as the survey has too few responses and numerous duplicates that are skewing results, you decide to repeat the survey in order to create a new, improved dataset. what is your first step?