- Email: [email protected]
Elementary school students’ perceptions of stuttering: A mixed model approach
Journal of Fluency Disorders 45 (2015) 1–11
Contents lists available at ScienceDirect
Journal of Fluency Disorders
Elementary school students’ perceptions of stuttering: A mixed model approach James Panico a,∗ , E. Charles Healey b,1 , Jennifer Knopik c a b c
Southern Illinois University Edwardsville, 1101 Founders Hall, Edwardsville, IL 62026, USA University of Nebraska-Lincoln, 253 Barkley Memorial Center, Lincoln, NE 68503, USA Southern Illinois University Edwardsville, Edwardsville, IL 62026, USA
a r t i c l e
i n f o
Article history: Received 25 February 2015 Received in revised form 11 June 2015 Accepted 14 June 2015 Available online 28 June 2015 Keywords: Stuttering Elementary Students Perceptions
a b s t r a c t Purpose: Past research studies have focused on perceptions of stuttering by various age groups and only a few have examined how children react to a peer who stutters. All of these studies used a quantitative analysis but only one included a qualitative analysis of elementary school age children’s responses to stuttering. The aim of this study was to further explore the perceptions of elementary school students toward a peer who stutters using both quantitative and qualitative analyses of three levels of stuttering. Methods: Participants included 88 elementary school children between 8 and 12 years of age. Each participant viewed one of four audiovisual samples of a peer producing fluent speech and mild, moderate, and severe simulated stuttering. Each participant then rated five Likert statements and answered three open-ended questions. Results: Quantitative and qualitative results indicated that negative ratings and the percentage of negative comments increased as the frequency of stuttering increased. However, the children in this study indicated that they were comfortable listening to stuttering and would be comfortable making friends with the peer who stutters. Conclusion: The findings of this study together with past research in this area should help clinicians and their clients appreciate the range of social and emotional reactions peers have of a child who stutters. Educational objectives: After reading this article, the reader will be able to: (a) discuss past research regarding children’s perceptions of stuttering; (b) summarize the need to explore the perceptions of elementary-aged children toward a peer who stutters; (c) describe the major quantitative and qualitative findings of children’s perceptions of stuttering; and (d) discuss the need for disseminating more information about stuttering to children and teachers. © 2015 Elsevier Inc. All rights reserved.
1. Introduction For approximately half a century, researchers have been interested in how school aged children react to stuttering. Knowing how children react to stuttering has important clinical implications about how children who stutter might be
∗ Corresponding author. Tel.: +1 618 650 5838. E-mail addresses: [email protected] (J. Panico), [email protected] (E.C. Healey). 1 Tel.: +1 402 304 1819. http://dx.doi.org/10.1016/j.jfludis.2015.06.001 0094-730X/© 2015 Elsevier Inc. All rights reserved.
2
J. Panico et al. / Journal of Fluency Disorders 45 (2015) 1–11
treated by their peers in and outside of the school setting as well as how they manage social relationships involving teasing and bullying (Langevin & Hagler, 2004). Unfortunately, there are only a few empirical studies on this topic. Giolas and Williams (1958) conducted the first study in this area to determine how a total of 120 five- and eight-year-old children reacted to audio samples of three adult speakers producing fluent and imitated stuttered speech as judged by three trained speech-language pathologists. They discovered that children as young as five years-of-age showed awareness of disfluent speech and reacted negatively when hearing part-word repetitions in the adult’s speech. The authors reported that the group of eight year-old children used the label “stuttering” when they heard an adult producing 10% part-word repetitions during the speech sample. Five-year-old children in the study did not use the word stuttering after hearing the disfluent speech. Culatta and Sloan conducted a similar study in 1977 and asked sixty children in the first through fourth grades (six to nine years-of-age) to listen to audio samples of stuttering and fluent speech produced by an adult speaker. The method by which the stuttering samples were created was not specified. The children were divided into two groups of 30: (1) first and second graders, and (2) third and fourth graders. As expected, both age groups preferred the fluent speech sample over the one that contained the stuttering. They also found that the first and second grade children did not use the term “stuttering” to refer to the disfluent speech but “stuttering” was used by about 37% of the third and fourth graders. Ambrose and Yairi (1994) examined perceptions of stuttering of 20 normally fluent and 20 disfluent children between the ages of two and five by asking them to choose between two types of speech produced by puppets; one who spoke fluently and the other who stuttered. The children who stuttered had to meet specific criteria set forth by the authors for inclusion in the study. The disfluent children who were three years of age identified closely with the puppet who stuttered while the fluent speaking three-year-olds selected the fluent puppet as the one that resembled how they spoke. This study showed that the awareness of stuttering begins at an early age and preschoolers are capable of differentiating between fluent and stuttered speech. These findings were replicated by Ezrati-Vinacour, Platzky, and Yairi (2001) who also noted that the term “stuttering” was not used by children under five years of age. Only a few of the seven year-old children labeled the disfluent speech as “stuttering.” Several reports have indicated that older children (9–14 years) respond with more negative reactions to stuttered speech. Franck, Jackson, Pimentel, and Greenwood (2003) had 75 9- to 11-year-old children listen to one fluent speech sample and one containing moderate stuttering from an adult who stuttered. The speakers’ stuttering severity level was determined through a Stuttering Severity Instrument -3 (Riley, 1994) rating from three experienced speech-language pathologists. Using a series of adjective pairs related to personality (outgoing–shy) and intelligence (intelligent–stupid), one group of children rated the fluent speaker and another group rated the speaker who stuttered. The results showed that fourth- and fifth-grade children had more negative perceptions of an adult who stuttered than the adult who did not stutter. Hartford and Leahy (2007) selected 80 children between six and 13 years-of-age and asked them to evaluate an audio sample of an adult speaker who produced fluent as well as moderate to severe pseudostuttering. The severity level of the pseudostuttering was determined by 22 speech-language pathology students. Using quantitative and qualitative measures, the researchers found that the six to eight year-old-children made fewer negative comments about the disfluent speaker than did children 9–13 years-of-age. Nonetheless, many of the children commented that the person who stuttered appeared “nervous,” “shy,” and “wouldn’t have many friends.” The consistent finding that elementary school-age children harbor negative perceptions of an adult who stutters provides evidence that children tend to react negatively toward a disfluent speaker at an early age. Langevin and Hagler (2004) were the first to study children’s perception of a disfluent peer relative to their beliefs, feelings, and attitudes toward a child who stutters. The major intent of their study was to create a scale to measure peer attitudes toward children who stutter. A sample of 267 elementary school children aged 8–13 years (grades four to six) were asked to rate 40 statements on a scale of 1–5 about a nine-year-old boy and an eight-year-old girl who stuttered. The boy stuttered on 16.5% and the girl stuttered on 43% of syllables produced in the sample. Each child also exhibited a variety of secondary coping behaviors such as head movements, broken eye contact, and rapid jaw jerking. Participants viewed one-minute samples of the boy and girl who stuttered and then rated statements relative to their affective (e.g., I would like having a kid who stutters live next door to me), behavioral (e.g., I would walk in the hall with a kid who stutters), and cognitive (e.g., kids who stutter are like normal kids) intentions toward children who stutter. The results of their study showed that fifth and sixth graders tended to have more positive perceptions of the two peers who stuttered than the fourth graders. They also found no differences in listeners’ ratings between the boy and girl who stuttered or between the boys and girls who served as participants. Additional results from the study focused on verifying the reliability and validity of the scale they had developed. Continued research into children’s and adolescents’ perceptions of stuttering was conducted by Evans, Healey, Kawai, and Rowland (2008). They examined the reactions and perceptions of 64 students between the ages of 10 and 14 toward a 16 year-old male peer who exhibited a very mild stuttering severity but was able to simulate a variety of stuttering types with a high level of control. The students were divided into four groups. Each group only listened to one type of video sample of the male peer speaking fluently or with simulated mild, moderate, or severe stuttering. All listeners were asked to rate 11 Likert statements using a five-point scale about the speaker shown in the video. The results of this study were similar to past studies involving children who rated adult speech samples in that, perceptional ratings of a person who stutters become increasingly more negative as the stuttering frequency increases. Additionally, the majority of the listeners reacted negatively even to mild stuttering although some students reported they were comfortable listening to and interacting with a peer who stuttered mildly. This finding implies that a peer with mild stuttering would be more likely to “fit in” at school than a peer who had moderate to severe stuttering.
J. Panico et al. / Journal of Fluency Disorders 45 (2015) 1–11
3
Although the Langevin and Hagler (2004) study provided the first quantitative investigation of children’s reactions to an elementary school age boy and girl who stutter, it is limited in that it did not provide any data concerning verbal statements or descriptions about how the children reacted to the samples of stuttering they heard. Eliciting peoples’ opinions, beliefs, and perspectives about a phenomenon (i.e., a qualitative research method) can provide a personal point of view of a given situation (Creswell, 1998). Because of the variety of ways stuttering behavior could be perceived, it would seem important to obtain children’s reactions to stuttering using a mixed model approach (i.e., combined quantitative and qualitative methods). Mixed methods has been used previously in the research by Susca and Healey (2001) and Panico, Healey, Brouwer, and Susca (2005) when investigating adults’ perceptions of adults who stutter, as well as Evans et al. (2008) who studied how children and adolescents react to a peer who stutters. The focus of these studies was to have listener’s rate statements about a person who stuttered of a scale of 1–5 as well as ask them to comment on their comfort in listening to stuttering, how they perceived the person who stuttered, and specific features about the way the person spoke. Another limitation of the Langevin and Hagler (2004) study is that they had their listeners rate only who stuttered moderately and severely. They stated that additional research is needed using children who have mild stuttering. Thus, a study that analyzes how elementary aged children react to a peer who has mild, moderate, or severe stuttering could expand the perception of stuttering across three levels of stuttering and provide a comparison of results to the Langevin and Hagler study. It could also provide insights for children who stutter about how peers might react to stuttering and why they might encounter teasing and bullying about the way they speak. The purpose of the present study was to address two research questions about children’s perceptions of stuttering: (1) How do children rate a peer who stutters relative to five statements (using a 7-point scale) about the peer’s speech? and (2) What types of comments do children produce when responding to a series of questions (e.g., how smooth was her speech? Was she a good speaker?) about a peer who speaks fluently as well as with mild, moderate, and severe simulated stuttering? 2. Methods 2.1. Participants A total of 88 children participated in this study and were recruited within the Edwardsville, Illinois (USA) community from the YMCA Kid’s Network Latchkey afterschool program. The participants ranged in age from 8 to 12 years and were in the third (n = 35), fourth (n = 27), or fifth (n = 26) grade. The study included 49 boys, ranging in age from 8 to 12 years (M = 9.85), and 39 girls, ranging in age from 8 to 11 years (M = 9.46). Each participant passed a hearing screening at 25 dB HL across 500, 1 K, 2 K, and 4 K frequencies using a Maico M39 portable audiometer. In order to be included in the study, the parent/guardian of each participant confirmed to the first author that his/her child had not previously received or was currently receiving speech and/or language therapy services. The participants’ parents/guardians also confirmed that his or her child had little to no knowledge, familiarity of, or contact with children who stutter. These two criteria were used to prevent bias among the participants when listening to stuttered speech if a child had extensive contact or exposure to a child who stutters. Langevin and Hagler (2004) found that children who had contact with a child who stutters have significantly more positive attitudes toward a child who stutters than children who don’t have contact with a child who stutters. 2.2. Construction of speech samples In order to control for confounding variables such as the type and duration of the disfluency, the speech samples were composed of simulated stuttering. A pilot study was conducted in order to determine if the types and number of simulated disfluencies were representative of mild, moderate, and severe stuttering. Five graduate students who had completed a graduate course in fluency disorders and were trained in the assessment of stuttering measures were asked to independently listen to three passages in which the first author read the passages in a conversational manner while incorporating simulated stuttering into each of the readings. The reading passage was taken from the Stuttering Severity Instrument-4 (Riley, 2009). The first author simulated the core behaviors of stuttering, including part-word repetitions, whole-word repetitions, and audible sound prolongations (Guitar, 2006) in an easy, rhythmical, and relaxed manner. The samples did not contain any concomitant behaviors during the simulated stuttering moments, such as facial or head movements. Because simulated secondary coping behaviors could have varied from one sample to the next, the authors chose not to include them for any speech sample. After watching each of the three audiovisual samples in a randomized order, each of the five graduate students was then asked to rate the sample from 1 to 3 (where 1 = mild stuttering; 2 = moderate stuttering; and 3 = severe stuttering). All five of the students independently classified the sample containing 5% stuttering in the forms above as having the least stuttering and was rated as mild stuttering. There was also unanimous agreement that the sample containing 10% stuttering was considered moderate in severity, and the sample containing 15% stuttering was rated as severe stuttering. This confirmed that the type and form of stuttering produced in the pilot study could be used as model forms of stuttering for the peer speaker to imitate. These percentages of stuttering (calculated as a percentage of words stuttered) and durations of stuttering words in the speech samples of the present study are consistent with stuttering severity classifications developed in previous studies (Evans et al., 2008; Susca & Healey, 2001).
4
J. Panico et al. / Journal of Fluency Disorders 45 (2015) 1–11
Upon completion of the pilot study, a total of four audiovisual speech samples were created. The samples consisted of the stuttering frequencies levels (5, 10, and 15%) tested in the pilot study along with a 0% frequency level, which served as a control condition. The peer speaker for the four audiovisual speech samples was a nine-year-old girl who was fluent and never had a history of stuttering. The girl was similar in age to the target age group of participants, was home schooled, and did not live in the same town as any of the participants. In addition, there were no anticipated problems using a girl as the speaker because one result from Langevin and Hagler (2004) showed no significant perceptual differences between a boy vs. a girl who stuttered. The girl in the present study willingly volunteered to participate and agreed to be trained by the first author to produce the same type and form of part-word repetitions, whole-word repetitions, and audible sound prolongations used in the pilot study. An extensive amount of time was spent instructing the girl and allowing her to produce and practice the three target forms of stuttering as many times as she wanted during oral reading of an experimental passage. The passage was a fourth grade reading-level story entitled “Johnny Appleseed” (Leslie & Caldwell, 2005). A combination of content and function words in the story were targeted so the peer could produce a specific form of simulated stuttering. Specifically, the speaker produced part- and whole-word repetitions that consisted of two unit repetitions (e.g., a-a-apples and he-he-he). Furthermore, her audible sound prolongations ranged from .58 s to .78 s with an overall average of .65 s across all three disfluent samples. After the training, the third author independently verified that the girl produced the correct form of stuttering on the target words. The accuracy and naturalness of the simulated stuttering also was evaluated by an independent group of three experienced clinicians, who confirmed that the girl’s stuttering was perceived as authentic stuttering behavior. Four audiovisual samples of the reading passage were recorded: (1) fluent (0% frequency of stuttering), (2) mild (5% stuttering frequency), (3) moderate (10% stuttering frequency), and (4) severe (15% stuttering frequency). All frequencies were calculated as a percentage of words stuttered. Every effort was made to have the peer speaker keep facial expressions, intonation, and vocal volume consistent across audiovisual samples. To assist the speaker in producing a conversational speech style during the taping, the passage was placed above and directly behind the camera. The placement of the passage also allowed for the speaker to have eye contact with the camera so that she could appear to be speaking in as natural way as possible. 2.3. Audiovisual samples The four speech samples for this study were recorded using a Sony DCR-HC46 mini-DV digital camera. The camera only showed the girl’s face and upper shoulder area. In addition, a plain, solid white background was used. Specifics about each sample are described below: 2.3.1. 0% sample In this audiovisual sample, the speaker read the Johnny Appleseed passage in a natural, conversational-style manner while exhibiting no stuttering. This sample was recorded in order to serve as a control condition against which the other disfluent samples could be judged. The total duration of this sample was 1 min 21 s. 2.3.2. 5% sample This sample contained 5% stuttering. In the passage, the words and, he, and of were selected for whole-word repetitions, apples, western, often, and Johnny for part-word repetitions, and Massachusetts, seeds, and walked for audible sound prolongations. The total duration of the sample was 1 min 36 s. 2.3.3. 10% sample A 10% frequency of stuttering was generated for this sample. In the passage, the words he, and, his, but, had, and of were selected for whole-word repetitions, apples, grow, western, went, often, planted, and Johnny for part-word repetitions, and Massachusetts, new, seeds, walked, and never for audible sound prolongations. The total duration of the sample was 1 min 45 s. 2.3.4. 15% sample The speaker produced a 15% frequency of stuttering for this sample. In the passage, the words he, and, of, his, to, but, and had were selected for whole-word repetitions, kinds, apples, apple, grow, people, western, went, crossed, often, planted, and Johnny for part-word repetitions, and Massachusetts, settlers, new, seeds, walked, and never for audible sound prolongations. The total duration of the sample was 1 min 50 s. 2.4. Procedure Each of the 88 participants was randomly assigned to watch one of the four audiovisual conditions so that there were 22 participants in each condition. A consistent set of instructions was used to ensure that the first and third authors provided the same instructions to each participant. The first or third author (both of whom do not stutter) showed the audiovisual sample to each participant using a portable DVD player. Each participant watched the audiovisual sample individually and was seated in a quiet room that posed no outside distractions. Specifically, each participant was given the following instructions:
J. Panico et al. / Journal of Fluency Disorders 45 (2015) 1–11
5
“I want you to watch this video of a girl telling a story. I’m not going to ask you to retell the story or ask you anything about the story. When the video is finished you will rate five statements and answer three general questions.” This facilitated more focused attention on the child’s speech rather than on the details of the story. They were also asked to let the first or third author know if they knew the girl and none of the participants acknowledged that he/she knew or had any contact with the speaker in the sample. After watching the audiovisual sample, each participant was given a sheet of paper that contained five statements that were to be rated on a 7-point Likert scale. The five statements were: (1) (2) (3) (4) (5)
This girl is a good speaker. This girl has smooth speech. This girl has an easy time telling the story. I felt comfortable listening to this girl. I would feel comfortable making friends with this girl.
A score of 1 corresponded to “strongly disagree” whereas a score of 7 corresponded to “strongly agree.” Each of the five statements contained a line-drawn representation of the emotion (frowning, neutral, and smiling face) that corresponded to the Likert rating of 1, 4, and 7, respectively in order to aid each participant’s comprehension of the Likert scale. Specifically, The Likert rating of 1 contained a frowning face (strongly disagree), a 4 contained a neutral face, and a rating of 7 contained a smiling face (strongly agree). The participants were also encouraged to ask questions if they were confused about the wording of a particular statement. After rating the Likert scale statements, each participant was asked three open-ended questions by the first or third authors, who were fluent speakers. These questions were similar to the ones used previously in the Susca and Healey (2001) study regarding listeners’ perceptions of a disfluent speaker. The questions were: (1) Tell me what you thought about the girl’s speech? (2) Was it easy or hard to listen to the way she talked? Why? (3) What did you notice about the way she talked? The participants’ responses were recorded using Marantz model #PMD660 digital recorder. If a participant provided a response that was brief or incomplete, the first or third author prompted the participants to add information by asking “Can you tell me more about that?” or “Is there anything else that you noticed?” After participants completed the study, they were asked if they had any questions and were also instructed to not discuss anything related to the study with their peers in order to prevent potential bias among future participants in the study. None of the participants reported that they were told anything about the study. 2.5. Data reliability and validity 2.5.1. Reliability Interjudge and intrajudge reliability were assessed for the coding of the positive/negative comment data. The first and third authors independently coded the data in the transcript for positive and negative comments. A unit-by-unit agreement ratio (Kazdin, 1982) was used to assess interjudge reliability. The interjudge reliability value was calculated to be 97%. In order to assess intrajudge reliability, after a two-week period the second author re-coded each verbal response in the transcript as positive or negative. An intrajudge reliability value of 93% for the positive/negative comment data was calculated using the same unit-by-unit agreement ratio. 2.5.2. Validity Validity of the qualitative data was established in this study through the use of three procedures described by Creswell (1998): (a) control of research bias, (b) triangulation to control for researcher bias, and (c) peer debriefing. First, research bias was controlled by limiting the amount of information regarding the nature of the study that was told to the participants. Participants were informed that that they were to watch and listen to a girl telling a story and that they would not be asked to retell the story or be asked anything about the story. The word “stuttering” was not used during the experiment in order to avoid bias toward the speaker in the speech sample. After each participant completed the study, they were asked not to discuss anything related to the study with their peers in order to prevent potential bias among future participants in the study. Second, a triangulation method involving the first and third authors as well a graduate student not involved with the study, was used to corroborate the findings obtained from the statement analysis. Using a random selection (through the use of a random number generator) of transcribed utterances, unanimous agreement was achieved for placing comments into the theme created for the data. A second-year graduate student (a different student from the one involved in triangulation) participated in the study as “peer debriefer.” The student was provided with 10% of the randomly selected statements from the transcript and asked to independently code each of the statements with one of the four possible theme clusters. The peer debriefer was in 100% agreement with the authors of the study in regard to her assignment of the randomly selected statements into appropriate theme clusters.
6
J. Panico et al. / Journal of Fluency Disorders 45 (2015) 1–11
3. Results The results of this study are presented in three segments: A one-way ANOVA of the Likert ratings for the five statements across the four conditions, a one-way ANOVA of the positive/negative comment data, and data from the qualitative analysis. This is followed by the descriptive statistics both the Likert and positive/negative comment data. 3.1. Data analysis Both quantitative and qualitative methods were used to analyze the data in this study. Participants’ responses to the open-ended questions were transcribed verbatim. Each verbal response in the transcript was individually coded as “positive,” “negative,” or “neutral.” A total of 463 statements were coded as positive if the main component of the statement conveyed a positive connotation (e.g., “She spoke very clearly,” and “It was easy to make sense of what she was saying and understand it”). Verbal responses were coded as negative when the main idea of the statement conveyed a negative connotation (e.g., “It was hard to listen to because she kept on stopping and repeating a couple of words”). Some statements consisted of both a positive component and a negative component (e.g., “The girl stuttered but knew the story very well”). These statements were divided into two parts, with the portion of the statement conveying a positive connotation coded as positive and the portion of the statement conveying a negative connotation coded as negative. Statements coded as neutral (e.g., “Her speech was OK”) were not included in the analysis because the focus was only on students’ positive and negative comments. After each verbal response was coded as positive, negative, or neutral, the percentage of positive and negative statements was calculated across the four different stuttering frequencies. The transcript was further analyzed to determine themes that emerged from what participants said about the audiovisual samples they observed using standard qualitative data analysis procedures (Creswell, 1998). 3.2. Likert ratings A one-way ANOVA was performed to determine the existence of significant differences across the five Likert statements in the one fluent speech and in the three stuttering frequency conditions. The results of the analysis revealed statistical significance for Likert statements 1, 2, and 3 but not for statements 4 (I felt comfortable listening to this girl) and 5 (I would feel comfortable making friends with this girl). Specifically, the significant difference were found for the statement, “This girl is a good speaker” was F(3, 84) = 6.128, p = .001, 2p = .180 and for the second statement, “This girl has smooth speech” [F(3, 84) = 6.041, p = .001, 2p = .177] across the four conditions. Statement 3 (This girl has an easy time telling the story) yielded a statistically significant difference as well, F(3, 84) = 8.509, p = <.001, 2p = .233. The partial eta-squared values (2p ) indicate small effect sizes for the three statements that yielded statistically significant differences (Cohen, 1988). Further post hoc testing (Tukey’s HSD) was performed to determine the location of significant differences between the stuttering frequency levels. Post hoc analyses revealed that participants’ ratings for Statement 1 were different between the fluent and the mild, moderate, and severe stuttering frequency samples, 0–5%, 0–10%, and 0–15%, respectively. Participants also rated Statement 2 differently for the fluent speech sample (0%) and the 15% severe stuttering sample. Furthermore, differences were found between the 0–10% and 0–15% conditions for Statement 3. No other statistically significant differences were found between the conditions. Descriptive statistics for each of the five Likert statements are shown in Table 1. The scores ranged from 1 to 5 for each speaking condition averaged across age groups, such that the larger the mean score, the more positive the perception. The mean scores for Statement 1 (This girl is a good speaker) ranged from 3.32 to 4.82 while the mean scores for Statement 2 (This girl has smooth speech) ranged from 2.36 to 4.36. For Statement 3 (This girl had an easy time telling the story), mean scores ranged from 2.86 to 5.00. Statement 4 (I felt comfortable listening to this girl) mean scores ranged from 4.14 to 5.32 and 4.18 to 5.00 for Statement 5 (I would feel comfortable making friends with this girl). These two statements (S4 and S5) elicited favorable ratings across the four conditions. Statement 3 within the 0% speech condition yielded the highest mean rating of 5.32 whereas the lowest mean rating of 2.36 was found in the 15% condition associated with Statement 2, which is expected given that listeners were evaluating the smoothness of the girl’s speech. 3.3. Qualitative analysis There were a total of 463 participant comments. Specifically, there were 101 comments in the 0% condition, 129 comments in the 5% condition, 123 comments in the 10% condition, and 110 comments in the 15% condition. The participants’ comments were transcribed and placed into theme clusters that best fit the nature of the comments. Four specific theme clusters emerged from the data and they were: (1) speech characteristics, (2) speaker characteristics, (3) listener comfort, and (4) listener comprehension. The following is an overview of each of the theme clusters including example comments from each. 3.3.1. Speech characteristics The speech characteristics theme cluster included comments related to various aspects of the speaker’s speech including speech rate, repetitions, and effort. For the fluent speech condition, sample comments included “She had good speech
J. Panico et al. / Journal of Fluency Disorders 45 (2015) 1–11
7
Table 1 Likert descriptive data. Likert statement
Sample (% stuttering)
M
SD
Statement 1
Control (0%) Mild (5%) Moderate (10%) Severe (15%)
4.82 3.55 3.36 3.32
1.53 1.10 1.00 1.64
Statement 2
Control Mild Moderate Severe
4.36 3.09 3.05 2.36
1.76 1.51 1.56 1.53
Statement 3
Control Mild Moderate Severe
5.00 4.27 3.36 2.86
1.38 1.70 1.36 1.64
Statement 4
Control Mild Moderate Severe
5.32 5.05 4.45 4.14
1.89 1.62 2.20 1.91
Statement 5
Control Mild Moderate Severe
4.73 5.00 4.59 4.18
2.10 1.54 2.44 1.92
Notes. The mean values are based on a 7-point Likert scale where higher values reflect more agreement with the given statement. Each sample contained 22 different participants for a total of 88 participants.
because she knew what she was saying,” “She seemed to have pretty good speech,” “She didn’t stumble on words,” and “Her words were pretty clear.” However, when the speech sample contained 5% stuttering, listeners reported more negative than positive comments. Examples of the negative comments were “It was a little rough,” “She said stuff over and over again,” and “She held the first letters and made words really long.” In the 10% condition, the comments were primarily negative such as “When she talked it was kind of bumpy,” “It was a little scratchy when she talked,” and “She would say part of a word a lot and then she would finally get to the end of it.” Finally, the 15% condition evoked comments including “It wasn’t really smooth because she had to keep repeating most of the words,” “She kept getting stuck on the words,” and “She kept saying words two or three times and some words she would hold out the sound.” 3.3.2. Speaker characteristics This theme cluster primarily dealt with the participants’ perceptions of the speaker’s comfort level in delivering the story. In the 0% condition, both positive and negative comments were made such as “I felt like she was pretty sure of what she was thinking,” “She seemed sure of herself because she knew a lot of facts about it,” and “She seemed kind of nervous while saying it,” respectively. One participant even commented, “It seemed like she was talking like a normal person would.” For the 5% condition, participants made comments including “From the video I think she had disabilities” and “It just seemed really hard for her to say some of the words,” and “She had a little bit of trouble but she is trying to do the best she can.” Participants also commented positively at this level including “She spoke very clearly,” and “Parts of the story she was really good at.” The 10% condition yielded primarily negative statements such as “She is probably not very good at talking,” “I don’t think she was really comfortable because when you stutter you’re not confident,” and “She talked kind of funny.” Within the 15%, one participant commented that “Maybe she has a disorder or something that doesn’t involve her communicating correctly with her speech.” Another participant stated, “She’s a good talker but sometimes she stutters because maybe she gets nervous or something.” Surprisingly, some participants commented positively on the speaker’s ability including “She was good at pronouncing words. She felt comfortable with what she was saying.” It was of interest as well to determine how many children used the word “stuttering” when referring to the speaker. A total of 11% (4 of 35) of the third graders (8–9 years old) and approximately 19% (5 of 27) of the fourth graders (9–10 years old) used the word “stuttering” when describing the speech of the speaker. This resulted in a combined total of 14.5% (9 out of 62) of the third and fourth graders used the word stuttering in a comment made about the peer who stuttered. 3.3.3. Listener comfort This cluster revolved around how comfortable the participants were in watching and listening to the speaker. Comments within this cluster consisted of positive and negative comments including “I thought it was easy to listen to because she enunciated well,” “She kept making the same sound which made it hard to listen to,” “It was fairly easy to listen to because she would tell me exactly what she was going to say,” “It was easy to listen to the way she talked,” and “It was hard to listen to because she kept on stopping and repeating a couple of words.” This pattern of positive and negative comments was consistent for all four stuttering frequency conditions.
8
J. Panico et al. / Journal of Fluency Disorders 45 (2015) 1–11
100 90 80
Positive Negative
Percent
70 60 50 40 30 20 10 0
0%
5%
10%
15%
Speech Sample Fig. 1. Percentage of positive and negative comments across the three open-ended questions for each of the four stuttering frequency conditions.
3.3.4. Listener comprehension The final theme cluster related to how well participants were able to follow the story being told by the speaker. The 0% condition consisted mostly of positive comments such as “I could understand all of her words” and “It was easy to understand her because I could follow all of her words.” On participant commented, “I could understand what she was saying because I have five cousins that talk like that.” In the 5% and 10% conditions participants commented that “When she said the whole word I could understand her” and “It was only the first couple of letters that were messed up but the rest just went through it so I could understand it.” Others stated “Sometimes her speech wasn’t smooth so it was hard to understand her,” and “I understood the story just fine.” Finally, in the 15% condition several participants made comments such as “It was easy to make sense of what she was saying and understand it,” and “Other than some bumps on her words the story was easy for me to understand” whereas other participants had some perceived difficulty with understanding the story as evidenced by comments such as “It was a little difficult to understand what she was saying,” “Not too easy because she kept getting stuck on words,” and “It was just hard to understand her because of her speech.” There were more negative comments for the 15% condition than the other three conditions. 3.3.5. Positive/negative comment data A one-way ANOVA was performed to determine if the percentage of positive comments made by participants across the four conditions were significantly different. Overall, the analysis revealed statistical significance, F(3, 84), p = .000, 2p = .403. Post hoc testing (Tukey’s HSD) was then conducted to determine the location of the significant differences between frequency conditions. Specifically, the percentage of positive comments was significantly higher between the 0% and 5%, the 0% and 10%, the 0% and 15% conditions. No other statistically significant differences were found between the other conditions. Fig. 1 shows the percentage of positive and negative comments made across the three open-ended questions. Overall, there is a clear pattern of decreasing positive comments and increasing negative comments as the stuttering frequency increased from 0% to 15%. Specifically, the 0% condition yielded the highest percentage of positive comments (67%/33% negative). In the 5% condition, the total percentage of positive comments decreased to 38% and there were 62% negative comments. In the 10% condition the percentage of positive comments decreased to 34% versus 66% negative comments. The 15% condition yielded 28% positive and 72% negative comments. 4. Discussion This study sought to examine elementary school-age children’s perceptions of stuttering using quantitative and qualitative research methodology. To our knowledge, this is the first study to use this methodological approach to determine how typically fluent children perceive an elementary school-age peer who stutters across three different frequencies of stuttering. The results are based on children who did not have extensive contact or exposure to other children who stutter in order to minimize any biasing that might have taken place in the participants’ responses. One of the major findings, derived from the children’s ratings for specific statements and many of the qualitative comments children made about the speaker, showed that the ratings decreased to become less favorable as the frequency of stuttering increased from 5% to 15%. This was particularly true for statements 1, 2, and 3. Similar results were found for
J. Panico et al. / Journal of Fluency Disorders 45 (2015) 1–11
9
students’ ratings for statements 1, 2 and 3 used in the Evans et al. (2008) study. This finding suggests that children as young as 8 years of age (third grade) react negatively to a child’s speech that is not smooth and when a child has difficulty telling a story. However, on a more positive note, there were no significant differences in children’s ratings across the four conditions when they were asked to rate two statements related to the degree of comfort they felt listening to the girl speak and how comfortable they felt making friends with the girl. The fact that children in this study reported feeling comfortable with a peer who stutters is encouraging for even at a young age, children seem to be able to separate how they feel about a child who stutters and that listening to stuttering is not uncomfortable. This is consistent with past studies by Healey, Gabel, Daniels, and Kawai (2007), Panico et al. (2005), Evans et al. (2008), and Susca and Healey (2001) who also showed that the major of listeners are comfortable listening to stuttering and are comfortable being a friend with a person who stutters. Of course being comfortable with or friendly toward those who stutter does not necessarily mean that listeners do not react negatively to stuttering behavior. As shown in this study, the number of negative qualitative comments children made toward the disfluent peer gradually increased as the stuttering frequency increased. There was a steady rise in the number of negative comments from the 5% to the 15% samples relative to (1) the frequency of stuttering, (2) how difficult it was for the listeners to listen to the girl stutter, and (3) how difficult it was to follow what the peer said. The finding is similar to the data reported by Susca and Healey (2001) for adults who stutter. A third finding of the present study related to the percentage of children in the third and fourth grades who used the term “stuttering” in their comments about the peer’s speech. Culatta and Sloan (1977) reported that 37% of their third and fourth grade participants used the word “stuttering” when referring to the adult’s speech. In the present study, only 14% of the total number of third and fourth graders used the label “stuttering.” This lower percentage of students who used “stuttering,” compared to the Culatta and Sloan study might suggest that children today could have been taught by parents, teachers, or others that it is unacceptable to use labels or negative terms when referring to adults or children. This current trend might be related to children in schools being made aware of and learning about children who have speech, language, hearing, cognitive, social, and physical disabilities. The term “disabilities” might also reflect what children are exposed to more often in today’s society. In fact, one child stated that, “From the video, I think she (the peer speaker) had disabilities.” In the context of the present study, it is interesting that the child used the term “disabilities” rather than “stuttering.” It is possible that the comments children made to the first and third authors about the girl’s speech was influenced by responding to an adult. It might not be polite or “correct” to use a label with an adult but it might be if a child was talking to another child. The results of the present study show by the time children reach third grade, they are developing negative perceptions and reactions to children who stutter (Evans et al., 2008; Franck et al., 2003; Hartford & Leahy, 2007; Langevin & Hagler, 2004). The relative consistency of this finding across research studies shows that stuttering is perceived negatively by the time children reach third grade. This suggests schools need to continue their efforts to help all children learn to appreciate the challenges children who stutter face in coping with the disorder. 4.1. Limitations of the study and future research Most of the children used in the present study were Caucasian (90%) with only a few minorities included in the sample. Future studies in this area also could include a more multi-culturally diverse sample of listeners given the influence social, cultural, and experiential factors have on perceptions of speakers. Studies have shown that negative perceptions and attitudes exist culturally and geographically (Abdalla & St Louis, 2011; Mayo, Mayo, Jenkins, & Graves, 2004). Thus, the findings of this study might not generalize to children from other cultural and ethnic backgrounds. It would also be informative to examine the presence of secondary coping behaviors within the speech samples to determine how those behaviors might impact listeners’ perceptions of stuttering. It is possible that eye closure, head turns, or other types of facial grimaces present during a disfluent moment might change how children perceive stuttering from a peer. This might have been the case with the participant responses in the Langevin and Hagler (2004) study when rating the disfluent speakers who had moderate and severe stuttering and accompanying secondary behaviors. It would be of interest if some form of struggle behavior and effort accompanying a low frequency level of stuttering could elicit more negative reactions than those found in the present study. Finally, it would also be of interest to investigate children’s ability to recall and comprehend information after listening to varying frequencies of stuttering spoken by a peer who stutters. There is some evidence that stuttering behavior influences recall and comprehension of information when adults listen to stories told by speakers who exhibit mild to severe stuttering (Panico & Healey, 2009) but examining whether the same effect would occur for children may also yield useful findings. Another limitation to this study was the extensiveness of the qualitative portion of the study. Although the qualitative methods used in the present study were similar to other studies examining listener perceptions of stuttering, one could argue that a greater number and broader range of questions could have been used. This might have led to additional or different types of thematic categories of the statements and interpretation of the positive and negative comments made about the three levels of stuttering. 5. Conclusions The purpose of the present study was to further investigate children’s perceptions of a peer who stutters using a mixed methodology. Quantitatively, the results indicated that perceptions become more negative as stuttering frequency increased
10
J. Panico et al. / Journal of Fluency Disorders 45 (2015) 1–11
which is consistent with previous research with older children and adults. The qualitative data complimented the quantitative findings and showed that participants commented on several aspects of the speakers’ speech pattern including rate, articulation, and effort in addition to the stuttering. By the time children reach third grade, they appear to be developing negative perceptions and reactions toward children who stutter. This suggests that children who stutter need emotional support from parents, SLPs, and teachers in helping them cope with the disorder. On a positive note, the findings of the present study revealed that although children were aware of the stuttering they still indicated a willingness to be friends with the peer. This finding should be encouraging to children who stutter in that his or her peers would accept them as friends. The findings of this study together with past research in this area should help clinicians and their clients appreciate the range of social and emotional reactions peers have toward a child who stutters.
Financial disclosures James Panico, E. Charles Healey, and Jennifer Knopik did not disclose any relevant financial relationships in support of the research reported in this article.
Non-financial disclosures James Panico, E. Charles Healey, and Jennifer Knopik did not disclose any relevant non-financial relationships used in support of the research reported in this article. CONTINUING EDUCATION Elementary school students’ perceptions of stuttering: A mixed model approach QUESTIONS (1) In general, past research that has examined children’s perceptions of stuttering has revealed which of the following? a. preschool-aged children are able to differentiate between fluent and stuttered speech b. the term “stuttering” was used overwhelmingly by children in third and fourth grades when listening to a disfluent speaker c. children responded positively to the disfluent speech produced by adult speakers d. findings have yielded significant differences toward perceptions of boys vs. girls who stutter e. that older children tend to respond with less negative reactions to stuttered speech (2) Comments related to rate, vocal characteristics, and effort are part of which theme cluster: a. speaker characteristics b. communication pragmatics c. speech characteristics d. listener comfort e. story comprehension (3) The features of stuttering used in the present study to create the samples were comprised of: a. dysrhythmic part- and whole-word repetitions b. two-unit part- and whole-word repetitions along with audible sound prolongations c. inaudible sound prolongations lasting approximately two seconds each in duration d. commensurate secondary behaviors e. three-unit repetitions along with audible sound prolongations lasting three seconds each in duration (4) Which of following is true regarding the findings of the present study? a. the number of negative qualitative comments gradually decreased as the stuttering frequency increased b. approximately half of the third and fourth grade children use the term “stuttering” when commenting about the speech of the speaker c. there were significant differences across the four conditions when asked to rate the level of comfort in listening to the speaker or making friends with the speaker d. the participants commented on several aspects of the speaker speech in addition to the presence of stuttering e. although the children were aware of the stuttering they did not indicate a willingness to become friends with the peer (5) Based on the findings of the present study, future research should: a. incorporate a more multiculturally diverse sample of participants b. create additional audio samples with higher frequencies of stuttering c. collect additional information on adult’s recall and comprehension when listening/viewing a disfluent sample d. not include secondary behaviors as they may cause the listener to focus more on the visual signal versus the stuttering moment itself e. solely on the use of quantitative methodology to examine listener perceptions of stuttering
J. Panico et al. / Journal of Fluency Disorders 45 (2015) 1–11
11
References Abdalla, F. A., & St Louis, K. O. (2011). Arab school teachers’ knowledge, beliefs and reactions regarding stuttering. Journal of Fluency Disorders, 37, 54–69. Ambrose, N., & Yairi, E. (1994). The development of awareness of stuttering in preschool children. Journal of Fluency Disorders, 19, 229–245. Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum. Creswell, J. W. (1998). Qualitative inquiry and research design. Thousand Oaks, CA: Sage Publications Inc. Culatta, R., & Sloan, A. (1977). The acquisition of the label stuttering by primary level schoolchildren. Journal of Fluency Disorders, 2, 29–34. Evans, D., Healey, E. C., Kawai, N., & Rowland, S. (2008). Middle school students’ perceptions of a peer who stutters. Journal of Fluency Disorders, 33, 203–219. Ezrati-Vinacour, R., Platzky, R., & Yairi, E. (2001). The young child’s awareness of stuttering-like disfluency. Journal of Speech, Language, and Hearing Research, 44, 368–380. Franck, A. L., Jackson, R. A., Pimentel, J. T., & Greenwood, G. S. (2003). School-age children’s perceptions of a person who stutters. Journal of Fluency Disorders, 28, 1–15. Giolas, T. G., & Williams, D. E. (1958). Children’s reactions to nonfluencies in adult speech. Journal of Speech and Hearing Research, 1, 86–93. Guitar, B. (2006). Stuttering: An integrated approach to its nature and treatment. Philadelphia: Lippincott, Williams & Wilkins. Hartford, E., & Leahy, M. M. (2007). The perceptions of primary school children of a person who stutters. In J. Au-Yeung, & M. M. Leahy (Eds.), Proceedings of the 5th world congress on fluency disorders (pp. 223–229). Dublin, Ireland: The International Fluency Association. Healey, E. C., Gabel, R. M., Daniels, D. E., & Kawai, N. (2007). The effects of self-disclosure and non-self-disclosure on listeners’ perceptions of a person who stutters. Journal of Fluency Disorders, 32, 51–69. Kazdin, A. E. (1982). Single case research designs: Methods for clinical and applied settings. New York: Oxford University Press. Langevin, M., & Hagler, P. (2004). Development of a scale to measure peer attitudes toward children who stutter. In A. K. Bothe (Ed.), Evidence-based treatment of stuttering: Empirical bases and clinical applications (pp. 139–171). Mahwah, NJ: Erlbaum. Leslie, L., & Caldwell, J. S. (2005). Qualitative reading inventory-4. Boston: Pearson. Mayo, R., Mayo, C., Jenkins, K., & Graves, L. (2004, September). Public knowledge of stuttering: Cross cultural perspectives. Speech Pathology.com. Retrieved from. http://www.speechpathology.com/articles/public-knowledge-stuttering-cross-cultural-1457 Panico, J., & Healey, E. C. (2009). The influence of text type, topic familiarity, and stuttering severity on listener recall, comprehension, and mental effort. Journal of Speech, Language, and Hearing Research, 52, 534–546. Panico, J., Healey, E. C., Brouwer, K., & Susca, M. (2005). Listener perceptions of stuttering across two presentation modes: A quantitative and qualitative approach. Journal of Fluency Disorders, 30, 65–85. Riley, G. D. (1994). Stuttering severity instrument for children and adults (3rd ed.). Austin, TX: Pro-ed. Riley, G. D. (2009). Stuttering severity instrument for children and adults (4th ed.). Austin TX: Pro-Ed. Susca, M., & Healey, E. C. (2001). Perceptions of simulated stuttering and fluency. Journal of Speech, Language, and Hearing Research, 44, 61–72.
James Panico is currently an associate professor at SIUE where he teaches undergraduate and graduate courses in fluency disorders, clinical methods, counseling, and professional issues. His research area of interest involves perceptions of stuttering. Dr. Panico has presented at professional meetings at the state, national, and international levels. Charlie Healey is a professor emeritus from the University of Nebraska-Lincoln. He has published extensively in the area of stuttering. Furthermore, he has given numerous presentations at the state, national, and international levels. Jennifer Knopik completed her undergraduate and graduate degree in speech-language pathology at SIUE. She has published and presented in the area of stuttering. Currently she is a speech-language pathologist for a private rehabilitation company.
Contents lists available at ScienceDirect
Journal of Fluency Disorders
Elementary school students’ perceptions of stuttering: A mixed model approach James Panico a,∗ , E. Charles Healey b,1 , Jennifer Knopik c a b c
Southern Illinois University Edwardsville, 1101 Founders Hall, Edwardsville, IL 62026, USA University of Nebraska-Lincoln, 253 Barkley Memorial Center, Lincoln, NE 68503, USA Southern Illinois University Edwardsville, Edwardsville, IL 62026, USA
a r t i c l e
i n f o
Article history: Received 25 February 2015 Received in revised form 11 June 2015 Accepted 14 June 2015 Available online 28 June 2015 Keywords: Stuttering Elementary Students Perceptions
a b s t r a c t Purpose: Past research studies have focused on perceptions of stuttering by various age groups and only a few have examined how children react to a peer who stutters. All of these studies used a quantitative analysis but only one included a qualitative analysis of elementary school age children’s responses to stuttering. The aim of this study was to further explore the perceptions of elementary school students toward a peer who stutters using both quantitative and qualitative analyses of three levels of stuttering. Methods: Participants included 88 elementary school children between 8 and 12 years of age. Each participant viewed one of four audiovisual samples of a peer producing fluent speech and mild, moderate, and severe simulated stuttering. Each participant then rated five Likert statements and answered three open-ended questions. Results: Quantitative and qualitative results indicated that negative ratings and the percentage of negative comments increased as the frequency of stuttering increased. However, the children in this study indicated that they were comfortable listening to stuttering and would be comfortable making friends with the peer who stutters. Conclusion: The findings of this study together with past research in this area should help clinicians and their clients appreciate the range of social and emotional reactions peers have of a child who stutters. Educational objectives: After reading this article, the reader will be able to: (a) discuss past research regarding children’s perceptions of stuttering; (b) summarize the need to explore the perceptions of elementary-aged children toward a peer who stutters; (c) describe the major quantitative and qualitative findings of children’s perceptions of stuttering; and (d) discuss the need for disseminating more information about stuttering to children and teachers. © 2015 Elsevier Inc. All rights reserved.
1. Introduction For approximately half a century, researchers have been interested in how school aged children react to stuttering. Knowing how children react to stuttering has important clinical implications about how children who stutter might be
∗ Corresponding author. Tel.: +1 618 650 5838. E-mail addresses: [email protected] (J. Panico), [email protected] (E.C. Healey). 1 Tel.: +1 402 304 1819. http://dx.doi.org/10.1016/j.jfludis.2015.06.001 0094-730X/© 2015 Elsevier Inc. All rights reserved.
2
J. Panico et al. / Journal of Fluency Disorders 45 (2015) 1–11
treated by their peers in and outside of the school setting as well as how they manage social relationships involving teasing and bullying (Langevin & Hagler, 2004). Unfortunately, there are only a few empirical studies on this topic. Giolas and Williams (1958) conducted the first study in this area to determine how a total of 120 five- and eight-year-old children reacted to audio samples of three adult speakers producing fluent and imitated stuttered speech as judged by three trained speech-language pathologists. They discovered that children as young as five years-of-age showed awareness of disfluent speech and reacted negatively when hearing part-word repetitions in the adult’s speech. The authors reported that the group of eight year-old children used the label “stuttering” when they heard an adult producing 10% part-word repetitions during the speech sample. Five-year-old children in the study did not use the word stuttering after hearing the disfluent speech. Culatta and Sloan conducted a similar study in 1977 and asked sixty children in the first through fourth grades (six to nine years-of-age) to listen to audio samples of stuttering and fluent speech produced by an adult speaker. The method by which the stuttering samples were created was not specified. The children were divided into two groups of 30: (1) first and second graders, and (2) third and fourth graders. As expected, both age groups preferred the fluent speech sample over the one that contained the stuttering. They also found that the first and second grade children did not use the term “stuttering” to refer to the disfluent speech but “stuttering” was used by about 37% of the third and fourth graders. Ambrose and Yairi (1994) examined perceptions of stuttering of 20 normally fluent and 20 disfluent children between the ages of two and five by asking them to choose between two types of speech produced by puppets; one who spoke fluently and the other who stuttered. The children who stuttered had to meet specific criteria set forth by the authors for inclusion in the study. The disfluent children who were three years of age identified closely with the puppet who stuttered while the fluent speaking three-year-olds selected the fluent puppet as the one that resembled how they spoke. This study showed that the awareness of stuttering begins at an early age and preschoolers are capable of differentiating between fluent and stuttered speech. These findings were replicated by Ezrati-Vinacour, Platzky, and Yairi (2001) who also noted that the term “stuttering” was not used by children under five years of age. Only a few of the seven year-old children labeled the disfluent speech as “stuttering.” Several reports have indicated that older children (9–14 years) respond with more negative reactions to stuttered speech. Franck, Jackson, Pimentel, and Greenwood (2003) had 75 9- to 11-year-old children listen to one fluent speech sample and one containing moderate stuttering from an adult who stuttered. The speakers’ stuttering severity level was determined through a Stuttering Severity Instrument -3 (Riley, 1994) rating from three experienced speech-language pathologists. Using a series of adjective pairs related to personality (outgoing–shy) and intelligence (intelligent–stupid), one group of children rated the fluent speaker and another group rated the speaker who stuttered. The results showed that fourth- and fifth-grade children had more negative perceptions of an adult who stuttered than the adult who did not stutter. Hartford and Leahy (2007) selected 80 children between six and 13 years-of-age and asked them to evaluate an audio sample of an adult speaker who produced fluent as well as moderate to severe pseudostuttering. The severity level of the pseudostuttering was determined by 22 speech-language pathology students. Using quantitative and qualitative measures, the researchers found that the six to eight year-old-children made fewer negative comments about the disfluent speaker than did children 9–13 years-of-age. Nonetheless, many of the children commented that the person who stuttered appeared “nervous,” “shy,” and “wouldn’t have many friends.” The consistent finding that elementary school-age children harbor negative perceptions of an adult who stutters provides evidence that children tend to react negatively toward a disfluent speaker at an early age. Langevin and Hagler (2004) were the first to study children’s perception of a disfluent peer relative to their beliefs, feelings, and attitudes toward a child who stutters. The major intent of their study was to create a scale to measure peer attitudes toward children who stutter. A sample of 267 elementary school children aged 8–13 years (grades four to six) were asked to rate 40 statements on a scale of 1–5 about a nine-year-old boy and an eight-year-old girl who stuttered. The boy stuttered on 16.5% and the girl stuttered on 43% of syllables produced in the sample. Each child also exhibited a variety of secondary coping behaviors such as head movements, broken eye contact, and rapid jaw jerking. Participants viewed one-minute samples of the boy and girl who stuttered and then rated statements relative to their affective (e.g., I would like having a kid who stutters live next door to me), behavioral (e.g., I would walk in the hall with a kid who stutters), and cognitive (e.g., kids who stutter are like normal kids) intentions toward children who stutter. The results of their study showed that fifth and sixth graders tended to have more positive perceptions of the two peers who stuttered than the fourth graders. They also found no differences in listeners’ ratings between the boy and girl who stuttered or between the boys and girls who served as participants. Additional results from the study focused on verifying the reliability and validity of the scale they had developed. Continued research into children’s and adolescents’ perceptions of stuttering was conducted by Evans, Healey, Kawai, and Rowland (2008). They examined the reactions and perceptions of 64 students between the ages of 10 and 14 toward a 16 year-old male peer who exhibited a very mild stuttering severity but was able to simulate a variety of stuttering types with a high level of control. The students were divided into four groups. Each group only listened to one type of video sample of the male peer speaking fluently or with simulated mild, moderate, or severe stuttering. All listeners were asked to rate 11 Likert statements using a five-point scale about the speaker shown in the video. The results of this study were similar to past studies involving children who rated adult speech samples in that, perceptional ratings of a person who stutters become increasingly more negative as the stuttering frequency increases. Additionally, the majority of the listeners reacted negatively even to mild stuttering although some students reported they were comfortable listening to and interacting with a peer who stuttered mildly. This finding implies that a peer with mild stuttering would be more likely to “fit in” at school than a peer who had moderate to severe stuttering.
J. Panico et al. / Journal of Fluency Disorders 45 (2015) 1–11
3
Although the Langevin and Hagler (2004) study provided the first quantitative investigation of children’s reactions to an elementary school age boy and girl who stutter, it is limited in that it did not provide any data concerning verbal statements or descriptions about how the children reacted to the samples of stuttering they heard. Eliciting peoples’ opinions, beliefs, and perspectives about a phenomenon (i.e., a qualitative research method) can provide a personal point of view of a given situation (Creswell, 1998). Because of the variety of ways stuttering behavior could be perceived, it would seem important to obtain children’s reactions to stuttering using a mixed model approach (i.e., combined quantitative and qualitative methods). Mixed methods has been used previously in the research by Susca and Healey (2001) and Panico, Healey, Brouwer, and Susca (2005) when investigating adults’ perceptions of adults who stutter, as well as Evans et al. (2008) who studied how children and adolescents react to a peer who stutters. The focus of these studies was to have listener’s rate statements about a person who stuttered of a scale of 1–5 as well as ask them to comment on their comfort in listening to stuttering, how they perceived the person who stuttered, and specific features about the way the person spoke. Another limitation of the Langevin and Hagler (2004) study is that they had their listeners rate only who stuttered moderately and severely. They stated that additional research is needed using children who have mild stuttering. Thus, a study that analyzes how elementary aged children react to a peer who has mild, moderate, or severe stuttering could expand the perception of stuttering across three levels of stuttering and provide a comparison of results to the Langevin and Hagler study. It could also provide insights for children who stutter about how peers might react to stuttering and why they might encounter teasing and bullying about the way they speak. The purpose of the present study was to address two research questions about children’s perceptions of stuttering: (1) How do children rate a peer who stutters relative to five statements (using a 7-point scale) about the peer’s speech? and (2) What types of comments do children produce when responding to a series of questions (e.g., how smooth was her speech? Was she a good speaker?) about a peer who speaks fluently as well as with mild, moderate, and severe simulated stuttering? 2. Methods 2.1. Participants A total of 88 children participated in this study and were recruited within the Edwardsville, Illinois (USA) community from the YMCA Kid’s Network Latchkey afterschool program. The participants ranged in age from 8 to 12 years and were in the third (n = 35), fourth (n = 27), or fifth (n = 26) grade. The study included 49 boys, ranging in age from 8 to 12 years (M = 9.85), and 39 girls, ranging in age from 8 to 11 years (M = 9.46). Each participant passed a hearing screening at 25 dB HL across 500, 1 K, 2 K, and 4 K frequencies using a Maico M39 portable audiometer. In order to be included in the study, the parent/guardian of each participant confirmed to the first author that his/her child had not previously received or was currently receiving speech and/or language therapy services. The participants’ parents/guardians also confirmed that his or her child had little to no knowledge, familiarity of, or contact with children who stutter. These two criteria were used to prevent bias among the participants when listening to stuttered speech if a child had extensive contact or exposure to a child who stutters. Langevin and Hagler (2004) found that children who had contact with a child who stutters have significantly more positive attitudes toward a child who stutters than children who don’t have contact with a child who stutters. 2.2. Construction of speech samples In order to control for confounding variables such as the type and duration of the disfluency, the speech samples were composed of simulated stuttering. A pilot study was conducted in order to determine if the types and number of simulated disfluencies were representative of mild, moderate, and severe stuttering. Five graduate students who had completed a graduate course in fluency disorders and were trained in the assessment of stuttering measures were asked to independently listen to three passages in which the first author read the passages in a conversational manner while incorporating simulated stuttering into each of the readings. The reading passage was taken from the Stuttering Severity Instrument-4 (Riley, 2009). The first author simulated the core behaviors of stuttering, including part-word repetitions, whole-word repetitions, and audible sound prolongations (Guitar, 2006) in an easy, rhythmical, and relaxed manner. The samples did not contain any concomitant behaviors during the simulated stuttering moments, such as facial or head movements. Because simulated secondary coping behaviors could have varied from one sample to the next, the authors chose not to include them for any speech sample. After watching each of the three audiovisual samples in a randomized order, each of the five graduate students was then asked to rate the sample from 1 to 3 (where 1 = mild stuttering; 2 = moderate stuttering; and 3 = severe stuttering). All five of the students independently classified the sample containing 5% stuttering in the forms above as having the least stuttering and was rated as mild stuttering. There was also unanimous agreement that the sample containing 10% stuttering was considered moderate in severity, and the sample containing 15% stuttering was rated as severe stuttering. This confirmed that the type and form of stuttering produced in the pilot study could be used as model forms of stuttering for the peer speaker to imitate. These percentages of stuttering (calculated as a percentage of words stuttered) and durations of stuttering words in the speech samples of the present study are consistent with stuttering severity classifications developed in previous studies (Evans et al., 2008; Susca & Healey, 2001).
4
J. Panico et al. / Journal of Fluency Disorders 45 (2015) 1–11
Upon completion of the pilot study, a total of four audiovisual speech samples were created. The samples consisted of the stuttering frequencies levels (5, 10, and 15%) tested in the pilot study along with a 0% frequency level, which served as a control condition. The peer speaker for the four audiovisual speech samples was a nine-year-old girl who was fluent and never had a history of stuttering. The girl was similar in age to the target age group of participants, was home schooled, and did not live in the same town as any of the participants. In addition, there were no anticipated problems using a girl as the speaker because one result from Langevin and Hagler (2004) showed no significant perceptual differences between a boy vs. a girl who stuttered. The girl in the present study willingly volunteered to participate and agreed to be trained by the first author to produce the same type and form of part-word repetitions, whole-word repetitions, and audible sound prolongations used in the pilot study. An extensive amount of time was spent instructing the girl and allowing her to produce and practice the three target forms of stuttering as many times as she wanted during oral reading of an experimental passage. The passage was a fourth grade reading-level story entitled “Johnny Appleseed” (Leslie & Caldwell, 2005). A combination of content and function words in the story were targeted so the peer could produce a specific form of simulated stuttering. Specifically, the speaker produced part- and whole-word repetitions that consisted of two unit repetitions (e.g., a-a-apples and he-he-he). Furthermore, her audible sound prolongations ranged from .58 s to .78 s with an overall average of .65 s across all three disfluent samples. After the training, the third author independently verified that the girl produced the correct form of stuttering on the target words. The accuracy and naturalness of the simulated stuttering also was evaluated by an independent group of three experienced clinicians, who confirmed that the girl’s stuttering was perceived as authentic stuttering behavior. Four audiovisual samples of the reading passage were recorded: (1) fluent (0% frequency of stuttering), (2) mild (5% stuttering frequency), (3) moderate (10% stuttering frequency), and (4) severe (15% stuttering frequency). All frequencies were calculated as a percentage of words stuttered. Every effort was made to have the peer speaker keep facial expressions, intonation, and vocal volume consistent across audiovisual samples. To assist the speaker in producing a conversational speech style during the taping, the passage was placed above and directly behind the camera. The placement of the passage also allowed for the speaker to have eye contact with the camera so that she could appear to be speaking in as natural way as possible. 2.3. Audiovisual samples The four speech samples for this study were recorded using a Sony DCR-HC46 mini-DV digital camera. The camera only showed the girl’s face and upper shoulder area. In addition, a plain, solid white background was used. Specifics about each sample are described below: 2.3.1. 0% sample In this audiovisual sample, the speaker read the Johnny Appleseed passage in a natural, conversational-style manner while exhibiting no stuttering. This sample was recorded in order to serve as a control condition against which the other disfluent samples could be judged. The total duration of this sample was 1 min 21 s. 2.3.2. 5% sample This sample contained 5% stuttering. In the passage, the words and, he, and of were selected for whole-word repetitions, apples, western, often, and Johnny for part-word repetitions, and Massachusetts, seeds, and walked for audible sound prolongations. The total duration of the sample was 1 min 36 s. 2.3.3. 10% sample A 10% frequency of stuttering was generated for this sample. In the passage, the words he, and, his, but, had, and of were selected for whole-word repetitions, apples, grow, western, went, often, planted, and Johnny for part-word repetitions, and Massachusetts, new, seeds, walked, and never for audible sound prolongations. The total duration of the sample was 1 min 45 s. 2.3.4. 15% sample The speaker produced a 15% frequency of stuttering for this sample. In the passage, the words he, and, of, his, to, but, and had were selected for whole-word repetitions, kinds, apples, apple, grow, people, western, went, crossed, often, planted, and Johnny for part-word repetitions, and Massachusetts, settlers, new, seeds, walked, and never for audible sound prolongations. The total duration of the sample was 1 min 50 s. 2.4. Procedure Each of the 88 participants was randomly assigned to watch one of the four audiovisual conditions so that there were 22 participants in each condition. A consistent set of instructions was used to ensure that the first and third authors provided the same instructions to each participant. The first or third author (both of whom do not stutter) showed the audiovisual sample to each participant using a portable DVD player. Each participant watched the audiovisual sample individually and was seated in a quiet room that posed no outside distractions. Specifically, each participant was given the following instructions:
J. Panico et al. / Journal of Fluency Disorders 45 (2015) 1–11
5
“I want you to watch this video of a girl telling a story. I’m not going to ask you to retell the story or ask you anything about the story. When the video is finished you will rate five statements and answer three general questions.” This facilitated more focused attention on the child’s speech rather than on the details of the story. They were also asked to let the first or third author know if they knew the girl and none of the participants acknowledged that he/she knew or had any contact with the speaker in the sample. After watching the audiovisual sample, each participant was given a sheet of paper that contained five statements that were to be rated on a 7-point Likert scale. The five statements were: (1) (2) (3) (4) (5)
This girl is a good speaker. This girl has smooth speech. This girl has an easy time telling the story. I felt comfortable listening to this girl. I would feel comfortable making friends with this girl.
A score of 1 corresponded to “strongly disagree” whereas a score of 7 corresponded to “strongly agree.” Each of the five statements contained a line-drawn representation of the emotion (frowning, neutral, and smiling face) that corresponded to the Likert rating of 1, 4, and 7, respectively in order to aid each participant’s comprehension of the Likert scale. Specifically, The Likert rating of 1 contained a frowning face (strongly disagree), a 4 contained a neutral face, and a rating of 7 contained a smiling face (strongly agree). The participants were also encouraged to ask questions if they were confused about the wording of a particular statement. After rating the Likert scale statements, each participant was asked three open-ended questions by the first or third authors, who were fluent speakers. These questions were similar to the ones used previously in the Susca and Healey (2001) study regarding listeners’ perceptions of a disfluent speaker. The questions were: (1) Tell me what you thought about the girl’s speech? (2) Was it easy or hard to listen to the way she talked? Why? (3) What did you notice about the way she talked? The participants’ responses were recorded using Marantz model #PMD660 digital recorder. If a participant provided a response that was brief or incomplete, the first or third author prompted the participants to add information by asking “Can you tell me more about that?” or “Is there anything else that you noticed?” After participants completed the study, they were asked if they had any questions and were also instructed to not discuss anything related to the study with their peers in order to prevent potential bias among future participants in the study. None of the participants reported that they were told anything about the study. 2.5. Data reliability and validity 2.5.1. Reliability Interjudge and intrajudge reliability were assessed for the coding of the positive/negative comment data. The first and third authors independently coded the data in the transcript for positive and negative comments. A unit-by-unit agreement ratio (Kazdin, 1982) was used to assess interjudge reliability. The interjudge reliability value was calculated to be 97%. In order to assess intrajudge reliability, after a two-week period the second author re-coded each verbal response in the transcript as positive or negative. An intrajudge reliability value of 93% for the positive/negative comment data was calculated using the same unit-by-unit agreement ratio. 2.5.2. Validity Validity of the qualitative data was established in this study through the use of three procedures described by Creswell (1998): (a) control of research bias, (b) triangulation to control for researcher bias, and (c) peer debriefing. First, research bias was controlled by limiting the amount of information regarding the nature of the study that was told to the participants. Participants were informed that that they were to watch and listen to a girl telling a story and that they would not be asked to retell the story or be asked anything about the story. The word “stuttering” was not used during the experiment in order to avoid bias toward the speaker in the speech sample. After each participant completed the study, they were asked not to discuss anything related to the study with their peers in order to prevent potential bias among future participants in the study. Second, a triangulation method involving the first and third authors as well a graduate student not involved with the study, was used to corroborate the findings obtained from the statement analysis. Using a random selection (through the use of a random number generator) of transcribed utterances, unanimous agreement was achieved for placing comments into the theme created for the data. A second-year graduate student (a different student from the one involved in triangulation) participated in the study as “peer debriefer.” The student was provided with 10% of the randomly selected statements from the transcript and asked to independently code each of the statements with one of the four possible theme clusters. The peer debriefer was in 100% agreement with the authors of the study in regard to her assignment of the randomly selected statements into appropriate theme clusters.
6
J. Panico et al. / Journal of Fluency Disorders 45 (2015) 1–11
3. Results The results of this study are presented in three segments: A one-way ANOVA of the Likert ratings for the five statements across the four conditions, a one-way ANOVA of the positive/negative comment data, and data from the qualitative analysis. This is followed by the descriptive statistics both the Likert and positive/negative comment data. 3.1. Data analysis Both quantitative and qualitative methods were used to analyze the data in this study. Participants’ responses to the open-ended questions were transcribed verbatim. Each verbal response in the transcript was individually coded as “positive,” “negative,” or “neutral.” A total of 463 statements were coded as positive if the main component of the statement conveyed a positive connotation (e.g., “She spoke very clearly,” and “It was easy to make sense of what she was saying and understand it”). Verbal responses were coded as negative when the main idea of the statement conveyed a negative connotation (e.g., “It was hard to listen to because she kept on stopping and repeating a couple of words”). Some statements consisted of both a positive component and a negative component (e.g., “The girl stuttered but knew the story very well”). These statements were divided into two parts, with the portion of the statement conveying a positive connotation coded as positive and the portion of the statement conveying a negative connotation coded as negative. Statements coded as neutral (e.g., “Her speech was OK”) were not included in the analysis because the focus was only on students’ positive and negative comments. After each verbal response was coded as positive, negative, or neutral, the percentage of positive and negative statements was calculated across the four different stuttering frequencies. The transcript was further analyzed to determine themes that emerged from what participants said about the audiovisual samples they observed using standard qualitative data analysis procedures (Creswell, 1998). 3.2. Likert ratings A one-way ANOVA was performed to determine the existence of significant differences across the five Likert statements in the one fluent speech and in the three stuttering frequency conditions. The results of the analysis revealed statistical significance for Likert statements 1, 2, and 3 but not for statements 4 (I felt comfortable listening to this girl) and 5 (I would feel comfortable making friends with this girl). Specifically, the significant difference were found for the statement, “This girl is a good speaker” was F(3, 84) = 6.128, p = .001, 2p = .180 and for the second statement, “This girl has smooth speech” [F(3, 84) = 6.041, p = .001, 2p = .177] across the four conditions. Statement 3 (This girl has an easy time telling the story) yielded a statistically significant difference as well, F(3, 84) = 8.509, p = <.001, 2p = .233. The partial eta-squared values (2p ) indicate small effect sizes for the three statements that yielded statistically significant differences (Cohen, 1988). Further post hoc testing (Tukey’s HSD) was performed to determine the location of significant differences between the stuttering frequency levels. Post hoc analyses revealed that participants’ ratings for Statement 1 were different between the fluent and the mild, moderate, and severe stuttering frequency samples, 0–5%, 0–10%, and 0–15%, respectively. Participants also rated Statement 2 differently for the fluent speech sample (0%) and the 15% severe stuttering sample. Furthermore, differences were found between the 0–10% and 0–15% conditions for Statement 3. No other statistically significant differences were found between the conditions. Descriptive statistics for each of the five Likert statements are shown in Table 1. The scores ranged from 1 to 5 for each speaking condition averaged across age groups, such that the larger the mean score, the more positive the perception. The mean scores for Statement 1 (This girl is a good speaker) ranged from 3.32 to 4.82 while the mean scores for Statement 2 (This girl has smooth speech) ranged from 2.36 to 4.36. For Statement 3 (This girl had an easy time telling the story), mean scores ranged from 2.86 to 5.00. Statement 4 (I felt comfortable listening to this girl) mean scores ranged from 4.14 to 5.32 and 4.18 to 5.00 for Statement 5 (I would feel comfortable making friends with this girl). These two statements (S4 and S5) elicited favorable ratings across the four conditions. Statement 3 within the 0% speech condition yielded the highest mean rating of 5.32 whereas the lowest mean rating of 2.36 was found in the 15% condition associated with Statement 2, which is expected given that listeners were evaluating the smoothness of the girl’s speech. 3.3. Qualitative analysis There were a total of 463 participant comments. Specifically, there were 101 comments in the 0% condition, 129 comments in the 5% condition, 123 comments in the 10% condition, and 110 comments in the 15% condition. The participants’ comments were transcribed and placed into theme clusters that best fit the nature of the comments. Four specific theme clusters emerged from the data and they were: (1) speech characteristics, (2) speaker characteristics, (3) listener comfort, and (4) listener comprehension. The following is an overview of each of the theme clusters including example comments from each. 3.3.1. Speech characteristics The speech characteristics theme cluster included comments related to various aspects of the speaker’s speech including speech rate, repetitions, and effort. For the fluent speech condition, sample comments included “She had good speech
J. Panico et al. / Journal of Fluency Disorders 45 (2015) 1–11
7
Table 1 Likert descriptive data. Likert statement
Sample (% stuttering)
M
SD
Statement 1
Control (0%) Mild (5%) Moderate (10%) Severe (15%)
4.82 3.55 3.36 3.32
1.53 1.10 1.00 1.64
Statement 2
Control Mild Moderate Severe
4.36 3.09 3.05 2.36
1.76 1.51 1.56 1.53
Statement 3
Control Mild Moderate Severe
5.00 4.27 3.36 2.86
1.38 1.70 1.36 1.64
Statement 4
Control Mild Moderate Severe
5.32 5.05 4.45 4.14
1.89 1.62 2.20 1.91
Statement 5
Control Mild Moderate Severe
4.73 5.00 4.59 4.18
2.10 1.54 2.44 1.92
Notes. The mean values are based on a 7-point Likert scale where higher values reflect more agreement with the given statement. Each sample contained 22 different participants for a total of 88 participants.
because she knew what she was saying,” “She seemed to have pretty good speech,” “She didn’t stumble on words,” and “Her words were pretty clear.” However, when the speech sample contained 5% stuttering, listeners reported more negative than positive comments. Examples of the negative comments were “It was a little rough,” “She said stuff over and over again,” and “She held the first letters and made words really long.” In the 10% condition, the comments were primarily negative such as “When she talked it was kind of bumpy,” “It was a little scratchy when she talked,” and “She would say part of a word a lot and then she would finally get to the end of it.” Finally, the 15% condition evoked comments including “It wasn’t really smooth because she had to keep repeating most of the words,” “She kept getting stuck on the words,” and “She kept saying words two or three times and some words she would hold out the sound.” 3.3.2. Speaker characteristics This theme cluster primarily dealt with the participants’ perceptions of the speaker’s comfort level in delivering the story. In the 0% condition, both positive and negative comments were made such as “I felt like she was pretty sure of what she was thinking,” “She seemed sure of herself because she knew a lot of facts about it,” and “She seemed kind of nervous while saying it,” respectively. One participant even commented, “It seemed like she was talking like a normal person would.” For the 5% condition, participants made comments including “From the video I think she had disabilities” and “It just seemed really hard for her to say some of the words,” and “She had a little bit of trouble but she is trying to do the best she can.” Participants also commented positively at this level including “She spoke very clearly,” and “Parts of the story she was really good at.” The 10% condition yielded primarily negative statements such as “She is probably not very good at talking,” “I don’t think she was really comfortable because when you stutter you’re not confident,” and “She talked kind of funny.” Within the 15%, one participant commented that “Maybe she has a disorder or something that doesn’t involve her communicating correctly with her speech.” Another participant stated, “She’s a good talker but sometimes she stutters because maybe she gets nervous or something.” Surprisingly, some participants commented positively on the speaker’s ability including “She was good at pronouncing words. She felt comfortable with what she was saying.” It was of interest as well to determine how many children used the word “stuttering” when referring to the speaker. A total of 11% (4 of 35) of the third graders (8–9 years old) and approximately 19% (5 of 27) of the fourth graders (9–10 years old) used the word “stuttering” when describing the speech of the speaker. This resulted in a combined total of 14.5% (9 out of 62) of the third and fourth graders used the word stuttering in a comment made about the peer who stuttered. 3.3.3. Listener comfort This cluster revolved around how comfortable the participants were in watching and listening to the speaker. Comments within this cluster consisted of positive and negative comments including “I thought it was easy to listen to because she enunciated well,” “She kept making the same sound which made it hard to listen to,” “It was fairly easy to listen to because she would tell me exactly what she was going to say,” “It was easy to listen to the way she talked,” and “It was hard to listen to because she kept on stopping and repeating a couple of words.” This pattern of positive and negative comments was consistent for all four stuttering frequency conditions.
8
J. Panico et al. / Journal of Fluency Disorders 45 (2015) 1–11
100 90 80
Positive Negative
Percent
70 60 50 40 30 20 10 0
0%
5%
10%
15%
Speech Sample Fig. 1. Percentage of positive and negative comments across the three open-ended questions for each of the four stuttering frequency conditions.
3.3.4. Listener comprehension The final theme cluster related to how well participants were able to follow the story being told by the speaker. The 0% condition consisted mostly of positive comments such as “I could understand all of her words” and “It was easy to understand her because I could follow all of her words.” On participant commented, “I could understand what she was saying because I have five cousins that talk like that.” In the 5% and 10% conditions participants commented that “When she said the whole word I could understand her” and “It was only the first couple of letters that were messed up but the rest just went through it so I could understand it.” Others stated “Sometimes her speech wasn’t smooth so it was hard to understand her,” and “I understood the story just fine.” Finally, in the 15% condition several participants made comments such as “It was easy to make sense of what she was saying and understand it,” and “Other than some bumps on her words the story was easy for me to understand” whereas other participants had some perceived difficulty with understanding the story as evidenced by comments such as “It was a little difficult to understand what she was saying,” “Not too easy because she kept getting stuck on words,” and “It was just hard to understand her because of her speech.” There were more negative comments for the 15% condition than the other three conditions. 3.3.5. Positive/negative comment data A one-way ANOVA was performed to determine if the percentage of positive comments made by participants across the four conditions were significantly different. Overall, the analysis revealed statistical significance, F(3, 84), p = .000, 2p = .403. Post hoc testing (Tukey’s HSD) was then conducted to determine the location of the significant differences between frequency conditions. Specifically, the percentage of positive comments was significantly higher between the 0% and 5%, the 0% and 10%, the 0% and 15% conditions. No other statistically significant differences were found between the other conditions. Fig. 1 shows the percentage of positive and negative comments made across the three open-ended questions. Overall, there is a clear pattern of decreasing positive comments and increasing negative comments as the stuttering frequency increased from 0% to 15%. Specifically, the 0% condition yielded the highest percentage of positive comments (67%/33% negative). In the 5% condition, the total percentage of positive comments decreased to 38% and there were 62% negative comments. In the 10% condition the percentage of positive comments decreased to 34% versus 66% negative comments. The 15% condition yielded 28% positive and 72% negative comments. 4. Discussion This study sought to examine elementary school-age children’s perceptions of stuttering using quantitative and qualitative research methodology. To our knowledge, this is the first study to use this methodological approach to determine how typically fluent children perceive an elementary school-age peer who stutters across three different frequencies of stuttering. The results are based on children who did not have extensive contact or exposure to other children who stutter in order to minimize any biasing that might have taken place in the participants’ responses. One of the major findings, derived from the children’s ratings for specific statements and many of the qualitative comments children made about the speaker, showed that the ratings decreased to become less favorable as the frequency of stuttering increased from 5% to 15%. This was particularly true for statements 1, 2, and 3. Similar results were found for
J. Panico et al. / Journal of Fluency Disorders 45 (2015) 1–11
9
students’ ratings for statements 1, 2 and 3 used in the Evans et al. (2008) study. This finding suggests that children as young as 8 years of age (third grade) react negatively to a child’s speech that is not smooth and when a child has difficulty telling a story. However, on a more positive note, there were no significant differences in children’s ratings across the four conditions when they were asked to rate two statements related to the degree of comfort they felt listening to the girl speak and how comfortable they felt making friends with the girl. The fact that children in this study reported feeling comfortable with a peer who stutters is encouraging for even at a young age, children seem to be able to separate how they feel about a child who stutters and that listening to stuttering is not uncomfortable. This is consistent with past studies by Healey, Gabel, Daniels, and Kawai (2007), Panico et al. (2005), Evans et al. (2008), and Susca and Healey (2001) who also showed that the major of listeners are comfortable listening to stuttering and are comfortable being a friend with a person who stutters. Of course being comfortable with or friendly toward those who stutter does not necessarily mean that listeners do not react negatively to stuttering behavior. As shown in this study, the number of negative qualitative comments children made toward the disfluent peer gradually increased as the stuttering frequency increased. There was a steady rise in the number of negative comments from the 5% to the 15% samples relative to (1) the frequency of stuttering, (2) how difficult it was for the listeners to listen to the girl stutter, and (3) how difficult it was to follow what the peer said. The finding is similar to the data reported by Susca and Healey (2001) for adults who stutter. A third finding of the present study related to the percentage of children in the third and fourth grades who used the term “stuttering” in their comments about the peer’s speech. Culatta and Sloan (1977) reported that 37% of their third and fourth grade participants used the word “stuttering” when referring to the adult’s speech. In the present study, only 14% of the total number of third and fourth graders used the label “stuttering.” This lower percentage of students who used “stuttering,” compared to the Culatta and Sloan study might suggest that children today could have been taught by parents, teachers, or others that it is unacceptable to use labels or negative terms when referring to adults or children. This current trend might be related to children in schools being made aware of and learning about children who have speech, language, hearing, cognitive, social, and physical disabilities. The term “disabilities” might also reflect what children are exposed to more often in today’s society. In fact, one child stated that, “From the video, I think she (the peer speaker) had disabilities.” In the context of the present study, it is interesting that the child used the term “disabilities” rather than “stuttering.” It is possible that the comments children made to the first and third authors about the girl’s speech was influenced by responding to an adult. It might not be polite or “correct” to use a label with an adult but it might be if a child was talking to another child. The results of the present study show by the time children reach third grade, they are developing negative perceptions and reactions to children who stutter (Evans et al., 2008; Franck et al., 2003; Hartford & Leahy, 2007; Langevin & Hagler, 2004). The relative consistency of this finding across research studies shows that stuttering is perceived negatively by the time children reach third grade. This suggests schools need to continue their efforts to help all children learn to appreciate the challenges children who stutter face in coping with the disorder. 4.1. Limitations of the study and future research Most of the children used in the present study were Caucasian (90%) with only a few minorities included in the sample. Future studies in this area also could include a more multi-culturally diverse sample of listeners given the influence social, cultural, and experiential factors have on perceptions of speakers. Studies have shown that negative perceptions and attitudes exist culturally and geographically (Abdalla & St Louis, 2011; Mayo, Mayo, Jenkins, & Graves, 2004). Thus, the findings of this study might not generalize to children from other cultural and ethnic backgrounds. It would also be informative to examine the presence of secondary coping behaviors within the speech samples to determine how those behaviors might impact listeners’ perceptions of stuttering. It is possible that eye closure, head turns, or other types of facial grimaces present during a disfluent moment might change how children perceive stuttering from a peer. This might have been the case with the participant responses in the Langevin and Hagler (2004) study when rating the disfluent speakers who had moderate and severe stuttering and accompanying secondary behaviors. It would be of interest if some form of struggle behavior and effort accompanying a low frequency level of stuttering could elicit more negative reactions than those found in the present study. Finally, it would also be of interest to investigate children’s ability to recall and comprehend information after listening to varying frequencies of stuttering spoken by a peer who stutters. There is some evidence that stuttering behavior influences recall and comprehension of information when adults listen to stories told by speakers who exhibit mild to severe stuttering (Panico & Healey, 2009) but examining whether the same effect would occur for children may also yield useful findings. Another limitation to this study was the extensiveness of the qualitative portion of the study. Although the qualitative methods used in the present study were similar to other studies examining listener perceptions of stuttering, one could argue that a greater number and broader range of questions could have been used. This might have led to additional or different types of thematic categories of the statements and interpretation of the positive and negative comments made about the three levels of stuttering. 5. Conclusions The purpose of the present study was to further investigate children’s perceptions of a peer who stutters using a mixed methodology. Quantitatively, the results indicated that perceptions become more negative as stuttering frequency increased
10
J. Panico et al. / Journal of Fluency Disorders 45 (2015) 1–11
which is consistent with previous research with older children and adults. The qualitative data complimented the quantitative findings and showed that participants commented on several aspects of the speakers’ speech pattern including rate, articulation, and effort in addition to the stuttering. By the time children reach third grade, they appear to be developing negative perceptions and reactions toward children who stutter. This suggests that children who stutter need emotional support from parents, SLPs, and teachers in helping them cope with the disorder. On a positive note, the findings of the present study revealed that although children were aware of the stuttering they still indicated a willingness to be friends with the peer. This finding should be encouraging to children who stutter in that his or her peers would accept them as friends. The findings of this study together with past research in this area should help clinicians and their clients appreciate the range of social and emotional reactions peers have toward a child who stutters.
Financial disclosures James Panico, E. Charles Healey, and Jennifer Knopik did not disclose any relevant financial relationships in support of the research reported in this article.
Non-financial disclosures James Panico, E. Charles Healey, and Jennifer Knopik did not disclose any relevant non-financial relationships used in support of the research reported in this article. CONTINUING EDUCATION Elementary school students’ perceptions of stuttering: A mixed model approach QUESTIONS (1) In general, past research that has examined children’s perceptions of stuttering has revealed which of the following? a. preschool-aged children are able to differentiate between fluent and stuttered speech b. the term “stuttering” was used overwhelmingly by children in third and fourth grades when listening to a disfluent speaker c. children responded positively to the disfluent speech produced by adult speakers d. findings have yielded significant differences toward perceptions of boys vs. girls who stutter e. that older children tend to respond with less negative reactions to stuttered speech (2) Comments related to rate, vocal characteristics, and effort are part of which theme cluster: a. speaker characteristics b. communication pragmatics c. speech characteristics d. listener comfort e. story comprehension (3) The features of stuttering used in the present study to create the samples were comprised of: a. dysrhythmic part- and whole-word repetitions b. two-unit part- and whole-word repetitions along with audible sound prolongations c. inaudible sound prolongations lasting approximately two seconds each in duration d. commensurate secondary behaviors e. three-unit repetitions along with audible sound prolongations lasting three seconds each in duration (4) Which of following is true regarding the findings of the present study? a. the number of negative qualitative comments gradually decreased as the stuttering frequency increased b. approximately half of the third and fourth grade children use the term “stuttering” when commenting about the speech of the speaker c. there were significant differences across the four conditions when asked to rate the level of comfort in listening to the speaker or making friends with the speaker d. the participants commented on several aspects of the speaker speech in addition to the presence of stuttering e. although the children were aware of the stuttering they did not indicate a willingness to become friends with the peer (5) Based on the findings of the present study, future research should: a. incorporate a more multiculturally diverse sample of participants b. create additional audio samples with higher frequencies of stuttering c. collect additional information on adult’s recall and comprehension when listening/viewing a disfluent sample d. not include secondary behaviors as they may cause the listener to focus more on the visual signal versus the stuttering moment itself e. solely on the use of quantitative methodology to examine listener perceptions of stuttering
J. Panico et al. / Journal of Fluency Disorders 45 (2015) 1–11
11
References Abdalla, F. A., & St Louis, K. O. (2011). Arab school teachers’ knowledge, beliefs and reactions regarding stuttering. Journal of Fluency Disorders, 37, 54–69. Ambrose, N., & Yairi, E. (1994). The development of awareness of stuttering in preschool children. Journal of Fluency Disorders, 19, 229–245. Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum. Creswell, J. W. (1998). Qualitative inquiry and research design. Thousand Oaks, CA: Sage Publications Inc. Culatta, R., & Sloan, A. (1977). The acquisition of the label stuttering by primary level schoolchildren. Journal of Fluency Disorders, 2, 29–34. Evans, D., Healey, E. C., Kawai, N., & Rowland, S. (2008). Middle school students’ perceptions of a peer who stutters. Journal of Fluency Disorders, 33, 203–219. Ezrati-Vinacour, R., Platzky, R., & Yairi, E. (2001). The young child’s awareness of stuttering-like disfluency. Journal of Speech, Language, and Hearing Research, 44, 368–380. Franck, A. L., Jackson, R. A., Pimentel, J. T., & Greenwood, G. S. (2003). School-age children’s perceptions of a person who stutters. Journal of Fluency Disorders, 28, 1–15. Giolas, T. G., & Williams, D. E. (1958). Children’s reactions to nonfluencies in adult speech. Journal of Speech and Hearing Research, 1, 86–93. Guitar, B. (2006). Stuttering: An integrated approach to its nature and treatment. Philadelphia: Lippincott, Williams & Wilkins. Hartford, E., & Leahy, M. M. (2007). The perceptions of primary school children of a person who stutters. In J. Au-Yeung, & M. M. Leahy (Eds.), Proceedings of the 5th world congress on fluency disorders (pp. 223–229). Dublin, Ireland: The International Fluency Association. Healey, E. C., Gabel, R. M., Daniels, D. E., & Kawai, N. (2007). The effects of self-disclosure and non-self-disclosure on listeners’ perceptions of a person who stutters. Journal of Fluency Disorders, 32, 51–69. Kazdin, A. E. (1982). Single case research designs: Methods for clinical and applied settings. New York: Oxford University Press. Langevin, M., & Hagler, P. (2004). Development of a scale to measure peer attitudes toward children who stutter. In A. K. Bothe (Ed.), Evidence-based treatment of stuttering: Empirical bases and clinical applications (pp. 139–171). Mahwah, NJ: Erlbaum. Leslie, L., & Caldwell, J. S. (2005). Qualitative reading inventory-4. Boston: Pearson. Mayo, R., Mayo, C., Jenkins, K., & Graves, L. (2004, September). Public knowledge of stuttering: Cross cultural perspectives. Speech Pathology.com. Retrieved from. http://www.speechpathology.com/articles/public-knowledge-stuttering-cross-cultural-1457 Panico, J., & Healey, E. C. (2009). The influence of text type, topic familiarity, and stuttering severity on listener recall, comprehension, and mental effort. Journal of Speech, Language, and Hearing Research, 52, 534–546. Panico, J., Healey, E. C., Brouwer, K., & Susca, M. (2005). Listener perceptions of stuttering across two presentation modes: A quantitative and qualitative approach. Journal of Fluency Disorders, 30, 65–85. Riley, G. D. (1994). Stuttering severity instrument for children and adults (3rd ed.). Austin, TX: Pro-ed. Riley, G. D. (2009). Stuttering severity instrument for children and adults (4th ed.). Austin TX: Pro-Ed. Susca, M., & Healey, E. C. (2001). Perceptions of simulated stuttering and fluency. Journal of Speech, Language, and Hearing Research, 44, 61–72.
James Panico is currently an associate professor at SIUE where he teaches undergraduate and graduate courses in fluency disorders, clinical methods, counseling, and professional issues. His research area of interest involves perceptions of stuttering. Dr. Panico has presented at professional meetings at the state, national, and international levels. Charlie Healey is a professor emeritus from the University of Nebraska-Lincoln. He has published extensively in the area of stuttering. Furthermore, he has given numerous presentations at the state, national, and international levels. Jennifer Knopik completed her undergraduate and graduate degree in speech-language pathology at SIUE. She has published and presented in the area of stuttering. Currently she is a speech-language pathologist for a private rehabilitation company.