We experimentally test for the effect of an additional alternative on the measured risk preferences of farmers in rural Peru. In our experiment, subjects revealed their risk preferences with a series of choices between two gambles. We added a third gamble, which was always dominated by one of the two existing gambles. We found that subjects chose this gamble nearly one quarter of the time, in some cases causing the subjects to appear to be more risk loving. We found that subjects in a traditional laboratory environment did not choose the dominated gamble, but their choices were affected by its presence.