How do you handle ambiguous scientific challenges when the data do not support your initial hypothesis?
