Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
The Internet activities, gratifications, and health consequences related to compulsive Internet use
(USC Thesis Other)
The Internet activities, gratifications, and health consequences related to compulsive Internet use
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
The Internet Activities, Gratifications, and Health Consequences Related to
Compulsive Internet Use
A Dissertation Presented to the
Faculty of the USC Graduate School
University of Southern California
In Partial Fulfillment of the
Requirements for the Degree
Doctor of Philosophy in Health Behavior Research
Institute for Health Promotion and Disease Prevention Research
Department of Preventive Medicine
University of Southern California
By
Jennifer Y. Tsai, M.P.H.
August 2018
i
Acknowledgements
This work would not have been possible without the village of people who mentored and
supported me throughout this process. My sincerest gratitude to my committee members: Drs.
Steve Sussman for encouraging me to strive for excellence and for teaching and showing me
what it is to be a great researcher (and person); Luanne Rohrbach for grounding me and offering
perspective and a sense of balance; Jimi Huh for her patience and continued stats guidance no
matter how many times I did not get it – my dissertation would be “insignificant” without your
mentorship; Adam Leventhal for his support and pushing me to think critically; and Julie
Cederbaum for her kind words and encouragement.
I would also like to extend my gratitude to other PM faculty, staff, and students. Dr.
Michael Cousineau has been instrumental in my professional and individual growth. His
generosity to his students and passion and dedication to his work and society have been
inspirational. Special thanks to Drs. Sue Kim and Kim Miller who constantly reminded me that
everything is going to be ok; Leah Medi whose cheerfulness and optimism always brightens my
mood; the TND and H&H teams for their relentless pursuit of data; the study participants for
their willingness to answer surveys and share their experiences; the MPH staff for never
questioning how many parking passes I needed and sneaking me free food – they have made my
life so much easier and heartier; the other HBR students for making Soto fun; and Marny
Barovich who not only helped me navigate PhD life but has become an invaluable confidante –
no one can acquire signatures as efficiently as she can.
Finally, I want to thank my family and friends. They provided laughter when I was in
tears and perspective when things got rough and have been my greatest advocates. To my
parents, I thank them for all the emotional and financial sacrifices they have made for me; I
promise retirement is near! More importantly, I thank my parents for instilling in me a strong
work ethic, strength in the face of adversity, compassion for others, and integrity of character,
without which I would not have made it this far.
ii
Table of Contents
Acknowledgements i
Table of Contents ii
List of Tables & Figures iv
Abstract v
SPECIFIC AIMS 1
CHAPTER 1: BACKGROUND AND SIGNIFICANCE 4
Definition of CIU 4
Prevalence of CIU among Adolescents 5
Adolescent Susceptibility to CIU 6
Sociodemographic Correlates 8
Mental Health and Neurobiological Correlates 9
Consequences associated with CIU 10
Theories used in CIU 11
Limitations to Current CIU Research 12
Introduction to Dissertation Studies 13
Uses and Gratification Theory (UGT) 14
UGT and CIU 16
Critiques and Limitations of UGT 17
CHAPTER 2: Identifying subgroups of Internet users and its association with CIU
and gender among a sample of high risk adolescents 19
Background 19
Methods 22
Research Questions 22
Data Collection 23
Measures 24
Analytic Plan 25
Results 27
Sample Characteristics 27
Assessment of Missing and Excluded Data 28
LCA Results 29
LCA w/ Covariates 33
Unadjusted & Adjusted Regression Models 36
Discussion 40
Limitations 44
Conclusion 46
CHAPTER 3: Evaluation of Pleasure & CIU using the Electronic Pleasurable Events
Schedule (E-PES)
47
Background 47
Methods 50
Research Questions 50
Data Collection 50
Measures 51
iii
Analytic Plan 54
Results 57
Sample 57
Assessment of Missing Data & Attrition 59
Results of EFA & CFA 59
Results of MLM 63
Discussion 66
Limitations 72
Conclusion 73
CHAPTER 4: Examining the Association between CIU and Alcohol, Tobacco,
Marijuana, and Other Drug Use Using a Latent Transition Analysis
75
Background 75
Methods 80
Research Questions 80
Data Collection 80
Measures 81
Analytic Plan 82
Results 84
Sample 82
Assessment of Missing Data & Attrition 82
LCA & LTA Results 85
Discussion 89
Limitations 92
Conclusion 94
CHAPTER 5: OVERALL DISCUSSION & CONCLUSION 95
Summary of Findings 95
Theoretical Implications 96
Methodological Implications 97
Programmatic Implications 97
Overall Limitations 99
Overall Conclusion 100
LITERATURE CITED 101
iv
List of Tables & Figures
List of Tables
Table 1 Model Fit Indices and Chosen Model for LTA of Last 7-day Internet
Activities 31
Table 2. Results of LCA Retaining Four Classes of Last 7 Day Internet Activities 32
Table 3. Results of LCA with Covariates 35
Table 4. Unadjusted and Adjusted Linear Regression Models 38
Table 5. Study 2 Sample Characteristics 58
Table 6. Results of E-PES Exploratory Factor Analysis 61
Table 7. Results of E-PES Confirmatory Factor Analysis 62
Table 8. Results of E-PES MLM Models 65
Table 9. Prevalence of Internet Addicts and Last 30-Day Alcohol, Nicotine,
Marijuana, and Other Drug Use 85
Table 10. Model Fit Indices and Chosen Model for LTA 87
Table 11. Results of LTA 88
List of Figures
Figure 1. Heuristic model of proposed research 14
Figure 2. Conceptual model of Latent Class Analysis 27
Figure 3. Latent Class Probabilities for Internet activities 30
Figure 4. Conceptual Model for Study 2 57
Figure 5. Results of Confirmatory Factor Analysis 62
Figure 6. Conceptual model of LTA 79
v
ABSTRACT
Background: The proliferation of the Internet and Internet related applications has consequently
increased the concern for compulsive internet use (CIU), particularly among adolescents. CIU
shares similar qualities and definitions with other substance and behavioral addictions, which are
often characterized by 1) preoccupation, 2) mood alterations, 3) tolerance, and 4) withdrawal.
CIU has been linked with several mental health issues, such as depression, anhedonia, and
ADHD and behavioral consequences, such as substance use, alcohol use, hostility, and suicidal
ideation. Understanding the risks and consequences associated with CIU may play an important
factor not only in CIU prevention but in the prevention of other associated mental and behavioral
problems.
Methods: This dissertation was guided by the Uses and Gratifications Theory. Study 1 assessed
CIU and Internet activities using latent class analysis among a sample of continuation high
school students participating in Project Towards No Drugs Abuse in Los Angeles, CA, USA.
Study 2 and Study 3 used data from the Happiness and Health Study, a study surveying regular,
youths attending high schools in Los Angeles, CA, USA. Study 2 assessed pleasure derived from
Internet use and its association with CIU using multi-level modeling. Study 3 assessed the
longitudinal association between CIU and substance use using latent transition analysis.
Results: Study 1 found that generalized Internet use and use of the Internet for specific activities
were associated with greater CIU. Study 2 found certain gratifications were associated with CIU
onset but not with changes in CIU. These relationships are moderated by anhedonia and gender
cross-sectionally but not longitudinally. Over time, baseline CIU and identifying as “Hispanic”
and “Other” ethnicity were the only variables that significantly predicted changes in CIU one
year later. Study 3 found Internet Addiction (IA) was an independent behavioral consequence as
well as a consequence associated with substance use behaviors among subsets of adolescents. IA
was a relatively stable behavior among adolescents exhibiting IA only, whereas those who
exhibited IA and substance use problems were more likely to change behaviors at follow-up.
Conclusion: These findings contribute to our understanding of factors that may be associated
with CIU etiology and consequence. These results suggest the importance of CIU prevention at
an early age and provide insight on how best to tailor CIU prevention and intervention programs,
as well as substance use programs, most effectively.
1
SPECIFIC AIMS
The proliferation of the Internet and Internet related applications have consequently
increased the concern for compulsive Internet use (CIU), also called Internet addiction (IA) and
problematic Internet use (PIU), particularly among adolescents. CIU shares similar qualities and
definitions with other substance and behavioral addictions which are often characterized by: 1)
appetitive effects, 2) satiation, 3) preoccupation, 4) loss of control, and 5) negative consequences
(Sussman, 2017). CIU has been linked with several mental health issues, such as depression
(Evren, Dalbudak, Evren, & Ciftci Demirci, 2014; Kimet al., 2006; Lin, Ko, & Wu, 2011),
anhedonia (Guillot et al., 2016), and ADHD (Bernardi & Pallanti, 2009; Evren et al., 2014), and
behavioral consequences, such as substance use (Liu, Desai, Krishnan-Sarin, Cavallo, &
Potenza, 2011), alcohol use (Chiao, Yi, & Ksobiech, 2014; Sun et al., 2012), hostility (Ko, Yen,
Yen, Chen, & Chen, 2012), and suicidal ideation (Kim et al., 2006). Understanding the risks and
consequences associated with IA may play an important factor not only in CIU prevention but in
the prevention of other associated mental and behavioral problems.
To address identified gaps in the in current CIU research, the proposed dissertation sought to
understand the activities that may instigate CIU among U.S. adolescents and examine the long-
term psychosocial and behavioral factors associated with CIU. Findings from the proposed
studies contribute to a more comprehensive understanding of the etiology of CIU and its
associated consequences.
Study 1: Identifying subgroups of Internet users and its association with CIU and gender
among a sample of high risk adolescents
Various activities are predicted to be related to CIU, but the clustering of these activities
among individual compulsive internet users is not well known. Identifying distinct subgroups of
adolescents who identified similar patterns of use and CIU highlight risk factors for CIU
2
screening and can be important to tailoring appropriate and effective CIU prevention and
intervention programs.
Aim 1: To conduct a latent class analysis (LCA) among a sample of continuation high
school students participating in project Towards No Drug Abuse (TND) to distinguish
subgroups of adolescents with similar patterns of Internet use behavior.
Hypothesis 1: Distinct subgroups of adolescents that share patterns of Internet
use behavior will be identified in the sample.
Aim 2: To examine differences in LCA class membership by CIU and gender.
Hypothesis 2: Conditional probabilities of belonging to the identified subgroups
will vary by CIU and gender. Aligned with the general consensus among the
literature, CIU will be associated with classes endorsing more Internet activities
and social interaction. Similar to studies in the literature showing that females and
males differ on their Internet use, we predict that females are more likely to be
classified in subgroups where Internet activities promote social interaction
specifically and males are more likely to be classified in subgroups that generally
endorse more types of Internet activities (both solitary and social online
activities).
Study 2: Evaluation of Pleasure & CIU using the Electronic Pleasurable Events Schedule
(E-PES)
Aim 3: To examine whether baseline pleasure scores associated with engagement in
Internet activities predict CIU approximately one year later accounting for baseline CIU,
age, gender, and ethnicity among a sample of regular high school youth participating in
the Happiness and Health study (H&H).
3
Hypothesis 3: Higher reported pleasure derived from electronic use is positively
associated with CIU one year later, controlling for baseline CIU, age, gender, and
ethnicity.
Aim 4: To examine whether anhedonia and gender moderates the relationship between
pleasure scores associated with engagement in Internet activities and CIU.
Hypothesis 4: Pleasure derived from Internet use will be strongly associated with
CIU among individuals who report greater levels of anhedonia. The relationship
between pleasure derived from Internet use and CIU will be more strongly
associated among males.
Study 3: Examining the Association between CIU and Alcohol, Tobacco, Marijuana, and
Other Drug Use Using a Latent Transition Analysis
Aim 5: To conduct a latent class analysis (LTA) among a sample of regular high school
students participating in the Happiness and Health study (H&H) to identify subgroups of
adolescents with CIU and alcohol, tobacco, marijuana, and other drug use behaviors.
Hypothesis 5: Distinct latent statuses of individuals demonstrating homogenous
behavior will be identified in our sample. At baseline, a subgroup of individuals
that exhibit CIU independent of drug behavior and other subgroups that cluster
based on ATOD behaviors.
Aim 6: To determine whether membership within these identified subgroups transition
over time.
Hypothesis 6: Identified latent statuses at Time 1 will shift over time. Individuals
identified in a CIU latent status group at baseline will likely transition into latent
status groups of ATOD behaviors one year later.
4
CHAPTER 1: BACKGROUND AND SIGNIFICANCE
The Internet has become ubiquitous in daily life. It is used for looking up information
(e.g., driving directions and news), performing everyday transactions (e.g., paying bills,
purchasing items), entertainment (e.g., playing online games, watching movies, listening to
music), and social interaction (e.g., chatting) (Fallows, 2004). In addition, the Internet and
Internet related activities have the added appeal of accessibility, affordability, anonymity,
approximation (to real-life), convenience, escapism, and disinhibition (Griffiths, Parke, Wood, &
Parke, 2006; Hertlein & Stevenson, 2015). Although the Internet has positively revolutionized
many aspects of communication and behavior, researchers and clinicians are becoming
increasingly concerned with the negative impacts of Internet use, such as compulsive Internet use
(CIU) (Griffiths, 2000; Young, 1998) – also interchangeably called Internet addiction and
problematic Internet use (PIU) (Cash, Rae, Steel, & Winkler, 2012) in the literature.
Definition of Compulsive Internet Use (CIU)
Compulsive Internet use (CIU) is defined by preoccupation with the Internet leading to
negative consequences (Cash et al., 2012). Internet addiction was first researched by Young
(1998), who noticed that online users were exhibiting symptoms of Internet addiction similar to
addicts of other behaviors and substances, most notably with gambling addicts (Young, 1998).
Indeed, Internet addicts share many of the same characteristics as other addicts: 1) appetitive
effects, 2) satiation, 3) preoccupation, 4) loss of control, and 5) negative consequences (Sussman,
2017). Yet despite growing literature arguing for the existence of CIU and the addition of
gambling addiction to the fifth edition of the Diagnostic and Statistical Manual of Mental
Disorders (DSM-V), CIU remains excluded from the DSM-V list of behavioral addictions (Cash
et al., 2012; Pies, 2009).
5
Prevalence of CIU among Adolescents
Prevalence of CIU among adolescents has been documented worldwide, with Asian
countries reporting higher prevalence of problematic Internet use (Frangos, Frangos, &
Sotiropoulos, 2011; Gámez-Guadix, Calvete, Orue, & Las Hayas, 2015; Ko et al., 2012; Kuss,
Griffiths, & Binder, 2013; Lin et al., 2011; Sussman, Lisha, & Griffiths, 2011). Studies
conducted among Taiwanese high school students have demonstrated CIU prevalence as high as
13.8% (Yang & Tung, 2007) and 20.1% (Ko, Yen, Liu, Huang, & Yen, 2009). Tran et al. (2017)
reported 21.2% of Vietnamese youth in their sample were addicted to the Internet. While one
study in Korea found that prevalence of CIU among youth was only 1.6%, 38% of the sample
were considered “possible” Internet addicts (Kimet al., 2006). Similarly, Jang, Hwang, & Choi
(2008) found that only 4.3% of Korean adolescents were addicted to the Internet, but 30% of
their sample were intermittent Internet addicts. Park, Kim, and Cho (2008) reported that 10% of
Korean high school youth in their sample were considered at high risk for CIU. China, too, has
reported high rates of CIU among their youth, with 10.2% of youth at least moderately addicted
to the Internet (Lam, Peng, Mai, & Jing, 2009). Siomos, Dafouli, Braimiotis, Mouzas, and
Angelopoulos (2008) found that 8.2% of Greek adolescents between the ages of 12 and 18 were
addicted to the Internet, with males who frequented Internet cafes most likely to be Internet
addicts.
A systematic literature review conducted by Moreno et al. (2011a) examining prevalence
of CIU among emerging adults and adolescents in the United States found that prevalence of
CIU ranged between 0%-26.3%. However, the authors noted the limited number of studies that
examined CIU solely among adolescent populations. Among the few studies conducted
specifically among adolescents in the United States, Liu et al. (2011) reported 4% of regular high
6
school youth in their sample were considered addicted to the Internet; Sussman et al. (2014)
reported that last 30 day prevalence of Internet addiction among a sample of continuation high
school students – students who have problems remaining in regular high schools – was 18.4%.
Prevalence of CIU among American youth attending regular high schools seems slightly
lower compared to other countries, such as Taiwan and China (Zhang, Amos, & McDowell,
2008). However, the U.S. has comparably high percentages of use – 92% of American
adolescents between the age of 13 and 17 report going online daily; 24% go online “almost
constantly”; and more than half go online several times a day (Lenhart et al., 2015). CIU in the
U.S. may be underreported or underdiagnosed. Discrepancies between prevalence of CIU in the
U.S. and other countries may be further explained by different locations of Internet accessibility.
That is, in Asian and European countries where Internet prevalence is higher, CIU may be
observed in PC Bangs (in Korea) (Huhh, 2008) and Internet cafes (in Taiwan) (Wu & Cheng,
2006). However, in the United States, most Internet activity occurs in the home, where CIU may
be less visible.
Adolescent Susceptibility to CIU
While emerging adults (ages 18-25) have higher self-reported prevalence of CIU
compared to other age groups (Young & De Abreu, 2010), adolescents (ages 13-17) may be a
population at risk for CIU. Habit forming behavior in childhood/early adolescents may become
addictions during late adolescence, emerging adulthood, and adulthood (Everitt, 2014). For
example, studies examining adolescent drug use have shown that initiations of such behaviors at
younger ages are associated with riskier behaviors and addiction as adolescents age (Chen, Storr,
& Anthony, 2009). Similar to drug use behaviors, higher Internet literacy and Internet use may
lead to higher probabilities of becoming addicted to the Internet (Leung & Lee, 2012), perhaps
7
across the lifespan. Schools, for example, have integrated information technology to
communicate with their students, search for subject related content, and submit homework
assignments (Kozma, 2003). Social networking (SNS) applications such as Facebook, Snapchat,
and Instagram have become an integral part of communicating and maintaining social
relationships among teens, with 71% of American teenagers reporting the use of multiple SNS
sites (Fallows, 2004). As a result, Internet use is becoming normalized during early adolescence.
Those particularly vulnerable, or given additional lifestyle opportunities, may develop CIU.
Another reason adolescents may be vulnerable to CIU is that cognitive processes
associated with self-regulation are not fully developed until late adolescence and/or emerging
adulthood (Pokhrel et al., 2013). Self-regulation is defined as the ability to control one’s
emotional and behavioral impulses and to accurately judge the appropriateness or normalcy of
one’s behavior. Deficient self-regulation has commonly been linked to several behavioral and
media addictions, including CIU (LaRose, Lin, & Eastin, 2003). Specifically, lower self-
regulation has been associated with general CIU (Wegmann, Stodt, & Brand, 2015) and
addiction to an Internet-specific activity, such as social networking (Caplan, 2010; Gámez-
Guadix, Villa-George, & Calvete, 2012; Wegmann et al., 2015) or online gaming (Liau et al.,
2015). Liau et al. (2015) found that self-regulation mediated the relationship between impulsivity
and pathological online video gaming among a sample of Singaporean adolescents. That is,
adolescents with higher impulsivity had lower self-regulation which led to higher pathological
video gaming (Liau et al., 2015). The importance of self-regulation development among
adolescents in preventing CIU is further highlighted by Mun and Lee (2015) who found that
interventions teaching self-regulation skills to Korean elementary school students at risk for CIU
successfully lowered CIU scores at follow-up. Without self-regulation skills instruction,
8
adolescents may be at higher risk for CIU than older individuals whose self-regulation skills are
more developed.
Sociodemographic Correlates
Sociodemographic correlates associated with higher CIU are gender, ethnicity, and socio-
economic status (SES). In the U.S., individuals who have reported problems with Internet use
tend to be male, White, and of higher SES (Young, 2007). However, as the Internet becomes
more available, affordable, and accessible, various minority and SES groups have greater access
to the Internet (Whiteley et al., 2012) – limitations to Internet use are no longer bound by
whether or not someone can afford it. In the near future, discrepancies in Internet use by
ethnicity and SES may not apply; the generalizability of current research results may be limited.
Liu et al. (2011), for example, found that problems with Internet use were higher among Asian
and Hispanic high schoolers than their Caucasian counterparts. Lucszak et al. (2016) conducted a
systematic review of 68 empirical papers published between 2000 and 2013 on Internet addiction
among adults and found that only one study reported significant associations between ethnicity
and Internet addiction – Asians reported higher Internet addiction than non-Asians. Additional
CIU research among diverse racial and SES populations are needed as the digital divide begins
to close.
The effect of gender on CIU is inconsistent. Many studies have found that males are more
likely to exhibit CIU (Lin et al., 2011; Yang & Tung, 2007). However, recent studies have
shown that females are just as likely to report CIU as their male counterparts dependent on the
Internet activity being studied. Males, for example, are more likely to be addicted to online
gaming (Bernardi & Pallanti, 2009; Frangos et al., 2011), pornography (Frangos et al., 2011;
Young, 2007), and downloading (Frangos et al., 2011) than females; conversely, females are
9
more likely to be addicted to social networking (Bernardi & Pallanti, 2009; Frangos et al., 2011)
and online shopping (Young, 2007) than males. Some studies have found no association between
gender and CIU (Gámez-Guadix et al., 2015; Liu et al., 2011), arguing that as Internet becomes
more prolific, Internet activities may become gender neutral. Additional studies examining the
effects of gender on general CIU and addiction to specific online activities is warranted as
adolescent Internet use becomes more common.
Mental Health and Neurobiological Factors
The presence of co-morbid psychological correlates influencing CIU is well-documented
in the literature. Significant psychological correlates include depression (Carli et al., 2012; Diddi
& LaRose, 2006; Evren et al., 2014; Lee, Han, Kim, & Renshaw, 2013), symptoms of attention
deficit and hyperactivity disorder (ADHD) (Carli et al., 2012; Evren et al., 2014), anxiety
(Bernardi & Pallanti, 2009; Carli et al., 2012), and symptoms of obsessive compulsion (Dong,
Lu, Zhou, & Zhao, 2011; Liau et al., 2015; Lin et al., 2011; Meerkerk, van den Eijnden, Franken,
& Garretsen, 2010). Other psychological correlates include hostility/aggression (Carli et al.,
2012), stress (Lin et al., 2011), social anxiety/phobia (Bernardi & Pallanti, 2009), and anhedonia
(Guillot et al., 2016). The directionality of these psychological correlates with IA has not been
well-examined. Dong et al. (2011) reported that symptoms of obsessive compulsion were present
among Chinese adolescents before the development of CIU, but depression, anxiety, hostility,
interpersonal sensitivity, and psychoticism became worse with increased CIU. Future research
examining the directionality and long-term effects of these associations is warranted to inform
treatment programs.
Individuals with these co-morbid problems may seek out the Internet as a coping
mechanism and as a means to elevate moods (Gordon, Juang, & Syed, 2007). Adolescents with
10
higher prevalence of anhedonia, for example, have problems achieving same levels of pleasure
from performing the same activities as their less anhedonic counterparts (Ferguson & Katkin,
1996; Huys, Pizzagalli, Bogdan, & Dayan, 2013; Szczepanik et al., 2017). As a result, anhedonic
individuals may turn to the Internet for immediate enhanced stimulation of reward pathways that
is similar to the stimulation effects of drug and alcohol use (Murali & George, 2007). Yet, the
hedonic value that is associated with initial Internet use and its effect on subsequent problematic
Internet behaviors and Internet addiction are unclear, particularly among adolescents who may
exhibit comorbid psychological risk factors. Among a sample of online gambling addicts,
gambling for mood regulation and enjoyment were both predictors of online gambling addiction.
Individuals with higher depression scores, however, found lower levels of enjoyment gambling
online, but both motivations were independently associated with higher probability of addiction
(Lloyd et al., 2010). The relationship between psychological correlates and hedonic values
associated with Internet activities, and their independent and comprehensive impact on Internet
addiction is an important aspect of understanding CIU etiology and treatment.
Consequences associated with CIU
Negative consequences associated with CIU are highlighted in the literature. Several
studies have demonstrated that CIU is associated with physical consequences: sedentary
behavior (Matusitz & McCormick, 2012), obesity (Matusitz & McCormick, 2012; Park & Lee,
2017), being underweight (Park & Lee, 2017), and poor sleeping patterns (Flisher, 2010).
Psychosocial consequences associated with CIU include poor academic performance (Yang &
Tung, 2007), becoming victims and perpetrators of cyberbullying (Gámez-Guadix, Orue, Smith,
& Calvete, 2013), increased hostility/aggression (Ko et al., 2009), delinquent behavior (Chiao et
al., 2014), suicidal ideation (Kim et al., 2006), and lower quality of life (Cao, Sun, Wan, Hao, &
11
Tao, 2011; Tran et al., 2017). Among global samples of adolescents reporting addiction, the co-
occurrence of CIU with other substance and behavioral addictions is high (Sussman, Arpawong,
et al., 2014; Tsai et al., 2016). CIU has been linked with other risky behaviors among adolescents
such as smoking (Chiao et al., 2014; Frangos et al., 2011), drinking alcohol (Sampasa-Kanyinga
& Chaput, 2016; Sun et al., 2012) and other drug use (Liu et al., 2011). However, the
interpretation of these results is often limited by the studies’ cross-sectional designs and
limitations in examining several problem behaviors at one time.
Theories used in CIU
Commonly referenced CIU theory has been guided by the addiction framework (Griffiths
et al., 2006; Young, 1998), which has been criticized for its inability to capture what people are
actually doing online (Caplan, 2002; Shaffer, Hall, & Vander Bilt, 2000). To address the
previously mentioned limitations, Davis (2001) proposed a cognitive-behavioral model of
pathological Internet use that distinguishes between generalized problematic Internet use and
specific (activity-oriented) problematic Internet use. In addition, Davis asserted that problematic
Internet use were manifestations of some maladaptive behavior associated with the
predisposition of psychosocial disorders (e.g., depression and anxiety). While studies have
shown that CIU is associated with several psychosocial variables (Caplan, 2002), comorbid
psychopathology of other disorders alone does not fully explain CIU (Van Rooij & Prause, 2014)
and the cognitive-behavioral model of pathological Internet does not account for other
consequences of Internet use beyond that of CIU itself.
More recently, Brand et al. (2016) suggested the Interaction of Person-Affect-Cognition-
Execution (I-PACE) model that integrates predisposing variables, affective and cognitive
responses to internal or external stimuli, executive and inhibitor control, decision-making
12
behavior resulting in the use of certain Internet applications and sites, and consequences of using
the Internet applications and sites of choice into one integrated model (Brand, Young, Laier,
Wölfling, & Potenza, 2016). While the I-PACE model is comprehensive and theoretically
oriented, it is empirically difficult to analyze and hard to control for all the variables proposed in
the model.
Limitations to Current CIU Research
Despite a growing concern for CIU among adolescents, several gaps in the literature
remain. First, the majority of CIU research has been conducted in populations outside of the
U.S., particularly in Asia. Trends in non-American countries show that CIU is indeed a growing
concern that affects their youth, but the relative impact of Internet use on American youth is not
as clear. While studies have documented notable prevalence of CIU in the U.S. (Liu et al., 2011;
Moreno et al., 2011a), parameters of Internet use in the U.S. require additional analysis. In
addition, the effect of racial differences associated with CIU warrants further investigation; most
studies examining CIU have occurred among homogenous populations. Those that have included
heterogeneous populations have found racial differences in CIU susceptibility (Liu et al., 2011;
Luczak et al., 2016), highlighting the importance of furthering the science. Second, questions
remain regarding generalized Internet addiction and Internet addiction associated with specific
activities. Greater understanding regarding the specific or general activities associated with CIU
among American youth is needed. A comparison of the Internet activities that are associated with
CIU among American youth compared to activities associated with CIU among youth abroad can
also help inform prevention efforts. Thirdly, the intrinsic motives associated with Internet use
that may spur addictive behavior, such as the amount of pleasure received from participating in
certain internet activities, is not fully understood. Additionally, the relationship between intrinsic
13
motivations and psychological correlates of CIU has not been examined. Although cross-
sectional studies correlating CIU with substance use populate current research, the long-term
behavioral health consequences associated with CIU that may impact negatively quality of life
are not as prolific.
A. INTRODUCTION TO DISSERTATION STUDIES
The need to examine CIU as a probable consequence of Internet use is increasingly critical,
particularly among adolescents. Globally, there has been an impetus for establishing CIU
treatment programs, particularly among Asian countries like China, Taiwan, and South Korea,
who view CIU as a serious public health threat (Young & De Abreu, 2010). Already, China has
taken measures to identify and curb CIU among their adolescents, such as creating special units
within hospitals to treat CIU (Moore, 2008). Although the United States lags behind other
countries in CIU research treatment, an increasing number of CIU treatment facilities have
opened nationwide to address this growing problem (Foran, 2015). A better understanding of
CIU from etiology to its relationship with other consequential behaviors can inform more
effective CIU screening and treatment programs.
The three studies were conducted using a cohesive theoretical framework: the Uses and
Gratifications Theory (see Figure 1). Study 1 identified patterns of Internet activity behavior
among a sample of high risk youth attending continuation high schools. The associations
between patterns of Internet use and gender, and patterns of Internet use and specific CIU
problems were also examined. Study 2 examined the cross-sectional and longitudinal
relationships between intrinsic motivation (i.e., pleasure) of electronic media use and CIU, and
whether this relationship was moderated by anhedonia and gender, controlling for age, ethnicity,
and baseline CIU (only in longitudinal model). Prior work has explored psychosocial correlates
14
and intrinsic motivation independently. However, because mental health distress and addictions
are highly comorbid, pleasure may be an underlying motivation for CIU, particularly among
individuals with mental health-related susceptibility to CIU. Study 3 examined the longitudinal
association of CIU with alcohol, nicotine, marijuana, and other drug use (i.e., stimulants,
prescription stimulant pills without a doctor’s advice, and prescription painkillers without a
doctor’s advice). Collectively, study 1 and study 2 investigated the motivations behind Internet
use that may lead to CIU as a consequence. Study 3 extended the research regarding negative
consequences of Internet use beyond CIU to include other behaviors that may be detrimental to
adolescent health.
Uses and Gratification Theory (UGT)
Compared to other theories examining CIU, the uses and gratifications theory is
advantageous because it allows for the examination of media use and progression to addiction
among normal, non-clinically diagnosed Internet users (LaRose, 2011). Uses and gratifications
15
theory (UGT) was first proposed in the 1940s (Blumler & Katz, 1974) and has been widely used
in communications research to study motivations behind individual media choice (Ruggiero,
2000). UGT posits that individuals choose to engage in media that fulfill their social and
psychological needs and that individual media use can lead to both positive and negative
consequences associated with the behavior (Blumler & Katz, 1974; Ruggiero, 2000). UGT
focuses on a user-driven approach and seeks to explain the “how and why” of media behavior
(Stafford, Stafford, & Schkade, 2004).
Previous research has used UGT to examine the use and addiction of newspaper, magazine,
radio, and television media (LaRose, Mastro, & Eastin, 2001; Sussman & Moran, 2013). More
recently, UGT has been used to examine Internet use: web browsing (Stafford et al., 2004),
online shopping (Stafford et al., 2004), e-mail (Ruggiero, 2000; Stafford et al., 2004), music
(Krause, North, & Heritage, 2014), and social networking sites/applications such as Facebook
(Alhabash, Chiang, & Huang, 2014; Ha, Kim, Libaque-Saenz, Chang, & Park, 2015; Raacke &
Bonds-Raacke, 2008), MySpace (Raacke & Bonds-Raacke, 2008), KaKao Talk (Ha et al., 2015),
and Twitter (Johnson & Yang).
Studies using UGT have found that Internet choice motivations are often driven by the
perceived gratifications that result from Internet use: content gratifications, process
gratifications, and social gratifications (Stafford et al., 2004). Content gratifications derived from
the Internet result from the messages related to specific websites (or Internet applications) and
are defined by information seeking or learning. Content gratification is meant to connect users to
the world outside of the media being used (Song, Larose, Eastin, & Lin, 2004). Alternatively,
process gratifications result from the actual use of the Internet and are defined by the pleasure
received from Internet use rather than the content presented (Song et al 2004). Process
16
gratifications tend to pull users away from the real world as more pleasure is derived from the
consumption of Internet use (Song et al., 2004). The Internet’s ability to foster an environment
where interactive communication can take place virtually (through online discussions, e-mail,
chatrooms, and social networking sites) has the potential has led to the proposal of a third
gratification called social gratification that is unique to the Internet. Social gratification is
achieved and sought through interacting online with new friends or maintaining existing
relationships (Stafford et al., 2004).
Under each of the three gratifications, researchers have further pinpointed more descriptive
gratifications. Lin (2001) reported entertainment (process gratification), escape (process
gratification), informational learning (content gratification), and interaction (social gratification)
as motivations that explained the majority of the variance for Internet adoption and use. Parker
and Plank (2000) found Internet use was associated with the need for relaxation and escape, both
process gratifications. Charney and Greenberg (2002) found eight gratification factors
influencing time adolescents spent on the internet: keeping informed (content gratification),
diversion and entertainment (process gratification), peer identity (content gratification), good
feelings (process gratification), communication (process gratification), sights and sounds
(content gratification), career (content gratification), and coolness (process gratification).
UGT and CIU
CIU has mostly been attributed to process gratifications, although some researchers assert
that all three gratifications may contribute to CIU. Rubin (2009) refers to process gratifications
as “ritualized” use. Because process gratifications are related to pleasure derived from the
consumption of internet use (i.e., substitution of real-life activities for virtual ones like social
interaction and entertainment), this could easily lead to excessive Internet use. Song et al. (2004)
17
reported both content and process gratifications were related to higher prevalence of CIU; they
were unable to define a clear distinction between process and content gratifications leading to
Internet addiction. The authors suggest that as the Internet becomes more complex and
entrenched in daily lives, the distinction between content and process gratifications may become
less relevant. Rather, the need to examine the motivations behind media use that may lead to
addiction is more important than the actual activities being performed. The influence of social
gratifications leading to CIU is less studied, but those that exist supported claims that virtual
social interactions are associated with increased Internet use and addiction (Li & Chung, 2006;
Tang, Chen, Yang, Chung, & Lee, 2016; Van Rooij, Schoenmakers, Van de Eijnden, & Van de
Mheen, 2010).
UGT in the 21st century has become more refined. Noting the limitations of the original
UGT to explain Internet addiction, LaRose and colleagues (2011) proposed the addition of
Internet self-efficacy and self-regulation, drawn from Bandura’s Social Cognitive Theory. The
addition of self-efficacy and self-regulation improved the predictive power of UGT to explain
the variance of Internet consumption behavior from 10% to 30-40% (LaRose, Lin, & Eastin,
2003). Thus, LaRose’s modified UGT suggests that gratifications sought and outcome
expectancies related to Internet use may lead to repetitive, non-conscious Internet behavior,
resulting in addiction (LaRose, 2011).
Critiques and Limitations of UGT
UGT has been widely criticized for being too simplistic. The primary critique is that
UGT does not account for macro media influences beyond the individual’s control that drive
media use (Ball-Rokeach, 1998). Critics argue that much of UGT research focuses on the
psychological and sociodemographic variables that influence media choice, which cannot detect
18
interpersonal influences in media choice. Although this was true for more traditional forms of
media (i.e., radio, television, and newspapers), proponents of UGT believe that UGT fits Internet
research; the Internet is a more user driven medium that requires individuals to actively engage
with the medium to seek information and activities. Furthermore, information presented on the
Internet is so vast that users must preemptively decide what they want from their Internet use
before navigating the world-wide web. Another critique of previous UGT research is that
researchers tend to label gratification statements without much consideration for how many
people actually reported that gratification (Ruggiero, 2000). The gratification statements are also
interpreted by researchers which may bias the interpretability of results. While UGT researchers
are making more of an effort to detect whether labeled gratifications sought correspond with
labeled gratifications obtained, limitations remain regarding the actual amount of satisfaction
received from use (LaRose et al., 2001; Palmgreen & Rayburn, 1985).
Despite improvements to UGT in Internet research, several limitations remain,
specifically pertaining to the motivations behind Internet use and the consequences associated
with Internet use warrant further investigation. Thus, while the primary goal of this dissertation
is to contribute to the overall understanding and effects of Internet addiction, a secondary goal is
to address the gaps in CIU research that uses UGT.
19
CHAPTER 2: Identifying subgroups of Internet users and its association with CIU and
gender among a sample of high risk adolescents
BACKGROUND
Compulsive Internet use (CIU) is prevalent among adolescent populations worldwide
(Frangos et al., 2011; Gámez-Guadix et al., 2015; Ko et al., 2012; Kuss et al., 2013; Lin et al.,
2011). Specific Internet activities may predispose individuals to become more addicted to the
Internet. Among adolescents and emerging adults, online chatting, online gaming, social
networking and downloading are common activities linked with problematic Internet use more so
than other online behaviors, such as online gambling (Király et al., 2014; Van Rooij et al., 2010).
Extensive use of social networking sites (SNS), web-based media platforms that allow users to
interact with friends and other users with similar interests via chat and public posting or sharing
of information, contributes to and perpetuates CIU (Kuss et al., 2013; Van Rooij et al., 2010).
For example, the number of times users logged on to Facebook was associated with more self-
reported problematic Internet use among college students (Kittinger, Correia, & Irons, 2012).
Individuals addicted to Twitter reported problems such as insomnia and trouble communicating
with people in real life (Saaid, Al-Rashid, & Abdullah, 2014). Similarly, online gaming and
more specifically, massively multiplayer online role-playing games (MMORPG), have been
associated with CIU (Van Rooij et al., 2010). MMORPG significantly explained a small
percentage of the overall variance for problematic Internet use among a sample of online game
players (Caplan, Williams, & Yee, 2009). Furthermore, addiction to these activities may differ
by gender – while some researchers have found social networking and online chatting addiction
to be higher among females, males are at greater risk for gaming addiction (Frangos et al., 2011;
Young, 2007).
20
Terms such as SNS addiction (Kuss & Griffiths, 2011), Facebook addiction (Andreassen,
Torsheim, Brunborg, & Pallesen, 2012), Twitter addiction (Saaid et al., 2014), and Internet
gaming addiction (Kuss & Griffiths, 2012b) have been identified as disorders specific to
Internet-related applications. Some researchers suggest that the terms “compulsive Internet use”,
“problematic Internet use” and “Internet addiction” may become obsolete as addictions on the
Internet become more nuanced (Starcevic & Billieux, 2017). However, these Internet related
disorders can arguably be considered extensions of CIU or subtypes of CIU since some of these
activities can only be conducted via the Internet (Griffiths, 2012).
Under the premise of the cognitive-behavioral model of pathological Internet use,
researchers acknowledge the need to distinguish between addiction to a specific activity
conducted on the Internet and general addiction to the Internet (Caplan, 2002; Davis, 2001).
Indeed, Davis (2001) theorized two types of Internet addiction that result from cognitive
maladaptive behavior– specific pathological Internet use (specific PIU) and generalized
pathological Internet use (generalized PIU). Specific PIU is defined as content-specific Internet
dependence (Davis 2001). Individuals with specific PIU are more likely to have an addiction
offline which is translated to its online counterpart or will concurrently develop an addiction to
that behavior offline compared to those without specific PIU. For example, Internet dependency
may be driven by online gaming or online pornography; in the absence of the Internet,
individuals with specific PIU are just as likely to be addicted to gaming and pornography offline.
Generalized PIU is defined as general overuse of the Internet where no clear objective for
Internet use has been established (Davis 2001). For example, an individual who exhibits
generalized PIU may find themselves spending aimless hours on chat or e-mail and lack a
specific Internet platform.
21
Aligned with the cognitive behavioral model, the uses and gratifications theory (UGT)
seeks to determine the specific activities that drive Internet consumption (Ruggiero, 2000).
Indeed, the UGT theory posits that users engage in specific Internet uses (or Internet activities)
that fulfill their needs (e.g., entertainment, connection with others, surveillance, and escapism)
(Rubin, 2009). These activities provide content gratifications (information provided by the
Internet), process gratifications (enjoyment received from using the Internet), and social
gratifications (social interaction provided by the Internet) which contribute to Internet
consumption (Stafford et al., 2004), where process and social gratifications are more indicative
of CIU (Song et al., 2004). As previously mentioned, the added advantage of using UGT
compared to the cognitive behavioral model of pathological Internet use is that it explains media
use among normal, nonclinical populations (LaRose, 2011).
Prior studies have examined the association between Internet specific activities and CIU
often using analytic approaches that cannot account for patterns of Internet use and its
association with CIU. Most studies examine Internet activities by assessing frequencies (e.g.,
Young, 2007), conducting MANOVAs (e.g., Tonioni et al., 2012), and including specific
Internet activities in logistic regression (e.g., Frangos et al., 2011). To our knowledge, no studies
have examined Internet activities and addiction using a latent class analysis. Indeed, researchers
have noted the importance of identifying subgroups of Internet users for prevention and
intervention (Frangos et al., 2011; Young, 2007). The aim of this first study was to improve our
overall understanding of the Internet activities that are associated with CIU using a latent class
analysis (LCA). Latent class analysis (LCA) has the ability to categorize individuals into
subgroups based on their response patterns. The primary advantage of using LCA is the ability to
assess multiple activities at once and classify individuals based on reported patterns of behavior.
22
Based on the latent classes identified, this study assessed if groups of individuals who have
similar patterns of utilizing specific Internet activities exist and if these patterns of Internet
utilization were characterized by some underlying gratification. This study also analyzed the
odds of latent class membership with CIU indicators and gender as covariates.
METHODS
Research Questions
The proposed research study addressed the following questions:
Q1: Do distinct subgroups of Internet users exist among this sample of continuation high
schoolers?
Hypothesis 1: Distinct subgroups of adolescents that share patterns of Internet
use behavior and CIU will be identified.
Q2: Will the conditional probabilities of belonging to the identified LCA groups vary by
CIU and gender?
Hypothesis 2: Conditional probabilities of belonging to the identified subgroups
will vary by CIU and gender. Aligned with the general consensus among the
literature, CIU will be associated with classes endorsing more Internet activities
and social interaction. Similar to studies in the literature showing that females and
males differ on their Internet use, we predict that females are more likely to be
classified in subgroups where Internet activities promote social interaction
specifically and males are more likely to be classified in subgroups that generally
endorse more types of Internet activities (both solitary and social online
activities).
23
Data Collection
Data for study 1 was obtained from Project Towards No Drug Abuse (TND-6), a
nationally recognized evidence-based drug abuse prevention program (Sussman, Sun, Rohrbach,
& Spruijt-Metz, 2012). TND curriculum comprised of 12 lessons that teach motivation, skills,
and decision-making aimed at preventing and reducing drug use and included an added
motivational interviewing booster component. Participants were enrolled in this study if they
attended one of the twenty-four southern California continuation high schools (CHS)
participating in TND. CHS are for youth who are unable to attend regular high school due to
difficulties achieving sufficient credits in the regular school system (Sussman, Arriaza, &
Grigsby, 2014). Such students are often displaced from traditional high schools due to deviancy
problems such as substance use and pregnancy, or may have left traditional high school settings
in order to fulfill familial responsibilities (e.g., work). Due to familial responsibilities and/or
instability in the home, CHS students represent an at-risk population for various problematic
behaviors (Lisha et al., 2014; Sussman, 2010). All study procedures were reviewed and approved
by the University of Southern California Institutional Review Board.
Baseline measures were collected before program implementation (February 2008 to
April 2009). Program implementation occurred between March 2008 and May 2009. Following
program implementation, participants were contacted for follow-up interviews at 1, 2, 3, and 4-
years post program implementation, with approximately 15 months between each follow-up
interview (October 2009 to April 2014).
Trained project staff administered paper and pencil surveys in one 50-minute classroom
session at baseline and 1-year follow-up. Participants who were absent the day of data collection
were contacted by phone and given the option to complete the interview verbally. At 2-year, 3-
24
year, and 4-year follow-up, participants completed surveys by phone, by mail-back, or in-person
(at school or via home visit). To be eligible to participate in all waves of data collection, all
participants under the age of 18 were required to provide informed assent and parental consent.
Participants over the age of 18 were required to provide informed consent. A total of 1,704
participants consented to be part of the study, but only 1,676 (98%) participants completed the
baseline survey.
Measures.
Demographic Information. Demographic information was collected on age (in years) and
gender (male vs female).
Internet Activity Measure. A multi-response Internet item measure was used to assess the
prevalence and co-occurrence of Internet activities conducted in the last 7 days. The measure
header was: “In the last 7 days, did you do any of the following computer and Internet activities?
(Mark all that apply).” The activities provided were: “read news online”, “e-mailed”, “social
networking (such as MySpace or Facebook)”, “read/participated in online forum discussions”,
“did homework online”, “played online games”, “played offline games”, “watched movies or
video clips online”, “chatted online (IM, audio or video) with people I know in real life”,
“chatted online with people I met online”, “chatted online with strangers”, “shopped online”,
“visited porn site(s)”, “downloaded music”, and “I did not go online in the last 7 days”. In
addition, participants were provided an item, “other (specify),” where they can indicate an
activity that was not listed on the list. Response choices were coded as 1 or “Yes” if participants
reported they conducted the activity in the last 7 days and 0 or “No” if participants did not report
that behavior.
25
Because the purpose of this study was to examine the Internet activities that preclude
compulsive Internet use, the items “played games offline” and “I did not go online in the last 7
days” were excluded from analysis. Seventy-seven students indicated an activity under the item
“other (specify).” Responses were recoded into one of the sixteen Internet activities if applicable;
otherwise, responses were excluded from analysis if the activity was too vague to determine if it
was conducted online or offline (e.g., used Microsoft Word, listened to music). Fourteen of the
indicated activities were recoded.
Compulsive Internet Use Scale (CIUS). CIU was measured using an abbreviated (36-
item) version of the Online Cognition Scale (Davis, Flett, & Besser, 2002) that was used among
a sample of N=211 undergraduate students (mean age=21.73 years , SD=4.40 years). In previous
studies, the Online Cognition Scale demonstrated good internal consistency (alpha = 0.84)
(Davis et al., 2002). For the present study, a four item-index response was used to measure
diminished impulse control associated with CIU (e.g., Guillot et al., 2016). The four items were:
“I use the Internet more than I ought to”; “I usually stay on the Internet longer than I had
planned”; “Even though there are times when I would like to, I can’t cut down on my use of the
Internet”; and “My use of the Internet sometimes seems beyond my control”. Each item response
was scored on a Likert scale: (1) Never, (2) Rarely, (3) Sometimes, (4) Most of the time, and (5)
Always. The response items were a continuous covariate in the model. This scale showed good
internal consistency among this sample (Cronbach’s alpha = 0.81). All CIUS items were
standardized for analysis.
Analytic Plan.
While TND data was longitudinal, this analysis only included baseline data. This was
because this was the only wave of data that asked questions related to specific Internet use
26
activities and CIU. Students who did not answer completely on key variables of interest in this
study (i.e., did not select any Internet activity; did not answer any of the CIU items; and did not
indicate gender) or who indicated they did not use the Internet in the past 7 days were excluded
from data analysis. The final sample size was 1,367 participants.
To address our first research aim, a latent class analysis (LCA) was conducted. All LCA
models were analyzed using the MPlus Version 6.0 software program (Muthén & Muthén,
2010). The number of assumed latent classes began at 1 and increased subsequently until model
fit indices and class interpretability indicated the best latent class solution (Muthén & Muthén,
2000). Fit indices used to analyze the model and class fit were: Akaike information criterion
(AIC), Bayesian information criterion (BIC), entropy (0-1), and Lo-Mendell Rubin (LMR) p-
values. Lower AIC and BIC values indicated better model fit. Higher entropy values (closer to 1)
indicated more distinct class solutions. LMR p-values tested whether the n-class solution was a
better fit than the n-1 class solution, where a significant p-value (<0.05) indicated the n-1 class
solution was better than the n class solution.
Following the identification of the most appropriate latent class solution, covariates were
added to the model to address my second research aim. The four CIUS items were each added to
the LCA model as a covariate to examine the association between the identified Internet activity
latent classes and CIU. Gender was also added as another covariate to the model. In addition, a
model with the standardized composite score of CIUS as the only covariate was also analyzed.
Results were interpreted as a multinomial logistic regression model to determine if covariates
predict latent class membership. Odds ratios were reported (Collins & Lanza, 2013). See Figure
2.
27
In addition to the LCA, a linear regression model was conducted to determine the
individual Internet activities that are associated with CIU. All 13 activities were included in the
model as independent variables and CIU was the dependent variable. The model also controlled
for gender. Analysis for the regression model was run in SAS 9.4 (SAS Institute, Cary, NC,
USA).
RESULTS
Sample Characteristics.
All descriptive statistics were run in SAS Version 9.4 (SAS Institute, Cary, NC, USA).
As aforementioned, only participants who answered all covariate questions (i.e., gender and
CIUS items) and who did not answer “I did not go online in the last 7 days” were included in the
28
model (N=1,367). The average age of the participants was 16.80 years old (SD = 0.91), 57%
were male, and 64% self-identified as “Non-White Hispanic”. Last 7 day prevalence of the 13
Internet items among this sample CHS youth from highest prevalence to lowest prevalence was:
social networking (72%), downloaded music (53%), watched movies or video clips online
(52%), e-mailed (43%), chatted online with people known in real life (38%), read news online
(34%), played games online (33%), did homework online (21%), chatted online with people met
online (17%), shopped online (15%), visited porn site (13%), read/participated in online forum
discussion (10%), and chatted online with strangers (5%). Mean scores for each compulsive
internet indicator on the CIUS were: 2.9 (SD=1.2) for staying on the Internet longer than
planned, 2.6 (SD=1.2) for using the Internet more than one ought to, 2.3 (SD=1.2) for not being
able to cut down on Internet use even when one wants to, and 2.0 (SD=1.1) for using the Internet
seems beyond one’s control. Mean scores for the composite CIUS was 9.81 (SD= 3.72).
Assessment of Missing and Excluded Data
Chi-square analysis and t-tests were used to assess if the excluded sample (N=160)
differed from the sample retained for analyses (N=1,367). Significant differences between those
with missing data and the final sample was found for age, and each of the CIUS consequences
independently and composite CIUS scores. Those who were excluded from analysis were more
likely to be older individuals (mean age 16.81 vs 16.64; p=0.004). They were also less likely to
indicate problems with their Internet use: mean scores of 2.86 vs 2.07 for staying on the Internet
longer than planned (p<0.0001), 2.56 vs 1.65 for using the Internet more than one ought to
(p<0.0001), 2.32 vs 2.01 (p=0.001) for not being able to cut down on Internet use even when one
wants to, 2.06 vs 1.82 for using the Internet seems beyond one’s control (p=0.006), and 9.81 vs
7.48 for composite CIUS scores (p<0.0001). There were no significant differences by gender.
29
LCA Results
Results of the LCA suggested a four-class model. We failed to find a significant
difference between the four class and five class solution (p = 0.16). Although the AIC continued
to decrease, BIC and interpretability of the classes indicated the four-class solution was optimal.
Indeed, Nylund, Tihomir, and Benght (2007) ran a series of Monte Carlo simulations and found
that BIC consistently identified the number of latent classes correctly and was more reliable than
other information criteria (see Table 1).
Item response probabilities of each latent class are presented in (Table 2 and Figure 3).
Given high endorsement of all Internet activities, Class 1 was labeled as “Dependent” (8%, N=
109). Conversely, Class 2, which was the most prevalent class, was labeled as “Functioning”
(48%, N= 656) due to low endorsement of Internet activities. Class 3 and Class 4 had high
endorsements of specific Internet activities and were labeled “Social” (6%, N= 82) and
“Entertainment and Surveillance” (38%, N= 519), respectively. Members of the “Social” class
reported high probabilities of chatting online with strangers (90%) and “chatting online with
people met online” (63%), while members of the “Entertainment and surveillance” class reported
high prevalence of playing games online (50%), watching movies or video clips online (79%),
downloading music (75%), and reading news online (50%).
30
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Item Response Probabilities
Figure 3. Latent Class Probabilities for Endorsement of Each Internet Behavior Among a Sample of
CHS Youth (N=1,367)
Dependent
Functioning
Social
Entertainment & Surveillance
31
Table 1. Indicators of Fit for Models Last 7 Day Internet Activities Among Continuation High School Youth (N=1,367)
Class AIC BIC
Sample Adjusted
BIC
Adjusted LMR p-
value
Entropy
1 18957.95 19025.82 18984.52 -- --
2 17692.31 17833.26 17747.50 0.00 0.72
3 17506.76 17720.79 17590.55 0.00 0.72
4 17427.00 17714.12 17539.41 0.04 0.75
5 17389.38 17749.58 17530.40 0.16 0.68
Note: Boldface indicates values of selected model. Abbreviations: AIC, Akaike information criterion; BIC, Bayesian information criterion;
LMR, Lo Mendell Rubin.
32
Table 2. Results of LCA Retaining Four Classes of Last 7 Day Internet Activities Among
Continuation High School Youth (N=1,367)
Class 1 Class 2 Class 3 Class 4
Dependent
(N= 109, 8%)
Functioning
(N= 656, 48%)
Social
(N = 82, 6%)
Entertainment
&
Surveillance
(N= 520, 38%)
Read News
Online
0.68 0.18 0.15 0.50
E-mailed 0.88 0.19 0.51 0.61
Social
Networking (e.g.,
Facebook)
0.96 0.58 0.79 0.84
Read/Participate
d in Online
Discussion
Forums
0.51 0.01 0.05 0.13
Did Homework
Online
0.41 0.13 0.12 0.28
Played Online
games
0.78 0.15 0.13 0.5
Watched Movies
or Video Clips
Online
0.93 0.25 0.47 0.79
Chatted Online
with People I
Know
0.96 0.14 0.63 0.51
Chatted Online
with People I
Met Online
0.91 0.02 0.90 0.08
Chatted Online
with Strangers
0.42 0.00 0.29 0.00
Shopped Online 0.31 0.05 0.15 0.25
Visited Porn Site 0.38 0.05 0.24 0.17
Downloaded
Music
0.89 0.30 0.41 0.75
33
LCA with Covariates
As proposed, each of the four CIUS items and gender were added as covariates to the
four-class LCA model simultaneously. However, there was not enough variation in one of our
classes to detect differences by gender. Thus, gender was subsequently excluded from our
analysis and only results for the CIUS items as covariates in the four-class solution are presented
here. In addition, another model which only included the composite CIUS score was analyzed.
Results of our analysis designated the “Functioning” class and “Social” class,
respectively, as reference groups (see Table 3). Relative to the “Functioning” class, students who
reported staying on the Internet longer than planned were significantly more likely to be
classified in all other classes: “Dependent” (OR= 2.03, p=0.003), “Entertainment and
Surveillance” (OR=1.47, p=0.02), and “Social” (OR=1.88, p=0.02). Similarly, those who
reported using the Internet more than they ought to were significantly more likely to be classified
in the “Functioning” (OR = 2.40, p<0.0001) and “Entertainment and Surveillance” classes
(OR=1.61, p<0.0001) when compared to the “Functioning” class. Those who reported Internet
use seemed beyond their control, however, were significantly more likely to be classified in the
“Functioning” class than the “Entertainment and Surveillance” class (OR = 0.73, p=0.02). No
differences in class membership were found for “not being able to cut down on Internet use”.
Similarly, when the composite CIUS score was included as a covariate in the model, having
higher CIUS scores was associated with higher odds of belonging in the “Entertainment and
Surveillance” class (OR= 1.50, p<0.001), “Social” class (OR = 2.28, p<0.001), and “Dependent”
class (OR = 2.65, p<0.001) compared to the “Functioning” class.
When the reference group was designated as the “Social” class, only those who were
more likely to report using the Internet “more than they ought to” were more likely to be
34
classified in the “Dependent” class than the “Social” class. No significant differences in class
membership were detected between the “Social” class and “Entertainment and surveillance”
class for any of the four CIUS items individually. When the composite CIUS score was included
as a covariate, we see that those with higher CIUS scores were less likely to be in the
“Functioning” class (OR=0.44, p<0.0001) and “Entertainment and Surveillance” class (OR
=0.66, p=0.04) than the “Social” class. There was no significant difference in class membership
between the “Social” class and “Dependent” class for composite CIUS scores.
35
Table 3. Odds Ratios for Predictors of Latent Class Membership among a sample of
Continuation High School Students Participating in Project Towards No Drug Abuse
(N=1,367)
Social Entertainment &
Surveillance
Dependent Functioning
1 2 3 4
N=105 (8%) N=530 (39%) N=133 (10%) N=598 (44%)
Intercept -1.79** -0.04 -1.75 REF
CIUS 1: Uses Internet more than ought to.
OR 1.43 1.61*** 2.40*** REF
CIUS 2: Stays on Internet longer than planned.
OR 1.88* 1.47* 2.03** REF
CIUS 3: Can’t cut down on Internet use.
OR 1.46 0.96 1.14 REF
CIUS 4: Internet use seems beyond their control.
OR 0.75 0.73* 0.66 REF
Using Social Class as reference
Intercept REF 1.75** 0.04 1.79**
CIUS 1: Uses Internet more than ought to.
OR REF 1.13 1.68* 0.70
CIUS 2: Stays on Internet longer than planned.
OR REF 0.79 1.08 0.53*
CIUS 3: Can’t cut down on Internet use.
OR REF 1.52 0.77 0.68
CIUS 4: Internet use seems beyond their control.
OR REF 0.98 0.88 1.34
Using Sum Score of CIUS as Covariate in Model
N = 87 (6%) N = 506 (37%) N = 113 (8%) N = 661 (48%)
Intercept -1.90*** -0.22 -1.90*** REF
CIUS Total
OR 2.28*** 1.50*** 2.65*** REF
Using Social Class as Reference and Sum Score of CIUS as a Covariate in Model
Intercept REF 1.68*** 0.01 1.90***
CIUS Total
OR REF 0.66* 1.16 0.44***
Note: Boldface indicates statistical significance (p< *0.05; p< **<0.01; p < ***<0.001)
36
Unadjusted & Adjusted Regression Models
Unadjusted linear regression models with Bonferroni adjustment (p=0.003) found all
activities except “reading news online” and “doing homework online” to be associated with
using the Internet more than one ought to, staying on the Internet longer than planned, and
composite CIUS scores (See Table 4). Social activities (i.e., e-mailing and chatting online with
people one knows in real life, people met online, and online strangers) were associated with not
being able to cut down on Internet use; only chatting online with people one knows in real life
and online strangers was associated with Internet use seems beyond one’s control. Gender was
not a significantly associated with any of the CIUS items or composite CIUS scores in the
unadjusted models.
Results of the adjusted linear regression model show that after controlling for gender,
using social networking applications (β=0.24, p<0.001; β=0.20, p<0.001; β=0.15, p=0.01),
playing online games (β=0.12, p= 0.04; β =0.15, p=0.01; β=0.14, p=0.02), chatting online with
people one met online (β=0.20, p=0.01; β=0.18, p=0.02; β=0.16, p=0.048), and chatting online
with strangers (β=0.30, p=0.02; β=0.38, p=0.02; β=0.44, p<0.001) were significantly associated
with using the Internet more than one ought to, staying on the Internet longer than planned, and
composite CIUS scores. Watching movies or video clips online was only associated with using
the Internet more than one ought to (β=0.13, p=0.02). Visiting porn sites (β=0.27, p<0.001;
β=0.20, p=0.01) was significantly associated with using the Internet more than one ought to and
staying on the Internet longer than planned. Chatting online with people one knows in real life
was the only activity that yielded positive, significant associations with all four of the CIUS
items and composite CIUS (see Table 4). Doing homework online was significantly, negatively
associated with three of the four CIUS items: staying on the Internet longer than planned (β=-
37
0.14, p=0.03), can’t cut down on Internet use (β=-0.20, p=0.004), and Internet use seems beyond
one’s control (β=-0.15, p=0.03); and composite CIUS (β=-0.19, p=0.004). Reading news online
was also significantly, negatively associated with can’t cut down on Internet use (β=-0.15, p=
0.02) and Internet use seems beyond one’s control (β=-0.15, p=0.01). Being female was
associated with a significant, positive association with using the Internet more than one ought to
(β=0.23, p<0.001), staying on the Internet longer than planned (β=0.19, p<0.001), and
composite CIUS scores (β=0.18, p=0.002).
38
Table 4. Standardized Coefficients of CIUS measures and Gender (N=1,367)
Unadjusted Models Adjusted Models
CIUS 1:
Uses
Internet
more
than
ought to.
CIUS 2:
Stays on
Internet
longer
than
planned.
CIUS 3:
Can’t cut
down on
Internet
use.
CIUS 4:
Internet
use
seems
beyond
their
control.
CIUS
Total
CIUS 1:
Uses
Internet
more
than
ought to.
CIUS 2:
Stays on
Internet
longer
than
planned.
CIUS 3:
Can’t cut
down on
Internet
use.
CIUS 4:
Internet
use seems
beyond
their
control.
CIUS
Total
β (SE) β (SE) β (SE) β (SE) β (SE) β (SE) β (SE) β (SE) β (SE) β (SE)
Female
0.15
(0.05)
0.12
(0.05)
0.03
(0.05)
0.06
(0.05)
-0.11
(0.05)
0.23
(0.05)***
0.19
(0.05)***
0.05
(0.06)
0.09
(0.06)
0.18
(0.06)**
Read News
Online
0.16
(0.06)
0.15
(0.06)
-0.09
(0.06)
-0.13
(0.06)
0.03
(0.06)
0.00
(0.06)
0.004
(0.06)
-0.15
(0.06)*
-0.15
(0.06)*
-0.09
(0.06)
E-mailed
0.34
(0.05)***
0.32
(0.05)***
0.16
(0.05)**
0.07
(0.05)
0.28
(0.05)***
0.07
(0.06)
0.07
(0.06)
0.10
(0.06)
0.02
(0.06)
0.09
(0.06)
Social
Networking (such
as Myspace or
Facebook)
0.46
(0.06)***
0.40
(0.06)***
0.12
(0.06)
0.05
(0.06)
0.32
(0.06)***
0.24
(0.06)***
0.20
(0.06)***
0.04
(0.06)
-0.2
(0.06)
0.15
(0.06)*
Read/Participated
in online forum
discussion
0.46
(0.09)***
0.42
(0.09)***
0.26
(0.09)
0.13
(0.09)
0.40
(0.09)***
0.10
(0.09)
0.06
(0.09)
0.12
(0.10)
0.06
(0.10)
0.11
(0.09)
Did homework
online
0.04
(0.07)
0.01
(0.07)
-0.15
(0.07)
-0.13
(0.07)
-0.07
(0.07)
-0.11
(0.06)
-0.14
(0.06)*
-0.20
(0.07)**
-0.15
(0.07)*
-0.19
(0.07)**
Played online
games
0.30
(0.06)***
0.30
(0.06)***
0.12
(0.06)
0.09
(0.06)
0.26
(0.06)***
0.12
(0.06)*
0.15
(0.06)*
0.07
(0.06)
0.09
(0.06)
0.14
(0.06)*
Watched movies
or video clips
online
0.36
(0.05)***
0.31
(0.05)***
0.03
(0.05)
0.01
(0.05)
0.23
(0.05)***
0.13
(0.05)*
0.10
(0.06)
-0.05
(0.06)
-0.03
(0.06)
0.05
(0.06)
Chatted online
(IM, audio or
video) with
0.52
(0.05)***
0.53
(0.05)***
0.27
(0.06)***
0.21
(0.06)**
0.48
(0.05)***
0.25
(0.06)***
0.30
(0.06)***
0.19
(0.06)**
0.19
(0.06)**
0.29
(0.06)***
39
people I know in
real life
Chatted Online
with people I met
online
0.57
(0.07)***
0.56
(0.07)***
0.33
(0.07)***
0.19
(0.07)
0.52
(0.07)***
0.20
(0.08)*
0.18
(0.08)*
0.10
(0.08)
0.02
(0.09)
0.16
(0.08)*
Chatted online
with strangers
0.80
(0.12)***
0.84
(0.12)***
0.61
(0.12)***
0.41
(0.12)**
0.84
(0.12)***
0.30
(0.13)*
0.38
(0.13)**
0.40
(0.14)*
0.31
(0.14)*
0.44
(0.13)***
Shopped online
0.27
(0.07)***
0.23
(0.07)***
0.04
(0.08)
-0.07
(0.08)
0.15
(0.08)
0.07
(0.07)
0.05
(0.07)
-0.01
(0.08)
-0.09
(0.08)
0.00
(0.07)
Visited porn site
0.44
(0.08)***
0.37
(0.08)***
0.07
(0.08)
0.08
(0.08)
0.31
(0.08)**
0.27
(0.08)***
0.20
(0.08)*
-0.02
(0.08)
0.05
(0.08)
0.16
(0.08)
Downloaded
music
0.31
(0.05)***
0.23
(0.05)***
0.06
(0.05)
0.04
(0.05)
0.20
(0.05)**
0.09
(0.05)
0.02
(0.06)
0.00
(0.06)
0.02
(0.06)
0.04
(0.06)
Note: Unadjusted models w/ Bonferroni correction significant when p<0.003; p< *0.05; p< **<0.01; p < ***<0.001
40
DISCUSSION
Results from the LCA and linear regression model support previous research suggesting
Internet use is associated with activities that provide content, process, and social gratifications
(LaRose, 2011; Stafford et al., 2004) and that process and social gratifications are more
indicative of CIU (Song et al., 2004). As hypothesized, the identification of four latent classes
suggest that distinct subgroups of Internet users exist and that these subgroups can further be
classified as classes exhibiting specific process and content gratifications (i.e., “Entertainment
and Surveillance” and “Dependent” classes) and social gratifications (i.e., “Social” class).
Activities associated with social interaction, such as social networking and chatting online with
people one met in real life, with people met online, and with online strangers were associated
with more problematic Internet use than other types of activities in the regression analysis.
Playing online games, watching movies or video clips online, and visiting porn sites – activities
that allow individuals to escape from reality (i.e., process gratifications) – were also associated
with specific CIU problems and higher composite CIU scores. The significance of these
activities and their associations with CIU is similar to those found in other studies (Anderson,
Steen, & Stavropoulos, 2016). Conversely, doing homework online and reading news online
which provides neither social nor process gratifications are found to deter CIU.
Furthermore, we found that problems associated with Internet use (i.e., staying on the
Internet longer than one planned and using the Internet more than one ought to) predict
membership in classes that demonstrate general reliance on the Internet for all activities (i.e.,
“Dependent” class) and Internet use for specific reasons (i.e., “Entertainment and Surveillance”
and “Social” classes). While some researchers have argued that that the terms “Internet
addiction”, “compulsive Internet use”, and “problematic Internet use” are too vague and
41
inadequate (Starcevic & Billieux, 2017; Van Rooij, Ferguson, Van de Mheen, & Schoenmakers,
2017), these results are aligned with previous studies that posit the existence of two types of
problematic Internet – generalized PIU and specific PIU (Caplan, 2010; Davis, 2001), and
provide additional support for maintaining the distinction between generalized PIU and specific
PIU (Montag et al., 2015; Pontes & Griffiths, 2014). Moreover, these results suggest that certain
CIU consequences may be more severe among individuals with generalized PIU than those who
use the Internet for specific purposes.
On the other hand, when CIU scores were computed by aggregating across many
consequences, higher general CIU was associated with higher probability of membership in the
“Social” class compared to those in the “Entertainment and Surveillance” class and there was no
statistically significant difference between the “Social” and “Dependent” classes. These findings
may indicate that overall CIU in the “Social” class may be just as problematic as in the
“Dependent” class. Future studies should examine the long-term consequences of CIU based on
class membership – whether these CIU consequences become progressively more problematic
over time and if consequences manifest in Internet specific classes, particularly the “Social”
class, with the absence of the Internet.
The most surprising observation was that those who reported lower loss of control of
Internet use were more likely to be classified in the “Entertainment and Surveillance” class than
the “Functioning” class even though the “Entertainment and Surveillance” class was associated
with more problematic Internet use otherwise. This finding may be explained by differences in
Internet use self-efficacy – confidence in the ability to carry out a task online (Bandura, 1982;
LaRose & Kim, 2006) – and gender.
42
More prevalent activities in the “Entertainment and Surveillance” class, such as online
gaming and downloading music, have been associated with higher levels of self-efficacy among
users. For example, involvement and interaction with the online gaming community was found
to be more common among individuals with higher Internet self-efficacy (Hopp, Barker, &
Schmitz Weiss, 2015) and positively predicted intentions to download pirated games (Phau &
Liang, 2012). Similarly, higher Internet self-efficacy was associated with downloading music
(LaRose, Lai, Lange, Love, & Wu, 2005) and intentions to continue downloading music from the
Internet despite the possibility of getting caught (LaRose & Kim, 2006). This finding highlights
the importance of examining self-efficacy as a mediator between Internet use and the
manifestation of CIU problems and suggests future studies examine whether “Internet use seems
beyond one’s control” accurately assesses CIU or unintentionally assesses Internet self-efficacy.
Previous studies have also found gender differences in Internet use and Internet self-
efficacy. Compared to females, males have traditionally been more ardent Internet users
(Bimber, 2000; Young, 2007), self-report higher levels of Internet self-efficacy (Durndell &
Haag, 2002; Torkzadeh & Koufteros, 1994), and are more likely to be addicted to online gaming
(Bernardi & Pallanti, 2009; Frangos et al., 2011) and downloading (Frangos et al., 2011). More
intense use of the Internet may contribute to increased self-efficacy and higher prevalence of
gaming and downloading among males may further explain why those reporting less loss of
control are in the “Entertainment and Surveillance” than the “Functioning” class. We were
unable to include gender as a covariate in our LCA model. Future studies should examine the
role of gender and self-efficacy in predicting class membership of similar latent classes,
particularly as more females are beginning to engage in what was previously considered
43
predominantly male Internet activities, such as online gaming (Taylor, 2003) and as the gender
digital divide closes (Fallows, 2005).
The identification of a generalized Internet class and two specific Internet activity classes
as well as the significance of different class memberships based on the four CIU items and
composite CIUS scores suggest the importance of creating tailored intervention and prevention
approaches for adolescents who fall in these different latent classes. Pontes and Griffiths (2014)
described generalized PIU as addiction to the Internet rather than addiction on the Internet. Thus,
those in the “Dependent” class may need tools to limit overall Internet use (e.g., applications that
restrict their time on the Internet), and parents, educators, and friends may play integral roles in
monitoring Internet use and re-evaluating social norms surrounding Internet behavior. To combat
the process gratifications offered by the activities in the “Functioning” class may also require
subscription to slower Internet connectivity to diminish the pleasure of Internet use experiences.
Innovative methods to make one aware of his/her Internet use , such as the incorporation of
mindfulness in prevention and intervention programs, may also deter generalized CIU,
especially in relation to consequences that result from prolonged and aimless use (e.g., conflicts
with parents, disengagement in other real-life activities).
Conversely, activity specific Internet use or addiction on the Internet, such as those in the
“Entertainment & Surveillance” and “Social” classes, may require interventions that not only
target the consequences of participating in these activities online but also offline and should
address underlying psychological and social motives for engaging in these activities (e.g.,
psychological and social determinants that drive use). Those who fall in the “Social” class, in
particular, may require more novel intervention and prevention approaches as social interaction
that occurs online is unique to the Internet and similar settings cannot be replicated offline (e.g.,
44
the ability to talk to multiple people at once, meet several people of various backgrounds without
much effort, no limitations to time and place) (Montag et al., 2015). Specifically, interventions
that target those in the “Social” class need to enlist the help of parents and other authority figures
to rigorously monitor the types of chatrooms and forums in which adolescents are participating
and need to address ways adolescents in this class can engage with peers offline in safe settings.
This includes educating parents and adolescents about the risks of CIU and chatting with online
strangers (e.g., sexual predation and harassment) (Falender, 2017; Soh, Chew, Koay, & Ang,
2018), making sure parents are just as self-efficacious as their children about computer use
(Correa, Straubhaar, Chen, & Spence, 2015; Van den Bulck, Custers, & Nelissen, 2016), and
providing curriculum that teaches adolescents skills to reduce social anxiety and increase self-
confidence in offline social situations. Since more frequent Internet use has been found to be
associated with increased instances of being solicited for sex online (Mitchell, Finkelhor, &
Wolak, 2001), adolescents who exhibit CIU and fall in the “Social” class may be more
vulnerable to such consequences. Cognitive behavior therapy may be important in treating
adolescents who fall in the “Social” class to target the psychological antecedents that contribute
to dependency on the Internet for social interactions. In addition, industrial and policy reform,
such as restrictions (e.g., age restrictions and time limitations) enforced by the creators of the
chatroom/online forum and the government, may be more effective in combating consequences
associated with behaviors in the “Social” class.
Limitations
Several limitations to the current study should be considered. Baseline data was collected
in 2008 and 2009. Since then, accessibility to the Internet and Internet applications have
extended beyond traditional computers with the advent and proliferation of the smartphone and
the variety of Internet applications have grown exponentially resulting in the emergence of other,
45
trendier applications not captured by the TND survey (Anderson & Jiang, 2018). Despite this
limitation, results from this study may offer insight of forthcoming consequences resulting from
Internet use given that new Internet applications and smartphones may make social connection
with strangers more accessible.
Second, LCA is a data-driven, exploratory analysis. Labeling of latent groups is driven by
the data analyzed and inferred by the data analyst(s). In addition, this study was conducted
among CHS youth who are more vulnerable to problematic behavior than normal high school
youth (Lisha et al., 2014). Thus, the generalizability of the classes may be limited to the present
population and may not apply to other continuation high school youths or regular high school
youth. Replication of this study among regular, ethnically diverse, adolescent populations is
needed to confirm the existence of a four-class solution.
The CIUS items in this study were limited to four items related to compulsion and does
not encompass the breadth of consequences that may be associated with CIU. Future studies
should consider exploring more varied CIU problems (e.g., conflicts with friends and family
members, difficulty related to functioning offline), which will provide a better perspective of the
most relevant and immediate Internet consequences related to these identified latent classes.
Last, Internet activity use was confined to the last seven days and a simple, binary
response (i.e., Yes/No). This brief assessment of Internet use activity does not account for
frequency and extended use of Internet activities which may be important indicators of CIU.
Future studies should consider using more nuanced measures (e.g., frequency and count data
over a longer period of time) to assess the association between Internet use activity and CIU.
Other research methods such as qualitative and event-level research may also provide more
insight regarding this association.
46
CONCLUSION
This is the first study to our knowledge that utilized LCA to examine Internet activities
and CIU. Given that Internet activities rarely occur independently, LCA allows researchers to
understand patterns of Internet use that are most associated with CIU. As with previous studies,
this study supports the notion that different types of Internet dependency exist. Future studies
and intervention and prevention programs should continue to account for these differences as
long-term effects of CIU are further explored among different modalities and platforms (e.g.,
smartphones).
47
CHAPTER 3: Evaluation of Pleasure & CIU using the Electronic Pleasurable Events
Schedule (E-PES)
BACKGROUND
The Uses and Gratifications theory (UGT) stems from communication research and
posits that individuals actively choose to engage with specific media content, such as the Internet
(Blumler & Katz, 1974; Stafford et al., 2004). A key tenet of UGT is the notion that individuals
choose to engage in certain media content on the Internet because it fulfills certain needs, or
(perceived) gratifications (Ruggiero, 2000). These (perceived) gratifications often lead to
initiation and habitual use of the Internet. For example, Lin (2001) found that gratification-
seeking motivations – specifically entertainment, informational learning, and escapism – are the
best predictors of online service adoption.
Given the wide range of services that can be offered through an online platform,
gratifications associated with traditional media (e.g., radio, newspaper, and television) have been
modified to encompass gratifications offered by the Internet that are otherwise unobtainable.
Stafford et al. (2004) identified three types of media gratifications specific to the Internet –
content gratifications, process gratifications, and social gratifications. Several studies have
broken these gratifications to more specific subcategories such as self-expression, surveillance
(information-seeking), entertainment, escapism, and social interaction (Alhabash et al., 2014;
Courtois, Mechant, De Marez, & Verleye, 2009; Parker & Plank, 2000; Whiting & Williams,
2013).
Process gratifications, gratifications related to the pleasure derived from the actual use of
the Internet, have most often been linked to compulsive Internet use (CIU). Process gratifications
have the capability of removing users from real life into an alternate (virtual) world as the user
48
becomes more entrenched in the Internet activity (Song et al., 2004). However, researchers have
suggested that social gratifications are also important motives for Internet use (Papacharissi &
Rubin, 2000) and can often lead to Internet dependency (Caplan, 2002; Leung, 2014; Tang et al.,
2016; Yang & Tung, 2007). Activities related to social media use (e.g., Facebook) (Kuss et al.,
2013; Ryan, Chester, Reece, & Xenos, 2014) and massively multiplayer online role playing
games (e.g., World of Warcraft) (Kuss, Griffiths, Karila, & Billieux, 2014; Tone, Zhao, & Yan,
2014) have been linked with CIU as they are able to fulfill both process gratifications (e.g.,
escapism) and social gratifications (e.g., social interaction with others).
Efforts have been made to distinguish the differences between gratifications sought (GS)
– perceived enjoyment of the media activity, and gratifications obtained (GO) – enjoyment
received from media use. Comparatively, UGT gratifications research is predominantly focused
on GS rather than on GO as it is considered to be a better predictor of (initial) media
consumption (LaRose, 2011). However, researchers argue that GO may player a greater role in
CIU (LaRose, 2011; Song et al., 2004). While the perceived gratifications initiate use,
gratifications or pleasure the individual receives from conducting the activity is more likely to
lead to habitual Internet use (Hicks et al., 2012; Song et al., 2004). As Internet use becomes more
habitual, gratifications originally sought may no longer pertain to the gratifications actually
received (LaRose, 2011); when addiction occurs, habits become non-conscious, automated
processes and participants may have difficulty reconciling their actions with initial GS. The role
that intrinsic motivations like GO, or pleasure, plays in CIU are not well understood and often
confounded by GS (LaRose & Eastin, 2004).
49
Vulnerability to CIU may be further compounded by neuropsychological predisposition
to sensitivity to reward cues; addicts are typically less sensitive to pleasant, less intense stimuli
due to dopaminergic dysfunctions in the brain (Rømer, Whybrow, & Kringelbach, 2015).
Internet addicts are commonly diagnosed with comorbid psychological diagnoses such as
depression (De Leo & Wulfert, 2013; Huang, 2010; Lee et al., 2013), obsessive-compulsive
disorder (Bernardi & Pallanti, 2009; Lin et al., 2011; Zhang et al., 2008), and (social) anxiety
(De Leo & Wulfert, 2013; Younes et al., 2016). Trait anhedonia, a key indicator of depression
that has been associated with lower reward sensitivity and the inability to experience pleasure
among non-clinical populations (Keller et al., 2013), significantly predicted Internet addiction
among a sample of continuation high school students (Guillot et al., 2016). The Internet’s ability
to provide an abundant amount of resources and immediate reward without much effort may spur
Internet addiction as individuals who are more anhedonic seek the Internet as a way to
manipulate reward systems in the brain and stimulate hedonic experience (Guillot et al., 2016).
The purpose of this study was to examine the relationship between the derived pleasure
from Internet activities (GO) and CIU accounting for trait anhedonia. This is the first study to
our knowledge that directly asks participants to rate how much pleasure was derived from
engaging in the activity and examines the relationship between GO and problematic Internet use.
Results from this study will help us better understand the intrinsic motivations associated with
different types of Internet activities and its association with CIU – an aspect of Internet research
using UGT that has been understudied in the literature. This information can provide pertinent
information for CIU prevention and intervention programs, particularly among anhedonic
adolescents who may be more susceptible to problematic Internet use.
50
METHODS
Research Questions
This dissertation addressed the following research questions:
Question 3: Do baseline pleasure scores associated with engagement in Internet activities
predict CIU approximately one year later accounting for baseline CIU, age, and
ethnicity?
Hypothesis 3: Higher reported pleasure derived from electronic use is positively
associated with CIU.
Question 4: Do anhedonia and gender moderate the relationship between pleasure
and CIU?
Hypothesis 4: Pleasure derived from Internet use will be more strongly
associated with CIU among individuals who report greater levels of anhedonia.
Pleasure derived from online game playing activities will be more strongly
associated with CIU among males while pleasure derived from social online
activities will be more strongly associated with CIU among females.
Data Collection
Data for Study 2 comes from the Happiness and Health (H&H) study (Leventhal et al.,
2015). H&H is a longitudinal survey of substance use and mental health among high school
students enrolled in ten public schools surrounding Los Angeles, CA, USA. The goal of H&H
was to follow participants from ages 14 to 18 (9
th
through 12
th
grade), a period of time when
adolescents begin to explore and engage in risky behaviors that may be detrimental to health.
Approximately 40 public high schools in Los Angeles were approached for participating in the
study due to their diverse demographics and proximity. Of the 40 schools approached, ten public
high schools agreed to participate.
51
To be enrolled in the study, students had to provide both written or verbal assent and
parental consent. Students who were taking special education (e.g., severe learning disabilities)
or English as a Second Language programs were excluded from participation. Five waves of data
have been collected thus far with approximately six months between each data collection.
Baseline and six-month follow-up measures were collected in Fall 2013 and Spring 2014,
respectively, when the students were in 9
th
grade. 12-month follow-up and 18-month follow-up
were collected in Fall 2014 and Spring 2015, respectively, when the students were in 10
th
grade.
Twenty-four month follow-up was collected in Fall 2016 when the students were in 11
th
grade.
At each wave of data collection, students were administered pencil and paper surveys across two
60-minute periods less than two weeks apart in the students’ classrooms onsite. If the students
were absent during data collection, phone or Internet surveys were given. All study procedures
were reviewed and approved by the University of Southern California Institutional Review
Board.
A total of 4,100 students were eligible to participate in the study. Of the 3,874 students
(94.5%) who assented to participate in the study, 3,396 students (87.7%) provided parental
consent. The final sample size at baseline was 3,383 participants who completed the survey.
MEASURES
Demographic variables. Self-reported items were used to assess age (in years), gender
(male vs. female), and ethnicity (Asian, White, Hispanic, Black/African American, American
Indian/Alaska Native, Native Hawaiian/Pacific Islander, Other, and I cannot choose only one
term). To increase the sample and statistical power, ethnicity was collapsed into three groups
based on similar CIU behaviors: Asian and Whites, Hispanic, and Other.
52
Electronic Pleasurable Events Schedule (E-PES). The E-PES was modeled after the
Pleasant Events Schedule (PES), a scale commonly used among geriatric, clinical and research
populations to measure the frequency of and subjective pleasure received from rewarding events
such as being at the beach, buying things for oneself, and watching TV (Lewinsohn & Graf,
1973). Similar to the PES, the E-PES was constructed to measure the frequency of and subjective
pleasure received from engaging in certain electronic or online activities over the last six months.
The E-PES consisted of 14 electronic and online events that were adapted from items assessing
electronic media use in Project Towards No Drug Abuse (Rohrbach, Sun, & Sussman, 2010):
checking social media sites, posting, liking or commenting, sharing, browsing, reading,
streaming videos, streaming music, chatting online, texting, video-chatting, online shopping,
playing games with your friends and family, playing games by yourself. Participants were first
asked to indicate the frequency of use in the past six months on a Likert scale: “never” (0), “a
little - one or two times” (1), or “everyday” (2). Then, participants were asked to indicate how
much pleasure they received from each activity using a Likert scale: “not at all” (0), “a little” (1),
and “a lot” (2). Internal consistency for each scale is good for the E-PES frequency component
(Cronbach’s alpha = 0.86) and excellent for the E-PES pleasure component (Cronbach’s alpha =
0.90). Given high multicollinearity between the E-PES frequency items and pleasure items, only
the pleasure items were analyzed in this study.
Anhedonia. Trait anhedonia was assessed using the Snaith–Hamilton Pleasure Scale
(SHAPS) (Snaith et al., 1995). Participants were asked fourteen items that inquired about
experiencing pleasure or enjoyment spanning various activities including sensory stimuli (e.g., “I
would find pleasure in the scent of flowers, or the smell of a fresh sea breeze, or freshly baked
bread”), social activities (e.g., “I would enjoy being with my family or close friends”), and
53
hobbies (e.g., “I would enjoy my favorite television or radio program”). Response items were
rated on a four point Likert scale ranging from “strongly disagree” (0) to “strongly agree” (3).
These items were reverse coded where “strongly agree” received a score of 0 and “strongly
disagree” received a score of 3. Higher sum scores indicated greater trait anhedonia with 56
being the highest score that can possibly be received. This scale showed excellent internal
reliability (Cronbach’s alpha = 0.94), consistent with other literature which demonstrated
adequate internal reliability and convergent validity among a sample of adolescents (Leventhal et
al, 2015).
Compulsive Internet Use Scale (CIUS). CIU was measured at Wave 3 and Wave 5 of the
H&H study using the Compulsive Internet Use scale (Meerkerk, Van Den Eijnden, Vermulst, &
Garretsen, 2009), a 14-item self- response measure. The CIUS was developed as a short and
easily administered scale to assess CIU derived from addiction literature and was tested among
samples of heavy and regular Internet users (Meerkerk et al., 2009). The CIUS has shown high
internal consistency from previous studies (Cronbach’s alpha ranging between 0.89-0.90)
(Meerkerk et al., 2009). Participants were asked, “How often… 1) …do you find it difficult to
stop using the Internet when you are online?; 2) …do you continue to use the Internet despite
your intention to stop?; 3) …do others (e.g., partner, children, parents) say you should use the
Internet less?; 4) ...do you prefer to use the Internet instead of spending time with others (e.g.,
partner, children, parents)?; 5) …are you short of sleep because of the Internet?; 6) …do you
think about the Internet, even when not online?; 7)… do you look forward to your next Internet
session?; 8) … do you think you should use the Internet less often?; 9) … have you
unsuccessfully tried to spend less time on the Internet?; 10) …do you rush through your
homework in order to go on the Internet?; 11) … do you neglect your daily obligations (work,
54
school, or family life) because you prefer to go on the Internet?; 12) … do you go on the Internet
when you are feeling down?; 13) … do you use the Internet to escape from your sorrows or get
relief from negative feelings?; and 14) … do you feel restless, frustrated, or irritated when you
cannot use the Internet?” Each item was scored on a 5-point Likert scale: Never (0), Seldom (1),
Sometimes (2), Often (3), Very often (4). A weighted sum score across the 14 items was
computed to assess endorsed CIU at both time points. The CIUS exhibited good internal
consistency in this sample at T1 (Cronbach’s alpha = 0.95) and T2 (Cronbach’s alpha = 0.95).
Analytic Plan
Although H&H has five waves of data collected, data for this study only utilized data
collected from wave 3, which will be referred to as Time 1 (T1), and wave 5, which will be
referred to as Time 2 (T2), as these are the only two time points that measure E-PES and CIUS.
All statistical analyses were run in SAS Version 9.4 (SAS Institute, Cary, NC, USA).
Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA).
An exploratory factor analysis (EFA) of the 14 E-PES measures was performed on half of
the sample (N=1,106). Factor loading coefficients of 0.4 or greater were used as a cutoff value,
and factors with eigenvalues greater than 1.0 were retained for analysis. Following the EFA, a
confirmatory factor analysis (CFA) was performed on the other half of the sample (N=1,106) to
verify the factor structure. Goodness-of-fit indices were used to evaluate the model fit of our
data: Goodness of fit Index (GFI), Bentler comparative fit index (CFI), and Normed Fit Index
(NFI), Bentler Non-normed Index (NNI) values above 0.90 and root mean squared error or
approximation (RMSEA) values close to and less than 0.06 indicate acceptable model fit (Hu &
Bentler, 1999). Weighted mean scores of each E-PES factor were used as predictors of the
multilevel regression model.
55
Hierarchical Linear Regression Model (HLM)
Hierarchical linear regression models (HLM) were conducted using the SAS PROC
MIXED function and maximum likelihood estimation (ML) to examine the association between
CIU and the three identified factors of E-PES, controlling for age and ethnicity at T1, T2, and
from T1 to T2.
1
The longitudinal model examining CIU change between T1 and T2 had an added
covariate of baseline CIU. These models also assessed the interactions between trait anhedonia
and gender and each of the three factors of CIU. Since our research questions are only mainly
interested in level-1 effects controlling for the random effect of school on CIUS (Peugh, 2010),
all measures except for gender and ethnicity were centered on its grand mean (GMC). See Figure
4.
First, intraclass correlation (ICC) was calculated to determine the proportion of CIU
variation that occurs across school.
Level 1:
Y
ij
= Π
0j
+ εij
Next, an a priori model was generated based on the literature. School was considered a
random effect or level-2 (macro unit, between-group) predictor in the model. All other individual
predictors and covariates in this model were considered fixed effects or level-1 (micro unit,
within-group) predictors. The final model is as follows, with random effects included for the
intercept across schools.
1
Socio-economic status (SES) was assessed using the free and/or reduced lunch variable. Preliminary chi-square
analysis demonstrated no significant differences in CIU among those who did not receive free and/or reduced
lunch, those who received reduced lunch, and those who received free lunch. In addition, SES was not a significant
predictor in any of the MLM models and was subsequently excluded from further analysis.
56
Level 1:
Y
ij
= Π
0j +
Π
1j
(CIUS
GMC ij
) + Π
2j
(E-PES1
GMC ij
) + Π
3j
(E-PES2
GMC ij
)
+ Π
4j
(E-PES3
GMC ij
) + Π
5j
(Anhedonia
GMCij
)
+ Π
6j
(Gender
ij
) +
Π
7j
(Age
GMCij
)
+ Π
8j
(Ethnicity
ij
) + Π
9j
(Anhedonia
GMCij
*E-PES1
GM ij
)
+ Π
10j
(Anhedonia
GMCij
*E-PES2
GMC ij
)
+ Π
11j
(Anhedonia
GMCij
*
E-PES3
GMCij
) + Π
12j
(Gender
ij
*
E-PES1
GMCij
)
+ Π
13j
(Gender
ij
*
E-PES2
GMCij
) + Π
14j
(Gender
ij
*
E-PES3
GMCij
) + [ε
ij
]
Level 2
Intercept: Π
0j =
γ
00
+ ξ
0j
Π
1j =
γ
10
Π
2j =
γ
20
Π
3j =
γ
30
Π
4j =
γ
40
Π
5j =
γ
50
Π
6j =
γ
60
Π
7j =
γ
70
Π
8j =
γ
80
Π
9j =
γ
90
Π
10j =
γ
100
Π
11j =
γ
110
Π
12j =
γ
120
Π
13j =
γ
130
Π
14j =
γ
140
57
Combined:
Y
ij
= [γ
00
+ γ10 (CIUS
GMC ij
) + γ
20
(E-PES1
GMC ij
) + γ30 (E-PES2
GMC ij
) + γ
40
(E-PES3
GMC ij
) +
γ
50
(Anhedonia
GMC ij
) + γ
60
(Gender
ij
)
j
+
γ
70
(Age
GMCij
) + γ
80
(Ethnicity
ij
)
+ γ
90
(Anhedonia
GMC ij
*
E-PES1
GMC ij
) + γ
100
(Anhedonia
GMC ij
*
E-PES2
GMC ij
)
+ γ
110
(Anhedonia
GMC ij
*E-PES3
GMC ij
) + γ
120
(Gender
ij
*E-PES1
GMC ij
)
+ γ
130
(Gender
ij
*E-PES2
GMCij
) + γ
140
(Gender
ij
*E-PES3
GMC ij
)] + [ε
ij
] + [ξ
0j
]
RESULTS
Sample
Only participants who answered key variables of interest in this study (i.e., at least one
activity in each of the three E-PES factors, all demographic measures, and had a weighted CIUS
score) at both T1 and T2 was retained for analysis in the sample (N=975). Average age of the
participants at T1 was 15.53 years old (SD=0.41); 61% were male. Almost half of the sample
Figure 4. Conceptual Model for Study 2.
58
self-identified as “Hispanic/Latino” (43%), followed by “Asian” (21%), “Other” (21%), and
“White” (15%). Average anhedonia score was 23.27 (SD=7.90) at T1 and 23.19 (SD=7.40) at
T2, with 23% and 18% of the sample exhibiting clinical levels of anhedonia, respectively (Snaith
et al., 1995). Average CIU scores at T1 and at T2 were 20.28 (SD=13.49) and 20.57 (SD=13.18),
respectively. Average E-PES1, E-PES-2, and E-PES-3 scores are 1.43 (SD=0.42), 1.64
(SD=0.41), and 1.63 (SD=0.48), at T1, respectively, and 1.45 (SD=0.41), 1.65 (SD=0.39), and
1.63 (SD=0.47), at T2, respectively. See Table 5.
Table 5. Study 2 Sample Characteristics (N=975)
Variables Mean (SD) or N (%)
Age at T1
15.53 (SD = 0.41)
Gender Male
595 (61%)
Ethnicity Hispanic/Latino
419 (43%)
Asian
205 (21%)
White
146 (15%)
Other (e.g., African
American, Native
American, Mixed, or
Other)
205 (21%)
T1 T2
CIUS
20.28 (SD=13.49) 20.57 (SD=13.18)
Anhedonia
23.27 (SD = 7.90) 23.19 (SD=7.40)
EPES-1: Social
Surveillance
1.43 (SD=0.42) 1.45 (SD=0.41)
EPES-2:
Entertainment
Surveillance
1.64 (SD=0.41) 1.65 (0.39)
EPES-3:
Entertainment
Game Playing
1.63 (SD=0.48) 1.63 (SD=0.47)
59
Assessment of Missing Data and Attrition
Several statistical analyses were used to assess potential sample bias introduced by
excluding individuals who did not have complete data at T1 and T2 (N=2,421, 71%). Chi-square
analysis and t-tests were used to assess if the excluded sample differed from the sample retained
for analyses (N=975) on 8 key measures of interest (i.e., CIU scores, E-PES-1, E-PES-2, E-PES-
3, gender, age, anhedonia, and ethnicity). Significant differences between those with missing
data and the final sample was found for T1 CIU (t=-4.04, p<0.0001), E-PES-1 (t= -9.23,
p<0.0001), E-PES-2 (t=-9.19, p<0.0001), E-PES-3 (t=-20.53, p<0.0001), gender (χ
2
=127.76,
p<0.0001), anhedonia (t=4.52, p<0.0001), and ethnicity (χ
2
=31.74, p<0.0001). The retained
sample had higher baseline (T1) CIU scores (18.11 vs 20.28), mean E-PES-1 scores (1.26 vs
1.43), mean E-PES-2 scores (1.46 versus 1.64), and mean E-PES-3 scores (1.06 vs 1.63), but had
lower anhedonia (24.72 vs 23.27). Similarly, differences were found between those with missing
data and the final samples for T2 CIU (t=-4.30, p<0.0001), E-PES-1 (t=-5.90, p<0.0001), E-
PES-2 (t=-6.27, p<0.0001), E-PES-3 (t=-15.37, p<0.0001), gender (χ
2
=134.97, p<0.0001), age
(t=-18.06, p<0.0001), anhedonia (t=23.19, p<0.0001), and ethnicity (χ
2
=31.74, p<0.0001). The
retained sample also had higher T2 CIU (18.28 vs 20.57), E-PES-1 (1.34 vs 1.45), E-PES-2 (1.53
vs 1.63), E-PES-3 (1.19 vs 1.63) but had lower anhedonia (24.44 vs 23.19). Males (39.15% vs
60.87%; 38.75 vs 60.83) and Asians (13.86% vs 21.51%) were more likely to be retained in the
final sample at both time points, and older individuals at T2 were more likely to be retained for
analysis (16.21 years old vs 16.50 years old).
Results of EFA and CFA
Three latent factors were identified from the EFA: a “Social Surveillance” factor, an
“Entertainment Surveillance” factor, and an “Entertainment Gaming Factor” (see Table 6).
60
Subsequently, results from the CFA demonstrated acceptable model fit. Although chi-square was
significant (chi-square=403.905, DF=74, and p<0.0001), other model fit criterion demonstrated
acceptable fit: GFI>0.90, CFI>0.90, NFI>0.90, NNI>0.90, and RMSEA≤0.06 (see Table 7 and
Figure 5). The CFA explained 56% of the common variance.
61
Table 6. Results of Electronic Pleasure Events Schedule (E-PES) Exploratory Factor Analysis
(N=1,106)
E-PES Pleasure Item
Factor 1:
Social
Surveillance
Factor 2:
Entertainment
Surveillance
Factor 3:
Entertainment
Game playing
Checking social media sites (Facebook,
twitter, Instagram, etc.).
0.69 -- --
Posting your own photos, images, vides, status
updates, or blogs
0.77 -- --
Liking or commenting on other people’s
statuses wall posts, pictures, etc.
0.78 -- --
Sharing other people’s photos, images, videos,
status updates, blogs, articles, news or
websites
0.70 -- --
Browsing or viewing photos, images or videos
(YouTube, Vine, Pinterest, Imgur, or Reddit,
etc.)
-- 0.71 --
Reading Blogs, articles, news, online forums,
or books on a phone, tablet, or computer
-- 0.50 --
Watching streamed television shows or
movies (Netflix, Hulu, iTunes, etc.)
-- 0.67 --
Streaming or downloading music (ITunes,
Pandora, YouTube, etc.)
-- 0.72 --
Chatting online (instant messaging, Facebook
messenger, etc.)
.60 -- --
Texting (text messaging)
.53 -- --
Video chatting (Skype, Facetime, Omegle,
etc.)
0.60 -- --
Online shopping or viewing products online
(clothes, electronics, games, etc.)
0.42 -- --
Playing games with your friends and family
on a console (Xbox, PlayStation, Wii),
personal computer, or cell phone
-- -- 0.86
Playing games by yourself on a console
(Xbox, PlayStation, Wii), personal computer,
or cell phone
-- -- 0.86
Note: Only Item factor loadings >0.40 are presented here (in bold).
62
Table 7. Results of Electronic Pleasure Events Schedule (E-PES)
Confirmatory Factor Analysis (N = 1,106)
Chi-Square (χ
2
) 403.905***
Degrees of Freedom (DF) 74
Goodness of Fit Index (GFI) 0.947
Bentler Comparative Fit Index (CFI) 0.938
Normed Fit Index (NFI) 0.925
Bentler Non-normed Index (NNI) 0.924
RMSEA 0.064 (0.058, 0.070)
Note: *** p<0.001
Figure 5. Results of Exploratory Factor Analysis of Electronic Pleasure
Events Schedule (N=1,106)
63
Results of the MLM
ICC was first conducted to determine how much variance occurs between schools. The
ICC for our model was 5% at T1 and 2% at T2, indicating that 5% and 2% of the total variance
in our model at T1 and T2, respectively, occurs between schools and should be accounted for in
our models.
Results of the MLM models are presented in Table 8. At T1, we see that E-PES-2
(Entertainment Surveillance) was associated with CIU controlling for age, gender, ethnicity, and
random intercept across school. For every one unit increase in E-PES-2 (Entertainment
Surveillance) scores, CIU increases by 7.43 units (p=0.001). That is, the more pleasure one
received from conducting the activities in the E-PES-2 factor, individual students were reporting
experiencing more frequent and severe CIU consequences. When evaluated at the mean value for
all E-PES scores, males had lower CIU scores than females (b=-3.48, p<0.001) and those who
self-reported as “Other” ethnicity had lower CIUS scores compared to Asians and Whites (b=-
3.15, p=0.001). Although anhedonia was a significant predictor in the model (b=0.29, p<0.001),
anhedonia did not moderate the relationship between any of the E-PES factors and CIU.
However, the association between E-PES-3 (Entertainment Game Playing) and CIU was
moderated by gender; the relationship between pleasure received from entertainment game
playing and CIU was stronger among males than females (b=4.72, p=0.02).
Similar to T1, E-PES-2 (Entertainment Surveillance) at T2 was associated with higher
CIU scores, controlling for age, ethnicity, and random intercept across schools. For every one
unit increase in E-PES-2 (Entertainment Surveillance) scores, CIU scores increased by 5.68 units
(p=0.02). When evaluated at the mean value for all E-PES scores males also had lower CIU
scores than females (b=-2.75, p=0.002) and those who self-identified as Hispanic (b=-3.72,
64
p=0.0001) or “Other” ethnicity (b=-4.12, p<0.0001) had lower CIU scores compared to Asians
and Whites. At T2, the relationship between pleasure from entertainment surveillance and CIU
was stronger among adolescents who were more anhedonic (b=0.38, p=0.04). However, gender
did not moderate any of the E-PES factors and CIU.
None of the predictors or the interaction terms in the longitudinal MLM model was
significant except for baseline CIU and ethnicity. Every one unit increase in baseline CIU score
is predictive of a 0.52 unit increase in CIU scores one year later (p<0.0001). Similar to T2,
identifying as Hispanic or “Other” ethnicity significantly predicts a 3.10 (p=0.001) and 2.84
(p=0.0004) unit decrease in CIU one year later compared to Asians and Whites, respectively.
The interaction effects of anhedonia and gender were also not statistically significant in this
model.
65
Table 8. Results of MLM Models for High School Students Participating in Happiness & Health
Study (N=975)
Wave 3 Wave 5 Wave 3 to 5
Intercept 24.04 (0.93)*** 25.47 (1.00)*** 22.76 (0.75)***
E-PES-1
(Social Surveillance)
1.08 (2.09) -1.52 (2.27) -1.07 (1.74)
E-PES-2
(Entertainment
Surveillance)
7.43 (2.29)** 5.68 (2.39)* 1.13 (1.92)
E-PES-3
(Entertainment Game
Playing)
-1.42 (1.54) 1.08 (1.46) 0.97 (1.28)
Male -3.48 (0.88)*** -2.75 (0.87)*** -0.33 (0.66)
Age -0.87 (1.03) -1.79 (1.02) -1.29 (0.85)
Ethnicity
Asian & White (REF) . . .
Hispanic -1.50 (1.15) -3.72 (1.13)** -3.10 (0.95)**
Other -3.15 (0.97)** -4.12 (0.97)*** -2.84 (0.79)***
Anhedonia 0.29 (0.06)*** 0.17 (0.06)** 0.07 (0.05)
Baseline CIUS -- -- 0.52 (0.03)***
Anhedonia*E-PES-1 0.29 (0.16) -0.14 (0.16) 0.22 (0.13)
Anhedonia*E-PES-2 -0.07 (0.17) 0.38 (0.19)* -0.05 (0.15)
Anhedonia*E-PES-3 0.01 (0.14) -0.13 (0.13) -0.08 (0.12)
Gender*E-PES-1 -1.16 (2.58) 2.76 (2.77) 0.20 (2.15)
Gender*E-PES-2 -4.15 (2.83) -4.56 (2.93) 0.25 (0.24)
Gender*E-PES-3 4.72 (2.06)* 1.65 (2.04) -0.38 (1.72)
Log Likelihood 7743.3 7714.5 7387.4
AIC 7777.3 7748.5 7423.4
AICC 7777.9 7749.1 7424.1
BIC 7782.4 7753.6 7428.8
Covariate 0.64 (1.22) 1.00 (1.33) 0.13 (0.58)
Residual 164.14 (7.48)*** 159.12 (7.25)*** 114.19 (5.20)***
Note: p<0.05*, p<0.01**, p<0.001***
66
DISCUSSION
Results from this study add to UGT in Internet use literature. This study specifically
addresses one of the limitations of UGT and provides critical information regarding how
pleasure is related to the development of CIU. In addition, it is the only scale that we know of,
that allows researchers to distinguish activities performed over a period of time, which is
important in distinguishing between active and non-active users. These findings may inform
future research on best approaches to ask about Internet use pleasure. Future studies should
continue to examine and compare the effectiveness of different scales in accurately assessing
GO.
While previous studies typically found at least four gratification factors (i.e.,
entertainment, information seeking, to pass time/relieve boredom, and social reasons) (LaRose,
2011), this study only identified three factors (Social Surveillance, Entertainment Surveillance,
and Entertainment Game Playing). Although only three factors were identified, the qualitative
labels of the factors were similar to those of previous studies (LaRose, 2011). The slight
difference in the number of identified factors may have resulted from differences in how the
items were assessed, the number of Internet activities assessed, and the age of the populations
sampled. Papacharissi and Rubin (2000), for example, examined 27 reasons for using the Internet
among college students using an Internet motivations scale; in contrast, this study evaluated 14
Internet activities among adolescents using a newly developed E-PES scale. Additional items
assessing information seeking and entertainment may result in the identification of an additional
factor, and information-seeking may be more prominent among college students than high
schoolers.
67
Results from the cross-sectional analyses demonstrated that E-PES-2, the Entertainment
Surveillance factor, was the only E-PES factor significantly associated with higher CIU scores at
both T1 and T2. This finding supports previous UGT studies that also found significant
associations between process gratifications and CIU (Caplan, 2002; Leung, 2014; Tang et al.,
2016; Yang & Tung, 2007). Activities in E-PES-2 (viewing photos, images or videos; reading
blogs, news, and online forums; streaming/watching TV shows and movies; and
downloading/streaming music) are activities that have replaced or are extensions of traditional
media (i.e., television, newspapers, and print magazine) rather than new activities specific to the
Internet as presented in the other two factors (e.g., social media and MMPORG). Motivations
that drove television addiction may also be motivations that are driving E-PES-2: learning (e.g.,
information seeking through reading blogs, news, and online forums), connection with real or
fictional people (e.g., streaming TV shows to compensate for loneliness), and affect motives
(e.g., TV shows for distraction) (Sussman & Moran, 2013). Furthermore, engagement in these
activities require little effort and focus. For users who may be seeking distractions from reality
with minimal effort, these activities may easily fulfill certain levels of gratification but may also
lead to aimless hours spent on the Internet and other CIU consequences. Additional research is
needed to identify the mechanisms contributing to engagement in these activities.
As previously mentioned, this is the first study to our knowledge that evaluated how
pleasure is associated with CIU using a scale that specifically asked participants to indicate how
much average pleasure was received among those who actually engaged in specific Internet
activities in the past 7 days. Previous studies asked participants how likely they were to agree
with reasons for media consumption as a proxy measure for assessing GO from the Internet, e.g.,
“I used the Internet… to participate in discussions… to meet new people… to look for
68
information” (LaRose, 2011; Papacharissi & Rubin, 2000). A major limitation of using such a
scale is the assumption that reasons for Internet use are equated with GO from use. Compared to
previous assessments of GO, the advantages of the newly developed scale used in this study are:
1) the ability to distinguish between participants who performed and did not perform the activity
over a certain period of time, and 2) the ability to evaluate actual GO from engaging in the
Internet activity. Thus, while previous studies found that GO was generally not associated with
CIU (LaRose, 2011), the results of this study found that GO obtained from E-PES-2
(Entertainment surveillance) was significantly associated with CIU cross-sectionally, which may
be explained by using a scale that more accurately captures GO.
Demographic factors were important indicators of CIU at T1 and T2. Similar to a
previous study conducted among high schoolers in Connecticut, non-Hispanic adolescents,
namely, Asians (Liu et al., 2011) and Whites (Young, 2007) had higher CIU scores than their
non-Asian and non-White counterparts. Gender differences in CIU scores were also detected.
Though males have been cited as more general problematic Internet users (Lin et al., 2011; Yang
& Tung, 2007), this study found that females had higher CIU scores than males. Given the
criterion for retention in our final sample, the females in our sample may generally be those who
tend to exhibit more problematic Internet use and engage in more Internet behaviors (e.g., online
game playing) than the females that were excluded from analysis. On the other hand, lower CIU
scores among males may also indicate the growing problem of CIU among females. In addition,
studies have found females as more problematic social networking users (Bernardi & Pallanti,
2009; Frangos et al., 2011) but that increased use of social networking has been associated with
greater depression (Moreno et al., 2011b). Greater levels of depression among more frequent
SNS users may have resulted in lower reported pleasure scores for EPES-1 (“Social
69
Surveillance”), which may explain why females have higher composite CIU scores than males
but the association between E-PES-1 (“Social Surveillance”) and CIU, specifically, is
insignificant in the MLM models. Furthermore, gender significantly moderated the relationship
between E-PES-3 (Entertainment Surveillance) and CIU at T1 but not at T2. This may be
explained by males as early adopters of game playing and as more avid gamers than females
(Kuss & Griffiths, 2012b). Though gender was not be a significant moderator of E-PES-3 and
CIU at T2, this may be indicative of the growing number of females playing games online
(Taylor, 2003), especially as males and females begin to fraternize during late adolescence.
Anhedonia was only a significant predictor in the model at T1 when all other E-PES
factors were evaluated at its mean, but significantly moderated the relationship between E-PES-2
(“Entertainment Surveillance”) and CIU at T2. Adolescents with more anhedonia may utilize the
Internet to seek instant gratification that is harder to obtain in real-life (Meerkerk et al., 2010;
Yau, Potenza, Mayes, & Crowley, 2015). At T1, adolescents with more anhedonia may be using
the Internet more compulsively just to reach similar levels of pleasure as reported by typical
adolescents. As adolescents age at T2, more anhedonic individuals may become more selective
in their Internet use than adolescents with lower anhedonia.
While the other two E-PES factors have some dependency on interaction with others
(e.g., social interaction requires others to respond to posts, game playing is often played with
others), Entertainment Surveillance activities can be carried out on one’s own, which may make
gratifications more instantaneous. Given that activities in the Entertainment Surveillance factor
can be considered extensions of previous media, such as television and radio, it is logical to
conclude that the relationship between pleasure derived from “Entertainment Surveillance” and
CIU is stronger among adolescents who are more anhedonic as television shows or YouTube
70
clips, which are easily accessible and abundant online, can provide excitement and escape from
negative moods (Potts & Sanchez, 1994). Indeed, a longitudinal study of Danish adolescents
found that television viewing, but not computer use, was associated with depressive symptoms
(Grøntved et al., 2015). Future studies should continue to examine differences in selected
Internet use, pleasure from online use, and CIU among adolescents with higher anhedonia
compared to adolescents with lower anhedonia.
Results of the longitudinal analysis failed to support hypothesis #3 and #4 that E-PES
factors would be significantly associated with CIU and that these relationships would be
moderated by anhedonia and gender. Baseline CIU, or CIU at T1, was one of the only three
significant predictors (i.e., ethnicity) of CIU at T2. The highly significant explanatory power of
baseline CIU predicting CIU one-year later may also negate the significant effects of the other
variables in model. However, this finding is aligned with a previous study examining CIU among
Cantonese adolescents that found baseline CIU was the only predictor of CIU one year later
despite significant demographic variables at baseline (i.e., age, gender, family economic status,
and immigration) (Shek & Yu, 2012). The significance of baseline behaviors as indicators of
future behaviors are also common findings among studies examining drug addiction (Sussman,
Dent, & Leu, 2000; Swift, Hall, & Copeland, 2000).
The only other significant variables in the longitudinal model other than baseline CIU
was identifying as Hispanic or “Other” ethnicity, which was predictive of lower CIU scores one-
year later. A protective effect on CIU for Hispanic and “Other” ethnicity adolescents may be
attributed to ethnic identity – sense of identity and community with a racial or ethnic group
(Rivas‐Drake et al., 2014). Ethnic identity has been associated with higher self-esteem, lower
depression, and mitigated the negative impact of racial discrimination online (Rivas‐Drake et al.,
71
2014). Involvement with ethnic minority advocacy and cultural organizations may also provide
additional opportunities to socialize offline that may not be as available to adolescents belonging
to broader ethnic populations, (i.e., White and Asian). Indeed, ethnic identity has been found to
be protective against various risky behaviors in adolescence, including substance use (Zapolski,
Fisher, Banks, Hensel, & Barnes-Najor, 2017) and risky sexual behaviors (Jeltova, Fish, &
Revenson, 2005) – though more research is needed to understand the association between
ethnicity, ethnic identity, and CIU particularly among American youth.
Differences between the significance of E-PES factors cross-sectionally and
longitudinally may be explained by duration of engagement in the Internet activities and a non-
significant relationship between anhedonia, pleasure, and CIU in the longitudinal model supports
the notion of CIU as a relatively stable behavior. Pleasure from entertainment surveillance may
be associated with CIU cross-sectionally when one first engages in specific activities or aspects
of the activity (e.g., starts a new show). However, as the novelty of the Internet activity
diminishes with continued use, the effect of pleasure on CIU may be attenuated over time.
Furthermore, without intervention efforts to address anhedonia among adolescents, CIU will
persist once CIU has been established, although anhedonia’s roles in the deterioration or
improvement of CIU behaviors remain to be examined as adolescents enter emerging adulthood.
Over time, other factors may predict CIU better than pleasure obtained from engaging in
certain Internet activities. Factors like the need to stay relevant or maintain online presence
(Christakis & Moreno, 2009), and “fear of missing out (FOMO)” (Elhai, Levine, Dvorak, &
Hall, 2016) may cause one to continue to engage in online activities even when the gratifications
obtained from engaging in the activities may have diminished. Further examination of
72
interaction between these factors and effects of these on pleasure received can help provide a
more comprehensive picture of pleasure and CIU.
Limitations
This study has several limitations that should be considered. First, the use of the E-PES
scale has never been validated elsewhere. The use of the scale in this present study may not fully
capture the range of pleasure received from engaging with electronic media. Specifically, a 3-
point Likert response for pleasure may not be distinct enough to capture differences in pleasure
and indicating use only in the last 7 days may be too narrow a time frame to capture use of
certain Internet activities (e.g., gaming, online shopping) that occur less frequently but may be
highly pleasurable and addictive. However, high Cronbach’s alpha of the E-PES scale indicated
high internal consistency.
Second, the E-PES is a self-reported, subjective measure that is hard to validate and
measure. Future studies should replicate the use of the E-PES scale among ethnically diverse
populations and age groups and consider utilizing more nuanced answer choices to help
distinguish differences between frequency of use and pleasure from use cross-sectionally and
over extended periods of time. Implementing event-level research analysis and qualitative
methods in conjunction with administering the E-PES scale can further validate and improve the
E-PES. In addition, these methods can help inform researchers about the accuracy of self-
reporting when using this scale as well as provide additional information regarding when
pleasure leads to consequences associated with Internet use and if/when pleasure ends but
consequences with Internet persist. Results from these types of analyses may be telling of the
effectiveness of E-PES in capturing real time pleasure associated with engaging in the Internet
73
activity. With a few additional modifications, the E-PES may become a reliable scale used to
measure electronic use pleasure and the fulfillment of gratifications among adolescents.
Lastly, generalizability of the study is limited. Results from this study may only be
applicable to adolescents of southern California and may not be applicable to adolescents from
other states or non-American populations. This study was also unable to use the full sample of
students that participated in this study as many students did not answer the E-PES or CIU
questions at both time points perhaps due to study fatigue and study design in surveying students
who were absent the day of data collection. Future studies may consider examining just two of
these factors or combining activities that may seem similar to decrease missing data. Replication
of this study among other populations is also needed, especially with the use of the pleasure scale
developed here.
CONCLUSION
These findings have important implications for CIU prevention and intervention. In
particular, the need to prevent CIU is most important as past CIU behaviors seem to predict
future CIU behaviors. Prevention and intervention programs should discuss the role of
psychosocial factors, namely anhedonia, in relation to CIU and restructure social norms
associated with Internet use. Demographic differences (i.e., race and gender) of CIU suggest that
programs may want to tailor curriculum to the needs of their specific audience. For adolescents
in general, ethnic identity may be a deterrent for CIU; programs need to identity novel ways of
increasing ethnic identity among adolescent populations offline, particularly among larger ethnic
groups that may lack ethnic solidarity. Finally, for adolescents, pleasures obtained from Internet
activities may not be the driving force of CIU and changes in CIU; thus, prevention and
intervention programs should address other factors that may diminish Internet use pleasure but
74
contribute to increasing CIU related problems, including the need to maintain an online presence
and FOMO.
In sum, the results of this study contribute to our overall understanding of the intrinsic
motivations that may instigate CIU among adolescents and the moderating roles of anhedonia
and gender in this relationship. GO through the engagement of Internet activities have been
understudied in the past and have been considered insignificantly associated with CIU (LaRose,
2011). Findings from this study, however, demonstrate that GO and CIU are perhaps associated
with the onset of but not changes in CIU and highlights the importance of CIU prevention as
early as possible. Future studies should continue to investigate the relationship between GO and
CIU in adolescents, particularly how this relationship may be impacted by other factors such as
Internet presence and FOMO and examine the effectiveness of the E-PES scale as a systematic
way of assessing GO among adolescents.
75
CHAPTER 4: Study 3: Examining the Association between CIU and Alcohol, Tobacco,
Marijuana, and Other Drug Use Using a Latent Transition Analysis
BACKGROUND
Adolescence (between the ages of 13 and 18) is a critical period of development that may
lead to behavioral problems during emerging adulthood and adulthood. Adolescence is
characterized by physical development, cognitive changes, psychosocial changes, and identity
development (Sadler, 2011). Cognitive changes and psychosocial changes, in particular, may
make some adolescents more vulnerable to risk-taking behaviors. Cognitive functioning, which
helps individuals consolidate their thoughts and actions with their goals, intentions, and values,
may be limited as the adolescent brain is still developing (Steinbeis & Crone, 2016). That is, the
maturation of the prefrontal cortex (PFC) which controls decision-making matures later than
other structures of the brain that are related to motivational and affective processes (Leshem,
2016). Stimulus-driven behavior may take precedence over executive cognitive decision-making
and the ability to control impulsivity and exercise rational judgement of behavior is lacking
(Leshem, 2016). Thus, risk-taking, sensation seeking, experimentation and poor decision-making
are often heightened during adolescence (Pokhrel et al., 2013).
Psychosocial changes coupled with limited cognitive functioning make youth particularly
vulnerable to high risk behaviors. During this period, youth become more autonomous and peers
play a more central role in their lives; behaviors become dictated by the adolescents’ social
context and/or perceived social norms (Steinbeis & Crone, 2016). Peer interaction is often
considered a positive developmental process that helps individuals cope with stress and
difficulties. However, adolescents who interact with youth that engage in risky behaviors are
highly susceptible to engaging in similarly risky behaviors, perhaps as a result of wanting to fit
76
in with their peers (Sanders, Munford, Liebenberg, & Ungar, 2017; Valente et al., 2007).
Delinquent youth, for example, tend to be friends with other delinquent youth (Valente et al.,
2007).
Given the neurocognitive changes and psychosocial changes that occur during
adolescence, several problematic behaviors can emerge during this period, including alcohol,
tobacco, and other drug use (ATOD). ATOD behaviors are typically initiated during adolescence
and escalate as youth enter emerging adulthood (between the ages of 18-25) (Patrick &
Schulenberg, 2014). Positive beliefs and attitudes toward ATOD also increase from 8
th
through
12
th
grade (Johnston, O’Malley, Bachman, & Schulenberg, 2013). ATOD experimentation, use,
and dependence at an earlier age increase the likelihood of ATOD dependence in adulthood,
particularly if initiation starts when the adolescent is still in their early teens (Chen et al., 2009;
King & Chassin, 2007; Lopez-Quintero et al., 2011). In addition, other adolescent risk-taking
behaviors, including sexual risk taking, delinquent behavior, aggression, anti-social behavior,
and school drop-out, have been associated with drug and alcohol use among adolescents (NIDA,
2018), which also have long term health and psychosocial consequences.
While efforts to reduce adolescence ATOD initiation, use, and dependence remain a
significant public health priority (NIDA, 2018), compulsive Internet use (CIU) has also become
increasingly concerning among youth. More than 98% of adolescents in American reportedly use
the Internet (Fallows, 2004), with 4% of American adolescents Internet addicted (Liu et al.,
2011). Worldwide, adolescent CIU prevalence rates are reportedly as high as 20% and higher
(Ko et al., 2009; Tran et al., 2017). Similar to many of the symptoms and consequences
associated with ATOD, preoccupation with the Internet has been linked with academic failure
(Yang & Tung, 2007), aggression (Ko et al., 2009), sexual risk-taking (Whiteley et al., 2012),
77
delinquency (Holtz & Appel, 2011), and family conflict (Yang & Tung, 2007). Several studies
have also highlighted the positive associations between CIU and ATOD behaviors (Anderson et
al., 2016). Among a sample of Taiwanese youths who have never smoked or used alcohol, using
the Internet at cybercafés increased the likelihood of smoking and alcohol use four years later,
particularly for females (Chiao et al., 2014). Fisoun, Floros, Siomos, Geroukalis and Navridis
(2012) found that among Greek youth, higher Internet use was associated with higher odds of
having used illicit substances. Among youths attending high school in Connecticut, lifetime
alcohol use, marijuana use, and cigarette were associated with problematic Internet use (Liu et
al., 2011).
Evidence from functional magnetic resonance imaging (fMRI) and positron emission
tomography (PET) studies further support the association between problematic Internet use and
substance use. They indicate similar neurobiological triggers occur among substance abusers also
occur among Internet addicts (Kuss & Griffiths, 2012a). Ko et al. (2013) compared fMRI
between subjects who were diagnosed as nicotine and Internet gaming disorder addicts (who also
fit the criteria for Internet addiction) and healthy subjects. Both addictive and healthy subjects
were shown images related to computer games, cigarette smoking, and other neutral stimuli. The
authors found that cue-induced computer gaming urges and smoking urges activated similar
patterns in the brain, namely the parahippocampus and the anterior cingulate. They conclude that
both Internet gaming and nicotine addiction may share similar mechanisms of activating parts of
the brain. Similarly, Kim et al. (2011) used PET to examine dopamine D2 receptor binding
potential among men with and without IA. The authors found that individuals with CIU had
lower levels of dopamine D2 receptors, an abnormality of the brain that is commonly observed in
78
individuals who are addicted to cocaine, marijuana, and alcohol (Volkow, Fowler, Wang, Baler,
& Telang, 2009).
As posited in the uses and gratifications theory (UGT), media consumption can lead to
negative, often unintended, consequences, especially as self-regulation to control the behavior
diminishes (LaRose, 2011). That is, gratifications sought from Internet use and expected
outcomes associated with Internet use may lead to habitual Internet use. As Internet habits
become nonconscious and as Internet use begins to alter dysphoric moods, CIU becomes
increasingly probable (LaRose, 2011). Internet addiction may eventually lead to other negative
consequences such as academic failure (Caplan, 2002) or drug use as the need to alter dysphoric
moods becomes more imminent and Internet use is no longer gratifying. Rücker, Akre,
Berchtold, & Suris (2015) found that problematic Internet use is associated with substance use
and purported that because CIU shares common characteristics as other risky behaviors, CIU
could act as a “gateway” behavior to other more problematic behaviors similar to the role
tobacco plays as a "gateway" drug to more illicit drug use. Indeed, aligned with UGT, problem
behavior theory suggests that behaviors often cluster, and that adolescents who exhibit one
problem behavior often exhibit other problematic behaviors as well (Jessor & Jessor, 1977).
Several studies have demonstrated that an association between CIU and ATOD exists,
but methodological limitations have hampered CIU and ATOD research. The biggest limitation
in CIU and ATOD research is the limited number of longitudinal studies. To our knowledge,
only two studies have looked at the long term consequences of the interaction between CIU and
ATOD. Both studies found that among adolescents, problematic Internet use at baseline was
predictive of alcohol use at follow-up (Gámez-Guadix et al., 2015; Sun et al., 2012). Another
limitation of these studies is the inability to assess multiple problematic behaviors
79
simultaneously. That is, studies often examine each ATOD behavior independently and cannot
fully assess clusters of problematic behaviors or how these clustered problematic behavior
transition over time.
To address the methodological gaps presented in the literature, the purpose of the third
study is to examine the longitudinal relationship between ATOD addiction and CIU using a
latent transition analysis. More specifically, we wanted to examine if CIU clusters with other
ATOD addiction behaviors at baseline and how identified behavior statuses at baseline transition
at one year follow-up. A better understanding of the ATOD behaviors associated with CIU and
how they can transition over time can help inform prevention efforts for both ATOD and CIU.
Latent transition analysis (LTA) is a longitudinal approach to latent class analysis (LCA).
It allows for researchers to examine how membership in identified latent classes, or statuses,
change over time. See Figure 6.
Figure 6. Conceptual Model of Latent Transition Analysis
80
METHODS
Research Questions:
The proposed research study addressed the following research questions:
Question 5: Do distinct groups of ATOD and CIU behavior exist in this sample?
Hypothesis 5: Distinct subgroups of homogenous behavior will be identified in
our sample. At baseline, we will find a subgroup of individuals that exhibit CIU
independent of drug behavior and other subgroups that cluster based on drug
behaviors.
Question 6: Will the prevalence of identified latent statuses transition across time
or will they remain stable?
Hypothesis 6: Identified latent statuses at Time 1 will shift over time. Individuals
identified in a CIU latent status group at baseline will transition into latent status
groups of ATOD behaviors one year later.
Data Collection
Data for study 3 comes from the Happiness and Health (H&H) study (Leventhal et al.,
2015), a longitudinal survey of substance use and mental health among high school students
enrolled in ten public schools surrounding Los Angeles, CA, USA. Students were eligible to
participate in the survey if they were not enrolled in special education or English as a second
language programs and provided both written or verbal assent and parental consent. Every six
months beginning in 9
th
grade (Fall 2013) through 12
th
grade (Spring 2018), participants were
administered pencil and paper surveys across two 60-minute periods less than two weeks apart in
the students’ classrooms onsite. Students absent during data collection were given surveys via
the phone or Internet. Of the 4,100 students eligible to participate in the program, 3,383
participants provided written or verbal assent and parental consent and completed the survey. All
81
study procedures were reviewed and approved by the University of Southern California
Institutional Review Board.
MEASURES
Compulsive Internet Use Scale (CIUS). CIU was measured using the 14-item Compulsive
Internet Use scale (Meerkerk et al., 2009). Participants were asked, “How often… 1) …do you
find it difficult to stop using the Internet when you are online?; 2) …do you continue to use the
Internet despite your intention to stop?; 3) …do others (e.g., partner, children, parents) say you
should use the Internet less?; 4) ...do you prefer to use the Internet instead of spending time with
others (e.g., partner, children, parents)?; 5) …are you short of sleep because of the Internet?; 6)
…do you think about the Internet, even when not online?; 7)… do you look forward to your next
Internet session?; 8) … do you think you should use the Internet less often?; 9) … have you
unsuccessfully tried to spend less time on the Internet?; 10) …do you rush through your
homework in order to go on the Internet?; 11) … do you neglect your daily obligations (work,
school, or family life) because you prefer to go on the Internet?; 12) … do you go on the Internet
when you are feeling down?; 13) … do you use the Internet to escape from your sorrows or get
relief from negative feelings?; and 14) … do you feel restless, frustrated, or irritated when you
cannot use the Internet?” Each item was scored on a 5-point Likert scale: Never (0), Seldom (1),
Sometimes (2), Often (3), Very often (4). A weighted sum score across the 14 items was
computed to assess endorsed CIU at both time points. Composite CIUS scores were further
dichotomized, where scores of 21 of higher on the CIUS were recoded as 1 or “Yes” indicating
Internet addiction (IA) (Guertler et al., 2014) and scores less than 21 were recoded as 0 or “No”
indicating regular Internet use. Similar to previous studies (Meerkerk et al., 2009), the CIUS
82
exhibited good internal consistency in this sample at T1 (Cronbach’s alpha = 0.95) and T2
(Cronbach’s alpha = 0.95).
Past 30-day Alcohol and Substance Use Measure. Items used to assess substance use at
Wave 3 and Wave 5 of H&H were taken from the Youth Behavior Risk Surveillance Survey
(YBRS) (Eaton et al., 2008) and Monitoring the Future questionnaire (MTF) (Johnston et al.,
2013). Participants were asked to indicate in the last 30 days, how many total days they used the
following substances: alcohol, cigarettes, e-cigarettes, marijuana, stimulants (e.g., cocaine, crack,
rock, meth, crystal, crank, ice), prescription stimulant pills without a doctor’s advice (e.g.,
Ritalin, Adderall, Dexedrine, uppers), and prescription painkillers without a doctor’s advice (e.g.,
Vicodin, OxyContin, Percocet, and Codeine). Responses were given on a Likert scale ranging
from “0 days”, “1-2 days”, “3-5 days”, “6-9 days”, “10-14 days”, “15-19 days”, “20-24 days”,
“25-29days”, and “all 30 days”. Cigarette and e-cigarette items were collapsed together to form a
“Nicotine” category. Stimulants, prescription stimulant pills without a doctor’s advice, and
prescription painkillers without a doctor’s advice were collapsed to create an “Other drug use”
category. Each substance use item was recoded so that a 1or “Yes” response indicated past 30
day use while a 0 or “No” response indicated no substance use in the past 30 days.
Analytic Plan
Only data from wave 3, referred to as Time 1 (T1), and wave 5, referred to as Time 2
(T2), were assessed in this study as these were the only two time points that measured CIU.
Following guidelines established by Collins and Lanza (2013), a series of analyses was
conducted to determine the final LTA model. Preliminary analysis was first conducted using
latent class analysis (LCA) to inform the latent statuses. Using results from the LCA as a guide,
latent transition analysis (LTA) across T1 and T2 with 2 to 5 latent statuses were compared to
83
determine the final model selection; results from the LTA were used to determine the final model
as it has the ability to detect additional number of classes due to the inclusion of multiple time
points that LCA could not (Collins & Lanza, 2013). Model selection for the LTA was assessed
using fit indices: Akaike information criterion (AIC), Bayesian information criterion (BIC), and
entropy. Lower AIC and BIC values indicated better model fit. Higher entropy values (closer to
1) indicated more distinct class solutions. Unlike LCA, LMR p-values were not used as criterion
to determine model fit as LTA contingency tables tend to be large rendering the G2 statistic and
p-values unreliable (Collins & Lanza, 2013). In addition, class interpretability, parsimony, and
conceptual appeal were assessed to determine the final model solution (Collins & Lanza, 2013;
Nylund, 2007). After the most parsimonious number of latent statuses was selected, a fully
invariant model was compared to a fully non-invariant model to determine whether parameters
should be constrained or freely estimated using model fit indices (i.e., AIC, BIC, and likelihood
ratio test [LRT]).
The LTA solution produced three sets of parameters of interest: item response
probabilities, prevalence of latent status membership at each wave, and transition probabilities
(Collins & Lanza, 2013). Item response probabilities indicated the probabilities of endorsing
behaviors conditional on latent class membership and provided the foundation for how the latent
statuses were interpreted and qualitatively labeled. Prevalence of latent status membership
provided the percentage (and number) of subjects likely to be in that latent status at each wave.
Transition probabilities represented the probabilities of transitioning to other latent statuses at
follow-up (T2) conditional on membership in a particular latent class at baseline (T1).
84
RESULTS
Sample
Of the 3,396 H&H participants measured at baseline, 2,793 participants had data on key
measures at both time points for the current analysis (82%). Average age of the participants was
15.07 (SD=0.41) years old and 46% of the sample was male. Most students reported they were
Hispanic/Latino (47%), followed by Asian (17%), White (15%), and “Other” (e.g.,
Black/African American, Native American Indian, Other, and mixed; 21%). Prevalence of
Internet Addiction was 42.89% at T1 and 44.63% at T2. Last 30-day prevalence of nicotine use,
alcohol use, marijuana use, and other drug use were 10.24%, 19.91%, 10.99%, and 3.69% at T1,
respectively, and 7.37%, 20.13%, 11.75%, and 3.63% at T2, respectively. See Table 9. All
descriptive statistics were conducted in SAS Version 9.4 (SAS Institute, Cary, NC, USA).
Assessment of Missing Data & Attrition
Chi-square analysis compared participants at T1 who were missing complete data and not
analyzed (N= 546) and participants who had complete data on key study measures (N=2,850),
including each of the substance use items, the Internet addiction item, age, and gender.
Significant differences were found regarding last 30 day use of nicotine (chi-square = 6.745,
df=1, p=0.009), alcohol (chi-square=7.141, df=1, p=0.008) and marijuana (chi-square=8.168,
df=1, p=0.004); age (chi-square=84.76, df=1, p<0.0001); and gender (chi-square=5.86, df=1,
p=0.015). Participants who used nicotine, alcohol, or marijuana in the last 30 days who were
younger than 15 years old and male were more likely to have incomplete data at baseline than
participants who did not use nicotine, alcohol, or marijuana, who were 15 years old or older and
female. In addition, attrition analysis compared participants who had full data at baseline and
who were missing on all key study measures one year later and could not be analyzed (N=57).
Similarly, significant differences were found regarding last 30 day use of nicotine (chi-square=
85
9.395, df=1, p=0.002), alcohol (chi-square=7.98, df=1, p=0.005), and marijuana (chi-
square=16.21, df=1, p<0.0001).
Table 9. Prevalence of Internet Addiction and Last 30-Day Alcohol, Nicotine,
Marijuana, and Other Drug Use Among a Sample of Regular High School Youth in
Los Angeles County (N=2,793)
T1 (%) T2 (%)
Internet Addiction
42.89 44.63
Alcohol Use
19.91 20.13
Nicotine (cigarette and e-
cigarette smoking)
10.24 7.37
Marijuana
10.99 11.75
Other Drugs
3.69 3.63
LCA & LTA Results
Preliminary LCA results suggested that a 2-class or 3-class model fit the data at both T1
and T2, respectively (see Table 9). However, subsequent LTA analysis suggested that the 4-
status or 5-status solutions were comparable. Although AIC continues to decrease, BIC increases
and identification of the latent statuses become problematic from the 4 to 5 status model. In
addition, Nylund, Tihomir, and Benght (2007) ran a series of Monte Carlo Simulations and
found that BIC consistently identifies the number of latent classes correctly and was more
reliable than other information criteria. Therefore, the 4-status solution was selected. Model fit
indices comparing a fully invariant and non-invariant 4-status model indicated the fully-invariant
4-status solution was the most parsimonious model: AIC and BIC values were lower for the
86
invariant model (model 2) compared to the non-invariant model (model 1) and freeing
parameters did not improve model fit. See Table 10.
Qualitative labels were applied based on the item-response probabilities. “Non-users”
were unlikely to use any ATOD or be addicted to the Internet at both time points. “Compulsive
Internet Users” were highly likely to be addicted to the Internet but not substance users.
“Compulsive Internet and Alcohol Users” were moderately likely to be Internet addicts but
highly likely to have used alcohol in the last 30 days. “Compulsive Internet and Poly Drug
Users” were moderately likely to be addicted to the Internet but highly likely to have used all
substances in the last 30 days, particularly marijuana and other drugs.
Transition probabilities are represented in Table 10. Values on the diagonal reflect the
stability of the latent statuses (i.e., staying the same across both time points). All other values in
the row represent the likelihood of transitioning from baseline into another latent status at T2.
Transition probabilities showed that the four latent statuses were relatively stable across time.
Approximately 11% of baseline “Non-users” transitioned to “Compulsive Internet Users” by
follow-up, but few progressed to “Compulsive Internet and Alcohol Users” or “Compulsive
Internet and Poly Drug Users”. “Compulsive Internet Users” at baseline were unlikely to
transition to any of the other three latent statuses, but if transitioning did occur it was to either
“Compulsive Internet and Alcohol Users” or “Compulsive Internet and Poly Drug Users”
statuses; none transitioned to “Non-users”. Approximately 35% of “Compulsive Internet and
Alcohol Users” and 32% of “Compulsive Internet and Poly Drug Users” at T1 were likely to
transition to any of the other three classes at T2. See Table 11.
87
Table 10. Model Fit Indices for Latent Transition Analysis Models With 2 to 5 Statuses
and Time Invariant Model of Chosen Latent Status
No. of
Latent
Statuses
DF AIC BIC Entropy LL
Model 1 2 991 18794 18930 0.87 -9374
3 976 18307 18532 0.84 -9116
4 957 18190 18517 0.81 -9040
5 939 18126 18566 0.86 -8989
Model 2 4 976 18184 18391 0.79 -9056
Notes: AIC = Akaike’s information criterion; BIC = Bayesian information criterion; LL = log
likelihood. Model 1: Identification of latent statuses. Model 2: Model tested simultaneously for
T1 and T2 with item-response probabilities constrained to be equal (fully invariant); selected
model is highlighted in bold.
88
Table 11. Item-Response Probabilities and Transition Probabilities for Selected Latent
Transition Analysis Model (N=2,793)
Latent Statuses
Non-users
Compulsive
Internet Users
Compulsive
Internet &
Alcohol Users
Compulsive
Internet & Poly
Drug Users
Proportion of sample constituted by each status
T1 0.358 0.424 0.122 0.095
T2 0.336 0.459 0.105 0.099
Item Response Probabilities
a
Nicotine use in last 30
days
0.008 0.010 0.287 0.498
Alcohol use in last 30
days
0.063 0.040 0.764 0.746
Marijuana use in last
30 days
0.010 0.015 0.106 0.942
Other Drug Use in
last 30 days
0.003 0.011 0.059 0.246
Internet Addiction 0.013 0.773 0.463 0.388
Transitions from baseline (rows) to One-Year follow-up (column) in LTA without covariates; values on
the diagonal reflect cases without transition
Non-users 0.865
b
0.114 0.002 0.018
Compulsive Internet
Users 0.000 0.933
b
0.038 0.028
Compulsive Internet
& Alcohol Users
0.129 0.125 0.652
b
0.095
Compulsive Internet
& Poly Drug Users
0.113 0.110 0.088 0.689
b
Note:
a
Item-response probabilities were fully invariant (constrained to be equal) across Time 1 and Time
2.
b
These values indicate transitional probabilities that represent stability (i.e., staying the same status
across both time points).
89
DISCUSSION
Results from this study suggest that IA is a more prevalent and stable condition among
adolescents than previously thought. While last 30-day prevalence of alcohol, nicotine,
marijuana and other drug use were generally consistent with the 2018 Monitoring the Future
report (Johnston et al., 2018), IA among our sample of adolescents was much higher than that
found in previous studies of American youth (Liu et al., 2011; Moreno et al., 2011a). In addition,
IA was prevalent across three out of the four identified latent statuses. High representation of IA
in these latent statuses further supports the possibility of increasing problematic Internet use
among adolescents. Alternatively, low representation of last 30-day drug use, particularly hard
drugs (e.g., stimulants and marijuana), in these latent statuses compared to IA may also indicate
that accessing nicotine, alcohol, and drugs among younger adolescents is more difficult than
accessing the Internet and may explain why there may be more evidence suggesting problematic
Internet behaviors but not substance use among adolescents in this age group. Future studies
should examine whether differences in accessibility to substances and the Internet and vice versa
play a role in problematic behaviors over time.
Previous research has demonstrated significant associations between substance use
behaviors and IA (e.g., Fisoun et al., 2012; Lee et al., 2013; Rücker et al., 2015). To our
knowledge, this is the first study that used LTA to analyze IA and drug use behaviors, and
similar to previous studies, this study found that IA was associated with substance use in two of
the four latent statuses. Since IA is present in both substance use statuses, it is reasonable to
conclude IA and substance use as comorbid behaviors among a subset of adolescents. Though
this study was unable to examine the specific association between substance use and IA, one may
speculate that the Internet and using the Internet facilitates substance use behaviors. For
90
example, the Internet may play an integral role in exposure to and education of substances and
substance use behaviors (e.g., lingo), in the acquisition of substances (e.g., online prescription
orders through unregulated websites), or in substance use maintenance (e.g., substance users join
communities of other substance users online). The specific role, if any, Internet use may play in
the adoption and/or maintenance of substance use is warranted. Furthermore, the environmental
context in which Internet use takes place may also contribute to substance use. Previous studies
have found positive associations between Internet use and substance use among adolescents who
frequented cybercafés where adolescents with IA who spend more time in cybercafés
subsequently have higher substance use (Chiao et al., 2014; Yang & Tung, 2007). In public
domains, such as cybercafés, drugs may be more easily accessible and substance use behaviors
may be a widely accepted social norm. Researchers should examine the environmental contexts
in which Internet use occurs among American adolescents.
Latent transition probabilities indicated minimal transitions in the “Non-users” and
“Compulsive Internet Users” statuses and more transitions in the “Compulsive Internet and
Alcohol Users” and “Compulsive Internet and Poly Drug Users” statuses. The high stability of
the “Compulsive Internet Users" status precluded any conclusions regarding whether IA
preceded substance use. However, the stability of the “Compulsive Internet Users” status and the
fluctuating “Compulsive Internet and Alcohol Users” and “Compulsive Internet and Poly Drug
Users” statuses may suggest: 1) subjects in this status have different characteristics than those in
the IA and substance use statuses, or 2) IA and substance use may be a normal part of adolescent
development (Galvan, 2010). Previous studies have found similar characteristics that predict
substance use among Internet addicts, such as sensation seeking, impulsiveness, and diminished
self-regulation (Hwang et al., 2014; Ko et al., 2006; Ko et al., 2008). Despite many psychosocial
91
similarities, one study found that adolescents in Taiwan who had IA differed from substance
users on harm avoidance (Ko et al., 2006). Students who scored high on harm avoidance tended
to be Internet addicts while those who had low harm avoidance were likely to exhibit comorbid
IA and substance use. Differences in harm avoidance may explain why transitions from the
“Compulsive Internet Users” status to other statuses with substance use are less likely to occur.
More research distinguishing the psychosocial indicators that drive IA specifically and IA and
substance use are needed. Furthermore, identification of these psychosocial indicators can be
incorporated as covariates in future LTA models.
Among those who were in either substance use statuses at T1, transition out of these
groups one-year later could result as adolescents: 1) grow out of the experimentation phase and
no longer engage in substance use behaviors (Arnett, 2005), 2) progress into harder drug use (if
they belonged in the “Compulsive Internet and Alcohol Users” status at T1 and transitioned into
the “Compulsive Internet and Poly Drug Users” status at T2) (Kandel & Kandel, 2015), and 3)
find substances too difficult to obtain or too risky to use and transition to more accessible and
possibly less risky substances and behaviors (e.g., alcohol and Internet use) (Sussman & Black,
2008; Toumbourou et al., 2007). Those who transition into or remain in the “Compulsive
Internet and Poly Drug Users” status at T2 warrant the most attention as this is seemingly the
status with the riskiest combination of behaviors and may elicit the most consequences in the
future should these behaviors persist. Further investigation of the stage-sequential progression
from non-user to “Compulsive Internet User” to “Compulsive Internet and Poly Drug User” is
warranted (Kandel, 2002). This relationship may become clearer as adolescents age especially as
they enter emerging adult hood where parental monitoring diminishes and accessibility to
92
substances increase (Arnett, 2005). Studies should replicate this study in other populations and
continue to monitor these behaviors throughout adolescence and into adulthood.
These findings have important implications for IA and substance use prevention and
intervention programs. Most IA prevention programs target universal populations; however,
results from this study support the creation and evaluation of selective and indicated IA
prevention programs (Vondráčková & Gabrhelík, 2016). As indicated by the LTA, groups of
adolescents exhibit different types of problematic behavior: IA only or IA and substance use.
Thus, tailored prevention and intervention programs should differentiate between Internet addicts
and comorbid Internet addicts and substance users. For those vulnerable to IA, specifically,
prevention and intervention programs may want to address harm avoidance and consequences
specific to the Internet (e.g., cyberbullying, online sexual solicitation, harassment) (Falender,
2017). Interventions that promote increased parental monitoring of adolescent Internet use and
provide parents the technical skills that make them as computer proficient as their children may
be needed.
On the other hand, more complex prevention programs that target several risk behaviors
may be more effective for those exhibiting IA and substance use. Increased parental bonding and
addressing beliefs regarding peer Internet and substance use, for example, may be needed to
deter IA and substance use. In addition, addressing the particular role the Internet may play in
facilitating substance use and vice versa may be critical components to successful IA and
substance use prevention and intervention programs.
LIMITATIONS
Several limitations should be addressed. As with most LCA and LTA studies,
generalizability of the study is limited as these methods are data driven. Prevalence of Internet
93
addiction at T1 and T2 in this study is generally higher than that of most studies, especially
studies conducted in the U.S. (Moreno et al, 2011a). However, some studies conducted among
Asian and European populations found similar IA prevalence rates: 37.9% in Hong Kong
(Leung, 2004), 37% in China (Xin et al., 2018), 42% in India (Krishnamurthy & Chetlapalli,
2015), and 36.7% in Italy (Milani, Osualdella, & Di Blasio, 2009). Although a conservative cut-
off point from the only study that examined the respective clinical diagnosis of IA using the
CIUS among a sample of problematic gamblers (Guertler et al., 2014) was used to determine IA
in our sample, validation of the clinical diagnosis of IA using the CIUS among regular
adolescent populations is needed. Additional research examining the clinical cut-off for IA using
the CIUS scale is needed before any definitive conclusions regarding increasing IA trends among
adolescents in the United States can be drawn. In addition, past 30-day substance use, was low
among this sample of students. Identification of latent statuses may differ among adolescent
populations who have lower prevalence of IA and higher last 30-day prevalence substance use.
All measures were self-reported and thus subject to response bias. Problems with Internet
use may be considered more socially acceptable and thus, more willingly reported. Substance
use, however, may be seen as a delinquent behavior and adolescents in this sample may be
reluctant to answer truthfully.
The age of our sample was a little young. Accessibility to substances may be limited
among younger adolescents as they are still residing at home and remain under the supervision of
their parents. As adolescents progress into emerging adulthood, substance use and IA behaviors
may increase and latent statuses may change. Greater transitioning may occur between latent
statuses examined at different time points (e.g., between the ages of 18 to 19 years old). Thus,
the results of this study may be limited to adolescents in this age range.
94
Finally, a portion of the participants in this study did not respond to all the CIUS and
ATOD questions perhaps as a result of study fatigue and the study design related to data
collection among students who were absent the day surveys were administered, which excluded
them from analysis. Several students were also lost to follow-up between T1 and T2. Missing
data and attrition analysis found that those who were more likely to use certain substances (i.e.,
nicotine, alcohol, and marijuana) were more likely to have been excluded from the original
analysis or excluded from follow-up. Replication of this study among diverse populations is
needed.
CONCLUSION
This study is a novel approach to examining IA and substance use and provides
additional insight into the association between these behaviors. While prior research has shown
that associations between substance use and Internet addiction exist, the current study identifies
four latent statuses that demonstrate Internet addiction as an independent risk behavior as well as
a comorbid behavior associated with varying levels of substance use. Continued research to
determine the prevalence of these latent statuses and if and when these transitions occur among
other population are needed as adolescents become more reliant on the Internet. Innovative and
tailored IA and substance use prevention and intervention programs will be needed to address the
needs for different subsets of the population who exhibit various risk behaviors.
95
CHAPTER 5: Overall Discussion and Conclusions
The research conducted in this dissertation contributes to the overall understanding of
CIU and its associated consequences, particularly among samples of ethnically diverse American
youth. Following the framework of UGT, specific attention was paid to the types of activities
that contributed to CIU, the pleasure received from specific Internet activities and its relationship
with CIU, and the association between CIU and other problematic behaviors during adolescence
– namely, substance use. Demographic variables that found mixed results in previous studies and
psychosocial variables (i.e., trait anhedonia) that were understudied in the literature were
examined to provide additional insight on their associations with CIU.
Summary of Findings
Study 1 found that adolescents who had higher CIU scores were more likely to belong in
the “Social”, “Entertainment Surveillance”, and “Dependent” classes than the “Functioning”
class, with those in the “Social” and “Dependent” classes having a higher association with
certain CIU consequences and general CIU. In addition, Study 1 identified distinct groups of
Internet users that demonstrate the existence of both generalized (“Dependent”) and specific PIU
(“Social” and “Entertainment and Surveillance”) – both of which were associated with CIU
consequences. Study 2 found that certain gratifications were associated with CIU onset but not
with changes in CIU and that anhedonia and gender moderated pleasure and CIU relationships
cross-sectionally but not longitudinally. Baseline CIU and identifying as “Hispanic” and “Other”
ethnicity were the only variables that significantly predicted changes in CIU one year later.
Study 3 found CIU as an independent consequence as well as a consequence that is associated
with substance use behaviors among subsets of adolescents. In sum, these findings contribute to
our understanding of factors that may be associated with CIU etiology and consequence and have
96
important implications for theory, future research methods, and CIU prevention and intervention
programs.
Theoretical Implications
Key findings from these studies support the use of UGT as a theoretical framework in
examining CIU among non-clinical populations. Aligned with UGT theory, adolescents
selectively choose to engage in specific online activities that fulfill gratifications and lead to
consequences associated with use. In particular, Study 2 specifically addresses one of the
limitations of UGT and provides critical information regarding how pleasure is related to the
development of CIU. Furthermore, these studies suggest that modifications to UGT suggested by
LaRose (2011), namely the addition of self-efficacy and deficient self-regulation, may be
important indicators associated with CIU that should be integrated into future research. Study 1,
for example, found equivocal results that may be explained by increased self-efficacy and
decreased self-regulation.
CIU researchers have debated whether CIU, or IA, are appropriate terms to describe
problems associated with online activities, leading some to suggest the elimination of said terms
in favor of more activity specific addiction (e.g., Facebook addiction, social networking
addiction, MMORPG addiction) (Griffiths, 2000; Van Rooij & Prause, 2014). However, a
significant finding from these studies suggests the existence of CIU into generalized and specific
PIU (Davis, 2001). Specifically, the identification of a “Dependent” class demonstrates use of
multiple Internet activities without an overarching purpose and/or theme. Thus, arguments which
state the persistence of these addictions offline (Van Rooij et al., 2017) may not be completely
valid since the platform of the Internet as an entity and the process gratifications it has to offer
seem to lend itself to problematic Internet behaviors and the diagnosis of Internet addiction.
97
Methodological Implications
Given the complexity of CIU, more robust measures and methods are needed to fully
comprehend the nuances that contribute to CIU. First, methods used in these studies should be
replicated. Since adolescents rarely engage in only one Internet activity, methods like LCA and
LTA are important ways to analyze patterns of Internet use that lead to CIU. Longitudinal
analysis of CIU and its predictors remain important to the understanding of how CIU changes
over time. Studies should investigate CIU across the lifespan particularly when youth are
beginning to use the Internet through emerging adulthood when evidence of CIU is most
prominent (Moreno et al., 2011a). In conjunction with more robust quantitative methods, using
qualitative and event-level methods will provide information regarding contextual factors
attributed to the development of CIU, which have been understudied in CIU literature (Anderson
et al., 2016), and will provide information regarding whether CIU can propagate psychosocial
dsyfunctioning among normal and susceptible adolescents (e.g., ADHD, anhedonia).
The measures used to assess CIU and correlated variables need to be reevaluated. Items
asked in the CIUS may not accurately gauge problematic use. The clinical cut-off for IA
(Guertler et al., 2014) using the CIUS needs to be assessed among non-clinical, adolescent
populations. Validation of the E-PES scale including its psychometric properties and efficiency
in assessing actual pleasure among other populations is warranted.
Programmatic Implications
Results from this dissertation have significant implications for future CIU screening and
for CIU prevention and intervention programs and reinforce the importance of CIU prevention as
early as possible (Shek & Yu, 2012). Screening adolescents by the types of activities performed
online can help determine CIU risk. Furthermore, screening should be conducted to distinguish
98
between adolescents with generalized or specific PIU as these have different implications for
intervention. This may involve the development of more refined measurements to assess these
differences, which seem to be lacking in the current literature.
More tailored CIU prevention and intervention programs should be created for selective
rather than universal audiences. Adolescents who are at risk for generalized PIU need tools that
help them cut-down on general Internet use. Limiting access to the Internet, such as
implementing tools that shut down the Internet or slow the speed of Internet connection after
certain periods of time, may be one of the most important prevention and intervention tools for
those with generalized PIU. Innovative methods (e.g., mindfulness) to prevent and combat CIU
should be considered. Role of parents and other authority figures as enforcers to limit duration of
use may be effective treatment methods, which will require additional Internet training. Policies
regulating the industry (e.g., age restrictions on use, periodic warning messages) may be
implemented as a means of combatting CIU and shifting some of the responsibilities for CIU
prevention to those who are instrumental in causing it.
Different prevention and intervention curriculums should be used for adolescents who are
engaging in specific types of online activities or who are at risk for specific PIU. Depending on
the types of activity, prevention and intervention programs should emphasize the respective risks
associated with specific Internet activities, such as online sexual solicitation, harassment, and
cyberbullying (Falender, 2017). Since specific PIU has been linked to offline problematic
behaviors (Van Rooij & Prause, 2014), addressing the underlying reasons for engagement in
those specific activities will be instrumental in prevention and treatment. Cognitive behavior
therapy may be important in tackling ways to reduce compulsive engagement in the activity both
online and offline.
99
Efforts should be made to separate CIU and ATOD prevention and intervention
programs. While study 3 did find an association between CIU and ATOD behaviors among two
subgroups of the sample, an additional latent class of just Internet addicts suggests differences
between Internet addicts and ATOD users who also are addicted to the Internet. Alternatively,
these distinct classes could also suggest differences in ease of access specific to each behavior.
As such, CIU programs need to address the differing characteristics that may drive CIU and
discuss CIU risk among very young populations since Internet use is easily accessible by youth
and integrated as common practice early on in life. Conversely, ATOD programs may want to
target an older group of adolescents and focus efforts on discussing the dynamic role of the
Internet in ATOD initiation and use. Furthermore, additional resources may be allocated to
populations that demonstrate IA and poly drug use as they are most at risk for diminished quality
of life.
Overall Limitations
Although limitations to each specific study have been addressed, there are broader
limitations that are pertinent to all three studies. First, this dissertation only examines a narrow
perspective of CIU and was unable to comprehend the full environmental context that may
contribute to the onset and persistence of CIU. Second, Internet activities change rapidly and so
the activities listed may no longer be relevant. Results currently presented may only be relevant
for a short duration of time until other innovative applications that have the ability to satisfy
several gratifications are invented. Third, distinctions between Internet use on computers versus
Smartphones and tablet devices were not made. This is important as more and more people are
using alternative devices to connect with the Internet and the implications for prevention and
treatment may differ based on the mode of access. In addition, applications available on
100
Smartphones may function differently from applications on the computer and having an easily
accessible Smartphone that connects to the Internet may cause more problematic behaviors than
being tied to an immobile computer.
OVERALL CONCLUSION
This dissertation sought to address gaps in CIU and UGT literature. Undoubtedly, the
Internet allows users to engage in various activities that were previously limited by geography
and time. Immediate process, content and social gratifications can be more easily sought and
obtained through the Internet. However, as demonstrated by the literature and results from these
studies, CIU is much more prevalent than previously thought, particularly among adolescents,
and seems to be increasing exponentially. CIU is not easily explained by pleasure obtained
through use and it poses risks as an independent platform and in conjunction with other risk
behaviors. Continued research is necessary in examining the factors that predict CIU, the
psychosocial correlates associated with CIU, and whether CIU actually causes harm and impairs
normal functioning, especially as society’s dependency on the Internet grows. Only with
increased understanding of the entire CIU career trajectory – from etiology to lifetime
consequence – can the formation of policies regulating Internet industry and standards of CIU
prevention and treatment be implemented.
101
LITERATURE CITED
Alhabash, S., Chiang, Y. H., & Huang, K. (2014). MAM & U&G in Taiwan: Differences in the
uses and gratifications of Facebook as a function of motivational reactivity. Computers in
Human Behavior, 35, 423-430.
Anderson, E. L., Steen, E., & Stavropoulos, V. (2017). Internet use and problematic Internet use:
A systematic review of longitudinal research trends in adolescence and emergent
adulthood. International Journal of Adolescence and Youth, 22(4), 430-454.
Andreassen, C. S., Torsheim, T., Brunborg, G. S., & Pallesen, S. (2012). Development of a
Facebook addiction scale. Psychological reports, 110(2), 501-517.
Arnett, J. J. (2005). The developmental context of substance use in emerging adulthood. Journal
of Drug Issues, 35(2), 235-254.
Ball-Rokeach, S. J. (1998). A theory of media power and a theory of media use: Different
stories, questions, and ways of thinking. Mass Communication and Society, 1(1-2), 5-40.
Bandura, A. (1982). Self-efficacy mechanism in human agency. American Psychologist, 37(2),
122.
Bernardi, S., & Pallanti, S. (2009). Internet addiction: a descriptive clinical study focusing on
comorbidities and dissociative symptoms. Comprehensive Psychiatry, 50(6), 510-516.
Bimber, B. (2000). Measuring the gender gap on the Internet. Social science quarterly, 868-876.
Blumler, J. G., & Katz, E. (1974). The uses of mass communications: Current perspectives on
gratifications research (Vol. 1974). Beverly Hills, CA: Sage Publications, Inc.
Brand, M., Young, K. S., Laier, C., Wölfling, K., & Potenza, M. N. (2016). Integrating
psychological and neurobiological considerations regarding the development and
maintenance of specific Internet-use disorders: An Interaction of Person-Affect-
102
Cognition-Execution (I-PACE) model. Neuroscience & Biobehavioral Reviews, 71, 252-
266.
Cao, H., Sun, Y., Wan, Y., Hao, J., & Tao, F. (2011). Problematic Internet use in Chinese
adolescents and its relation to psychosomatic symptoms and life satisfaction. BMC Public
Health, 11(1), 802.
Caplan, S. E. (2002). Problematic Internet use and psychosocial well-being: development of a
theory-based cognitive–behavioral measurement instrument. Computers in Human
Behavior, 18(5), 553-575.
Caplan, S. E. (2010). Theory and measurement of generalized problematic Internet use: A two-
step approach. Computers in Human Behavior, 26(5), 1089-1097.
Caplan, S. E., Williams, D., & Yee, N. (2009). Problematic Internet use and psychosocial well-
being among MMO players. Computers in Human Behavior, 25(6), 1312-1319.
Carli, V., Durkee, T., Wasserman, D., Hadlaczky, G., Despalins, R., Kramarz, E., . . . Brunner,
R. (2012). The association between pathological internet use and comorbid
psychopathology: a systematic review. Psychopathology, 46(1), 1-13.
Cash, H., Rae, C.D., Steel, A.H., & Winkler, A. (2012). Internet addiction: A brief summary of
research and practice. Current Psychiatry Reviews, 8(4), 292-298.
Charney, T., & Greenberg, B. S. (2002). Uses and gratifications of the Internet. Communication
technology and society: Audience adoption and uses, 379-407.
Chen, C.Y., Storr, C. L., & Anthony, J. C. (2009). Early-onset drug use and risk for drug
dependence problems. Addictive Behaviors, 34(3), 319-322.
103
Chiao, C., Yi, C.-C., & Ksobiech, K. (2014). Adolescent Internet use and its relationship to
cigarette smoking and alcohol use: A prospective cohort study. Addictive Behaviors,
39(1), 7-12.
Christakis, D. A., & Moreno, M. A. (2009). Trapped in the net: will internet addiction become a
21st-century epidemic? Archives of Pediatrics & Adolescent Medicine, 163(10), 959-960.
Collins, L. M., & Lanza, S. T. (2013). Latent class and latent transition analysis: With
applications in the social, behavioral, and health sciences (Vol. 718). Hoboken, New
Jersey: John Wiley & Sons, Inc.
Correa, T., Straubhaar, J. D., Chen, W., & Spence, J. (2015). Brokering new technologies: The
role of children in their parents’ usage of the internet. New Media & Society, 17(4), 483-
500.
Courtois, C., Mechant, P., De Marez, L., & Verleye, G. (2009). Gratifications and seeding
behavior of online adolescents. Journal of Computer‐Mediated Communication, 15(1),
109-137.
Davis, R. A. (2001). A cognitive-behavioral model of pathological Internet use. Computers in
Human Behavior, 17(2), 187-195.
Davis, R. A., Flett, G. L., & Besser, A. (2002). Validation of a new scale for measuring
problematic Internet use: Implications for pre-employment screening. CyberPsychology
& Behavior, 5(4), 331-345.
De Leo, J. A., & Wulfert, E. (2013). Problematic Internet use and other risky behaviors in
college students: An application of problem-behavior theory. Psychology of Addictive
Behaviors, 27(1), 133-141.
104
Diddi, A., & LaRose, R. (2006). Getting hooked on news: Uses and gratifications and the
formation of news habits among college students in an Internet environment. Journal of
Broadcasting & Electronic Media, 50(2), 193-210.
Dong, G., Lu, Q., Zhou, H., & Zhao, X. (2011). Precursor or sequela: pathological disorders in
people with Internet addiction disorder. PloS One, 6(2), e14703.
Durndell, A., & Haag, Z. (2002). Computer self efficacy, computer anxiety, attitudes towards the
Internet and reported experience with the Internet, by gender, in an East European
sample. Computers in Human Behavior, 18(5), 521-535.
Eaton, D. K., Kann, L., Kinchen, S., Shanklin, S., Ross, J., Hawkins, J., . . . Chyen, D. (2008).
Youth risk behavior surveillance--United States, 2007. Morbidity and Mortality Weekly
Report. Surveillance Summaries (Washington, DC: 2002), 57(4), 1-131.
Elhai, J. D., Levine, J. C., Dvorak, R. D., & Hall, B. J. (2016). Fear of missing out, need for
touch, anxiety and depression are related to problematic smartphone use. Computers in
Human Behavior, 63, 509-516.
Everitt, B. J. (2014). Neural and psychological mechanisms underlying compulsive drug seeking
habits and drug memories–indications for novel treatments of addiction. European
Journal of Neuroscience, 40(1), 2163-2182.
Evren, C., Dalbudak, E., Evren, B., & Ciftci Demirci, A. (2014). High risk of internet addiction
and its relationship with lifetime substance use, psychological and behavioral problems
among 10th grade adolescents. Psychiatria Danubina, 26(4), 0-339.
Falender, J. (2017). Bernadette H. Schell: Online Health and Safety: From Cyberbullying to
Internet Addiction.
105
Fallows, D. (2004). The Internet and daily life: Many Americans use the Internet in everyday
activities, but traditional offline habits still dominate: Pew Internet & American Life
Project.
Fallows, D. (2005). How women and men use the Internet. Pew Internet & American Life
Project, 28, 1-45.
Ferguson, M. L., & Katkin, E. S. (1996). Visceral perception, anhedonia, and emotion.
Biological Psychology, 42(1-2), 131-145.
Fisoun, V., Floros, G., Siomos, K., Geroukalis, D., & Navridis, K. (2012). Internet addiction as
an important predictor in early detection of adolescent drug use experience—implications
for research and practice. Journal of Addiction Medicine, 6(1), 77-84.
Flisher, C. (2010). Getting plugged in: an overview of internet addiction. Journal of Paediatrics
and Child Health, 46(10), 557-559.
Foran, C. (2015, November 5). The Rise of the Internet Addiction Industry. The Atlantic.
Retrieved from: https://www.theatlantic.com/technology/archive/2015/11/the-rise-of-the-
internet-addiction-industry/414031/.
Frangos, C. C., Frangos, C. C., & Sotiropoulos, I. (2011). Problematic internet use among Greek
university students: an ordinal logistic regression with risk factors of negative
psychological beliefs, pornographic sites, and online games. Cyberpsychology, Behavior,
and Social Networking, 14(1-2), 51-58.
Galvan, A. (2010). Adolescent development of the reward system. Frontiers in Human
Neuroscience, 4, 6.
106
Gordon, C. F., Juang, L. P., & Syed, M. (2007). Internet use and well-being among college
students: Beyond frequency of use. Journal of College Student Development, 48(6), 674-
688.
Griffiths, M. (2000). Does Internet and computer" addiction" exist? Some case study evidence.
CyberPsychology and Behavior, 3(2), 211-218.
Griffiths, M., Parke, A., Wood, R., & Parke, J. (2006). Internet gambling: An overview of
psychosocial impacts. UNLV Gaming Research & Review Journal, 10(1), 27-39.
Griffiths, M. D. (2012). Facebook addiction: concerns, criticism, and recommendations—a
response to Andreassen and colleagues. Psychological Reports, 110(2), 518-520.
Grøntved, A., Singhammer, J., Froberg, K., Møller, N. C., Pan, A., Pfeiffer, K. A., & Kristensen,
P. L. (2015). A prospective study of screen time in adolescence and depression symptoms
in young adulthood. Preventive Medicine, 81, 108-113.
Guertler, D., Rumpf, H.-J., Bischof, A., Kastirke, N., Petersen, K. U., John, U., & Meyer, C.
(2014). Assessment of problematic internet use by the compulsive internet use scale and
the internet addiction test: A sample of problematic and pathological gamblers. European
Addiction Research, 20(2), 75-81.
Guillot, C. R., Bello, M. S., Tsai, J. Y., Huh, J., Leventhal, A. M., & Sussman, S. (2016).
Longitudinal associations between anhedonia and internet-related addictive behaviors in
emerging adults. Computers in Human Behavior, 62, 475-479.
Gámez-Guadix, M., Calvete, E., Orue, I., & Las Hayas, C. (2015). Problematic Internet use and
problematic alcohol use from the cognitive–behavioral model: A longitudinal study
among adolescents. Addictive Behaviors, 40, 109-114.
107
Gámez-Guadix, M., Orue, I., Smith, P. K., & Calvete, E. (2013). Longitudinal and reciprocal
relations of cyberbullying with depression, substance use, and problematic internet use
among adolescents. Journal of Adolescent Health, 53(4), 446-452.
Gámez-Guadix, M., Villa-George, F. I., & Calvete, E. (2012). Measurement and analysis of the
cognitive-behavioral model of generalized problematic Internet use among Mexican
adolescents. Journal of Adolescence, 35(6), 1581-1591.
Ha, Y. W., Kim, J., Libaque-Saenz, C. F., Chang, Y., & Park, M.-C. (2015). Use and
gratifications of mobile SNSs: Facebook and KakaoTalk in Korea. Telematics and
Informatics, 32(3), 425-438.
Hertlein, K. M., & Stevenson, A. (2015). The seven “As” contributing to Internet-related
intimacy problems: A literature review. Cyberpsychology: Journal of Psychosocial
Research on Cyberspace, 4(1).
Hicks, A., Comp, S., Horovitz, J., Hovarter, M., Miki, M., & Bevan, J. L. (2012). Why people
use Yelp. com: An exploration of uses and gratifications. Computers in Human Behavior,
28(6), 2274-2279.
Holtz, P., & Appel, M. (2011). Internet use and video gaming predict problem behavior in early
adolescence. Journal of Adolescence, 34(1), 49-58.
Hopp, T., Barker, V., & Schmitz Weiss, A. (2015). Interdependent self-construal, self-efficacy,
and community involvement as predictors of perceived knowledge gain among
MMORPG players. Cyberpsychology, Behavior, and Social Networking, 18(8), 468-473.
Hu, L. t., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis:
Conventional criteria versus new alternatives. Structural Equation Modeling: A
Multidisciplinary Journal, 6(1), 1-55.
108
Huang, C. (2010). Internet use and psychological well-being: a meta-analysis. Cyberpsychology,
Behavior, and Social Networking, 13(3), 241-249.
Huhh, J.S. (2008). Culture and business of PC bangs in Korea. Games and Culture, 3(1), 26-37.
Huys, Q. J. M., Pizzagalli, D. A., Bogdan, R., & Dayan, P. (2013). Mapping anhedonia onto
reinforcement learning: a behavioural meta-analysis. Biology of Mood & Anxiety
Disorders, 3(1), 12.
Hwang, J. Y., Choi, J.S., Gwak, A. R., Jung, D., Choi, S.-W., Lee, J., . . . Jung, H. Y. (2014).
Shared psychological characteristics that are linked to aggression between patients with
Internet addiction and those with alcohol dependence. Annals of general psychiatry,
13(1), 6.
Jang, K. S., Hwang, S. Y., & Choi, J. Y. (2008). Internet addiction and psychiatric symptoms
among Korean adolescents. Journal of School Health, 78(3), 165-171.
Jeltova, I., Fish, M. C., & Revenson, T. A. (2005). Risky sexual behaviors in immigrant
adolescent girls from the former Soviet Union: Role of natal and host culture (vol 43, pg
3, 2005). Journal of School Psychology, 43(2), 171-171. doi:10.1016/j.jsp.2005.04.001.
Jessor, R. & Jessor, S. L. (1977). Problem behavior and psychosocial development: A
longitudinal study of youth. Academic Press : San Diego, CA.
Johnson, P. R., & Yang, S. (2009). Uses and gratifications of Twitter: An examination of user
motives and satisfaction of Twitter use. Paper presented at the Communication
Technology Division of the annual convention of the Association for Education in
Journalism and Mass Communication in Boston, MA.
Johnston, L. D., Miech, R. A., O’Malley, P. M., Bachman, J. G., Schulenberg, J. E., & Patrick,
M. E. (2018). Monitoring the Future national survey results on drug use: 1975-2017:
109
Overview, key findings on adolescent drug use. Ann Arbor: Institute for Social Research,
The University of Michigan.
Johnston, L. D., O’Malley, P. M., Bachman, J. G., & Schulenberg, J. E. (2013). Monitoring the
Future national results on drug use: 2012 Overview, Key Findings on Adolescent Drug
Use. Ann Arbor: Institute for Social Research, The University of Michigan.
Kandel, D., & Kandel, E. (2015). The Gateway Hypothesis of substance abuse: developmental,
biological and societal perspectives. Acta Paediatrica, 104(2), 130-137.
Kandel, D. B. (2002). Stages and pathways of drug involvement: Examining the gateway
hypothesis. New York, New York: Cambridge University Press.
Keller, J., Young, C. B., Kelley, E., Prater, K., Levitin, D. J., & Menon, V. (2013). Trait
anhedonia is associated with reduced reactivity and connectivity of mesolimbic and
paralimbic reward pathways. Journal of Psychiatric Research, 47(10), 1319-1328.
Kim, K., Ryu, E., Chon, M.Y., Yeun, E.J., Choi, S.Y., Seo, J.S., & Nam, B.W. (2006). Internet
addiction in Korean adolescents and its relation to depression and suicidal ideation: a
questionnaire survey. International Journal of Nursing Studies, 43(2), 185-192.
Kim, S. H., Baik, S.-H., Park, C. S., Kim, S. J., Choi, S. W., & Kim, S. E. (2011). Reduced
striatal dopamine D2 receptors in people with Internet addiction. Neuroreport, 22(8),
407-411.
King, K. M., & Chassin, L. (2007). A prospective study of the effects of age of initiation of
alcohol and drug use on young adult substance dependence. Journal of Studies on
Alcohol and Drugs, 68(2), 256-265.
Király, O., Griffiths, M. D., Urbán, R., Farkas, J., Kökönyei, G., Elekes, Z., . . . Demetrovics, Z.
(2014). Problematic internet use and problematic online gaming are not the same:
110
findings from a large nationally representative adolescent sample. Cyberpsychology,
Behavior, and Social Networking, 17(12), 749-754.
Kittinger, R., Correia, C. J., & Irons, J. G. (2012). Relationship between Facebook use and
problematic Internet use among college students. Cyberpsychology, Behavior, and Social
Networking, 15(6), 324-327.
Ko, C.H., Liu, G.C., Yen, J.Y., Yen, C.F., Chen, C.S., & Lin, W.C. (2013). The brain activations
for both cue-induced gaming urge and smoking craving among subjects comorbid with
Internet gaming addiction and nicotine dependence. Journal of Psychiatric Research,
47(4), 486-493.
Ko, C. H., Yen, J. Y., Chen, C. C., Chen, S. H., Wu, K., & Yen, C. F. (2006). Tridimensional
personality of adolescents with Internet addiction and substance use experience. The
Canadian Journal of Psychiatry, 51(14), 887-894.
Ko, C. H., Yen, J. Y., Liu, S. C., Huang, C. F., & Yen, C. F. (2009). The associations between
aggressive behaviors and Internet addiction and online activities in adolescents. Journal
of Adolescent Health, 44(6), 598-605.
Ko, C. H., Yen, J. Y., Yen, C. F., Chen, C. S., & Chen, C. C. (2012). The association between
Internet addiction and psychiatric disorder: a review of the literature. European
Psychiatry, 27(1), 1-8.
Ko, C. H., Yen, J. Y., Yen, C. F., Chen, C. S., Weng, C. C., & Chen, C. C. (2008). The
association between Internet addiction and problematic alcohol use in adolescents: the
problem behavior model. CyberPsychology & Behavior, 11(5), 571-576.
Kozma, R. B. (2003). Technology and classroom practices: An international study. Journal of
Research on Technology in Education, 36(1), 1-14.
111
Krause, A. E., North, A. C., & Heritage, B. (2014). The uses and gratifications of using
Facebook music listening applications. Computers in Human Behavior, 39, 71-77.
Krishnamurthy, S., & Chetlapalli, S. K. (2015). Internet addiction: Prevalence and risk factors: A
cross-sectional study among college students in Bengaluru, the Silicon Valley of India.
Indian Journal of Public Health, 59(2), 115-121.
Kuss, D. J., & Griffiths, M. D. (2011). Online social networking and addiction—a review of the
psychological literature. International Journal of Environmental Research and Public
Health, 8(9), 3528-3552.
Kuss, D. J., & Griffiths, M. D. (2012a). Internet and gaming addiction: a systematic literature
review of neuroimaging studies. Brain Sciences, 2(3), 347-374.
Kuss, D. J., & Griffiths, M. D. (2012b). Internet gaming addiction: A systematic review of
empirical research. International Journal of Mental Health and Addiction, 10(2), 278-
296.
Kuss, D. J., Griffiths, M. D., & Binder, J. F. (2013). Internet addiction in students: Prevalence
and risk factors. Computers in Human Behavior, 29(3), 959-966.
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(25), 4026-4052.
Lam, L. T., Peng, Z.-w., Mai, J.C., & Jing, J. (2009). Factors associated with Internet addiction
among adolescents. CyberPsychology & Behavior, 12(5), 551-555.
LaRose, R. (2011). Uses and Gratifications of Internet Addiction. In K. S. Young, & De Abreu,
C. N. (Ed.), Internet Addiction: A Handbook and Guide to Evaluation and Treatment.
Hoboken, New Jersey: John Wiley & Sons, Inc.
112
LaRose, R., & Eastin, M. S. (2004). A social cognitive theory of Internet uses and gratifications:
Toward a new model of media attendance. Journal of Broadcasting & Electronic Media,
48(3), 358-377.
LaRose, R., & Kim, J. (2006). Share, steal, or buy? A social cognitive perspective of music
downloading. CyberPsychology & Behavior, 10(2), 267-277.
LaRose, R., Lai, Y. J., Lange, R., Love, B., & Wu, Y. (2005). Sharing or piracy? An exploration
of downloading behavior. Journal of Computer-Mediated Communication, 11(1), 1-21.
LaRose, R., Lin, C. A., & Eastin, M. S. (2003). Unregulated Internet usage: Addiction, habit, or
deficient self-regulation? Media Psychology, 5(3), 225-253.
LaRose, R., Mastro, D., & Eastin, M. S. (2001). Understanding Internet usage a social-cognitive
approach to uses and gratifications. Social Science Computer Review, 19(4), 395-413.
Lee, Y. S., Han, D. H., Kim, S. M., & Renshaw, P. F. (2013). Substance abuse precedes internet
addiction. Addictive Behaviors, 38(4), 2022-2025.
Lenhart, A., Duggan, M., Perrin, A., Stepler, R., Rainie, H., & Parker, K. (2015). Teens, social
media & technology overview 2015. Pew Research Center [Internet & American Life
Project]. Retreived from: http://www.pewinternet.org/2015/04/09/teens-social-media-
technology-2015/
Leshem, R. (2016). Brain development, impulsivity, risky decision making, and cognitive
control: Integrating cognitive and socioemotional processes during adolescence—An
introduction to the special Issue. Developmental Neuropsychology, 41:1-2, 1-5, DOI:
10.1080/87565641.2016.1187033.
Leung, L. (2004). Net-generation attributes and seductive properties of the Internet as predictors
of online activities and Internet addiction. CyberPsychology & Behavior, 7(3), 333-348.
113
Leung, L. (2014). Predicting Internet risks: A longitudinal panel study of gratifications-sought,
Internet addiction symptoms, and social media use among children and adolescents.
Health Psychology and Behavioral Medicine: an Open Access Journal, 2(1), 424-439.
Leung, L., & Lee, P. S. N. (2012). The influences of information literacy, internet addiction and
parenting styles on internet risks. New Media & Society, 14(1), 117-136.
Leventhal, A. M., Strong, D. R., Kirkpatrick, M. G., Unger, J. B., Sussman, S., Riggs, N. R., . . .
Audrain-McGovern, J. (2015). Association of electronic cigarette use with initiation of
combustible tobacco product smoking in early adolescence. JAMA, 314(7), 700-707.
Lewinsohn, P. M., & Graf, M. (1973). Pleasant activities and depression. Journal of Consulting
and Clinical Psychology, 41(2), 261.
Li, S.M., & Chung, T.M. (2006). Internet function and Internet addictive behavior. Computers in
Human Behavior, 22(6), 1067-1071.
Liau, A. K., Neo, E. C., Gentile, D. A., Choo, H., Sim, T., Li, D., & Khoo, A. (2015).
Impulsivity, self-regulation, and pathological video gaming among youth: testing a
mediation model. Asia Pacific Journal of Public Health, 27(2), NP2188-NP2196.
Lin, C. A. (2001). Audience attributes, media supplementation, and likely online service
adoption. Mass Communication & Society, 4(1), 19-38.
Lin, M.P., Ko, H.C., & Wu, J. Y.W. (2011). Prevalence and psychosocial risk factors associated
with Internet addiction in a nationally representative sample of college students in
Taiwan. Cyberpsychology, Behavior, and Social Networking, 14(12), 741-746.
Lisha, N. E., Grana, R., Sun, P., Rohrbach, L., Spruijt-Metz, D., Reifman, A., & Sussman, S.
(2014). Evaluation of the psychometric properties of the Revised Inventory of the
114
Dimensions of Emerging Adulthood (IDEA-R) in a sample of continuation high school
students. Evaluation & The Health Professions, 37(2), 156-177.
Liu, T. C., Desai, R. A., Krishnan-Sarin, S., Cavallo, D. A., & Potenza, M. N. (2011).
Problematic Internet use and health in adolescents: data from a high school survey in
Connecticut. The Journal of Clinical Psychiatry, 72(6), 836.
Lloyd, J., Doll, H., Hawton, K., Dutton, W. H., Geddes, J. R., Goodwin, G. M., & Rogers, R. D.
(2010). How psychological symptoms relate to different motivations for gambling: An
online study of internet gamblers. Biological Psychiatry, 68(8), 733-740.
Lopez-Quintero, C., de los Cobos, J. P., Hasin, D. S., Okuda, M., Wang, S., Grant, B. F., &
Blanco, C. (2011). Probability and predictors of transition from first use to dependence
on nicotine, alcohol, cannabis, and cocaine: Results of the National Epidemiologic
Survey on Alcohol and Related Conditions (NESARC). Drug and Alcohol Dependence,
115(1), 120-130.
Luczak, S. E., Khoddam, R., Yu, S., Wall, T. L., Schwartz, A., & Sussman, S. (2017).
Prevalence and co‐occurrence of addictions in US ethnic/racial groups: Implications for
genetic research. The American journal on addictions, 26(5), 424-436.
Matusitz, J., & McCormick, J. (2012). Sedentarism: the effects of Internet use on human obesity
in the United States. Social Work in Public Health, 27(3), 250-269.
Meerkerk, G. J., van den Eijnden, R. J. J. M., Franken, I. H. A., & Garretsen, H. F. L. (2010). Is
compulsive internet use related to sensitivity to reward and punishment, and impulsivity?
Computers in Human Behavior, 26(4), 729-735.
Milani, L., Osualdella, D., & Di Blasio, P. (2009). Quality of interpersonal relationships and
problematic Internet use in adolescence. CyberPsychology & Behavior, 12(6), 681-684.
115
Mitchell, K. J., Finkelhor, D., & Wolak, J. (2001). Risk factors for and impact of online sexual
solicitation of youth. JAMA, 285(23), 3011-3014.
Montag, C., Bey, K., Sha, P., Li, M., Chen, Y. F., Liu, W. Y., . . . Keiper, J. (2015). Is it
meaningful to distinguish between generalized and specific Internet addiction? Evidence
from a cross‐cultural study from Germany, Sweden, Taiwan and China. Asia‐Pacific
Psychiatry, 7(1), 20-26.
Moore, M. (2008, November 11). China Offers Therapy to 4 Million Internet Addicts. The
Telegraph. Retreived from:
https://www.telegraph.co.uk/news/worldnews/asia/china/3437769/China-offers-therapy-
to-4-million-internet-addicts.html.
Moreno, M. A., Jelenchick, L., Cox, E., Young, H., & Christakis, D. A. (2011a). Problematic
internet use among US youth: a systematic review. Archives of Pediatrics & Adolescent
Medicine, 165(9), 797-805.
Moreno, M. A., Jelenchick, L. A., Egan, K. G., Cox, E., Young, H., Gannon, K. E., & Becker, T.
(2011b). Feeling bad on Facebook: Depression disclosures by college students on a social
networking site. Depression and Anxiety, 28(6), 447-455.
Mun, S. Y., & Lee, B. S. (2015). Effects of an integrated Internet addiction prevention program
on elementary students' self-regulation and Internet addiction. Journal of Korean
Academy of Nursing, 45(2), 251-261.
Murali, V., & George, S. (2007). Lost online: an overview of internet addiction. Advances in
Psychiatric Treatment, 13(1), 24-30.
116
Muthén, B., & Muthén, L. K. (2000). Integrating person‐centered and variable‐centered analyses:
Growth mixture modeling with latent trajectory classes. Alcoholism: Clinical and
Experimental Research, 24(6), 882-891.
Muthén, L. K., & Muthén, B. O. (2010). Mplus user's guide: Statistical analysis with latent
variables: User's Guide: Muthén & Muthén.
NIDA. (2018, January 17). Principles of Drug Addiction Treatment: A Research-Based Guide
(Third Edition). Retrieved from https://www.drugabuse.gov/publications/principles-drug-
addiction-treatment-research-based-guide-third-edition on 2018, April 10.
Nylund, K. L. (2007). Latent transition analysis: Modeling extensions and an application to peer
victimization (Doctoral dissertation, University of California, Los Angeles). Retreived
from: http://www.statmodel.com/download/Nylund%20dissertation%20Updated1.pdf.
Palmgreen, P., & Rayburn, J. D. (1985). A comparison of gratification models of media
satisfaction. Communications Monographs, 52(4), 334-346.
Papacharissi, Z., & Rubin, A. M. (2000). Predictors of Internet use. Journal of Broadcasting &
Electronic Media, 44(2), 175-196.
Park, S., & Lee, Y. (2017). Associations of body weight perception and weight control behaviors
with problematic internet use among Korean adolescents. Psychiatry Research, 251, 275-
280.
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(172), 895-909.
Parker, B. J., & Plank, R. E. (2000). A uses and gratifications perspective on the Internet as a
new information source. American Business Review, 18(2), 43-49.
117
Patrick, M. E., & Schulenberg, J. E. (2014). Prevalence and predictors of adolescent alcohol use
and binge drinking in the United States. Alcohol Research: Current Reviews, 35(2), 193-
200.
Peugh, J. L. (2010). A practical guide to multilevel modeling. Journal of School Psychology,
48(1), 85-112.
Phau, I., & Liang, J. (2012). Downloading digital video games: predictors, moderators and
consequences. Marketing Intelligence & Planning, 30(7), 740-756.
Pies, R. (2009). Should DSM-V designate" Internet addiction" a mental disorder? Psychiatry
(Edgmont), 6(2), 31-37.
Pokhrel, P., Herzog, T. A., Black, D. S., Zaman, A., Riggs, N. R., & Sussman, S. (2013).
Adolescent neurocognitive development, self-regulation, and school-based drug use
prevention. Prevention Science, 14(3), 218-228.
Pontes, H. M., & Griffiths, M. D. (2014). Internet addiction disorder and internet gaming
disorder are not the same. Journal of Addiction Research & Therapy, 5(4).
Potts, R., & Sanchez, D. (1994). Television viewing and depression: No news is good news.
Journal of Broadcasting & Electronic Media, 38(1), 79-90.
Raacke, J., & Bonds-Raacke, J. (2008). MySpace and Facebook: Applying the uses and
gratifications theory to exploring friend-networking sites. Cyberpsychology & Behavior,
11(2), 169-174.
Rømer Thomsen, K., Whybrow, P. C., & Kringelbach, M. L. (2015). Reconceptualizing
anhedonia: novel perspectives on balancing the pleasure networks in the human
brain. Frontiers in Behavioral Neuroscience, 9, 49.
118
Rivas‐Drake, D., Seaton, E. K., Markstrom, C., Quintana, S., Syed, M., Lee, R. M., . . . Yip, T.
(2014). Ethnic and racial identity in adolescence: Implications for psychosocial,
academic, and health outcomes. Child Development, 85(1), 40-57.
Rohrbach, L. A., Sun, P., & Sussman, S. (2010). One-year follow-up evaluation of the Project
Towards No Drug Abuse (TND) dissemination trial. Preventive Medicine, 51(3), 313-
319.
Rubin, A. M. (2009). Uses and gratifications. The SAGE handbook of media processes and
effects. Thousand Oaks, CA: Sage Publications Inc.
Ruggiero, T. E. (2000). Uses and gratifications theory in the 21st century. Mass Communication
& Society, 3(1), 3-37.
Ryan, T., Chester, A., Reece, J., & Xenos, S. (2014). The uses and abuses of Facebook: A review
of Facebook addiction. Journal of Behavioral Addictions, 3(3), 133-148.
Rücker, J., Akre, C., Berchtold, A., & Suris, J. C. (2015). Problematic Internet use is associated
with substance use in young adolescents. Acta Paediatrica, 104(5), 504-507.
Saaid, S. A., Al-Rashid, N. A. A., & Abdullah, Z. (2014). The impact of addiction to Twitter
among university students. In Future information technology (pp. 231-236). Springer,
Berlin, Heidelberg.
Sadler, K. (2011). Normal adolescent development. The Mass General Hospital for Children
Adolescent Medicine Handbook, 23-26.
Sampasa-Kanyinga, H., & Chaput, J. P. (2016). Use of social networking sites and alcohol
consumption among adolescents. Public Health, 139, 88-95.
Sanders, J., Munford, R., Liebenberg, L., & Ungar, M. (2017). Peer paradox: the tensions that
peer relationships raise for vulnerable youth. Child & Family Social Work, 22(1), 3-14.
119
Shaffer, H. J., Hall, M. N., & Vander Bilt, J. (2000). "Computer addiction": a critical
consideration. American Journal of Orthopsychiatry, 70(2), 162-168.
Shek, D. T. L., & Yu, L. (2012). Internet addiction phenomenon in early adolescents in Hong
Kong. The Scientific World Journal, 2012.
Siomos, K. E., Dafouli, E. D., Braimiotis, D. A., Mouzas, O. D., & Angelopoulos, N. V. (2008).
Internet addiction among Greek adolescent students. CyberPsychology & Behavior,
11(6), 653-657.
Snaith, R. P., Hamilton, M., Morley, S., Humayan, A., Hargreaves, D., & Trigwell, P. (1995). A
scale for the assessment of hedonic tone the Snaith-Hamilton Pleasure Scale. The British
Journal of Psychiatry, 167(1), 99-103.
Soh, P. C.H., Chew, K. W., Koay, K. Y., & Ang, P. H. (2018). Parents vs peers’ influence on
teenagers’ Internet addiction and risky online activities. Telematics and Informatics,
35(1), 225-236.
Song, I., Larose, R., Eastin, M. S., & Lin, C. A. (2004). Internet gratifications and Internet
addiction: On the uses and abuses of new media. CyberPsychology & Behavior, 7(4),
384-394.
Stafford, T. F., Stafford, M. R., & Schkade, L. L. (2004). Determining uses and gratifications for
the Internet. Decision Sciences, 35(2), 259-288.
Starcevic, V., & Billieux, J. (2017). Does the construct of Internet addiction reflect a single
entity or a spectrum of disorders?. Clinical Neuropsychiatry, 14(1), 5-10.
Steinbeis, N., & Crone, E. A. (2016). The link between cognitive control and decision-making
across child and adolescent development. Current Opinion in Behavioral Sciences, 10,
28-32.
120
Sun, P., Johnson, C. A., Palmer, P., Arpawong, T. E., Unger, J. B., Xie, B., . . . Sussman, S.
(2012). Concurrent and predictive relationships between compulsive Internet use and
substance use: Findings from vocational high school students in China and the USA.
International journal of environmental research and public health, 9(3), 660-673.
Sussman, S. (2010). Emerging adulthood and substance abuse. In L. V. Berhardt (Ed.), Advances
in medicine and biology (Vol. 6, pp. 221-231). Hauppauge, NY: Nova Science
Publishers.
Sussman, S. (2017). Substance and behavioral addictions: Concepts, causes, and cures. New
York, NY: Cambridge University Press.
Sussman, S., Arpawong, T. E., Sun, P., Tsai, J., Rohrbach, L. A., & Spruijt-Metz, D. (2014).
Prevalence and co-occurrence of addictive behaviors among former alternative high
school youth. Journal of Behavioral Addictions, 3(1), 33-40.
Sussman, S., Arriaza, B., & Grigsby, T. J. (2014). Alcohol, tobacco, and other drug misuse
prevention and cessation programming for alternative high school youth: a review.
Journal of School Health, 84(11), 748-758.
Sussman, S., Dent, C. W., & Leu, L. (2000). The one-year prospective prediction of substance
abuse and dependence among high-risk adolescents. Journal of Substance Abuse, 12(4),
373-386.
Sussman, S., Lisha, N., & Griffiths, M. (2011). Prevalence of the addictions: a problem of the
majority or the minority? Evaluation & the Health Professions, 34(1), 3-56.
Sussman, S., & Moran, M. B. (2013). Hidden addiction: television. Journal of Behavioral
Addictions, 2(3), 125-132.
121
Sussman, S., Sun, P., Rohrbach, L. A., & Spruijt-Metz, D. (2012). One-year outcomes of a drug
abuse prevention program for older teens and emerging adults: Evaluating a motivational
interviewing booster component. Health Psychology, 31(4), 476.
Swift, W., Hall, W., & Copeland, J. (2000). One year follow-up of cannabis dependence among
long-term users in Sydney, Australia. Drug & Alcohol Dependence, 59(3), 309-318.
Szczepanik, J. E., Furey, M. L., Nugent, A. C., Henter, I. D., Zarate Jr, C. A., & Lejuez, C. W.
(2017). Altered interaction with environmental reinforcers in major depressive disorder:
Relationship to anhedonia. Behaviour Research and Therapy, 97, 170-177.
Tang, J.H., Chen, M.C., Yang, C.Y., Chung, T.Y., & Lee, Y.A. (2016). Personality traits,
interpersonal relationships, online social support, and Facebook addiction. Telematics
and Informatics, 33(1), 102-108.
Taylor, T. L. (2003). Multiple pleasures: Women and online gaming. Convergence, 9(1), 21-46.
Tone, H.J., Zhao, H.R., & Yan, W.S. (2014). The attraction of online games: An important factor
for Internet Addiction. Computers in Human Behavior, 30, 321-327.
Tonioni, F., D'Alessandris, L., Lai, C., Martinelli, D., Corvino, S., Vasale, M., . . . Bria, P.
(2012). Internet addiction: hours spent online, behaviors and psychological symptoms.
General Hospital Psychiatry, 34(1), 80-87.
Torkzadeh, G., & Koufteros, X. (1994). Factorial validity of a computer self-efficacy scale and
the impact of computer training. Educational and Psychological Measurement, 54(3),
813-821.
Toumbourou, J. W., Stockwell, T., Neighbors, C., Marlatt, G., Sturge, J., & Rehm, J. (2007).
Interventions to reduce harm associated with adolescent substance use. The Lancet,
369(9570), 1391-1401.
122
Tran, B. X., Hinh, N. D., Nguyen, L. H., Le, B. N., Nong, V. M., Thuc, V. T. M., . . . Ho, R. C.
M. (2017). A study on the influence of internet addiction and online interpersonal
influences on health-related quality of life in young Vietnamese. BMC Public Health,
17(1), 138.
Tsai, J., Huh, J., Idrisov, B., Galimov, A., Espada, J. P., Gonzálvez, M. T., & Sussman, S.
(2016). Prevalence and Co-Occurrence of Addictive Behaviors Among Russian and
Spanish Youth: A Replication Study. Journal of Drug Education, 46(1-2), 32-46.
Valente, T. W., Ritt‐Olson, A., Stacy, A., Unger, J. B., Okamoto, J., & Sussman, S. (2007). Peer
acceleration: effects of a social network tailored substance abuse prevention program
among high‐risk adolescents. Addiction, 102(11), 1804-1815.
Van den Bulck, J., Custers, K., & Nelissen, S. (2016). The child-effect in the new media
environment: Challenges and opportunities for communication research. Journal of
Children and Media, 10(1), 30-38.
Van Rooij, A. J., Ferguson, C. J., Van de Mheen, D., & Schoenmakers, T. M. (2017). Time to
abandon Internet Addiction? Predicting problematic Internet, game, and social media use
from psychosocial well-being and application use. Clinical Neuropsychiatry, 14(1), 113-
121.
Van Rooij, A. J., & Prause, N. (2014). A critical review of “Internet addiction” criteria with
suggestions for the future. Journal of Behavioral Addictions, 3(4), 203-213.
Van Rooij, A. J., Schoenmakers, T. M., Van de Eijnden, R. J. J. M., & Van de Mheen, D. (2010).
Compulsive internet use: the role of online gaming and other internet applications.
Journal of Adolescent Health, 47(1), 51-57.
123
Volkow, N. D., Fowler, J. S., Wang, G. J., Baler, R., & Telang, F. (2009). Imaging dopamine's
role in drug abuse and addiction. Neuropharmacology, 56, 3-8.
Vondráčková, P., & Gabrhelík, R. (2016). Prevention of Internet addiction: A systematic review.
Journal of Behavioral Addictions, 5(4), 568-579.
Wegmann, E., Stodt, B., & Brand, M. (2015). OR-94: Decision making under risk and self-
regulation predict tendencies towards Internet addiction and addictive use of social
networking sites. Journal of Behavioral Addictions, 4(S1), 42-43.
Whiteley, L. B., Brown, L. K., Swenson, R. R., Valois, R. F., Vanable, P. A., Carey, M. P., . . .
Romer, D. (2012). African American adolescents meeting sex partners online: Closing
the digital research divide in STI/HIV prevention. The Journal of Primary Prevention,
33(1), 13-18. doi:10.1007/s10935-012-0262-3.
Whiting, A., & Williams, D. (2013). Why people use social media: a uses and gratifications
approach. Qualitative Market Research: An International Journal, 16(4), 362-369.
Wu, C.S., & Cheng, F.F. (2006). Internet café addiction of Taiwanese adolescents.
CyberPsychology & Behavior, 10(2), 220-225.
Xin, M., Xing, J., Pengfei, W., Houru, L., Mengcheng, W., & Hong, Z. (2018). Online activities,
prevalence of Internet addiction and risk factors related to family and school among
adolescents in China. Addictive Behaviors Reports, 7, 14-18.
Yang, S. C., & Tung, C.J. (2007). Comparison of Internet addicts and non-addicts in Taiwanese
high school. Computers in Human Behavior, 23(1), 79-96.
Yau, Y. H. C., Potenza, M. N., Mayes, L. C., & Crowley, M. J. (2015). Blunted feedback
processing during risk-taking in adolescents with features of problematic Internet use.
Addictive Behaviors, 45, 156-163.
124
Younes, F., Halawi, G., Jabbour, H., El Osta, N., Karam, L., Hajj, A., & Khabbaz, L. R. (2016).
Internet addiction and relationships with insomnia, anxiety, depression, stress and self-
esteem in university students: a cross-sectional designed study. PloS One, 11(9),
e0161126.
Young, K. S. (1998). Internet addiction: The emergence of a new clinical disorder.
Cyberpsychology & Behavior, 1(3), 237-244.
Young, K. S. (2007). Cognitive behavior therapy with Internet addicts: treatment outcomes and
implications. CyberPsychology & Behavior, 10(5), 671-679.
Young, K. S., & De Abreu, C. N. (2010). Internet addiction: A handbook and guide to
evaluation and treatment. Hoboken, New Jersey: John Wiley & Sons, Inc.
Zapolski, T. C. B., Fisher, S., Banks, D. E., Hensel, D. J., & Barnes-Najor, J. (2017). Examining
the protective effect of ethnic identity on drug attitudes and use among a diverse youth
population. Journal of Youth and Adolescence, 46(8), 1702-1715.
Zhang, L., Amos, C., & McDowell, W. C. (2008). A comparative study of Internet addiction
between the United States and China. CyberPsychology & Behavior, 11(6), 727-729.
Abstract (if available)
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Prospective associations of stress, compulsive internet use, and posttraumatic growth among emerging adults
PDF
Energy drink consumption, substance use and attention-deficit/hyperactivity disorder among adolescents
PDF
A penta-dimensional longitudinal analysis of the predictors of compulsive internet use among adolescents using linear mixed model (LMM)
PDF
The role of social support in the relationship between adverse childhood experiences and addictive behaviors across adolescence and young adulthood
PDF
Sociocultural stress, coping and substance use among Hispanic/Latino adolescents
PDF
Distress tolerance and mindfulness disposition: associations with substance use during adolescence and emerging adulthood
PDF
Using a structural model of psychopathology to distinguish relations between shared and specific features of psychopathology, smoking, and underlying mechanisms
PDF
Pathways of drug use among people who inject drugs
PDF
Social self-control and adolescent substance use
PDF
The dynamic relationship of emerging adulthood and substance use
PDF
Role transitions, past life events, and their associations with multiple categories of substance use among emerging adults
PDF
Motivational interviewing with adolescent substance users: a closer look
PDF
The influence of contextual factors on the processes of adoption and implementation of evidence-based substance use prevention and tobacco cessation programs in schools
PDF
Understanding the methodological limitations In the ecological momentary assessment of physical activity
PDF
The role of depression symptoms on social information processing and tobacco use among adolescents
PDF
Adolescent social networks, smoking, and loneliness
PDF
The effects of mindfulness on adolescent cigarette smoking: Measurement, mechanisms, and theory
PDF
Psychosocial and cultural factors in the primary prevention of melanoma targeted to multiethnic children
PDF
Understanding the dynamic relationships between physical activity and affective states using real-time data capture techniques
PDF
Cultural risk and protective factors for tobacco use behaviors and depressive symptoms among American Indian adolescents in California
Asset Metadata
Creator
Tsai, Jennifer Yo-ka
(author)
Core Title
The Internet activities, gratifications, and health consequences related to compulsive Internet use
School
Keck School of Medicine
Degree
Doctor of Philosophy
Degree Program
Preventive Medicine (Health Behavior Research)
Publication Date
07/26/2018
Defense Date
05/04/2018
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
compulsive internet use,drug use,gratifications,Internet activities,OAI-PMH Harvest
Format
application/pdf
(imt)
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Rohrbach, Luanne (
committee chair
), Sussman, Steve (
committee chair
), Cederbaum, Julie (
committee member
), Huh, Jimi (
committee member
), Leventhal, Adam (
committee member
)
Creator Email
jyktsai@gmail.com,tsaijy@usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c89-32244
Unique identifier
UC11669032
Identifier
etd-TsaiJennif-6512.pdf (filename),usctheses-c89-32244 (legacy record id)
Legacy Identifier
etd-TsaiJennif-6512.pdf
Dmrecord
32244
Document Type
Dissertation
Format
application/pdf (imt)
Rights
Tsai, Jennifer Yo-ka
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Tags
compulsive internet use
drug use
gratifications
Internet activities