UCR DATA ARE EXAMINED IN TERMS OF LANGUAGE, COLLECTION PROBLEMS, PECULIARITIES IN DEFINITION, ANALYTICAL PROBLEMS THEY PRESENT, AND SYSTEMATIC BIASES. IT IS ARGUED THAT UCR DATA ARE MEANINGFUL ONLY WHEN THEY ARE PART OF AN ELABORATE SET OF INTERRELATED TENDENCY STATEMENTS OR LAWS (FACTOR THEORY). THIS TYPE OF LAW DOES NOT CURRENTLY EXIST AT THE LEVEL OF CRIMINAL OR DEVIANT BEHAVIOR. THE EXTENT TO WHICH UCR DATA COULD BE USED AS VARIABLES IN A DEVELOPING FACTOR THEORY AND THE ADEQUACY OF THE STATUTORY LANGUAGE FOR USE IN THEORY CONSTRUCTION ARE EXAMINED. IT IS CONCLUDED THAT UP UNTIL THE 1974 REVISION OF UCR CATEGORIES, UCR LANGUAGE WAS INADEQUATE FOR ANY SCIENTIFIC RESEARCH IN CRIME AND DEVIANCE. IT IS ARGUED THAT THERE ARE LOGICAL AND EMPIRICAL GROUNDS FOR BELIEVING THAT POLICE PROCEDURES LEAD TO THE CREATION OF SPURIOUS PATTERNS OF OFFENSE RELATIONSHIPS; THE SAME STRUCTURAL EFFECTS INFLUENCE POLICE AND THE REPORTING PUBLIC AND MAKE DATA UNRELIABLE; AND CONTEXTUAL DIFFERENCES IN THE SOCIAL, POLITICAL, AND DEMOGRAPHIC MAKEUP OF COMMUNITIES AND LOCAL POLICING STYLES WILL PRODUCE REGIONAL DIFFERENCES IN CORRELATIONS BETWEEN SOCIOLOGICAL VARIABLES AND UCR DATA. THE RESIDUAL PROBLEMS OF UCR ANALYSIS AND USE, SUCH AS AGGREGATION BIAS AND COMPOSITIONAL EFFECTS, ARE CONSIDERED. INITIAL STEPS FOR IMPROVING THE RATIONAL USE OF UCR DATA ARE PROPOSED AND REFERENCES ARE PROVIDED. (AUTHOR ABSTRACT MODIFIED--DAG)