Mechanical Turk for Social Scientists

Here are slides from a recent symposium on how to use Amazon's Mechanical Turk for social science research. These cover everything from the similarities and differences between Mturk samples, representative sample, and convenience sample, the surprising downsides of manipulation and attention checks and higher payments to subjects, and practical advice on how to use the service for pretty sophisticated surveys, experiments, and other types of research. Particularly interesting is the possibility of conducting surveys and other research in multiple countries for very low cost. I'm hoping to run some survey experiments on Mturk soon and these slides, as well as the other resources to which they link, are a great intro.