Computer Game Research

Discussion

Hypothesis 1 : Relationships Between Problem Video Game Playing, Social Skills and Self Esteem

Overall, there were not large and pronounced relationships found between a multidimensional measure of social skills, self esteem and problems associated with electronic game play. This suggests that social difficulties, specifically caused by a lack of social skills, do not underpin problematic play. Considering the strong relationships established between social difficulties, network size, competence and SSI scores, this finding casts doubt on the theory that avid (or perhaps dependant) video game players turn to games to compensate or cope with social difficulties. This finding gains credence from the sample’s high playing time per week and PVP total scores, indicating that they were avid gamers. An unlikely but possible alternative explanation is that the PVP is not measuring problems associated with video game play, or is not appropriately weighting the most deleterious problems, and thus only minute relationships were found between the two scales. If this is the case, further research into the construct of problem video game playing, including the enhancement of scales that weight intensity of associated problems, may confirm whether social skills, self esteem and video game playing problems are more strongly related.

Emergent Relationships

Nonetheless, some statistically significant relationships did emerge between the social, but not emotional, subscales of the SSI and the PVP total score. The correlation matrix indicates that those who experience more problematic play communicate no differently on non-verbal dimensions than their less problematic counterparts. However, higher problematic play was related to mild decreases in SE (r = -.134, N = 621, p = .001) and SC (r = -.217, N = 621, p < .001), and an increase in SS (r = .224, N = 621, p < .001). As expected, there was an inverse relationship between problematic play and the Rosenberg Self Esteem Scale (r = -.234, N = 621, p < .001). The degree of this relationship is similar to relationships found between self esteem and playing time in previous studies. However, the relationship found in this study between hours of play per week and the self esteem score is smaller than those found previously (r = -.091, N = 621, p < .01), suggesting problems associated with play are more strongly related to self esteem than playing time alone.

Interpretation Using Subscale Descriptions

According to the subscale definitions, the relationship with SE suggests that those who had higher problematic play scores are less socially and verbally engaging than those with lower PVP scores. The relationship found with SC suggests those with higher PVP scores are less adaptable to varying social situations, and less able to act in differing social roles. SS showed the strongest relationship with PVP of all subscales, and unlike others the direction was positive (r = .224, N = 621, p < .001). This indicates that as PVP scores increase, so does the participants’ skill in listening to others and awareness of social norms and rules. However, the correlation coefficients indicate that these relationships only explain a minimal 1.7, 4.7 and 5 percent, respectively, of the variance between the two measures. The correlation coefficient of the relationship between self esteem and PVP scores indicated that 5.4% of variance was shared.

Nonetheless, this pattern is particularly interesting, as to quote the definition of SS from the scale manual (Riggio & Carney, 2003):

Extremely high scores on this scale, in conjunction with moderate to low scores on Social Expressivity and Social Control, may indicate self-consciousness that may inhibit participation in social interaction (p. 4).

If taken together these relationships suggest those with higher problematic scores have mildly lower self-esteem, are less verbally expressive, less comfortable in varying social situations, better listeners and more sensitive to social norms; although all relationships are minute in magnitude. Causal directionality can only be hypothesized, it may be that increased time commitment associated with problematic play reduces avenues for face-to-face socialisation, or these characteristics represent risk factors preceding problematic play. To further explore relationships between these variables, a multiple regression was conducted.

Hypothesis 2: Predicting Problem Play with Self Esteem and Social Skills

Previous studies have questioned whether some gamers who have social difficulties may turn to game worlds to escape from their situation, or even complement socialisation. To explore this hypothesis, a prediction of problematic video game play was attempted using a measure of diverse social skills, as well as other variables of interest, including play time, sex and self esteem. Employing multiple regression (enter method), a significant model emerged: F (9, 611) = 15.051, p < .001, accounting for 16.9% of the variance in PVP total scores (adjusted R² = .169). The predictor accounting for the most variance in PVP scores was hours of play per week (β = .314, p < .001). Only one SSI subscale significantly predicted PVP scores - Social Sensitivity (SS), although the percentage of variance it accounted for was small (β = .184, p < .001). The RSES score also emerged as a significant predictor, but a smaller one (β = -.103, p < .03). These two predictors accounted for a small but significant prediction of problematic play scores, and may be tested further in models that already predict problematic play (Yee, in Press). Considering SS was the only SSI subscale to significantly predict PVP scores, this component of social skills comprises a small but important element of problematic play and warrants further analysis.

Understanding Social Sensitivity: Convergent Constructs and Previous Findings

By reviewing constructs related to SS, it is possible to speculate as to what other social or personality characteristics may be related to problematic video game play. In studies building validity for the SSI, relationships between SS, psychosocial health and personality were examined.

The inverse relationship between SS and psychosocial health variables.

Previous studies have indicated that the relationship between SS and psychosocial health deviates from the positive relationships found with the other five SSI subscales. SS was found to be inversely related to self esteem and well-being, and positively related to social anxiety (Riggio, Throckmorton, & DePaola, 1990). Specifically, SS was positively correlated with the subscales interaction anxiety (r = .25, N = 121, p < .01) and audience anxiety (r = .31, N = 121, p < .01) of the Leary Social Anxiety Inventory; it was also negatively correlated with the Rand Well-Being Scale (r = -.25, N = 121, p < .05) and the Coopersmith Self Esteem inventory (r = -.34, N = 121, p < .05). This finding was confirmed in the present study, as the RSES total score was negatively correlated with SS (r = -.39, N = 621, p < .001), but positively correlated with every other subscale. To quote the original definition of the subscale, Riggio (1986) states:

In extremes, high socially sensitive persons’ concern over appropriate social behaviour may lead to self-consciousness and social anxiety, which may inhibit the persons’ participation in social interaction (p. 651).

This explanation is further validated by relationships found between SS and the Fenigstein, Scheier and Buss measures of Public Self-Consciousness (r = .58, N = 149, p<.001) and Social Anxiety (r = .37, N = 149, p<.001). The relationship between Public Self-Consciousness and SS was the strongest in magnitude of all other psychosocial health variables tested (Riggio & Carney, 2003).

Personality variables and SS.

Research on how the SSI subscales relate to the 16PF and the Eysenck Personality Inventory yielded personality traits associated with each subscale (Riggio, 1986; Riggio & Carney, 2003). Those higher in SS are more affected by feelings (r = -.41, N = 149, p < .05), shy (r = -.26, N = 149, p < .05), astute (r = .40, N = 149, p < .05), apprehensive (r = .60, N = 149, p < .05), conservative (r = -.27, N = 149, p < .05) and tense (r = .54, N = 149, p < .05), which corroborate the definition of SS. Oddly, there is a mild inverse relationship between SS and the 16PF domain careless of social rules (r = -.23, N = 149, p < .05). This may indicate those with high SS scores are highly aware of social rules in the self and others, but do not necessarily conform to them.

Applied to problem game playing.

These relationships may give an indication of social and personality aspects predictive of a small amount of problematic play. In order of the magnitude of correlations with SS, those with high PVP scores may be more: apprehensive, publicly self-conscious, tense, affected by feelings, astute, socially anxious, conservative and shy; and have lower self esteem and less well-being. Ergo, as highlighted by the definition of the subscale, those higher in SS can become over-concerned with social norms and rules, as well as their public appearance, and thus exhibit elements of social anxiety or even possible social hesitance (Riggio, 1986; Riggio & Carney, 2003; Riggio, Throckmorton, & DePaola, 1990). The prediction of PVP scores by SS may indicate that a very small component of problematic play hinges on these personality characteristics. Although no correlations emerged between time commitment and SSI subscale scores while PVP scores did, in the multiple regression analysis time commitment accounted for more variance in the model than PVP scores, suggesting that when all variables are held constant, time commitment is more important. This may indicate that causal directionality is more related to a lack of face-to-face socialization due to playing time commitment, but this conclusion is only speculative.

It must be noted that dependable inferences of what social and personality characteristics perpetuate problematic play cannot be gained by investigating was has correlated with SS in past studies. It must again be stressed that the amount of variance in PVP scores explained by SS was very small. Future investigations are required to determine if any of these social and personality traits are directly important to problematic play.

Overview of Social Skills, Self Esteem and Problem Electronic Game Play

On face value, a pattern of relationships emerged between PVP scores and social characteristics that suggest a social impairment, resonating with the notion of games compensating or providing escape from social difficulties. However, the magnitude of these relationships was very small. In light of the lack of larger relationships between the SSI subscales and PVP scores, as well as the relatively small proportion of variance explained by SS in the multiple regression analysis, it must be stressed that problem gamers do not seem to have severe social difficulties. However, as SS was the only subscale to emerge as a significant positive predictor from six diverse subscales, and as it is related to increases in public self-consciousness, apprehensiveness and social anxiety, and decreases in well-being and self esteem, then this single predictive subscale does corroborate previous studies that found relationships between problem game play, a desire to escape, introversion and social anxiety (Barnett et al., 1997; Lo, Wang, & Fang, 2005; Wan & Chiou, 2006; Wood, Gupta, Derevensky, & Griffiths, 2004; Yee, in Press). It is possible that a very small amount of problematic play is encouraged or maintained by traits representing a social hesitancy. This dovetails with explanations of social electronic games reducing anxiety cues and making communication less confronting and more personal (McKenna & Bargh, 2000; Young, 2004). However, the proportion explained by these social traits is minimal, and future research is required to elaborate variables related to problematic play.

Hypothesis 3: Social Skills and Time Spent Playing

Surprisingly, correlational analyses revealed no indication that time spent playing online games is related to social skills. This lack of significant relationships between time spent playing may be an indication that a measure of traditional social skills is unable to detect skills learnt through online socialisation, or that online environments are not as socially enriching as suggested by some (Steinkuehler & Williams, 2006). Further research on the role of virtual communications on psychological development is recommended to explore the relationship between time spent playing and social skills.

Further Explorations: Gender, Age, Age at First Played Electronic Games and Genre

To elaborate on this contemporary area of study, comparisons of genre on all variables of interest were performed. They indicate that MMORPG players had the highest time commitment, PVP scores and age, significantly higher than other genres. This suggests MMORPGs are popular among an older audience compared to other games, confirming recent demographic studies (Griffiths, Davies, & Chappell, 2003).

The finding that MMORPGs had significantly higher PVP scores and time commitment indicates that there is something particularly compelling or problem-inducing about MMORPGs compared to other genres, corroborating evidence that interactive internet functions are more ‘addictive’, and validating the current research focus on MMORPGs (Armstrong, Phillips, & Saling, 2000; McKenna & Bargh, 2000). However, this finding may also simply represent the most popular genre of game at the present time, or within this sample. It is logical that the highest indications of problematic play and time commitment will be present in the most popular genre of the time, as that genre is likely to attract the most avid gamers. Interestingly, despite the gender discrepancy in the sample, males did not show significantly higher time commitment compared to females according to t-tests. This may be an indication of the further proliferation of electronic games.

Limitations and Suggestions for Future Research

Sample Bias

Although forums catering for a range of gaming systems were utilized, the sample obtained does not represent as broad a cross section of gamers as sought. Even though many systems are identified as secondary and tertiary systems played, the PC makes up the great majority of the primary systems identified (82.8%). This is likely because the medium of the study was a computer. The average hours spent playing per day (M = 3.42, SD = 2.48) and per week (M = 23.78, SD = 16.88) were generally higher than previous studies. Similarly, the mean PVP score is higher than would have been expected judging by prevalence of problematic play in past studies (M = 4.17, SD = 2.19).

This could be the result of the study invitations being posted on forums devoted to gaming, hence targeting gamers who spend time in meta-game environments that allow for further exploration of game content, and who are therefore likely to be more committed gamers. According to the endorsement of individual PVP items, 56.5% of the sample spent time outside of play persistently thinking and planning their game playing, suggesting play was a salient and meaningful activity to much of the sample. Those that play console games may be less likely to populate meta-game environments, as they would require a shift of medium to a computer to access them, although this may change with the advent of internet browsing capability in home game consoles. These sample characteristics suggest that the results are more applicable to highly involved PC gamers rather than less avid or console players, and generalisations to all gamers must be made cautiously.

Problem Video Game Playing Scale

The problem video game playing scale (or PVP) is dichotomous and unable to measure intensity of problems affirmed. Also, without extensive validation, it cannot be verified that the scale is measuring all elements of the construct, including negative effects. More investigations into the construct of problematic play are required to determine its epidemiology, longitudinal fluctuations, negative consequences and mediating variables. Qualitative research tapping problematic players’ perspectives may help to identify risk factors, including what real life problems are commonly being escaped from; symptomatology, including what negative consequences of problematic play are most intrusive; as well as potential co-morbidities and maintenance factors which may then help to inform scale development. Scale development efforts could draw from qualitative studies, use a response method that is sensitive to degrees such as Likert-type scales, and perform factor analyses on the pooled items of all previous “video game addiction” scales with self-reported addicts to further pinpoint what composes problematic play.

Research Collaboration

Considering the lack of research and consequent research tools concerning electronic game play, collaboration is required among those pioneering investigations in this field. Currently the largest base of psychologically research on electronic games is based online, and not peer reviewed! (Collaboration may help to focus future research direction by prioritizing issues concerning electronic game play, including scale development, providing a base from which future researchers can build from, and more broadly to stimulate investigations into the roles that electronic games have in the lives of those that use them.

Conclusion

This study aimed to investigate the relationship between problematic electronic game play, social skills and self esteem in an adult sample. Confirming past studies, an inverse relationship (and prediction) was present between problematic play and self esteem, but the magnitude was small (Barnett et al., 1997; Colwell, Grady & Rhaiti, 1995; Colwell & Pain, 2000). Also resonating with past studies, a pattern of relationships emerged between the social skills subscales and problematic play indicating a social hesitancy (Barnett et al., 1997; Lo, Wang & Fang, 2005). Further, the subscale SS (which is related to public self-consciousness and social anxiety) accounted for a small prediction of problem video game playing scores. Taken together, these results may indicate a deleterious consequence, or possible risk factor associated with problematic play, although considering the high PVP scores and time commitment within the sample, stronger relationships would have been expected to justify the hypothesis that problematic play is impelled by social difficulties. Ultimately the magnitude of the relationships and predictions found was minimal; suggesting that other factors aside from social skills and self esteem will better predict problematic play.

The findings of this study cast doubt on the notion that gamers with social difficulties turn to game worlds to cope with or escape from these problems. This resonates with past studies that found little negative consequences to the construct of problematic or dependant play (Fisher, 1994; Griffiths, 2000; Griffiths & Hunt, 1998). Despite anecdotal accounts and clinical examples (Allison, Von Wahlde, Shockley, & Gabbard, 2006; Griffiths, 2000), this consistent lack of severe negative consequences suggests that research has either overlooked important areas of this construct, or taken an unnecessarily pathological approach.


Overview Participants Materials Method Results Discussion References Links Contact	Discussion Discussion Hypothesis 1 : Relationships Between Problem Video Game Playing, Social Skills and Self Esteem Overall, there were not large and pronounced relationships found between a multidimensional measure of social skills, self esteem and problems associated with electronic game play. This suggests that social difficulties, specifically caused by a lack of social skills, do not underpin problematic play. Considering the strong relationships established between social difficulties, network size, competence and SSI scores, this finding casts doubt on the theory that avid (or perhaps dependant) video game players turn to games to compensate or cope with social difficulties. This finding gains credence from the sample’s high playing time per week and PVP total scores, indicating that they were avid gamers. An unlikely but possible alternative explanation is that the PVP is not measuring problems associated with video game play, or is not appropriately weighting the most deleterious problems, and thus only minute relationships were found between the two scales. If this is the case, further research into the construct of problem video game playing, including the enhancement of scales that weight intensity of associated problems, may confirm whether social skills, self esteem and video game playing problems are more strongly related. Emergent Relationships Nonetheless, some statistically significant relationships did emerge between the social, but not emotional, subscales of the SSI and the PVP total score. The correlation matrix indicates that those who experience more problematic play communicate no differently on non-verbal dimensions than their less problematic counterparts. However, higher problematic play was related to mild decreases in SE (r = -.134, N = 621, p = .001) and SC (r = -.217, N = 621, p < .001), and an increase in SS (r = .224, N = 621, p < .001). As expected, there was an inverse relationship between problematic play and the Rosenberg Self Esteem Scale (r = -.234, N = 621, p < .001). The degree of this relationship is similar to relationships found between self esteem and playing time in previous studies. However, the relationship found in this study between hours of play per week and the self esteem score is smaller than those found previously (r = -.091, N = 621, p < .01), suggesting problems associated with play are more strongly related to self esteem than playing time alone. Interpretation Using Subscale Descriptions According to the subscale definitions, the relationship with SE suggests that those who had higher problematic play scores are less socially and verbally engaging than those with lower PVP scores. The relationship found with SC suggests those with higher PVP scores are less adaptable to varying social situations, and less able to act in differing social roles. SS showed the strongest relationship with PVP of all subscales, and unlike others the direction was positive (r = .224, N = 621, p < .001). This indicates that as PVP scores increase, so does the participants’ skill in listening to others and awareness of social norms and rules. However, the correlation coefficients indicate that these relationships only explain a minimal 1.7, 4.7 and 5 percent, respectively, of the variance between the two measures. The correlation coefficient of the relationship between self esteem and PVP scores indicated that 5.4% of variance was shared. Nonetheless, this pattern is particularly interesting, as to quote the definition of SS from the scale manual (Riggio & Carney, 2003): Extremely high scores on this scale, in conjunction with moderate to low scores on Social Expressivity and Social Control, may indicate self-consciousness that may inhibit participation in social interaction (p. 4). If taken together these relationships suggest those with higher problematic scores have mildly lower self-esteem, are less verbally expressive, less comfortable in varying social situations, better listeners and more sensitive to social norms; although all relationships are minute in magnitude. Causal directionality can only be hypothesized, it may be that increased time commitment associated with problematic play reduces avenues for face-to-face socialisation, or these characteristics represent risk factors preceding problematic play. To further explore relationships between these variables, a multiple regression was conducted. Hypothesis 2: Predicting Problem Play with Self Esteem and Social Skills Previous studies have questioned whether some gamers who have social difficulties may turn to game worlds to escape from their situation, or even complement socialisation. To explore this hypothesis, a prediction of problematic video game play was attempted using a measure of diverse social skills, as well as other variables of interest, including play time, sex and self esteem. Employing multiple regression (enter method), a significant model emerged: F (9, 611) = 15.051, p < .001, accounting for 16.9% of the variance in PVP total scores (adjusted R² = .169). The predictor accounting for the most variance in PVP scores was hours of play per week (β = .314, p < .001). Only one SSI subscale significantly predicted PVP scores - Social Sensitivity (SS), although the percentage of variance it accounted for was small (β = .184, p < .001). The RSES score also emerged as a significant predictor, but a smaller one (β = -.103, p < .03). These two predictors accounted for a small but significant prediction of problematic play scores, and may be tested further in models that already predict problematic play (Yee, in Press). Considering SS was the only SSI subscale to significantly predict PVP scores, this component of social skills comprises a small but important element of problematic play and warrants further analysis. Understanding Social Sensitivity: Convergent Constructs and Previous Findings By reviewing constructs related to SS, it is possible to speculate as to what other social or personality characteristics may be related to problematic video game play. In studies building validity for the SSI, relationships between SS, psychosocial health and personality were examined. The inverse relationship between SS and psychosocial health variables. Previous studies have indicated that the relationship between SS and psychosocial health deviates from the positive relationships found with the other five SSI subscales. SS was found to be inversely related to self esteem and well-being, and positively related to social anxiety (Riggio, Throckmorton, & DePaola, 1990). Specifically, SS was positively correlated with the subscales interaction anxiety (r = .25, N = 121, p < .01) and audience anxiety (r = .31, N = 121, p < .01) of the Leary Social Anxiety Inventory; it was also negatively correlated with the Rand Well-Being Scale (r = -.25, N = 121, p < .05) and the Coopersmith Self Esteem inventory (r = -.34, N = 121, p < .05). This finding was confirmed in the present study, as the RSES total score was negatively correlated with SS (r = -.39, N = 621, p < .001), but positively correlated with every other subscale. To quote the original definition of the subscale, Riggio (1986) states: In extremes, high socially sensitive persons’ concern over appropriate social behaviour may lead to self-consciousness and social anxiety, which may inhibit the persons’ participation in social interaction (p. 651). This explanation is further validated by relationships found between SS and the Fenigstein, Scheier and Buss measures of Public Self-Consciousness (r = .58, N = 149, p<.001) and Social Anxiety (r = .37, N = 149, p<.001). The relationship between Public Self-Consciousness and SS was the strongest in magnitude of all other psychosocial health variables tested (Riggio & Carney, 2003). Personality variables and SS. Research on how the SSI subscales relate to the 16PF and the Eysenck Personality Inventory yielded personality traits associated with each subscale (Riggio, 1986; Riggio & Carney, 2003). Those higher in SS are more affected by feelings (r = -.41, N = 149, p < .05), shy (r = -.26, N = 149, p < .05), astute (r = .40, N = 149, p < .05), apprehensive (r = .60, N = 149, p < .05), conservative (r = -.27, N = 149, p < .05) and tense (r = .54, N = 149, p < .05), which corroborate the definition of SS. Oddly, there is a mild inverse relationship between SS and the 16PF domain careless of social rules (r = -.23, N = 149, p < .05). This may indicate those with high SS scores are highly aware of social rules in the self and others, but do not necessarily conform to them. Applied to problem game playing. These relationships may give an indication of social and personality aspects predictive of a small amount of problematic play. In order of the magnitude of correlations with SS, those with high PVP scores may be more: apprehensive, publicly self-conscious, tense, affected by feelings, astute, socially anxious, conservative and shy; and have lower self esteem and less well-being. Ergo, as highlighted by the definition of the subscale, those higher in SS can become over-concerned with social norms and rules, as well as their public appearance, and thus exhibit elements of social anxiety or even possible social hesitance (Riggio, 1986; Riggio & Carney, 2003; Riggio, Throckmorton, & DePaola, 1990). The prediction of PVP scores by SS may indicate that a very small component of problematic play hinges on these personality characteristics. Although no correlations emerged between time commitment and SSI subscale scores while PVP scores did, in the multiple regression analysis time commitment accounted for more variance in the model than PVP scores, suggesting that when all variables are held constant, time commitment is more important. This may indicate that causal directionality is more related to a lack of face-to-face socialization due to playing time commitment, but this conclusion is only speculative. It must be noted that dependable inferences of what social and personality characteristics perpetuate problematic play cannot be gained by investigating was has correlated with SS in past studies. It must again be stressed that the amount of variance in PVP scores explained by SS was very small. Future investigations are required to determine if any of these social and personality traits are directly important to problematic play. Overview of Social Skills, Self Esteem and Problem Electronic Game Play On face value, a pattern of relationships emerged between PVP scores and social characteristics that suggest a social impairment, resonating with the notion of games compensating or providing escape from social difficulties. However, the magnitude of these relationships was very small. In light of the lack of larger relationships between the SSI subscales and PVP scores, as well as the relatively small proportion of variance explained by SS in the multiple regression analysis, it must be stressed that problem gamers do not seem to have severe social difficulties. However, as SS was the only subscale to emerge as a significant positive predictor from six diverse subscales, and as it is related to increases in public self-consciousness, apprehensiveness and social anxiety, and decreases in well-being and self esteem, then this single predictive subscale does corroborate previous studies that found relationships between problem game play, a desire to escape, introversion and social anxiety (Barnett et al., 1997; Lo, Wang, & Fang, 2005; Wan & Chiou, 2006; Wood, Gupta, Derevensky, & Griffiths, 2004; Yee, in Press). It is possible that a very small amount of problematic play is encouraged or maintained by traits representing a social hesitancy. This dovetails with explanations of social electronic games reducing anxiety cues and making communication less confronting and more personal (McKenna & Bargh, 2000; Young, 2004). However, the proportion explained by these social traits is minimal, and future research is required to elaborate variables related to problematic play. Hypothesis 3: Social Skills and Time Spent Playing Surprisingly, correlational analyses revealed no indication that time spent playing online games is related to social skills. This lack of significant relationships between time spent playing may be an indication that a measure of traditional social skills is unable to detect skills learnt through online socialisation, or that online environments are not as socially enriching as suggested by some (Steinkuehler & Williams, 2006). Further research on the role of virtual communications on psychological development is recommended to explore the relationship between time spent playing and social skills. Further Explorations: Gender, Age, Age at First Played Electronic Games and Genre To elaborate on this contemporary area of study, comparisons of genre on all variables of interest were performed. They indicate that MMORPG players had the highest time commitment, PVP scores and age, significantly higher than other genres. This suggests MMORPGs are popular among an older audience compared to other games, confirming recent demographic studies (Griffiths, Davies, & Chappell, 2003). The finding that MMORPGs had significantly higher PVP scores and time commitment indicates that there is something particularly compelling or problem-inducing about MMORPGs compared to other genres, corroborating evidence that interactive internet functions are more ‘addictive’, and validating the current research focus on MMORPGs (Armstrong, Phillips, & Saling, 2000; McKenna & Bargh, 2000). However, this finding may also simply represent the most popular genre of game at the present time, or within this sample. It is logical that the highest indications of problematic play and time commitment will be present in the most popular genre of the time, as that genre is likely to attract the most avid gamers. Interestingly, despite the gender discrepancy in the sample, males did not show significantly higher time commitment compared to females according to t-tests. This may be an indication of the further proliferation of electronic games. Limitations and Suggestions for Future Research Sample Bias Although forums catering for a range of gaming systems were utilized, the sample obtained does not represent as broad a cross section of gamers as sought. Even though many systems are identified as secondary and tertiary systems played, the PC makes up the great majority of the primary systems identified (82.8%). This is likely because the medium of the study was a computer. The average hours spent playing per day (M = 3.42, SD = 2.48) and per week (M = 23.78, SD = 16.88) were generally higher than previous studies. Similarly, the mean PVP score is higher than would have been expected judging by prevalence of problematic play in past studies (M = 4.17, SD = 2.19). This could be the result of the study invitations being posted on forums devoted to gaming, hence targeting gamers who spend time in meta-game environments that allow for further exploration of game content, and who are therefore likely to be more committed gamers. According to the endorsement of individual PVP items, 56.5% of the sample spent time outside of play persistently thinking and planning their game playing, suggesting play was a salient and meaningful activity to much of the sample. Those that play console games may be less likely to populate meta-game environments, as they would require a shift of medium to a computer to access them, although this may change with the advent of internet browsing capability in home game consoles. These sample characteristics suggest that the results are more applicable to highly involved PC gamers rather than less avid or console players, and generalisations to all gamers must be made cautiously. Problem Video Game Playing Scale The problem video game playing scale (or PVP) is dichotomous and unable to measure intensity of problems affirmed. Also, without extensive validation, it cannot be verified that the scale is measuring all elements of the construct, including negative effects. More investigations into the construct of problematic play are required to determine its epidemiology, longitudinal fluctuations, negative consequences and mediating variables. Qualitative research tapping problematic players’ perspectives may help to identify risk factors, including what real life problems are commonly being escaped from; symptomatology, including what negative consequences of problematic play are most intrusive; as well as potential co-morbidities and maintenance factors which may then help to inform scale development. Scale development efforts could draw from qualitative studies, use a response method that is sensitive to degrees such as Likert-type scales, and perform factor analyses on the pooled items of all previous “video game addiction” scales with self-reported addicts to further pinpoint what composes problematic play. Research Collaboration Considering the lack of research and consequent research tools concerning electronic game play, collaboration is required among those pioneering investigations in this field. Currently the largest base of psychologically research on electronic games is based online, and not peer reviewed! (Collaboration may help to focus future research direction by prioritizing issues concerning electronic game play, including scale development, providing a base from which future researchers can build from, and more broadly to stimulate investigations into the roles that electronic games have in the lives of those that use them. Conclusion This study aimed to investigate the relationship between problematic electronic game play, social skills and self esteem in an adult sample. Confirming past studies, an inverse relationship (and prediction) was present between problematic play and self esteem, but the magnitude was small (Barnett et al., 1997; Colwell, Grady & Rhaiti, 1995; Colwell & Pain, 2000). Also resonating with past studies, a pattern of relationships emerged between the social skills subscales and problematic play indicating a social hesitancy (Barnett et al., 1997; Lo, Wang & Fang, 2005). Further, the subscale SS (which is related to public self-consciousness and social anxiety) accounted for a small prediction of problem video game playing scores. Taken together, these results may indicate a deleterious consequence, or possible risk factor associated with problematic play, although considering the high PVP scores and time commitment within the sample, stronger relationships would have been expected to justify the hypothesis that problematic play is impelled by social difficulties. Ultimately the magnitude of the relationships and predictions found was minimal; suggesting that other factors aside from social skills and self esteem will better predict problematic play. The findings of this study cast doubt on the notion that gamers with social difficulties turn to game worlds to cope with or escape from these problems. This resonates with past studies that found little negative consequences to the construct of problematic or dependant play (Fisher, 1994; Griffiths, 2000; Griffiths & Hunt, 1998). Despite anecdotal accounts and clinical examples (Allison, Von Wahlde, Shockley, & Gabbard, 2006; Griffiths, 2000), this consistent lack of severe negative consequences suggests that research has either overlooked important areas of this construct, or taken an unnecessarily pathological approach.