In search of "Favorite-Long Shot Bias" : An experimental study of the demand for sweepstakes

This paper studies experimentally the demand for a variable-prize sweepstakes in which a single winner receives 90% of the total receipts, and whether such demand would exhibit the favorite-longshot bias (FLB) widely reported in the racetrack betting literature. We find significant incidence of sweepstakes purchase over population sizes ranging from 2 to 141, a greater tendency for FLB among those who exhibit longshot preference (LSP) over fixed-odds lotteries, but mixed support for FLB per se. In particular, the demand on average for 28-subject sweepstakes exceeds that of the largest one with 141 subjects including those who are averse to taking even-chance bets. Further and intriguingly, we observe significant demands for 2-person sweepstakes even among risk adverse subjects. Taken together, our observations reveal an incremental demand for sweepstakes arising from its interactive nature that can reinforce the effects of LSP as well as counter the effects of risk aversion