ALTHOUGH MANY JURISDICTIONS ALLOW VOTERS TO DIRECTLY PARTICIPATE IN PUBLIC PERSONNEL DECISIONS BY MEANS OF REFERENDUMS, SUCH PROCESSES MAY NOT BE THE OPTIMUM METHOD FOR SETTLING COMPLEX ISSUES. A STUDY OF 28 LOCAL CAMPAIGNS EXAMINES THE WORKABLITY OF REFERENDUMS FOR DECISIONMAKING. DATA WERE GATHERED IN 28 CITIES WHOSE CITIZENS VOTED ON THE TEXAS FIRE AND POLICE EMPLOYEE RELATIONS ACT (FPERA), WHICH PROVIDES COLLECTIVE BARGAINING RIGHTS FOR POLICE AND FIRE FIGHTERS IF SUCH RIGHTS ARE GRANTED IN A LOCAL OPTION REFERENDUM. THE PRIMARY PROPONENTS OF THE FPERA ADOPTION REFERENDUMS WERE THE FIRE AND POLICE ASSOCIATIONS DIRECTLY INVOLVED, BUT THE LOCAL LABOR COUNCILS GENERALLY PROVIDED NO ACTIVE SUPPORT. OPPOSITION TO THE REFERENDUMS CAME FROM DIVERSE GROUPS INCLUDING THE LOCAL GOVERNMENTS AND AD HOC GROUPS, BUT IN 11 OF THE REFERENDUMS THERE WAS NO OPPOSITION. CAMPAIGN EXPENDITURES VARIED FROM 1.1 TO 33.3 CENTS PER VOTER, WITH THE HEAVIER EXPENDITURES REFLECTING THE USE OF THE BROADCAST MEDIA. PRESS RELEASES WERE THE MOST COMMON FORM OF ATTEMPTING TO INFLUENCE VOTERS, FOLLOWED BY RADIO ADVERTISING AND NEWSPAPER ADVERTISING. ALTHOUGH MOST PERSONS KNEW FEW DETAILS OF THE LABOR ISSUE, THE CAMPAIGNS DID LITTLE TO ACTUALLY INFORM PEOPLE. INSTEAD, BOTH SIDES APPEALED TO ALREADY EXISTING BIASES. CONSEQUENTLY, THE OUTCOMES OF THE RERERENDUMS WERE NOT THE RESULT OF INFORMED CHOICES SINCE THE VOTERS HAD LITTLE OR NO HARD INFORMATION UPON WHICH TO BASE THEIR DECISIONS. TABLES OF THE CAMPAIGN METHODS AND ARGUMENTS ARE PROVIDED. (TWK)