I’ve released the Stuttering Analysis App on the Windows App Store free of charge as a way to give back to my profession. No strings, it’s free. It took me half a year to write and hopefully you’ll find it useful!
Stuttering Analysis App Reference Manual
The Speech Squirrel Stuttering Analysis App is © 2013 Seth Koster, M.S. CCC-SLP.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDER “AS IS” AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL SETH KOSTER, SPEECH SQUIRREL SOFTWARE OR WWW.SPEECHSQUIRREL.COM BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
Dysfluency must be assessed in an appropriate environment by a qualified Speech-Language Pathologist, as such this software may only be used by a Speech-Language Pathologist or delegate (CFY, graduate clinician, research assistant, etc.) under appropriate supervision as specified by ASHA’s Code of Ethics and Scope of Practice. No assessment tool is a replacement for informed clinical judgment, only use this software if you agree to look critically at any results and use your own judgment as to whether they’re valid for your particular patient/client/student/subject.
Table of Contents
- General Use
- Disfluency Frequency in Syllables
- Disfluency Frequency in Words
- Speech Rate
- Naturalness Rating
- Effort Rating
- Awareness Rating
- Secondary Behaviors
- Other Sections
- Output File
- Frequency Breakdown
- Duration Breakdown
- Normative Data Referenced
This work is, first and foremost, dedicated to my mother.
She is the finest clinician I’ve ever met and the person who instilled in me a love of the profession.
My highest professional aspiration is to become a clinician worthy of carrying on her legacy.
I also wish to express my deep appreciation to the current and historical experts in fluency.
Without their contributions to our field I would not have the understanding or normative data necessary to create this application.
It is with great professional respect with which I present ideas and analysis based on their research.
The Stuttering Analysis App is designed to combine several commonly used dysfluency measures into one easy to use app. Strongly influenced by the Real-Time Analysis of Speech Fluency (Yaruss, 1998), this application is designed to be used in real time while either interacting with a patient/student/client/subject (hereafter referred to as subject) or analyzing video or audio recorded speech.
Spending some time reading and understanding the Real-Time Analysis of Speech Fluency, particularly the discussion of acquiring a representative sample, will be extremely useful in determing how best to use this application. While researchers may prefer to opt for as much accuracy as possible at the expense of time, diagnosing clinicians may be better served by capturing enough data to appropriately assess while not increasing data granularity to levels which will not serve the assessment needs.
As a clinician I don’t have a large fluency caseload, nor an excess of available time for evaluations, so it was important to me to write an application which allowed me to capture accurate information within a reasonable timeframe even though I might not use the tool on a regular basis. It is my hope that you will also be well served by this set of requirements.
Time of assessment: Maximum of 17.5 minutes of subject speech.
Using a tablet or touchscreen computer (for app versions which can run on non-touchscreen computers, such as the Windows Store version) is highly recommended. The wise clinician will practice using the application on recorded speech prior to attempting a real-time evaluation. Particularly the Frequency of Disfluent Speech in Syllables phase of the assessment may be intensive since so many different types of disfluencies must be kept track of if one wishes to get complete data.
As stated, this application is designed to be used in real time, as such as much as possible has been automated. This means, though, that measures are interrelated. For example if you do not accurately complete the Frequency of Disfluent Speech in Syllables phase, including stopping and starting the timer appropriately, you will not get an accurate Syllables Per Minute Speech Rate. In some instances you may find that you’re presented with more data than you need, that is by design. My design document stated that I would make available as much information as possible and allow the user to disregard what they do not need. By including all data and analysis my hope is to meet the needs of both diagnosing clinicians as well as researchers.
Disfluency Frequency in Syllables
Duration of phase: 300 syllables or 10 minutes, whichever comes first.
Important Note: Syllable count is only increased when you click the ‘Syllable’ button.
This phase is designed to capture a variety of information, following these directions carefully will ensure that the data you capture is as accurate as possible.
Use the ‘Start Timer’ button to start the timer and enable the rest of the buttons. For each syllable click the ‘Syllable’ button, and for each disfluency hold down the correct disfluency button for the duration of that disfluency in order to capture duration data. Be sure to click the ‘Stop Timer’ button whenever the subject stops speaking for longer periods than mere prosodic pauses and click the ‘Start Timer’ button again when they start speaking again.
Spend some time carefully going over More Typical vs Less Typical disfluency types (Gregory & Hill, 1999) in order to be able to easily tell the difference during the assessment, errors in this area can greatly impact the validity of your results.
Keep the gestalt view of cultural differences in mind. As Robinson and Crowe (1987) found, normal disfluencies such as word and phrase repetition may be higher in non-stuttering African Americans than their Caucasian counterparts. While I do not have relevant research for other ethnic group comparisons, it is important to be mindful of the fact that these differences in typical disfluency rates can impact the validity of our assessment results. This is particularly true if we’re not careful of which fluency classification we ascribe to a particular disfluent event.
Disfluency Frequency in Words
Duration of phase: 5 minutes of subject speech.
Important Note: Word count is only increased when you click the ‘Word’ button.
Use the ‘Start Timer’ button to start the timer and enable the rest of the buttons. For each word the subject utters click the ‘Word’ button, and for each disfluency click the ‘Disfluency’ button. Unlike the Disfluency Frequency in Syllables phase, there is no need to hold down the ‘Disfluency’ button during this phase of the assessment. Be sure to click the ‘Stop Timer’ button whenever the subject stops speaking and click the ‘Start Timer’ button again when they start speaking again.
When the results file is saved, norms based on Guitar (1998) will be displayed for comparison with the frequency of disfluency exhibited.
Important Note: Speech Rate is automatically calculated during other phases of assessment.
Speech Rate is reported both in Words Per Minute as well as Syllables Per Minute. The age of the subject will determine which norms are shown in the output file you’ll save at the end of the assessment.
Words Per Minute norms are based on Pindzola, Jenkins & Lokken (1989), Davis & Guitar (1976), and data summarized by Andrews & Ingham (1971).
Syllables Per Minute norms are based on data summarized by Andrews & Ingham (1971). I’m currently unable to locate SPM norms for children based on age as opposed to grade level, I hope to include this information when I can track it down. For now the application reports only adult norms.
Duration of phase: 2.5 minutes of subject speech
Important Note: As soon as you rate the previous 15 seconds of speech, the timer begins again. Be prepared.
The Naturalness Rating is a nine point ordinal rating shown to have high inter-rater reliability (Martin, Haroldson & Triden 1984).
Over 10 trials you’ll listen to the subject’s speech for 15 seconds and when the alarm sounds you’ll rate the previous 15 seconds on a scale of one to nine. One corresponds to most natural and nine corresponds to least natural.
The Effort Rating is a four point ordinal rating based on Guitar (1998).
To appropriately use this rating it’s important to thoroughly read Dr. Guitar’s text ‘Stuttering: Considerations in the evaluation of treatment’ which will explain his viewpoint on the levels of disfluency, from normal disfluency to advanced stuttering. This rating is one piece of a larger assessment paradigm, and the information gleaned should be integrated into a complete picture of the subject by considering all aspects of the suspected dysfluency.
The Awareness Rating is a five point ordinal rating based on Guitar (1998).
The Awareness Rating is part of the same holistic assessment system as the Effort Rating, which is designed to maintain a wide focus involving the entirety of the subject.
The Secondary Behaviors section is a criterion referenced checklist.
Guitar (1998) specified several behaviors which are indicative of a specific level of stuttering. These behaviors are listed and the application will automatically list both the level of stuttering of which some associated behaviors were exhibited, as well as which level of stuttering of which all behaviors were exhibited.
Additional secondary behaviors which I felt might be pertinent are also included in the checklist, and all specific behaviors not listed should be noted in the ‘Notes’ section of the application.
Important Note: The only required field is Age, this is to ensure that appropriate norms can be reported if they are available.
The other sections involved in this application include text fields to fill in Name (of subject), Age, Clinician Name, and Notes, as well as a drop-down Date chooser.
The Name field is not required but will be included in the output file if you choose to include it, and will be used as part of the default name of the output file. For example, if you use the name Bob, when you save the file the default name it will save as (which can be changed) will start with Bob.
The Age field is required and only accepts whole numbers. If you enter a number with a decimal point the remainder will be stripped and the age will be rounded down. Age will be reported in the output file and will be used to determine which norms are reported in the output file.
The Clinician Name field is not required, but will be reported in the output file if entered.
The Notes field is not required; however it is strongly recommended.
This field is freeform, you can type whatever you like in it and it will be faithfully reproduced in the output file. You cannot, however, paste information from other sources into it and expect faithful reproduction.
The drop-down Date chooser defaults to the current date, and should only be used if analyzing audio or video recorded on a previous date, and even then with a deeply critical eye as to whether this is appropriate considering your paradigm of use.
The results of the evaluation can be saved to a location on your computer. This maintains the integrity of Private Health Information/Personally Identifying Infromation. This maintenance of privacy cannot be controlled if you choose to save to a cloud location, such as Dropbox, which is not compliant with HIPAA.
When deciding what format in which to save text, the Rich Text Format (.rtf) was chosen to ensure maximum portability and reliability. The .rtf format allows for enough format options to create a readable report while still relying on an open standard. All personal computer systems I’m aware of are able to open, edit and save .rtf files without purchasing any additional software.
Important Note: The Frequency Breakdown refers to the frequency of disfluencies exhibited during the phase when syllables are counted.
The Frequency Breakdown area looks at each individual type of disfluency and gives some specific information.
Syllable Repetitions: 32
Percent of total disfluencies: 28.46%
Percent of classification (lt): 62.73%
Frequency in total syllables: 1.32%
*Examples are not representative of actual subject data, numbers may not add up correctly.
The first line shows the number of that specific type of disfluencies captured. In this case the evaluator clicked the ‘Syllable Repetitions’ button 32 times.
The ‘Percent of total disfluencies’ line shows the percentage of all of the disfluencies. In this case if you look at all of the disfluencies the subject exhibited, 28.46% of them were syllable repetitions.
The ‘Percent of classification’ line shows a percentage of the disfluencies in the same classification. In this case if you look at all of the less typical disfluencies, 62.73% of them were syllable repetitions. The ‘(lt)’ in this line signifies that this is in the ‘Less Typical’ category as opposed to the ‘More Typical’ category.
The ‘Frequency in total syllables’ shows what percentage of the total number of syllables had an instance of the specific disfluency. In this case if you look at all of the syllables, 1.32% of them had a syllable repetition.
Important Note: The Duration Breakdown refers to the duration of disfluencies exhibited during the phase when syllables are counted.
The Duration Breakdown area looks at each individual type of disfluency and gives some specific information.
Total duration of disfluencies in this category: 8.32 seconds
Percent of total speech duration: 1.39%
Percent of total disfluent duration: 12.45%
Percent of disfluency classification: 63.88%
Average duration of all disfluencies: 1.55 seconds
Average duration of three longest disfluencies: 2.84 seconds
Longest disfluency duration: 3.42 seconds
Second longest disfluency duration: 2.87 seconds
Third longest disfluency duration: 2.23 seconds
*Examples are not representative of actual subject data, numbers may not add up correctly.
The ‘Total duration’ line shows the duration of all events of this specific type of disfluency added together. In this case the total time the evaluator held down the ‘Syllable Repetitions’ button adds up to 8.32 seconds.
The ‘Percent of total speech duration’ line shows the percentage of the total duration of disfluency of this type compared to the total speaking time of the subject. In this case if the subject spoke for 600 seconds (10 minutes) all of the syllable repetition durations added together were 1.39% (or 8.32 seconds, as indicated on the line above).
The ‘Percent of total disfluent duration’ line shows the percentage of the total duration of disfluency of this type compared to the total disfluent time of the subject. In this case if the subject was disfluent for 100 seconds, 12.45 of those seconds would have been syllable repetitions.
The ‘Percent of disfluency classification’ line shows the percentage of the total duration of disfluency of this type compared to the total disfluent time of the subject in this classification. In this case if the subject exhibited less typical disfluencies for a total of 10 seconds when added together, 6.39 of those seconds would have been syllable repetitions.
The ‘Average duration of all disfluencies’ line shows how long the average duration was for this type of disfluency.
The ‘Average duration of three longest disfluencies’ line shows how long the average duration was for the three longest disfluent events of this type.
The ‘Longest disfluency duration’ line shows how long the actual duration was for the longest disfluent event of this type.
The ‘Second longest disfluency duration’ line shows how long the actual duration was for the second longest disfluent event of this type.
The ‘Third longest disfluency duration’ line shows how long the actual duration was for the third longest disfluent event of this type.
Normative Data Referenced
This application has not been independently normed, instead the primary purpose is to gather information and allow the evaluator to apply norms already developed to the information gathered. Wherever possible I’ve added normative data from published sources so that the evaluator can compare results with established research.
First, this application is just a tool. Clinicians, with their education, qualifications and expertise should never treat this application as anything other than an aid to help them organize data to then use with their own clinical judgment.
This application is designed not to email information or data. If you choose to save data to cloud storage services, such as Dropbox, be sure to understand the privacy compliance (or lack thereof) provided and the ethical and legal implications.
This application is designed not to maintain internal data between uses. If you close the application it will ‘forget’ the data you recorded so that if another user opens the application PII/PHI will not be exposed.
Andrews, G., & Ingham, R. (1971). Stuttering: Considerations in the evaluation of treatment. International Journal of Language & Communication Disorders, 6(2), 129-138.
Davis M & Guitar R (1976). Speech rate of elementary school children in Vermont. Graduate student research paper. University of Vermont, Burlington, VT.
Gregory, Hugo H., & Diane Hill. “Differential Evaluation – Differential Therapy for Stuttering Children.” Stuttering and Related Disorders of Fluency. By Richard F. Curlee. New York: Thieme, 1999. 22-29. Print.
Guitar, Barry. Stuttering: An Integrated Approach to Its Nature and Treatment. 2nd ed. Baltimore, MD: Williams & Wilkins, 1998. Print.
Martin, R. R., Haroldson, S. K., & Triden, K. a. (1984). Stuttering and speech naturalness. The Journal of speech and hearing disorders, 49(1), 53–8. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/6700202
Pindzola, R. H., Jenkins, M. M., & Lokken, K. J. (1989). Speaking rates of young children. Language, Speech, and Hearing Services in Schools, 20(2), 133–139. Retrieved from http://psycnet.apa.org/psycinfo/1989-36074-001
Robinson, T.L. Jr., & Crowe, T.A. (1987). A comparative study of speech disfluencies in nonstuttering Black and White college athletes. Journal of Fluency Disorders. 12,147-156.
Yaruss, J.S. (1998). Real-time analysis of speech fluency: Procedures and reliability training. American Journal of Speech-Language Pathology, 7(2), 25-37.