In this paper the authors report results of an initial attempt to test the theory of expressive voting. Their experience involves requiring subjects to vote between receiving $A in cash or having $B donated to charity on their behalf. Across subjects they vary the probability that their vote will decide which disbursement is made. As the probability of being decisive is lowered, the opportunity cost of giving expression to charitable sentiment decreases; hence, the likelihood of voting in accordance with such sentiment is expected to increase. They find weak support for the hypothesis, but the results do not replicate when they alter the parameter settings. Limitations of their design are discussed. Copyright 1992 by Kluwer Academic Publishers
