THE DISCREPANCY BETWEEN PUBLIC SUPPORT FOR GUN REGISTRATION AS MEASURED IN NATIONAL SURVEYS AND LEGISLATIVE INACTION ON THE SAME ISSUE LEADS TO QUESTIONS ABOUT THE ADEQUACY OF TRADITIONAL POLL MEASURES OF PUBLIC OPINION. HYPOTHESES ABOUT THREE KINDS OF SHORTCOMINGS ARE TESTED: (1) GUN REGISTRATION SENTIMENT TENDS TO VARY APPRECIABLY WITH QUESTION WORDING, AND THEREFORE IS LESS CRYSTALLIZED THAN SURVEY DATA SUGGEST; (2) ANTIGUN REGISTRATION OPINIONS ARE HELD WITH GREATER INTENSITY THAN ARE PROGUN REGISTRATION OPINIONS, AND THEREFORE HAVE DIFFERENT CONSEQUENCES FOR POLITICAL ACTION; AND (3) OPPOSITION TO GUN REGISTRATION IS LOCATED PARTICULARLY AMONG THOSE WITH GREATER POLITICAL KNOWLEDGE AND INFLUENCE, AND THUS HAS A DISPROPORTIONATE IMPACT ON LEGISLATORS. ALTHOUGH THERE IS SOME EVIDENCE IN SUPPORT OF THE FIRST TWO HYPOTHESES, IT DOES NOT APPEAR SUFFICIENT TO EXPLAIN THE DISCREPANCY. HOWEVER, THE STUDY PRODUCES A VALUABLE FINDING ABOUT THE USE OF INTENSITY MEASURES FOR MAXIMUM SENSITIVITY TO NONLINEAR RELATIONS. TABULAR DATA AND REFERENCES ARE INCLUDED. (AUTHOR ABSTRACT MODIFIED--BAC)