factor Wald χ2 OR 95% CI

∗∗∗
Impulsivity 60.383 1.095 1.071–1.121
Refusal self-efficacy of Internet use 49.174∗∗∗ 0.974 0.966–0.981
Positive outcome expectancy of Internet use 20.816∗∗∗ 1.022 1.012–1.031
Disapproving attitude of Internet use by others 19.507∗∗∗ 1.166 1.089–1.248
Depressive symptoms 15.417∗∗∗ 1.032 1.016–1.049
Subjective well-being 15.268∗∗∗ 0.944 0.916–0.971
Frequency of others' invitation to use Internet 8.561∗∗ 1.080 1.026–1.136
Virtual social support 7.947∗∗ 1.028 1.008–1.047

OR implies odd ratios; CI, confidence interval.

∗∗
p < 0.01; ∗∗∗p < 0.001.

the enhancement of delayed gratification and improve skills of building behavior plans in order to decrease the development of IA.
The results of this study also pointed out that several cognitive factors were crucial to understanding and preventing IA.
Regression analyses revealed that low refusal self-efficacy and high positive outcome expectancy of Internet use were important
contributing psychosocial factors in the development of IA among other psychosocial categories, which is consistent with findings
from previous studies (Lin et al., 2008, 2011; Wu et al., 2016a; Wu et al., 2016b). We suggest that high school IA prevention plans try
to reduce positive outcome expectancy with regard to Internet use by using methods such as multisession experiential expectancy
challenges, assistance or courses that teach refusal skills under high risk situations, and make efforts to enhance refusal self-efficacy
(Lin et al., 2011).
On the other hand, we also found that depressive symptoms, subjective well-being, frequency of others' invitation to use the
Internet, and high virtual social support made independent contributions to the prediction of IA above and beyond the effects of other
psychosocial factors, respectively. This finding is consistent with the findings of previous studies (Lee et al., 2014; Mei et al., 2016;
Wu et al., 2016b; Yeh et al., 2008). Interestingly, we found that the disapproving attitudes of Internet use by others positively
predicted IA, which is inconsistent with the findings from a previous study (Wu et al., 2016b). One explanation for this is that when
people in your surrounding environment (parents, surrounding relatives, siblings, peers and best friend) become aware of the
adolescent's IA problems, and start to hold a disapproving attitude and forget to use more favorable regulating methods, the issue is
likely to rebound just for the sake of opposition, and may exacerbate the adolescent's use of the Internet. However, since this study
design was cross-sectional and no causal relationship can be illustrated, further studies should examine whether disapproving atti-
tudes of Internet use by others is an important factor or index for determining IA.
This study revealed that alexithymia, frequency of others' Internet use, and Internet use accessibility were significantly higher,
and self-esteem and actual social support were significantly lower in the Internet addicted group compared to the non-Internet
addicted group, which corresponded with previous research on IA (Dalbudak et al., 2013; Lin et al., 2011; Nie et al., 2017; Scimeca
et al., 2014; Wu et al., 2016, 2016b; Yeh et al., 2008). Nevertheless, these factors were unable to significantly predict IA when entered
after other psychosocial factors in the regression analyses. It is probable that other psychosocial factors may have mediated the
relationship. Comprehensively, the main feature of this study falls on the incorporation of multiple psychosocial factors associated
with IA, and revealed that behavioral and cognitive factors greatly influenced IA compared to other factors.
A number of limitations should be noted. First, since the present study was a cross-sectional design, no causal relationships could
be examined. Second, all information was obtained from a paper and pencil self-report, which may result in a possibility of response
biases. In the future, interviews, informants, and multiple assessments might provide a richer and more thorough understanding of
IA. Third, although participants were recruited from senior high schools throughout Taiwan using both stratified (by school type (i.e.
regular high school or vocational high school) and clustered (by class) sampling, potential confounding effects due to the nesting of
the data have not been controlled (participants embedded within classrooms/schools/geographical regions). Fourth, a time range of
over ten years has passed since Ko et al. (2005b)’s research on the IA diagnosis being a 63/64 cut-off point of the CIAS, which may
result in an overestimate of the IA prevalence. Future studies can reexamine the IA cut-off point that is more suitable to the present
generation. Finally, the use of a secondary school student sample limits the generalizability of results, and findings may not be
applicable to all Internet addicted individuals of similar age who do not attend school.
Despite limitations, this study examined the prevalence of IA and investigated the associations between IA and a variety of
psychosocial factors from a large representative sample of secondary school students in Taiwan. Our study suggests that IA is fairly
prevalent among secondary school students in Taiwan, and in particularly, behavioral and cognitive risk factors should be the focus of
IA preventive strategies when targeting this age group population. Our results may help educational agencies and mental health
organizations design suitable IA prevention programs geared toward the high school population.

Disclosure statement

All the authors declare that they have no conflict of interests.

44
M.-P. Lin et al. Journal of Adolescence 62 (2018) 38–46

Acknowledgements

This research was funded in part by the Ministry of Science and Technology, Taiwan (Grant No. MOST 106-2511-S-003-034-MY2
and MOST 102-2511-S-003-016-MY3), the National Natural Science Foundation of China (Grant No. 31771228), the National Social
Science Foundation of China (Grant No. 14ZDB159), and the Major Projects of the Humanities and Social Science Research Base of
Ministry of Education (Grant No. 16JJD190001). This study was also supported by Research Center for Crisis Intervention and
Psychological Service of Guangdong Province, South China Normal University and the Base of psychological services and counseling
for "Happiness" in Guangzhou.

References

Ak, S., Koruklu, N., & Yilmaz, Y. (2013). A study on Turkish adolescent's Internet use: Possible predictors of Internet addiction. Cyberpsychology, Behavior, and Social
Networking, 16, 205–209. https://doi.org/10.1089/cyber.2012.0255.
American Psychiatric Association (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: American Psychiatric Publishing.
American Psychiatric Association (2000). Diagnostic and statistical manual of mental disorders, 4th edition, text revision. Washington, DC: American Psychiatric
Publishing.
Bagby, R. M., Parker, J. D. A., & Taylor, G. J. (1994). The twenty-item Toronto alexithymia scale: I. Item selection and cross-validation of the factor structure. Journal of
Psychosomatic Research, 38, 23–32. https://doi.org/10.1016/0022-3999(94)90005-1.
Chang, S. H. (2016). Testing a model of codependency for college students in Taiwan based on Bowen's concept of differentiation. International Journal of Psychology.
Epub ahead of print https://doi.org/10.1002/ijop.12271.
Chen, Y., Kang, Y., Gong, W., He, L., Jin, Y., Zhu, X., et al. (2016). Investigation on internet addiction disorder in adolescents in Anhui, People's Republic of China.
Neuropsychiatric Disease and Treatment, 12, 2233–2236. https://doi.org/10.2147/NDT.S110156.
Chen, S. H., Weng, L. C., Su, Y. J., Wu, H. M., & Yang, P. F. (2003). Development of Chinese Internet addiction scale and its psychometric study. Chinese Journal of
Psychology, 45, 279–294.
Chien, L. L., Ko, H. C., & Wu, J. Y. (2007). The five-factor model of personality and depressive symptoms: One-year follow-up. Personality and Individual Differences, 43,
1013–1023. https://doi.org/10.1016/j.paid.2007.02.022.
Chiu, S. H., Ko, H. C., & Wu, J. Y. (2007). Depression moderated the effect of exposure to suicide news on suicidality among college students in Taiwan. Suicide & Life-
threatening Behavior, 37, 585–592. https://doi.org/10.1521/suli.2007.37.5.585.
Chun, J. S. (2015). Determinants of tobacco use among Korean female adolescents: Longitudinal test of the theory of triadic influence. Children and Youth Services
Review, 50, 83–87. https://doi.org/10.1016/j.childyouth.2015.01.009.
Cohen, J. (1988). Statistical power analysis for the behavioral science ( (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum.
Dalbudak, E., Evren, C., Aldemir, S., Coskun, K. S., Ugurlu, H., & Yildirim, F. G. (2013). Relationship of Internet addiction severity with depression, anxiety, and
alexithymia, temperament and character in university students. Cyberpsychology, Behavior, and Social Networking, 16, 272–278. https://doi.org/10.1089/cyber.
2012.0390.
Defoe, I. N., Semon Dubas, J., Somerville, L. H., Lugtig, P., & van Aken, M. A. (2016). The unique roles of intrapersonal and social factors in adolescent smoking
development. Developmental Psychology, 52, 2044–2056. https://doi.org/10.1037/dev0000198.
Flay, B. R., & Petraitis, J. (1994). The influence of triadic influence: A new theory of health behavior with implications for preventive intervention. In G. Albrecht (Vol.
Ed.), Advances in medical sociology: A reconsideration of health change models: Vol. 4, (pp. 19–44). Greenwich, CT: JAI Press.
Flay, B. R., Petraitis, J., & Hu, F. B. (1999). Psychosocial risk and protective factors for adolescent tobacco use. Nicotine & Tobacco Research, 1, S59–S65.
Fu, A. T., Ko, H. C., Wu, J. Y., Cherng, B. L., & Cheng, C. P. (2007). Impulsivity and expectancy in risk for alcohol use: Comparing male and female college students in
Taiwan. Addictive Behaviors, 32, 1887–1896. https://doi.org/10.1016/j.addbeh.2007.01.003.
Grigsby, T. J., Forster, M., Unger, J. B., & Sussman, S. (2016). Predictors of alcohol-related negative consequences in adolescents: A systematic review of the literature
and implications for future research. Journal of Adolescence, 48, 18–35. http://dx.doi.org/10.1016/j.adolescence.2016.01.006.
Johansson, A., & Götestam, K. G. (2004). Internet addiction: Characteristics of a questionnaire and prevalence in Norwegian youth (12-18 years). Scandinavian Journal
of Psychology, 45, 223–229. https://doi.org/10.1111/j.1467-9450.2004.00398.x.
Kilic, M., Avci, D., & Uzuncakmak, T. (2016). Internet addiction in high school students in Turkey and multivariate analyses of the underlying factors. Journal of
Addictions Nursing, 27, 39–46. https://doi.org/10.1097/JAN.0000000000000110.
Király, O., Nagygyörgy, K., Koronczai, B., Griffiths, M. D., & Demetrovics, Z. (2015). Assessment of problematic Internet use and online video gaming. In E.
Aboujaoude, & V. Starcevic (Eds.). Mental health in the digital age: Grave dangers, great promise (pp. 46–68). Oxford: Oxford University Press.
Klein Velderman, M., Dusseldorp, E., van Nieuwenhuijzen, M., Junger, M., Paulussen, T. G., & Reijneveld, S. A. (2015). Cultural, social and intrapersonal factors
associated with clusters of co-occurring health-related behaviours among adolescents. European Journal of Public Health, 25, 31–37. http://dx.doi.org/10.1093/
eurpub/cku051.
Ko, C. H., Yen, J. Y., Chen, C. C., Chen, S. H., & Yen, C. F. (2005a). Proposed diagnostic criteria of Internet addiction for adolescents. The Journal of Nervous and Mental
Disease, 193, 728–733. https://doi.org/10.1097/01.nmd.0000185891.13719.54.
Ko, C. H., Yen, J. Y., Yen, C. F., Chen, C. C., Yen, C. N., & Chen, S. H. (2005b). Screening for internet addiction: An empirical study on cut-off points for the Chen
internet addiction scale. Kaohsiung Journal of Medical Sciences, 21, 545–551.
Kuss, D. J., Griffiths, M. D., & Binder, J. F. (2013a). Internet addiction in students: Prevalence and risk factors. Computers in Human Behavior, 29, 959–966. https://doi.
org/10.1016/j.chb.2012.12.024.
Kuss, D. J., Griffiths, M. D., Karila, L., & Billieux, J. (2014). Internet addiction: A systematic review of epidemiological research for the last decade. Current
Pharmaceutical Design, 20, 4026–4052.
Kuss, D. J., Van Rooij, A. J., Shorter, G. W., Griffiths, M. D., & van de Mheenb, D. (2013b). Internet addiction in adolescents: Prevalence and risk factors. Computers in
Human Behavior, 29, 1987–1996. https://doi.org/10.1016/j.chb.2013.04.002.
Lai, C. M., Mak, K. K., Watanabe, H., Jeong, J., Kim, D., Bahar, N., et al. (2015). The mediating role of internet addiction in depression, social anxiety, and psychosocial
well-being among adolescents in six Asian countries: A structural equation modelling approach. Public Health, 129, 1224–1236. https://doi.org/10.1016/j.puhe.
2015.07.031.
Lee, J. Y., Shin, K. M., Cho, S. M., & Shin, Y. M. (2014). Psychosocial risk factors associated with Internet addiction in Korea. Psychiatry Investigation, 11, 380–386.
https://doi.org/10.4306/pi.2014.11.4.380.
Leung, L., & Lee, P. S. N. (2012). The influences of information literacy, Internet addiction and parenting styles on internet risks. New Media & Society, 14, 117–136.
https://doi.org/10.1177/1461444811410406.
Li, C. H., Ko, H. C., Weng, L. J., Liau, L. C., & Lu, R. B. (2002). The development of an impulsiveness scale: Psychometric properties and relation to antisocial
personality disorder. Chinese Journal of Psychology, 44, 109–119.
Lin, M. P., Hu, W. H., Lin, Y. P., & Wu, Y. W. (2013). Internet use related cognitions scale: Development, confirmatory factor validation and psychometric properties.
Paper presented at the 44th society for psychotherapy research international annual meeting, Brisbane, Australia.
Lin, M. P., Ko, H. C., & Wu, J. Y. (2008). The role of positive/negative outcome expectancy and refusal self-efficacy of Internet use on Internet addiction among college
students in Taiwan. Cyberpsychology, Behavior, and Social Networking, 11, 451–457. https://doi.org/10.1089/cpb.2007.0121.
Lin, M. P., Ko, H. C., & Wu, J. Y. (2011). Prevalence and psychosocial risk factors associated with internet addiction in a nationally representative sample of college

45
M.-P. Lin et al. Journal of Adolescence 62 (2018) 38–46

students in Taiwan. Cyberpsychology, Behavior, and Social Networking, 14, 741–746. https://doi.org/10.1089/cyber.2010.0574.
Li, Q., Tian, M., Taxer, J., Zheng, Y., Wu, H., Sun, S., et al. (2016). Problematic Internet users' discounting behaviors reflect an inability to delay gratification, not risk
taking. Cyberpsychology, Behavior, and Social Networking, 19, 172–178. https://doi.org/10.1089/cyber.2015.0295.
Lu, L. (2006). Cultural fit: Individual and societal discrepancies in values, beliefs and SWB. Journal of Social Psychology, 146, 203–221. https://doi.org/10.3200/SOCP.
146.2.203-221.
Mei, S., Yau, Y. H., Chai, J., Guo, J., & Potenza, M. N. (2016). Problematic Internet use, well-being, self-esteem and self-control: Data from a high-school survey in
China. Addictive Behaviors, 61, 74–79. https://doi.org/10.1016/j.addbeh.2016.05.009.
Nie, J., Zhang, W., & Liu, Y. (2017). Exploring depression, self-esteem and verbal fluency with different degrees of Internet addiction among Chinese college students.
Comprehensive Psychiatry, 72, 114–120. https://doi.org/10.1016/j.comppsych.2016.10.006.
Pallanti, S., Bernardi, S., & Quercioli, L. (2006). The shorter PROMIS questionnaire and the Internet addiction scale in the assessment of multiple addictions in a high-
school population: Prevalence and related disability. CNS Spectrums, 11, 966–974. https://doi.org/10.1017/S1092852900015157.
Parker, J. D. A., Taylor, G. J., & Bagby, R. M. (2003). The 20-item Toronto Alexithymia Scale III. Reliability and factorial validity in a community population. Journal of
Psychosomatic Research, 55, 269–275. https://doi.org/10.1016/S0022-3999(02)00578-0.
Park, S. K., Kim, J. Y., & Cho, C. B. (2008). Prevalence of Internet addiction and correlations with family factors among South Korean adolescents. Adolescence, 43,
895–909.
Petraitis, J., Flay, B. R., & Miller, T. Q. (1995). Reviewing theories of adolescent substance use: Organizing pieces in the puzzle. Psychological Bulletin, 117, 67–86.
https://doi.org/10.1037/0033-2909.117.1.67.
Poli, R., & Agrimi, E. (2012). Internet addiction disorder: Prevalence in an Italian student population. Nordic Journal of Psychiatry, 66, 55–59. http://dx.doi.org/10.
3109/08039488.2011.605169.
Rosenberg, M. (1965). Society and the adolescent self-image. New Jersey: Princeton University Press.
Scimeca, G., Bruno, A., Cava, L., Pandolfo, G., Muscatello, M. R., & Zoccali, R. (2014). The relationship between alexithymia, anxiety, depression, and Internet
addiction severity in a sample of Italian high school students. Scientific World Journal, 2014(504376)https://doi.org/10.1155/2014/504376.
Shek, D. T. L., & Yu, L. (2016). Adolescent Internet addiction in Hong Kong: Prevalence, change, and correlates. Journal of Pediatric and Adolescent Gynecology, 29,
S22–S30. https://doi.org/10.1016/j.jpag.2015.10.005.
Stavropoulos, V., Alexandraki, K., & Motti-Stefanidi, F. (2013). Recognizing Internet addiction: Prevalence and relationship to academic achievement in adolescents
enrolled in urban and rural Greek high schools. Journal of Adolescence, 36, 565–576. https://doi.org/10.1016/j.adolescence.2013.03.008.
Tang, J., Yu, Y., Du, Y., Ma, Y., Zhang, D., & Wang, J. (2014). Prevalence of Internet addiction and its association with stressful life events and psychological symptoms
among adolescent Internet users. Addictive Behaviors, 39, 744–747. https://doi.org/10.1016/j.addbeh.2013.12.010.
Tsitsika, A., Critselis, E., Kormas, G., Filippopoulou, A., Tounissidou, D., Freskou, A., et al. (2009). Internet use and misuse: A multivariate regression analysis of the
predictive factors of internet use among Greek adolescents. European Journal of Pediatrics, 168, 655–665. https://doi.org/10.1007/s00431-008-0811-1.
Veen, W., & Vrakking, B. (2006). Homo zappiens: Growing up in a digital age. London: Network Continuum Education.
Wegmann, E., & Brand, M. (2016). Internet-communication disorder: It's a matter of social aspects, coping, and Internet-use expectancies. Frontiers in Psychology,
7(1747)https://doi.org/10.3389/fpsyg.2016.01747.
Weinstein, A., & Lejoyeux, M. (2010). Internet addiction or excessive Internet use. American Journal of Drug and Alcohol Abuse, 36, 277–283. https://doi.org/10.3109/
00952990.2010.491880.
Wu, J. Y. W., Ko, H. C., Tung, Y. Y., & Li, C. C. (2016a). Internet use expectancy for tension reduction and disinhibition mediates the relationship between borderline
personality disorder features and Internet addiction among college students – one-year follow-up. Computers in Human Behavior, 55, 851–855.
Wu, J. Y. W., Ko, H. C., Wong, T. Y., Wu, L. A., & Oei, T. P. (2016b). Positive outcome expectancy mediates the relationship between peer influence and Internet
gaming addiction among adolescents in Taiwan. Cyberpsychology, Behavior, and Social Networking, 19, 49–55. https://doi.org/10.1089/cyber.2015.0345.
Wu, X. S., Zhang, Z. H., Zhao, F., Wang, W. J., Li, Y. F., Bi, L., et al. (2016). Prevalence of Internet addiction and its association with social support and other related
factors among adolescents in China. Journal of Adolescence, 52, 103–111. https://doi.org/10.1016/j.adolescence.2016.07.012.
Yang, S. C., & Tung, C. J. (2007). Comparison of Internet addicts and non-addicts in Taiwanese high school. Computers in Human Behavior, 23, 79–96. https://doi.org/
10.1016/j.chb.2004.03.037.
Yeh, Y. C., Ko, H. C., Wu, J. Y. W., & Cheng, C. P. (2008). Gender differences in relationships of actual and virtual social support to Internet addiction mediated through
depressive symptoms among college students in Taiwan. Cyberpsychology, Behavior, and Social Networking, 11, 485–487. https://doi.org/10.1089/cpb.2007.0134.
Younes, F., Halawi, G., Jabbour, H., El Osta, N., Karam, L., Hajj, A., et al. (2016). Internet addiction and relationships with insomnia, anxiety, depression, stress and
self-esteem in university students: A cross-sectional designed study. PLoS One, 11, e0161126. https://doi.org/10.1371/journal.pone.0161126.
Zhang, Y., Mei, S., Li, L., Chai, J., Li, J., & Du, H. (2015). The relationship between impulsivity and internet addiction in Chinese college students: A moderated
mediation analysis of meaning in life and self-esteem. PLoS One, 10, e0131597. https://doi.org/10.1371/journal.pone.0131597.

46