In our method, a child is trained to perform a play-based task when a signal is heard. An experimenter—called an observer—watches the child’s behavior to determine when the signal was played. In contrast to clinical testing methods, the signal is randomly placed in one of two intervals for each trial by the computer program.