The reliability of paleomagnetic data

A set of seven reliability criteria has been applied to a previously published Phanerozoic paleopole database for Europe and North America and a Late Precambrian data set for Africa. A quality factor (0 ⩽ Q ⩽ 7) is assigned to a result, based on the number of criteria satisfied. Three criteria, dealing with age reliability, structural control and laboratory demagnetization analysis are deemed the most important; for the Phanerozoic results these are satisfied by a large majority of the results, whereas for the majority (up to 80%) of the African Late Precambrian results such criteria are not met. Criteria based on tests that constrain the age of the magnetization, such as those dealing with folds, conglomerates, contacts or reversals, enhance the reliability of a result; for the Phanerozoic, they are generally satisfied by about one third of the data, but for the Precambrian only a few results incorporate such tests. The assertion is made in this study that these criteria indeed qualitatively describe the reliability of results in broad terms, so that a data set satisfying on average most of the criteria (Q ⩾ 4) can be described as more robust than a data set with average Q = 2. Statistical evaluations illustrate the difference in robustness of paleopole data sets between the well-studied Phanerozoic Era and the much more uncertain Late Precambrian.