Background and aims: The components model of addiction outlines six criteria shared by all addictive disorders. This proposal has been widely applied to conceptualize behavioral addictions, including problematic social media use (PSMU). However, certain criteria can be defined as ?core" (e.g., mood modification, relapse, withdrawal, conflict), reflecting problematic involvement, while others as ?peripheral" (e.g., salience, tolerance), reflecting non-problematic involvement. We evaluated whether a two-factor model distinguishing between core and peripheral criteria provides a better fit than the unifactorial model in PSMU. Additionally, we examined whether core and peripheral criteria exhibit different patterns of association with psychological measures. Methods: A total of 2,761 adolescents (M 5 14.80 years, SD 5 1.91 years) completed the Bergen Social Media Addiction Scale (BSMAS), the Social Media Disorder Scale (SMD), and measures of depression, anxiety, loneliness, life satisfaction, and self-esteem. Confirmatory factor analyses compared one-factor and two-factor models for the BSMAS and SMD. Associations were evaluated using structural equation models. Results: A two-factor model that distinguished core (i.e., mood modification, relapse, withdrawal, conflict) and peripheral (i.e., salience, tolerance) criteria provided a better fit than the unifactorial model for both scales. Core criteria were positively associated with depression, anxiety, and loneliness, and negatively associated with life satisfaction and self-esteem. Opposite patterns were observed for peripheral criteria: they were negatively associated with depression, anxiety, and loneliness, and positively associated with life satisfaction and self-esteem. Discussion and Conclusions: Our findings suggest that the components model of addiction may not be valid for assessing PSMU, promoting overdiagnosis and pathologization.  Kulcsszavak: components model of addiction, confirmatory factor analysis, core and peripheral criteria, problematic social media use, structural equation modeling