2003 IFA Congress: Montreal, Canada

Conversation Analysis of Children Who Stutter and Parents Who Stutter

John A. Tetnowski1, Jack S. Damico1, Jennifer A. Bathel1 & Thomas C. Franklin2
1University of Louisiana at Lafayette, R0. Box 43] 70, Lafayette, LA 70504-3170 USA
2Southwest Missouri State University, Communication Sciences and Disorders, 901 S. National Avenue, Springfield, MO 65804 USA


Due to the inheritance patterns of stuttering, children who stutter are likely to have a parent who also stutters. Despite this, most parent-child interactional studies in the field of stuttering are between a child who stutters and an adult parent who does not stutter. This study uses a modified qualitative approach to describe the complex interactions that occur between a child who stutters and her parent, who also stutters. The parent modifies his communication patterns significantly when speaking with his daughter that greatly reduce his stuttering behaviors. Descriptions and examples of conversation are described in detail.

  1. Introduction
When attempting to understand the complexity of the communication process- regardless of whether the focus is on a disorder- it is beneficial to actually record and analyze authentic communication interaction in as many contexts as possible. Such a focus on authentic settings that involve different contextual variables often reveal important factors that may influence the behaviors under scrutiny and our understanding of the role those variables play. For example, many researchers have investigated the effects of parental speech behaviors on their stuttering children (e.g., Kelly & Conture, 1992; Meyers & Freeman, 1985 a,b,c: Zebrowski & Conture, 1989). The research findings have been erratic in their conclusions regarding the impact of parental behaviors (see Nippold & Rudzinski, 1995 for a review). Despite the inconclusive findings regarding parental behaviors in relation to children’s stuttering behaviors, advice given to parents of disfluent children generally supports the notion that parental speech attitudes and behaviors affect the speech of their children.

It is common practice for a speech-language pathologist to advise parents to slow their rate of speech, use simpler sentences, not interrupt the child’s speech, and generally create a safe and relaxing environment for the child to speak (e.g., Ainsworth & Fraser, 1988; Bernstien Ratner, 1992; Guitar, et a1., 1992). While this sounds like logical advice for parents, it is subscribing to an etiologic theory of stuttering that has not been empirically conclusive in the literature.

One reason these data may be inconclusive is that most of the studies have used non-stuttering parents to study their interactions, even though evidence on stuttering inheritance patterns would suggest that many of the children who stutter will indeed have a parent who also stutters (e.g., Ambrose, Cox, & Yairi, 1997; Felsenfeld, 1996). Findings from the nonstuttering parent studies indicate that 1) these parents communicate differently when interacting with their child who stutters and 2) children interact differently with their mothers than with their fathers. It is likely that the speech behaviors of the child differ dependent upon whom ever the conversational partner happens to be. It is just as likely, therefore, that the speech of the conversational partner may differ dependent on the speech of the child. This should not be unexpected since communication is a synergistic interaction that vacillates as various internal and external variables (i.e., presence of disfluencies, rate of speech, identity of conversational partner, response time latency, presence of interruptions, linguistic complexity) present themselves (Goffman, 1967; Smith, 1999).

Since communication is such a complex system, a more descriptive means of assessment may be warranted. One system of analysis, conversation analysis (see Tetnowski & Damico, 2001, for details) will allow for the observed interactional behaviors between speakers to be clearly identified and investigated. Therefore, the purpose of this study is to describe the interactions of a child who stutters, while interacting with her father, who also stutters, by studying the lesser-analyzed content- but one that offers much potential for analyzing contextual variation. An open-stance (i.e., qualitative, see Damico & Simmons-Mackie, 2003) viewpoint will be utilized here to provide significantly more depth to the analysis.

  1. The Study
There seems to be a tendency in current research paradigm to only include in studies those parents of children who stutter that do not stutter themselves. This is surprising given that stuttering tends to run in families (Ambrose et al., 1997; Felsenfeld, 1996). Despite this knowledge, sufficient attention has not been given to the analysis of communicative interactions between children who stutter and their stuttering parent. Despite the likelihood of a child who stutters may be interacting with a parent who stutters, these interactions have been sparse in the literature. In addition, all available literature has controlled for particular parental speech behaviors (e. g., speaking rate, turn taking, frequency of interruption, response time latency, etc.) and compared such behaviors with that of nonstuttering parent/child dyads through quantitative measures. This type of quantitative analysis supposes that all stuttering children and their parents are alike. That is, most studies have made the assumption that people who stutter are a homogeneous group, not a dynamic and synergistic population. This is interesting because many stuttering texts and descriptive narratives describe people who stutter as having uniquely different patterns and development (e. g., Van Riper, 1982). Most experimental research is based on the methodology of controlling for all variables and presenting findings that assume a group of stutterers are a homogeneous group. In a recent publication by Manning (2001) the preface to the book quotes Bannister (1966) stating that humans are “notoriously nonsensical and unfit for scientific scrutiny.” We do not mean to challenge the experimental paradigm that has become so successful in building the scientific underpinning of our profession. However, we do feel that additional methodologies such as qualitative research methodologies can provide additional and different data. Additionally, qualitative methods can address some variables as they interact rather that attempt to control them. Qualitative methods may allow for a greater level of naturalistic study of communication interaction in people who stutter as has been called for in the literature (see Ingham & Riley, 1998).

In light of previous conflicting research and findings that indicate linguistic complexity of the speaker to be significant, studying the interaction between parents who stutter and children who stutter is fertile ground for further understanding the complex relationship between parents and children who stutter. As previously discussed, children who stutter may be interacting daily with a parent who also stutters. Consequently, it would be beneficial to examine the effects of parental speech behaviors of parents who stutter on the speech of their children who also stutter, and vice versa.

  1. Participants
The participants are a 40-year-old father who stutters and his 2.7-year-old daughter who was beginning to show signs of stuttering. During the interview portion of his daughter’s (KL) diagnostic session the father (ML) revealed that he was concerned with his daughter’s fluency because he was noticing patterns in her speech that resemble his own stuttered speech. He reported that he had stuttered most of his life and had brief ineffective therapy during middle school and a brief period of private therapy since that time. While interacting with his daughter, the clinician noticed that he did not seem to stutter. When asked to estimate how much he stutters when speaking to his daughter he guessed about 50-60% of the time, thus reporting little awareness of his speech behaviors when speaking to his daughter.

Since the evaluation KL has been involved in therapy to reduce her frequency of stuttering using a modified version of the Lidcombe stuttering program. ML was determined to be a severe stutter with a range of syllables stuttered from 2.5% to 40% in structured tasks and levels averaging 22 stuttered syllables per minute during conversation tasks during a complete stuttering evaluation. ML also began therapy shortly after the diagnostic session.

  1. Method
The methodology employed is a modified qualitative approach incorporating interactional analysis with several ethnographic data collection and analysis procedures (Creswell, 1998; Damico & Simmons-Mackie, 2003). Specifically, an interactional analysis of videotaped conversations was undertaken and then triangulated with data obtained via interviews and collection of discussion data during support group meetings. Through this qualitative approach, quantitative disfluency measures (i.e., percent of stuttered syllables) were incorporated with other behaviors and variables that act to enrich the interactional descriptions and to suggest more potential and viable interactional causes and effects.

Data for this study was collected over a period of 8 months and involved audio-recording, participant observation, and ethnographic interviewing (see Damico & Simmons-Mackie, 2003; Tetnowski & Franklin, 2003). These data collection strategies were employed during a diagnostic session in the participants’ home, and in-clinic evaluation, a support group meeting, and during several therapeutic sessions within the clinic. The key sample for this study was gathered during the 5”â  therapeutic session with ML. The data analyzed spanned approximately 4 minutes and includes interactions between ML, two clinicians, and KL. Data during this sampling period was recorded and then analyzed using video playback. More specific and detailed analyses were completed using the Computerized Speech Research Environment (CSRE) and a desktop computer.

  1. Results
Quantitative Analysis of Speech and Language

Percentage and type of stuttering was calculated for the father and daughter as well as qualitative descriptions of each subjects’ speech. KL displayed 3% stut:ered syllables when speaking to the clinician during the diagnostic session, 4% stuttered syllables when speaking with her mother at home and during the diagnostic session, and --% stuttered syllables when speaking to her father during the diagnostic session. ML displayed 7% stuttered syllables when conversing with the clinician during his daughter’s diagnostic session, 10.32% when conversing with the clinician and other adults during a support group meeting, and 0.9% when speaking to his daughter during the diagnostic session. These preliminary results indicated that the father stutters with lower frequency when speaking to his daughter than when speaking to the clinician and other adults. This has been reported in the literature (e.g., Van Riper, 1982), but not described in detail. These results are summarized in the Table 1.


Table 1. Description of stuttering behaviors.

At this point in therapy ML had been instructed to not use his techniques outside the therapy session and ML reported to not be using them except in the therapy session and when doing his structured homework away from the clinic. Intrarater reliability was ensured when both authors agreed to frequency, type, and transcription assignment with 100% agreement for each utterance (using either the audio or video that was available). The transcription method was based on a combination of strategies that would allow the authors to describe the conversational sample as completely as possible. Methods were based upon transcription methods recommended by Tetnowski & Damico (2001). At this point no other transcription method and analysis was deemed suitable for the depth of analysis required for this project. Once the data were recorded, the tapes were cyclically reviewed and transcribed.

The transcript was orthographic, with every utterance accounted for (including unintelligible utterances). Non-speech behaviors that might provide important interactional data were included in the transcript (e.g., laughter, body movements, change in bodily location). After transcription, moments of stuttering were recorded and coded as were moments of non-stuttered disfluency.

During this procedure, the stuttering-like-disfiuencies and the nonstuttering-like-disfluencies were then categorized according to type (e.g., word repetition, block, filler, etc.). The stuttering-like- disfluencies included single syllable word repetitions, blocks, and prolongations, while the non-stuttering-like-disfluencies included fillers (e.g., um, uh, etc), phrase repetitions, broken words, and revisions. Following this coding procedure, frequency counts were obtained in addition to the calculation of Mean Length of Utterance (MLU) for descriptive purposes only. Due to the highly subjective nature of transcript analysis a set of rules were then outlined to describe how the authors came to their categorical decisions (see notes above Example 1; for a more in-depth description, see Tetnowski & Damico, 2001). The sample was then analyzed to determine particular patterns and points of interest. These initial codings and analyses were completed by the first author.

Fifty-one utterances were analyzed during a therapeutic session. Twenty-five utterances were analyzed when ML was speaking to J T (second author) and OB (KL’s clinician) and 26 utterances were gathered when ML was speaking to his daughter (KL). During this 4 minute sample ML displayed a 7.27 %SS, 11.77 %DIS, and an MLU score of 11.56 when speaking to the clinicians. He displayed 2.08 %SS, 2.08 % DIS, and an MLU score of 3.69 when speaking to his daughter. ML stuttered with less frequency when speaking to his daughter consistent with previous findings during his daughter’s diagnostic session. In addition, ML did not display any nonstuttering-like disfluencies (NSD) (i.e., fillers, phrase repetitions, broken words) when speaking to his daughter. It can also be noted that ML’s MLU was dramatically reduced when speaking to his daughter, indicating that ML uses shorter utterances (and possibly less complex utterances) when speaking to KL. Results are summarized in Table 2 below.


Table 2. Description of ML’s speaking with clinicians and child (KL).

Qualitative Analysis of Speech and Language

To enrich our understanding of the interactional complexity, further analysis of ML’s communication was conducted using qualitative methodologies. Analyses of conversation using these qualitative strategies can provide more depth than traditional measures (Tetnowski & Damico, 2001; Susca & Healey, 2002).

In the analysis of ML’s speech behaviors, child directed speech (CDS), also known as motherese was taken into account. Child directed speech has been de:ermined to include an increase in pitch, increased vowel duration, and decreased rate (Werker et al., 1994). Qualitative judgments regarding the characteristics of ML’s speech when speaking to his daughter were consistent with typical CDS and included increased pitch, exaggerated prosodic variation, rhythmic variations, decreased rate, and an increase in the use of questions and repetitions. These changes in speech behavior were also noted to occur in the speech of the two clinicians when speaking to KL.


Table 3. Description of stuttering and disfluency when using child directed speech (CDS) or not (NCDS).

In order to assess the influence of CDS speech behaviors and their impact on ML’s fluency, ML’s utterances were categorized according to his use of CDS and non-use of CDS (NCDS). CDS speech was judged and labeled by the authors using the same method of agreement previously described. In other words, ML’s increase in pitch, exaggerated prosodic variation, and slower rate clearly marked the onset of CDS. As a result of this differentiation, an in-depth analysis was completed using only utterances when ML was speaking to KL. This analysis was completed since ML did not use CDS in all instances when talking with KL.

In instances when ML was directing his talk to KL (and judged as CDS), his MLU was 3.57, and his %SS was 1.47. In instances where ML was talking to KL and was not using CDS, his MLU was 3.85, and his %SS was 3.7. ML’s MLU was about the same whether he used CDS or not when talking with KL. However, ML’s %SS was lower when he used CD8 and talking to KL than when he did not use CDS and was talking to KL.

Qualitative Descriptions

There are times when ML is speaking to his daughter that he slows his rate of speech, increases

the duration of individual words, uses rising intonation more often, and seems to use higher pitch. His speech is generally faster when speaking to adults and perceptually slower when speaking to his daughter. The following examples demonstrate his strategic use of CDS and how it affects his speech. In Example 1, ML is speaking to the clinician (JT) when his daughter (KL) is not in the room.


Example 1
(Note: All disfluencies are in bold; the line above indicates eye gaze; descriptions of disfluency type are above the utterance, sswr=single syllable word repetition,pWr=part Word repetition, bwzbroken word, revzrevision, filzinterjection, phr=phrase repetition,bl=blockage; *=unintelligible; $=child directed speech)

In this example, ML exhibits his typical pattern while interacting with another adult. He produces a spontaneous complex utterance that is quite long and complex. Example 1 lies in stark contrast to the patterns that emerge when ML speaks to his daughter. Example 2 illustrates this quite clearly. In this first example, the utterance is quite long in length and complexity when talking to J T, but shorter when talking to KL (in example 2). Also noted in Example 1 are several instances of stuttering, such as the single syllable word repetitions in lines 2354, 2355, 2356, 2357 and 2360.


Example 2
In this example, ML continues his explanation from the previous example and remains disfluent until KL enters the room. When KL enters the room he immediately and with ease switched into CDS and is successful at maintaining fluency. In line 2369, “Hey sweetie” is a short utterance,

but typical of ML’s utterances when speaking to his daughter as observed in multiple observations of conversational interactions between ML and KL in various settings. In addition, ML strategically, yet unconsciously, uses CDS to maintain the attention of his daughter and to elicit desired responses from his daughter. This strategic application is demonstrated in Example 3.


Example 3
In the first line (2426) ML is asking KL what she did in therapy that day and KL responds (line 2427) with her teacher’s name; an undesired response to “what did you do?”. In line 2428 ML switches into CDS at the end of his sentence on the word “do” and then proceeds to continue to use CDS throughout the conversational exchange. He also provides KL with the correct response in order to obtain the desired answer to line 2426 by using a simple recast (Weiss, 2002) to elicit more information about the book reading activity. It is also known that KL_likes Blues Clues. Lines 2434- 2435 is an example of ML’s use of repeated utterances and is typical of his speech when speaking to his daughter. It is also interesting to note that ML will maintain the use of CDS when speaking to an adult if he is still focused on KL (e.g., she is in his lap or he still intends to speak to her). He does this to maintain fluency.

  1. Summary & Discussion
Literature investigating parent-child dyads have examined how children who stutter respond to the speech behaviors of their non stuttering parents (Bernstein Ratner, 1992; Kelly & Conture, 1992; Myers & Freeman, l985a,b,c; Nippold & Rudzinski, 1995; Yaruss & Conture, 1995). The literature has not been able to consistently determine if and how parental speech behaviors affect the speech behaviors of children. Despite KL’s diagnosis of mild stuttering, she did not display any stuttering behavior during this sample. It is possible; therefore, that KL responded favorably to ML’s use of CDS and was subsequently more fluent when conversing with her father. Our data has also shown that ML speaks differently when speaking to his daughter than when speaking to adults. When ML speaks to his daughter using both simplified language and altered speech characteristics (i.e., the strategic use of child-directed speech) he stutters with less frequency and severity. It has been our observation that alteration of both speech and language variables are required for the decreased stuttering effect to be at its maximum. Further studies using qualitative analysis procedures, in addition to more quantitative studies in natural conditions are warranted to understand these complex interactions that take place in authentic communications.

Finally, these results are based on a single case documenting the use of CDS. It is not intended to be predictive of the entire population of people who stutter. However, this study has strong implications for the reason people who stutter do not stutter when speaking to young children. More importantly, it also has strong implications regarding how parents speak to their stuttering children and how stuttering parents interact with their stuttering children, something neglected until now. Further research should include a recurrent and closer look at CDS and it’s implications for intervention and parental training.

Ainsworth, S., & Fraser, J. (1988). If your child stutters: A guide for parents ( 3rd ed. ). Memphis, TN: Speech Foundation of America.
Ambrose, N.G., Cox, N .J ., & Yairi, E. (1997). The genetic basis of persistence and recovery in stuttering. American Speech-Language Hearing Association, 40, 567-580.
Bernstein Ratner, N. (1992). Measurable outcomes of instructions to modify normal parent-child verbal interactions: Implications for indirect stuttering therapy. Journal of Speech and Hearing Research, 35, 14-20.
Creswell, J .W. (1998). Qualitative inquiry and research design: Choosing among five traditions. Thousand Oaks, CA: Sage Publications.
Damico, J.S. & Simmons-Mackie, N.N. (2003). Qualitative research and speech-language pathology: A tutorial for the clinical realm. American Journal of Speech-Language Pathology, 12, 131-143.
Felsenfeld, S. (1996). Progress and needs in the genetics of stuttering. Journal of Fluency Disorders, 21, 77-103.
Goffman, E. (1967). Interaction ritual: Essays in face to face behavior. Garden City, NY: Doubleday.
Guitar, B., Schaefer,  Donahue_Hilburg, G., & Bond, L. (1992). Parental verbal interactions and speech rate: A case history in stuttering. Journal of Speech and Hearing Research, 35, 742-754.
Ingham, J .C., & Riley, G. (1998). Guidelines for documentation of treatment efficacy for young children who stutter. Journal of Speech Language Hearing Research, 41, 753-770.
Kelly, E.M., & Conure, E.G. (1992). Speaking rates, response time latencies, and interrupting behaviors of young stutterers, nonstutterers, and their mothers. Journal of Speech and Hearing Research, 35, 1256-1267.
Manning, W.H. (2001). Clinical Decision Making in Fluency Disorders. San Diego, CA: Singular.
Miles, S., & Bernstien Ratner, N. (2001). Parental language input to children at stuttering onset. Journal
Myers, S.C., & Freeman, F.J. (1985a). Are mothers of stutterers different? An investigation of social communicative interaction. Journal of Fluency Disorders, 10, 193-209.
Myers, S.C., & Freeman, F.J. (1985b). Interruptions as a variable in stuttering and disfluency. Journal of Speech and Hearing Research, 28, 428-435.
Myers, S.C., & Freeman, F.J. (19850). Mother and child speech rates as a variable in stuttering and disfluency. Journal of Speech and Hearing Research, 28, 436-444.
Nippold, M.A. & Ruzinski, M. (1995). Parent’s speech and children’s stuttering: A Critique of the literature. Journal of Speech and Hearing Research, 38, 978-998.
Smith, A. (1999). Stuttering: A unified approach to a multifactorial, dynamic disorder. In N. Bernstein Ratner & E.C. Healy (Eds.), Stuttering Research and Practice: Bridging the Gap (pp. 28-44). New Jersey: Lawrence Erlbaum Assoc.
Tetnowski, J. A, & Damico, J .S. (2001). A demonstration of the advantages of qualitative methodologies in stuttering research. Journal of Fluency Disorders, 26, 17-42. Section 5. Language, Speech and Discourse 279
Tetnowski, J.A. & Franklin, T.C. (2003). Qualitative research: Implications for description and assessment. American Journal of Speech-Language Pathology, 12, 155-164.
Van Riper, C. (1982). The Nature of Stuttering, 2"“' ed. Englewood Cliffs, N.J.: Prentice Hall.
Weiss, A.L. (2002). Recasts in parents’ language to their school-age children who stutter: A preliminary study. Journal of Fluency Disorders, 27, 243-266.
Werker, J .F., Pegg, J .E.,& McCleod, P.J. (1994). A cross-language investigation of infant preference for infant-directed communication. Infant Behavior and Development, 17, 323-333.
Yairi, E; Ambrose, N . & Cox, N. (1996). Genetics of stuttering: A critical review. Journal of Speech and Hearing Research, 39, 771-784.
Yaruss, J .S., & Conture, E.G. (1995). Mother and child speaking rates and utterance lengths in adjacent fluent utterances: Preliminary observations. Journal of Fluency Disorders, 20, 257- 278.
Zebrowski, P.M., & Conture, E.G. (1989). Judgments of disfluencies by mothers of stuttering & normally fluent children. Journal of Speech and Hearing Research, 32, 625-634.
join button


The IFA implemented Japanese translations of some pages on the site for the 2018 Joint World Congress. Choosing Japanese below to see these translations.

Not all pages are translated, but you can use Google translate to see a machine translation using the switch below

Google Translate


Journal of Fluency DisordersBrowse the current issue

The official journal of the International Fluency Association
IFA Members receive online access to JFD as a member benefit.

Read more: JFD

Twitter Feed @official_IFA

IFA on Facebook