In statistics, usually hypotheses are created before the data is examined. However, in today's high data output environment, many findings may be examined further after the data is analyzed. Do you think that it is a good approach to generate such hypotheses after the data has been analyzed? Why or why not?