Sun, 30 September 2018
Today I discuss one of my own papers: "Instructions" by Freeman, Kimbrough, Petersen, and Tong. This research project on experimental instructions has been ongoing for years, but it was recently conditionally accepted for publication. I tell the story of how the research came together and detail some of the results.
A survey of instruction delivery and reinforcement methods in recent laboratory experiments reveals a wide and inconsistently-reported variety of practices and limited research evaluating their effectiveness. Thus we experimentally compare how methods of delivering and reinforcing experiment instructions impact subjects' understanding and retention. We report a one-shot individual decision task in which mistakes can be unambiguously identified in behavior and find that mistakes are prevalent in our base-line treatment which uses plain, but relatively standard experimental instructions. We find combinations of reinforcement methods that can eliminate half of subjects' mistakes, and we find that we can induce a similar reduction in mistakes via enhancements to the content of instructions. Residual mistakes suggests this may be an important source of noise in experimental studies.