Online gaming has become increasingly popular. However, this has led to concerns that these games might induce serious problems and/or lead to dependence for a minority of players. Aim: The aim of this study was to uncover and operationalize the components of problematic online gaming.
A total of 3415 gamers (90% males; mean age 21 years), were recruited through online gaming websites. A combined method of exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) was applied. Latent profile analysis was applied to identify persons at-risk.
EFA revealed a six-factor structure in the background of problematic online gaming that was also confirmed by a CFA. For the assessment of the identified six dimensions – preoccupation, overuse, immersion, social isolation, interpersonal conflicts, and withdrawal – the 18-item Problematic Online Gaming Questionnaire (POGQ) proved to be exceedingly suitable. Based on the latent profile analysis, 3.4% of the gamer population was considered to be at high risk, while another 15.2% was moderately problematic.
Citation: Demetrovics Z, Urbán R, Nagygyörgy K, Farkas J, Griffiths MD, Pápay O, et al. (2012) The Development of the Problematic Online Gaming Questionnaire (POGQ). PLoS ONE 7(5): e36417. https://doi.org/10.1371/journal.pone.0036417
Editor: Jerson Laks, Federal University of Rio de Janeiro, Brazil
Received: January 26, 2012; Accepted: March 31, 2012; Published: May 10, 2012
Copyright: © 2012 Demetrovics et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported by the Hungarian Scientific Research Fund (grant number: 83884) (http://otka.hu). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
The popularity of online gaming has spread at an increasing rate since its introduction . However, given the increasing number of studies on the phenomenon of problematic gaming , , , , , there have been growing concerns about negative consequences for a small minority. These gamers spend more time with gaming than planned while ignoring other important activities causing negative effects on their performance , , social relationships , , and withdrawal symptoms , . There are different names present regarding this phenomenon in scientific literature. The phenomenon is referred to as addiction , , problematic use , , excessive game use  or engagement . However, all these authors – irrespective of the name – agree in that there exists an excessive form of online gaming that shows a problematic pattern and that is related to behavioral addictions , . The present authors propose to use the name problematic gaming. This term describes both the quintessence of the phenomenon (i.e., that the behavior is not only excessive but gaming-related problems are also expected to be present), while avoiding the notion of dependency (as the exact definition and diagnostic criteria have not yet been clarified or agreed).
The issues outlined above also suggest that precise assessment and screening of the phenomenon is an urgent matter. In order to measure the problematic nature of gaming, some authors have developed questionnaires based on the general phenomenon of internet addiction , while others have attempted to operationalize the behavioral addiction model of Griffiths , . However, the limitation of these current questionnaires is that many of them typically target users of Massively Multiplayer Online Role-Playing Games (MMORPG) , , , , . Though this type of game is the most popular among online games, the total population of gamers is more diverse and it is therefore necessary to develop a measure that is suitable for the assessment of other genres and gamer populations such as those who play online Real-Time Strategy (RTS) games and online First Person Shooter (FPS) games.
The aim of our study was therefore twofold. Firstly, to explore what components comprise problematic online gaming. Secondly, to make these dimensions measurable, to develop such a scale on which the identified dimensions can be assessed. Contrary to earlier studies, the objective here was to create a questionnaire suitable for all types of Massively Multiplayer Online Games. A further intention was to carry out an empirical based analysis to ensure that all components of problematic gaming remain in focus.
The study was approved by the Institutional Review Board of the Eötvös Loránd University. Informed consent was obtained via our online system from all subjects. After introducing the goals of the study in details the subjects were asked to tick into a box if they agreed to continue and participate in the study.
Sample, Procedure and Participants
All Hungarian websites that facilitate the playing online games were identified (n = 18). All 18 sites were contacted by the research team and were asked information about the number of visitors, and requested their cooperation in the planned study. All sites responded. Based on this information, the number of (ever) registered users was estimated to be approximately 30,000. However, many of these users may have simultaneously registered on multiple sites. Furthermore, it is likely that many formerly registered users were not currently active. All sites agreed to publish a call for participation in the present study on their home sites or via a newsletter. In the call for participation, gamers were asked to visit the study website, to sign in with a password provided by the researchers, and to complete a questionnaire. A total of 4390 questionnaires were started but not everyone completed the whole survey. This left 3415 completed questionnaires. In addition to answering the general questions regarding online gaming habits, the respondents were asked about online gaming problems.
Major socio-demographic characteristics of the gamers (gender, age, qualification, marital status, school, work) and characteristics regarding their online gaming activities were recorded. Additionally, the survey contained a 26-item questionnaire that listed several problems regarding online gaming. The 26 items were created by means of (i) a comprehensive literature review supplemented with (ii) interviews with online gamers. Firstly, in relation to the literature review, a full search was carried out in the databases Web of Science, Science Direct, PsycINFO, and Medline using the following keywords: online gam*, MMO, MMORPG, multiplayer, FPS, First Person Shooter, RTS, Real Time Strategy. A total of 199 hits were found. However, 115 was excluded because they were irrelevant regarding specific aspects and characteristics of online gaming. The remaining 84 papers were read carefully and all items were listed out that could be considered as reflecting problematic aspects of online gaming. A total of 42 characteristics were identified this way. Secondly (and concurrently with the literature review), 15 online gamers were asked to list problems they had noticed in themselves and/or others as result of online gaming. These gamers listed 32 problems. Following exclusion of duplicates and similar items, the list was reduced to 26 items.
Statistical analysis comprised an exploratory factor analysis (EFA) with robust maximum-likelihood estimation (MLR) in MPLUS 6.0. The goodness of fit was assessed by the root-mean-square error of approximation (RMSEA) and its 90% confidence interval (90% CI), and p value smaller than 0.05 for test of close fit (Cfit>.05). The factor solution was selected based on fit statistics and interpretability of factors.
The factor structure based on EFA was confirmed through confirmatory factor analyses (CFA) with independent samples. CFA was performed with robust maximum-likelihood estimation (MLR) in MPLUS 6.0. The goodness of fit was evaluated using RMSEA and its 90% confidence interval (90% CI), p value smaller than 0.05 for test of close fit, standardized root-mean-square residual (SRMR), comparative fit index (CFI), and Tucker-Lewis Fit Index (TLI). As Brown (2006) and Kline (2005) recommended, multiple indices were selected in order to provide different information for evaluating model fit.
To carry out the above analyses, four non-overlapping groups from the sample were randomly selected. Sample 1 (n = 600) was used to perform an initial EFA on the original 26 items. Sample 2 (n = 600) was used to conduct a separate EFA to cross-validate the factor structure found in the first analysis. Sample 3 (n = 600) was used to conduct CFA analysis. After the inspection of modification indices, we also cross-validated the final CFA model with sample 4 (n = 1615).
In order to identify the groups of users with high risk of problematic use of online gaming a person-oriented statistical framework was selected, seeking subtypes of gamers that exhibited similar patterns of symptoms of problematic use. Therefore a latent profile analysis was performed with 1 to 6 classes with the full sample (n = 3415). The latent profile analysis ,  is a latent variable analysis with a categorical latent variable – in this case problematic gamers – and continuous manifest indicators such as factor scores of POGQ. In the process of determining the number of latent classes, the Bayesian information criteria parsimony index was used, alongside the minimization of cross-classification probabilities, entropy and the interpretability of clusters. In the final determination of the number of classes, the likelihood-ratio difference test (Lo-Mendell-Rubin Adjusted LRT Test) was also used. This compares the estimated model with a model having one less class than the estimated model . A low p value (<.05) indicates that the model with one less class is rejected in favor of the estimated model.
To determine the cut-off point for POGQ a sensitivity analysis based on membership in the most problematic group in the latent profile analysis was carried out. Considering membership in this group as a gold standard, the sensitivity and specificity values for all POGQ cut-off points was calculated. Thus, the accuracy of the POGQ by calculating the proportion of participants classified as being at high risk for problematic gaming versus other gamers could be assessed. The sensitivity (i.e., the proportion of true positives belonging to the most problematic group based on LPA) and specificity (i.e., the proportion of true negatives) was defined as suggested by Altman and Bland  and Glaros and Kline . In order to explore the probability that the POGQ would give the correct “diagnosis”, the positive predictive values (PPV), the negative predictive values (NPV), and the accuracy values for each possible POGQ cut-off points was calculated. PPV was defined as the proportion of participants with positive test results who are correctly diagnosed , . The NPV was defined as the proportion of patients with negative test results who are correctly diagnosed , .
90% of our sample (n = 3072) was male. Mean age was 21 years (SD = 5.85 years). Slightly more than one-tenth of the participants had graduate education, while 39.4% had secondary education. The majority (61.9%) were primarily students, but approximately one-quarter worked full time (24.3%). Almost two-thirds of the participants were single (64.9%), and a further 24.3% were in a relationship but did not live with their partner (see Table 1).
Slightly more than one-third (35.7%) played a maximum 14 hours per week and approximately the same number played between 15 and 28 hours per week. One in ten played more than 42 hours a week (6 hours per day on average) (Table 1). Approximately a half spent money on gaming, although most did not spend more than $25(US) per month. The majority of the sample participants were individual gamers (38.6%) but there was also a relatively high ratio of different level organized gamers (Table 1).
Exploratory Factor Analyses
An exploratory factor analysis was performed with maximum-likelihood estimation which is robust to non-normality and promax rotation to evaluate the factor structure of 26 items on Sample 1 (n = 600). Acceptability of the factor solution was based on goodness of fit index (RMSEA <0.08, Cfit (90% CI) <0.08, pclose >0.05), the interpretability of the solution and salient factor loadings (>0.30). A total of 1–8 factor solutions was examined. The six-factor solution provided the first adequate RMSEA value based on the criteria (χ2 = 409.8, df = 184 p<0.0001; RMSEA = 0.045 [0.039–0.051] pclose >0.90). The exploratory factor analysis on Sample 2 (n = 600) was repeated. As in Sample 1, a six-factor solution also provided the first adequate and interpretable factor solution (χ2 = 457.7 df = 184 p<0.0001; RMSEA = 0.050 [0.044–0.056] Cfit = 0.514). Factor loadings are presented in Table 2.
For the further development of this scale, items with the following rules were selected. First, items were excluded that had factor loadings lower than 0.40 at least in one of the two analyses. Second, items with salient cross loadings were excluded. If a cross loading only in one of the two parallel EFAs was identified, the cutoff 0.50 was used. In case of more than two cross-loadings, a 0.30 as a cutoff was used to exclude items from further analyses. The excluded items are crossed out in Table 2. As result of the above criteria, 18 of the original 26 items were retained (see Appendix S1).
Confirmatory Factor Analysis
Based on the previous analyses on the samples 1 and 2, a six-factor solution was tested on Sample 3 (n = 600) with confirmatory factor analysis. This model provided an optimal fit to the data (χ2 = 256.0 df = 120 p<0.0001; CFI = 0.965; TLI = 0.956; RMSEA = 0.043 [0.036–0.051] Cfit>0.90; SRMR = 0.037). We cross-validated this model with Sample 4 (n = 1615) and found adequate level of fit (χ2 = 512.8 df = 120 p<0.0001; CFI = 0.962; TLI = 0.952; RMSEA = 0.045 [0.041–0.049] Cfit>0.90; SRMR = 0.036). The factor loadings, factor reliabilities, internal consistencies, means, and SDs are presented in Table 3.
Professional vs. Non-professional Gamers
The possible bias stemming from the inclusion of professional gamers in the total sample was checked. The level of fit of the measurement models without professional gamers (N = 2857) was satisfactory (χ2 = 763.5 df = 120 p<0.0001; CFI = 0.965; TLI = 0.956; RMSEA = 0.043 [0.040–0.046]; SRMR = 0.034). The level of fit of the measurement models only among professional gamers (N = 528) was also satisfactory (χ2 = 290.0 df = 120 p<0.0001; CFI = 0.948; TLI = 0.934; RMSEA = 0.052 [0.044–0.059]; SRMR = 0.034). Furthermore, a multi-group analysis with non-professional (N = 2857) and professional (N = 528) gamers was performed. In this analysis, the factor loadings and intercepts were set equal in both groups. The level of fit was satisfactory (χ2nonprofessional = 758.6, χ2professional = 302.4 df = 264, p<0.0001; CFI = 0.963; TLI = 0.957; RMSEA = 0.042 [0.040–0.045]; SRMR = 0.036) and the means of latent variables were not statistically different in either group.
Labels of Factors
In the first factor, two items belonged that referred to obsessive thinking and daydreaming on the online game. This dimension was named preoccupation. The second factor contained items concerning the excessive use of online games. The three items belonging here referred to noticing gaming related problems, elongated gaming time, and the difficulties in controlling time spent on gaming. This factor was named overuse. The third factor was named immersion as these four items indicated dealing excessively with online games, immersion in gaming, and losing track of time. The fourth factor indicated damage to social relationships, and the preference of gaming over social activities. This three-item dimension was named social isolation. The two items of the fifth factor referred to the comments of the player’s social environment on overuse of online games and the related conflicts, so this factor was named interpersonal conflicts. Finally, the four items of the sixth factor concerned the appearance of withdrawal symptoms in cases when players experienced difficulties in gaming as much as they wanted. This dimension got the name withdrawal.
Latent Profile Analysis
A latent profile analysis was performed on the dimensions of problematic online gaming, and, a four-class solution was found according to the decision criteria. As Table 4 demonstrates that the AIC, BIC and sample-size adjusted BIC continued to decrease as more latent classes were added. However, a leveling-off after the four-latent-class solution was noted. In inspection of entropy, the two-class solution reached the maximum level, but the four-class solution also provided also an adequate level of entropy. Based on the L-M-R test, the four-class solution was accepted.
The features of each class are presented in Figure 1. The first class represents those gamers (47.8% of the total sample) that scored on dimensions of problematic use below the average. The second class of gamers (33.7%) represents the low risk of problematic use. The third class (15.2%) represents the medium risk of problematic use. Finally, the fourth class (3.4%) represents the high risk of problematic use. In this latter group, the ‘social isolation’ factor and ‘withdrawal symptoms’ factor especially showed an elevated level compared to other dimensions.
. The latent profile analysis resulted four classes. The first class represents those gamers that scored on POGQ below the average, while the second class represents the low risk, the third class represents medium risk of problematic use. The fourth class (3.4%) represents the high risk of problematic use.
Determination cut-off score to be classified a problematic gamer: Sensitivity and specificity analyses.
Based on membership in the fourth class (i.e., being at high risk for problematic gaming) as a “gold standard”, the sensitivity, specificity, PPV as well as NPV, and accuracy of the POGQ at all possible cut-off points (Table 5) were calculated.
Based on this analysis, a cut-off score of 65 points is suggested as an ideal cut-off to be classed as problematic gamer. In this case, specificity is 100%, while sensitivity is 96%. This means that practically none of the negative (i.e., non-problematic) cases are considered as problematic, while only 4% of the true problematic cases are not recognized. Accuracy, as well as NPV in this case is 100%, while PPV is 90%. Increasing of the cut-off score would result in the growing number of the false negative cases, whereas decreasing would lead to more false positive cases.
Due to the growing number of indicated problems concerning online gaming it has become an absolute necessity to develop a tool with adequate psychometric characteristics for the measurement of the extent of gaming-related problems. The Problematic Online Gaming Questionnaire (POGQ) developed in this study, based on the results of the analyses, appears to fulfill those requirements that are expected from a measure like this. The POGQ was created in a way that it is applicable for all types of online games and its empirical basis makes it possible to cover all problems experienced by the players.
The results of these empirically-based analyses are at the same time very much supported by the fact that the six dimensions identified in the background of problematic online gaming fit closely to the available theoretical frameworks. Griffiths  proposed a “components” model for addictions that assumes the six classical symptoms for addiction behaviors in general that is salience, mood modification, tolerance, withdrawal, conflict, and relapse. The withdrawal and preoccupation components can be identified with our equally named factors, while conflict dimension is partly covered by the interpersonal conflicts factor and partly by the factor overuse (intrapersonal conflicts). It is interesting though that items explicitly representing salience dimension fell out during analysis (see item 2, 5, 11 in Table 2.), but the component is still present in overuse, preoccupation, and social isolation factors. The relapse component appears in the overuse dimension while mood modification dimension is primarily present in the withdrawal factor (item 13 in Table 2).
In another approach, the DSM-IV criteria for psychoactive substance use dependence by the American Psychiatric Association  that are generally regarded as the definitional basis of behavioral addictions can be considered. These dimensions – withdrawal, lack of control (unsuccessful attempt to quit), much time spent on the activity, behavior continues despite knowledge of adverse consequences, more intensive use and for longer period than intended – are all clearly reflected in the POGQ. The more than adequate psychometric properties of the POGQ and the wide empirical content it is based on, is reassuring regarding the future use of the scale. However, further tasks include the cross-cultural validation of the POGQ and clinical validation of the scale.
It is an issue in relation to all behavioral addictions not present in DSM-IV-TR whether those individuals who engage in a specific behavior excessively should be regarded as having a disorder, and with which criteria we should identify individuals functioning on other pathologic levels. Considering the lack of consensus regarding definitions, the authors of the present study insisted on using the expression ‘problematic gaming’ instead of the more ambiguous gaming dependence. However, the latent profile analysis performed also indicated that a segment of the online gaming population (3.4% in this study) significantly exceeded the whole population and characteristically showed more problems than others. A further 15.2% of the population also showed moderately elevated level of problems. One of the most important tasks in future research is the detailed analysis of this at-risk population and to explore which background factors may carry high risk concerning problematic gaming. For forthcoming studies, the results here highlight two significant dimensions, withdrawal and social isolation, that showed elevated levels in case of these gamers, while obsession and overuse seemed to be the less indicative dimensions. Therefore, it seems that intensive actual (overuse) or imaginary (obsessive) gaming is less indicative of problematic gaming in itself. These results coincide with the results observed concerning problematic internet use . In contrast, neglecting social relationships and especially the presence of withdrawal symptoms (feeling depressed or irritable, getting restless, anxious or upset when not able to game) appear to carry the highest risks. Furthermore, it is important to note that both dimensions include reducing or neglecting other activities which are key characteristics of addictions according to the results of many other studies.
One limitation of the present study is that it was carried out among Hungarian gamers thus results should be cautiously generalized for other cultures. However, it is hoped that future studies can confirm the findings presented here in other cultures. Another important issue is that current results were based on self-report data. It is again a challenge for future studies to investigate and confirm the identified problem-dimensions in clinical and/or observational studies. In conclusion, and based on all these assumptions, it is hoped that creation of the POGQ will facilitate and enhance further research, and that the instrument will serve as a valid and reliable tool in future studies.
Conceived and designed the experiments: ZD. Performed the experiments: ZD KN JF OP KF GK. Analyzed the data: RU AO. Wrote the paper: ZD UR KN JF MDG KF.
- 1. ComScore (2007) ComScore2007Worldwide online gaming community reaches 217 million people. Retrieved 16 January, 2012, from http://www.comscore.com/Press_Events/Press_Releases/2007/07/Worldwide_Online_Gaming_Grows. Worldwide online gaming community reaches 217 million people. Retrieved 16 January, 2012, from http://www.comscore.com/Press_Events/Press_Releases/2007/07/Worldwide_Online_Gaming_Grows.
- 2. Caplan S, Williams D, Yee N (2009) Problematic internet use and psychosocial well-being among MMO players. Comput Human Behav 25: 1312–1319.S. CaplanD. WilliamsN. Yee2009Problematic internet use and psychosocial well-being among MMO players.Comput Human Behav2513121319
- 3. Griffiths MD (2008) Diagnosis and management of video game addiction. New Directions in Addiction Treatment and Prevention 12: 27–41.MD Griffiths2008Diagnosis and management of video game addiction.New Directions in Addiction Treatment and Prevention122741
- 4. Ng BD, Wiemer-Hastings P (2005) Addiction to the internet and online gaming. Cyberpsychol Behav 8: 110–113.BD NgP. Wiemer-Hastings2005Addiction to the internet and online gaming.Cyberpsychol Behav8110113
- 5. Wan CS, Chiou WB (2006) Why are adolescents addicted to online gaming? An interview study in Taiwan. Cyberpsychol Behav 9: 762–766.CS WanWB Chiou2006Why are adolescents addicted to online gaming? An interview study in Taiwan.Cyberpsychol Behav9762766
- 6. Young KS (2009) Understanding online gaming addiction and treatment issues for adolescents. Am J Fam Ther 37: 355–372.KS Young2009Understanding online gaming addiction and treatment issues for adolescents.Am J Fam Ther37355372
- 7. Smyth JM (2007) Beyond self-selection in video game play: an experimental examination of the consequences of massively multiplayer online role-playing game play. Cyberpsychol Behav 10: 717–721.JM Smyth2007Beyond self-selection in video game play: an experimental examination of the consequences of massively multiplayer online role-playing game play.Cyberpsychol Behav10717721
- 8. Cole H, Griffiths MD (2007) Social interactions in massively multiplayer online role-playing gamers. Cyberpsychol Behav 10: 575–583.H. ColeMD Griffiths2007Social interactions in massively multiplayer online role-playing gamers.Cyberpsychol Behav10575583
- 9. Peters CS, Malesky LA (2008) Problematic usage among highly-engaged players of massively multiplayer online role playing games. Cyberpsychol Behav 11: 481–484.CS PetersLA Malesky2008Problematic usage among highly-engaged players of massively multiplayer online role playing games.Cyberpsychol Behav11481484
- 10. Griffiths MD (2010) The role of context in online gaming excess and addiction: Some case study evidence received. Int J Ment Health Addict 8: 119–125.MD Griffiths2010The role of context in online gaming excess and addiction: Some case study evidence received.Int J Ment Health Addict8119125
- 11. Young KS (1998) Caught in the Net: How to recognize the signs of Internet addiction and a winning strategy for recovery. New York, NY: Wiley. KS Young1998Caught in the Net: How to recognize the signs of Internet addiction and a winning strategy for recovery.New York, NYWiley
- 12. Lemmens JS, Valkenburg PM, Peter J (2009) Development and validation of a game addiction scale for adolescents. Media Psychol 12: 77–95.JS LemmensPM ValkenburgJ. Peter2009Development and validation of a game addiction scale for adolescents.Media Psychol127795
- 13. Charlton JP, Danforth IDW (2007) Distinguishing addiction and high engagement in the context of online game playing. Comput Human Behav 23: 1531–1548.JP CharltonIDW Danforth2007Distinguishing addiction and high engagement in the context of online game playing.Comput Human Behav2315311548
- 14. Kim MG, Kim JE (2010) Cross-validation of reliability, convergent and discriminant validity for the problematic online game use scale. Comput Human Behav 26: 389–398.MG KimJE Kim2010Cross-validation of reliability, convergent and discriminant validity for the problematic online game use scale.Comput Human Behav26389398
- 15. Brockmyer HJ, Fox CM, Curtiss KA, McBroom E, Burkhart KM, et al. (2009) The developement of the Game Engagement Questionnaire: A measure of engagement in video game-playing. J Exp Soc Psychol 45: 624–634.HJ BrockmyerCM FoxKA CurtissE. McBroomKM Burkhart2009The developement of the Game Engagement Questionnaire: A measure of engagement in video game-playing.J Exp Soc Psychol45624634
- 16. Griffiths MD (1995) Technological addictions. Clinical Psychology Forum 76: 14–19.MD Griffiths1995Technological addictions.Clinical Psychology Forum761419
- 17. Grant JE, Potenza MN, Weinstein A, Gorelick DA (2010) Introduction to behavioral addictions. Am J Drug Alcohol Abuse 36: 233–241.JE GrantMN PotenzaA. WeinsteinDA Gorelick2010Introduction to behavioral addictions.Am J Drug Alcohol Abuse36233241
- 18. Griffiths MD (2005) A ‘components’ model of addiction within a biopsychosocial framework. J Subst Use 10: 191–197.MD Griffiths2005A ‘components’ model of addiction within a biopsychosocial framework.J Subst Use10191197
- 19. Smahel D, Blinka L, Ledabyl O (2008) Playing MMORPGs: connections between addiction and identifying with a character. Cyberpsychol Behav 11: 715–718.D. SmahelL. BlinkaO. Ledabyl2008Playing MMORPGs: connections between addiction and identifying with a character.Cyberpsychol Behav11715718
- 20. Chou TJ, Ting CC (2003) The role of flow experience in cyber-game addiction. Cyberpsychol Behav 6: 663–675.TJ ChouCC Ting2003The role of flow experience in cyber-game addiction.Cyberpsychol Behav6663675
- 21. King DL, Delfabbro PH, Zajac IT (2011) Preliminary validation of a new clinical tool for identifying problem video game playing. Int J Ment Health Addict 9: 72–87.DL KingPH DelfabbroIT Zajac2011Preliminary validation of a new clinical tool for identifying problem video game playing.Int J Ment Health Addict97287
- 22. De Prato G, Feijóo C, Nepelski D, Bogdanowicz M, Simon JP (2010) Born Digital/Grown Digital: Assessing the future competitiveness of the EU video games software industry. Luxembourg. G. De PratoC. FeijóoD. NepelskiM. BogdanowiczJP Simon2010Born Digital/Grown Digital: Assessing the future competitiveness of the EU video games software industry.Luxembourg
- 23. Collins LM, Lanza ST (2010) Latent class and latent transition analysis. Hoboken, NJ: Wiley. LM CollinsST Lanza2010Latent class and latent transition analysis.Hoboken, NJWiley
- 24. Vermunt JK, Magidson J (2002) Latent Class Cluster Analysis. In: Hagenaars JA, McCutcheon AL, editors. Advances in Latent Class Analysis. Cambridge: Cambridge University Press. JK VermuntJ. Magidson2002Latent Class Cluster Analysis.JA HagenaarsAL McCutcheonAdvances in Latent Class AnalysisCambridgeCambridge University Press
- 25. Muthén LK, Muthén BO (1998) Mplus User’s Guide. Fifth Edition. Los Angeles, CA: Muthén & Muthén. LK MuthénBO Muthén1998Mplus User’s Guide.Fifth Edition. Los Angeles, CAMuthén & Muthén
- 26. Altman DG, Bland JM (1994b) Diagnostic tests. 1: Sensitivity and specificity. BMJ 6943: 1552.DG AltmanJM Bland1994bDiagnostic tests. 1: Sensitivity and specificity.BMJ69431552
- 27. Glaros AG, Kline RB (1988) Understanding the accuracy of tests with cutting scores: the sensitivity, specificity, and predictive value model. J Clin Psychol 44: 1013–1023.AG GlarosRB Kline1988Understanding the accuracy of tests with cutting scores: the sensitivity, specificity, and predictive value model.J Clin Psychol4410131023
- 28. Altman DG, Bland JM (1994a) Diagnostic tests 2: Predictive values. BMJ 6947: 102.DG AltmanJM Bland1994aDiagnostic tests 2: Predictive values.BMJ6947102
- 29. American Psychiatric Association (2000) Diagnostic and Statistical Manual of Mental Disorders (4th ed., text revision). Washington, DC: American Psychiatric Association. Association American Psychiatric2000Diagnostic and Statistical Manual of Mental Disorders (4th ed., text revision).Washington, DCAmerican Psychiatric Association
- 30. Koronczai B, Urban R, Kokonyei G, Paksi B, Papp K, et al. (2011) Confirmation of the three-factor model of problematic internet use on off-line adolescent and adult samples. Cyberpsychol Behav Soc Netw 14: 657–664.B. KoronczaiR. UrbanG. KokonyeiB. PaksiK. Papp2011Confirmation of the three-factor model of problematic internet use on off-line adolescent and adult samples.Cyberpsychol Behav Soc Netw14657664