The purpose of this study was to assess whether stuttered speech could be simulated by normal-speaking individuals and to identify significant parameters used in the simulation process. The subjects were 24 normal-speaking graduate students in speech-language pathology. The subjects were assigned the task of simulating stuttered speech given specific levels of severity. Then they were asked to listen to the simulated samples and identify the number and types of dysfluencies presented. Analysis showed agreement among subjects on the number of dysfluencies, with the exception of samples simulating severe stuttering behavior. Judging the number of dysfluencies in the simulated samples of severe stuttering appeared to be a more difficult task. The types of dysfluencies noted by the listeners were restricted primarily to repetitions and prolongations. Simulated stuttered speech was therefore judged to be composed of fewer types of stuttering behavior than actual stuttered speech is reported to be.