constitute stuttering must be clearly defined and adhered to beforehand (see below). Second, this requires the physical manipulation of a stopwatch and possibly two event counters simultaneously. All of this has to be done whilst paying attention to the client’s speech. This can be tricky, particularly if the speech sample is taken from conversation, when turn taking can lead to lots of clicking on and off with the stopwatch. Once the speech sample data have been collected, the clinician uses the syllable count and stuttering count data
to calculate rate of speech and the percentage of stuttering. The ability to rate speech samples accurately is a skill and can only be achieved through prac¬tice. Being able to rate samples accurately “online” and without later recourse to audio or video feedback is particularly difficult. Ideally, a number of speech samples should be taken across a range of speaking situations. Commonly, these include conversation and monologue, and with older clients oral reading (where word avoidance cannot be hidden less easily) and telephone conversations.
What to count and what not to count
This can be awkward, both in terms of the units to be counted (syllables or words) and what is regarded as normal as opposed to stuttered. For example, some clinicians include all interjections and phrase revisions as stuttering in their fluency count as normal, whilst others may not. To complicate the matter further, it could well be that within the same speech sample the inter¬jection might be seen as normal disfluency. “Oh … 26 multiplied by 20 is … er … 520”, or stuttered as in “my name is …. er … John”. Similarly, “he needs the will to live . . . the will to succeed” may sound plausible as attempt¬ing to repair a genuine slip of the tongue, whereas “I want to go to the … I want to watch a film” is more likely to be connected to word avoidance. For the clinician, the most important thing is to state explicitly the criteria on which the judgements of fluency and stuttering are made, and at the least to make a note of any behaviour which is considered unusual, even if the clin¬ician is unsure as to whether it should be considered as genuine stuttering. Data then needs to