When I assess variability and consistency I use specific operational definitions to define each behavior. The following is related to the perceptual analysis level (basically what our ear tells us is happening). Variability refers to the tendency to make different errors during multiple productions of a given sound in the exact context (same word or phrase, same position in the word). For example, suppose a child was to repeat the word “pen” multiple times and on three consecutive trails said “fen”, then “shen”, then “cren.” Since the error on the initial “p” sound varied from time to time in the same place in the word and in the same context this would be considered a variable production. Consistency refers to the tendency to produce errors in the same place on repeated trials in the same context. For example, if the target production was “pen” and the child said “fen,” “pef,” and then “pof” these productions would be considered inconsistent since the errors changed positions on repeated trials. In the first example above, the child would be considered to have variable productions as noted, but consistent since all of the errors occurred in the initial position of the words.

A critical issue is to consider whether children with apraxia of speech (CAS) are variable and inconsistent. In my experience I believe that children with CAS are consistent in that they tend to produce errors in the same place on repeated trials. However, there is debate as to whether they are variable. My view is that they are invariant in the type of error they produce. There are multiple ways that speech-language pathologists classify errors.

Errors can be called:

• substitutions (one sound is substituted for another as in the examples above)
• additions (a sound is added such as spulash for splash)
• distortions (sounds are produced but are not quite accurate such as too much air during the production of s)
• and omissions (sounds are left out such as pe for pen).

Children with apraxia mostly produce distortions or substituted distortions. Thus the error type in CAS is relatively consistent (a distortion). But how the sound is distorted is variable from trial to trial. Please note that other error types occur, but the predominant error type is the distortion.

Finally, it is important to consider that the underlying movement patterns or acoustic patterns (the speech waveform) may be variable (and data suggest it is in apraxia) even if perceptually the childs production sounds accurate. For example, we have found that even when sounds are judged to be accurately produced in apraxia of speech, the underlying movements are more variable than normal as are the underlying acoustic patterns. Indeed, we have found that speech intelligibility in apraxia can be predicted reasonably well by the amount of variability in the underlying acoustic or movement pattern for sounds that are perceptually judged as accurate. Our view is that the amount of underlying variability may be a index of the stability of the speech motor system the more variability, the less stable, the more severe the apraxia or the more guarded the prognosis.

(Donald A. Robin, Ph.D., BC-NCD, is a Professor, Department of Communicative Disorders, at San Diego State University and the Joint Doctoral Program in Language and Communicative Disorders, San Diego State University/University of California, San Diego. Dr. Robin has worked with and studied children with childhood apraxia (motor speech disorders) for over 25 years. He has published over 70 papers, co-authored textbooks, and presented workshops on motor speech disorders in adults and children. Dr. Robin is chairperson of the Writing Committee on Childhood Apraxia of Speech for the Academy of Neurogenic Communication Disorders and Sciences (ANCDS). He has recently retired as the Editor of the American Journal of Speech-Language Pathology. He is a coauthor of the popular text on childhood apraxia, Developmental Apraxia of Speech: Theory to Clinical Practice.)