bogus observation

A fictitious observation inserted into an analysis that is believed to improve the analysis.