IT WAS HYPOTHESIZED THAT THREE VARIABLES DETERMINE FEAR OF CRIME AND CONSEQUENT BEHAVIOR CHANGES: (1) THE TYPE OF CRIME (PERSONAL VERSUS PROPERTY), (2) THE LOCATION OF THE CRIME (IN ONE'S OWN SURROUNDINGS VERSUS ELSEWHERE IN THE COMMUNITY), AND (3) THE FREQUENCY WITH WHICH THE CRIME OCCURS (RECURRING CRIMES VERSUS AN ISOLATED INCIDENT). THE FIRST VARIABLE IS A QUALITATIVE ONE. THE OTHERS ARE USED BY PEOPLE TO ASSESS THE PROBABILITY THAT THEY THEMSELVES WILL BE VICTIMIZED. TO TEST THE HYPOTHESES, TWO SAMPLES OF WOMEN--128 STUDENTS OF A PENNSYLVANIA COLLEGE--WERE INTERVIEWED IN THEIR RESIDENCES. THE SUBJECTS READ FICTITIOUS CRIME STORIES, WHICH THEY WERE TOLD WERE ACCOUNTS OF ACTUAL CRIMES THAT HAD NOT BEEN REPORTED BY THE NEWS MEDIA. AFTER READING THE STORIES, THE SUBJECTS COMPLETED SCALES DESIGNED TO MEASURE THEIR EMOTIONAL RESPONSE AND POTENTIAL BEHAVIORAL RESPONSE (PLANS TO TAKE PRECAUTIONS). THE CRIME OF PHYSICAL ASSAULT PRODUCED GREATER EMOTIONAL AND POTENTIAL BEHAVIORAL RESPONSE THAN DID THE CRIME OF BURGLARY. RECURRING CRIMES CAUSED PEOPLE TO CONSIDER TAKING PRECAUTIONS MORE OFTEN THAN DID ISOLATED CRIMES. THERE WAS SOME EVIDENCE THAT CRIMES OCCURRING IN THE SUBJECTS' OWN IMMEDIATE SURROUNDINGS GENERATED GREATER CONCERN THAN CRIMES OCCURRING ELSEWHERE IN THE COMMUNITY. SUPPORTING DATA AND A LIST OF REFERENCES ARE INCLUDED. (LKM)