From dissociative experience and internet use to happiness: A Taiwan Birth Cohort Study

Functional and excessive use of internet is hard to distinguish, and internet use can affect adolescents’ development of self-identity. The aim of our study was to investigate the pathway relationship of the risk and protective factors of internet use, including mother care, absorptive dissociative trait, having been bullied, exercise, self-perceived depressive mood and happiness of 12-year-old adolescents. Methods The Taiwan Birth Cohort Study dataset, which used a national household probability sampling method and included 17,694 12-years-old adolescents, was used for this study. Our results showed 5.3% adolescents reported to spend more than ve hours online during schooldays. Additionally, adolescents that spend more than ve hours online during schooldays tended to have higher absorptive trait, perceived less care from mothers, more likely to have been bullied, expressed higher level of depressed mood, which leads to lower level of perceived happiness.


Background
Internet networking has transformed how people communicate, entertain, behave, and educate. However, along with the convenience technology has brought., it has also brought growing adverse effects, such as pathological issues. Internet usage have found to affect the brain and cognitive process [1], this in uence on the "digital natives" is especially important. Adolescents are at stage of development where they are on the search for their identities, and formation of self-image, along with the development of executive function [2]. Internet use could interfere with these developments [2]. However, functional and excessive use of internet is hard to distinguish, and 45% of the adolescents in the United States have reported to be "almost constantly" online, and 95% having access to a smartphone [3]. Internet addiction was proposed to be included in the fth edition of the Diagnostic and Statistical Manual of Mental Disorders. Alongside with this debate was whether the diagnosis should be towards generalized internet use and/or potentially addictive activities that can be engaged on the internet (e.g., gaming, social media, pornography, information seeking and shopping). The nal DSM-5 introduced internet gambling disorder as a subtype of pathological gambling under the category of Substance-Related and Addictive Disorders.
Internet addiction has been reported to be a public health concern in both China and South Korea [4]. In Germany, a 1.16% prevalence was found for internet gaming disorder in adolescents [5], however, a much higher prevalence of 6% was found in Korea [6]. Internet gaming disorder has been found to be related to internalizing and externalizing problems in teenagers, including depression, anxiety, impulsivity and aggression [6]. Besides using internet for gaming, increased use of social media screen time has also been found to be a risk factor for depression and suicide [7]. Showing that excessive internet use can have damaging effect on mental health.
Besides internet use, bullying has also been found to effect the psychological well-being of adolescents [6]. Although internet use has its downside, however, other studies have found that social media use can boost well-being [8], depending on the motivation and frequency of social media use [3]. Shaw and Grant proposed that people with low self-esteem are more likely to decrease loneliness, modify their mood, and seek social support online [9]. Kurniasanti et al. also proposed that internet gaming can be used a as method of mood modi cation to adjustment their mood, get away from reality and overcome depression [2]. Adolescents who experience bullying in real life may seek social support online, however, over a third of children who have experienced cyberbullying also reported to have experienced bullying in real life [11].
Nevertheless, whether adolescents seek social support online because they have experienced interpersonal di culties in real life, or vice versa, is yet to be determined.
In addition to the possibility of seeking social support, the individual's own trait may also increase their risk to becoming addicted online. A previous study found association between dissociative experiences and internet addiction [12]. Furthermore, absorptive trait within the dissociative experience have been found to be associated with smartphone addiction in adolescents [13], hence it is included in our investigation as a potential risk factor for internet addiction.
Besides the risk factor on internet addiction, we were also interested in the protective factors which may prevent internet usage from in uencing the daily activity and social interaction of adolescents. Family guidance and care can prevent adolescents' involvement in problematic and hazardous behaviors. In the contrary, families with high levels of con ict and dysfunction were strong predictors for internet addiction [14]. Besides family, physical exercise can also improve psychological adjustment and well-being [15]. Exercise along with cognitive-behavioral therapy have been found to improve the symptoms of patients with depression [16]. Therefore, whether mother care and exercise are resilience factors to internet addiction will also be investigated in our study.
Therefore, the aim of our study was using a national birth cohort dataset to investigate the risk and protective factors of internet use. Including the pathway relationship of the mother care, absorptive dissociative trait, bullying, exercise, internet use, self-perceived depressive mood and happiness of 12- year-old adolescents.

Participants
This study used the Taiwan Birth Cohort Study (TBCS) 12-year-old dataset. Aimed to build a sample representative of the children in Taiwan, the TBCS used household probability sampling method. Babies born in the year 2005 were randomly selected with no exclusion criteria [17]. A two-stage strati ed random sampling was used, at the rst stage, the primary sampling unit was cities and town [17]. Eightyve townships were systematic randomly selected, by grouping all 369 townships in Taiwan into 12 strata according to four levels of size of the settlement in which the subjects were resident and three levels of total fertility rate. In the second stage, newborns were proportionally selected according to the rate of births from the 85 selected settlements [17]. At when the children were 6-months, 21 248 babies and families (11.7% selection rate) were selected [17]. The 12-years-old is the fth stage dataset, 188,14 families agreed to participate, and within these 17 694 (94.05%) adolescents completed the questionnaires. The protocol of the study was approved by the institutional review board of a teaching hospital. Written informed consent was obtained at each stage of the study after a detailed explanation of the study.

Materials
All factors analyzed in this study were of adolescents' self-report.
Absorptive trait. The TBCS included three items from the Adolescent Dissociative Experience Scale. "I get so wrapped up in watching TV, reading, or playing a video game that I don't have any idea what's going on around me." "I am so good at lying and acting that I believe it myself." "I can't gure out if things really happened or if I only dreamed or thought about them." Response choices ranged from 1 (strongly agree) to 4 (strongly disagree). After recoding the responses, with higher total score implying higher absorptive trait, the Chronbach's alpha of the three items resulted in .559.
Happiness. The self-perceived happiness of the adolescents was measured using the seven item Chinese Oxford Happiness Questionnaire, which has shown good psychometric properties in community adolescents in Taiwan [18]. Higher scores implied better perceived happiness.
Internet use. The TBCS asked two questions "How many hours do you spend online during schooldays?", and another question asking "How many hours do you spend online on days when you don't have to go to school?" In addition to the continuous variable derived from these questions, the hours adolescents spent online were further dichotomized into those who spend less than an hour, and those who spend more than 5 hours a day online. Since Twenge, Joiner, Rogers and Martin 8 found adolescents who spent more than 5 hours a day on social media and smartphones had 66% increased risk in suicide compared to those who spent less than an hour online.
Depressive state. The item "I felt depressed" from the Center for Epidemiological Studies-Depression was used to measure adolescents' depressive state. Response choices were never = 1, once in a while = 2, sometimes = 3, often = 4, always = 5.
Maternal care. Adolescents' self-perceived level of maternal care was measured by asking the adolescents three questions: 1) "Does your mother (or the person who mainly takes care of you) know what you do on your free time?" 2) "Does your mother (or the person who mainly takes care of you) know who you normally hang out with (Eg. When you go out to play, exercise, shopping or do homework etc.?" 3) "Does your mother (or the person who mainly takes care of you) know when you go to bed?". These items were answered in a ve-point Likert scale of "always", "often", "sometimes", "once in a while", or "never". The ratings of these items were combined to form the mother care factor, which has a Cronbach's alpha of .702.

Statistical analysis
The demographic distribution of the adolescents and parents were analyzed using Statistical Package for the Social Sciences (SPSS) 20.0 for Windows software (SPSS Inc., Chicago, USA). Bayesian analysis, a multiple imputation method based on item response theory, accounts for multiple sources of correlation was used to replace missing data.
Structural equation model (SEM) was used to investigate the pathway relationship of internet use, dissociative absorption experiences, being bullied, and perceived level of happiness of these adolescents.
The SEM was analyzed using the Analysis of a MOment Structures 7.0 statistical software package (SPSS Inc., Chicago, USA). SEM models with a p value greater than .5, adjusted goodness-of-t index (AGFI) greater than .9, root mean square error of approximation (RMSEA) less than .08 implies that the null model approximates the real structure.

Results
The demographic distribution of the adolescents and parents is shown in Table 1. Results showed 75.7% teens reported to have never been bullied, which means approximately a quarter of the children have ever experienced bullying. Adolescents reports to go online on average of 1.55 (standard deviation=1.87) hours during schooldays, and 3.62 on days without school. Additionally, 5.3% of the teens reports to go online more than 5 hours on schooldays and 25.9% on days without school. The majority of the teens reported to exercise on a regular basis (82.1%).
Two SEMs were analyzed to investigate the pathway relationship of internet use, absorptive trait, bullying, depressive state and level of happiness in teenagers, with maternal care and exercise being the resilience factors within the model. The rst model included internet use as a continuous variable, and in the second model internet use was dichotomized into those who were online less than 1 hour or more than 5 hours a day (those who reported to go online between 1 and 5 hours were excluded in this model). The second SEM investigated the pathway differences of sex, maternal care, absorptive trait, exercise, internet use during schooldays and depressive state in teens that spent more than ve hours and less than an hour online during schooldays, as shown in gure 2. The model resulted in a good t, with a p value of .862, AGFI of 1.000, and RMSEA of less than .001. Similar to the results of the rst gure, females, those who perceived lower maternal care, have been bullied and/or higher absorptive trait were more likely to spend more than ve hours online per day during schooldays (β=.05, p<.001; β=.03, p=.011; β=-.10, p<.001; β=.09, p<.001). This model further found, teens who spend more than ve hours online schooldays were less likely to exercise, and more likely to be depressed (β=-.05, p<.001; β=.05, p<.001). Those who perceived higher maternal care, lower absorptive trait, less frequency of being bullied, exercised more, and less depressed reported to have higher level of happiness (β=.33, p<.001; β=-.18, p<.001; β=-.06, p<.001; β=.11, p<.001; β=-.19, p<.001).
The third SEM investigated the pathway differences of sex, maternal care, absorptive trait, exercise, internet use in teens that spent more than ve hours and less than an hour online during off schooldays. However, since the nding of this model was similar to that of the continuous internet use model ( gure 1), the result of this model was not presented.

Discussion
Our national birth cohort study found among the 17,694 adolescents that participated, 5.3% reported to spend more than ve hours online during schooldays. Adolescents that spend more than ve hours online during schooldays reported to have higher absorptive trait, perceived less care from mothers, more likely to have been bullied, expressed higher depressed mood, which leads to lower level of perceived happiness. Adolescents that spend more than 5 hours online during schooldays, compared to those that spent less than an hour online, were more likely to have been bullied and affect their level of happiness, showing that they may be a more pathologic group effecting their occupational function (which for this group of students is being in school).
Approximately ve percent of the teenagers in our study reported to spend more than ve hours online during school days. Although currently there is still no consistent set of criteria to de ne internet addiction, our study used the cutoff point of spending more than ve hours online during schoolday, and a 5.3% prevalence was found, which is within the range of 1.5-8.2% found in a review study [19].
With regards to bullying, approximately a quarter (24.7%) of the adolescents in our study reported to have been bullied, however, within teenagers who goes online more than 5 hours a day, 31.15% reported to have been bullied in school. The reported rate of bullying is similar to the 25.3% found in the Taiwan Birth Cohort pilot Study [20], which is similar to the 16.9% found in the United States [21]. Furthermore, our study found teenagers that were bullied in school were more likely to be in this high risk group for excessive internet use. Lam, Peng, Mai and Jing found recent stressful event as a risk factor for internet addiction [22], and Young proposed that excessive use of internet as a coping mechanism to avoid negative emotions and problems [23]. Therefore, being bullied in school maybe one of the stressful events adolescents experience which increases their risk of being addicted online.
In addition to the stressful event of being bullied, the personal characteristics of high absorptive dissociative trait also increases their risk of becoming addicted online. This is in line with previous studies which found dissociative experiences to be associated with internet addiction [12]. Furthermore, absorptive trait within the dissociative experience have been found to be associated with smartphone addiction [13]. A meta-analysis also found "escape from self" to be an important intrapersonal risk factor for internet addiction [24]. Adolescents may experience great distress in the process of identify formation, those who are unable to overcome the stress in reality may turn to the internet for transient escape from real life and satisfaction from the virtual self, leading to internet addiction [24].
Besides the risk factor of absorptive trait and being bullied, we found maternal care and regular exercise were protective factors to being addicted online. Parental care and protection can decrease the adolescents' motivation of to participate in social networking and becoming addicted online [25]. It also serves as a protective factor preventing teenagers from participating in problematic and harmful behaviors [26]. In contrast, a previous study found teens with internet addiction had a higher rate of con ict with parents compared to non-addicts [27]. For those who experience con ict at home or bullying in school may seek interpersonal connection and a sense of belonging online. Besides care from family, regular exercise was another protective factor. Exercise itself can boost mood and the psychological well-being of an individual, and a previous study also found those not addicted online more frequently attended stress-releasing leisure activities compared to those who were addicted online [28].
As for the consequences of being addicted online. Our study found, those who spent more time online were more likely to experience depressed mood and perceived lower level of happiness. The association between internet addiction and depressed mood have also been found in a previous study [7]. Twenge, Joiner, Rogers and Martin found adolescents who spent more time on new media screen activities (such as smartphone devices and social media) were more likely to report mental health issues, and those who reported to have spent more time on nonscreen activities (such as exercise or in-person interaction) were less likely to report mental health issues [28]. The negative relationship between internet use and exercise was also found in our study, adolescents that spent more time online, were less likely to exercise regularly. Furthermore, those that exercised regularly perceived higher level of happiness, and those who spent more time online perceived lower level of happiness.
A limitation of our study was that only the hours adolescents spent online were collected, and no information regarding the behavioral consequences of internet addiction was collected. The current proposed diagnostic criteria for internet addiction includes the four components of: 1) excessive use, 2) withdrawal, including feelings of anger, anxiety, or depression when the access to a computer or smartphone is limited, 3) tolerance, including the need for more hours of use, and 4) adverse consequences, including fatigue, social isolation, or poor academic or occupational performances [4]. However, since there are still no o cial diagnostic criteria for internet addiction, therefore no diagnostic consensus has been reached. Furthermore, the general use of internet was collected, and no information regarding the habit of use was collected. Young and Cristiano has proposed internet usage to be separated into four categories of gaming, social media, pornography, and information seeking [29]. Further information regarding the purpose of internet use may provide us with more information regarding the different subtype of internet addiction and its effects.

Conclusions
Our large national birth cohort study showed those with higher absorptive dissociative trait, perceived less care from mothers and have been bullied in school were more likely to spend more than ve hours online during schooldays, expressed higher depressed mood and lower level of perceived happiness. Since parental care is a protective factor for internet addiction, parents should provide care and guidance, while respecting the autonomy of the adolescents, reducing the motivation of adolescents to become involved with social networking and risk of internet addiction [13,24]. Furthermore, adolescents who do not have an adequate coping strategy when faced with stressful situation are more likely to choose avoidant methods, which increases their risk of becoming addicted. Exercise rehabilitation has been promoted by Kim to increase physical and mental health condition of those with internet addiction [30]. Therefore, it should be promoted to all adolescents, especially those in the high risk group such as those who have experienced bullying or with absorptive trait, to prevent them from being addicted online.  Figure 1 The path relationships among dissociative trait, bullying, internet use, depressive mood and happiness in