Selection model approaches evaluate whether the pattern of results that have been accumulated from a number of studies suggests an underlying filtering process, such as the non-publication of a given percentage of results that had not reached formal statistical significance [14–16]. These methods have been less widely applied than small-study effect tests, even though they may be more promising. One application of a selection model approach examined an entire discipline (Alzheimer’s disease genetics) and found that selection forces may be different for first/discovery results versus subsequent replication/refutation results or late replication efforts [17]. Another method that probes for data “fiddling” has been proposed [18] to identify selective reporting of extremely promising p-values in a body of published results.
