RESEARCH CONDUCTED WITH THE SUPPORT OF LEAA INDICATES THAT SOCIO-ECONOMIC CONDITIONS OF VICTIMS ARE CORRELATED WITH DECISIONS TO REPORT ROBBERIES. IF CERTAIN POPULATION GROUPS, SUCH AS WHITES, FEMALES, AND THE ELDERLY, ARE MORE LIKELY TO CALL THE POLICE THAN OTHERS, THEN ROBBERY STATISTICS OF THE UNIFORM CRIME REPORTS (UCR) MAY NOT REFLECT ACTUAL CRIME RATES. RATHER, A HIGH ROBBERY RATE FOR A CITY REPORTED BY THE UCR MAY ONLY INDICATE A DISPROPORTIONATE GROUP OF PERSONS PREDISPOSED TO SEEK OFFICIAL ASSISTANCE. A MORE ACCURATE MEASURE OF PUBLIC SAFETY INVOLVES USE OF THE NUMBER OF NONCOMMERCIAL ROBBERIES COMMITTED BY STRANGERS EXPRESSED AS A RATIO TO THE TOTAL POPULATION OF VULNERABLE RESIDENTS. COMPARISON OF BALTIMORE POLICE DATA AND RESULTS OF A VICTIMIZATION SURVEY OF 23,467 RESIDENTS INDICATE THAT ROBBERIES INVOLVING THE USE OF A GUN WERE HIGHLY REPORTED TO THE POLICE AND WERE UNIFORMLY REPORTED BY ALL GROUPS WITHIN THE POPULATION. NONCOMMERCIAL ROBBERIES ARE CHARACTERIZED BY THEIR OCCURRENCE IN THE EVENING AND IN PLACES OF PUBLIC ACCESS. MOST WERE COMMITTED BY SMALL GROUPS OF YOUNG BLACK MALES AGAINST LONE VICTIMS WHO WERE PHYSICALLY ASSAULTED IN SOME MANNER. OF THE 13,681 INCIDENTS UNCOVERED, ONLY 7,685 (56.2 PERCENT) HAD BEEN REPORTED. VARIABLES ASSOCIATED WITH REPORTING ARE SERIOUSNESS OF THE CRIME AND SOCIODEMOGRAPHIC CHARACTERISTICS OF THE VICTIM. IMPLICATIONS AND RECOMMENDATIONS ARE DISCUSSED. SPECIFICATIONS FOR ASSIGNING SERIOUSNESS SCORES IN DATA ANALYSIS ARE APPENDED. TABULAR DATA AND BIBLIOGRAPHY ARE PROVIDED. (TWK).