Knowing whether data are reliable is of fundamental importance to the establishment of knowledge, the formulation of explanatory hypotheses, and the development of effective policy.Yet there appear to be no standard, established tests to enable users to judge whether they should accept a given statistic as a fact.Numerous internationally agreed documents set out principles and practices King Bookcase Bed to promote sound statistics, but they offer no direct guidance on whether to accept data as presented.Other documents discuss statistical quality, Activity Toy but focus largely on utilitarian considerations such as availability and timeliness; when they do discuss accuracy, they again consider processes (lists of good practices) rather than results (are the data correct?).
This paper is a plea for, and a first attempt at, identifying some characteristics of data that may be accepted as true.