Gaming Disorder Scale Reveals Two Faces of Problematic Gaming in U.S. Players

A widely used screening tool for internet gaming disorder may be capturing two separate things at once, according to a new open-access study published in the International Journal of Mental Health and Addiction. When researchers administered the Internet Gaming Disorder Scale-Short Form, known as the IGDS9-SF, to 620 current U.S. gamers, their statistical analysis pointed to a two-factor structure rather than the single dimension assumed in most prior validation work. One factor clustered around core addictive symptoms such as preoccupation and loss of control, while the other gathered items describing functional impairment in relationships, work, or study. The authors label these Symptom-related Internet Game Disorder (SIGD) and Impairment-related Internet Game Disorder (IIGD).

The IGDS9-SF was developed by Pontes and Griffiths in 2015 and mirrors the nine core criteria that the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders uses to define internet gaming disorder. Respondents rate nine items on a five-point scale from Never to Very Often, producing a total score from 9 to 45, with higher scores indicating greater severity. The scale was designed to measure the severity of the disorder and its harmful effects in both online and offline gaming-related activities, and although it was built for research purposes, its developers suggested that people scoring 36 or above out of 45 could be classified as disordered gamers. Until now, validation studies across different languages and regions have repeatedly supported treating the instrument as unidimensional.

The new study set out to test that assumption among a more recent and geographically specific population. Participants were recruited between March and December 2023 through Reddit and Steam, with the survey distributed across 45 Steam discussion boards and subreddits devoted to video games and related topics, including communities centered on titles such as Overwatch, Dota 2, and League of Legends. Only participants who identified the United States as their country of residence were included. The authors argue that U.S. gamers differ from players in other regions in device and game access, time spent gaming and on game-related social platforms, and gaming culture, and that the post-pandemic centrality of social and online play may have shifted how problematic gaming presents itself.

The sample was heavily skewed toward men, at 86 percent, with a mean age of 27.28 years and a range from 18 to 59. Participants reported spending an average of 22 hours per week playing online games, with a standard deviation of 14.7, and roughly 15.6 hours per week on game-related social platforms such as Reddit, Steam, Twitch.tv, and Discord. Team-based action games were the most popular favorite genre at 41 percent, followed by massively multiplayer online games at 13 percent, sports games at 12 percent, and sandbox games at 10.5 percent, with the remaining quarter of the sample preferring battle royale, multiplayer online battle arena, action role-playing, and strategy games. The survey, which took about 15 to 20 minutes to complete, also included the Metacognitions about Online Gaming Scale, a 12-item measure of maladaptive beliefs about gaming developed by Spada and Caselli in 2017.

Methodologically, the researchers split their data in two. An exploratory subset of 116 participants was used for Exploratory Factor Analysis, while the remaining 504 were reserved for a Multidimensional Graded Response Model, an item-response-theory approach suited to polytomous Likert-type items. In the exploratory phase, Horn’s parallel analysis indicated retaining two factors, because only the first two observed eigenvalues exceeded those generated from random data. The Kaiser-Meyer-Olkin measure of sampling adequacy was .86, which the authors describe as meritorious, and Bartlett’s test of sphericity was significant, indicating that the items shared underlying latent factors. The two-factor solution explained around 54 percent of the total variance.

The first factor contained six items and was labeled SIGD, reflecting self-centered struggles with gaming, while the second contained three items and was labeled IIGD, reflecting difficulties with others such as family and friends. The two factors correlated at .66. Reliability estimates were satisfactory, with McDonald’s Omega of .80 for SIGD and .71 for IIGD. In the confirmation phase, the multidimensional model fit the data substantially better than the unidimensional one, with RMSEA of .037, SRMSR of .039, CFI of .991, and TLI of .988, compared with RMSEA of .080, SRMSR of .062, CFI of .957, and TLI of .943 for the single-factor model. A chi-square difference test favored the two-dimensional solution, and all relative information criteria, including AIC, BIC, and sample-size-adjusted BIC, pointed the same way. Item discrimination estimates ranged from 0.763 to 2.641, suggesting the items distinguish well between high and low levels of the construct, and item-level fit statistics were acceptable.

The authors connect their two factors to a long-standing distinction in addiction research between core addictive symptoms, such as preoccupation, tolerance, withdrawal, and loss of control, and functional impairment, such as jeopardizing relationships or academic and occupational functioning. They note that the item content of the second dimension, covering continuation despite psychosocial problems, deception, and jeopardized or lost relationships or opportunities, parallels the social-impairment domain recognized within the DSM-5 substance use disorder framework from which the gaming disorder criteria were adapted. The pattern also echoes earlier work: a network analysis by Gomez and colleagues in 2022 found that the three symptoms corresponding to the study’s second-factor items form a closely related subset, and Gisbert-Pérez and colleagues in 2025 identified clusters associated with cognitive-emotional symptoms and relational difficulties. The authors caution, however, that the divergence from prior unidimensional validations may reflect methodological differences, since earlier confirmatory factor analyses tested item covariance against a single-factor specification, and they call for direct comparative testing rather than treating either finding as definitive.

Correlational analyses added further validity evidence. SIGD was positively associated with time spent playing online games (r = .236), time spent on game-related social media platforms (r = .137), negative metacognitions about the uncontrollability of online gaming (r = .523), negative metacognitions about the dangers of online gaming (r = .468), and positive metacognitions about online gaming (r = .377). IIGD was positively correlated with time spent gaming (r = .152), negative metacognitions about uncontrollability (r = .490), negative metacognitions about danger (r = .454), and positive metacognitions (r = .13). Notably, SIGD but not IIGD correlated with time on game-related social media. The authors suggest that exposure to online gaming communities may be associated with problematic symptoms given their pervasiveness, while also observing that such communities can foster social inclusion and belonging, meaning they do not always translate into functional impairment. These are associations in a cross-sectional sample, so no causal direction can be inferred.

The study’s limitations are considerable and the authors state them plainly. The sample of 620 was randomly divided into two halves for different analyses, and although each phase met established sample-size recommendations, the authors acknowledge that rules of thumb in item analysis remain debated and that larger samples would strengthen confidence. Recruitment through Reddit and Steam likely excluded casual gamers and players who avoid online discussions, and the predominantly male sample limits generalizability. Whether the two-factor solution holds for gamers of other nationalities, cultures, and time frames requires further validation, differential item functioning analyses were not possible due to sample size restrictions, and all data were self-reported, leaving open the possibility of response bias.

Even with those caveats, the findings carry practical weight. If problematic gaming genuinely splits into symptom-driven and impairment-driven components, screening could eventually distinguish players experiencing internal struggles with gaming from those whose relationships or responsibilities are visibly suffering, informing earlier and more targeted interventions. The authors suggest that practitioners may test and revise the two subscales as a possible screening tool pending further psychometric evaluation, and that future research should examine the subfactors’ predictors and relationships in more detail, including the role of gaming-related social media. Because no prior published study has reported the IGDS9-SF’s multidimensionality, replication in different contexts and populations is the essential next step before the two-factor structure can be treated as established.

Subject of Research: Validation of the Internet Gaming Disorder Scale-Short Form and its two-factor structure among U.S. gamers

Article Title: Investigating Problematic Gaming Among Current U.S. players: Exploring the Multidimensionality of the Internet Gaming Disorder Scale-Short Form

Article References: Gandolfi, E., & Soyturk, I. (2026). Investigating Problematic Gaming Among Current U.S. players: Exploring the Multidimensionality of the Internet Gaming Disorder Scale-Short Form. International Journal of Mental Health and Addiction. https://doi.org/10.1007/s11469-026-01757-6

Image Credits: AI Generated

DOI: 10.1007/s11469-026-01757-6

Keywords: internet gaming disorder, IGDS9-SF, psychometric validation, graded response model, factor analysis, problematic gaming, U.S. gamers, functional impairment, gaming metacognitions, DSM-5, online gaming, scale reliability