By 1983, the project had conducted four statewide sample survey audits of records as well as one telephone survey of recordkeeping practices. The audits verified accurate data and indicated some of the reasons behind the inconsistencies and inaccuracies that had been a problem for years. The audit revealed that providing statistics as a byproduct of automated operational systems does not in itself improve data quality. Only where uniformity with State requirements was explicitly built into the local system was data quality improved by automation. While filing information has been reliable and timely, dates of events have been accurate, and criminal caseload information has become more consistent, a wide range of error levels and year-to-year shifts by individual courts have characterized the termination statistics. In addition to uniform records, regular training and flexible technology are needed. Four notes, two tables, and one figure are supplied.