The Over-Inference Paradox (OIP): An Expectation–Absence Detection–Inference (EADI) Framework

Under conditions of information asymmetry, evaluators often draw inferences from information that is expected to be present but is absent. However, existing research on absence-based inference generally assumes that evaluators and targets share the same underlying norms, paying limited attention to the systematic misjudgments that arise when this assumption breaks down across institutional contexts. This conceptual paper integrates disclosure economics, auditing methodology, and intelligence analysis, while engaging with signaling theory, sensemaking, and ecological rationality, to develop the Expectation–Absence Detection–Inference (EADI) framework. The framework explains how evaluators may systematically misinterpret normative differences as indicators of poor quality. Specifically, the paper introduces the Over-Inference Paradox (OIP), proposing that the defensibility of an expectation suppresses systematic false positives when norms are shared but, paradoxically, increases systematic false positives when norms are not shared, thereby producing opposite effects under the two conditions. The proposed mechanism suggests that defensibility increases cognitive confidence, reduces the perceived necessity of procedural verification, and leads evaluators to bypass the Sharedness Check, a prerequisite for the third‑stage inference verification process, thereby preventing inference verification from being initiated. The paper further develops three theoretical propositions together with their operationalization and research designs. By reconceptualizing norm sharedness from an implicit background assumption into a procedural step requiring independent verification, it introduces the Sharedness Check and verification gates as mechanisms for reducing systematic false positives in cross‑institutional evaluations. The central theoretical insight is that defensibility is not equivalent to applicability. This distinction reconstructs the analytical framework of absence‑based inference and offers theoretical and practical implications for cross‑border evaluation, managerial judgment, and decision quality under conditions of information asymmetry.