[Air-L] Establishing rationale for rejecting completed MTurk tasks

Can anyone point in the right direction for literature establishing some sound guidelines for rejecting poor quality completions from MTurk workers? For example, my survey had an average completion time of 9:30, but I have respondents “complete” the survey in less than a minute. I would think it would be appropriate to reject these workers responses? Additionally, I have responses that I have reviewed where a worker selected the middle answer for all survey questions. I would assume that’s valid to reject this completion as well. Many thanks in advance, Cory -- Stephen Cory Robinson cory.robinson@colostate.edu<mailto:cory.robinson@colostate.edu> Office: Clark C258A http://colostate.academia.edu/StephenCoryRobinson