The Identity-Centered Attributional Reversal Theory (ICART)

AbstractThis paper proposes the Identity-Centered Attributional Reversal Theory (ICART), aconditional theoretical framework for investigating how strongly identity-central beliefs,ideologies, or group memberships may influence the attribution of bias to external individuals,institutions, and information systems. Its primary case concerns a politically identifiedobserver who interprets apparently balanced media coverage as favouring an opposing politicalparty. The proposed mechanism extends beyond the initial perception of media bias: theobserver's attribution may change subsequent source evaluation, evidence weighting, memory,and judgment in ways that favour the observer's own identity. Under specified conditions,those changes may reinforce the original attribution, producing a self-reinforcing feedbackprocess termed the Attributional Reversal Loop (ARL).The framework distinguishes independently audited directional bias, perceived directionalbias, signed attribution discrepancy, hostile-intent attribution, identity-congruent informationprocessing, and longitudinal reinforcement. It presents a layered ontology, operationaldefinitions, bounded statistical models, a dynamic feedback formulation, eight testablehypotheses, competing explanations, and a staged empirical research programme. Fourresearch components are proposed: observational measurement, controlled experimentation,longitudinal feedback testing, and cross-domain replication.ICART is presented as a research proposal rather than an established universal law.Its central contribution is the hypothesis that perceived outgroup-favouring bias can bean antecedent of the observer's own identity-congruent processing, which may in turn re-inforce subsequent bias attribution. The existence, direction, magnitude, causal structure,generalisability, and novelty of this complete feedback mechanism remain open empiricalquestions.Keywords: identity centrality; identity salience; attributional reversal; hostile media perception;motivated reasoning; directional bias; epistemic reweighting; information processing; feedbackdynamics; social identity; causal inference.