The Bergen Facebook Addiction Scale: its psychometric properties and invariance among women with eating disorders

OBJECTIVE: Facebook addiction is increasing, giving rise to limited real-life social networks, loneliness, poor work and academic performance, psychopathology, and low well-being. Facebook entails numerous factors that increase the risk for disordered eating attitudes and behaviors (e.g., use time and Facebook activities such as social grooming and photo sharing). This study aimed to evaluate the psychometric properties of the Bergen Facebook Addiction Scale (BFAS) among patients with eating disorders (EDs) given lack of validation of Facebook addiction measures in this population. METHODS: A cross-sectional study involving 123 inpatient and outpatient women with EDs (Mean age = 27.3, SD = 10.6, range = 14–59 years) used confirmatory factor analysis (CFA), multigroup CFA, structural equation modeling (SEM), Spearman’s rho Spearman’s analysis, McDonald's Omega (ω), Cronbach’s alpha (α), and item-total correlations to examine the structure, invariance, criterion validity, reliability, and discriminant validity of the BFAS. RESULTS: Correlating the residuals of items 2, 3, and 5 resulted in an excellent fit of a one-factor structure of the BFAS (χ(2)(7) = 8.515, p = .289, CFI = .998, TLI = .996, RMSEA = .042, SRMR = .0099). The BFAS was invariant at the configural, metric, and scalar levels across groups of EDs, age, education, and marital status. High values of ω and α (.96) as well as item-total correlations (.851–.929) indicated excellent reliability and high discrimination index of the BFAS. Criterion validity is noted by strong positive correlation with the Six-item Internet Addiction Test (S-IAT, r = .88) and SEM using the S-IAT to predict the BFAS (χ2(49) = 103.701, p = .001, CFI = .975, TLI = .966, RMSEA = .096, SRMR = .0317).. CONCLUSION: The BFAS is a reliable unidimensional measure. Its high discrimination index and invariance across different groups make it useful for detecting Facebook addiction among patients with ED. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s12905-022-01677-2.