Speech-Language Pathology/Stuttering/Measurement of Stuttering

Measuring stuttering is notoriously difficult. You can measure several aspects of stuttering:

• Frequency of disfluencies. I.e., disfluencies per hundred words or syllables. The "average" stutterer is dysfluent on 10 percent of words. People who don't stutter are disfluent on about 2% of words.
• Duration of disfluencies. The "average" disfluency lasts about 1 second.
• Speaking rate, or word per minute. The average speaking rate for people who don't stutter is 167 words per minute (Darley, 1940). The average speaking rate for stutterers is 123 words per minute. The average stutterer speaks about 25% slower than the average non-stutterer.
• Types of disfluencies. Wendell Johnson developed eight categories of disfluencies in 1959. The first three are common to both stutterers and non-stutterers. The last five are stuttering disfluencies.
1. Interjections ("uh," "er").
2. Revisions ("I was - I am going").
3. Incomplete phrases.
4. Part-word repetitions ("ba-ba-ba-baseball").
5. Word repetitions.
6. Phrase repetitions ("I was I was going").
7. Broken words ("I was g - (pause) - oing home").
8. Prolonged sounds.
• Psychological effects.

The preferred test questionnaire for measuring stuttering is the "Stuttering Severity Instrument for Children and Adults," 3rd edition, by Glyndon Riley.

A study by Young in 1961 found that ordinary listeners judged stuttering to be "severe" when there were part-word repetitions, prolongations, broken words, and a new category, "undue stress or tension." But many speech samples were judged "severe" with relatively few of these disfluencies, suggesting that there are yet more, unidentified types of stuttering disfluencies.

There are many electronic instruments that can measure aspects of stuttering. See "Instrumentation For The Assessment And Treatment Of Stuttering", by Klaas Bakker (1996) and "Automatic Recognition of Dysfluencies" (Howell, 1997).

Problems Measuring Stuttering

1. Differences between judges. Different people listening to the same tape will hear different numbers of disfluencies, even if they are trained speech pathologists.
2. Counting normal disfluencies. Everyone has normal disfluencies. Does having lots of normal disfluencies make a person a stutterer? Or is stuttering something different? Should normal disfluencies be counted as stuttering?
3. Children vs. adults. Children stutter in different ways than adults: some disfluencies are normal for children but are abnormal in adults.
4. Hidden stuttering. Some stutterers can substitute words and appear to never stutter.
5. Speaking conditions. Some stutterers can read out loud perfectly fluently. Others stutter most severely when reading. Some stutterers are most fluent in conversation. Others are most disfluent. Some are most fluent in stressful situations, on telephones, etc., while others are the opposite. You can't judge how severely a person stutters from a single speaking situation.
6. Internal or psychological stuttering is difficult to quantify. For many stutterers the internal effects of stuttering are greater than their physical symptoms. Several stuttering therapies focus on improving the self-esteem and overall communication skills of stutterers, with little or no attention to improving physical fluency.

Different people stutter in different ways, and each person has multiple ways of stuttering. There is no single behavior common to all stuttering. Be skeptical when anyone claims that a therapy reduced stuttering 85%, etc. Ask how stuttering was measured, under what conditions, etc.