Since the 70s, amazing successes of behavioral medicine stimulated development of investigation in the area of stuttering behavioral therapy (Azrin et al, 1980; Jones et al, 1997; Wagaman et al, 1993). During 3 months, we carried our research on 60 stuttering children from 2 to 15 years old, suffered by a light and average stuttering. The selection criterion was the ability to repeat short sentences (up to 4 words) from second or third attempt without stuttering. The participants were divided into 4 groups.

1st group (n=10). An absence of reinforcement. Parents were keeping the speech of children under observation during the 3 months without any reactions and negative repercussions.

2nd group (n=10). Verbal positive and negative reinforcement. Parents marked fluent speech by praise and in the cases of stuttering they constituted a fact of stuttering by oral remarks.

3rd group (n=12). Negative reinforcement. During the 3 months, children were not praised for good speech, but after stuttering manifestation they had to repeat the sentence (1-4 words) several times, as long as they pronounced it fluently.

4th group (n=8). Probabilistic reinforcement. Originally, children were included in the 3rd group, but in accordance with different circumstances, (the lack of parental attention, child caprice or unwillingness to repeat) they got "punishment" in the form of repetition in 50%-80% of all stuttering cases.

The speech of participants was tested in the beginning and at the end of 1st, 2nd, 3rd and 6th months of the experiment. The testing consisted of quantification of stuttering units in certain speech fragments. It was revealed that:

1. An absence of reinforcement does not influence speech. After the 3 months, some children demonstrated statistically insignificant speech worsening because of endured somatic diseases or stresses.
2. Verbal positive and negative reinforcement is ineffective.Verbal positive and negative reinforcement is ineffective, at the expiration of 3 months; the speech of patients was still kept approximately on the same level.
3. The negative reinforcement leads to a gradual facilitation of speech fluency at pre-school children since 2-3 days. The school age children's speech did not change within 3-5 days, and then, usually, abrupt improvement began. Later, 5 younger children which corrected 90%-100% of all fluency errors stopped to stutter towards the end of the first month. Further negative reinforcement with the rest of the participants enabled the elimination of stuttering within 1-4 months from the beginning of the experiment. Obtained results remained till the last testing after 6 months.
4. Probabilistic reinforcement did not lead to significant alterations in speech.

Thus, absence of reinforcement, verbal positive and negative, and probabilistic reinforcement did not influence the stuttering development. Only negative reinforcement (punishment in the form of repetition) represents a successful methodological approach to the elimination of stuttering. That is why we recommend a negative reinforcement as an effective method of behavioral therapy for children with a light and average stuttering.