THE NCS IS A NATIONWIDE PROBABILITY SAMPLING OF HOUSEHOLDS. STUDY DATA WERE TAKEN FROM THE CENTRAL CITIES SAMPLE OF THE NCS WHICH PROVIDES DATA FROM 26 MAJOR CITIES. DATA WERE COMPARED FOR THE VIOLENT CRIMES OF RAPE, AGGRAVATED ASSAULT, AND ROBBERY AND THE PROPERTY OFFENSES OF LARCENY, BURGLARY, AND MOTOR VEHICLE THEFT FROM BOTH DATA SOURCES. AS EXPECTED, VICTIM-REPORTED CRIME RATES EXCEEDED OFFICIAL REPORTS FOR EVERY CRIME CATEGORY EXCEPT MOTOR VEHICLE THEFT. ALL DIFFERENCES NOTED BETWEEN THE TWO CRIME RATES WERE SIGNIFICANT. FURTHER, CORRELATIONS BETWEEN THE TWO SETS OF STATISTICS SHOWED MODERATELY STRONG POSITIVE CORRELATIONS, WITH THE EXCEPTION OF AGGRAVATED ASSAULT. THIS SIMILARITY IN DISTRIBUTIONAL PATTERNS BETWEEN THE TWO SOURCES SUGGESTS THAT DESPITE THE SUBSTANTIAL DIFFERENCES IN ABSOLUTE AMOUNT OF CRIME REVEALED BY EACH DATA SOURCE, UCR CRIME RATES DO HAVE SOME APPLICABILITY AS ROUGH INDICATORS OF ACTUAL CRIME RATES. IT IS CONCLUDED THAT DESPITE THE MINOR PROBLEMS INVOLVED IN COMPARING OFFICIAL WITH SURVEY CRIME RATES, SUCH A COMPARISON IS USEFUL IN THAT IT CASTS SOME DOUBT ON THE ACCURACY OF OFFICIAL STATISTICS. REFERENCES ARE PROVIDED. (JAP)