If you discovered a significant error in your analysis just before a presentation, what would you do?
