Taiwan's first nation-wide referendum, the "Peace Referendum," was successfully held on 20 March 2004. However, its two propositions were rejected due to low voting rates that failed to meet the minimum threshold required by law, namely, one half of total eligible votes. One frequently-cited explanation for the failed referendum results is that opposition parties discouraged their supporters from participating in the "peace referendum." However, this explanation is based on the simple hypothesis that the gap in party cleavage is sufficient to explain the tumout/approval rate of the referendum. Using voting records provided by the Central Election Commission, matched with aggregate data of revenue/expenditure and social conditions at the township/district level, in this paper, this paper attempts to verifY the hypothesis above, while clarifying other significant factors that may have affected the results of the peace referendum. The findings show that the impact of party cleavageon the tumout rate for the referendum was both significant and dominant, although its effect on the approval rate of the two referendum propositions was less obvious. In addition to party c;eavage,the results suggest that the higher the education level, the larger the proportion of net-immigrants, and the greater the concentration of Hoklo (Min-nan) people all led to a higher turnout rate for the referendum. Further examination of the survey data is required, however, in order to determine whether the above aggregate-level findings reflect the actual process of voting intention and orientation at the individual level.