**4. Discussion**

In the present study, we firstly adapted the Gaming Motivation Scale (GAMS; [22]) to the Italian adolescents' context, examining its psychometric properties in a sample of young Italian students and then assessed the ability of this instrument to predict the presence of possible psychopathological aspects in the young population.

To these aims, after having verified the factorial structure of the GAMS-it and compared it with the original version, we run a Confirmatory Factor Analysis on a subsample of 388 participants with low (LG) and high (HG) frequency of gaming. Results support the six-factor structure of GAMS-it reporting satisfactory statistics fit. As shown in Table 1, all GAMS-it items reported a significant factor loading on the expected factor and all factors correlated significantly and in the expected direction (Table 2); in addition, the reliability also resulted in being satisfactory. In addition, high correlation between Integrated Regulation and Identified Regulation (θ = 0.945) did emerge, as also reported in the original publication [22]. Moreover, the multi-sample CFA highlighted the equality between the

various distributions of the correlation between the indicators and their dimensions, so the same six factor model with the same loadings in the two groups is tenable. The results confirmed the hypothesis of an invariant structure for the two groups. As a whole, such results showed adequate levels of validity and reliability of GAMS-it, as in the original validation study [22].

The comparison between two subgroups with respect to the six factors of GAMS-it and the result obtained from CESD and STAI showed a significantly higher score on Intrinsic motivation, Integrated regulation, Identified Regulation, Introjected Regulation, and External Regulation factor in HG, but not Amotivation. Moreover, HG showed a significantly lower score in the STAI, but not in the CESD (Table 3), indicating that a grea<sup>t</sup> amount of exposure to video games does not cause an increase of anxiety levels as usually believed.

Another interesting result is the one emerging from Latent Regression Structural Equation Model correlating CESD and STAI scores with latent six factor of GAMS-it, showing satisfactory fit (SB-χ2(144) = 248.22, *p* < 0.01; Table 3). Regarding CES-D scores, only the GAMS-it' Amotivation factor predicts significantly and positively the depression level (b = 0.247, se = 0.081, *p* < 0.001). This indicates that Amotivation levels can predict the depression level in a directly proportional way. Regarding STAI score, it emerged that not only Amotivation has a positive and significant effect (b = 0.390, se = 0.074, *p* < 0.001), but also that Intrinsic motivation has a negative and significant effect (b = −0.688, se = 0.238, *p* = 0.004). This indicates that Amotivation levels predict different levels of anxiety in a directly proportionate way; conversely, Intrinsic motivation levels predict indirectly the proportional anxiety levels. Finally, the two clinical measures, CES-D and STAI, reported a significant and positive latent correlation (r = 0.717, se = 0.047, *p* < 0.001). The present results only partially confirm the hypothesis of the study about a general correlation between motivation and anxiety/depression.

All these results allow to make some interesting conclusions. From SEM results, three conclusions can be drawn. As a first, we can confirm that the original GAMS factor structure can be generalized to a different culture (i.e., the Italian one). This means that factors underlying motivations to play video-games are the same in the two cultures. Therefore, this scale can be used to evaluate gaming motivation, providing additional statistic support to GAMS and introducing in the Italian context a new scale to be used as a potential screening instrument of gaming motivation. Nonetheless, further research is mandatory in order to establish and clarify psychometric property of the GAMS-it. Secondarily, multi-sample SEM confirmed that the GAMS factor structure is invariant in Heavy and Light gamers; this does mean that motivations to play video-games in both populations did not differ in terms of types of basic drive and that the differences between HG and LG are limited to a quantitative difference, along the same factor structure: Heavy gamers show higher average values with respect to Light gamers on all GAMS dimensions, but one (i.e., Amotivation). Finally, consequent to this, the Amotivation factor is the only one that significantly predicts psychopathological traits like depression and anxiety. So the present study, in conclusion, shows (and adds to the literature) that motivation of both Heavy and Light gamers is the same, that it differs somewhat on average between the two groups, and that these differences may be used to discriminate between them; of greater importance, our study shows that Amotivation is not critical for differentiating HG from LG, but is crucial as a precursor of depression and anxiety, in the sense that high levels of Amotivation predict high levels of depression and anxiety scores, namely, lower gaming motivation can be related to high level of depression and anxiety. These data are in line with current literature on depression and anxiety symptoms in adolescence, where anhedonia, apathy and loss of interest for all kind of activity can predict psychopathological outcomes [41]. An opposite effect has instead been reported for Intrinsic Motivation factor (limitedly to STAI scale), showing that high levels of Intrinsic Motivation can predict a reduced level of anxiety.

The present data contribute to explain the not ever consistent literature on the effects of exposure to videogames on psychopathological components as depression and anxiety. For example, the relationship between videogaming and psychopathological symptoms is still unclear; is psychopathology to induce the approach to VG (and thus possible excessive use and addiction) or does the exposure to VG tend to unmask a latent condition of depression and anxiety? In our opinion, the results obtained in the present study can help to clarify the problem. Indeed, if a lack of gaming motivation predicts high levels of depression and anxiety, we could suppose that persons with depression and anxiety will not be interested in videogaming, as they are unmotivated. Therefore, a state of anxiety and depression cannot be considered a factor able to encourage the use of video games. This assertion is also confirmed by negative relationship between high Intrinsic Motivation to video games use and anxiety. Healthy participants seek wellness in video games, because technology increases well-being-inducing vigor and resistance in players [42].

Furthermore, research based on Self Determination Theory (SDT) revealed that self-determined forms of motivation induce adaptive consequences as pleasure, persistence and wellness [43]. This statement could also be extended to pathological gambling: Thus, people generally anxious and/or depressed would be not encouraged to play videogames or gambling because of their reduced levels of motivation. In this line, anxiogenic and depressive symptoms of patients with Internet Gaming Disorder (IGD) may emerge after exposure to video games. However, the multifactorial aspects of videogaming do not allow us to consider these definitive conclusions, and more research is needed to clarify the effective relationship among the different factors involved.

As a limitation of the study, the sample size should be highlighted. Further validation work should test the GAMS-it on larger samples in order to obtain a greater statistical validity. For example, a possible future direction could be a full cross-sectional sample, including also a group of "medium level" players. Moreover, another limitation arises from the limited availability of participants to complete psychopathological questionnaires. Also, the two subgroups (HG and LG) considered were unbalanced with respect to gender composition; this did not allow us to investigate the issue of gender weight, although it should be borne in mind that the differences between LG and HG with respect to psychopathological traits could arise by sample unbalancing. Finally, all data were collected using self-reports, which could lead to common problems of counterfeiting. Further studies are needed not only to replicate our results but also for testing the GAMS-it with multiple reports (i.e., relatives, friends) or with behavioral and objective measures.

Future prospects may be the application of GAMS-it on participants with IGD diagnosis, or to investigate the relationship between GAMS-it subscales, gender, and the preference for different kind of video games. Also, the correlation with other factors as aggression, impulsivity or addiction could be studied. Another issue to be studied is the correlation between videogames motivation and specific VG characteristics, namely graphic characteristics, immersivity, game option and history, and their potential link with mental and behavioral health [44]. Finally, it would be interesting to evaluate gaming motivation in different age ranges, in order to have an overview of gaming motivation at every time of life.
