Analyses of 212 reports of crime incidents from the 1974 Portland, Oregon, victimization survey were conducted. These were matched with official crime reports of the same incidents. The record check described and analyzed the following: differences between survey and police data; the frequency of "don't know" responses; patterns of telescoping and their relation to victim characteristics; survey incidents not found in police data; and implications regarding the utility of survey and police data and methods used to collect victimization survey information. Conclusions include: (1) Information obtained through victimization surveys is sufficiently similar to that reported to police so that most crimes are classified the same way by the two sources of data; (2) The survey data contained higher estimates of the dollar loss from the crime; (3) The reliability of validity of the survey data depended upon the type of information considered; (4) For most of the types of information considered, accuracy or completeness did not decline as a function of the time lag between occurrence and interview; (5) The age of the victim was not related to the amount or type of error in the data. (Author)