Generically, this is any analysis (systematic or otherwise) in which one looks for discrepant data points. In the context of NANOGrav, we use this to describe automated methods for TOA rejection. The primary algorithm used is described in Vallisneri & van Haasteren (2017).