A conceptual and methodological critique of internet addiction research: Towards a model of compensatory internet use Daniel Kardefelt-Winther

Tag layer of the DHSS Meta-Tagging Project. The article text is reproduced from the staged copy named in sourceDesc and is not altered; the tags are held separately in standOff so that overlapping and nested spans are preserved.

 A conceptual and methodological critique of internet addiction research: Towards a model of compensatory internet use 2014 Computers in Human Behavior 10.1016/j.chb.2013.10.059 kardefeltwinther2014-critique-internet-addiction-research kardefeltwinther2014-critique-internet-addiction-research.txt, the project's staged copy 8f330849d5762c2556a99921838632c1bfa2f481 repository-deposit-not-redistributable media-and-communication-studies

the paper's own words

the tag layer; a sentence written by the tagger about the paper, always anchored to a quotation but not itself a quotation

 46 spans located by searching the stored text for the quote's exact words, allowing any whitespace between them internet addiction research itself - its conceptual foundations and its reliance on direct-effects models why a field with so much collected data has produced so little theory, and whether compensation rather than compulsion better explains excessive internet use proposes the compensatory internet use model - the internet is used to alleviate a negative life situation - and argues the field must move from direct effects to mediation and interaction between psychosocial well-being and motivations critique people continuing to spend heavy time online despite experiencing outcomes that harm them conceptual-analysisconceptual-analysis a lack in real life produces a motivation to go online, an application affords the missing thing, and the same act yields both the relief and the damage motivations MEDIATE between psychosocial well-being and excessive use - which is why a direct-effects design cannot see the process the harm scales with how much compensation the person needs, so a permanent real-life difficulty implies a permanent demand the model's testable prediction: escapism should predict problematic outcomes MORE strongly in high-stress people than low-stress ones the compensatory model is proposed and argued here, not tested - the paper asks for the empirical work rather than reporting it the paper leaves open whether the phenomenon should be called addiction at all the field it critiques cannot isolate causes because nearly everything tested came out significant the compensation idea is old and repeatedly stated; what is missing is that almost nobody has tested it Kubey and Csikszentmihalyi's finding that people watch heavy TV in dysphoric states, carried across to the internet as self-medication the historical move from direct effects to mediating variables is proposed as the transition internet addiction research must now make Ingleby's argument that the governing principles of a field are philosophical rather than empirical, applied to why more findings have not produced theory controlling for stress when testing loneliness and social anxiety against excessive online gamingeliminates: the possibility that loneliness and social anxiety are independent predictors rather than artefacts of being tested alone · outcome: failed controlling for psychosocial well-being when relating gaming motivations to internet addictioneliminates: the possibility that the motivation-addiction association is spurious · outcome: held internet addictioninternet addiction compensatory internet usecompensatory internet use direct effectsdirect effects motivationsmotivations psychosocial well-beingpsychosocial well-being escapismescapism coping strategycoping strategy affordancesaffordances interaction effectsinteraction effects internet addiction - the pathological sense, an impulse-control disorder with criteria borrowed from pathological gambling internet addiction - the compensatory sense, a practical response to an unmet need that sometimes costs more than it returns internet addiction - the normative-shift reading, where the symptom is a new way of living that only has pathological interpretations available `internet addiction` - THREE COMPETING LABELS FOR THE SAME THING, DECLARED IN THE PAPER'S FIRST FOOTNOTE: `excessive internet use, compulsive internet use, problematic internet use - labels that have been used interchangeably to describe more or less the same concept`. And `more or less` is doing work: the paper's whole argument is that the concept is not stable. critiques → Young (1998), who built internet addiction on the DSM-IV criteria for pathological gambling contradicts → the direct-effects tradition in internet addiction research builds-on → Ingleby (1981) on epistemological issues in psychiatry builds-on → Wood (2008), that addiction concerns the interaction between individual, culture and environment builds-on → Kardefelt-Winther (2014), the author's own finding that escapism and achievement mediate stress and excessive gaming critiques → Young's edited volume (2011), offered as evidence that the field has explanations but no consensus direct effects —contrasted-with→ interaction effectsthe relation itself is separately evidenced · strength of the relation: 0.9 · the relation stated: denying a direct relationship is exactly what makes interaction the analytical focus Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but the model does not assume a direct relationship, so interaction effects are its focus; contrasted-with is loose.direct effectsinteraction effects loneliness and social anxiety stopped predicting excessive online gaming once stress was controlled for escapism and achievement mediated the relationship between stress and excessive online gaming nearly every psychosocial factor tested came out a significant predictor, which is why none could be isolated as a cause adding the activity and the motivation to the model lets a result state WHY someone is online rather than only that they are at risk internet addiction as a lost-control state with problematic outcomesinternet addiction the basic tenet of compensatory internet use - the locus of the problem is the reaction, not the applicationcompensatory internet use the compensatory model stated as a proposal about negative life situations giving rise to a motivationcompensatory internet use

A conceptual and methodological critique of internet addiction research: Towards a model of compensatory internet use Daniel Kardefelt-Winther ⇑ Department of Media & Communications, London School of Economics and Political Science, WC2A 2AE London, United Kingdom a r t i c l e i n f o Article history: Available online 23 November 2013 Keywords: Internet addiction Compulsive internet use Problematic internet use Compensatory internet use Motivations for internet use a b s t r a c t Internet addiction is a rapidly growing field of research, receiving attention from researchers, journalists and policy makers. Despite much empirical data being collected and analyzed clear results and conclu- sions are surprisingly absent. This paper argues that conceptual issues and methodological shortcomings surrounding internet addiction research have made theoretical development difficult. An alternative model termed compensatory internet use is presented in an attempt to properly theorize the frequent assumption that people go online to escape real life issues or alleviate dysphoric moods and that this sometimes leads to negative outcomes. An empirical approach to studying compensatory internet use is suggested by combining the psychological literature on internet addiction with research on motiva- tions for internet use. The theoretical argument is that by understanding how motivations mediate the relationship between psychosocial well-being and internet addiction, we can draw conclusions about how online activities may compensate for psychosocial problems. This could help explain why some peo- ple keep spending so much time online despite experiencing negative outcomes. There is also a method- ological argument suggesting that in order to accomplish this, research needs to move away from a focus on direct effects models and consider mediation and interaction effects between psychosocial well-being and motivations in the context of internet addiction. This is key to further exploring the notion of internet use as a coping strategy; a proposition often mentioned but rarely investigated. 2013 Elsevier Ltd. All rights reserved. 1. Introduction Internet addiction1 is typically described as a state where an individual has lost control of the internet use and keeps using inter- net excessively to the point where he/she experiences problematic outcomes that negatively affects his/her life (Young & Abreu, 2011). Examples of such outcomes are cases where individuals lost sleep or skipped meals because they were spending time on the internet, or where internet use has resulted in conflicts with family members or led to the detriment of a job or educational career. Most research on internet addiction is based on initial research by Young (1998), who conceptualized internet addiction as an impulse-control disorder, deriving diagnostic criteria from the Diagnostic and Statis- tical Manual of Mental Disorders (DSM-IV) diagnosis for pathological gambling. Since addictions were not acknowledged in DSM-IV, Young contended that the diagnosis of pathological gambling was most akin to the pathological nature of internet use and adopting the criteria would be helpful in clinical settings and stimulate further research (Young, 1998). The subsequent empirical work has not been successful in terms of agreeing on a definition or the diagnostic criteria, nor in the explanations of what leads to or follows from internet addic- tion. Researchers have also been unable to agree on who is at greater risk, unable to agree on whether the problems are persis- tent and unable to determine whether the proposed methods for treatment are successful. Despite finding many associations be- tween psychosocial well-being and internet addiction researchers have been unable to agree on a general theory about the etiology. Traditionally, research on internet addiction has focused on direct effects models exploring the associations between psychological vulnerabilities and internet addiction. Studies have explored vulnerabilities such as depression (Kim et al., 2006), low self-esteem (e.g., Fioravanti, Dèttore, & Casale, 2012) and high sensation-seeking (Armstrong, Phillips, & Saling, 2000; Velezmoro, Lacefield, & Roberti, 2010; Widyanto & McMurran, 2004), loneliness and shyness (e.g., Caplan, 2002, 2003, 2005; Kim, LaRose, & Peng, 2009), locus of control and online experience (Chak & Leung, 2004), attention-deficit/hyperactivity/impulsivity symp- toms (Yoo, Cho, & Ha, 2004) and suicidal ideation (Kim et al., 2006). Studies have also explored the association with psychoso- cial well-being (e.g., Young & Abreu, 2011; Caplan, Williams, & 0747-5632/$ - see front matter 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.chb.2013.10.059 ⇑Tel.: +44 7946567850. E-mail address: d.a.kardefelt-winther@lse.ac.uk 1 Or excessive internet use, compulsive internet use, problematic internet use – labels that have been used interchangeably to describe more or less the same concept (Widyanto & Griffiths, 2006). Computers in Human Behavior 31 (2014) 351–354 Contents lists available at ScienceDirect Computers in Human Behavior journal homepage: www.elsevier.com/locate/comphumbeh

Yee, 2009; Lemmens et al., 2011; Van Rooij, 2011) as well as the association with various personality traits (e.g., Leung, 2007; Lo et al., 2005; Whang, Lee, & Chang, 2003), interpersonal skills and intelligence (Byun et al., 2009). This psychologically oriented approach to studying internet addiction has yielded plenty of statistically significant results. However, because most factors were found to be significant predic- tors it has not been possible to make any claims about unique risk factors which has made it difficult to isolate the causes behind internet addiction. Furthermore, as I discussed in a recent article (Kardefelt-Winther, 2014) the associations for both loneliness and social anxiety with excessive online gaming lost significance when stress was controlled for. This result cautions that vulnera- bilities posited as significant predictors of internet addiction may only be significant by virtue of being examined in isolation from other factors. A direct effects approach has not allowed researchers to explore the significant predictors of internet addiction while controlling for interactions with other influencing psychosocial conditions or mediating variables. Therefore, in terms of theory building the psychological approach has not contributed much to a better understanding of why some people keep using the internet despite experiencing problematic outcomes. The lack of theoretical development is evident in Young’s edited book (2011) where each chapter suggests different causes for internet addiction. Although it constitutes an important effort to summarize the existing research it leaves the reader with many possible explanations but no consensus. Considering the amounts of data that have been collected and the efforts made, the lack of progress indicates that there are issues somewhere along the way that makes theoretical development difficult. Ingleby’s (1981) review of epistemological issues in psychiatry suggests that researchers sometimes delude themselves that all that is needed for theoretical development is just ‘‘more findings’’. He further sug- gests that ‘‘the literature on mental disorders is quite out of propor- tion to the adequacy of our knowledge about them’’ (p. 23). What matters, Ingleby argues, are the fundamental principles which gov- ern the acquisition and interpretation of ‘‘findings’’; and these principles, although they are governed by matters of fact, are not themselves discovered empirically – they are as much philosophi- cal as scientific ones (p. 24). What is needed, then, is not more find- ings but a reappraisal of the kind of explanations we should be looking for. Following Ingleby’s ideas, there may be much to gain by considering alternative theories for internet addiction that do not only take the literature on mental disorders as its starting point. Early speculation by Young suggested that internet addiction may occur when the internet is used to cope with difficult real life situations (1998). This has been repeatedly mentioned in the liter- ature on internet addiction (e.g., Armstrong et al., 2000; Bessière, Kiesler, Kraut, & Boneva, 2008; Chak & Leung, 2004; Kim et al., 2009; Kuss, Louws, & Wiers, 2012; Shen & Williams, 2011; Whang et al., 2003; Widyanto & Griffiths, 2006; Young, 2009; Young & Abreu, 2011) but rarely empirically investigated. The tenet of Young’s speculation is that internet use has a propensity to allevi- ate dysphoric moods and may therefore be used to cope with or compensate for real life problems. Similar ideas about the compen- satory potential of media use was suggested in an early study by Kubey and Csikszentmihalyi (1990), who claimed that people were more likely to engage in bouts of heavy TV watching when they were in dysphoric states. Bessière, Kiesler, Kraut, and Boneva (2004) stated that if this logic applies to the internet as well, it would suggest that people who are feeling bad are using online entertainment as a form of self-medication (p. 31). Indeed, in later works Caplan and High (2011) also suggested that through the ex- change of online messages, users compensate for what they may lack in real life. In the context of internet addiction, Young and Abreu (2011) discussed whether an individual becomes addicted to Facebook because they are using Facebook to fulfill missing so- cial needs (p. 12). These recent discussions take the idea of com- pensatory internet use to a more detailed level where applications are assumed to have different compensatory potential depending on their affordances. However, while plenty of speculation has surrounded the idea of compensatory internet use few studies have empirically investi- gated the compensatory potential of internet applications in the context of excessive internet use. Crucially, researchers have not investigated whether a theory of internet addiction based on the idea of compensation may better explain why people spend so much time online that they experience problematic outcomes, compared to the dominant theory of internet addiction as compul- sive behavior and a mental disorder. While there is a theoretical basis for investigating internet addiction as mental disorder, no theoretical model exists to support research on compensatory internet use in this area. The lack of theoretical support may be one reason for why the idea of compensation is repeatedly men- tioned but rarely followed up empirically. It would also explain why the conceptualization of internet addiction as a mental disor- der is still dominating research despite the apparent shortcomings in terms of furthering the development of internet addiction theory (e.g., Widyanto & Griffiths, 2006). In this paper I will develop the claim that internet addiction can be usefully approached from a perspective of compensation rather than compulsion. Instead of the compulsive, pathological, nature that internet addiction is ascribed in the literature, I argue that it can be better understood as a coping strategy grounded in under- standable (but not always healthy) motivations. This follows on Wood’s (2008) observation that theories of addiction are increas- ingly moving away from a focus on the activity or substance as a causal factor and instead suggesting that ‘‘addiction concerns the interaction between the individual, their culture and their environ- ment’’ (p. 177). In his paper, Wood (2008) recommends a dose of healthy skepticism towards the idea of video game addiction, but maintains that a minority of people do play excessively. A model of compensatory internet use recognizes and seeks to understand this minority outside a framework of pathology and mental disor- ders. I will discuss how researchers can develop this theory by combining existing research on the psychological antecedents of internet addiction with knowledge from research concerned with the attractions and compensatory potential of the internet. 2. Towards a model of compensatory internet use This paper proposes a theory of compensatory internet use where negative life situations can give rise to a motivation to go online to alleviate negative feelings. The basic tenet of the theory of compensatory internet use is that the locus of the problem is a reaction by the individual to his negative life situation, facilitated by an internet application. As an example, if real life is character- ized by a lack of social stimulation the individual reacts with a motivation to go online to socialize which is facilitated by an appli- cation where socializing is afforded, such as an online game or a so- cial networking site. This can then have positive and negative outcomes: positive in the sense that the individual feels better be- cause he gets the desired social stimulation and negative because he may not go out and make new offline friends, which in the long run means he could become dependent solely on the internet for social stimulation. This scenario would be labeled as an internet addiction when approached through a pathological perspective, but has little to do with the compulsive nature of addictions. It is an understandable and practical way to acquire social stimulation when there is a lack of it (e.g., Chappell, Eatough, Davies & Griffiths, 352 D. Kardefelt-Winther / Computers in Human Behavior 31 (2014) 351–354

2006), but this habit may sometimes lead to negative conse- quences and addiction-like symptoms due to the amount of com- pensation required to alleviate negative feelings. For individuals with permanent real life issues, such as physical handicaps or dis- abilities, the need for compensation may be constant as case study evidence provided by Griffiths (2000) would suggest. For less se- vere cases, such as temporary school or work related stress, a few hours of compensatory internet use may be beneficial and lead to fewer problematic outcomes (e.g., Leung, 2007). With internet becoming ubiquitous in society, it is clear that some of the alleged ‘‘symptoms’’ of internet addiction can be inter- preted as a normative shift in how younger generations entertain or communicate and as a testament to the embeddedness of inter- net use in everyday life rather than pathological behaviour. Smahel and Blinka (as cited in Young & Abreu, 2011) have suggested that what is treated by researchers as pathological behaviour may be a new way of life for which researchers currently have only path- ological interpretations. Against this background, it seems all the more important to understand the contexts, purposes and motiva- tions for internet use, as these are likely to have a strong impact on the outcomes (Shen & Williams, 2011). Motivations for going online have been explored in some stud- ies on internet addiction, primarily in the context of online gaming addiction. Utilizing Yee’s (2006, 2007) framework for gaming moti- vations as a starting point, researchers have investigated whether motivations for playing online games are associated with internet addiction (e.g., Caplan et al., 2009; Kuss et al., 2012). Importantly, in the study by Caplan et al. (2009) psychosocial well-being was controlled for as they suspected that associations between psycho- social well-being or motivations for play and internet addiction may be spurious. Indeed, following on Caplan et al.’s (2009) study, Kardefelt-Winther (2014) demonstrated empirically that the moti- vations escapism and achievement mediated the relationship between stress and excessive online gaming. This suggests that motivations for play and psychosocial well-being may be usefully explored in conjunction rather than separately. There seems to be an opportunity here to combine the psychological approach with the motivations approach through the model of compensa- tory internet use. Theoretically, in the compensatory model the motivations for use are grounded in psychosocial problems or un-met real life needs. In terms of research operationalization, this can be tested by exploring whether the association between moti- vations and internet addiction vary depending on the level of psy- chosocial well-being. Methodologically, this is explored by interaction effects between psychosocial problems and potentially alleviating motivations for use. For example, people high on social anxiety may compensate for feelings of loneliness by socializing in a game or on a social networking site because online environments feel safer due to the sense of anonymity (McKenna, Green, & Gleason, 2002). In such a case, where the motivation to go online is grounded in an un-met real life need and where the internet use alleviates the real life problem, an individual may feel a strong desire to spend more time online which could lead to problematic outcomes. Whether this is what we wish to call internet addiction or not can be debated, as can the compulsive nature of such inter- net use, but to suggest that this is a mental disorder seems to be a stretch. 3. Conclusions This paper will conclude by summarizing the elements of the compensatory internet use theory and suggest some implications for interpretation of results, empirical work and further theory building. The psychological approach to internet addiction used in most studies consists of psychosocial vulnerabilities (1) and problematic outcomes of internet use (2). A typical conclusion from empirical research using this model is: ‘‘Internet users high on social anxiety (1) and loneliness (1) are at risk of neglecting schoolwork (2) and hav- ing conflicts with parents due to their engagement with the internet (2)’’. This paper suggests an inclusion of two additional elements that have been mentioned in this paper: the online activity and its affordances (a) and motivations for going online (b). Using this model researchers could describe the observed situation in greater detail: ‘‘a player of World of Warcraft (a) who wants to socialize (b) and chat (b) with other players, and is high on social anxiety (1), may be at increased risk of neglecting schoolwork (2) and of having conflicts with parents due to their engagement with the internet (2)’’. This provides an explanation for excessive use and negative outcomes without framing the behaviour as pathological. It allows researchers to understand what the user is using the internet for and interpret the problematic outcomes against the background of the motivations for going online and the real life context of the user. Essentially, it enables researchers to say something about why a person spends so much time online without resorting to speculation. This has been a missing component in most research to date because direct effects models are restrictive by nature and do not allow the researcher to consider the impact of other variables and therefore masks underlying processes that may be vital in explaining excessive use. Exploring motivations in conjunc- tion with psychosocial well-being allows us to elaborate on why someone goes online by contextualizing the motivation for exces- sive use in the presence of psychosocial problems. This affords a discussion of whether the internet use may be beneficial and understandable as an effective coping strategy, despite the occur- rence of problematic outcomes. As with any model, it is important to consider its empirical functionality in addition to the theoretical contribution. The model of compensatory internet use suggests that researchers need to empirically investigate the relationship between motiva- tions and psychosocial well-being in the context of excessive internet use. Is the use socially motivated – indicating loneli- ness? Achievement oriented – indicating frustration over lacking real life success? Is it about domination and competition – indic- ative of anger or a desire for accomplishment? Is it primarily escapist – indicative of a stressful real life situation? Because the model does not assume a direct relationship between psycho- social well-being and internet addiction, interaction effects be- tween the elements in the model constitute the analytical focus. Williams, Yee, & Caplan (2008) have argued that game re- search needs to undergo the same transition as early communi- cation research did, where the direct effects model was made more nuanced by the addition of mediating variables. I propose that a similar transition is needed for internet addiction research and that motivations for play could be a useful starting point. By considering why the user is motivated to go online in relation to their psychological and contextual reality it may be possible to determine if individuals use the internet to cope with distressing real life situations and provide the first step towards identifying these situations at the same time. As was mentioned earlier, Kardefelt-Winther (2014) found that the relationship between stress and excessive online gaming was mediated by escapism, a motivation for play that is often linked to problematic out- comes (Caplan et al., 2009; Kuss et al., 2012). This suggests that some people may play online games to escape from stress, which is associated with more negative outcomes. While this is not a surprising finding in itself, the broader implication is that when motivations are preceded by psychosocial problems, the risk for negative outcomes may be higher. This is the core idea of a model of compensatory internet use. D. Kardefelt-Winther / Computers in Human Behavior 31 (2014) 351–354 353

The next step for future research would be to explore interac- tion effects between, for example, stress and escapism in the con- text of excessive online gaming. According to a model of compensatory internet use, the relationship between escapism and problematic outcomes should be stronger for people with higher levels of stress, compared to those with lower levels of stress. Eventually patterns may be found where certain motiva- tions often precede problematic outcomes when the individual suf- fers from particular life difficulties. This would be a useful starting point for discussing preventive actions.