ABSTRACT
This chapter discusses common concerns that researchers have with Mechanical Turk (MTurk), reviewing the evidence that bears upon each concern. It suggests that readers are knowledgeable about best practices in web study design. To motivate MTurkers to be attentive, studies should follow general principles of web survey design and be no longer or tedious than necessary. Although a large amount of MTurkers are from the United States or other predominantly English-speaking nations. However, in English language tasks, they nevertheless give poorer-quality responses than American MTurkers. Experienced MTurkers may learn that deception is common in social science research and presume that researchers are likely to be deceptive, which may change the way they respond. A variation of this concern is that repeated participation in a specific paradigm can lead participants to acquire knowledge or experience that may change how they respond. Survey platforms may sometimes not measure attrition unless participants answer at least one question, leading attrition rates to be underestimated.
