The Bergen Social Media Addiction Scale Validity in a Romanian Sample Using Item Response Theory and Network Analysis

The validation of the Bergen Social Media Addiction Scale (BSMAS) in different cultural contexts is contributory in adding to the extant research body on psychometric qualities of the scale. The main purpose of this study was to examine the validity of the BSMAS, based on classical test theory and advanced psychometric testing, namely item response theory (IRT) and network analysis in a Romanian sample of 705 participants (39% males, aged 18–54 years, M = 30.24, SD ± 9.15). Composite reliability and McDonald’s omega proved the very good internal consistency of the BSMAS-RO. The CFA highlighted the unifactorial model fitted well and measurement invariance across gender was revealed. Females obtained significantly higher scores on social media addiction (SMA) than males. As for concurrent validity, SMA correlated with social media intensity use, anxious attachment, social anxiety, need to belong, self-esteem, and happiness. IRT analysis proved adequate discrimination and difficulty parameters for all items. The highest discrimination was registered by salience criterion and the highest difficulty by conflict, withdrawal, and mood modification criteria. Relapse criterion had the lowest values of both parameters. Network analysis revealed that salience and withdrawal criteria had the highest indicators of centrality and relapse the lowest. Our findings revealed that the most relevant criteria to the diagnosis of SMA in Romanian people are salience, conflict, withdrawal, and mood modification. The robust psychometric properties of the BSMAS-RO provide health professionals a valid instrument for assessing SMA. Programs to prevent SMA must focus on dispositional traits as AATT, social anxiety, respectively unsatisfied need to belong, and promoting the ability to initiate and maintain rewarding social relationships.