I get asked a lot of questions from my students and mentees about using MTurk. How do editorial boards view it in terms of credibility? What are the good and bad sides of it in terms of data collection, cost, and publishability?
Christine, you might check Mason & Suri, 2012. We cited them in two new internet-based papers in Developmental Psychology, both Luthar & Ciciolla, (2015)
Who mothers mommy? Factors that contribute to mothers' well being, and What it feels like to be a mother: Variations by children's developmental stages.
They found that that parts of the MTurk sample heavily participate in research studies. Hence, they might not be "nonnaive" as the paper states. Furthermore they found that workers often do not pay attention to tasks, risking overall data quality of studies.