How would you design an experiment to test whether a new bug-reporting interface improves the quality of submissions at Applause?
