Computerized Classification Testing In More Than Two Categories Using Stochastic Curtailment

The truncated sequential probability ratio test (TSPRT) has been widely studied as a decision algorithm in COT for 2 or more categories. Finkelman (2003) added an algorithm to the TSPRT in the form of stochastic  curtailment to classify an examinee in an even earlier stage of testing. This stochastically curtailed SPRT (SCSPRT) halts testing when a change of classification is possible, but unlikely. Finkelman {2003) adapted the algorithm for 2 categories and with fixed item ordering.
The present study replicated his results, then replicated them in realistic settings, and subsequently  generalized the SCSPRT to 3 categories while using adaptive item selection. The results show increased efficiency when using 1 and 2 cut points. Different item selection methods are discussed.