Research has suggested that negative emotions are associated with problematic Internet use. This study investigated the role of metacognitions as a mediator of the relationship between negative emotions and problematic Internet use. A sample of 97 university students completed the following questionnaires: Hospital Anxiety and Depression Scale, Boredom Proneness Scale, Metacognitions Questionnaire 30, and Internet Addiction Test. All dimensions of metacognition were found to be positively and significantly correlated with problematic Internet use. Positive and significant correlations were also observed between problematic Internet use and negative emotions (anxiety, depression and boredom). Structural equation modeling was used to test a mediation model in which negative emotions predicted metacognitions which in turn predicted problematic Internet use. The results supported the hypothesis that the relationship between negative emotions and problematic Internet use is fully mediated by metacognitions, suggesting that metacognitive theory may be relevant to understanding problematic Internet use. The implications of these findings are discussed.

Keywords: Anxiety; Boredom; Depression; Metacognitions; Negative emotions, problematic Internet use

Introduction

The last decade has seen a world-wide surge in Internet use. This has been accompanied by a growing concern over what researchers have variously termed problematic, pathological or addictive Internet use (e.g. [Beard and Wolf, 2001], [Scherer, 1997], [Wallace, 1999] and [Young, 1998a]). Problematic Internet use (PIU) can be defined as “use of the Internet that creates psychological, social, school and/or work difficulties in a person’s life” (Beard & Wolf, 2001, p. 378).

Anecdotal and research evidence suggests that PIU affects a significant number of people with incidence rates (determined by frequent or daily Internet usage meeting varying criteria for pathological use) ranging between 4.6% and 18.3% ([Kaltiala-Heino et al., 2004], [Morahan-Martin & Schumacher, 2000] and [Niems et al., 2005]). These figures exceed the estimated incidence rates for other impulse control disorders, the category of disorders which PIU has most often been compared to ([LaRose et al., 2003] and [Shapira et al., 2003]). Pathological gambling, for example, is estimated to affect 1–3% of the US population and other impulse control disorders have even lower incidence rates (American Psychiatric Association, 2000).

A variety of studies have investigated the relationship between PIU and psychological dysfunction, revealing that anxiety and mood problems are common among problematic Internet users ([Kim et al., 2006], [Morgan and Cotten, 2003], [Sanders et al., 2000], [Whang et al., 2003] and [Young and Rogers, 1998]). Additionally, negative emotions appear to differentiate problematic from non-problematic Internet use with a reduction of negative emotions driving problematic use, and an enhancement of positive emotions driving non-problematic use (Wan & Chiou, 2006). As such, using the Internet may represent a way of coping with real life problems that takes the form of problem solving, entertainment and challenge in non-problematic use, and of reassurance and avoidance in problematic use. If this were the case then PIU could be conceptualized as a specific form of maladaptive self-regulatory strategy (LaRose et al., 2003).

The emergence of cognitive theories of psychological dysfunction (e.g., Beck, 1976) has led to a growing interest in the characteristics of cognition and its regulation. Recent work by Davis (2001), applying Beck’s (1976) cognitive therapy paradigm, has highlighted the potential role of cognitions in the development and maintenance of PIU. More recently, a particular line of theoretical work has begun emphasizing the role of metacognitions as central to the development and persistence of psychological dysfunction ([Wells, 2000], [Wells and Matthews, 1994] and [Wells and Matthews, 1996]).

Metacognitions refer to the information individuals hold about their own cognition and internal states, and about coping strategies that impact on both ([Flavell, 1979], [Wells, 2000], [Wells and Matthews, 1994] and [Wells and Matthews, 1996]). Examples of information individuals hold about their own cognition and internal states may include beliefs concerning the significance of particular types of thoughts, e.g. “It is bad to think thought X” or “I need to control thought X”. Examples of information individuals hold about coping strategies that impact on cognition and internal states may include beliefs such as “Smoking will help me get things sorted out in my mind” or “Drinking will help me solve the problem”. According to the metacognitive theoretical tenet ([Wells, 2000], [Wells and Matthews, 1994] and [Wells and Matthews, 1996]) metacognitions are fundamental in predisposing individuals to develop response patterns to thoughts and internal events that are characterised by heightened self-focused attention, recyclic thinking patterns, avoidance and thought suppression, and maladaptive behaviors. One such maladaptive behavior may be indeed PIU.

The role of metacognitions has been explored using the Metacognitions Questionnaire ([Cartwright-Hatton and Wells, 1997] and [Wells and Cartwright-Hatton, 2004]), which was developed to assess metacognitions related to psychopathology. The questionnaire consists of five distinct sub-scales: (1) positive beliefs about worry, measuring the extent to which a person believes that perseverative thinking is useful; (2) negative beliefs about worry concerning uncontrollability and danger, which assesses the extent to which a person thinks that perseverative thinking is overwhelming and dangerous; (3) cognitive confidence, which assesses confidence in attention and memory; (4) beliefs about the need to control thoughts; and finally (5) cognitive self-consciousness, measuring the tendency to monitor one’s own thoughts and focus attention inwards.

Support for the link between individual dimensions of metacognition and psychological dysfunction has come from a wide range of studies that show cross-sectional and directional relationships between dimensions of metacognition and psychological disturbance including, depression (Papageorgiou & Wells, 2003), generalised anxiety disorder ([Wells, 2000] and [Wells and Matthews, 1994]), hypochondriasis (Bouman & Meijer, 1999), obsessive–compulsive symptoms ([Emmelkamp and Aardema, 1999], [Hermans et al., 2003], [Myers and Wells, 2005] and [Wells and Papageorgiou, 1998]), pathological procrastination (Spada, Hiou, & Nikcevic, 2006), pathological worry (Wells & Papageorgiou, 1998), post-traumatic stress disorder (Roussis & Wells, 2006), predisposition to auditory hallucinations ([Baker and Morrison, 1998] and [Morrison et al., 2000]), problem drinking ([Spada et al., 2007], [Spada and Wells, 2005], [Spada and Wells, 2006] and [Spada et al., 2007]), psychosis (Morrison, French, & Wells, 2007), smoking dependence ([Spada et al., 2007], [Spada et al., 2007] and Spada et al., 2007 M.M. Spada, M. Zandvoort and A. Wells, Metacognitions in problem drinkers, Cognitive Therapy and Research 31 (2007), pp. 709–716. Full Text via CrossRef[Spada et al., 2007]), and test-anxiety ([Matthews et al., 1999] and [Spada et al., 2006]).

To date, no attempt has been made to investigate the possible links between PIU, metacognitions and negative emotions. Two lines of reasoning suggest that metacognitions should be associated with PIU. Firstly, PIU has been conceptualized as a means for regulating negative cognition and affect, and as such it represents, at least in part, a strategy that should be driven by metacognitions. Secondly, negative emotions are associated with PIU, and according to metacognitive conceptualizations of psychological dysfunction, emotional disturbance should be linked to maladaption in metacognitions.

In this study, we tested a mediation model in which negative emotions predict metacognitions which in turn predict PIU. Because all the variables in the model were expected to be highly correlated and measured with large errors, the test was performed using structural equation modeling.