The dual architecture of digital fidelity: empirical evaluation and refinement of a theoretical model, and examination of gender invariance

Introduction This study aims to empirically evaluate and refine a theoretical model that conceptualizes digital fidelity based on Individual Ethics and Relational Boundaries (The Dual Architecture). Methods The model’s structural, discriminant, and cross-gender invariance validity were examined using both Classical Test Theory (CFA, WLS-MV) and Modern Test Theory (Rasch). Results CFA empirically revised the two-factor Dual Architecture structure, and multi-group CFA confirmed strict invariance across genders. Nomological network analysis (relationship satisfaction, attachment styles) provided evidence for the discriminant validity of F1 and F2. Conversely, the Rasch analysis revealed a critical methodological challenge: The F1 (Digital Fidelity Ethics) sub-dimension showed unacceptably low person-separation reliability (PSI = 0.432) and a severe targeting issue (ceiling effect). Discussion The findings indicate that the Dual Architecture model shows promising theoretical alignment and comparability; however, the Ethics dimension’s measurement using current self-report items presents significant methodological difficulties, likely due to social desirability bias.