Reference-Point Dependence of Judgment: How the Baseline Against Which an Event Is Measured Changes What It Is Judged to Be

This conceptual paper introduces Reference-Point Dependence of Judgment, the mechanism by which changing a baseline, comparison class, or expected alternative can change the relative assessment of an event held fixed. The event can remain unchanged while the judgment changes. The account distinguishes the dependence built into relative judgments from the empirical question of whether a reference changes what people conclude or do. It assesses references by their suitability to the purpose of the judgment and by whether they are disclosed. A changed assessment does not by itself establish misleading use. The paper relates the mechanism to framing and prospect theory and examines reference periods in global temperature reporting. The case shows that reported anomalies can change while trends and rankings remain unchanged; it does not establish that readers were misled. A related discussion distinguishes changing a reference from expressing the same comparison through a different measure. Part of the Generative Operations series within On Human Modeling, the paper offers questions for identifying references, assessing their relevance, and examining who chooses them. These questions can inform human–AI research on how comparison baselines shape the interpretation of AI outputs and whether those references are visible and appropriate to users’ purposes.