Internet Addiction and Maladaptive Schemas: The Potential Role of Disconnection/Rejection and Impaired Autonomy/Achievement.

INTRODUCTION AND OBJECTIVES
Problematic internet use (PIU) has become public health concern, particularly among adolescents and emerging adults. There is growing interest concerning the potential impacts of early maladaptive schemas (EMSs) on PIU and its most severe manifestation internet addiction (IA). However, a deeper understanding of these relationships is needed regarding of effects of schemas on IA. The purposed of the present study was to explore the role of EMSs among adults.


METHODS
The sample comprised 714 Iranian participants who completed a self-report survey comprising sociodemographic variables, the Internet Addiction Test (IAT), and the Young Schema Questionnaire-Short Form (YSQ-SF). The data were analyzed using partial least squares structural equation modeling (PLS-SEM).


RESULTS
Findings indicated that there was a positive and significant relationship between EMSs domains and IA. The results confirmed that disconnection/rejection domains and impaired autonomy/achievement schema domains were significantly related with IA. The results of the analysis of convergent validity and discriminant validity were acceptable among the nine reflective constructs.


CONCLUSION
Findings of the present study indicated that existence of underlying EMSs may be a vulnerability factor for developing IA and adds to the growing body of the cyberpsychology literature that has examined the relationships between the EMSs and IA.

However, despite the many benefits of the internet (Wiederhold, 2017), research has indicated that a small minority of individuals overuse the internet and may be at risk of developing internet addiction (Ostovar et al., 2016).

| Internet addiction
Over the past two decades, the internet has revolutionized the way individuals communicate with each other. Moreover, it has become an integral, popular and essential part of many individual's daily lives Gervasi et al., 2017). M. D. Griffiths (1998Griffiths ( , 1999 argued that internet addiction is an umbrella term, due to the fact that the vast majority of individuals experiencing problems report that their concerns relate to specific online activities (e.g., online gaming). The advent of smartphones with affordable and instant internet access has expedited individuals' adoption of the internet . The accessibility of online activities can lead individuals to spend more time online than they originally intend (H. M. Pontes et al., 2015). Increased access and persistent internet use increase the risk of developing online addictive behaviour  that can negatively impact on psychological development , physical health, education (e.g., poor grades), and relationships (e.g., interpersonal difficulties), as well as association with other behaviours (e.g., comorbidity with substance addictions; Chi et al., 2020;Kojima et al., 2019).
Between 2000 and 2020, the number of internet users increased almost tenfold worldwide, with an estimated 4.7 billion individual internet users worldwide and 3.2 billion individuals using online social network sites. More than half of current internet users are young individuals from the Middle East (Internet World Stats, 2019). In Asia, there has been a high variation in prevalence of internet use (ranging from 8.1% to 50.9%) among young people and adolescents (Internet World Stats, 2019). In a study on the worldwide prevalence of internet addiction (IA), Xin et al. (2018) reported that in Europe and the United States, rates of IA ranged from 7.9% to 25.2% among adolescents, whereas prevalence rates of IA in the Middle East (ME) and Africa ranged from 17.3% to 23.6%. Asia had a higher variation of IA among young individuals and adolescents, ranging from 8.1% to 50.9% (Hassan et al., 2020). In developed countries, internet penetration is 81% among the population, compared with 40% in developing countries and 15% in the least developed countries (International Telecommunication Union, 2016).
In 2016, some experts reported that IA had become a serious issue in Iran (Shajari et al., 2016) and there was internet overuse among different groups of Iranian cohorts (i.e., adolescent, youth and adults; Shajari et al., 2016). Iran is one of the developing countries where internet and social networks have markedly increased (Yarahmadi et al., 2020). In first quarter of 2017, the use of internet tripled, and more than 67.6 million Iranians in the population (80.5%) were using internet and social media sites (Internet World Stats, 2019). Consequently, increasing numbers of young and adult individuals in Iran are believed to be addicted to internet use.
Addiction is a complex behaviour and hard to identify particularly if it is a behavioural addiction (Griffiths & Larkin, 2004). Acquisition, development and maintenance of addictive behaviours result from the interplay between many different interacting factors (M. D. Griffiths, 2005) including the psychological and biological factors related to the individual including beliefs, attitudes, personality traits, genetic predispositions, and unconscious motivations, as well as the social environment in which the individual lives and the activity itself (Parke & Griffiths, 2006). The phenomenon of IA has been debated since the mid-1990s (M.D. Griffiths, , 1995Griffiths, , , 1996 and has often been used interchangeably with terms such as 'problematic internet use' (PIU; Caplan, 2002), 'internet addiction disorder' (IAD; Jiang, 2014), 'pathological internet use' (Davis, 2001), 'excessive Internet use' (Douglas et al., 2008), 'internet dependency' (ID; Mafe & Blas, 2006), and 'compulsive internet use' (CIU; Meerkerk et al., 2009). The present authors view IA as a more extreme manifestation of PIU (i.e., all internet addicts are problematic internet users but not all problematic internet users are internet addicts).
In general, these terms share conceptual notions of underlying pathology, although it is related to specific diagnostic criteria (Petry et al., 2014). Several diagnostic criteria for IA have been proposed such as salience and total preoccupation with the internet, obsessive thoughts about the internet, withdrawal symptoms when unable to use the internet, compromising of relationships, education and/or occupation because of excessive internet use, loss of interest in other activities due to spending too much time on the internet, and use of the internet to escape or relieve dysphoric mood states (M. Griffiths, 2000Griffiths, , 2005Tao et al., 2010). IA is also characterized by loss of control (e.g., impulse-control disorder and/or a behavioural addiction) and continued usage despite negative consequences (Shapira et al., 2003). One theoretical model that has received empirical support to explain PIU and IA is the cognitive-behavioural model (CBM; Davis, 2001;Montag et al., 2015) and associated components (Gámez-Guadix et al., 2013). These components comprise a preference for online social relationships, mood regulation through internet use, deficient self-regulation, and negative consequences. Also, the CBM drew a distinction between specific problematic internet use (SPIU) and generalized (GPIU) problematic internet use (GPIU) and identified the role of maladaptive cognitions (MCs) in the development of pathways to GPIU.
According to Davis' model, IA is related to maladaptive cognitions and social issues (such as obsessive thoughts or cognitive distortions,

Key Practitioner Messages
• Problematic internet use and internet addiction (IA) are an issue of concern among youth.
• Early maladaptive schemas (EMSs) are important contributory factors in IA.
• Practitioners could use schema therapy in helping to overcome IA.
• Practitioners should target specific EMSs (disconnection/ rejection and impaired autonomy/performance). depression, loneliness, social isolation, and lack of social support).
Maladaptive cognitions lead an individual to believe that through online existence, they can realize the reinforcement and social contact lacking in their offline lives. The need for social contact and reinforcement obtained online produces an increased desire to remain within a virtual social world (Davis, 2001). Furthermore, MCs describe those distortions in thoughts about the self or the world which bring about ruminations and constant thinking that relate to the individual's internet use (Davis, 2001). SPIU is viewed as the overuse of specific online activities such as online gambling (M. Brand, Young, & Laier, 2014), online pornography use and online gaming (M. Brand et al., 2016;Griffiths et al., 2016). Moreover, the behaviour is assumed to be the consequence of pre-existing psychopathology that is related to online activity. Research has identified that individuals who are psychologically vulnerable are especially at high risk of using the internet in a problematic way (M. Brand et al., 2014Brand et al., , 2016. Previous studies have demonstrated the potential negative effects that arise from excessive use and abuse of the internet on psychological health (H.M. Pontes & Griffiths, 2017) as well as physical consequences such as vision problems, decreased physical fitness and obesity, musculoskeletal problems, sleep deprivation, etc. Individuals with IA exhibit similar signs and symptoms to those of more established addictions (e.g., addiction to drugs, alcohol and gambling) and that IA has significant negative effects on individuals' daily lives (H.M. Pontes & Griffiths, 2017). In addition, IA appears to be more prevalent among the adolescents and emerging adults and can accompany or play a role in other psychiatric disorders (Öztürk et al., 2015).
Moreover, recent studies have noted that IA (compared with those without IA) (i) impacts negatively on adolescents' psychological well-being (e.g., low self-esteem), (ii) is associated with psychosocial risk (e.g., low perceived social support; Feng et al., 2019;Sharma & Sharma, 2018;Yavuz, 2019), and (iii) is associated with greater difficulty in communication, as well as cognitive and emotion regulation.
Many studies have indicated that among several factors associated with IA, personality traits stand out as one of the most important factors (Chang et al., 2019;Shi & Du, 2019). Individuals with personality disorders (such as obsessive-compulsive, schizoid, antisocial, borderline, histrionic, narcissistic, and schizotypal disorders) show higher levels of IA than individuals without such personality disorders (Zadra et al., 2016). Research into has found that individuals with Cluster B personality disorders and IA is more likely due to impulsivity and that individuals Cluster C personality disorders and IA is more likely due to low self-esteem (Zadra et al., 2016). Moreover, other studies have reported a relationship between IA and borderline personality disorders (Choudhary & Gupta, 2020).

| Early maladaptive schemas
One factor that may influence the aetiology of IA is early maladaptive schemas (EMSs). Such wasted time interferes with daily functioning and/or the carrying out of responsibilities, resulting in psychological pressures. Among other influential psychological variables in an individual's life, EMSs are of a great importance (Harris & Curtin, 2002). EMSs operate at the deepest level of cognition and are defined as broad, extremely stable and enduring themes or cognitive patterns regarding the individual self and an individual's relationships with others (J.E. . The concept of a schema is grounded in Beck's (1987) cognitive theory, which noted that maladaptive schemas can lead to a vulnerability to psychological distress (directly or indirectly). Schemas concern what to expect from experiences and situations and how to behave with respect to the self, others and the world (Baldwin, 1999). As described by J.E. , schemas are formed from early life and affect individuals throughout their lifespan and (when maladaptive) are dysfunctional to a significant degree. Individuals with noxious and negative schema distort their view of events in their lives in order to maintain the validity of their schemas (K.S. . Furthermore, EMSs remain dormant until they are activated by situations relevant to that particular schema (Idrissi et al., 2018). However, EMSs are not necessarily activated at every moment (H.M. Nordahl & Thimm, 2012). Typically, EMSs are activated by unpleasant and negative situations (Barazandeh et al., 2016) which can lead to negative mood states (such as depression and anxiety), interpersonal difficulties, dysfunctional reactions and emotional disorders (Flink et al., 2018;Kunst et al., 2020).
J.E.    Spending time on the internet may be an escape from feelings that would otherwise be emotionally painful (Bakhshi Bojed & Nikmanesh, 2013). Moderate and high internet use could be employed as a way to cope with or relieve psychological distress symptoms (such as depression and anxiety; Dalbudak et al., 2013).
Excessive internet use and maladaptive behaviours may also result in further psychological symptoms (Tonioni et al., 2012). Davis (2001) argued that GPIU is rooted in proximal contextual factors (e.g., specific maladaptive cognitions) and that poor psychosocial health was an important factor contributing to GPIU.
Individuals with this problem may use the internet to forget about their responsibilities and have problems with procrastination. Moreover, anxiety and depression may serve to mediate the relationship between IA and mental aspects of life such as abandonment and social isolation. In schema therapy, EMSs and maladaptive coping responses are proposed to underlie personality disorders and some recurrent or chronic disorders (e.g., depression, anxiety and substance abuse), as well as milder psychological distress (e.g., poor relationship or problems at work; J. E. Young, 1990). Flink et al. (2018) also found that EMSs were associated with a broad range of psychiatric diagnoses and psychological distress, particularly personality disorders.
Likewise, when EMSs become active, they can lead to several cognitive and behavioural problems such as anger symptoms, neuroticism, social phobia, low academic achievement, substance abuse, addiction, interpersonal conflicts, and psychosomatic disorders (Estévez et al., 2019;Gómez-Odriozola & Calvete, 2020;Kunst et al., 2020;Vasilopoulou et al., 2020). Lacking adaptive strategies with the negative affect inherent in EMSs, individuals have difficulties regulating negative emotions and, over time, might perceive them as intolerable (Eslampanah et al., 2020;Marengo et al., 2019). J. E.  posit that EMSs often generate high levels of negative affect and self-defeating behavioural patterns, which are believed to underlie the development of a range of mental health problems.
However, individuals with positive schemas have more capabilities in coping with mental issues, and when they encounter stressful events, they suffer less from mental disorders and drug dependencies (Saddichha et al., 2012).
Recently, studies specifically targeting the role of MSs on many risky behaviours have been carried out (e.g., drug abuse, addiction, heavy drinking, sexual issues, and illegal behaviour), particularly among women (Marengo et al., 2019). The research to date has noted that individuals drinking alcohol can be a means of coping with the suffering triggered and sustained by MS symptoms (Chodkiewicz & Gruszczy nska, 2018) and that the schemas concerning female substance use tend to exhibit more pronounced dysfunctionality than those of male substance use (Shorey et al., 2015). Shaghaghi et al. (2011) noted that MSs among substance users are higher than healthy individuals. Research examining alcohol dependence has shown that most individuals with alcohol use disorder have more MSs in comparison to those without the disorder (Karami et al., 2015). An accumulating body of research has established that drug abuse (Riso et al., 2006) and IA (Bakhshi Bojed & Nikmanesh, 2013) are among the coping strategies that individuals use to avoid negative effects of the activated EMSs. Therefore, a significant number of individuals with IA have some psychopathology and dysfunctional beliefs that result in excessive internet use (Taymur et al., 2016). Hypothesis schema domain and the specific EMSs 2: Impaired autonomy/performance has a positive association with internet addiction Bojed & Nikmanesh, 2013). Recently, Arpaci (2019) examined the effect of EMSs on smartphone addiction and found a positive relationship between individuals with EMSs and smartphone addiction. It has also been found that borderline personality disorder (BPD) has a relationship with more severe endorsement compared with major depression (MD), particularly in the disconnection/rejection domain (Barazandeh et al., 2016), and that the severity of these EMSs has been suggested to be specific to BPD (Cohen et al., 2016). Zamirinejad et al. (2018)

| THEORETICAL BACKGROUND AND RESEARCH HYPOTHESES
The present study developed a theoretical model based on cognitive theories. More specifically, it utilizes schema theory (ST; J. E.  and the cognitive-behavioural model of PIU (Davis, 2001) to explain psychological mechanisms underlying IA. Figure 1 shows the theoretical model of the relationships between early maladaptive schemas and IA utilized in the present study (variance and latent models are presented without item parcels or error terms, for ease of understanding

| Internet Addiction Scale
The IAS comprises 20 items and contains four subscales: lack of control (CON: 'How often do you find yourself saying "just a few more The total score is calculated by summing up the scores of the 20 items ranging from 20 to 100. Here, scores of 20-49 = no IA; 50-79 = mild IA, and 70-100 = severe IA. In the present study, the Cronbach's alpha and omega coefficient (ω) were .93 and .94 respectively (Young, 1998).  There are many different methods that can be used to examine relationships between constructs when the analysis comprises several exogenous (independent) and endogenous (dependent) variables, as in the present study. However, the best method is using a structural equation model (SEM; García-Alcaraz et al., 2014). Partial least squares path modelling (PLS-PM) is a useful method to identify the relationships and to analyse the influence of the different aspects of the complex analysed. Using the second order or higher-order constructs helps researchers to extend the PLS-PM to more advanced and complex models. The PLS-PM approach allows researchers to estimate and test the higher-order in reflective and formative type models (Crocetta et al., 2020). The analysis was carried out to identify the relationships among these variables to test the hypotheses. The proposed framework included nine exogenous constructs and four were characterized as an endogenous composite construct.

| RESULTS
Initially, by using the expectation maximization (EM) method from SPSS 26, there were no missing data found. The normality of data distribution was checked based on skewness and kurtosis values. The high value of kurtosis influences the result of analysis and the data need to have variability to ensure the partial least square methods is valid to apply. The skewness and kurtosis values were larger than ±2 (Ryu, 2011). In addition, the histograms indicated the data were relatively skewed. Specifying the measurement and structural models, the PLS-SEM method using a path-weighting scheme was conducted to assess the reliability and validity of the construct measures in the inner model. In the present study, the model included 65 items describing thirteen constructs (including two domains out of five domains; nine schemas out of 15 schemas). The two domains were The outer loadings of all indicators of constructs ranged between .71 and .93 and were significant ( value, therefore no factors (F1-F13) were deleted. Consequently, the measurement model was shown to have convergent validity (see Table 6).
Moreover, J. Hair et al. (1998) Table 4 shows that all the values on the diagonals were higher than the threshold value, adequate that DV was established (J. F. Hair et al., 2011) and that DV was confirmed and statistically sufficient to support the model in the present study.
There is also another approach to assess the discriminant validity as posited by J. Henseler et al. (2015) through multitrait and multimethod matrix, namely the Heterotrait-Monotrait Ratio (HTMT). Two methods of using the HTMT value can be used to test the discriminant validity: (i) when using it as a criterion, if HTMT value is greater than 0.85 (J. Henseler et al., 2015;Kline, 2011) or 0.90 (Gold et al., 2001), then there is a problem with discriminant validity, and (ii) when using statistical test to assess the HTMT ratio inference, if the confidence interval of HTMT ratio for the structural paths contains the value of 1, it indicates a lack of discriminant validity (J. Henseler et al., 2016). If the value of 1 falls outside the interval's range, it suggests that the constructs are empirically distinct. The HTMT ratios of correlation results are shown in Table 5. Based on the results in Table 5, all HTMT values were lower than the required threshold value of HTMT .85 suggested by Kline (2011). This indicates that discriminant validity was acceptable for the present study. The results demonstrate that neither lower or higher confidence intervals included a value of 1.   (Blissett et al., 2006;Simard et al., 2011), personality pathology (B. Bach et al., 2016;B. Bach et al., 2017), play a mediating role between childhood adversities and personality disorders (Carr & Francis, 2010) Based on the results, all constructs fulfil the guidelines of the confidence test that is based on composite reliability (CR) greater than 0.70 (J. F. Hair et al., 2017) and Cronbach's alpha greater than 0.70 (Nunnally & Bernstein, 1994). The results of the analysis of convergent (AVE) and discriminant (HTMT) validity for each con-

| CONCLUSION
IA is a growing problem in the field of addictive behaviour. The present study offers further empirical evidence that there is a relationship between early maladaptive schemas and internet addiction. A better knowledge of the relationship between EMSs and IA can lead to better understanding of the aetiology and pathogenesis of this addictive behaviour. It provides new insights into the development of specific psychological and psychosocial approaches aimed at the prevention and treatment of IA. Moreover, the findings suggest that specific schemas, namely, abandonment, mistrust/abuse, emotional deprivation, social isolation/alienation, defective/shame, dependence/incompetence, vulnerability to harm and illness, enmeshment/undeveloped self, and failure with internet addiction. These results are consistent with J. E. Young et al.'s (2003) conceptualization framework of the association between EMSs with chronic psychological distress, as well as preliminary evidence of differential patterns of cognitive biases.
The present study has a number of limitations that provide opportunities and challenges for future research. First, the risk of internet addiction was assessed at only one time-point over a relatively short time.
Therefore, long-term developmental tendency of the behaviour can-