Construct validation of the Use, Abuse and Dependence on the Internet inventory

Abstract

This study aims to validate the structure of the Use, Abuse and Dependence on the Internet (UADI) inventory on a sample of 1056 high school (n = 820) and university (n = 236) students. In particular, we performed exploratory and confirmatory factor analyses on sub-samples (cross-validation), tested the construct convergent validity, and tested correlations of UADI dimensions with HEXACO personality inventory (HEXACO-PI, where HEXACO stands for the six assessed dimensions: Honesty–Humility, Emotionality, eXtraversion, Agreeableness, Conscientiousness and Openness to experience) and external criteria (daytime internet use, etc.). Five dimensions (Compensatory Escape, Dissociation, Real Life Impact, Experience Making and Addiction) were best captured by a second-order factor structure with a factor reflecting Real Life Impact and Dependence saturated by the other four dimensions. This latter factor converged with other traditional measures of internet dependence. The results are discussed in light of the relationships with validity criteria.

Introduction

The diffusion of the internet represents a big innovation and a sort of revolution in many contexts (media communication, work, social life, etc.). Its daily use has greatly increased and become increasingly widespread. Although the internet has great value for modern society, the problem of abnormal internet use has also developed. This tendency has been labeled in several ways, including internet addiction, internet addiction disorder, internet dependence and problematic, pathological, excessive or compulsive internet use (Chang and Man Law, 2008, Widyanto and Griffiths, 2007); however, it is not yet unanimously defined. Young (1999) claimed that internet addiction includes many different behaviour and impulse control problems (e.g., cybersexual addiction, cyber-relationship addiction, net compulsions, information overload and computer addiction).

Byun et al. (2009) conducted a meta-synthesis of research on internet addiction and found that researchers provided different, incomplete or often inaccurate definitions. Starting from Beard and Wolf’s (2001) suggestion (and even inviting authors to work towards developing a coherent and shared definition), Byun et al. (2009) proposed the following guideline: “an individual is addicted when an individual’s psychological state, which includes both mental and emotional states, as well as their scholastic, occupational and social interactions, is impaired by the overuse of the medium” (p. 204).

Since psychologists and researchers began focusing on internet addiction, several tools for assessing this construct have been proposed. Most of them rest on DSM-IV (APA, 1994) criteria for addiction, including the Internet Addiction Test (IAT) (Young, 1998; see also Beard & Wolf, 2001), the Internet-Related Addictive Behaviour Inventory (IRABI) (Brenner, 1997), the Internet Addictive Disorder (IAD) (Goldberg, 1996) and many others (Anderson, 2001, Kubey et al., 2001, Pratarelli et al., 1999). Lacking a shared definition, characteristics of these measurements often differ; for example, each one assesses different sub-dimensions of internet addiction (e.g., online preoccupation; social interactions, comfort or benefits; diminished impulse control; loneliness/depression; distraction; mood alteration; negative outcomes; compulsive use; excessive time online; withdrawal; tolerance; and interpersonal, familial, work-related, social and financial problems).

Chang and Man Law (2008) categorized these sub-dimensions into four macro-areas: compulsive internet use, excessive time spent online and failure to control it; withdrawal symptoms when being restricted from internet use; using the internet for social comfort; and negative social, academic or work consequences related to internet use. In addition, Griffiths (2000) identified the following problems: measures failing to assess disorder severity or a temporal dimension; measures tending to overestimate the problem incidence; and measures tending to neglect the context in which the internet is used. Furthermore, Chang and Man Law (2008) argued that these measures often test different sub-dimensions, lack psychometric validation, and are largely based on exploratory analyses. This strategy may have proved useful when internet addiction was first being studied, as it helped researchers gain a thorough knowledge of the phenomenon; however, a more coherent and comparable corpus of instruments is now needed. The authors were involved in a project devoted to exploring and confirming the structure of one of these instruments, namely the Internet Addiction Test (IAT). Their findings show that IAT items test three dimensions: “Withdrawal and Social Problems,” “Time Management and Performance” and ‘‘Reality Substitution.”

Del Miglio, Gamba, and Cantelmi (2001) proposed a scale to assess internet-correlated diseases called the Use, Abuse and Dependence on Internet (UADI) inventory. The UADI aimed to detect adaptive and maladaptive features of the internet usage. The Internet has remarkably positive functions, including serving as a supplementary instrument adaptive to reality, effective for self-realization, experimentation and research, but also has the capability to induce addiction.

The UADI is composed of five dimensions: Compensatory Escape, Dissociation, Real Life Impact, Experience Making and Addiction. Compensatory Escape measures the extent to which individuals use the internet to avoid daily difficulties, compensating for these difficulties through mood regulation, personal self-efficacy and online social relationships. Dissociation describes the presence of alienation and dissociative symptoms, such as peculiar sensations, depersonalization or derealization. Real Life Impact measures the consequences of internet use on daily life, including the modification of previous habits, social relations or mood. Experience Making concerns the extent to which individuals use the internet as tool to experiment with the self, including social experiences and searching for and testing emotions. Finally, Addiction analyzes symptoms of dependence, including tolerance, withdrawal and compulsiveness.

In the study (Del Miglio et al., 2001), the five cited dimensions were extracted through a rotated factorial analysis and had adequate internal consistency (.76 < α < .93), but the correlations among them are not reported. The authors asserted that the Real Life Impact represents the fundamental missing area (not assessed by Dependence items) to test, analyze and diagnose addiction, because it includes the consequences of internet use on the social and occupational sphere.

The UADI scale was conceived and partially validated in Italy (Del Miglio et al., 2001). The object of the present study is to carry out a more complete validation process, testing convergent and criterion validity of UADI scale. This validation process includes conducting a confirmatory analysis on different sub-samples (cross-validation) and testing its correlation with scales assessing the same concept, with personality traits, and between each UADI dimension and other external criteria often considered to be related to internet addiction, such as the amount of time spent online, holding and using other technological devices, etc. In social science and psychometrics, indeed, construct validity refers to the ability of a measure to assess the psychological construct which the scale aims to measure, namely it tests whether the scale actually assess the concept for which it has been conceived. The procedure to test construct is based on the idea that scales assessing the same construct or construct linked to the to-be-assessed one must show moderate or high correlation values, whereas concepts far from the considered one must show low correlations (Campbell, 1960). To test UADI validity, we refer to other scales assessing internet addiction (convergent validity) and to different variables linked to internet addiction (criterion validity).

Furthermore, the relationships between the UADI dimensions and the HEXACO 6-factor personality test (Lee & Ashton, 2004) will be investigated, in order to test discriminant validity and the association between internet addiction and different personality traits (the acronym stands for the six assessed dimensions: Honesty–Humility – H, Emotionality – E, Extraversion – X, Agreeableness – A, Conscientiousness – C, Openness to experience – O). Studying the association between specific patterns of personality and the tendency to develop an internet addiction provides information about the potential greater vulnerability of some clusters of people rather than others. The HEXACO personality test is an instrument validated across several cultures and languages (Ashton et al., 2006). It measures personality traits clustered in six areas covering dimensions beyond the traditional Big Five factors (Lee, Ogunfowora, & Ashton, 2005); one dimension is added to the traditional Big Five questionnaire assessment of personality. Three of the dimensions are quite similar to those measured by the Big Five: Conscientiousness, emphasizing orderliness, work effort and impulse control; Emotionality, concerning anxiety, emotional reactivity, sentimentality and lack of courage; and Openness to Experience, measuring intellectuality, imagination and unconventionality. Two factors represent a variation of Big Five traits: Extraversion, concerning sociability, talkativeness and liveliness (only secondarily including bravery, self-assurance and toughness, which is associated with this dimension on the Big Five questionnaire) and Agreeableness, indicating gentleness, good-naturedness, tolerance and patience (and also including, in contrast with its traditional Big Five definition, even-temper and lack of irritability). The sixth dimension is labeled Honesty/Humility and emphasizes trustworthiness, modesty, lack of greed and lack of slyness (Ashton et al., 2004).

These dimensions are considered relevant to the investigation of the relationship between personality characteristics and the development of an internet addiction. The comorbidity of internet addiction and several other variables, including psychological and psychiatric disorders, has been investigated (Black et al., 1999, Griffiths, 2000, Shaffer et al., 2000, Shapira et al., 2000); however, few studies have considered the relationship between internet addiction and specific personality characteristics. Those that have investigated this relationship only marginally have included personality characteristics. For example, Young and Rodgers (1998) found that dependent internet users were characterized by high scores on self-reliance, emotional sensitivity and reactivity, vigilance and non-conformism. Petrie and Gunn (1998) also found that people who define themselves as addicted to the internet were more likely to score highly and positively on depression and introversion.

Other studies conducted on internet usage in general pay no particular attention to addiction but are still relevant to this topic. Yang and Lester (2003) found that extraversion was positively associated with internet use, whereas neuroticism was negatively associated. Other authors (Engelberg and Sjöberg, 2004, Hills and Argyle, 2003, Jackson et al., 2003) have found no association between personality and internet usage, no differences in personality associated to the chosen internet functions (Everton, Mastrangelo, & Jolton, 2005) or to the consequences of internet usage (Kraut, Kiesler, & Boneva, 2002). However to date, we have no knowledge of any research that has analyzed the correlations between basic personality traits and different dimensions of internet use and abuse.

Finally, as Byun et al. (2009) pointed out, the participants of many studies about internet addiction have been high school students (adolescents) or university students, two populations that seem to be most exposed to the risk of internet dependence. No study yet has tried to confirm either the presence of a factorial structure shared by the two populations or to validate a scale considering the dimensions as potentially diverse.

The present study aims to explore and confirm the UADI scale structure in a sample of Italian high school students and a sample of university students. The resulting dimensions are then tested for reliability and convergent validity. Finally, correlations between these dimensions and HEXACO personality traits, along with many criteria and socio-demographic variables, are examined.