Tuesday, November 2, 12:30 pm Berkman Center, 23 Everett Street, second floor
The internet provides an unprecedented opportunity for social scientists to recruit large number of subjects quickly, cheaply and virtually effortlessly. Online labor markets, such as Amazon Mechanical Turk (MTurk), allow researchers to easily recruit and pay subjects from around the world to participate in studies which are monetarily incentivized (ie pay depends on choices in the study, rather just a flat rate). These labor markets also facilitate field studies, where 'subjects' are unaware they are in an experiment, but instead think they are just completing normal work tasks. The speed and easy of online experimentation has the potential to increase the rate of scientific progress by orders of magnitude. In this talk Dave will describe how we go about designing and running experiments using MTurk, some successful experiments we have had (mostly involving cooperative social dilemmas), and the lessons we have learned thus far.
Dave Rand is a Cooperation Fellow at the Berkman Center for Internet & Society, as well as a Harvard's Program for Evolutionary Dynamics and a FQEB Prize Fellow in the Psychology Department. Dave's work focuses on the evolution of human behavior, with a particular emphasis on cooperation, generosity and altruism. His approach combines (i) empirical observations from behavioral experiments with (ii) predictions generated by evolutionary game theoretic math models and computer simulations. David's research has been published in Science, Nature and PNAS, and featured on NPR's All Things Considered and Earth & Sky as well as in a range of print media.
Horton J, Rand DG, Zeckhauser RJ. (2010) The Online Laboratory: Conducting Experiments in a Real Labor Market National Bureau of Economic Research Working Paper Series No. 15961. (PDF)