TWO TYPES OF MEMORY BIAS OVER TIME -- FORGETTING AND FORWARD TELESCOPING (PROCHRONISTIC DISTORTION OF TIME) -- ARE KNOWN TO DISTORT CRIME STATISTICS WHEN VICTIMS ARE INTERVIEWED LONG AFTER THE EVENT. BOUNDED INTERVIEWING IS DESIGNED TO MINIMIZE TELESCOPING, BUT ITS EXPENSE PROHIBITS USE BY MOST STATE AND LOCAL CRIMINAL JUSTICE AGENCIES. THE IDEAL SITUATION WOULD BE TO DEVELOP A MODEL OF MEMORY DECAY OR BIAS HAVING ADJUSTMENT FACTORS ALLOWING 'CORRECTION' OF VIRTUALLY ALL VICTIMIZATION SURVEY PROCEDURES. THE AUTHORS ANALYZED BURGLARY VICTIMIZATION DATA FROM 18 CITIES AS COLLECTED BY THE LEAA SURVEYING EFFORT, AND FOCUSED ON MEMORY BIAS AMONG DIFFERENT CITIES ON THE POSSIBILITY OF CONSTRUCTING A GENERALLY APPLICABLE MODEL IF SUFFICIENT SIMILARITY AMONG CITIES WERE FOUND, AND ON PATTERNS OF BIAS FOR REPORTED VERSUS UNREPORTED CRIMES. THE STUDY INCLUDED ONLY INCIDENTS COMMITTED AGAINST CITY RESIDENTS (NOT TOURISTS OR COMMUTERS), AND ONLY A 10-MONTH PERIOD OUT OF ONE YEAR WAS ANALYZED DUE TO DATA OMISSIONS. THE LEAA VICTIM SURVEY DATA WERE COMPARED WITH OFFICIAL BURGLARY DATA (UNIFORM CRIME REPORTS) FOR A GIVEN MONTH TO DERIVE PATTERNS OF OVER- OR UNDERRECALL OF BURGLARIES DURING EACH MONTH. IT WAS NOTED THAT VICTIMS PLACED CRIMES IN SUMMER MONTHS IN THOSE CITIES WHERE SUMMER WAS A DISTINCTIVE SEASON, BUT THIS 'SEASONAL CITIES HYPOTHESIS' NEEDS FURTHER STUDY AND REPLICATION TO BE VALID. THE DATA INDICATE THAT REPORTED AND UNREPORTED INCIDENTS ARE TELESCOPED OR FORGOTTEN IN ABOUT THE SAME WAY, BUT UNREPORTED INCIDENTS HAVE A SOMEWHAT MORE EXTREME PATTERN OF BIAS THAN REPORTED ONES. THE LIKELIHOOD OF DEVELOPING A GENERALLY APPLICABLE FACTOR FOR CORRECTING VICTIM SURVEY DATA IS VERY LOW. HOWEVER, THE EVIDENCE PRESENTED SUPPORTS THE CONTENTION THAT MEMORY BIAS HAS A GREATER IMPACT ON UNREPORTED CRIMES AND THAT SEASONAL INFLUENCES ON MEMORY AFFECT THE UNREPORTED EVENTS MORE THAN THE REPORTED ONES. A SURVEY SCHEME FOR IMPROVED VICTIMIZATION ESTIMATES IS PROPOSED, MODELED AFTER ONE DEVELOPED BY ANOTHER RESEARHER THAT HAS PROVEN EFFECTIVE IN COLLECTING AND CORRECTING VITAL EVENT DATA OBTAINED FROM SURVEYS. (DJM)