A failure to communicate: an experimental investigation of the effects of advice on strategic play

We investigate why two person teams beat the truth wins benchmark in signaling games (Cooper and Kagel, 2005 ;  Cooper and Kagel, 2009) by studying an advice treatment where advisees have the benefit of two individuals׳ insight but not bilateral communication. The TW benchmark states that the performance of a team for problems with a demonstrable correct solution should be no worse than the performance of its most able individual. If one individual solves the problem, the team solves it as well. Advisees whose advisors play strategically have significantly higher levels of strategic play, but fall well short of the truth wins benchmark as (i) many advisors who play strategically do not provide advice to this effect, and (ii) many advisees fail to follow sound advice. Effect (i) is largely attributable to female advisors who are far less likely to provide advice than men. Whether or not advice contains a clear explanation for its effectiveness has no effect on the likelihood of advisees’ strategic play, a result at odds with the idea that economic agents regularly consider all the available alternatives.