The Social Risk Factors for Internet Addiction in College Undergraduate Students

Objective: The current study aims to explore precipitating and risk factors for internet addiction (IA) in university undergraduate students, and to provide evidence for interventions and the early prevention of IA in this population. Method: Four thousand eight hundred and fty-eight college sophomores completed an online survey on their internet use-related behaviours and social risk factors. Result: Compared to female students (5.4%), more male students (8.3%) were in the moderate and severe IA groups. Male and female students tend to engage in different types of online entertainment. There are some shared and some different precipitating factors for internet use and risk factors for IA between the different gender groups. Conclusion: By understanding the differences and similarities in risk factors for IA in both gender groups of the university undergraduate students, interventions can be more precisely designed to target specic behaviours in this populations.


Background
The internet use has grown exponentially during the past two decades. The development of the internet has brought convenience but also problems such as internet addiction (IA). Young (1998) Green eld, 1999). Furthermore, IA is associated with an increased rate of mental health problems. People with heavy IA, such as pathological heavy online game consumers, are likely to develop psychiatric disorders such as depression, social anxiety, and somatic symptoms (Wei et al., 2012).
The use of internet-based entertainment differs by gender and geographic region. In the United States, male internet users spend more time on pornography, and females usually spend time on online shopping (Young, 2013). In Europe, male internet users usually play single-player online games, while females spend most of their time on social media (Durkee et al., 2012). In Korea, 56.3% of males play online multiplayer games (Ha and Hwang, 2014). In Hong Kong, 30% of participants play online games (Wang et al., 2015). Similar to other addictive behaviours, people engage in online entertainment for different reasons. Shiffman and Rathbun (2011) found that people use smoking as a way to cope with negative affect [2]. IA has a clinical presentation similar to that of substance abuse and behavioural abuse, such as withdrawal symptoms, tolerance, and craving [3]. It is worth noting the difference between males and females in addictive behaviours. For example, women were more likely than men to use smoking to cope with negative affect, and a more recent study suggested that men are more likely to report somatic symptoms such as "ill, in pain, or uncomfortable" as triggers for using drugs (Kennedy et al, 2012).
Previous studies suggested that the severity of IA is positively correlated with the amount of time individuals spend on the internet (Rooij et al; Schoenmakers et al, 2010) [4]. However, currently, as technology use skyrockets, internet-based entertainment has become more accessible through smartphones, tablets, and other portable gaming devices (e.g., Nintendo Switch). Consequently, it is more di cult to assess the amount of time and the frequency of use in the traditional manner [5,6]. In a more recent study, Li and colleagues (2015) found that sustained time spent on online entertainment tends to increase as IA severity increases [7]. The duration of sustained online entertainment for each episode of internet use might be a better indicator to measure the amount of time of internet use.
Existing studies on risk factors for IA were conducted under the biopsychosocial model. Under this model, the family environment was considered one of the most important social factors. Single-parent, frequent migration, left-behind by parents during childhood, and deceased parents were the most common risk factors (Guo et al; Chen et al, 2012). Boys who experience negative life events are more susceptible to IA [8]. In addition to the family environment, interpersonal interaction with peers has also been studied. As a platform for sharing similar activities, the internet is a place where adolescents socialize with each other.
As a result, the in uence of friends could exacerbate adolescents' dependence on the internet [9,10].
Previous studies have suggested that biological factors, social environment factors (e.g. family, friends, etc.), and other psychological conditions (e.g., depression and social anxiety) together impact individuals' internet use behaviour [11]. Positive emotions such as happiness, relaxation, con dence, and the sense of achievement that people receive from online activities could serve as moderators that exacerbate IA [12].
The ndings are mixed on the effect of socioeconomic status (SES) and geographic location on IA. Family SES was found to be a predictor of IA. Speci cally, children from higher SES families were more vulnerable to IA [13][14][15]. Parents' unemployment status and low education levels were also found to be associated with a higher risk of IA ( The current study aims to explore social risk factors for IA in college undergraduate students and to provide evidence for interventions and early prevention that target this speci c population.

Participants
The investigation was carried out at a comprehensive university in mainland China. All sophomores (9,367) were invited to complete a questionnaire online in November 2013. A total of 4,858 (51.86%) responded and completed the survey.

Demographics
The demographics collected in this self-made questionnaire included gender, age of participants, and social risk factors below in this survey.
The social risk factors included number of relocations, childhood trauma experience, whether father and/or mother deceased, infatuation with the internet before college, similar online activities in roommates, a romantic relationship, and unpleasant events on campus. They were de ned as follows:

Number of relocations
Refers to the number of times the family's permanent residence was relocated. For the convenience of statistical analysis, those who had no relocation experience were assigned a value of 1, 1 relocation experience was assigned a value of 2, 2 relocation experiences were assigned a value of 3, and 3 or more relocation experiences were assigned a value of 4.

Childhood trauma experience
In this study, childhood trauma experience covers neglect and physical abuse before the age of 16. The Childhood Trauma Questionnaire has been used to evaluate childhood traumatic experience and veri ed as having good reliability and validity [16,17].

Internet use behaviours
We recorded the types of online entertainment activities (playing games, browsing information, chatting/posting comments, watching movies/TV series), the duration of sustained online entertainment Young's Internet Addiction Test (IAT) revised edition by Young (2011), a 20-item self-evaluation scale including 6 levels (0 = not applicable, 1 = almost never, 2 = occasionally, 3 = sometimes, 4 = often, and 5 = almost), was used in the present study. An IAT score of 50 is commonly considered the cut-off score for IA, which means that individuals who scored 50 or higher were considered to have IA, and those who scored lower than 50 were considered being without IA [18,19]. The total IAT score can be categorized into four classes of severity (normal: 0 to 30; mild: 31-49; moderate: 50-79; severe: 80-100) [20].
Young's IAT (2011) has been translated into different languages and has been veri ed to have good reliability and validity [21,22]. The Chinese version was adopted for the present study (Cronbach's alpha: 0.9085) [23].

Statistical analysis
The participants were separated into three groups (normal group: 0-30, mild IA group: 31-49, and moderate and severe IA group: 50-100) based on the scores using the latest revision of the IAT (Young, 2011). SPSS19 was used for data analysis. Two-sample t-tests, chi-squared tests and analysis of variance (ANOVA) were used for continuous variables and categorical variables as appropriate. Multiple logistic regression analysis was applied to predict the risk of mild IA and moderate and severe IA. The statistical signi cance level was set as p < .05 (two-tailed). The Bonferroni correction and the corrected pvalue were used to adjust the p-value while conducting the Chi-square analysis.

Prevalence of IA and IAT mean scores
The Cronbach's alpha coe cient and Guttman split-half coe cient of IAT were 0.915 and 0.845, respectively, in the present study.
The prevalence of IA in total was 6.81% (331/4858). The IA rate was 8.3% in males (197/2377) and 5.4% in females (134/2481). A higher IA rate in male participants (x 2 = 23.115, p < 0.001) was found in the current study among the 4,858 respondents. Furthermore, there were signi cant differences in the mean IAT score between the three IA severity groups (F = 4.344, p < 0.001) ( Table 1). Online gaming is the main online entertainment activity among male students in the moderate and severe IA groups (55.8%), whereas only 27% of male students in the normal group play online games.
The main trigger of online entertainment was boredom (male > 60%, female > 70%); however, negative affect was the most prevalent trigger of online entertainment activities in the moderate and severe IA groups (male 18.2%, female 12.7%, p < 0.001).
There is no statistically signi cant difference between the two gender groups in the duration of sustained online entertainment (males 2.72 ± 3.234, females 2.58 ± 2.983; t = 1.527, p = 0.527). Compared to the normal group, the moderate and severe groups signi cantly differed in online entertainment activities, triggers, and duration of sustained online entertainment (p < 0.001). (Table 2). Students in the moderate and severe IA groups were more likely to report childhood trauma, being infatuated with internet entertainment before college, not in a romantic relationship, and suffering unpleasant events on campus (p < 0.05). Males in the moderate and severe IA groups were more likely to report relocation and father and/or mother deceased.
Additionally, more than 70% of male and female students reported that their roommates engaged in similar online entertainment activities, but the difference between the moderate and severe IA group and the normal group was only marginally signi cant (p = .051) ( Table 3).    Table 4 about here***

Discussion
In the present study, we found that more male than female students had moderate and severe IA. In the moderate and severe IA group, both male and female students tended to continuously use the internet for more than 5 hours once they started; the main online activity for the male students in the moderate and severe IA groups was online gaming, while for the female students in the moderate or severe IA groups, it was online streaming (e.g., watching movies and dramas). Boredom was the most prominent trigger for using internet-based entertainment; however, negative affect-triggered internet use was more likely to become moderate or severe IA. Infatuation with the internet before college and adjustment issues after entering college were shared risk factors for both gender groups in the mild and moderate IA groups.
Roommates engaging in similar internet-based entertainment was a risk factor only for male students' IA and not being in a romantic relationship was only a risk factor for female students' IA. Relocation, the death of one or both parents, and childhood traumatic experience only affected mild IA.

Type of internet use
The main type of internet-based entertainment was different between male and female participants. Similar to previous research, our study found that online gaming is the main type of IA for male students [24]. For the female students in the moderate and severe IA group, 51.5% chose to use online streaming services as their main internet-based entertainment. This is possibly due to the relatively easy accessibility of online streaming services. It is also a risk factor for IA in female students, suggesting that research on female college students' IA should focus on this subtype. The different choice of internetbased entertainment between male and female students may be due to their unique needs and goals in using the internet.

Duration of sustained online entertainment
The duration of sustained online entertainment was used as an indicator of prolonged internet use for IA.
It varied from 2 hours in the normal group to 5 hours in the moderate and severe IA groups. Previous studies usually use "how many hours per day" and "how many days per week" to assess the severity of IA [5,6]. Wei and colleagues (2012) suggest that women tend to have shorter duration of sustained online entertainment, which is contradictory to our ndings [24]. Our study found no statistically signi cant difference between gender groups, which suggests that it could potentially serve as a universal indicator of the severity of IA for both men and women.

Triggers of internet-based entertainment
Boredom, stress, and negative affect were common triggers of massive internet use [25,26]. Boredom was the main trigger of mild IA. Participants in the moderate and severe IA group reported a signi cantly higher incidence of negative affect-triggered internet use than boredom-triggered use. This suggests that online entertainment serves as a coping mechanism for individuals with moderate to severe IA. Female students rarely report invitation from friends as a trigger for online entertainment, while it is the second most reported trigger from male students.

Risk factors for IA
History of using the internet (infatuation with the internet) is considered a predictor of IA. Most individuals with internet game addiction reported having their rst contact with online games in primary school and with internet overuse in middle school [7]; the time point of rst contact might in uence the severity of IA [24]. Our study yields a similar nding that experience with online games during early adolescence has an important impact on internet game addiction. Additionally, infatuation with the internet before college and adjustment issues at the beginning of college are common risk factors across gender groups.
IA could be a result of genes and environmental interactions. First-degree relatives of pathological gamblers showed signi cantly higher incidences of addictive behaviours compared to control groups [27] [28]. A twin study discovered that IA is 20 to 48 percent in uenced by genetic factors [29] [30] and over 50 percent by environmental factors [30]. The current study found that infatuation with the internet before college and early encounters with the internet at a young age are prominent risk factors for IA. This suggests that genetic vulnerabilities combined with the in uence of the family environment may play an important role in early contact with online entertainment and IA in childhood or adolescence.
Clinical implication Since the current study found that boredom is a common trigger of internet use for the majority of college students, it is important to create interventions that target this problem. In China, high school students commonly spend an enormous amount of time preparing for college entrance exams. After the exam, students who were successfully admitted by colleges suddenly have much free time and wonder what is the source of their boredom. Facilitating and helping students engage in meaningful campus activities might help to reduce it. This further reduces the likelihood of students being addicted to the internet.
Additionally, schools should pay close attention to the differences in triggering events between gender groups and design interventions accordingly.
Because online entertainment could potentially serve as a coping mechanism for negative affect, teaching students emotional regulation and stress management skills could be bene cial. Furthermore, preventions should be considered starting earlier, since early contact with the internet is associated with IA in college. Parents and teachers in elementary and middle schools should be prepared to be the rst line of defence. Preventions could strive to put restrictions on the duration and type of online activities of children and internet use for adolescents.

Limitations
Since our study is conducted on college undergraduate students, the results may have the limitation of being generalizable to the broader population. Future research on this topic could include participants' genetic information and family environment (e.g., how parents use online entertainment or whether parents have any kind of addiction) on a more diverse population. Consent for publication Written informed consent for publication was obtained from all participants.

Availability of data and materials Not applicable
Competing Interests The authors declare that they have no con ict of interest.
Funding This work was supported by the Education Department of Sichuan Province, China (11SB142) played the role in the design of the study;, the Department of Science and Technology of Sichuan provincial government (2019YFS0153) granted the data collection; the support of analysis came from the