In November 1972, electorates in California, Colorado, Michigan, and Oregon decisively rejected consitutional amendments that (according to their supporters) would have reduced or eliminated reliance on the property tax as a means of financing education. School finance reformers were perplexed by these defeats. This study sets out to explain the gross inconsistency between the desire, as sampled by widely-accepted polls, of citizens to eliminate the property tax and their behavior in the voting booth. The study's research design combines two methodologies: case studies, including interviews and analysis of secondary sources (campaign literature, etc.), and aggregate voting analysis, using both correlation and multiple regression techniques in order to link support of amendments to census characteristics of localities. The findings indicate two major explanations for the amendments' defeat: the content of the proposals and the constitutional amendment process itself. Socioeconomic status variables were not found to be important explanatory variables. (Author/DN)