Thirty listeners used the 9-point speech naturalness rating scale described by Martin, Haroldson, and Triden (1984) to score 1-minute spontaneous speaking samples from 15 normally fluent speakers and 15 stutterers who had completed the initial phases of a prolonged speech treatment program. The same listeners were later asked to judge whether each sample was from a stutterer or from a normal speaker. Reliable listeners identified almost identical numbers of samples from both speaker groups as "normal speakers," but the stutterers received significantly higher scores on the rating scale, indicating that their speech was judged more unnatural sounding. The speech samples incorporated those used in a related study by Ingham and Packman (1978), thus permitting a reevaluation of the findings of that study.

Three young children were taught correct articulation of an error phoneme on a program administered by their mothers in their homes. During the course of the treatment program each child was periodically brought into the speech clinic where the generalization of correct articulation learned in the home treatment setting could be measured in a series of four settings which differed from the treatment setting in varying degrees. Stimulus words which had not been taught during the home articulation program were also included in generalization testing. The results indicated (1) that correct articulation generalized to a high extent to the nontreatment settings investigated; and (2) that responses to new stimulus words were less adequate than responses to words which had received previous training. The clinical implications of these findings are discussed.

No abstract

The purpose of this study was to determine whether different topographies of disfluent behavior form a response class. A within-subject, repeated reversals (ABAB) design was used to analyze the stuttering behavior of three adult stuttering speakers. A single type of stuttering behavior was punished for each subject while frequencies of occurrence of other types were concurrently measured. The results showed that: (1) stuttering behaviors displayed direct behavioral covariation for all subjects, illustrating the existence of a response class; and (2) the response classes observed included both kernel and accessory features of stuttering. The results are discussed in terms of the literature on response classes and two-factor learning theory of stuttering with special emphasis on the implications of these results for our understanding of the development of stuttering.

A successful application of time-out procedures is illustrated by three case studies with high school and college-age male stutterers. The method of administering the time-out procedure is described in detail, and typical results are illustrated by data collected from clinical sessions. Modifications of the basic procedure are tested and their results discussed.