Patterns of Choices on Video Game Genres and Internet Addiction

Abstract

In this study, we attempt to identify motives pivotal in choices made regarding online and offline game genres and assess whether they are meaningful predictors of Internet addiction. A separate goal was to determine how an assessment of Internet-related cognitions compares to criteria-based research instruments on Internet addiction, and demonstrate its clinical usefulness. We are using data from a cross-sectional study of the entire adolescent student population aged 12–18 of the island of Kos. Results indicate that specific game genres and motives for choosing a game are important predictors of Internet addiction, even after accounting for sociodemographic and Internet-use-related variables. Gender differences are not statistically important when we account for genre choices and motives for choosing them. Boys' thought content tends to revolve more around their Internet use compared to girls. Several patterns of motives to play specific genres were identified with the help of a canonical correlation analysis, demonstrating the fact that the simple observation of overt behavior is not enough to deduce the meaning of these actions for the individual. Using multiple measures of online addiction, examining genre preferences and underlying motives for choosing a game can be helpful in reaching a better understanding of the individual game player.

Introduction

Online gaming has risen as a multibillion dollar entertainment industry with a percentage of players demonstrating significant addictive tendencies. Recent large scale research in Holland¹ and in Germany² have shown that while a large number of players may admit some negative impact in their lives, only a smaller percent can be classified as addicted, with boys outnumbering girls as much as 10 to 1. Absolute figures varied from 9.3 percent in a 2004 study limited to players of a single multiuser game,³ to 3 percent of the male and 0.3 percent of the female adolescent students in a larger scale study with a controlled sample in 2010.² This variation is attributed to the nature of each sample and the method employed to demarcate video game addiction from lighter involvement with game playing.

It is worth noting that although two individuals may manifest the same behavioral expressions of over indulgence with online gaming, we should still remember that they are individuals with their own motives and surrounding environment. The issue of the context in which a behavior is observed was noted by Griffiths in the presentation of two individuals who outwardly demonstrated excessive preoccupation with gaming; one did not had any negative consequences, while the other did.⁴ Their underlying motives for playing, however, differed considerably; while the first had little to do socially and was killing time, the other player tried to escape from existing problems.

Previous efforts to comprehend why players play specific online games have established that players favor both personal and social interaction, with and within the game.⁵ According to Yee three components, the achievement, social, and immersion component were instrumental in participation in MMORPGs.⁶ The importance of motivational differences for engaging in online gaming has been demonstrated in two related studies in a self-determination framework.^7,8 According to them, the appeal of gaming lies in the satisfaction of basic psychological needs, like autonomy, competence, and relatedness. Liu and Peng⁹ demonstrated the important roles that psychological dependency and deficient self-regulation play in negative consequences associated with online gaming. Their results also indicated that psychological dependency on massively multiplayer online games was predicted by a cognitive preference for a virtual life—a construct that is negatively related to social control skills.

Davis proposed a framework for problematic Internet use (PIU), which for the first time focused on cognitions; PIU results from problematic cognitions coupled with behaviors that either intensify or maintain the maladaptive response.¹⁰ Specific and generalized PIU was discerned in this model on the basis of whether the subject engaged in a single activity or had multiple favorite online activities. The construct of PIU was multifaceted with four factors described¹¹: social comfort, loneliness/depression, diminished impulse control, and distraction. Each factor relates to a pattern of associated cognitions that are more prominent in the subject's Internet use. Davis and colleagues devised a scale, the Online Cognitions Scale (OCS), to measure these constructs, claiming that it was a valid test of Internet addiction in an organizational sample.¹¹ This scale focuses on internal cognitions rather than external behavior and may provide a clinician with a list of maladaptive cognitions to target in a cognitive behavioral intervention. Its relatedness to online games has not been researched so far. This is a different approach to the one taken initially by Young with criteria-based scales like the Young Diagnostic Questionnaire (YDQ)¹² and advanced by separate efforts by Shapira¹³ and Tao,¹⁴ among others. A criteria-based diagnosis tends to focus on readily observed behaviors with explicit items (for example, a YDQ item directly queries about having lied to important others), which may tempt the individual to answer untruthfully, while the OCS has more innocuous statements like, “The Internet is an important part of my life.”

It is thus apparent that online video game playing contributes to a large extent to the notion of “Internet addiction,” especially among adolescent boys. The common element with those addicted to another online activity would be their cognitions, their thought content. This should point to the underlying etiology for turning to their particular set of online activities, be it feeling of social comfort, alleviation of loneliness or depression, diminished impulse control, and distraction from everyday problems. The purpose of this study is to examine this link in an adolescent sample so as to gain a better understanding of the typology and individual motives that lie in between deep rooted needs and overt behavior.

Research questions were formulated as follows:

(a) What are the favorite game genres of the adolescents and the relevant motives?

(b) How does an assessment of Internet-related cognitions compare to criteria-based research instruments on Internet addiction?

(d) Are those specific genres and motives helpful in predicting the levels of Internet use after controlling for related factors?

Materials and Methods

Study design and population

The study is part of a larger research project, Hippocrates 2010, focusing on online and offline behaviors of the youth in the island of Kos. It was designed by the Hellenic Association for the Study of Internet Addiction Disorder in collaboration with the Drug abuse prevention center Hippocrates of the Greek Organization against Illicit Drugs (OKANA). Previous research in Kos has indicated high percentages of Internet addiction symptomatology, that correlated with offline antisocial behaviors¹⁵ and chemical drug use experience.¹⁶ This study was of a cross-sectional design. All of the island's high schools participated in the study. The research material was distributed in schools; participation was voluntary and confidential. A target sample thus included all high-school students, with the research sample consisting of 2,017 teen students between 12 and 19 years of age, who were those students present at classes, an estimated 90 percent of the total high-school student population.

Measures

Students were handed the demographics questionnaire, a questionnaire on motives for favoring a game genre, the OCS and the YDQ. Demographics questionnaires included questions on sex, age, family background, school performance, and related goals.

The motives for favoring a particular game questionnaire contains 13 items rated on a six-point Likert scale, ranging from zero points, for total lack of interest for the item, to five points, for considering the item as essential to the gaming experience. It was formed after a pilot study, where high-school students of the island who attended oral presentations on Internet safety were asked to anonymously identify reasons for playing games in free text.

The OCS is a theory-driven, multidimensional measure of PIU.¹¹ The scale includes four factors: social comfort, loneliness/depression, diminished impulse control, and distraction. This is the first application of this scale in a Greek population after an appropriate license for its adaptation was acquired by the original author. It is also the first application of the scale in an adolescent sample.

The eight-question YDQ (Diagnostic Questionnaire for Internet Addiction)¹² is based on criteria from the 4th version of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) on gambling; it is the oldest and most widely used set of criteria for Internet addiction.

Data analysis

This included demographics, reliability, and validity statistics with the employment of a confirmatory factor analysis (CFA), a receiver-operating characteristic (ROC) analysis for cut-off determination, correlational and regression statistics, and canonical correlation^17,18 that was employed to discern the commonalities between motives for choosing a game and specific game types, so as to have a clearer picture of who plays what and why. The IBM SPSS Statistics 20 package¹⁹ was used for all data analysis, except for CFA analyses where the EQS software package²⁰ is used.

Results

Demographics and questionnaire results

Two thousand seventeen students were asked to participate in the survey. One thousand nine hundred seventy-one reported using the Internet (97.71 percent) and were included. One thousand nineteen were boys (51.7 percent, mean age 15.06 years, standard error [SE]=0.053) and 952 girls (48.3 percent, mean age 15.09, SE=0.055). Sample demographics, frequency of online gaming, and mean values for the OCS scale are presented in Table 1.

Table 1.

Demographics and Internet Preferences of Adolescent Participants

Variable	Boys	Girls
Sex	1,019 (51.7%)	952 (48.3%)
Age (mean, SE)	15.06 (0.053)	15.09 (0.055)
12 years	34 (3.3%)	36 (3.8%)
13 years	175 (17.2%)	163 (17.1%)
14 years	208 (20.4%)	186 (19.5%)
15 years	211 (20.7%)	166 (17.4%)
16 years	178 (17.5%)	187 (19.6%)
17 years	131 (12.9%)	143 (15%)
18 years	60 (5.9%)	61 (6.4%)
19 years	22 (2.2%)	10 (1%)
Having used the Internet for gaming
Almost never	112 (11.4%)	280 (30.2%)
Occasionally during the year	59 (6%)	107 (11.5%)
Couple of times in a month	107 (10.9%)	119 (12.8%)
At least once a week	163 (16.6%)	159 (17.1%)
Almost daily	226 (23.1%)	153 (16.5%)
A lot of times during a day	312 (31.9%)	110 (11.9%)
OCS mean scores (SD)
Total score	105.7226 (43.12)	87.9 (40.44)
Social comfort PIU	36.665 (16.8)	29.277 (14.99)
Lonely/depressed PIU	16.346 (8.188)	13.239 (7.47)
Impulsive PIU	30.489 (12.259)	25.444 (12.49)
Distraction PIU	22.785 (9.77)	20.563 (10.29)

SE, standard error; SD, standard deviation; OCS, Online Cognitions Scale; PIU, problematic Internet use.

Reliability and validity analyses

Cronbach's alpha for the YDQ was 0.772 and for the motives for choosing a particular video game 0.892. The OCS total score had a Cronbach's alpha of 0.951, whereas social comfort PIU had 0.886, lonely/depressed 0.793, impulsive 0.838, and distraction 0.831. Those results were similar to the original validation study¹¹ and other adaptation efforts.²¹ Results from the YDQ and OCS scale correlated significantly; Spearman r_s for the total score was 0.701 for p<0.001.

We examined total fit of the OCS four-factor structure by way of CFA. Results indicated that the four-factor structure provided excellent fit to the data at hand (χ²=3.197, 2 df, p=0.202) Fit indices were: Adjusted Goodness of Fit Index=0.995, Comparative Fit Index=0.999, standardized Root Mean Square Residual=0.003, Root Mean Square Error of Approximation=0.019 (90 percent confidence intervals 0.001–0.0056).

Determination of cut-off points is typically carried out by comparison with a golden standard measure. An ROC curve analysis was performed on the total OCS score with the cut-point set at the corresponding value of 5+ in the YDQ total score, a value that corresponds to addiction. Results indicated considerable variance between boys and girls, a finding replicated in other studies with the OCS²¹; thus, we decided to use separate cut-off criteria for the two sexes. Girls' results were straightforward, with the area under the curve at 0.888 (SE 0.018), 95 percent confidence intervals between 0.854–0.923. With a cut-off value set at 97.5, sensitivity was 90.4 percent and specificity 76.3 percent. With boys, however, the area under the curve was 0.793 (SE 0.019), 95 percent confidence intervals 0.756–0.830, and the best compromise between sensitivity and specificity was reached at 119.5 (sensitivity 75.5 percent and specificity 71.7 percent). Apparently, boys are a harder population to diagnose with just a questionnaire. An alternate cut-off value for higher sensitivity (89.9 percent) would be 86.5, but that would lead to low specificity (47.1 percent) and considerable overestimation of true positives. This wide variability in boys' scores demonstrates the fact that boys are at greater risk for addictive Internet use, but their baseline of normal use may be higher than girls' or, better stated, their thought content tends to revolve more around their Internet use when compared to girls.

Differences between the sexes on gaming habits

Boys were more avid gamers (Table 1). We thus run separate analyses on the preferred genre and stated motives to choose a video game. Results are presented in Table 2, ranked by effect sizes for the differences. Boys tended to be more frequent gamers in all genres except for karaoke-style gaming. Games on Facebook, life management, and thought games were equally likely to be played by either sex. Boys were also more highly motivated to play with higher means on every related question.

Table 2.

Comparison of Mean Scores Between Boys and Girls on Favorite Video Games and Their Game-Related Preferences

Favorite video game	Boys mean (SE)	Girls mean (SE)	t-test	Df	p Value	Effect size (d)
Combat simulation (e.g., Counterstrike, Call of duty)	2.57 (0.060)	0.43 (0.037)	29.364	1,909	<0.001	1.386
Sports (FIFA10, Pro Evolution, NBA 2010)	2.82 (0.060)	0.85 (0.047)	25.528	1,893	<0.001	1.192
Driving Simulation (e.g., Need for Speed, Colin Mac Rae)	2.28 (0.059)	0.80 (0.047)	19.422	1,862	<0.001	0.913
Adventure games (e.g., Grand Theft Auto, Assasin's Creed)	2.36 (0.060)	0.95 (0.051)	17.659	1,867	<0.001	0.827
MMORPG (e.g., World of Warcraft, AION, Everquest)	1.45 (0.060)	0.29 (0.030)	16.850	1,843	<0.001	0.806
Strategy games (e.g., Total War, World in conflict)	1.69 (0.059)	0.55 (0.041)	15.524	1,837	<0.001	0.734
Martial arts (Street Fighter, Tekken)	1.64 (0.058)	0.60 (0.043)	14.229	1,854	<0.001	0.671
Flight simulation (e.g., Hawx, Flight Simulator)	1.20 (0.054)	0.41 (0.036)	11.905	1,830	<0.001	0.565
RPG (Dragon Age, Diablo)	1.20 (0.054)	0.43 (0.035)	11.708	1,853	<0.001	0.553
Games in mobile phones and handhelds	1.93 (0.060)	1.38 (0.055)	6.688	1,878	<0.001	0.310
Platform games/riddle solving (e.g., Mario, Tomb Raider)	1.39 (0.055)	1.05 (0.050)	4.562	1,845	<0.001	0.214
Music simulators—karaoke (e.g., Beatles, Guitar Hero)	1.02 (0.052)	1.19 (0.054)	−2.282	1,844	0.023	0.106
Facebook games (e.g., farmville, cafeworld)	1.82 (0.062)	1.76 (0.063)	0.653	1,870	0.514	0.030
Life management (e.g., Second Life, Sims)	1.07 (0.051)	1.11 (0.052)	−0.596	1,854	0.552	0.028
Thought games (e.g., chess, tetris, item spotting)	1.26 (0.051)	1.25 (0.051)	0.177	1,863	0.859	0.008
Motivation for favoring a particular online video game
Enjoy the competition with other players	3.24 (0.052)	1.67 (0.053)	21.043	1,887	<0.001	0.969
Enjoy being a member of the best teams and considered an elite player	2.83 (0.059)	1.40 (0.055)	17.686	1,860	<0.001	0.824
Enjoy the competition against the computer	2.14 (0.055)	0.94 (0.046)	16.570	1,874	<0.001	0.774
Like cooperation with other players	3.32 (0.048)	2.16 (0.053)	16.288	1,874	<0.001	0.751
Become renown for my gaming successes	2.70 (0.059)	1.38 (0.056)	16.029	1,867	<0.001	0.746
Like being able to break the rules of the game	1.84 (0.060)	0.73 (0.045)	14.571	1,853	<0.001	0.687
Enjoy doing something that is not allowed in real life	1.75 (0.058)	0.73 (0.045)	13.649	1,851	<0.001	0.642
Like making friends through the game	2.52 (0.057)	1.71 (0.056)	10.144	1,858	<0.001	0.472
Seek the creation of an alternate identity	1.68 (0.056)	0.96 (0.049)	9.605	1,848	<0.001	0.450
Enjoy feeling alive, intensity, absorbed in its flow	2.38 (0.059)	1.83 (0.060)	6.590	1,857	<0.001	0.306
Like following clear, just rules	2.62 (0.057)	2.33 (0.061)	3.460	1,859	0.001	0.161
Seek to forget everyday life	2.15 (0.057)	1.95 (0.060)	2.419	1,872	0.016	0.112
Enjoy living in an imaginary world	2.08 (0.059)	1.94 (0.061)	1.660	1,857	0.097	0.077

Relationship between individual motives, game genres, and OCS

Correlational statistics were performed to find whether having increased motivation to play related to a change in the OCS factor scores. The target significance value was set at 0.001 or lower for an alpha of 0.05 to compensate for multiple comparisons within each factor score. Results, presented in Table 3, indicate that with each increase in the score of the individual motives, we had statistically significant rises in the OCS factor scores, the sole exception was the case of playing video games in search of an environment that provided clear and just rules to the gamer (vs. the unpredictable environment in real life). Similar results were obtained when we performed this analysis for the motives to play individual games. R values serve as effect sizes, ranking individual motives and game types by importance.

Table 3.

Comparison of Online Cognitions Scale Factor Scores Between Groups of Motives to Favor Specific Video Game Genres

Correlational statistics with OCS
Spearman's r value, all values statistically significant at p=0.05 unless stated
Favorite video game	Social comfort	Lonely/depressed	Impulsive	Distraction
Combat simulation (e.g., Counterstrike, Call of duty)	.322	.295	.302	.253
RPG (Dragon Age, Diablo)	.315	.263	.262	.216
MMORPG (e.g., World of Warcraft, AION, Everquest)	.296	.270	.266	.213
Life management (e.g., Second Life, Sims)	.216	.202	.197	.188
Strategy games (e.g., Total War, World in conflict)	.279	.263	.275	.203
Sports (FIFA10, Pro Evolution, NBA 2010)	.258	.244	.244	.195
Flight simulation (e.g., Hawx, Flight Simulator)	.288	.232	.247	.200
Driving Simulation (e.g., Need for Speed, Colin Mac Rae)	.278	.258	.262	.220
Adventure games (e.g., Grand Theft Auto, Assassin's Creed)	.302	.266	.273	.231
Platform games/riddle solving (e.g., Mario, Tomb Raider)	.215	.196	.200	.179
Martial arts (Street Fighter, Tekken)	.283	.264	.276	.202
Thought games (e.g., chess, tetris, item spotting)	.119	.122	.111	.097
Music simulators—karaoke (e.g., Beatles, Guitar Hero)	.189	.162	.174	.155
Facebook games (e.g., farmville, cafeworld)	.267	.273	.291	.258
Games in mobile phones and handhelds	.262	.228	.259	.214
Motivation for favoring a particular online video game
Enjoy the competition against the computer	.262	.244	.250	.201
Enjoy the competition with other players	.329	.287	.316	.272
Like cooperation with other players	.240	.183	.213	.196
Seek the creation of an alternate identity	.324	.280	.323	.291
Enjoy doing something that is not allowed in real life	.332	.312	.333	.279
Like being able to break the rules of the game	.384	.352	.372	.308
Like making friends through the game	.344	.298	.327	.290
Like following clear, just rules	.067^*	.051^*	.032^*	.047^*
Enjoy living in an imaginary world	.296	.251	.257	.272
Become renown for my gaming successes	.405	.367	.371	.340
Enjoy being a member of the best teams and considered an elite player	.383	.331	.358	.327
Seek to forget everyday life	.285	.259	.299	.338
Enjoy feeling alive, intensity, absorbed in its flow	.336	.300	.328	.334

Nonstatistically significant.

Regression outcomes for predicting total OCS score

A stepwise regression analysis was performed to determine whether motivations to play online games could predict a tendency to indulge in addictive Internet use, as measured by the total OCS score. The independent variables entered included sociodemographic factors (age, gender, and ethnicity), factors regarding Internet use (life time experience in Internet and PC use and availability of Internet at home), school performance outcomes (grade average, progression of grades during the last semester, and truancy), a simple Likert index of general happiness, and the individual motives to play online games. Results are presented in Table 4 including only statistically significant variables. The adolescent with a higher score in the OCS is more likely to be younger, feel unhappy, go online more often and spent more hours online, prefer combat simulations, life management and Facebook games, dislike cooperation with other players (although enjoying making friends through gaming), seek to forget everyday life, while becoming renown for successes in gaming, dislikes limits in gaming, and enjoys the sense of flow while gaming.

Table 4.

Linear Regression Model for Online Cognitions Scale Total Score

	Unstandardized estimates			95% CIs for B		Hypothesis test
Parameter	B	SE	Standardized beta	Lower bound	Upper bound	t	Sig.
(Intercept)	131.012	12.476		106.532	155.492	10.501	<0.001
Age	−3.188	0.663	−0.128	−4.488	−1.888	−4.812	<0.001
Happiness	−4.386	1.395	−0.077	−7.122	−1.650	−3.145	0.002
Frequency of online use	4.776	0.618	0.197	3.563	5.989	7.728	<0.001
Hours spent online	4.854	0.568	0.216	3.738	5.969	8.539	<0.001
Become renown for my gaming successes	2.376	0.741	0.106	0.923	3.829	3.208	0.001
Like being able to break the rules in the game	2.347	0.754	0.094	0.868	3.826	3.114	0.002
Seek to forget everyday life	2.242	0.725	0.095	0.820	3.664	3.094	0.002
Seek the creation of an alternate identity	1.503	0.780	0.057	0.028	3.035	1.926	0.049
Enjoy feeling alive, intensity, absorbed in the game's flow	1.710	0.737	0.075	0.263	3.156	2.319	0.021
Like making friends through the game	1.516	0.729	0.063	0.086	2.946	2.079	0.038
Like cooperation with other players	−2.623	0.830	−0.103	−4.252	−0.993	−3.158	0.002
Facebook games (e.g., farmville, cafeworld)	2.301	0.582	0.103	1.160	3.443	3.956	<0.001
RPG (Dragon Age, Diablo)	1.755	0.819	0.058	0.148	3.363	2.142	0.032

Adj. R²=0.381, F(17, 1094)=41.196, p<0.001.

CI, confidence interval.

Identification of specific patterns of motives to play games

Table 5 presents the dimension reduction analysis results for the canonical analysis with the number of significant canonical dimensions and percentage of variance explained in the canonical variate. This percentage corresponds to variance explained within the relationship of the two sets of variables. Each analysis offers a first dimension, which explains the largest chunk of the variance in the canonical correlate; this dimension includes in a single axis the most prominent reasons, choices, and cognitions. The other dimensions offer less prominent sets of variables that are explaining a unique portion of variance.

Table 5.

Dimension Reduction Analysis Results for the Canonical Analysis with % of Variance Explained ( r ²)

The Dimension reduction analysis for types of games and motives involved in choices of games
Roots	Wilks Lambda	F	Hypoth. DF	Error DF	Sig. of F	Canonical correlation	% of variance
1 OF 13	0.34844	8.40261	195	14560.05	<0.001	0.70976	50.375
2 OF 13	0.70215	3.16187	168	13523.64	<0.001	0.36331	13.199
3 OF 13	0.80892	2.21361	143	12485.92	<0.001	0.23828	5.678
4 OF 13	0.85762	1.90662	120	11446.8	<0.001	0.21653	4.688
5 OF 13	0.89981	1.58548	99	10406.12	<0.001	0.17704	3.134
6 OF 13	0.92892	1.36861	80	9363.65	0.016	0.15238	2.322
Total variance in the canonical correlation explained:							79.396

Table 6 presents results from the canonical analysis. Those results include both the standardized canonical coefficients (which represent the weight of each item in each canonical dimension) and also the cross-loadings (which represent the correlations of each item with each opposite canonical variable). Comparative strength of each cross-loading within each dimension points to specific clusters of similarities, while those similarities are vertically divided in two opposite planes by the sign preceding them, so that a group of positive correlates will carry the opposite meaning for two separate subpopulations than the group of negative correlates does; certain individuals will for example prefer action games, but for a wildly different set of motives and may in fact be puzzled as to why their opposites may play action games for their particular (polar opposite) set of reasons. The largest correlates in each dimension (both positive and negative) are in bold so that the reader may follow the conclusions.

Table 6.

Canonical Correlations Between Motivations for Types of Favorite Video Games and Choosing a Specific Video Game

	Standardized canonical coefficients						Cross-loadings
Predicted set—game types	1	2	3	4	5	6	1	2	3	4	5	6
Combat simulation (e.g., Counterstrike, Call of duty)	−0.329	−0.496	−0.334	−0.402	0.219	−0.342	−0.598	−0.122	−0.004	−0.014	0.002	−0.015
RPG (Dragon Age, Diablo)	−0.044	0.361	−0.31	0.298	−0.387	0.053	−0.47	0.137	−0.002	0.039	−0.036	0.009
MMORPG (e.g., World of Warcraft, AION, Everquest)	−0.129	−0.089	0.306	0.086	0.366	0.196	−0.48	0.019	0.039	0.02	0.021	0.019
Life management (e.g., Second Life, Sims)	−0.077	0.459	0.62	−0.693	0.039	−0.371	−0.365	0.235	0.066	−0.074	0.003	−0.019
Strategy games (e.g., Total War, World in conflict)	−0.042	−0.039	−0.032	0.054	−0.234	0.386	−0.517	0.05	0.013	0.014	−0.019	0.026
Sports (FIFA10, Pro Evolution, NBA 2010)	−0.182	−0.333	−0.191	−0.427	0.343	−0.26	−0.531	−0.1	−0.02	−0.033	0.025	−0.021
Flight simulation (e.g., Hawx, Flight Simulator)	−0.048	0.147	0.433	0.478	0.49	−0.207	−0.459	0.095	0.055	0.069	0.043	−0.016
Driving Simulation (e.g., Need for Speed, Colin Mac Rae)	−0.081	−0.201	0.522	0.317	0.006	0.091	−0.569	−0.038	0.043	0.037	−0.013	0.003
Adventure games (e.g., Grand Theft Auto, Assassin's Creed)	−0.195	0.025	−0.449	0.175	−0.706	−0.023	−0.579	0.005	−0.035	0.03	−0.062	0.001
Platform games/riddle solving (e.g., Mario, Tomb Raider)	0.008	0.036	−0.434	0.026	−0.143	0.241	−0.397	0.125	−0.043	0.009	−0.027	0.017
Martial arts (Street Fighter, Tekken)	−0.076	0.05	0.599	0.084	−0.306	0.24	−0.516	0.039	0.056	0.024	−0.042	0.018
Thought games (e.g., chess, tetris, item spotting)	−0.004	0.159	−0.241	−0.407	−0.271	−0.52	−0.268	0.162	−0.055	−0.049	−0.016	−0.044
Music simulators - karaoke (e.g., Beatles, Guitar Hero)	−0.05	0.195	−0.505	0.428	0.574	−0.043	−0.264	0.182	−0.088	0.046	0.076	−0.001
Facebook games (e.g., farmville, cafeworld)	−0.078	0.067	−0.112	−0.468	0.23	0.882	−0.319	0.112	−0.038	−0.076	0.044	0.09
Games in mobile phones and handhelds (PSP–Nintendo DS)	−0.067	0.179	−0.116	0.239	0.099	−0.319	−0.422	0.105	−0.031	0.017	0.015	−0.02
Motivations for choosing a video game
Enjoy the competition against the computer	−0.301	0.111	−0.255	−0.104	−0.495	−0.397	−0.549	0.006	−0.02	−0.021	−0.054	−0.032
Enjoy the competition with other players	−0.149	−0.543	0.165	−0.164	0.171	0.213	−0.572	−0.106	0.007	−0.037	−0.001	−0.005
Like cooperation with other players	−0.122	−0.281	0.106	−0.432	−0.289	−0.661	−0.488	−0.072	−0.017	−0.075	−0.007	−0.043
Seek the creation of an alternate identity	−0.09	0.474	0.692	0.023	0.543	−0.188	−0.456	0.156	0.12	0.006	0.033	−0.002
Enjoy doing something that is not allowed in real life	−0.053	−0.094	0.589	0.182	−0.332	0.172	−0.465	0.045	0.115	0.039	−0.027	0.023
Like being able to break the rules in the game	−0.283	0.008	−0.291	0.491	0.098	−0.258	−0.522	0.033	0.037	0.065	−0.002	0.015
Like making friends through the game	−0.053	0.192	−0.323	−0.507	0.369	0.671	−0.438	0.072	−0.058	−0.068	0.049	0.055
Like following clear, just rules	−0.061	0.204	−0.365	0.147	0.323	−0.437	−0.197	0.076	−0.113	−0.025	0.048	−0.054
Enjoy living in an imaginary world	0.002	0.567	−0.181	0.142	0.035	−0.293	−0.32	0.23	−0.045	0.013	−0.009	−0.011
Become renown for my gaming successes	−0.037	0.034	−0.216	0.3	−0.337	0.064	−0.517	0.019	−0.047	0.021	0.002	0.03
Enjoy being a member of the best teams and considered an elite player	−0.239	−0.328	0.019	0.04	0.743	0.322	−0.551	−0.037	−0.03	0.005	0.044	0.041
Seek to forget everyday life	0.027	0.287	0.155	−0.878	−0.348	0.206	−0.305	0.161	0.013	−0.087	−0.064	0.046
Enjoy feeling alive, intensity, absorbed in its flow	−0.023	−0.052	−0.302	0.635	−0.5	0.438	−0.391	0.085	−0.048	0.034	−0.069	0.053

Largest correlates in each dimension are employed in the analysis and are presented in bold.

The first dimension, which explains most of the canonical variance, offers a categorization of the reasons for enjoying video games by order of saliency. The other five dimensions offer two polar extremes each; that is, two groups of players who prefer games that they perceive as very different in their essence for reasons that they themselves would consider opposite.

The specific themes that emerged from the first analysis were primarily power, success, and dominance over opponents. Also, the contrasts between seeking and being indifferent to coplayers, competitiveness versus aloofness, rebelliousness versus conformation, help-seeking from others versus finding peace of mind through escapism, seeking out human contact versus turning to the machine world, and turning to human coplayers for friendship or for cooperation.

Dimension 2 shows the contrast between those players who favor an action-packed game to satisfy their need for competition and cooperation with other human players to those favoring a slower gaming environment with emphasis on thought and fantasy to forget their actual reality and live out an alternate one. Dimension 3 offers two possibilities for an alternate reality—those players who would play second life to try something that is not allowed, and their opposites who would like to live out a fantasy of becoming music stars, famous, but conforming to notions of normality.

Dimension 4 offers two extremes, those who would engage in games in Facebook or play second life trying to forget their actual life predicaments and seek company in the game, and those nonconformers who prefer the thrill of flying away or becoming rock stars to try something out that is unavailable to their everyday lives. Dimension 5 offers the contrast of gamers who engage in social games (music and Facebook based) seeking friendship on simple terms to those who prefer to escape from their world into a man-versus-machine world. The final sixth dimension offers a distinction between those playing simple SNS games to forget their daily lives through engagement and hope for new friends, to those who seek cooperation rather than friendship in the structured and predictable classical thought game.

Discussion

Challenges in the measurement of Internet addiction

This study confirmed that the employment of the OCS scale in a Greek adolescent population has high reliability and internal validity, while the comparison with the results from the YDQ scale was used as a method of external validation and fixing of cut-off criteria. Answering questions on any questionnaire assumes a level of candidness and self-awareness that those who are addicted may lack, especially in the early stages (e.g., if you do not make any attempts to cut down on your level of use, then you are deducted two points from the maximum YDQ score). We sought to negate the apprehension of candidly answering a criteria-based instrument with the anonymity of our sample. The OCS may be more useful in a clinical assessment, since the items included are formulated in an innocuous manner, although a comparison of these cut-off criteria to a clinical population's is the only way to prove this hypothesis.

A related finding was that boys are a harder population to diagnose with the typical means of a questionnaire. Maturity to self-observe a behavior may be lacking in young boys who are slower to mature than girls. It may be thus appropriate for any future research to present both possible results with separate cut-off marks for each gender noting that the higher cut-off for boys may be underestimating the problem. The thought content of boys who play games tends to revolve more around their Internet use, demonstrating a higher propensity for addictive use of the medium.

Regression findings on patterns of cognitions and motives

The primary focus was directed at the patterns of video gaming as they intertwine with relevant motives. The predictor variables that were statistically significant in our regression model included lower age and subjective feelings of happiness. Despite the fact that boys tended to present with higher OCS scores, gender was not an important predictor variable in this equation. This reflects the fact that they presented with higher cut-off values in the ROC-curve analysis. Young boys are at a higher risk for pathologic indulgence with the new technologies, especially online gaming²²; a question for future research, however, is whether this higher baseline level of involvement directly corresponds to Internet addiction, or is a cultural artifact. Caution is advised against reaching fast conclusions on individual cases, and the best course of action before pronouncing anyone as an online game addict would be a thorough clinical examination from a mental health professional.

Statistically significant motives were relevant to issues like virtual popularity versus real-life seclusion, avoidance of real-life problems by escaping to a reality where the identity of the self and the rules of existence itself can be altered, broken, and reshaped. Adolescence is an age where the young adult will experiment with the notion of self-hood, try on various beliefs to see if they fit, and demonstrate rebelliousness against prohibitions that stand in the way. Our model shows that the combination of those inherent traits of an adolescent coupled with social isolation and other real-life problems will have an additive effect to his/her turning to the Internet for solutions. The context of game playing is thus very important if we seek to understand the person rather than just apply a label of an addict.

Our research demonstrates that reaching a conclusion as to whether an adolescent is addicted to the new technologies or not is fraught with methodological difficulties. Large-scale surveys ideally should include clinically tested measures to have realistic cut-off points, and the researchers should resist over-generalization of results. There are significant differences between the sexes as to their game-playing habits, which should be taken into account in related research. A single individual should be assessed clinically with the aid not only of an addiction scale, but also with an exhaustive series of suitable questions on game-related motives and preferences.

Footnotes

Author Disclosure Statement

The authors have no conflict of interest.

References

van Rooij

, Schoenmakers

, van de Eijnden

RJJM

et al. Compulsive Internet use: the role of online gaming and other Internet applications. Journal of Adolescent Health, 2010; 47:51–57.

Rehbein

, Kleimann

, Mößle

. Prevalence and risk factors of video game dependency in adolescence: results of a German nationwide survey. Cyberpsychology, Behavior, and Social Networking, 2010; 13:269–277.

Griffiths

, Davies

MNO

, Chappell

. Demographic factors and playing variables in online computer gaming. CyberPsychology & Behavior, 2004; 7:479–487.

Griffiths

. The role of context in online gaming excess and addiction: some case study evidence. International Journal of Mental Health and Addiction, 2010; 8:119–125.

Choi

, Kim

. Why people continue to play online games: in search of critical design factors to increase customer loyalty to online contents. CyberPsychology & Behavior, 2004; 7:11–24.

Yee

. Motivations for play in online games. CyberPsychology & Behavior, 2006; 9:772–775.

Przybylski

, Rigby

, Ryan

. A motivational model of video game engagement. Review of General Psychology, 2010; 14:154–166.

King

, Delfabbro

. Motivational differences in problem video game play. Journal of Cybertherapy and Rehabilitation, 2009; 2:139–149.

Liu

, Peng

. Cognitive and psychological predictors of the negative outcomes associated with playing MMOGs (massively multiplayer online games) Computers in Human Behavior, 2009; 25:1306–1311.

10.

Davis

. A cognitive-behavioral model of pathological Internet use. Computers in Human Behavior, 2001; 17:187–195.

11.

Davis

, Flett

, Besser

. Validation of a new scale for measuring problematic Internet use: implications for pre-employment screening. CyberPsychology & Behavior, 2002; 5:331–345.

12.

Young

. Internet addiction: the emergence of a new clinical disorder. CyberPsychology and Behavior, 1996; 1:237–244.

13.

Shapira

, Lessig

, Goldsmith

et al. Problematic internet use: proposed classification and diagnostic criteria. Depression and Anxiety, 2003; 17:207–216.

14.

Tao

, Huang

, Wang

et al. Proposed diagnostic criteria for internet addiction. Addiction, 2010; 105:556–564.

15.

Fisoun

, Floros

, Geroukalis

et al. Internet addiction in the island of Hippocrates: the associations between internet abuse and adolescent off-line behaviours. Child and Adolescent Mental Health, 2012; 17:37–44.

16.

Fisoun

, Floros

, Siomos

et al. Internet addiction as an important predictor in early detection of adolescent drug use experience-implications for research and practice. Journal of Addiction Medicine, 2012; 6:77–84.

17.

Hair

, Black

, Babin

et al. 2010. Multivariate data analysis, 7th. New Jersey: Prentice Hall.

18.

Thompson

. 1984. Canonical correlation analysis: uses and interpretation, 47. Thousand Oaks, CA: Sage Publications Inc.

19.

Nie

, Hull

, Bent

. 2011. IBM statistical package for the social sciences (SPSS Version 20) Computer software. Chicago, IL: SPSS.

20.

Bentler

. 2003. EQS 6.1 for Windows [Computer software] Encino, CA: Multivariate Software.

21.

Özcan

, Buzlu

. Internet use and its relation with the psychosocial situation for a sample of university students. CyberPsychology & Behavior, 2007; 10:767–772.

22.

Griffiths

, Davies

MNO

, Chappell

. Online computer gaming: a comparison of adolescent and adult gamers. Journal of Adolescence, 2004; 27:87–96.