Effect Online Game

Running Head: Online Game Addiction

Perception of and Addiction to Online Games as a Function of Personality Traits

Searle Huh
University of Southern California and Nicholas David Bowman
Michigan State University

Online Publication Date: April 26, 2008
Journal of Media Psychology, V 13, No. 2, Spring, 2008

Abstract

With the growing popularity of online video games, there have been anecdotal reports suggesting that these games are highly addictive, with some gamers spending in excess of 40 to 50 hours per week playing. Thus, research into the individual characteristics that lead to excessive play is warranted. This paper examines two individual variables – personality and perceptions of media – and explores how they relate to online game play, specifically online game addiction. By presenting a revised metric for online game addiction, this paper explores the relationship between addiction and both personality and perception. Online addiction is presented in this paper as a process addiction with four unique factors: perceived social sanctions, excessive play, uncontrollable play, and displacement. Both personality and perception are found to be significantly associated with online game addiction. These results are interpreted and discussed, and future research direction is suggested.

Keywords: online game addiction, Big Five personality traits, media perception, MMOs
Perception and Addiction of Online Games as a Function of Personality Traits Computer games as a leisure activity have become an ever-increasing part of many young people’s day-to-day lives (Griffiths & Davis, 2005; Durkin, 2006). More recently, with the rapid diffusion of broadband Internet services and high-end graphic cards for computers and console systems, online video games – games played over certain online networks (primarily the Internet) – have become more popular and attractive than ever before (Sherry & Bowman, in press). According to a white paper from the Korea Game Industry Agency (2007), the world market for online video games increased from $ 2.1 billion in 2003 to $ 5.7 billion in 2006, representing a nearly three times market increase in less than half a decade. A recent AC Nielsen study reported that, of the 65 million active online gamers, over 15 million are over the age of 45 (as cited by Gonsalves, 2006), and over 64 percent – almost two-thirds – are female (as cited by Klepek, 2006). The same report found that, of the leisure time available to adolescents (about 55 hours per week), nearly 25 percent of this time was spent with video games (as cited by PRNewswire, 2006). In short, online gaming has swiftly emerged as a popular and successful source of entertainment and play for people of all ages. The majority of video game research has focused on perceived negative effects of video game play due to the content of the games, as social scientists have focused their efforts on investigating the proposed relationship between violent content and aggressive outcomes (Anderson & Bushman, 2001; Calvert & Tan, 1994; Jansz, 2005; Sherry, 2001a). Although online games often contain similar acts of violence, recent anecdotal evidence has suggested another negative behavioral effect that these games may pose, that of addiction. The Washington Post reports that, in 2005, at least 10 people in Korea died as a result of excessive game play, including one man who was found dead in an Internet café after allegedly playing for over 50 hours with few breaks (Khazan, 2006). Stories such as these have raised concerns from government agencies and citizens groups, who wish to better understand the dynamics of online game play, especially those variables that lead to online game addiction. The present study investigates the potential for personality traits – which have been shown to significantly predict media use patterns – to similarly predict individual’s perceptions of and dependency on online video games.
Attributes of Online Games Compared to traditional video games, online games have several distinguishable characteristics at both the technical and play levels. On a technical level, perhaps the most critical aspect of online games is that many people can play them through different online networks (Kim, Park, & Kim, 2002). Whereas the traditional video games only allow multi-player arrangements amongst a few co-located gamers, online games allow play between thousands of gamers located around the world simultaneously via the Internet. Whereas it would be nearly impossible to organize more than a few people in one room to play a particular game at one time simultaneously, online gamers need only to log in from any broadband Internet connection to play. Notably, massive multi-playability is just one of the advantages of online games. Although most online games that are popular for gamers and highlighted in the academia are MMORPGs (Massively Multiplayer Online Role-Playing Game, e.g, World of Warcraft) and MMOFPSs (Massively Multiplayer Online First-Person Shooter, e.g. Halo 2), our understanding of online games is not restricted to these. In fact, the only requirement for a video game to be understood as an online game is that it is played through some sort of computer network, although it is usually the case that the motivation for playing these games is to interact and play with other, non-collocated persons (see Schultheiss, Bowman, and Schumann, 2008). In terms of play mechanics, the online gamer – while desiring similar levels of realism and dramatic fantasy (Kim et al, 2002) – appears to have an enhanced need for connectivity and competition with actual human beings instead of computer-generated, artificially intelligent avatars. In other words, online gamers differ from traditional gamers in that they are playing not only for challenge and competition motives as suggested by Sherry, Lucas, Greenberg, and Lachlan (2006), but for an enhanced sense of social interaction not otherwise offered when playing a more traditional console game. Put another way, the degree of connectivity that online games offer to other players is much larger than traditional games, which only allow interactions with a few players.
Media Addiction as a Process Addiction Addiction is defined by the Gale Encyclopedia of Medicine (1999) as “a dependence, on a behavior or substance (emphasis added) that a person is powerless to stop.” Although the term is often used when defining a medical condition in which an individual has a dependency on a substance, Horvath (2004) posits that addiction can be applied to all types of excessive behavior. Furthermore, there is not a clear consistency in how the term addiction is used as the terms addiction, dependency and habitual use are often used interchangeably, all intended to refer to the same construct of addiction. Soule, Shell, and Kleen (2003) even suggest that the word addiction has been, to some extent, replaced by the word dependence when discussing substance abuse, although the reasons for this linguistic shift may be every bit political as they are semantical. Shaffer, Hall, & Bilt (2000) explain that While there are simple working definitions of addiction, the essence of the construct remains elusive…Thus, addiction is essentially a lay term, although scientists often use it. (p. 162).

Addiction can be understood in terms of substance addition and process addiction. Substance addiction has been associated with habitual patterns of behavior related to a chemical dependency. Common examples of substance addiction are alcoholism, drug abuse, and smoking (Lee & Perry, 2004). Conversely, process addiction can be understood as habitual patterns of behavior related to an activity, and can include gambling, spending, shopping, eating, and sexual addictions. Griffiths (1996, 1998) argues that media addiction should be understood as a process addiction, further labeling media addiction as an excessive human-machine interaction. Perhaps the most well-known application of this addiction paradigm is Horvath (2004), who created a measure for television addiction that identified four reliable factors: heavy viewing, problem viewing, craving for viewing, and withdrawal. Although the Horvath measure was designed for television, the fundamental structure of the scale – and by proxy his conceptualization of addiction – may well work for identifying video game addiction, as both are process addictions related to media use.
Personality and Media Use An individual’s personality is a relatively stable precursor of behavior in a micro level, as it indicates an enduring style of one’s thinking, feeling, and acting in different situations (McCrae & Costa, 1999; Guthrie, Ash, & Stevens, 2003). There are several different psychodynamic perspectives to human personality in the literature, including the psychoanalytic approach championed by Freud, the analytic approach proposed by Jung, and the life-span approach fronted by Erikson. Following a different tack, the notion of trait theory and of trait personality as a predictor of human behavior began in the 1950s, with Allport and Cattel’s effort to categorize trait personality dimensions (see Shultz & Shultz, 2005, for a summary of this work). From this trait approach, studies involving the so-called ‘Big Five’ (also called the five-factor index, or FFI) personality traits of extroversion, neuroticism, openness to experience, agreeableness, and consciousness (Costa & McCrae, 1988, 1992) have been plentiful. In a narrow sense, the FFI is not a theory of personality per se, but an empirical generalization about the covariation of different personality traits (McCrae & Costa, 1999). Interest in this model has increased largely due to research documenting empirical linkage between those traits identified by the FFI and a variety of behavioral measures (Guthrie et al. 2003) and several studies have supported the consistency of the Big Five factors across different populations of individuals (i.e. Costa & McCrae, 2004; Costa et al., 2004). Although the FFI was originally developed to characterize laypersons who have a rich vocabulary for describing themselves and others, in terms of relatively enduring patterns of thoughts, feelings, and actions (McCrae & Costa, 1999), many scholars in communication have additionally argued that aspects of an individual’s personality are associated with various media consumption processes (Rosengren, 1974; Rosengren, Wenner, & Palmgreen, 1985; Wober, 1986; Finn, 1997; Weaver, 2000; Hamburger & Ben-Artzi, 2000). The personality traits identified by the FFI have been used to explain the use of different forms of media. Finn (1997) found significant positive correlations between recreational reading and openness to experiences. Weaver (2000) also found support for relationships between the FFI and media use, reporting that whereas extraverts tend to prefer face-to-face communication, neurotics are partial to mediated communication, especially television. Hamburger and Ben-Artzi (2000) found that people high in extraversion participate in more recreational activities, such as online gaming and Internet chat rooms. In terms of media addiction, Young (1996), Griffiths, (1998) and Duran (2003) have all used personality traits to explain an individual’s susceptibility to Internet Addiction Disorder (IAD), a process addiction related to excessive Internet use. The goal of the present study is to further investigate the role of personality and media use, especially to the extent that personality traits identified by the FFI can explain differential patterns in the amount of media use, specifically those patterns that are indicative of an online gaming addiction.
Perception and Media Use Another important variable in explaining media use patterns is individuals’ perceptions of a particular medium. Because the decision to adopt a medium is largely dependent on perceptions of the medium [see Rogers & Singhal (1996) for a discussion on the diffusion of innovation (DoI) process, and Lehman-Wilzig & Cohen-Avigdor (2004) for an application of DoI to new media use], research involving excessive media use must look at individual’s perceptions of the medium (Trevino & Webster, 1992) as well as how these perceptions relate to actual media use (Rosengren, 1974; Sherry, 2001b; Vorderer, Hartmann, & Klimmt, 2003; Bowman & Sherry, 2006; Sherry, Rosaen, Bowman, & Huh 2006). Moreover, our understanding of the different attributes of a particular medium is limited only by the attributes that can be identified by theorists and researchers (Eveland, 2003), as well as the general audience of users. Although media differ in specific and concrete attributes, a user’s perception of the relative usefulness of particular attributes of a medium are anything but specific and concrete. Just as Jack can think text books are boring because they have no audio and video, Jill may find television to be too distracting and loud, preferring to read the printed version of a novel rather than watch the Hollywood production of the same. Thus, individual perceptions of attributes of a medium should affect how much time one spends consuming that medium. An often-discussed element of perceived attributes is “convenience” which is understood as the relative ease-of-use of the medium, often discussed in tandem with the medium’s interface. Interface refers to all parts of a media system that users contact with. The more complex a medium’s interface is, the less convenient the medium is perceived to be, and vice versa. A traditional television that is controlled by remote control could be understood to have an easier interface than the Internet, which is operated by a mouse and keyboard. Of course, this evaluation of convenience will vary according to each user such that an individual with computer skills would hardly notice a difference in complexity, while a computer novice would. Accordingly, convenience may be composed of evaluation of complexity and the corresponding ease of use. Ease of use is defined as one’s belief about how effortless a system is to use (Trevino & Webster, 1992). This definition further implies the importance of individual perception about technological attributes of the media (Trevino & Webster, 1992). Other dimensions of perceptions of media in the literature are usefulness and interest. Usefulness is conceptualized as an individual’s perception of the relative gain of using the medium over other media. Interest is understood as an individual’s willingness to become oriented to the medium. Teo, Lim, and Lai (1999) claim that ease of use may affect perceived usefulness and perceived enjoyment, and the perceived attributes influence the amount of web use. Lederer, Maupin, Sena, & Zhuang (2000) found that ease of understanding and convenience of use affect the amount of web use. With respect to video games, Gao (2004) suggested that ease of use is associated with attitude toward the game and intention to continue playing. When studying the effectiveness of language acquisition via computer games, deHann (2005) found that learning outcomes were positively related associated to perceived usefulness. In addition, most game-flow experience researchers have dealt with enjoyment as a key element of video game play (e.g. Sweetser & Wyeth, 2005). Table 1 offers a summary of previous research about perceived media attributes, especially related to internet use and video game play. The attributes are divided into three distinct groups of the perceptions that may influence media use: convenience, usefulness, and interest.
Table 1.
Some studies about perceived attributes of media and online activity or video game play
|Type of attributes |Independent variable |Dependent variable(s) |
|Convenience |Ease of use |Web use (Teo, Lim, & Lai 1999; Lederer, Maupin, Sena, & Zhuang 2000; |
| | |Cheung, Chang, & Lai 2000) |
| | |Game enjoyment and addiction (Gao 2004; Wan & Chiou 2007) |
| | | |
| | | |
| | |Internet exposure, intimacy, satisfaction (Papacharissi & Rubin 2000) |
| |Complexity |Web use (Cheung et al. 2000) |
|Usefulness |Usefulness |Web use (Teo et al. 1999; Lederer et al.2000) |
| |Consequence |Web use (Cheung et al. 2000) |
| |Video game play |Language acquisition (deHann 2005) |
|Interest |Enjoyment |Web use (Teo et al. 1999.) |
| | |Video game flow (Sweetser & Wyeth 2005) |
| |Affect |Web use (Cheung et al. 2000) |

These studies found that different perceptions of media are related to media use. If process addiction is defined as an otherwise acceptable, normal behavior gone bad (e.g., if a “shopaholic” is somebody who compulsively purchases goods [see Benson, 2000 for a comprehensive review of the condition]), it is reasonable to assume that perceptions of media could be related to compulsive use of media. Just as a shopaholic perceives shopping to be a very convenient, highly useful and interesting activity, an individual’s excessive media use may well be associated with their perceptions of the media as convenient, useful and interesting. Huh (2004) found that ritualized media use was correlated with perceptions of intimacy to the medium and ease of use, and Rubin (1984) found that heavy media users were more likely to continue their media use regardless of content. Thus, this study will investigate the potential for a relationship between individual’s perceptions of the video game medium and addiction to online video games.
Research Questions Correlations between the Big Five personality traits and patterns of addictive behaviors -- both substance and process addictions -- have been evidenced in the extant literature. However, these studies have not been extended to understanding addiction to video games. The present research is aimed at investigating possible correlations between personality dimensions and online game addiction by posing the following research question: RQ1: What is the nature of the relationship between the Big Five personality characteristics and game addiction? Specifically with respect to media use, individual perceptions of different attributes of a medium are associated with patterns of media use, including addictive usage; again, this has not been studied extensively with relation to online game addiction. This leads to the second research question of the current study: RQ2: What is the nature of the relationship between perceptions of media and game addiction?
Method
Participants Surveys containing items measuring addiction, perceptions of online games, and personality traits were distributed to students at a large, urban university in South Korea. A total of 173 students responded to the questionnaire. The students were enrolled in one of five social science classes in the university. Of 173 survey distributed, 142 were returned, and, nine of these were eliminated because they were not filled out correctly; thus, the final usable response rate was 133 of 173 (77 percent).
Procedure
Participants were recruited from undergraduate Communication courses at a large, Korean university to participate in a short survey about the popularity of online video games. To distinquish online games from traditional console games, online games were exemplified in the introductory text of the survey using popular commercial titles (e.g., World of Warcraft and Lineage) Importantly, no formal definition of online games was provided to participants; that is, participants were free to consider any video games played via the Internet when responding to survey questions. There was no compensation for participation in the survey, and those who agreed to participate in the research filled out a consent form and took the main survey. All participants completed the same survey questions, which took approximately ten minutes to complete.
Measures
Addiction measurement. The addiction measure was based on two separate addiction scales found in Horvath’s (2004) television addiction measure and Griffiths and Hunt’s (1998) computer game dependence measure. The Horvath scale identifies seven factors of addiction, identified as: tolerance, withdrawal, unintended use, cutting down on playing time, time spent playing, displacement and continued use. Each factor has five items associated with it, resulting in a 35-item, seven-factor scale. On inspection, four items were dropped because of redundancy issues (the items were nearly identical to each other; in these cases, we retained only one of the items), difficulty translating to Korean, and non-applicability to the video game medium. Also, items from the Griffiths and Hunt scale were fused with the Horvath scale, as both scales had a high degree of redundancy. Because the scales were modified to better fit the goals of this study, a principal components analysis (PCA) was conducted on the new scale to examine model fit. Several factor solutions were considered, but a four-factor solution was retained due to higher factor loadings, lower cross-loadings, and a large amount of variance explained. Of the original scale items, 15 were retained. These items loaded onto factors as follows: social sanctions (α = .920), uncontrollable play (α = .844), excessive play (α = .884), and displacement (α = .831); this solution explained 74 percent of the variance in the addiction measure. Factor means were: social sanctions (M = 1.94, SD = .883), uncontrollable play (M = 2.85, SD = .961), excessive play (M = 2.33, SD = 1.202), and displacement (M = 3.19, SD = 1.12). All items and their corresponding factor loadings are listed on Appendix 1. Personality measurement. Costa and McCrae’s (1992) NEO-Five Factors Inventory (NEO-FFI) was used in this study. The full version of this scale contains 50 -100 statements, including reverse scales for reliability testing. However, this study has a shortened version of the scale that removed the reverse-coded items (for ease of administration to participants) and inadequate statements for Asian cultural customs. Responses to each item are on a five-point scale, from 1 (Never) to 5 (Highly so). Items on each factor were summed, and reliabilities were calculated. Cronbach’s alpha for each dimension of the scale were: neuroticism (α = .768), extraversion (α = .682), openness to experience (α = .671), agreeableness (α = .621), and conscientiousness (α = .651). Factor means were: neuroticism (M = 2.87, SD = .777), extraversion (M = 3.28, SD = .568), openness to experience (M = 3.34, SD = .623), agreeableness (M = 3.51, SD = .511), and conscientiousness (M = 3.28, SD = .509). Perception of game media measurement. Using the different studies listed on Table 1, the researchers settled on three recurring factors understood to be indicators of an individual’s perception of media: convenience, usefulness, and interest. For each of the three dimensions, five-point semantic differential scales were used; these are listed in Table 2 below. Reliabilities of the dimensions were well within acceptable ranges: convenience (α = .797), usefulness (α = .785), and interest (α = .780). Factor means were: convenience (M = 3.55, SD = .794), usefulness (M = 2.86, SD = .623), and interest (M = 3.77, SD = .829).
Table 2. Adjective sets for the dimensions of perceived attributes of a medium.
|Factor |Negative Extreme |Positive Extreme |
|Convenience (α = .797) | | |
| |Hard to find information |Easy to find information |
| |Easy to use |Hard to use |
| |Hard to understand |Easy to understand |
| |Easy to learn how to use |Hard to learn how to use |
|Usefulness (α = .785) | | |
| |Harmful |Useful |
| |Rich in information |Lack in information |
| |Offering incorrect information |Offering correct information |
| |Untrustworthy |Credible |
| |Inessential in life |Essential in life |
|Interest (α = .780) | | |
| |Boring |Exciting |
| |Dull |Stimulating |
| |Unpleasant |Pleasant |

Results
Demographics
The mean age for the study sample was 22.99, SD = 2.515, and included 76 males and 57 females. Average time spent with media included 1.98 hours of television watching per day, and 3.26 hours of computer use per day. Of the time spent using computers, nearly 2/3 of that – 2.03 hours – was spent playing online video games. Participants also reported playing online games between three and four days per week, M = 3.43.
Addiction and Personality RQ1 investigated the nature of the relationship – if any – between the Big Five personality characteristics and game addiction. Because both of these constructs are multi-dimensional, a canonical correlation was performed to measure the strength and direction of association between our two constructs addiction and personality. The analysis of covariance between the addiction set and the personality set resulted in four roots, only one had a coefficient high enough (Rc > .30) to warrant interpretation (also refer to Tabachnick & Fidell, 1989; Tucker & Chase, 1980 for procedural explanation). The first root (Rc = .44, p = .003) accounted for 67 percent of the common variance between addiction and personality. High loadings on the social sanction (-.902) and uncontrollable play (-.812) dimensions of media process addiction were associated with high loadings on the neuroticism (-.626) and extroversion (-.535) dimensions of personality. In other words, the addiction dimensions of social sanctions and uncontrollable play covary positively and significantly with the personality dimensions of neuroticism and extraversion , see Table 3 below. Further analysis of the bivariate correlation matrix found significant correlations between social sanctions and both neuroticism (r = .219, p < .05) and extraversion (r = .220, p < .05), and uncontrollable play was associated with neuroticism (r = .258, p < .01) and extraversion (r = .208, p < .05). Displacement was also found to be significantly correlated with agreeableness, but this correlation was weak (r = .172, p < .05).
Table 3.
Canonical correlation between addiction and personality.
| |Root1 |
| |Correlation |Coefficient |
|Addiction set | | |
| |Social sanctions |-.902 |-.802 |
| |Uncontrollable play |-.812 |-.568 |
| |Excessive play |-.481 |.312 |
| |Displacement |-.226 |.157 |
|Variance |43.91% | |
| |Redundancy |8.5% | |
|Personality set | | |
| |Neuroticism |-.626 |-.634 |
| |Extraversion |-.535 |-.800 |
| |Openness to experience |-.229 |-.110 |
| |Agreeableness |.316 |.475 |
| |Conscientiousness |-.008 |.037 |
|Variance |16.6% | |
| |Redundancy |3.2% | |
|Canonical correlation |.440 | |
|Common variance |67.46% | |

Addiction and Perception

RQ2 investigated the nature – if any - of the relationship between perceptions of media and game addiction. Again, because both of these constructs are multi-dimensional, another canonical correlation analysis was performed to establish covariance between perceptions of media and game addiction. Of three roots, only one was significant (Rc = .370, p