Evaluation of enhancements to convective nowcasting systems, and particularly nowcasts of convective initiation, is hampered by the impacts of large-scale storm characteristics as well as small-scale errors in timing and location. As part of an effort to evaluate the impacts of human forecasters on an automated convective nowcasting system (AutoNowCaster; ANC), some new approaches and evaluation tools have been applied, which provide remedies for some of these difficulties. This evaluation focused on a forecaster-interactive capability that was added to the ANC to allow forecasters to enhance the performance of one-hour nowcasts of convective storm initiation and evolution produced every six minutes. This Forecaster-Over-The-Loop (FOTL-ANC) system was tested at the National Weather Service Ft. Worth/Dallas Texas Forecast Office during daily operations from 2005-2010. The forecaster's role was to enter the locations of surface convergence boundaries into the ANC prior to dissemination of nowcasts to the Center Weather Service Unit. To identify the nowcast performance changes that might result from the forecasters' input, the evaluation approach aimed to isolate these changes from larger-scale performance variations. The approaches and methods used for the evaluation include (a) focusing the evaluation on small regions rather than the broader domain of the ANC; (b) stratifying the cases into similar forcing categories; (c) examination of time series of performance for specific cases; and (d) use of contingency table performance diagrams to simultaneously investigate changes in multiple forecast performance attributes. These approaches provided clear insights into the generally positive performance changes associated with the forecasters' input. The approaches could also have applicability to other forecast evaluation studies.