Hide all but Video Show rest of page.

Note: Mac Safari users may have problems loading some videos. If so, use either Google Chrome or Mozilla Firefox to continue.

Jump to

Example 16
Example 17
Example 18
Example 19
Example 20
Example 21
Example 22
Example 23
Example 24
Example 25
Example 26
Example 27
Example 28
Example 29
Example 30

Part 2 Manual

Show Manual

Please note, this is the text from the manual placed here for your convenience. The manual is also a work book and you will want to refer to it for other useful resources.


Counting Stutters is arranged in the same format that you have become familiar with in Part One. Its purpose is to teach you to reliably and accurately count the frequency of stuttering while a person is talking, that is, as an online measurement. Practice begins with low frequency stuttering and advances to moderate and then high frequency stuttering. The frequency of a person's stuttering is obviously an important measure for assessment, treatment, and post-treatment phases of the clinical process. Many clinicians and most stutterers view as a major clinical goal the substantial reduction or complete elimination of moments of stuttering. Therefore, a valid and convenient measure of stuttering frequency is required.

The speech samples of Part Two are all taken from the speech of stuttering persons who vary in age and stuttering severity. These samples also provide a good variety of forms of stuttering behavior. As in Part One, the samples increase in length and difficulty beginning with 1-minute samples of low, medium, and high stuttering frequency and progressing to 2- and 3-minute samples of mixed stuttering frequency.

Before we get started it is necessary to address the complicated issue of the definition of a moment of stuttering. It is difficult to provide exact descriptions/definitions of stutters because they take so many different forms within and across speakers and because the same behavior (e.g., a phrase repetition) might be a moment of stuttering on one occasion and not on another, or for one person but not for another. So, while it is usually not a particularly difficult clinical task to determine whether someone is a stutterer, it is clearly not as simple to determine exactly when a person is stuttering (Curlee, 1981).

Rather than attempt to describe a priori exactly what audible and visible behaviors constitute moments of stuttering, and count only those (Wingate, 1964), our experience has shown that a better way to recognize and count stuttering moments is simply to rely on the perceptual judgment of the listener. Possibly because listeners are also speakers, it appears that listeners have a natural ability to discriminate between the range of disruptions in the flow of speech that are acceptable (generally referred to as normal disfluencies), and those that are perceived to be outside the range of normal, i.e., those that appear to be stuttered. Therefore, it has been found that a good way to identify moments of stuttering is simply to listen to the speech of the person in question and count as stutters those instances of speech flow disruption that exceed one's perceptual threshold of normalcy. While this threshold may not be exactly the same for every listener, research has shown that measures of stuttering frequency based on such perceptual judgments of listeners are at least as reliable as those based on a priori definitions that describe particular categories of behavior as stutters (Curlee, 1981; Martin & Haroldson, 1981). Further, in our clinical experience we have found that even when a priori definitions are provided, clinicians tend to disregard them and rely more on their perceptual judgment than on those definitions, although formulation of such behavior-specific and client-specific descriptions of particular forms of stuttering can be useful to the clinician once stutters have been initially identified perceptually.

For the speech sample recordings that follow, it will be your task to listen to each and indicate, by pressing the right mouse button, whenever you judge a moment of stuttering to have occurred and release the button when you judge the moment of stuttering is finished, using your own internalized, perceptual definition of stuttering. There will be more about this below.

Following is a list of directions explaining how to get started counting stutters. Read through the list before performing any of the operations. Then return to #1 and carry out the directions.

  1. As above, be sure the computer has been set up properly, as in Part One. Add each sample as a new subject according to the sample number and set the run time appropriately (1, 2, or 3 minutes). No other changes to the Run Configuration screen will be needed.
  2. Listen to the sample to be judged for 30 seconds or so, until you are familiar with the speaker's stuttering style and the kinds of stuttering behaviors you are likely to see and hear.
  3. Now you are nearly ready to try your hand (ear) at counting the stutters produced by the first speaker. Whenever you hear what you perceive to be a moment of stuttering, depress the right mouse button for the duration of that moment of stuttering. (For the time being, ignore the left mouse button previously used for counting syllables. During Part Two you will count only stutters.) As discussed above, whether a particular interruption in the flow of speech is or is not counted as a moment of stuttering will be based upon your own perceptual judgment regardingZ whether that behavior or chain of behaviors exceeds your threshold of what sounds normal. Beyond this, a few special rules should be followed.
    • a. Tap the right mouse button one time for each syllable that contains a moment of stuttering, no matter how long the stuttering event is, or how many repetitions it contains, or how many different forms of behavior are included. As long as you perceive the speaker to be making one attempt to produce a certain syllable, one stutter is all that is counted. In the sentences that follow, the underlined sections indicate just a few of the varieties of forms that a moment of stuttering might take. Each underlined section would be counted as one moment of stuttering (i.e., one tap of the right button).
      • Huh-huh- huh-he is coming with us. (This form of stuttering is commonly referred to as a syllable or part-word repetition.)
      • He is coming wwwwwwith us. (This form of stuttering is usually referred to as a sound prolongation.)
      • He is com- he is com-, he is coming with us. (Note here than an entire episode of phrase repetition is counted as one moment of stuttering, because all of it is one attempt to say the syllable com- in the word coming.)
      • He is kuh-kuh-kuh, (THROAT CLEARING) kuh-kuh-kuh, um, well um, he is, he was coming with us. (Note here that a chain of different behaviors is considered one moment of stuttering because it is all one continuous attempt to produce the syllable com- in the word coming.)
      • He is kuh-kuh-kuh-kuh-kuh mmmmmmming with us. (Note here that the speaker produces two different stutters, one on each syllable of the word coming. Therefore, two stutters are counted.)
      • He is k- (silent prolongation of articulatory position) -- PAUSE, SWALLOW. He is kuh-kuh-coming with us. (Note here that the speaker stutters once on the syllable com-, a silent prolongation, sometimes referred to as a hard contact or block, then the speaker completely stops talking and starts afresh, stuttering again on that syllable. Therefore, two stutters are counted because the speaker made two separate stuttered attempts to produce the first syllable of the word coming.
    • b. Remember that you are interested in measuring the frequency of stuttering, not the frequency of all speech interruptions. That is, speakers will also display normal disfluencies and other oddities in their speech (e.g., prolonged speech, monotone), which, while notably different from perfectly normal-sounding, fluent speech, would not be judged to be moments of stuttering and thus should not be counted. (If such differences were highly noticeable and made a speaker's speech sound abnormal, even though they were not reflected in speech rate or stuttering frequency measures, they would be revealed by a measure of speech naturalness (see Part Four).
    • c. Be sure to watch as well as listen to the sample. Some forms of stuttering will be more noticeable visibly than audibly.
  4. When you are ready to count stutters for the upcoming sample, the procedure is essentially the same as that used in Part One, but it is repeated here for good measure. Click the sample. When the sample begins, immediately click on the SMS program to start the SMS timer and begin counting stutters by depressing the stutter mouse button once for each perceived stuttering event. Samples 16-24 are one minute in duration and the SMS will automatically stop accepting stutter counts at the end of one minute (because you will have set the run time for 60 seconds).
  5. At the end of the sample the SMS will automatically display the total stutter counts on the screen in the box labeled STUT. (Because every stutter counted also registers as a syllable spoken, the SYL box will contain the same number during this exercise. However, because non-stuttered syllables are not being counted during this activity, the computer cannot correctly calculate %SS, which will appear as 100% during this exercise. Ignore these numbers for now.
  6. Turn to the section of the workbook that contains the summary data for Part Two (at the end of these Part Two instructions). There you will find a data sheet to be used for recording your data for the sample. On the part of the data sheet labeled Step 1, Sample 16 (this is found in the pdf manual/work book), record the number of stutters you counted in the 1 column (first attempt). Then compare your number to the Target Range of acceptable counts shown on the data sheet for that sample. To determine the standard for the number of stutters that occurred in each sample, each of five experienced listeners independently marked perceived stutters on a transcript of each speech sample.3 Wherever at least four of the five listeners agreed, a moment of stuttering was marked. The Target Ranges noted on the data sheets represent the range from 5% below to 5% above the number of stutters upon which the listeners agreed for each sample.
  7. If your count is not within the acceptable range, follow the procedures described previously for repeating syllable counts:
    • Click the sample in order to play/count it again. On the SMS screen in the box labeled Continue, click on Make another run with the same setting and click OK. Score the sample again and record the number of syllables you counted in the 2 column of the data sheet (second attempt). Once again compare your number to the acceptable range. If your count is still not in that range, repeat the entire process a third time and record your completed syllable count in the 3 column on the data sheet, once again comparing your data with the Target Range.
    • If, after three attempts, your count is still not within the acceptable range, click the sample once again. Then, find the transcript of that sample in the workbook (following the data summary page). Replay the sample and recount the stutters while reading along on the transcript. Each stuttered syllable is underlined. Note that each continuous underlining indicates one moment of stuttering, i.e., one continuous depression of the right mouse button. You may want to stop periodically to see that your count is closely matching that shown on the transcript or to try to figure out the source of differences between your perceptions and the stutter judgments indicated on the transcript. Record your total count of stutters in column 4 of the data sheet.
    • After you have worked through the transcript, click the sample once again and recount the stutters, without benefit of the transcript. Record the data in column 5 of the data sheet. Your count should be within the acceptable range by now.
  8. When your count is within the Target Range of acceptability, you are ready to go on to the next sample. As you have done previously, add a new subject according to the sample number and begin counting stutters when the sample begins playing. When each sample is completed, record and compare your data on the summary data sheet. Repeat the sample as instructed above whenever your counts are not within the specified range. Continue counting stutters in this fashion for Steps 1 through 5, Samples 16 through 30, then continue on in the workbook for instructions regarding Part Three. NOTE: Speech samples in Step 4 (Samples 25-27) are two minutes long; speech samples in Step 5 (Samples 28-30) are 3 minutes long. Remember to change the run length accordingly on the Run Configuration screen when you prepare the setup for these samples. In order to do this, at the completion of Samples 24 and 27, click on Make another run with new settings, which will produce the Run Configuration screen so that you can change the run time for the next set of samples.