Performance measures depend on the clarity and consistency of their metrics. The standard unit of analysis of court proceedings is the “case.” Yet the definition of a case varies widely. It may include multiple accused persons and charges, a count of accused persons, the number of informations (a set of charges), or the number of charges. Likewise, whether new cases, completed cases, or the number of cases heard in court are counted make a tremendous difference. In theory, many dozens of caseload measures are possible. Here, a matrix of twelve measures is proposed which balances theoretical exhaustion with practicality. Using data for one year of adult criminal cases in the Provincial Court of British Columbia it is shown that the choice of measure introduces variability of workload over 400 per cent. The matrix is intended as a heuristic device for policy-makers to scrutinise case-based indicators.
