An experiment was performed with computer-generated data to investigate some of the operational characteristics of tailored testing as they are related to various provisions of the computer program and item pool. With respect to the computer program, two characteristics were varied: the size of the step of increase or decrease in item difficulty for successive items, and the range in difficulty levels within which items might be considered acceptably close to a specified level. With respect to item pools, the two characteristics were varied: the number of items in the pool, and the shape of the item difficulty distribution. Simulated test data were generated by computer for various values of the four parameters (step size, acceptance range, number of items, and item difficulty distribution) and for various hypothetical ability levels from plus three to minus three. The resulting expected values and standard errors were tabulated and are presented as a guide for those involved in setting up tailored testing procedures. (Author/CTM)