A socio-cognitive model of video game usage.

The video game has become one of the most popular and pervasive forms of entertainment. On a regular basis, more than half of all Americans age 6 and older play some form of electronic digital interactive video games, including ones played in arcades, on handhelds, on game consoles, on personal computers, and on the Internet. The average game player is 33 years old and the average adult man and woman players play 7.6 and 7.4 hours per week, respectively (Entertainment Software Association, 2004). United States retail sales of video games, including portable and console hardware, software and accessories, reached $10.5 billion in 2005, surpassing motion-picture box-office figures in consumer entertainment expenditures (1) (NPD Group, 2006). This popularity is expected to increase as video games become equipped with enhanced speed, more detailed graphics, and increased online network functionality (Williams, 2002).

The popularity of video games is accompanied by social concerns regarding excessive video game use, sometimes hyperbolically called "video game addiction." Both popular and scholarly articles drew attention to the problem of excessive video game consumption by associating it with psychiatric conditions such as substance abuse dependency (van Grinsven, 2003), the so-called addictive personality (Griffiths & Dancaster, 1995), and pathological behavior (Fisher, 1994; Griffiths, 1992; Griffiths & Hunt, 1998; Phillips, Rolls, Rouse, & Griffiths, 1995).

However, the term "addiction" is problematic, particularly in understanding excessive media use. According to Shaffer, Hall, and Vander Bilt (2000), addiction is a lay term rather than a scientifically defined term, leading to the conceptual confusion surrounding excessive media use. Similarly, Peele (1995) pointed out that the term addiction may be abused in that it tends to generate a sense of urgency about psychological problems with the (often self-serving) purpose of alarming the lay public. Now leading media addiction researchers have adopted the term "problematic" media use (Caplan, 2005, p. 721; Kubey & Csikszentmihalyi, 2002, p. 74) in its stead. Many so-called media "addicts" would not be clinically diagnosed as such because their usage, no matter how excessive it might appear to the casual observer, did not have the dire but necessary consequences of broken families or ended careers attached to it (cf. Shaffer et al., 2000). And, many people overcome the symptoms of media addiction without professional intervention (Hall & Parsons, 2001).

LaRose, Lin, and Eastin (2003) argued that media addiction was overstated and that in many cases its "symptoms" may be understood as benign problems that are within the individual's capability to correct rather than malignant problems requiring professional intervention. Drawing on the social cognitive theory of self-regulation (Bandura, 1991), they proposed a model of unregulated media usage that ranges from normally impulsive media consumption patterns to extremely problematic behavior. Unregulated media consumption may affect any media user to some degree at various times, and may become problematic even at relatively low absolute levels of use while remaining unproblematic at high levels.

From this perspective, the present study explores socio-cognitive mechanisms of self-regulation in a model of video game consumption behavior. It extends previous research by integrating Csikszentmihalyi's (1975) theory of flow experience, which has also been proposed as an explanation of video game consumption (Sherry, 2004). By investigating the linkage between flow experience and self-regulation in decision-making processes in media usage, this study attempts to explicate sociocognitive media consumption mechanisms.

A Social Cognitive Perspective of Video Game Usage

Social cognitive theory is a comprehensive theoretical framework for understanding human behavior. It offers an agentic perspective: human agents intentionally make things happen through their actions by exercising forethought, reflecting on their behavior, and applying self-reactive motivating influences (Bandura, 2001). Social cognitive theory highlights the self-regulatory mechanism through which individuals observe their own behavior (self-observation), judge it in relation to personal and social standard or norms (normative judgmental process), and adjust their own behavior to environment by applying self-reactive incentives (self-reactive influence) (Bandura, 1991). Through the mechanism of self-regulation, individuals use their self-regulatory capabilities to predict, control, and manage their own behavior.

LaRose et al. (2003) conceptualized media attendance along a continuum of unregulated media behavior that lies between normally impulsive and problematically excessive consumption patterns. They proposed deficient self-regulation and habit formation as failures in normal self-regulation of media consumption behavior that often lead to patterns of mounting usage. Unregulated media consumption behavior may be initiated by individuals' conscious desire to regulate their negative psychological states, such as stress, boredom, loneliness, or the like. For example, stressed or bored people are likely to manage their psychological states by playing video games. Such incentives to initiate media consumption patterns parallel "pass time" and "relieve boredom" gratifications in uses and gratifications tradition. However, defining them instead as self-reactive outcome expectations has improved both their conceptual clarity in socio-cognitive terms and their explanatory power (LaRose & Eastin, 2004).

As individuals come to rely on their video game use to counter their psychological states, they are likely to form habits, defined as "situation-behavior sequences that are or have become automatic, so that they occur without self-instruction" (Triandis, 1980, p. 204; see also Bargh & Gollwitizer, 1994). Media habits may be established by past thinking about outcome expectations-cum-gratifications (Rosenstein & Grant, 1997; Stone & Stone, 1990). Through repetition, media consumers become ever less conscious of expected consequences of their media consumption and stop actively reasoning about their media consumption patterns. Individuals become no longer subject to active consideration and the performance of the media behavior may then become a conditioned response triggered by a sensory stimulus or a recurring situation, such as the sight of one's video game console upon returning from class. Consequently, media consumption patterns become automatic over time. Media habits thus may be regarded as a failure of self-observation (LaRose & Eastin, 2004). From this perspective, the automaticity (cf. Bargh & Gollwitzer, 1994) of habitual media consumption may be distinguished from the so-called ritualistic gratifications (Rubin, 1984) which still assume an active media selection process (e.g., to gratify needs to pass the time).

Most media habits are subject to self-control; for example, readers may curtail their morning newspaper "habit" if they notice that they are late to work or if it detracts from family interaction (e.g., Diddi & LaRose, 2006). Deficient self-regulation is defined as a state in which self-control is diminished (LaRose et al., 2003). Media consumers lose control of their habits and become deficient in self-regulation when the other two subprocesses of self-regulation, judgmental process and self-reactive influence, begin to fail. For example, people no longer judge their behavior against acceptable personal or social standards for "normal" amounts of game play and no longer apply self-reactive influences, such as self-administered rewards for moderating consumption or indulging feelings of guilt for excessive play. This is likely to happen when media consumption becomes a conditioned response to negative psychological states. If negative life consequences of excessive consumption, such as playing games so much that one flunks out of school, cause those negative moods, then a downward spiral into what might truly be considered a video game addiction, or more properly, problematic video game usage, may occur (LaRose et al., 2003). Similarly, Lee and Perry (2004) found that college students were more likely to be preoccupied with and lose control of instant message software use as their self-regulation became more deficient.

Although deficient self-regulation may deepen established media habits, it may also help initiate them to have a direct influence on media consumption. Impulsive thoughts, such as those triggered by the excitement of the release of a new video game, may also overwhelm one's judgment, self-reactive influences, and rational consideration of the merits/expected outcomes/gratifications of the game itself. In their study on the Internet, LaRose et al. (2003) found significant relationships among self-reactive outcome expectations, deficient self-regulation, habit strength, and Internet usage. Applying this reasoning to the present context of video game usage, the following hypotheses are proposed:

[H.sub.1]: Video game habit strength will be positively related to video game usage.

[H.sub.2]: Deficient self-regulation of video game consumption will be positively related to video game usage.

[H.sub.3]: Self-reactive outcome expectations will be positively related to video game usage.

[H.sub.4]: Deficient self-regulation of video game consumption will be positively related to video game habit strength.

[H.sub.5]: Self-reactive outcome expectations will be positively related to video game habit strength.

[H.sub.6]: Self-reactive outcome expectations will be positively related to deficient self-regulation of video game consumption.

Social Cognitive Perspective of Flow Experience

Flow is a concept that has been proposed to explain enjoyable experiences that can be produced from one's immersive engagement in everyday activities. Defined as "the holistic sensation that people feel when they act with total involvement" (Csikszentmihalyi, 1975, p. 36), flow is described as a psychological state in which an individual experiences a feeling of transcendence, or oneness, with one's activity so that nothing else seems to matter (Csikszentmihalyi, 1990). When people are in the flow state, they are so intensely focused on their present activity that they lose reflective self-consciousness, feel in control of their environment, sense merging of their actions and awareness, experience temporal distortion, and are intrinsically rewarded by the activity they are engaged in (Csikszentmihalyi, 1975). In this process, they continue to seek more complex challenges and perfect their skills during their course of action and, thus, their awareness and motivation of the performance become their activity, which eventually becomes an end in and of itself (Csikszentmihalyi, 1988). Flow has been applied to understanding media consumption behavior in a variety of interactive media environments (Ghani & Deshpande, 1994; Hoffman & Novak, 1996; Koufaris, 2002; Novak & Hoffman, 1997; Trevino & Webster, 1992; Webster, Trevino, & Ryan, 1993).

It has been suggested that flow experience instigates media users to participate in media consumption repeatedly and excessively. Hoffman and Novak (1996, p. 57) argued "flow is 'the glue' holding people" in highly interactive mediated communication. In a subsequent study, Novak and Hoffman (1997) applied the concept of flow to Internet usage and found that users' flow experience was significantly related to the frequency and duration of Web site visits, promoting "stickiness" to their media consumption behavior. Flow experience, often characterized as concentration and intrinsic enjoyment, has been found to predict media usage (Ghani & Deshpande, 1994). Focused concentration positively influenced the overall experience of computer users (Novak & Hoffman, 1997) and their intentions to use a system repeatedly (Webster et al., 1993).

The relationship between flow experience and media usage can be explained in social cognitive terms as well. Flow experiences provide enjoyable activity incentives (cf. Bandura, 1986) that motivate media consumption, video game usage in the present instance. For example, either by directly experiencing pleasurable flow states while playing video games themselves (enactive learning in social cognitive terms) or by observing the flow states of others (vicarious learning) video game players expect that the same enjoyable, immersive feelings of "oneness" will be visited upon them the next time they play a game. Further, video game players seek to fulfill their self-reactive outcome incentives to regulate their psychological states through their flow experiences. This assumption is consistent with Sherry's (2004) idea that flow experience offers an opportunity to seek out emotional pleasure, such as an escape to a fantasy, by both arousing and relaxing media users.

Once the relationship between media usage and enjoyable activity/self-reactive outcome incentives is well established through repeated flow experiences over time, media consumers may gradually cease to actively consider media consumption decisions. As a result, the enjoyable activity incentive should directly cause repeated media consumption that will lead to habit formation. Thus, flow experience should be directly related to habit strength.

Individuals also lose reflective self-consciousness by applying intense self-reactive incentives to their media behaviors through video game-induced flow experiences. As individuals experience the flow states, their self-regulation becomes deficient and the self-regulatory functions of judgmental process and self-reactive influence cease to moderate their gaming behavior. They thus become engaged/immersed in media consumption much longer than originally planned because two important mechanisms of self-regulation are temporarily disengaged. The third subfunction of self-regulation, self-observation of behavior, would seem to be also fully immobilized perhaps accounting for the intensity of concentration and "oneness" of the flow experience that temporarily blocks judgmental process and self-reactive influence as well as awareness of the passage of time. Consequently, garners may therefore fall into a pattern of mounting game consumption and resort to the enjoyment of the flow experience to dissipate the negative mood that follows from the harmful life consequences (e.g., flunking out of college) that are linked to excessive usage. Thus, the following hypotheses are formally stated:

[H.sub.7]: Flow experience will be positively related to video game habit strength.

[H.sub.8]: Flow experience will be positively related to deficient self-regulation of video game consumption.

[H.sub.9]: Flow experience will be positively related to self-reactive outcome expectations.

[H.sub.10]: Flow experience will be positively related to video game usage.

The flow state is attained only when the level of congruence between skill and challenge is above a certain threshold. As Ellis, Voelkl, and Morris (1994, p. 338) put it, "flow results from experience contexts characterized by a match between challenge and skills only when both challenges and skills exceed the level that is typical for the day to day experiences of the individual." Unless the match reaches the "above-threshold" point, a person is not motivated to become involved in the given activity, even if the person's perceived skill and challenge are matched. If the match lies below a person's typical level, then the person is likely to be in a state of apathy (Csikszentmihalyi & Csikszentmihalyi, 1988). Consequently, the best moment for the flow experience is realized when one is intrinsically motivated to strive to achieve something active, difficult, and worthwhile, not something passive, receptive, or relaxing (Csikszentimihaly, 1990). Thus, only the optimal balance activates intrinsic interests that require further concentration and involvement to gratify individuals' internal goals for their activity. Because video games are a highly active medium, requiring intense concentration and physical activity (Dominick, 1984), as well as concrete goals, clear feedback, and rich visual and aural information (Sherry, 2004), video game play will continue only if the game players' effortful involvement optimally matches the demanding levels of the contents of the game. In this sense, the flow state should be understood as a consequence of an optimal balance rather than a simple match between skill and challenge in a given situation.

The balance of skill and challenge is encapsulated succinctly in the social cognitive construct of self-efficacy, defined as "beliefs in one's capabilities to organize and execute the courses of action required to produce given attainments" (Bandura, 1997, p. 3). Thus, perceived skill may be equated with self-efficacy (see Koufaris, 2002). Based on this reasoning, the following hypothesis is stated:

[H.sub.11]: Those with high self-efficacy and high challenge will be more likely to experience flow experience than all others with high self-efficacy and low challenge, with low self-efficacy and high challenge, and with low self-efficacy and low challenge.

In an attempt to test the direct and indirect relationships as hypothesized above, this study proposed a structural model (See Figure 1) that integrates the underlying mechanisms of Bandura's (1991) social cognitive theory of self-regulation and Csikszentmihalyi's (1975)theory of flow experience. In the proposed model, video game use was directly and indirectly affected by self-reactive outcome expectations, habit strength, deficient self-regulation, flow experience, and optimal balance between perceived self-efficacy and perceived challenge.

[FIGURE 1 OMITTED]

Research Method

Respondents and Procedures

College students are an important population in studying video game consumption behavior because they are not only the first generation of home-console video game players (e.g., Nintendo/Super Nintendo, Sega Genesis, Sony PlayStation, etc.) (Lucas & Sherry, 2004), but also still enthusiastic gamers today. According to a Pew Internet & American Life Project survey (2) (Jones et al., 2003), video games have become a part of college life: nearly every college students had played a video game (e.g., console, computer, online games, etc.) at some point in their lives and about two out of three still play as a regular or occasional game player in many accommodating settings of campus, such as computer labs and dormitories. The same survey also reported that close to half of the population agreed that gaming keeps them from studying "some" or "a lot." Known to be more susceptible to depression than other populations (Rich & Scovel, 1987), college students themselves tend to have some characteristics that can provoke particular risk to obtain excessive video game habits, which may often stem from their tendency for unregulated "online" usage as well as compulsive buying behavior (LaRose & Eastin, 2002). As such, college students' video gaming is likely to tell much about underlying socio-cognitive mechanisms of video game consumption behavior.

A Web-based survey was employed for data collection. Respondents were a sample of 538 college students enrolled in two undergraduate communication and one advertising class at a large Midwestern university. Initially, the sample consisted of 538 student respondents. 52% of them were male and 48% were female. However, 28% of the sample (150 out of 538 respondents) reported that they did not play video games, yielding a sample size of 388. The final sample of 388 respondents consisted of 59% male and 41% female. The sample was 33.2% freshman, 22.4% sophomore, 28.1% junior, and 16.2% senior students with a median age of 19 years. The average number of minutes spent on video games in the typical weekday was reported to be 93 minutes (SD = 97.56), and typical weekend day use was 109 minutes (SD = 112.02).

Operational Measures

Measures were adopted from previous research and modified to fit the context of this research. Although each of the measures has been used in previous studies, confirmatory factor analysis (CFA), guided by Anderson and Gerbing's unidimensionality of the measurement scales (1984, 1988), (3) was also performed to validate and refine the measures by using established goodness of fit criteria (Kelloway, 1998; Kline, 1998), such as .8 or above for adjusted goodness-of-fit index (AGFI), closer to 1.0 for comparative fit index (CFI), and less than or equal to .08 for root mean square error of approximation (RMSEA). After the CFA procedure, the multiple indicators of each construct were summed and averaged to create each index. Finally, Cronbach's alpha ([alpha]) reliability coefficients were computed to assess the internal consistency among the indicators in each construct. Table 1 shows the results of CFA and reliability test with the means and standard deviations of the scales and their component items.

Flow Experience. Flow experience was measured as a second-order factor (4) using retrospective self-reports adapted from prior flow research (e.g., Koufaris, 2002; Novak, Hoffman, & Yung., 2000). In line with the previous self-report survey method, the present study used the retrospective measures consisting of nine 7-point Likert-type items, which captured three aspects of flow experience: enjoyment, merge of action and awareness, and concentration in playing video games (see Table 1). Factor loadings in CFA were found to be statistically significant (p < .001) with a good model fit ([chi square] = 18.4, df = 11, p = .073, AGFI = .969, CFI = .966, RMSEA = .042). (5) Cronbach alpha coefficients were .86 for intrinsic enjoyment, .87 for merge of action and awareness, and .87 for concentration.

Self-Reactive Outcome Expectations. Self-reactive outcome expectations were measured with an additive index of five 7-point Likert-type items, which assessed the degree to which respondents played video games to regulate psychological state so that they can relieve boredom, lessen loneliness, pass the time, or seek for an escape. (6) These items are also shown in Table 1. Factor loadings in the CFA model were found to be statistically significant (p < .001) with a good model fit [chi square] = 8.0, df = 5, p = .158, AGFI = .976, CFI = .995, RMSEA = .039). Cronbach alpha was .80 for self-reactive outcome expectations.

Deficient Self-Regulation and Habit Strength. Deficient self-regulation and habit strength initially consisted of 12 (6 items for deficient self-regulation and 6 for habit strength) 7-point Likert type-items. These items were adapted from prior research on Internet consumption behavior (LaRose et al., 2003; LaRose & Eastin, 2004). Although LaRose and Eastin (2004) made a clear distinction between deficient self-regulation and habit strength using an exploratory factor analysis, the present study performed a confirmatory factor analysis to obtain more stringent discriminant validity between the two constructs. Two theoretically distinctive factors were confirmed. (7) Factor loadings in the CFA were found to be statistically significant (p < .001) with a good fit of the two-factor model ([chi square] = 8.8, df = 8, p = .359, AGFI = .98, CFI = 1.00, RMSEA = .016). The relevant items are shown in Table 1. Cronbach alpha coefficients were .88 for deficient self-regulation and .88 for habit strength.

Perceived Self-Efficacy and Challenge. Optimal balance between perceived self-efficacy and challenge was operationalized and measured as a dichotomous dummy variable using the measures of perceived video game self-efficacy and video game challenge. As where 7 was strongly agree and 1 was strongly disagree, in the manner that is conceptualized in Bandura (1986), and recommended and employed by LaRose et al. (2003), video game self-efficacy was measured using an index of four (8) 7-point Likert-type items, which assess respondents' perceived belief or judgment in their capability to use skills for the particular course of video gaming. Factor loadings in the CFA were found to be statistically significant (p < .001) with a good fit of the model ([chi square] = 0.6, df= 2, p = .748, AGFI = .99, CFI = 1.00, RMSEA = .001). Video game challenge was also measured using an index of three (9) 7-point Likert-type items, 7 for strongly agree and 1 for strongly disagree, which was adapted from prior flow research (e.g., Koufaris, 2002). Cronbach alpha coefficients were .91 for video game self-efficacy (M = 5.22, SD = 1.26) and .86 for challenge (M = 3.78, SD = 1.52).

According to Massimini and Carli's (1988) conceptual flow model, the starting point for flow experience lies above the personal average of skills and challenges. Based on this conceptualization, the present study created optimal balance between self-efficacy and challenge as a dichotomous variable by assigning 1 to the survey respondents who rated both perceived self-efficacy and challenge higher than the median scores (5.50 for self-efficacy and 4.00 for challenge) of the two variables (n =134, high self-efficacy/high challenge) and by assigning zero to those who rated the two variables otherwise (n = 254, other combinations of self-efficacy and challenge, such as high self-efficacy/low challenge, low self-efficacy/high challenge, and low self-efficacy/low challenge).

Video Game Usage. Video game usage was an additive three-item index consisting of the number of minutes participants reportedly played video games on a typical weekday, a typical weekend day, and the day previous to the survey. Log10 (1 + value) transforms were applied to each value prior to summing the items in the index because an inspection of the distribution of these items scores was found to be skewed to the right due to outliers. Cronbach alpha for the resulting composite index was .67 (M = 4.08, SD = 1.72)

Data Analysis

First, Pearson product-moment correlations were calculated by SPSS 13. Then, with AMOS 5.0 using maximum likelihood estimation, a path-analytic structural equation modeling technique was performed to test the main 11 hypotheses in the proposed structural model while controlling for the other variables. Although longitudinal data is necessary to determine causality of the relationships, this analysis was chosen not only to examine the direct and indirect relationships of the hypotheses, but also to demonstrate whether or not the current data support causal associations in the proposed model as theorized. Finally, t tests were also performed to confirm the final hypothesis of this study.

Results

Table 2 shows a matrix of Pearson product-moment correlation coefficients. All variables were significantly correlated with all the others.

However, the results of structural equation modeling indicate that 10 out of 11 hypotheses were supported (see Figure 2). Unlike the matrix of Pearson correlation coefficients, flow experience did not emerge as a significant predictor of video game usage ([beta] = .08, p > .10) in the proposed structural model. (10) An inspection of the modification indices also suggested an unexpected causal link from optimal balance between video game self-efficacy and challenge to deficient self-regulation in video game usage ([beta] = .23, p < .001). Thus, the proposed model was revised and retested by deleting the nonsignificant path from flow experience to video game usage and adding the significant path from optimal balance to deficient self-regulation. The revised model fits the data very well [chi square] = 18 .6, df= 14, p = .179, AGFI = .969, CFI = .997, RMSEA = .029 (90% C.I: .001-.060). The significant path coefficients in the final model are shown in Figure 2. From this final model, 51% of the total variance in video game use was jointly explained by the causes specified in the final model.

Finally, t tests supported the final hypothesis that flow state is more likely to be experienced when perceived self-efficacy is optimally matched with perceived challenge. Compared to those with nonoptimal balance in playing video games (M = 3.94, SD = 1.42 for enjoyment; M = 4.05, SD = 1.50 for merge of action and awareness; M = 4.43, SD = 1.40 for concentration), people with optimal balance rated their flow experience significantly higher (M = 5.64, SD = 1.00, t(358.46) = -13.42, p < .001 for enjoyment; M = 5.41, SD = 1.01, t(363.45) =-10.6, p < .001 for merge of action and awareness; M = 5.62, SD = 1.04, t(321.88) = -9.0, p < .001 for concentration). The path coefficient from optimal balance to flow experience also confirmed the final hypothesis ([3 = .56, p < .001).

Discussion

Consistent with prior research on Internet usage (LaRose et al., 2003; LaRose & Eastin, 2004), the findings of the present study highlight the importance of self-regulatory mechanisms in the consumption of interactive media. That is, people tend to spend a substantial amount of time playing video games because their video gaming can provide self-reactive incentives to relieve boredom, lessen loneliness, pass the time, or provide an escape. Further, people are likely to engage in more video game playing by pursuing self-reactive incentives, which not only promote the loss of self-control over their media consumption but also trigger a repeated pattern of video game play. Consequently, people are no longer responsive to active consideration for their own video gaming and tend to play much longer than originally intended.

This study also demonstrated that video game players' flow experience instigated self-reactive outcome expectations. These in turn fostered deficient self-regulation and habit strength, the two main components of self-regulatory mechanisms. The findings also suggest that video game players' flow experience is likely to promote the loss of self-control. Thus, flow experience may not only activate self-reactive outcome expectations but also may make video game self-regulation less effective so that their video game consumption becomes repetitive and out of control. The present research also confirmed a basic tenet of the theory of flow experience: that the flow state is more likely to be experienced when perceived self-efficacy is optimally matched with perceived challenge.

[FIGURE 2 OMITTED]

The emergence of the unexpected causal path from optimal balance to deficient self-regulation in the final path diagram may also suggest that people are more likely to lose their self-control and continue their game play to an excessive extent particularly when their gaming skills optimally correspond to the level of the challenge of the video game. This significant casual link may not be surprising because optimal balance between self-efficacy and challenge can easily motivate players to continue to engage in their gaming so that they can strive to achieve self-administered reward for their performance. As such, optimally balanced skills and challenge in video game play may not only contribute to inducing flow experience, but also influence deficient self-regulation directly. Thus, this significant path should be regarded as a theoretical extension to the proposed model.

The present study also showed that video game players' flow experience was positively correlated with the amount of video game usage. However, the direct influence of flow experience on the amount of video game usage was not evident in the proposed model. The null relationship between flow experience and video game usage in the model may counter the findings of several flow studies (e.g., Ghani & Deshpande, 1994; Novak & Hoffman, 1997; Novak et al., 2000; Trevino & Webster, 1992; Webster et al., 1993) all suggesting that flow experience is directly related to the amount of time spent on media consumption. However, none of the prior studies incorporated the full range of variables found to predict media behavior here and so might have misidentified flow experience as an important predictor of usage.

Another possible explanation for the lack of direct path from flow experience to video game usage is that flow is so fleeting that it does not directly affect overall amount of time spent on video games. Instead, the flow state may only affect the duration of the play session in which it is achieved. Perhaps players reinstate their self-control over their video game play by restoring their self-observation, normative judgment, and self-reactive influence between game playing sessions. The temporary slip in self-control may be a temporary setback brought about by strong situational inducements that are eventually avoidable and alterable (Bandura, 1997).

People's ability to reinstate self-control to resist excessive video game consumption was apparent. Among people who agreed with none of the three addiction symptoms of the deficient self-regulation measure, and thus fell well below the criterion of video game addiction, there was still significant correlations between deficient self-regulation and video game usage (r = .48, p < .001, n = 283). This finding is consistent with LaRose et al.'s (2003) notion of unregulated media use. Unlike the all-or-nothing phenomenon of media addiction, deficient self-regulation was significantly related to media usage within even a normal population. Further, only 10 out of 388 video game players were possible "addicts" in this study, if one follows a criterion for media addiction suggested by Shaffer et al. (2000), in which harmful life consequences are a necessary condition. This finding suggests that concern about video game addiction may be exaggerated and that excessive video game play should be self-correctable for most people.

The present socio-cognitive view of problematic video game use also sheds light on the excessive use of traditional mass media, including television (e.g., Kubey & Csikszentmihalyi, 2002), as well as other interactive media in the Internet environment (e.g., Caplan, 2005; Diddi & LaRose, 2006; LaRose et al., 2003; Lee & Perry, 2004). As suggested in the present study, addictive television consumption behavior also may be triggered by internal cues (e.g., boredom, loneliness, or depression) and external cues (the sight of TV monitor or remote control device, or the incessant barrage of sensory cues of TV programs), which in turn promote TV viewers' self-reactive outcome expectations to relieve or lessen the unpleasant moods, habitual pattern of TV viewing, and loss of self-regulation over TV viewing. Parallel to this view, mood management theory (Bryant & Zillmann, 1984; Zillmann, 1988) also predicts that TV viewing is associated with dysphonic moods, which is indicative of problematic TV consumption behavior. Similarly, Anderson, Collins, Schmitt, and Jacobvitz (1996) found that stress is significantly linked to TV addiction. Consequently, TV viewing becomes rewarding itself by providing a means of escape from unpleasant moods and TV viewers get into downward spirals of unregulated media consumption behavior (LaRose et al., 2003).

Limitations

The present study employed a student sample that may be not generalizable to other populations who have different levels of video game usage and experience. However, the socio-cognitive mechanisms of video game consumption behavior uncovered in the current college student sample can be relevant for testing a nascent theoretical model, particularly when addressing the underlying mechanisms of the emergent media consumption behavior (Caplan, 2005; Pingree et al., 2001). As video games become integrated into the daily life of college students with new communication technologies (e.g., Jones et al., 2003), college students can be a suitable audience for understanding problematic video gaming habits that are an important conceptual issue of this study.

The self-reported recall measures of time spent playing video games also may be a limitation of the present study. In the first Middletown Studies, Papper, Holmes, and Popovich (2004) contended that self-report measures of surveys and diaries were less accurate than direct observations in reporting media use because people tend to underestimate their media use when using the self-report measures. The combination of the self-report measures and the direct observations could strengthen the validity of the result in future research.

The present study implemented cross-sectional data which provided limited support for the longitudinal assumption of the direct and indirect relationships hypothesized in the proposed model. A future study with a solid longitudinal design will be capable of demonstrating how the relationships among the involved variables develop in long-term processes of self-regulation in video game consumption behavior.

Suggestions for Future Research

The relationship between habit and deficient self-regulation should be further investigated. The two constructs may be closely related subdimensional indicators of an unobserved higher level of construct, such as self-regulation. Further research should develop alternative measures of deficient self-regulation and habit strength to achieve better discriminant validity.

It also remains uncertain whether the relationships unobserved (i.e., direct path between flow experience and usage) as well as observed in this study can be found across categories of video games (e.g., classic board game, action-adventure, role playing, puzzle, quiz, racing, strategy, etc.). For example, Sherry et al. (2006) found three underlying dimensions of video game genre preference (imagination, traditional, and physical enactment games) which corresponded to different types of gratification outcomes (e.g., social interaction, status, novel sensory, pass time, etc.).

In particular, Massive Multiplayer Online Role Playing Games (MMORPGs) would seem to have a different profile of expected outcomes than console-based games, engaging social interaction and status incentives to a much greater degree than their offline counterparts. For example, Voiskounsky, Mitina, and Avetisova (2004) investigated patterns of the Multi-User Dimension (MUD) online game players' behavior and found six dimensions of the behavioral patterns, such as flow experience, achievement, activity/passivity, interaction, and thoughtfulness/spontaneity. These findings may suggest that the various genres differentially activate other types of expected outcome expectations as well as different levels of flow experience that could have differential effects on deficient self-regulation, habit strength, and video game usage. Because different genres of video games often require different types or levels of skills and challenges, future research should explore the effect of flow experience on self-regulation in video games across different genres of video games.

References

Anderson, D., Collins, P., Schmitt, K., & Jacobvitz, R. (1996). Stressful life events and television viewing. Communication Research, 23, 243-260.

Anderson, J., & Gerbing, D. (1984). The effect of sampling error on convergence, improper solutions, and goodness-of-fit indices for maximum likelihood confirmatory factor analysis. Psychometrika, 49, 155-173.

Anderson, J., & Gerbing, D. (1988). Structural equation modeling in practice: A review and recommended two-step approach. Psychological Bulletin, 103, 411-423.

Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, N J: Prentice Hall.

Bandura, A. (1991). Social cognitive theory of self-regulation. Organizational Behavior and Human Decision Processes. 50, 248-287.

Bandura, A. (1997). Self-efficacy: The exercise of control. New York: W. H. Freeman.

Bandura, A. (2001). Social cognitive theory: An agentic perspective. Annual Review of Psychology. 52, 1-26.

Bargh, J. A., & Gollwitzer, P. M. (1994). Environmental control of goal directed action: Automatic and strategic contingencies between situation and behavior. In W. D. Spaulding (Ed.), Nebraska symposium on motivation, (Vol. 41, pp. 71-124). Lincoln: University of Nebraska Press.

Bryant, J., & Zillmann, D. (1984). Using television to alleviate boredom and stress: Selective exposure as a function of induced excitational states. Journal of Broadcasting, 28, 1-20.

Caplan, S. E. (2005). A social skill account of problematic Internet use. Journal of Communication, 55, 721-736.

Csikszentmihalyi, M. (1975). Beyond boredom and anxiety. San Francisco, CA: Jossey-Bass.

Csikszentmihalyi, M. (1988). Introduction. In M. Csikszentmihalyi & I. S. Csikszentmihalyi (Eds.), Optimal experience: Psychological studies of flow in consciousness (pp. 3-14). New York: Cambridge University Press.

Csikszentmihalyi, M. (1990). Flow: The psychology of optimal experience. New York: Harper & Row.

Csikszentmihalyi, M., & Csikszentmihalyi, I. (1988). Introduction to part IV. In M. Csikszentmihalyi & I. S. Csikszentmihalyi (Eds.), Optimal experience: Psychological studies of flow in consciousness (pp. 251-265). New York: Cambridge University Press.

Diddi, A., & LaRose, R. (2006). Getting hooked on news: Uses and gratifications and the formation of news habits among college students in an Internet environment. Journal of Broadcasting & Electronic Media, 50, 193-210.

Dominick, J. (1984). Videogames, television violence, and aggression in teenagers. Journal of Communication, 34, 136-147.

Ellis, G. D., Voelkl, J. E., & Morris, C. (1994). Measurement and analysis issues with explanation of variance in daily experience using the flow model. Journal of Leisure Research, 26, 337-385.

Entertainment Software Association (2004). Facts and research: Game player data. Retrieved May 21,2005, from http://www.theesa.com/facts/gamer_data.php

Fisher, S. (1994). Identifying video game addiction in children and adolescents. Addictive Behaviors, 19, 545-553.

Ghani, J. A., & Deshpande, S. P. (1994). Task characteristics and the experience of optimal flow in human-computer interaction. Journal of Psychology, 128, 381-391.

Griffiths, M. D. (1992). Pinball wizard: The case of a pinball machine addict. Psychological Reports, 71, 160-162.

Griffiths, M. D., & Dancaster, I. (1995). The effect of Type A personality on physiological arousal while playing computer games. Addictive Behaviors, 20, 543-548.

Griffiths, M. D., & Hunt, N. (1998). Dependence of computer games by adolescents. Psychological Reports, 82, 475-476.

Hair, J. F., Anderson, R. E., Tatham, R. L., & Black, W. C. (1998). Multivariate data analysis. Upper Saddle River, NJ: Prentice Hall.

Hall, A. S., & Parsons, J. (2001). Internet addiction: College student case studying best practices in cognitive behavior therapy. Journal of Mental Health Counseling, 23, 312-327.

Hoffman, D. L., & Novak, T. P. (1996). Marketing in hypermedia computer-mediated environments: Conceptual foundations. Journal of Marketing, 60, 50-68.

Jones, S., Clarke, L. N., Cornish, S., Gonzales, M., Johnson, C., Lawson, J. N. et al. (2003, July 6). Let the games begin: Gaming technology and entertainment among college students. Washington, DC: Pew Internet and American Life Project. Retrieved December 24, 2005, from http://www.pewinternet.org/pdfs/PIP_College_Gaming_Reporta.pdf

Kelloway, E. K. (1998). Using Lisrel for structural equation modeling. Thousand Oaks, CA: Sage.

Kline, R. B. (1998). Principles and practices of structural equation modeling. New York: Guilford Press.

Koufaris, M. (2002). Applying the technology acceptance model and flow theory to online consumer behavior. Information Systems Research, 13, 205-223.

Kubey, R., & Csikszentmihalyi, M. (2002). Television addiction is no mere metaphor. Scientific American, 286, 74-80.

LaRose, R., & Eastin, M. S. (2002). Is online buying out of control? Electronic commerce and consumer self-regulation. Journal of Broadcasting & Electronic Media, 46, 549-564.

LaRose, R., & Eastin, M. S. (2004). A social cognitive theory of Internet uses and gratifications: Toward a new model of media attendance. Journal of Broadcasting & Electronic Media, 48, 358-377.

LaRose, R., Lin, C. A., & Eastin, M. S. (2003). Unregulated internet usage: Addiction, habit, or deficient self-regulation? Media Psychology, 5, 225-253.

Lee, K. C., & Perry, S. (2004). Student instant message use in a ubiquitous computing environment: Effects of deficient self-regulation. Journal of Broadcasting & Electronic Media, 48, 399-420.

Lucas, K., & Sherry, J. L. (2004). Sex differences in video game play: A communication-based explanation. Communication Research, 31,499-523.

Massimini, F., & Carli, M. (1988). The systematic assessment of flow in daily experience. In M. Csikszentmihalyi & I. S. Csikszentmihalyi (Eds.), Optimal experience: Psychological studies of flow in consciousness (pp. 266-287). New York: Cambridge University Press.

Novak, T. P., & Hoffman, D. L. (1997). Measuring the flow experience among Web users. Retrieved December 24, 2005, from http://elab.vanderbilt.edu/Research/papers.html

Novak, T. P., Hoffman, D. L., & Yung, Y. (2000). Measuring the customer experience in online environments: A structural modeling approach. Marketing Science, 19, 22-44.

NPD Group (2006). The NPD group reports annual 2005 U.S. video game industry retail sales. Retrieved September 23, 2006, from http://www.npd.com/dynamic/releases/press_060117.html

Papper, R., Holmes, M., Popovich, M. (2004). Middetown media studies: Media, multitasking ... and how much people really use the media. The International Digital Media and Arts Association Journal, 1, 9-50.

Peele, S. (1995). Diseasing of America: How we allowed recovery zealots and the treatment industry to convince us we are out of control. Lexington, MA/San Francisco: Jossey-Bass.

Phillips, C. A., Rolls, S., Rouse, A., & Griffiths, M. D. (1995). Home video game playing in school children: A study of incidence and pattern of play. Journal of Adolescence, 18, 687-691.

Pingree, S., Hawkins, R. P., Bush Hitchon, J. C., Gilligan, E., Radler, B., Kahlor, L. et al. (2001). if college students are appointment television viewers ... Journal of Broadcasting & Electronic Media, 45, 446-463.

Rich, A. R., & Scovel, M. (1987). Causes of depression in college students: A cross-lagged panel correlational analysis. Psychological Reports, 60, 27-30.

Rosenstein, A. W., & Grant, A. E. (1997). Reconceptualizing the role of habit: A new model of television audience activity. Journal of Broadcasting & Electronic Media, 41, 324-344.

Rubin, A. M. (1984). Ritualized and instrumental television viewing. Journal of Communication, 34, 67-77.

Shaffer, H. J., Hall, M. N., & Vander Bilt, J. (2000). "Computer addiction": A critical consideration. American Journal of Orthopsychiatry, 70, 162-168.

Sherry, J. L. (2004). Flow and media enjoyment. Communication Theory. 14, 328-347.

Sherry, J. L., Lucas, K., Greenberg, B., & Lachlan, K. (2006). Video game uses and gratifications as predictors of use and game preference. In P. Vorderer & J. Bryant (Eds.), Playing computer games: Motives, responses, and consequences (pp. 213-224). Mahwah, N J: Lawrence Erlbaum Associates.

Stone, G., & Stone, D. B. (1990). Lurking in the literature: Another look at media use habits. Mass Communication Review, 17, 25-33.

Trevino, L. K., & Webster, J. (1992). Flow in computer-mediated communication: Electronic mail and voice mail, evaluation and impacts. Communication Research, 19, 539-573.

Triandis, H. C. (1980). Values, attitude and interpersonal behavior. In H. E. Howe & M. M. Page (Eds.), Nebraska symposium on motivation. (Vol. 37, pp. 195-259). Lincoln: University of Nebraska Press.

van Grinsven, L. (2003). Video games are addictive-scientists. Retrieved July 21, 2004, from http://www.wellspan.org/Health News/Reuters/20031104elin010.htm

Voiskounsky, A. E., Mitina, O. V., & Avetisova, A. A. (2004). Playing online games: Flow experience. Psychology Journal, 2, 259-281.

Webster, J., Trevino, L. K., & Ryan, L. (1993). The dimensionality and correlates of flow in human-computer interactions. Computers in Human Behavior, 9, 411-426.

Williams, D. (2002). A structural analysis of market competition in the U.S. home video game industry. International Journal on Media Management, 4, 41-54.

Zillmann, D. (1988). Mood management through communication choices. American Behavioral Scientist, 31,327-340.

Notes

(l) It excludes sales of DVDs and VCRs and rental and television revenue.

(2) The survey consists of 1,162 students at 25 colleges and universities.

(3) The primary principle is for each indicator to load on only one construct because it allows the most unambiguous assignment of meaning to the estimated constructs.

(4) The second-order factor model is a variation of factor analysis that parallels the first-order factor model mathematically (Kline, 1998). In the present study, the single second-order latent factor of flow experience accounts for the substantial covariation between the three first-order latent factors of enjoyment, merge, and concentrations.

(5) At first, the CFA test indicated that the overall goodness-of-fit indices of the measurement model were not acceptable. In respecifying the initial CFA model, two items were removed through reference to standardized residuals, cross-loadings, and modification indices as recommended by Hair, Anderson, Tatham, and Black