A comparison of self-report of criminal involvement and official arrest records

Self-reports of arrests and official arrest records were compared for 250 m ale and 80 female participants in the ongoing Hyperactivity Follow-Up Study of Young Adults at University of California at Berkeley, For males, kappa coefficients indicated good statistical agreement between arrest records an d self-report measures for 7 of the 12 types of crimes. Statistical agreeme nt, however, may not be the most meaningful yardstick to judge concordance. Additional analyses reveal that for almost all of the crimes examined, one third or more of those with an arrest record for the crime failed to revea l this information on the self-report measure. Arrests for both some high-f requency, less serious crimes and some low-frequency, more serious crimes w ere not revealed in the subjects' self-reports. An examination of the uniqu e information gained from self-reports found that the subjects most often r eported committing public disorder crimes for which they were not caught (o ver 30%) and were least likely to self-report crimes against people (less t han 10%). Furthermore, subjects with more convictions were more likely to a ccurately self-report their criminal involvement. Since both self-report an d official records each contribute unique information and reflect different sources of error, it is proposed that a combination of both types of recor ds is the most inclusive indicator of criminal activity in adulthood. (C) 2 001 Wiley-Liss, Inc.