A predictive adaptive testing (PAT) strategy was developed based on statistical predictive analysis, and its feasibility was studied by comparing PAT performance to those of the Flexilevel, Bayesian modal, and expected a posteriori (EAP) strategies in a simulated environment. The proposed adaptive test is based on the idea of using item difficulty and past information (observed data) about an examinee to acquire the probability of answering further items correctly. Development of the PAT model is described with reference to: (1) initial items; (2) scoring method; (3) selection of subsequent items to be administered; and (4) terminating criteria. The model was compared to the Flexilevel, Bayesian modal, and EAP strategies in a Monte Carlo simulation study in which the ability levels of 999 examinees were generated using a 71-item test. The strategies performed similarly at the low ability level. At the medium level, the Bayesian modal and EAP strategies were the most efficient. At the high level, the Bayesian modal strategy required fewer items than did the PAT and the EAP strategies. The three strategies produced similar results in terms of error variance and ability estimates. The PAT is potentially useful, particularly in small classroom testing. There are 12 tables of study data, 2 figures, and a 14-item list of references. (SLD)