Selection of Encountered Evidence: How Attention, Filtering, and Sampling Change the Evidence a Person Meets

The evidence a person meets is a selection. Attention, filters, and sampling rules decide which cases and features are recorded, presented, or noticed, and the encountered evidence can differ in composition from the population or events it is used to describe, even when those are unchanged. This paper names that operation the Selection of Encountered Evidence. Its general standard is adequacy for the specified inference; for claims about a population, that means representativeness or a justified adjustment for how the cases were selected. Selection is not bias by itself. It threatens an inference when the selection process affects the target claim in a way the analysis does not adequately account for. The paper relates the operation briefly to inattentional blindness and to the statistics of sample selection, and uses Abraham Wald's wartime analysis of damage to returning aircraft as a documented case. Part of the On Human Modeling series on Generative Operations. The paper is conceptual.