The Assessment of Basic Learning Abilities (ABLA) test has proven to be a useful assessment and training tool for staff responsible for training persons with developmental disabilities. The ABLA test assesses a person's ability to learn six tasks that are hierarchically ordered in level of difficulty. Two auditory matching tasks, a prototype visual-auditory nonidentity matching (VANM) task and a prototype auditory-auditory nonidentity matching (AANM) task have been demonstrated to be more difficult than ABLA level 6, and are hierarchically ordered in difficulty in relationship to each other. As a preliminary step to adding the VANM and AANM tasks to the ABLA test, Experiment I assessed the predictive validity of the VANM and AANM prototype tasks, as well as a third task, auditory-auditory identity matching (AAIM). A Chi-square analysis showed that the prototype VANM, AAIM, and AANM tasks demonstrate predictive validity. Experiment 2 examined where echoics, tacts, and mands fit in level of difficulty in relation to the ABLA tasks and the auditory matching tasks. Using order analysis, echoics, tacts and mands were all found to be more difficult than ABLA levels 3, 4, and 6. Pairwise comparisons found echoics to be less difficult than tacts and mands. Limitations of the experiments and areas of future research are discussed.