An investigation into factors associated with adolescent internet gaming disorder: utilizing machine learning and...

With the evolving structure of modern internet gaming, adolescent internet gaming disorder (IGD) has become an increasingly serious public health concern demanding societal attention. Recent research has shifted towards employing more robust machine learning algorithms to explore key risk factors for adolescent IGD. Consequently, this study aims to investigate the mechanisms underlying the onset and progression of adolescent IGD through the application of multiple machine learning algorithms and path analysis model. Consequently, this study utilized four machine learning algorithms to screen 15 risk factors from a large sample of 4501 adolescents (the average age was 14.23 ± 1.43 years, 52.3% girls) in a Chinese province, identifying the top four significant features (craving, avatar identification, character attachment, and ego-identification). Subsequently, a path analysis model grounded in the I-PACE theoretical framework was constructed to elucidate the developmental pathways of adolescent IGD. Results revealed one direct pathway (accounting for 42.2% of total effect) and five indirect pathways (mediation effect comprising 57.8%), substantiating the key pathway mechanisms of avatar identification and character attachment. The findings suggest that avatar identification represents a novel risk factor for adolescent IGD warranting greater attention. Adolescents seeking ego-identity within virtual gaming spaces may develop cognitive and emotional responses that constitute a novel pathway to gaming addiction.