The purpose of the investigation was to obtain some relationships among (1) test lengths, (2) shape of domain-score distributions, (3) advancement scores, and (4) several criterion-referenced test score reliability and validity indices. The study was conducted using computer simulation methods. The values of variables under study were set to be typical of those often used or obtained in practice. The reliability and validity indices (decision consistency, kappa, decision accuracy, predictive validity, and efficiency) are among the most useful indices for criterion-referenced test developers and evaluators. Practical guidelines are offered for using results obtained from the investigation to determine suitable lengths of criterion-referenced tests in specific assessment situations. (Author)