The discussion in this article is directed at a brief review of the basic components of single-subject experimental designs. It sets out to present some of the technical and practical advantages in developing single-subject designs to evaluate potential treatment variables and treatments. The point is made that the designs are specifically structured for applied, intervention research. In addition to explaining how the basic AB components function to allow for a controlled evaluation, a short review is included of other factors important to the scientific method such as operational definitions, reliability, repeated measures, and internal and external validity. The article serves as a foundation for the following two articles that are directed at demonstrating the flexibility of single-subject experimental studies.

Investigators have proposed that children with functional articulation disorders should be relabelled phonologically disordered. To support this proposal, evidence has been presented in the literature demonstrating that children's error patterns reflect the operation of phonological processes. No quantitative or qualitative criteria have been offered to differentiate these processes from surface error patterns. The purpose of the present descriptive study was to determine if differences would be found when two kinds of process analyses were employed: a nonquantitative criteria analysis as conducted in the studies reported in the literature, and a quantitative criteria analysis. Speech samples were obtained from 13 children with functional articulation problems. Their errors were submitted to the two analysis procedures. Results indicated that the number of identified processes were reduced when minimum quantitative criteria were used from the number identified when no quantitative criteria were imposed. The decrease occurred in individual children's patterns as well as across the patterns of the 13 children. It is suggested that there is a need to establish reasonable quantitative and qualitative criteria for phonological process identification.

Articulation errors of several children with severe articulation problems were analyzed for distinctive feature errors. On the basis of the analysis, children were administered training programs to teach production of features lacking in their repertoires. The study was concerned with generalization of trained features across phonemes in which the feature had been in error. During training, probes on selected phonemes were administered to test for feature generalization. Results for three children are presented to represent generalization from training on three different features. During training, feature errors deceased by 69% to 84%, indicating that a feature generalizes across several phonemes although training in only one phoneme is administered. The degree of generalization varied across phonemes and across phonemes in different positions in words. The study indicated that features have generality and that children’s feature errors are consistent and systematic. For clinical purposes, the results suggest that detailed analyses will aid in selecting items for training that may increase the efficiency of articulation training.

Five children who produced /θ/ for /s/ substitutions as a Disarticulation were trained to produce /s/ correctly in three syllables. Untrained exemplars of syllables and words were tested throughout baseline and training. The 60 probe items contained both spontaneous and imitated words and syllables combined with high, low, front, back vowels, and consonants. A functional analysis reversal design was used, and the generalization patterns were analyzed. The effect of context was found to be less influential than expected while other factors such as stimulability, amount of training, and subject characteristics appeared as important variables in generalization.

The articulation of 10 children with severe misarticulations was subjected to a feature analysis. The 13 distinctive features of English phonology as proposed by Jakobson, Fant, and Halle (1952) and Chomsky and Halle (1968) were used for the study. Phonetic transcriptions of responses on the McDonald Deep Test of Articulation formed the basis for the analysis. Two sets of data were compiled: the children’s feature systems in comparison to the English system and a traditional articulation evaluation of phoneme articulation. Results indicated that children’s feature errors were consistent across phonemes which contained the feature. It was further determined that misarticulations can be only partially described as a function of absence of features. Many of the errors occurred in the way features were used in particular combinations or contexts by the children. Errors resulted when children applied rules for feature usage which were different from the phonological rules in English. It is suggested that a distinctive feature analysis may offer a more efficient approach to articulation training.