Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Palatometry intervention in relation to body (structure and function), activity, and participation 2000

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
ubc_2000-0380.pdf [ 11.16MB ]
Metadata
JSON: 1.0089719.json
JSON-LD: 1.0089719+ld.json
RDF/XML (Pretty): 1.0089719.xml
RDF/JSON: 1.0089719+rdf.json
Turtle: 1.0089719+rdf-turtle.txt
N-Triples: 1.0089719+rdf-ntriples.txt
Citation
1.0089719.ris

Full Text

PALATOMETRY INTERVENTION IN RELATION TO BODY (STRUCTURE AND FUNCTION), ACTIVITY, AND PARTICIPATION by KAREN ELIZABETH DERRY B. A. (Hons.), The University of British Columbia, 1998 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (School of Audiology and Speech Sciences) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA August, 2000 © Karen Elizabeth Derry, 2000 UBC Special Collections - Thesis Authorisation Form http://www.library.ubc.ca/spcoll/thesauth.html I n p r e s e n t i n g t h i s t h e s i s i n p a r t i a l f u l f i l m e n t o f t h e r e q u i r e m e n t s f o r an advanced degree a t t h e U n i v e r s i t y o f B r i t i s h C o l u m b i a , I agree t h a t t h e L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r r e f e r e n c e and s t u d y . I f u r t h e r agree t h a t p e r m i s s i o n ' f o r e x t e n s i v e c o p y i n g o f t h i s t h e s i s f o r s c h o l a r l y p u rposes may be g r a n t e d by t h e head o f my department o r by h i s o r he r r e p r e s e n t a t i v e s . ; I t i s u n d e r s t o o d t h a t c o p y i n g o r p u b l i c a t i o n ' o f . t h i s t h e s i s 1 . ' f o t ' " f i n a n c i a l g a i n s h a l l not be a l l o w e d w i t h o u t my w r i t t e n p e r m i s s i o n . Department o f JdrVT?^ 0<(\ T^A^lcjCyj \ 5f^A <5b^e/V*UCV. The U n i v e r s i t y o f B r i t i s h C olumbia Vancouver, Canada 1 of 1 7/13/00 11:54 A M ABSTRACT The purpose of this study was to evaluate outcomes of palatometry therapy for a twenty-two-year-old woman with a bilateral, severe-to-profound hearing loss, a severe speech sound production disorder, and oral musculature weakness. Outcomes addressed levels of functioning at the body, activity, and participation levels (World Health Organization, 1999). At the level of the body, improvement was noted between pre- and post- therapy narrow phonetic transcriptions of speech production, and for pre- and post-therapy palatograms. Specifically, post-therapy palatograms for therapy targets (/g/, III, I til, and Id^l) and generalization targets (Is/, Izl, and fkf) showed improved linguapalatal contact patterns compared with pre-therapy palatograms. To obtain a measure of intelligibility (an activity level measure), five listeners who were unfamiliar with disordered speech were asked to complete open- and closed-set word identification tasks and an open-set sentence identification task. Word identification improved for both the open-set and closed-set tasks. An average gain of 34% was noted for word identification in the closed-set task. Identification of target phonemes within the open-set word identification task was worse than within the closed-set word identification task. In addition, each listener orthographically transcribed six sentences. I l l Level of participation (WHO, 1999) was assessed on the basis of responses from questionnaires administered to the speaker and on the basis of listeners' perceptions. Generally, responses from the participant indicated improved self- confidence but the presence of some participation restrictions related to her hearing impairment. Listeners' perceptions suggested that the speaker would likely encounter more participation restrictions in situations that involved speech as the primary mode of communication compared to situations involving primarily non-oral modes of communication. Overall, the results suggested that palatometry therapy had a positive effect on reduction in speech impairment, on reducing activity limitations, and in improving the participant's self-confidence and self-awareness as a participant within society. TABLE OF CONTENTS ABSTRACT ii TABLE OF CONTENTS iv LIST OF TABLES : : ix LIST OF FIGURES xv ACKNOWLEDGMENTS xvi 1. INTRODUCTION... 1 INTERNATIONAL CLASSIFICATION OF FUNCTIONING AND DISABILITY: BODY (FUNCTIONS AND STRUCTURES), ACTIVITY, AND PARTICIPATION 2 The World Health Organization's ICIDH-2 Beta-2 Draft Framework in Relation to Speech Therapy 4 PALATOMETRY 6 Palatometry as a Speech Therapy Tool for Persons with Hearing Impairments 8 Palatometry as a Speech Therapy Tool for Persons with Motor Speech Disorders 11 Summary 14 BACKGROUND PALATOMETRY STUDY 14 Assessment Tools and Measures.. 14 Results 15 INTELLIGIBILITY 16 Listener Identification Tasks and Factors Affecting Intelligibility Scores 17 PARTICIPATION 23 FOCUS OF THE CURRENT STUDY... 26 RESEARCH QUESTIONS AND HYPOTHESES 28 Level of Body 28 Level of Activity 30 Level of Participation.. 33 2. METHOD SPEAKER INFORMATION 34 1996 Pre- and Post-Treatment Assessment Audio-Recordings 37 THERAPY 40 Therapy and Generalization Targets 40 Session Structure 41 OUTCOME MEASURES 47 Body Level 47 Activity Level 62 Listeners 63 Participation Level 74 vi 3. RESULTS 80 Body Level 80 Pre-therapy assessment results 81 Post-therapy assessment results 94 Palatometry therapy results 114 Palatogram sound file recording results 123 Palatogram results '. 131 Activity Level 156 Open-set word identification task 156 Open-set sentence identification task 158 Phoneme identification in the open-set word task 161 Closed-set word identification task 163 Phoneme identification in the closed-set word task 164 Words identified by each listener for each task 167 Qualitative comments by listeners 169 Participation Level 172 Palatometry research project questionnaire 173 Questionnaire: This study , 176 Speech and hearing questionnaire results 178 Listeners' responses 182 4. DISCUSSION 184 Body Level 184 Pre- and Post-Therapy Word List Transcriptions 185 Treatment Target Phonemes in Therapy 194 Pre- and Post-Therapy Palatometry Data 198 Summary and Significance of Findings to the Body Level 219 Activity Level 222 vii Open- and Closed-Set Word Identification 223 Goal Phoneme Identification in the Open-Set and Closed-Set Word Identification Tasks 224 Word and Sentence Identification Tasks 227 Participation Level 231 Samantha's Responses 232 Listeners' Perceptions 236 APPLICATION AND SIGNIFICANCE OF FINDINGS TO THE WORLD HEALTH ORGANIZATION MODEL (1999) 237 LIMITATIONS OF THIS STUDY AND DIRECTIONS FOR FUTURE RESEARCH 241 CONCLUSION 244 5. REFERENCES 247 APPENDIX A '. 259 EXPERIMENTAL STIMULI 259 APPENDIX B 262 THERAPY DATA 262 APPENDIX C 263 viii QUESTIONNAIRES 263 A P P E N D I X D 275 P A L A T O G R A M R E S U L T S 275 i x LIST OF TABLES Table 1.0: Characteristics of listeners . 64 Table 1.1: Order of events during listener sessions one and two 70 Table 2.0: Samantha's pre-assessment consonant inventory over all word positions 83 Table 2.1: Pre-therapy vowel inventory 84 Table 2.2: Pre-therapy consonant and vowel percentage match with adult targets 84 Table 2.3: Samantha's post-assessment consonant inventory over all word positions 94 Table 2.4: Post-therapy vowel inventory 95 Table 2.5: Consonant and vowel percentage match with adult targets pre- and post-therapy assessments 95 Tables 2.6a-j: Percentage match of Samantha's forms with adult targets at pre- and post-therapy assessments 96 Table 2.6a: Consonants organized by manner: overall accuracy 96 Table 2.6b: Consonants organized by place: overall accuracy 97 Table 2.6c: Consonants organized by voicing: overall accuracy 97 Table 2.6d: Percentage match of Samantha's therapy and generalization targets with adult targets 98 Table 2.6e: Place accuracy independent of other features for target consonants 98 Table 2.6f: Place accuracy independent of other features for generalization consonants 100 Table 2.6g: Manner accuracy independent of other features for target consonants 101 X Table 2.6h: Manner accuracy independent of other features for generalization consonants 102 Table 2.6i: Voicing accuracy independent of other features for target consonants 104 Table 2.6j: Voicing accuracy independent of other features for generalization consonants 105 Table 2.7: Pre- and post-therapy target Igl transcriptions for blocks one and two 115 Table 2.8: Pre- and post-therapy target M 3 / transcriptions for blocks one and two 116 Table 2.9: Pre- and post-therapy target It J/ transcriptions for blocks one and two 117 Table 2.10: Pre- and post-therapy target /£/ transcriptions for blocks one and two 119 Table 3a: Pre- and post-assessment palatogram sound file recording transcriptions for word-initial /tj/ tokens 123 Table 3b: Pre- and post-assessment palatogram sound file recording transcriptions for word-final /tj/ tokens 124 Table 3c: Pre- and post-assessment palatogram sound file recording transcriptions for word-initial /$/ tokens 125 Table 3d: Pre- and post-assessment palatogram sound file recording transcriptions for word-final /$/ tokens 126 Table 3e: Pre- and post-assessment palatogram sound file recording transcriptions for word-initial /d3/ tokens 126 Table 3f: Pre- and post-assessment palatogram sound file recording transcriptions for word-final /CI3/ tokens 127 Table 3g: Pre- and post-assessment palatogram sound file xi recording transcriptions for word-initial Igl tokens 128 Table 3h: Pre- and post-assessment palatogram sound file recording transcriptions for word-final Igl tokens 128 Table 3i: Pre- and post-assessment palatogram sound file recording transcriptions for word-initial and word-final /k/ tokens 129 Table 3j: Pre- and post-assessment palatogram sound file recording transcriptions for word-initial and word-final /z/ tokens 129 Table 3k: Pre- and post-assessment palatogram sound file recording transcriptions for word-initial and word-final Is/ tokens 130 Table 4a: Word-initial and word-final stop closure phase results for each criterion of It 11 133 Table 4b: Word-initial and word-final fricative phase results for each criterion of It ll 134 Table 5a: Word-initial and word-final maximum contact phase results for each criterion of III 135 Table 5b: Word-initial and word-final release phase results for each criterion of III 136 Table 6a: Word-initial and word-final stop closure phase results for each criterion of/d.3/ 138 Table 6b: Word-initial and word-final fricative phase results for each criterion of /C13/ 139 Table 7a: Word-initial and word-final maximum contact phase results for each criterion of Igl 140 Table 7b: Word-initial and word-final release phase results for each criterion of Igl 141 xii Table 8a: Word-initial and word-final fkJ token maximum contact phase results 142 Table 8b: Word-initial and word-final fkl token release phase results 143 Table 9a: Word-initial and word-final Izl token maximum contact phase results 144 Table 9b: Word-initial and word-final Izl token release phase results 145 Table 10a: Word-initial and word-final Is/ token maximum contact phase results 147 Table 10b: Word-initial and word-final I si token release phase results 148 Table 11a: Percentage of criteria noted to have changed for the stop phase of the affricate treatment targets, M3/ and/tj/ 149 Table lib: Percentage of criteria noted to have changed for each degree of change for the fricative phase of the affricates 149 Table 11c: Percentage of criteria noted to have changed for each degree of change for the maximum contact phase of III 150 Table l id: Percentage of criteria noted to have changed for each degree of change for the release phase phase of/J/ 151 Table lie: Percentage of criteria noted to have changed for each degree of change for the maximum contact phase of/g/ 151 Table llf: Percentage of criteria noted to have changed for each degree of change for the release phase of/g/ 152 Xlll Table 12a: Percentage of criteria noted to have changed for each degree of change for the maximum contact phase of the generalization target words 153 Table 12b: Percentage of criteria noted to have changed for each degree of change for the release phase of the generalization tar get words 153 Table 13: Number of words correctly identified pre- and post-therapy in the open-set word identification task (out of 20) 157 Table 14: Difference between the percentage of words correctly identified pre- and post-therapy in the open-set word identification task 157 Table 15a: Percentage of words per sentence correctly identified pre- therapy in the sentence identification task 158 Table 15b: Percentage of words per sentence correctly identified post-therapy in the sentence identification task 158 Table 15c: Difference between the percentage of words per sentence correctly identified pre- and post-therapy in the sentence identification task 159 Table 16: Percentage of treatment targets correctly identified pre-therapy and post-therapy in the open-set word identification task 160 Table 17: Difference between the percentage of treatment targets correctly identified pre- and post-therapy in the open-set word identification task 160 Table 18: Percentage of generalization targets correctly transcribed pre-therapy and post-therapy in the open-set word identification task 161 Table 19: Difference between the percentage of generalization targets correctly identified pre- and post-therapy in the open-set word identification task 162 Table 20: Number of words correctly identified pre- and post-therapy x i v in the closed-set word identification task (out of 20) 163 Table 21: Difference between the percentage of words correctly identified pre- and post-therapy in the closed-set word identification task 163 Table 22: Percentage of treatment targets correctly identified pre-therapy and post-therapy in the closed-set word identification task 164 Table 23: Difference between the percentage of treatment targets correctly identified pre- and post-therapy in the closed-set word identification task 165 Table 24: Percentage of generalization targets correctly identified pre-therapy and post-therapy in the closed-set word identification task 166 Table 25: Difference between the percentage of generalization targets correctly identified pre- and post-therapy in the closed-set word identification task 166 Tables 26a-c: Words correctly identified by each listener for each task.... 167 Table 26a: Open-set word identification task 167 Table 26b: Closed-set word identification task 167 Table 26c: Open-set sentence identification task 168 Table 27a: Listeners' understanding ratings in the open-set word identification task for pre-therapy words 169 Table 27b: Listeners' understanding ratings in the open-set word identification task for post-therapy words 170 Table 28: Listeners' understanding ratings in the open-set sentence identification task for pre- and post-therapy words 172 Table 29: Questions asked to each listener ; 181 Table 30: Listeners' responses 182 \ XV LIST OF FIGURES Figure 1.0: Model Palatogram 54 Figure 1.1: Percent Correct After Palatometry and Non-Palatometry Therapy, Block One 121 Figure 1.2: Percent Correct After Palatometry and Non-Palatometry Therapy, Block Two 122 Figure 1.3: Total Number of + +, + N, + - , - +, and - - For All Treatment and Generalization Targets in All Phases 155 xvi ACKNOWLEDGEMENTS I am indebted to the following people: Dr. Barbara Bernhardt for her invaluable support, editing, suggestions, and mentoring. Thank you for making this project possible and for encouraging me to further explore palatometry. Dr. Kathy Pichora-Fuller for her helpful suggestions and careful editing. The listeners for their participation, commitment, and interest. Samantha and her mother. This project would not have been possible if it weren't for their determination and participation. My family and dearest friend Chris for encouraging me to always take the more challenging road. I have grown both professionally and personally. 1 CHAPTER ONE INTRODUCTION The current study evaluated the outcomes of palatometry therapy (or electropalatography) conducted with a twenty-two-year-old woman with a congenital bilateral, severe-to-profound hearing loss, a severe speech sound production disorder, and oral musculature weakness. Outcomes from this study address levels of functioning at the body, activity, and participation levels (World Health Organization [WHO], 1999). Palatometry intervention, an oral communication treatment at the level of the body, can potentially effect an individual's activity and participation. The introduction will begin with a description of the WHO's (1999) framework for categorizing human functioning and disability, followed by a discussion of the relation between this framework and speech therapy outcomes. Subsequently, a description of the palatometer and of palatometry as a speech therapy tool will be provided. This will be followed by a discussion on activity limitations and issues in intelligibility. Finally, the purpose of this study will be discussed, as it relates to the previous literature review. 2 INTERNATIONAL CLASSIFICATION OF FUNCTIONING AND DISABILITY: BODY (FUNCTIONS AND STRUCTURE), ACTIVITY AND PARTICIPATION The World Health Organization (1980) originally established the International Classification of Impairments, Disabilities, and Handicaps (ICIDH). In 1997, the framework was revised to the International Classification of Impairment, Disability, and Handicap-2 (ICIDH-2). The terms "disabilities" and "handicaps" were redefined as "activity limitations" and "participation restrictions" respectively. These latter terms were used because they contained a more positive dimension as opposed to the negative dimensions contained under the terms "disabilities" and "handicaps." Also contained in the ICIDH-2 are contextual factors both at the environmental (physical, social, attitudinal) and the personal (age, gender, personality, lifestyle habits) level. Such factors can positively facilitate one's activity level and participation or act negatively as a barrier. In July 1999, the WHO released the Beta-2 draft of the International Classification of Functioning and Disability (ICIDH-2, Beta-2). The dimensions of "impairment," "disability" and "handicap" represented within the 1997 classification system, are replaced with the terms "functioning" (a more neutral dimension) and "disability" (a negative dimension). These latter terms are umbrella terms covering three health-related positive dimensions: (1) body functions and structures; (2) activities at the individual level; and (3) 3 participation in society. The concept of body functions is defined as, "the physiological or psychological functions of body systems" (ICIDH-2 Beta-2 Short Version, 1999, p. 41). The term body structure is defined as "anatomical parts of the body such as organs, limbs, and their components" (p. 61). Impairments are "problems in body function or structure as a significant deviation or loss" (p. 61). According to the WHO (1999), activity is "the performance of a task or action by an individual. Activity limitations are difficulties in performance of activities" (p. 65). The concept of participation is "an individual's involvement in life situations in relation to health conditions, body functions or structures, activities, and contextual factors. Participation restrictions are problems an individual may have in the manner or extent of involvement in life situations" (p. 83). As in the ICIDH-2, the ICIDH-2 Beta-2 also considers contextual factors both at the environmental and personal levels. The aims of the ICIDH-2 Beta-2 are to provide a common language and framework for describing human functioning and disability; to improve communication between health care workers and people with disabilities; to permit comparison of data across countries and health care disciplines; and to provide a scientific basis for understanding and studying the functional states associated with health conditions (p. 9). Importantly, this classification system is not just about disability; it is about all people with a health condition. In addition, the WHO framework integrates both the medical and the social models into a bio-psycho-social approach which emphasizes health, functioning and well- 4 being (Larson, 1999). Traditional medical models focus on the impairment or the disease itself; social models, however, focus on the disability and participation levels. The WHO framework combines both models in order to examine human functioning at all levels; the body functions and structures, a person's activity level, a person's participation within society, and the physical environment (ICIDH-2 Beta-2, 1999). The World Health Organization's ICIDH-2 Beta-2 Draft Framework in Relation to Speech Therapy The ICIDH-2 Beta-2 model (1999) from the World Health Organization is useful as a tool for researchers measuring speech therapy outcomes. The focus of therapy and its results for persons with speech disorders, can be analyzed at all levels of functioning (body, activity, and participation). For example, speech therapy focusing on improving place of articulation targets the level of the body. Furthermore, focusing on improving articulation of a particular treatment target for everyday conversational speech, may help improve a person's ability to communicate with others, thereby decreasing activity limitations. Lastly, if improvement of speech at the conversational level is evident, then a person's level of participation within society may increase. Examining all levels of human functioning helps to derive a more functional and purposeful therapy program, and it encourages more meaningful client-focused outcome measurements. Thus, 5 outcomes can be specified that correspond to each of the levels in the WHO framework. Speech therapy outcome studies generally focus on speech measures at the level of the body or activity limitations. For example, measures tailored to the level of the body are more diagnostic and instrumental forms of measurement that focus on the deficit itself. Such measures include articulation scores, phonetic errors determined by transcription, fluency scores, deviant movements determined by videofluoroscopic examinations, and in the case of palatometry therapy, deviation from normal on palatograms (e.g., Michi, Suzuki, Yamashita, and Imai, 1986; Dagenais, 1992; Michi, Yamashita, Imai, Suzuki, and Yoshida 1993; Whitehill, Stokes, Hardcastle and Gibbon, 1995). These types of measures identify structural and or functional abnormalities, and they can be compared pre- and post-therapy. Activity level measures focus on specific communication tasks in everyday contexts, for example, producing intelligible speech, conversing with family and friends, reading the newspaper, and using the telephone. If outcomes evaluate participation (which is rare), then measurements usually include "health-related quality of life measures or handicap inventories, usually designed as self-administered questionnaires that capture, social, environmental, and economic disadvantages from the perspective of the client or family" (Frattali, 1998, p. 12). In addition, the ICIDH-2 Beta-2, considers personal and environmental factors, such as the communication environment and 6 familiarity with the communication partner. These factors can facilitate a particular communication situation or act negatively as a barrier. The next section discusses palatometry, the speech therapy tool used for this study. PALATOMETRY Palatometry, or electropalatography (EPG), is a computer-based biofeedback tool that provides an on-line, dynamic display of the tongue's contact on the hard palate. The client wears a thin, custom molded acrylic plate (a "pseudopalate"), that covers the hard palate and, in the case of the Kay Elemetrics Palatometer Model 6300 (the palatometer used for this study), maxillary teeth are also covered. Embedded within the pseudopalate are 60 to 90 electrodes which are connected to external circuitry through many fine wires. The electrodes are generally concentrated along the alveolar ridge area with some along the gingival margins of the premolars and molars. When the tongue contacts the pseudopalate (linguapalatal contact), an electronic circuit is completed via the electrodes. These contact points are then electronically transmitted onto a computer screen so that a visual representation of the speech sound is displayed. The contact points can operate as a dynamic palatograph, with the varying contact patterns reflecting the continually changing tongue contacts, or as a static palatograph, with the contact points remaining static after linguapalatal contact is made. In addition, the palatometer has a playback modality which presents the physiologically-based feedback with auditory/visual feedback. Contact patterns and sound files can be saved for future analysis and evaluation. The client compares his/her speech output with previously stored visual templates that he/she has produced, or to target patterns produced by the clinician. Computer files and printouts of contact patterns (palatograms) allow for comparisons and analysis at different points during the course of intervention. Research indicates that the palatometer has been an effective speech therapy tool for the treatment of a variety of speech impairments due to etiologies such as cleft palate (Michi, Yamashita, Imai, Suzuki & Yoshida, 1993), hearing impairment (Dagenais & Critz-Crosby, 1991; Dagenais, 1992; Dagenais, Critz-Crosby, Fletcher & McCutcheon, 1994; Parsloe, 1998), motor speech impairment (Dent, Gibbon & Hardcastle, 1995; Howard & Varley, 1995), and primary articulation disorder (Dent, Gibbon & Hardcastle, 1995; Gibbon, Hardcastle, Dent & Nixon, 1996). This section will focus on the use of palatometry with individuals with hearing impairment and motor speech impairment because the client for this study has both of these impairments. 8 Palatometry as a Speech Therapy Tool for Persons with Hearing Impairments Fletcher, Dagenais and Critz-Crosby (1991) conducted palatometer therapy with five children with profound hearing impariments who had previously received traditional aural/oral therapy (e.g., Ling, 1976). After a three to four week period of daily therapy, all children acquired better linguapalatal contact patterns for It, d, k, g, s, z, ll, and listeners judged the sounds to be more intelligible. Furthermore, a faster rate of progress was demonstrated in comparison to traditional therapy. However, whether the results were maintained over time is not reported. Subsequently, after a three-phase study, Dagenais (1992) reported that consonant production was better learned and retained (as measured immediately and six months post-treatment) with a group of children with hearing impairments who received palatometer training for one month compared to a traditionally trained group. In the third study, four children with profound hearing impairments received both palatometer training and traditional aural/oral therapy over a two- year period; however, each child received three half-hour sessions per week for the first year and two half-hour sessions for the second year. Results showed that lingualpalatal contact scores and listener judgments for stops It, d, k, gl, and alveolar sibilants I si and 17,1, peaked at the end of the first year and then levelled 9 off. The authors suggested that the children may have been less interested in the visual feedback in the second year. On the basis of a subsequent short-term, but equally intensive study, it was concluded that "electropalatographic techniques provided a viable alternative to the traditional type of training" and that the changes could be stable six months post-treatment (Dagenais, Critz-Crosby, Fletcher, & McCutcheon, 1994, p. 697). Parsloe (1998) also reported positive outcomes three months post- intervention in a case that is similar to the case of this study. Electropalatography was used as a tool to help a 10-year-old girl with a profound hearing-impairment. The participant, Sophie, had difficulty obtaining the correct placement and manner for the alveolar/velar voiced stop contrast, and for velars, palato-alveolars, and alveolars in all word-positions. Immediately after and three months post-intervention, significant improvement was evident for all sounds in all word positions. In addition, post-therapy EPG recordings showed that distinct phonological contrasts were evident between the alveolar, palatal, and velar categories. All of these studies have utilized intensive treatment sessions, all short- term with the exception of the third stage of the Dagenais study (1992). Interestingly, all of the studies discussed above primarily addressed impairment measures as intermediate outcomes (i.e., outcomes that determine whether treatment is benefiting the client; Frattali, 1998). 10 The speech measure outcomes used in the previous studies were based on palatograms and narrow phonetic transcription by expert listeners. These studies, however, do not indicate whether the outcomes generalized to daily life activities so that activity and participation levels increased. Bernhardt, Pichora-Fuller, & Williams (in press) addressed speech production outcomes after traditional therapy and after palatometry, at both the impairment (body level) and activity level for one subject, Dora, who had a hearing loss. Palatometry treatment consisted of 20 45-60-minutes sessions distributed across a five-month period, a time period more representative of standard clinic procedure. Target speech sounds, based on the post-baseline palatometry speech sample and analysis, included Igl, /dV, Isl, /$/, It J/, and /cj.3/ across word positions, and Irl in clusters. Perceptual evaluation included narrow phonetic transcription of a single word list by a trained clinician (an impairment level evaluation), orthographic transcription (of both words and sentences) by untrained individuals (an activity level evaluation), and pre- and post-treatment connected speech sample judgments by untrained individuals (an activity level evaluation, noted in Williams, 1998). Instrumental evaluation involved pre- and post-treatment palatogram analysis (an additional impairment level evaluation). Results indicated that participant Dora's speech sound production abilities showed faster progress during five months of palatometry therapy than 11 during her four years of traditional speech sound treatment program. In addition, Dora's activity limitations, based on untrained listeners ability to understand Dora's speech, appeared to have decreased significantly post- palatometry. However, whether improvement was noted in Dora's participation within society was not measured. Palatometry as a Speech Therapy Tool for Persons with Motor Speech Disorders Palatometry has also been used to treat motor speech disorders resulting from impairments such as cleft lip or palate (e.g., Hardcastle, Gibbon & Jones, 1991), neurological impairment (i.e. dysarthria), or delayed speech sound production (e.g., Dagenais, 1995; Hardcastle, Gibbon & Jones, 1991). Studies have shown that details of tongue placement were better detected using the EPG or palatometry system compared with auditory-based transcription. For example, in Hardcastle, Gibbon and Jones (1991), it was reported that 4-year-old participant, MB, was backing alveolar stops and as result was producing "homonyms" in which two different and distinct words were perceived as being identical (i.e., date, gate, and Kate were produced as [geik]). EPG analyses revealed that MB was in fact using two distinct places of articulation (i.e., alveolar and velar places); however, location of the release phase of both targets was in the posterior or velar region of the palate, giving the percept of 12 velar. More importantly, however, the EPG results indicated that MB was actually selecting the correct phoneme. Thus, subphonemic contrasts were present and the problems were at the phonetic production level, in the release phase. Similar results were reported by Hardcastle, Gibbon and Jones (1991) for a participant who suffered a left hemisphere cerebrovascular accident (CVA) which resulted in apraxia of speech. EPG data showed that participant M, was producing double articulations at the beginning of his words; however, M's double articulations were not noted perceptually. Additionally, EPG data indicated that M's articulatory patterns were highly variable, characteristic of impaired motor planning or apraxia of speech. Thus, the EPG results provided a closer and more precise examination of the variability of M's productions. EPG treatment may be valuable for the client who presents with apraxia of speech since misdirected articulatory gestures may be reduced because of the visual feedback provided. Hardcastle, Gibbon and Jones (1991) also discussed the benefits of using EPG for production of fricatives by persons with speech difficulties. The authors stated that "[descriptively, fricatives are not only difficult to characterize acoustically and aerodynamically, but they are also problematic to transcribe" (p. 63). It was concluded that EPG was useful in providing a more detailed description of the articulatory placement of fricatives. 13 In summary, EPG in conjunction with auditory-based transcription assessments aids in the detection of deviant articulations. For example, speech patterns that may be undetected by the listener may be evident with EPG. The use of EPG with persons with motor speech disorders may provide a clearer and more precise understanding regarding the nature of the disorder, thereby influencing the treatment procedure. Summary The use of palatometry (EPG) as a diagnostic and therapeutic tool for persons with hearing impairments and for persons with motor speech disorders has been described in the above discussion. It was shown that EPG therapy was successful immediately post-treatment and three to six months post-treatment. Furthermore, evidence from studies involving persons with various speech disorders showed that subphonemic contrasts that may not be perceived by listeners can be seen with EPG analysis. Lastly, EPG therapy has been shown to reduce activity limitations as was illustrated in the case of a hearing-impaired child. Outcome studies on the effects of palatometer therapy with regard to activity level are sparse. As mentioned earlier, the focus has traditionally been on outcomes at the impairment (body) level. 14 The research reported in this study follows up on an outcome study by Williams (1998) which is discussed in the next section. BACKGROUND PALATOMETRY STUDY Prior to the present study, a palatometry outcomes study was conducted (Williams, 1998). The goal of this study was to determine the effectiveness of speech therapy using palatometry on activity limitation (Williams, 1998; World Health Organization, 1997). Seven speakers acted as subjects, each presenting with a variety of impairments: cleft lip and palate, cleft palate, bilateral cleft lip and palate, brain injury, cochlear implant, motor speech disorder (two cases), and severe-to-profound hearing loss (two cases). One of the latter cases became the participant in the present study. Assessment Tools and Measures Twenty sessions of palatometer therapy were provided and palatograms and audio-recordings were made pre- and post-therapy. Audio-recordings consisted of 164 single words (Bernhardt, 1990), the Assessment of Intelligibility for Dysarthric Speakers (AIDS; Yorkston & Beukelman, 1981), the Rainbow Passage, and story retellings. 15 Perceptual evaluations at the impairment level included narrow phonetic transcription of the single word list by a trained clinician, and at the level of activity limitation, orthographic transcription (of both words and sentences) by untrained listeners, and judgments of pre- and post-treatment connected speech samples by untrained listeners. Palatogram analysis was an instrumental evaluation at the impairment level. (See Williams 1998 for further details). Results Post-therapy results at the level of impairment, showed improvement in phonetic accuracy (based on narrow transcription analysis results) and post- therapy palatograms indicated that all speakers made "sound approximations which were closer to normal" (Williams, 1998, p. 95). At the activity limitation level, there was statistically significant improvement in the identification of the goal phonemes in words by 16 untrained listeners for five of seven of the speakers. For the sentence identification task, a statistically significant improvement in accuracy of goal phoneme identification was evident for three of seven speakers. Judgment task results for two speakers indicated that post- therapy samples were easier to understand significantly more often than pre- therapy samples; however, for the remaining five subjects, no significant preference was found for pre- or post-therapy samples by untrained listeners. The participant of the present study improved in all of the above measures; 16 overall, perceptions of untrained listeners suggested that palatometry therapy provided an effective method of reducing limitations on the activity of producing intelligible speech for speakers with a variety of speech disorders (1998, iii). Extending the discussion of activity limitations, the next section concentrates further on intelligibility. INTELLIGIBILITY Intelligibility may be defined as "the match between the intention of the speaker and the response of the listener to the speech passed through the transmission system" (Schiavetti, 1992). Intelligibility may be affected by the presence of any of a range of disorders such as dyspraxia (Till, Yorkston, & Beukelman, 1994), dysarthria (Till, Yorkston, & Beukelman, 1994), cleft palate (Creaghead, Newman, & Secord, 1989), or hearing impairment (Maxon & Brackett, 1992; Northern & Downs, 1991; Osberger, 1992). Factors such as linguistic context, utterance length, phonological complexity, syllabic structure, word position in an utterance, and degree of fluency, may also account for variance in intelligibility scores (Boothroyd, 1985; Till, Yorkston, & Beukelman, 1994). Concomitantly, variables pertaining to the speaker and the listener such as level of fatigue, listener skill, and content familiarity may also affect intelligibility (Till, Yorkston, & Beukelman, 1994). 17 Measures of intelligibility are used to assess communication as opposed to the physical production (i.e., oral-motor structures and functions) of speech. Thus, intelligibility measures assess an individual's level of activity rather than the level of the body or the impairment (WHO, ICIDH-2, 1999). Intelligibility may be viewed as a measure of communicative effectiveness, "the ability to use speech to communicate effectively" (Osberger, 1992, p. 234). Intelligibility measures could even go beyond activity level to examine one's level of participation within society; for example, listeners' perceptions of a speaker's communicative effectiveness within society. One limitation to intelligibility measures, however, is that no information is obtained with respect to the underlying causes for reduced intelligibility. The following section outlines measures of intelligibility that are applicable to this thesis and discusses a number of factors that may influence intelligibility scores. Listener Identification Tasks and Factors Affecting Intelligibility Scores In an identification task, listeners are required to listen to speech samples and record what they hear. The aim of the identification task may be phoneme, word, or even sentence identification. The outcome measure associated with such tasks is the degree of accuracy such as the percentage of correctly identified target phonemes within words. Outcome information on intelligibility is 18 particularly useful when devising an intervention programme. For example, percent consonants correct (PCS), an outcome measure, can assess severity of involvement. Consonants that are particularly difficult for a client to produce will generally have a higher severity rating and hence a lower PCS, than consonants that pose less difficulty. When deciding which phonemes to target during therapy, the more severe consonants are more likely to be targeted than those with a less severe rating. There are both open-set and closed-set identification tasks. Open-set identification tasks require listeners to write down what they think they hear. In a closed-set identification task, however, listeners are required to select the word they think they hear from a number of alternatives. With an open-set task a percentage correct score can be yielded; the stimuli are scored as either correct or incorrect and a percentage of the stimuli understood by the listener can be derived. Open-set word identification in sentences may also be weighted so that, for example, function words are given a lower score than content words (Monsen, 1978). The Assessment of Intelligibility of Dysarthric Speech (AIDS) (Yorkston & Beukelman, 1981a) is an example of an open-set identification task. Examples of closed-set identification tasks which are used clinically include the Speech Intelligibility Test (SPINE; Monsen, 1981) and the Speech Pattern Contrast Tests (SPAC; Boothroyd, 1985). Boothroyd (1985) demonstrated that closed-set identification tasks provided speech production measures that were virtually independent of listener experience (Boothroyd, 1985). Osberger (1992) concluded 19 that closed-set identification tasks provided more meaningful intelligibility information for speakers with more severe hearing impairments compared with open-set identification tasks. With closed-set tasks, context is increased for very unintelligible speakers. Another advantage of using a closed-set task is that it provides analytic detail about a speaker's production of phonetic contrasts (Boothroyd, 1985). For example, word initial consonant placement can be examined by having the speaker say chump and having the listener chose between chump and jump. Yorkston and Beukelman (1981b) provided support for the use of closed- set identification tasks in their study involving one normal speaker and 13 dysarthric adults. Because their study is highly relevant for the current study, details of the study follow. Yorkston and Beukelman examined the single word and sentence intelligibility scores and speaking rates across the speakers in an attempt to combine the two measures and to quantify communication efficiency for dysarthric speakers (Yorkston & Beukelman, 1981). The authors defined communication efficiency as "a combination of the rate and accuracy with which spoken messages are transmitted from speaker to listener" (1981, p. 297). Yorkston and Beukelman (1981b) found that sentence intelligibility scores across the speakers ranged from 1 to 99% correct, and single word intelligibility scores ranged from 9 to 88% correct. For the mildly dysarthric speakers, sentence intelligibility scores tended to be higher than single word intelligibility scores. Speaking rates, as measured in words per minute (WPM), ranged from 170 WPM 20 to 17 WPM for the dysarthric speakers. The speaking rate of the normal speaker was 180 WPM. Speaking rate and intelligibility were not strongly correlated across the speakers; however, the authors concluded that a stronger relationship may have existed between rate and intelligibility within a single speaker. For example, if a speaker decreases his/her speech rate it may result in an increase in his/her speech intelligibility, thereby suggesting that slower speech may be more efficient in communication. On the other hand, a higher speaking rate may be more efficient in communication if a slower speech rate does not affect intelligibility. Yorkston and Beukelman (1981b) stress the importance of selecting tasks based on severity of dysarthria and the clinical issues being addressed (1981b). For example, for persons with severe dysarthria, closed-set word identification tasks may be most appropriate because subtle changes in performance due to factors such as fatigue and alteration in health status, may occur in time. Moderately and mildly dysarthric speakers may benefit from sentence intelligibility measures or from a combination of measures. With a variety of measures many factors affecting intelligibility and communication efficiency can be examined more closely. An important consideration when using word and sentence identification tasks is the degree of linguistic competency of the listener. Words or sentences produced by persons with hearing impairments may be affected by the linguistic competency of the listener and this in turn may affect the intelligibility of the speaker with a hearing impairment. If a listener has an awareness of the typical 21 syntactic errors of speakers with hearing impairments then identification of words in a sentence could presumably be easier. McGarr (1983) showed that persons experienced in listening to the speech of the hearing-impaired obtained higher intelligibility scores than unexperienced listeners, when the speech material used was sentences as opposed to words. Similarly, Monsen (1983) determined that differences between experienced and unexperienced listeners were higher for less intelligible than for more intelligible speakers. Speech intelligibility scores may also be affected by the phonological and syntactic complexity of the speech material. Monsen (1978) reported an average sentence intelligibility score of 76% for 67 hearing impaired adolescents. Twenty-three subjects, each with a severe hearing impairment had an average score of 91%, and 44 of the subjects, each with a profound hearing impairment, had an average sentence intelligibility score of 67%. Previous studies, however, have reported an average intelligibility of between 19% to 21% for children with profound hearing impairments (Markides, 1970; Smith, 1973 as reported in Monsen, 1978). Monsen accounts for his higher intelligibility percentages to differences in the speech material. Monsen used shorter and less syntactically complex sentences; the sentences used in Smith's study (1973), however, were twice as long and syntactically more complex as those used by Monsen (as reported in Monsen, 1978). In a follow-up study, Monsen (1983) found that sentence intelligibility for hearing-impaired children improved by 17% with simple syntax, familiar vocabulary, and a reduced number of consonant clusters 22 and polysyllabic words. In addition, Monsen (1983) determined that intelligibility scores improved by approximately 7% when listeners heard two presentations rather than only one presentation of the test sentence. The predictiveness of semantic and syntactic content has also been found to influence intelligibility scores of dysarthric speech. Carter et al. (1996) for example, determined that sentences containing high grammatical predictiveness such as the subject-verb-object structure, resulted in higher intelligibility scores than did sentences with low grammatical predictiveness. Garcia and Cannito (1996a), studied the influence of semantic predictiveness, (i.e., low versus high predictiveness), on sentence intelligibility scores for speakers with severe dysarthria. The authors reported that higher predictive sentences were 60% intelligible and sentences with low predictive content were only 15% intelligible. It was concluded that sentences containing predictive information provided listeners with additional semantic information that improved their ability to decode the message. Thus, speech intelligibility scores may be affected by the rate of speech, by the type of speech material used, the linguistic competence of the listeners performing the intelligibility evaluation, by the phonological and syntactic complexity of the stimuli, and by the predictiveness of sentence content. Other factors that may influence intelligibility scores include access to auditory and visual cues and characteristics of the speaker. 2 3 Furthermore, speech intelligibility is influenced by many factors some of which are under the control of the examiner and others of which are inherent in the speaker. For example, Smith (1975) found that as the level of the hearing loss of the speaker increased, speech intelligibility decreased. These factors should be considered before decisions are made about which identification task to choose and about the participant's communication efficiency. Based on the discussion above, research suggests closed-set identification tasks may provide more information about intelligibility than can be obtained with open-set tasks. This is especially important when the speech intelligibility of speakers with severe hearing impairments or with severe dysarthric speech is being examined. I now turn to a brief discussion on level of participation. PARTICIPATION Assessing level of participation is important in order to learn more about a person's social participation and ability to access services and opportunities within society. If, for example, a person has a severe motor speech impairment s/he may not be able to function in communication situations that require understandable and efficient speech. Thus, social activity participation may be restricted. Some of the goals of speech therapy, however, are to improve speech intelligibility and to enable speakers to communicate with greater ease. If improvement is perceived by the speaker then s/he may be more inclined to 24 participate in life situations such as participating in a conversation or asking for directions. Listeners' perceptions also provide a way to measure outcomes at the level of participation. If perceptions by untrained listeners are negative, then Samantha's participation in life situations may be restricted in an attempt to avoid dealing with such negative perceptions. Participation level measurements include "health-related quality of life measures or handicap inventories, usually designed as self-administered questionnaires that capture, social, environmental, and economic disadvantages from the perspective of the client or family" (Frattali, 1998, 12). Questionnaires can also provide an indication of the degree of motivation for changing one's condition and it has been reported that they have good face validity because self- report instruments reflect personal experiences (Erdman, 1994 in Metz, Caccamise, & Gustafson, 1997; Frattali, 1998). Questionnaires have been developed which focus on activity and participation limitations imposed by motor speech impairments and by hearing impairments. For example, Yorkston, Bombarier, and Hammen created a questionnaire for speakers with mild, moderate, and severe dysarthria (in Till, Yorkston, & Beukelman, 1994). This 100-item questionnaire was developed to obtain information in the following areas: characteristics of the disorder, situational difficulty, compensatory strategies, and perceived reactions of others. Another self-administered instrument focusing on limitations imposed by a speech impairment includes Erickson's S-24 Scale which was developed to 25 measure communication attitudes that distinguish stutterers from nonstutterers (S-24 Scale, Erickson, 1969). Erickson's original S-scale was "unsuitable for repeated administration as a measure of attitude change within stuttering therapy" (Andrews & Cutler, 1974, p. 317). Thus, a revised shorter form, the S-24 scale, was developed as a more reliable and valid measure capable "of describing aspects of the disability distinct from the specific speech handicap" (1974, 317). The S-24 Scale examines a person's self-concept as a speaker and changes in speech behaviour that may or may not have resulted because of therapeutic intervention (Andrews & Cutler, 1974). Thus, it assesses changes in communication attitudes of stutterers. The Communication Profile for the Hearing Impaired (CPHI, Demorest & Erdman, 1986, 1987) was developed as a self-report inventory of communication behaviours, their perceived effectiveness, and the presence of maladaptive accommodations to hearing loss (Demorest & Erdman, 1986, 1987). The CPHI assesses many areas of interpersonal functioning that may or may not be influenced by a hearing loss, and it examines specific hearing related events that may be of concern to a person with a hearing loss. Such events include, participating in a dinner conversation, talking on the telephone, and attending social gatherings. Additionally, it focuses on communication strategies and reported difficulties by persons with hearing losses. Lastly, it may also be advantageous to investigate the affect of the hearing impairment and the types of participation restrictions a speaker may encounter 26 according to the listener's point of view. Literature on this issue is sparse; however, it should be investigated because hearing impairment affects "others directly in verbal interactions and indirectly when sharing the same sound environment such as the home" (Stephens & Hetu, 1991, 196). In conclusion, examination of level of participation according to the speaker and the listener provides a better understanding of the impact of an impairment on an individual's life and role within society. It involves considering the relationships between the impairments (level of body), the daily life activities, and the sociocultural and physical environment of the individual (Frattali, 1998; Stephens & Hetu, 1991). I now turn to a discussion of the focus of the present study. Rationales for the study and the research questions asked are provided. FOCUS OF THE CURRENT STUDY Research has shown that EPG may be an effective speech therapy tool in reducing impairment. Few studies, however, have examined the relationship between the effects of EPG therapy on activity limitations and level of participation within society. Activity and participation are particularly relevant for clients. If, for example, EPG therapy can lead to an improvement at the impairment level, then perhaps a reduction may be evident with respect to a client's activity limitations and in turn an increase may be seen at the level of 27 participation within society. The question addressed by the current study was whether or not palatometry therapy was effective at the impairment level and more importantly, was such therapy able to reduce a speaker's activity limitations and increase his/her level of participation. The subject, Samantha, was one of the subjects from Williams' study. Prior to this study, Samantha was involved in the palatometry study discussed above (reported in Williams, 1998) and she received one year of additional non-palatometry therapy. According to Williams, improvement was noted for Samantha on the word identification, sentence identification, and judgment tasks (see previous discussion for details pertaining to the tasks). Williams states that "there appeared to be little doubt that palatometry was the primary factor in [Samantha's] improvement" (Willliams, 1998, 97). Untrained listeners were able to understand more of the information Samantha was transmitting through her speech. Since these listeners were better able to understand Samantha, it was assumed that daily interactions with others would be more successful for Samantha, thereby indicating a reduction in activity limitations. However, one aspect that was not investigated was whether or not palatometry therapy had a positive effect on Samantha's level of participation within society. This aspect acted as one impetus for the present study. Additionally, both Samantha and the previous palatometry clinicians (i.e., Williams and Bernhardt) felt that Samantha might benefit from more palatometry therapy. Since improvement had been seen using the palatometer it 28 was decided that perhaps more positive gains could be achieved. Furthermore, since Samantha was beginning to attend post-secondary classes, it was felt that she would be an appropriate case for which to investigate the effects of palatometry therapy at the level of participation within society. Samantha requested more palatometry therapy and agreed to participate in this current study. RESEARCH QUESTIONS AND HYPOTHESES Presently, there are no studies that have investigated outcomes for individuals with severe hearing impairments and concomitant motor speech disorders in terms of all three levels of functioning: body, activity, and participation levels. The majority of research has concentrated primarily on the impairment (body) level; however, there are a few studies focusing on the activity level. Thus, this current study is unique since all three levels of functioning are considered. Level of Body Body level measurements included traditional type measures so that results could be compared with those from other studies: narrow transcriptions of pre- and post-therapy assessment word lists, narrow transcriptions of audio- 29 taped speech and computer sound files, and pre- and post-therapy palatograms. Of particular interest were speech sound classes which were identified as areas of difficulty in other studies. These included lingual stops, fricatives, and affricates. Research questions and their associated hypotheses were as follows: 1. Do narrow transcriptions of the post-therapy single word list show improvement in terms of percent consonants correct and percent vowels correct, and in terms of feature matches with adult targets, compared to pre-therapy single word list? Null Hypothesis 1: There will be no difference between narrow transcriptions of the post-therapy single word list and the pre-therapy single word list in terms of percent consonants correct, percent vowels correct, and for feature matches with adult targets. 2. Do narrow transcriptions of the goal target phonemes in palatometry therapy show improvement from goal target phonemes in non-palatometry therapy? Null Hypothesis 2: There will be no difference between the number of correctly produced goal target phonemes in palatometry therapy compared to the number produced in non-palatometry therapy. 30 3. Do post-therapy palatograms show a reduction in impairment compared to pre-therapy palatograms? Null Hypothesis 3: There will be no difference between pre- and post-therapy palatograms in terms of contact patterns. 4. Do post-therapy palatogram sound file recordings show a reduction in impairment compared to pre-therapy palatogram sound file recordings? Null Hypothesis 4: There will be no difference between post-therapy and pre- therapy palatogram sound recordings in terms of narrow phonetic transcriptions. Activity Level Intelligibility measures used for this study were similar to the measures reported by Monsen (1981), Williams (1998), and Yorkston and Beukelman (1981a). For example, Williams (1998) used both open-set word and sentence identification tasks to investigate level of activity. However, because both Monsen (1981) and Yorkston and Beukelman (1981a) concluded that closed-set identification tasks were useful for speakers with severe hearing-impairments or with severe dysarthria (such as Samantha), this study also included a closed-set word identification task. If results from untrained listeners indicated improvement in intelligibility post-therapy, then it would suggest that activity limitations might be reduced 31 and that therapy facilitated Samantha's speech production at the communication level. Research questions and their associated hypotheses were as follows: 5. Do untrained listeners correctly identify more post-therapy words than pre- therapy words in the open-set word identification task? Null Hypothesis 5: There will be no difference in word identification accuracy between the pre- and post-therapy samples. 6. Do untrained listeners correctly identify more post-treatment sentences than pre-treatment sentences in the open-set sentence identification task? Null Hypothesis 6: There will be no difference in sentence identification scores between the pre- and post-treatment samples. 7. Do untrained listeners correctly identify more post-treatment target phonemes than pre-treatment target phonemes in the open-set word identification task? Null Hypothesis 7: There will be no difference in phoneme identification accuracy between the pre- and post-treatment samples. 8 . Do untrained listeners correctly identify more post-treatment generalization target phonemes than pre-treatment generalization target phonemes in the open-set word identification task? 32 Null Hypothesis 8: There will be no difference in phoneme identification accuracy between the pre- and post-treatment samples. 9. Do untrained listeners choose more post-therapy words than pre-therapy words as being correct in the closed-set identification task? Null Hypothesis 9: There will be no difference in word choice between the pre- and post-therapy samples. 10. Do untrained listeners correctly identify more post-therapy treatment phonemes than pre-therapy treatment phonemes in the closed-set word identification task? Null Hypothesis 10: There will be no difference in phoneme identification accuracy between the pre- and post-therapy samples. 11. Do untrained listeners correctly identify more post-treatment generalization target phonemes than pre-treatment generalization target phonemes in the closed-set word identification task? Null Hypothesis 11: There will be no difference in phoneme identification accuracy between the pre- and post-treatment samples. 33 Participation Level Participation level measures included questionnaires administered to listeners and to Samantha. Research has shown that self-administered questionnaires provide an indication of a person's activity and participation limitations and perceptions of his/her impairment; however, few studies, if any, have investigated listeners' perceptions of the person's speech and or hearing impairment. Therefore, this study addressed both the speaker's and the listeners' perceptions. Research questions and their associated hypotheses were as follows: 12. Does Samantha indicate an improvement due to changes in her speech on her level of participation, as based on responses from questionnaires? Null Hypothesis 12: There will be no improvement indicated in Samantha's responses to questionnaires addressing level of participation. 13. How do listeners perceive their level of participation with Samantha based on their perceptions from the listeners' tasks? Null Hypothesis 13: There will be no evidence of participation with Samantha in contexts involving oral modes of communication. 34 CHAPTER TWO METHOD This chapter describes the experimental set-up and methods used for the evaluation of each of the three areas: body (structure and function), activity, and participation. The first section briefly describes the participant of this study, Samantha. Subsequently, details on Samantha's 1996 pre- and post-treatment assessment audio-recordings are provided. SPEAKER INFORMATION In order to learn more about individual difference with regard to impairment (body), activity, and participation levels, it was decided that the focus of the present study would be with a single subject. Large group design studies are useful in learning about general trends; however, single subject studies can provide more detailed information in terms of individual difference. Furthermore, using one subject would allow for more information to be gathered on a variety of issues (i.e., level of activity and participation within society). Samantha, one of the speakers from Williams' (1998) study, participated in the current study. Samantha arrived in Canada at the age of ten years. Her first language is Cantonese, but she has been learning English since the age of 35 nine years. Samantha's mother reported that she was exposed to rubella during pregnancy with Samantha. As a consequence, Samantha has had bilateral severe-to-profound hearing loss accompanied by oral musculature weakness since birth. She has worn bilateral hearing aids since the age of two years. She began wearing a palatal lift at fifteen years due to paralysis of the soft palate. The objective of the palatal lift is to improve resonance by displacing the soft palate to the level of normal palatal elevation (Esposito, Mitsumoto, & Shanks, 2000). With the palatal lift in place, hypernasality should be eliminated or reduced. She received speech-language services for three years while in her native country and for the past ten years in Canada. From age 12 to 17 years, Samantha attended a School for the Deaf where total communication training was emphasized, and later, from age 17 to 20 years, she attended highschool classes where the focus was primarily oral communication. At the age of 21 years, Samantha began studies at a community college where she is currently continuing her studies. Samantha received palatometry therapy at the University of British Columbia's School of Audiology and Speech Sciences for an eight-month period in 1996 (Bernhardt, Loyst, & Muir, 1996). Palatometry goals for Samantha in 1996 included the following speech sounds: /r/, fkl, III, and Isl. Subsequently, Samantha received speech services at a local hospital for 20 months, during which time, therapy goals focused on the targets Samantha had achieved with the palatometry project in more conversational contexts, lxl, fkl, iy, Is/, along with voiced stops, Idl and Igl. At the conclusion of this therapy, the intelligibility of Samantha's production of single words (given a multiple choice response format) and sentences were 66% and 61% respectively, based on results from the AIDS (Yorkson & Beukelman, 1981a). Samantha reported to the Speech-Language Pathologists that she would like to focus more on vowels, Igl, /j/ and M3/. A summary of Samantha's case history is outlined below: • Prenatal exposure to the rubella virus. • Low birth weight. • Intensive care for one month post-natally. • Identification of congenital heart disease, severe hearing loss, and left side facial paralysis post-natally (specifics of impairments unknown). • Received bilateral hearing aids at approximately two years of age. • Received a palatal lift at fifteen years due to paralysis of soft palate. • Received speech-language therapy for three years from age six to nine in her native country, and for ten years in Canada from age 12 to 22. • Attended School for the Deaf from age 12 to 17, where total communication training was emphasized. • Began an oral communication program at a local highschool at the age of 17 years. • Began attending a community college at the age of 21 years. 37 • Received hospital-based therapy intermittently from age 21 to 23 years. • Participated in palatometry therapy at age 18 and again at 23 years. Samantha's 1996 Pre- and Post-Treatment Assessment Audio-Recordings This section describes the narrow transcriptions from Samantha's pre- and post-assessment audio-recordings that occurred in 1996 (Bernhardt, Loyst, & Muir, 1996). The purpose of this examination is to provide additional background information with respect to Samantha's speech sound production disorder. Narrow Transcriptions of the Pre- and Post-Assessment Word Lists Recorded in 1996 Thirteen words from the pre-therapy assessment session were narrowly transcribed by two experienced transcribers. The words were reportedly extremely difficult to transcribe. The only correctly produced phonemes were [f] in off and [t] in boots. The following substitutions were made: Itl --> [p'] (slightly ejective), /rj/ --> [n] (short) and [xkx], /ml --> [ph'] (ejective), IQI --> [f], Ibl --> [p], Isl - •> [S<] (forward, ungrooved), Itl -> [d], Izl -> [dj ], /!/ -> [1-] (barred 1, lateral fricative), IkJ -> [qX], InJ --> [x], /sp/ --> [ts], Ihl --> [X], Igl --> bilabial fricative, /br/ --> [p], III --> [ts], lxl --> bilabial fricative + [w], and /w/ --> 0 . There were no correct 38 vowel matches with the adult target. The types of errors Samantha made have been reported to be common in the speech of persons with severe to profound hearing impairments (Dagenais, 1992; Cowie & Douglas-Cowie, 1992; Parsloe, 1998). Samantha received palatometry therapy for an eight month period. Palatometry goals at this time included Iri, fkl, III, and Is/. The post-assessment word list contained 26 words which were transcribed by the previous two experienced transcribers. A comparison of her pre- and post-assessment transcriptions indicates the types of gains Samantha achieved during the eight months of therapy in 1996. Generally, Samantha improved her productions in all sound classes. For comparative purposes, we focus here on stops, fricatives, and affricates because they were targets in this study. In terms of oral stops, Samantha tended to substitute [p] for Itl and Ihl pre- therapy. Post-therapy however, Samantha correctly produced a [b] two out of four times and there were no instances in which [p] replaced Itl. The velar stops, fkl and Igl were clearly not established pre-therapy but fkl was emerging post- therapy (e.g., [k] for snake and an ejective fkl for candle). Fricatives also showed improvement from pre- to post-therapy. The fricative lit was already established. Her post-assessment consonant inventory included additions of [9], [s], [z], and [t$] (tense). Clearly, Samantha broadened 39 her place of articulations for the fricative sound class. The palatoalveolar fricative [$] and the affricate [t$] were either produced with no grooving or were very tense. Based on her productions for these two sounds, it may be assumed that Samantha was more aware of the correct place of articulation; however, she had difficulty combining the correct place (e.g., [coronal]) with the correct amount of affrication or frication (e.g., [grooving]). Improvement was also noted for the feature [voice]. Pre-therapy, Samantha produced the voiceless cognate for targeted voiced consonants. For example, for both boot and brush. Samantha produced a [p] word-initially. If a voiced consonant occurred word-finally, Samantha would omit the consonant (e.g., spoon [t S>ipu] where > represents forward place of articulation). Post-therapy however, Samantha correctly produced all voiced labials (e.g., boot [but ]) and she rarely omitted a word-final voiced sonorant (e.g., spoon [supun]). Thus, some improvement was made from the 1996 pre-assessment to the post-assessment recording. Samantha still however, presented with a very severe speech sound production disorder at the end of the treatment period. A comparison of Samantha's consonant and vowel inventory from 1996 with her pre-assessment inventories from this study, shows that each of the missing phonemes at the end of that study, except for /d.3/, were present at least once during this study's pre-assessment recording (e.g., /rj/, /0/, /ts/, /d/-clusters, 4 0 /b/-clusters/, /g/-clusters, /s/-clusters, Iri, /r/-clusters, /v/, Izl, /eel, lal, l\>l (barred i), loi, /ou/, /au/ and /ai/). Thus, this suggests that therapy provided between the end of the 1996 treatment period to the beginning of this study helped Samantha to learn to produce the missing phonemes listed above, but that she continued to have articulatory needs. THERAPY The following section describes the therapy and generalization targets and the procedures used in the treatment sessions. Therapy & Generalization Targets Four phoneme targets were selected for therapy: It ll, III, M3/, and Igl. Generalization targets included fkl, Is/, and Izl. (The rationale for selecting these particular targets is provided in the Results Chapter following the pre-therapy assessment discussion. The targets were chosen based on results from the pre- therapy assessment and therefore are explained in greater detail following the pre-therapy discussion). Over an eight week period, therapy targets were incorporated so that each target could be re-addressed. Mastery of a given sound was not required before the introduction of a subsequent sound; this approach appeared to reduce frustration and anxiety. The order of goals was as follows: Therapy Session Phoneme Goal Targeted in Therapy 1 /g/ 2 ity 3 /dz/ 4 5 ity 6 /dz/ 7 in 8 /g/ Session Structure The first four therapy sessions began with a seven-minute tongue warm-up exercise in front of a mirror in order to increase tongue movement, tongue strength, and to improve articulatory awareness (as suggested by Hardcastle, Gibbon, & Jones, 1991). The mirror was used to provide visual feedback regarding tongue movement and placement. Exercises included elevating the tip and the back of the tongue, moving the tongue from side to side, tongue protrusion, and touching the back of the upper and lower teeth with the blade of the tongue (Dworkin & Culatta, 1996; as suggested in Rammage, 1996). 42 Samantha was asked to practice these exercises everyday, at home, for five to ten minutes. After the warm-up exercises, audio-recordings were made of ten monosyllabic words which included the target phoneme for that session. Audio- recordings were conducted in a clinician's office because of lack of availability of the sound-proof booth. All target phonemes were in word-initial position because research has shown that palatogram analyses of the onset, closure, and release phases of word-final consonants may be more difficult than in word-initial position. Barry (1993) stated that ". . . in word-final positions the beginning and end points of the phases are not so easy to define" (Barry, 1993, 277). Furthermore, these points are according to Barry (1993), even "more difficult to identify with data from speech disordered subjects are being analysed" (1993, 277). The ten therapy words were different from the words used for the pre- and post-therapy assessment audio- or palatometer recordings, in order to avoid any learning effects (see Appendix B for therapy monosyllabic word list). In addition, these words were selected to sample a variety of vowel sounds. These first ten words were recorded without any verbal cues or feedback from the clinician. Then therapy began without the palatometer, using more traditional, aural/oral production techniques and imitation of the experimenter (e.g., Ling, 1976). For approximately twenty minutes, the first five words from the monosyllabic word list were taught and practiced. Visual feedback was made available by means of a mirror and by articulation modeling by the clinician. Verbal and tactile cues 43 were incorporated such as pointing to parts of the tongue, as well as positive and corrective verbal feedback. Instruction was required regarding place of articulation (i.e., tongue blade back and up, tongue tip raised, etc.), manner of articulation (i.e., voicing, nasality, air-flow, rate of speech, etc.), posture (i.e., sitting in an upright position to improve breath support), and listening (i.e., listening and watching the clinician model the target words). To facilitate articulation of the target phoneme the target sound was at first isolated from the remaining sounds within the word. For example, for the therapy word, gum. Samantha had to first articulate the Igl, pause, and then say "um." Eventually, the separation time between the target phoneme and the remaining sounds in the word was reduced so that the word was vocalized at a more normal rate. Samantha reported that the separation of the initial consonant from the vowel facilitated her articulation of the therapy words. Once all five words had been practiced, therapy commenced with the palatometer. Since Samantha had participated in a previous palatometer study, she was familiar with the workings and use of the palatometer. The goal of this part of the session was to improve articulation of the remaining five words and of the target phoneme from the therapy word list for that session. Generally, Samantha's tongue-palate contact for each goal phoneme had to be modified slightly from the model in order to resemble the target contact pattern. For example, because Samantha had left-side tongue weakness, she consistently had to increase left-side lingual-palate contact both laterally and medially in order to 44 better match the target contact pattern. Palatometry therapy began with Samantha articulating three CV syllables in which the consonant was the target phoneme and the vowels were /a/, Ixxl, or /i/ (i.e., /gi/, /ga/, and /gu/) (Dagenais, 1995). The purpose of this task was to increase articulatory awareness of the word-initial phoneme in order to facilitate phoneme production. Additionally, practicing target sounds in different vowel contexts was considered important because it facilitated awareness of the different palate regions (i.e., alveolar, palatal, and dorsal), and it provided an estimate of the variability of Samantha's CV productions (Hardcastle, Gibbon, & Jones, 1991). Furthermore, by using the playback speaker to listen to all CV articulations, auditory awareness of Samantha's productions was increased. Lastly, the CV practice allowed Samantha to practice comparing her palatograms to the static target palatograms made by the experimenter's supervisor (both palatograms were displayed adjacent to each other on the computer screen) (Dagenais, 1995). Each CV target was articulated and compared to the clinician's target about two or three times. Once completed, the remaining five words from the therapy word list for that session, were practiced. Verbal cues, feedback and instruction remained the same as for the non-palatometer therapy. Samantha's palatograms were compared to the clinician's, and target contact and non-contact patterns were indicated by the clinician. When necessary, palatograms from the palatometry manual were shown to Samantha to supplement the palatograms displayed on 45 the computer screen. These visual templates indicated typical target contact patterns of various vowels and of the target phonemes. Acceptability of Samantha's target words was judged by the clinician by comparing her palatograms to the adjacent target templates, and by comparing Samantha's audio productions with the clinician's auditory productions (via the playback speaker). Judgment was based on whether a production was in close approximation of the linguapalatal contact pattern. As Samantha participated in more sessions, she became better able to judge whether her own palatograms closely matched the target palatograms. All palatograms produced by Samantha were saved along with their sound files. Post-therapy audio-recordings of the ten words for that particular session were then made at the end of the session. As in the pre-therapy recordings, verbal cues or feedback from the clinician were not provided at this time. The remaining four therapy sessions generally followed the same procedures as in the first four sessions except for a few modifications: palatometry therapy was conducted before non-palatometry therapy; an additional five minute warm-up exercise was introduced; and target words were incorporated into a two-word phrase. The following rationales apply to the above: • In order to determine more accurately the effectiveness of therapy both with and without the palatometer and to minimize the possibility of a practice effect (i.e., memorizing a pattern), the order in which the therapies were conducted was reversed for the latter therapy sessions. 46 The purpose of the additional five-minute "warm-up" exercise was to improve sensory awareness of the tongue and palate along the midline and sides, and to facilitate tongue and lip muscle awareness (Dworkin & Culatta, 1996). For example, the experimenter gently stroked Samantha's tongue and palate, with the end of a tongue depressor. This was repeated four to five times. The "warm-up" exercises were conducted because it was felt by the experimenter following the initial four therapy sessions that Samantha may benefit from oral-motor-sensory exercises that she could learn and practice at the therapy session and then practice at home. Target word lists for the last four sessions consisted of ten two-word phrases instead of monosyllabic words (see Appendix B). The purpose of employing two-word phrases was to increase linguistic complexity (Hardcastle, Gibbon, & Jones, 1991). The particular phrases were chosen because they incorporated one of the target sounds (Igl, III, /tj/ or Id^l), which were the focus of the treatment sessions, and the remaining sounds within each phrase were ones that Samantha did not have difficulty producing (as determined from the pre-therapy assessment analysis). Five of the ten two-word phrases consisted of a determiner plus the therapy word (i.e.. the jam, that Jim): the other five consisted of the therapy word plus another word that began with the same phone as the phone at the end of the therapy word (i.e., gel light, gum mom). The identical phone was used at the beginning of the second word because it 47 allowed for easier articulation from the first word into the second word. Furthermore, if the second word began with a phone that did not match the phone at the end of the initial word, then Samantha would have had to concentrate on producing two different phones. It was decided by the clinicians that using meaningless phrases was in fact more desirable since Samantha's concentration would be more on the target phoneme than on the phrase itself. OUTCOME MEASURES Measure Of Body (Structure And Function) Research questions and their associated hypotheses from the Introduction Chapter are presented along with the perceptual and instrumental measures which were used to investigate each question. Research Questions 1. Do narrow transcriptions of the post-therapy single word list show improvement in terms of percent consonants correct and percent vowels correct, and in terms of feature matches with adult targets, compared to pre-therapy single word list? Null Hypothesis 1: There will be no difference between narrow transcriptions of the post-therapy single word list and the pre-therapy single word list in terms of 48 percent consonants correct, percent vowels correct, and for feature matches with adult targets. 2. Do narrow transcriptions of the goal target phonemes in palatometry therapy show improvement from goal target phonemes in non-palatometry therapy? Null Hypothesis 2: There will be no difference between the number of correctly produced goal target phonemes in palatometry therapy compared to the number produced in non-palatometry therapy. 3. Do post-therapy palatograms show a reduction in impairment compared to pre-therapy palatograms? Null Hypothesis 3: There will be no difference between pre- and post-therapy palatograms in terms of contact patterns. 4. Do post-therapy palatogram sound file recordings show a reduction in impairment compared to pre-therapy palatogram sound file recordings? Null Hypothesis 4: There will be no difference between post-therapy and pre- therapy palatogram sound recordings in terms of narrow phonetic transcriptions. 49 Perceptual and Instrumental Measures 1. Perceptual measure: narrow transcriptions of the pre- and post-single word list data by trained speech-language pathologists (S-LPs), the experimenter and her supervisor. 2. Perceptual measure: narrow transcriptions by trained S-LPs of the goal target phonemes in therapy (palatometry and non-palatometry, traditional articulation therapy). 3. Perceptual measure: narrow transcription of the pre- and post-palatogram sound file recordings by trained S-LPs. 4. Instrumental measure: pre- and post-treatment palatogram analysis by trained S-LPs. The following section describes how the measures were evaluated. Narrow Transcriptions of Pre- And Post-Treatment Assessment Single Word Lists Pre-and post-therapy assessments were conducted three months apart. Stimuli were recorded with a Marantz tape recorder and a PZM 33-1090B microphone (located approximately six inches from the speaker's mouth) in a sound-attenuating, double-walled Industrial Acoustical Company (IAC) booth. Pre- and post-therapy recordings were made by having the participant read the Rainbow Passage (Fairbanks, 1960) and the Basic Word List for Elicitation consisting of 164 words (Bernhardt, 1990) (see Appendix A for stimuli). Narrow 50 transcriptions of the pre- and post-treatment word lists were conducted by listeners familiar with disordered speech (i.e., the experimenter and her supervisor). Transcription of data from the Rainbow Passage did not occur due to Samantha's low intelligibility. The student S-LP (experimenter) who collected Samantha's recordings completed an initial transcription, using the International Phonetic Alphabet (1989), along with some specific diacritics to characterize Samantha's speech. The supervising professor then transcribed the data. A composite transcription was derived by comparing the two transcriptions while both the student and her supervisor relistened to the tape together. Narrow Transcription of the Goal Target Phonemes in Therapy Narrow transcription (using the International Phonetic Alphabet, 1989) of the goal and generalization target phonemes from all therapy word lists was completed initially by the experimenter, and then by the thesis supervisor. It was decided that because of Samantha's very low intelligibility and because transcriptions from the assessments took a long time, broad transcription (i.e., transcription with less detail) was to be used for the vowels and for any other non-target consonants within the words. Narrow Transcription of the Pre- & Post-Palatogram Sound File Recordings The palatogram recordings were also transcribed. Again, broad transcription was used for vowels and non-target consonants, whereas goal phonemes were transcribed narrowly. 51 Pre- & Post-Treatment Palatometry Assessments And Palatogram Analysis Thirty-six words were selected from a word list for palatometric assessment recordings (Bernhardt, Muir, & Loyst, 1996). The word-list was designed to elicit a range of consonants in both word-initial and word-final positions. Selection of the 36 words was based on the chosen therapy and generalization targets (see Results section for rationale of therapy and generalization targets). The 36 words included the four targets (/$/,/t$/, M3/ , Igl) and the three generalization targets (f\d, I si, and Izf) in word-initial (WI) and word-final positions (WF) (see Appendix A for list of palatometry assessment stimuli). The chosen 36 words were composed of six words with lt$l (three WI; three WF), five words with /J/ (two WI; three WF), three words with /d3/ (two WI; one WF), six words with Igl (three WI; three WF), five words with Ik/ (three WI; two WF), six words with Is/ (three WI; three WF), and five words with Izl (two WI; three WF). Recordings for both assessments were made using the Kay Elemetrics Palatometer Model 6300. Palatograms files and their respective sound files were saved for future analysis. Feedback was not provided during the assessments. Palatogram Analysis Pre-treatment assessment (Tl) and post-treatment assessment (T2) goal and generalization target palatograms and their respective sound recordings, were transcribed and compared with model palatograms for place of articulation 52 and manner of production. In addition, the point of maximum contact (MAX) and the release (REL) phases of Ig, k, s, J, and z/ were analyzed (Hardcastle, 1972; Hardcastle, Gibbon & Jones, 1991). The point of maximum contact phase, also termed the constriction phase or the hold phase, is the phase "where the maximum contact is held with no systematic change" (Barry, 1993, 276). The release phase is "where the amount of contact is seen to steadily diminish or change towards the configuration of the next segment" (1993, 276). For the affricate targets /d3 and t $/, the stop contact phase (STOP) rather than the maximum contact phase was analyzed because each of these targets began with stop closure. The REL phase was also analyzed for both of these targets. Since software was not available to quickly extract the number of correct and incorrect palatometric contacts, analysis of the palatograms was done by hand. Linguapalatal contact patterns were examined for target contact points within the normal region of articulation for a phoneme (according to Dagenais, 1992; Byrd, 1994; Dent, Gibbon, & Hardcastle, 1995; Hardcastle, Gibbon & Jones, 1991; and the thesis supervisor's stored models). The following region of articulation criteria were used for analyzing the goal target palatograms for the palatal sibilants, /t$/, / $/, and /d3/. Criteria were based on descriptions by Dagenais (Kay Elemetrics Manual), from Hardcastle, Gibbon, and Nicolaidis (1991), palatograms of the thesis supervisor, and from analysis of Samantha's palatograms. Some criteria were created to capture Samantha's palatal shape 5 3 spefically because her palatal shape was slightly smaller than the supervisor's palatal shape and secondly, Samantha's left side tongue weakness affected all articulations. Samantha continually had difficulty obtaining maximum left side tongue-palate contact for all targets. Thus, it was decided by the experimenters that the criteria would allow for both right and left contact to be examined rather than examining the palatal area as a whole. The reader should refer to the Figure 1.0 below when reading the criteria. Figure 1.0: Model Palatogram Left Right Q Medial Region • Lateral Region 55 Pre- and post-assessment palatograms were compared using the criteria mentioned below and each criterion was rated according to the following: change, with change being positive (+ +) or negative (- +); no change but with this being positive (- +) or negative (- -); or a neutral change (+ N), meaning that change occurred but the change was neither positive nor negative. To determine whether change had occurred or not occurred from the pre-treatment (Tl) to the post- treatment (T2) palatograms, both the experimenter and the experimenter's supervisor independently analyzed each palatogram by using the set of criteria and model palatograms. Criteria for therapy targets /t$/, / and /d?/ • groove width: This applies to the release (REL) phase of the affricates and to the maximum (MAX) contact and REL phase of III. Groove width should be approximately 8-10mm in width with the voiced cognate being about 6-8mm in width (Flectcher, Dagenais & Critz-Crosby, 1991). Tongue groove does not necessarily have to be made at the vertical center line (Dagenais, no date, Kay Elemetrics Manual); however, groove width for the release phase (REL) of the affricates should be about 4-5mm more posterior from the location of the groove width for III (Dagenais, Kay Elemetrics Manual). Groove width for the REL phase palatograms of III should be greater in width (i.e., about 9-llmm) 56 and the groove width location should be even further back than for the MAX phase palatograms (Byrd, 1994). • alveolar tip region contact: This was measured by counting down the middle column from the most anterior sensor down to the fifth sensor and then counting along the sides (left and right) from the most anterior middle sensor down to the third sensor on the outside edge. Alveolar tip region contact for STOP and MAX phase palatograms should have more sensors contacted in the alveolar region than for REL phase palatograms (Byrd, 1994). For the affricates, all sensors should be contacted in the posterior region of the alveolar region during the STOP phase in order to create stop closure (Dagenais, no date, Kay Elemetrics Manual). Tongue contact for the stop components of the affricates should be located in the posterior region of this region and may even enter the proceeding palatal region (Fletcher, 1989). For the palatal sibilants, REL phase alveolar region contact should be further back within the region (or into the next region) than for the alveolar sibilants (Dagenais, no date, Kay Elemetrics Manual). See Appendix B for examples. • alveolar and palatal region medial contact, right and left side: The palatal region was measured from the base of the alveolar region to the eighth sensor down the middle column. Thus, the palatal region consisted of three sensors in the middle of the middle column (sensors six, seven, and eight). The lateral edges of the palatal region were measured from the outside sensors found at the base of the alveolar region to the sixth sensor on each 57 side, left and right. Thus, the palatal region contained sensors four, five, and six on both sides and the sensors between the sides from the base of the alveolar region to the eighth sensor down the middle column. Medial contact was measured for both the alveolar and palatal regions on the right and left sides. The medial contact region consisted of all sensors located within the medial region (see figure 1.0 above). Generally, the anterior part of the medial region was located at the base of the third sensor down the middle column, and the bottom of the medial region was located at the base of the palatal contact region (around the "middle" of the palate). Maximum medial contact should be made for Ity, I ll, and M3/. However, because /t$/, / ll, and M3/ are produced with a groove (i.e., no contact) down the center of the palate, contact in the center of the alveolar and palatal contact region should not occur (Dagenais, no date, Kay Elemetrics Manual). Thus, contact was measured along the right and left sides of the medial region (i.e., to the right of the center column and to the left of the center column). As previously discussed, the MAX and the STOP phase palatograms should have more sensors contacted along the left and right sides of the alveolar and palatal medial contact region than the REL phase palatograms (Byrd, 1994). • alveolar and palatal region lateral contact, right and left side: This was determined to be the first three rows of electrodes from the outside of the palate towards the outside perimeter of the alveolar and palatal medial 58 region (see figure 1.0 above). Maximum lateral contact on the right and left sides of the palate from the base of the alveolar contact region to the base of the palatal contact region is desired (Hardcastle, Gibbon & Jones, 1991). MAX and STOP phase palatograms should have more forward lateral contact than REL phase palatograms (Byrd, 1994). • dorsal medial region, right and left contact: This was measured from the base of the palatal region down to the last row of sensors on the palate, excluding the outside three columns of sensors. Thus, only the middle dorsal portion of the palate was measured. Since Samantha has left side tongue weakness (discussed in greater detail in the results section), less left dorsal medial sensor contact was expected than right dorsal medial sensor contact. Contact down the center row is not desirable because Itl I, I If and M3/ are produced with a central groove that runs the length of the palate (Dagenais, no date, Kay Elemetrics Manual). MAX and REL phases should have similar dorsal medial contact. STOP phase palatograms should have no dorsal medial contact (Dagenais, no date, Kay Elemetrics Manual). • dorsal lateral region, right and left contact: This was determined as the outer three columns of sensors from the base of the palatal region to the last row of sensors. Once again, due to Samantha's left side tongue weakness, a minimum of two outside columns on the left side were expected to be contacted. Contact should be from the base of the palatal region to the back 59 row of sensors and MAX, STOP and REL phases should have similar dorsal lateral contact patterns (Hardcastle, Gibbon & Jones, 1991). Criteria for the alveolar sibilant generalization targets /s/ and Izl: • groove width: This should be approximately 5-7mm in width with the voiced cognate being slightly less in width (Dagenais, Kay Elemetrics Manual, no date). Tongue groove does not necessarily have to be made at the center of the palate. Dagenais stated that "few people make the tongue groove exactly at the midline" (Dagenais, Kay Elemetrics Manual, no date). Groove width for REL phase palatograms should be slightly greater in width because there is less tongue-palatal contact (see discussion above, Byrd, 1994). • alveolar tip region contact: See discussion above for details on the alveolar region contact. For the alveolar sibilants however, alveolar tip region contact for MAX phase palatograms should have more sensors contacted in the tip contact region than for REL phase palatograms because less contact is found during the release phase (Byrd, 1994). Hardcastle et al. stated that for I si and Izl, "there is a narrow grooved configuration in the anterior two or three rows" (1991, 51). • alveolar and palatal region medial contact, right and left side: See above discussion for details on how this contact region was measured. For the alveolar sibilants, Is/ and Izl, alveolar and palatal medial contact is desired as 60 long as there is a groove configuration in or near the midline of the palate (Hardcastle et al, 1991). • alveolar and palatal region lateral contact, right and left side: Read the previous discussion on this contact region for details on measurements. Maximum alveolar and palatal lateral contact is desired (Hardcastle et al., 1991). The amount of lateral contact should vary with the phonetic context. For example, there is more lateral contact when /s/ is followed by a front vowel such as /if in 'see' (Hardcastle et al., 1991). MAX phase palatograms should have more forward lateral contact than REL phase palatograms because during the release phase there is less contact (Byrd, 1994). • dorsal medial region, right and left contact: Measurements for this region are the same as those discussed above. As mentioned previously, because Samantha has left side tongue weakness (discussed in greater detail in the Results section), less left dorsal medial sensor contact was expected than right dorsal medial contact. MAX and REL phases should have similar dorsal medial contact since less contact during the release phase is found in the alveolar and palatal regions only (Byrd, 1994). • dorsal lateral region, right and left contact: Measurements for this region are the same as those discussed above. Due to Samantha's left side tongue weakness, a minimum of two outside columns on the left side were expected to be contacted. Contact should be from the base of the palatal region to the back row on the lateral margins. MAX and REL phases should 61 have similar dorsal lateral contact since less contact during the release phase is found in the alveolar and palatal regions only (Byrd, 1994). Criteria for therapy target /g/. and for generalization target fkf • symmetry: This was measured by examining the contacts made on the left side versus on the right side. Similar left and right side contacts are desired for the velars because if contact is not symmetrical, perceptually the phoneme may be slightly distorted; however, research has suggested that in "both normal and pathological speakers, patterns of lingual contact are rarely symmetrical" (Hardcastle, Gibbon, and Nicolaidis, 1991, 258). • dorsal medial region, right and left contact: See discussion above for details on measurements for this region. Due to Samantha's left side tongue weakness, less left dorsal medial sensor contact was expected than right dorsal medial sensor contact. MAX phase should have more medial contact than the REL phase because during the release phase of velars less contact is produced (Byrd, 1994). Contact within this region is dependent upon the vowel environment. For example, fkl or Igl followed by a front vowel will have more dorsal medial contact than fkl or Igl followed by a back vowel (Dagenais, Kay Elemetrics Manual, no date). • dorsal lateral region, right and left contact: Measurements for this region are the same as those discussed above. Once again, due to Samantha's left side tongue weakness, a minimum of two outside columns on the left side 62 were expected to be contacted. MAX phase should have more anterior lateral contact than the REL phase because during the release phase of velars less forward contact is produced (Byrd, 1994). As mentioned above, contact within this region will depend on the phonetic environment. For example, fkl or Igl followed by a front vowel will have more dorsal lateral contact than fkl or Igl followed by a back vowel (Dagenais, Kay Elemetrics Manual, no date). • back row contact: Back row contact consisted of all sensors along the bottom/back row. Ideally, all sensors on the back row should be contacted for both phases (Dagenais, Kay Elemetrics Manual, no date; Byrd, 1994). MEASURE OF ACTIVITY In order to measure level of activity, five listeners from Williams' study (1998) were asked to complete three tasks: open-set word identification tasks, an open-set sentence identification task, and a closed-set word identification task. The following section describes the requirements of the listeners. Subsequently, the research questions and their associated hypotheses for the activity level are presented. The experimental set-up for the listener tasks is discussed and each task is described in detail. 63 LISTENERS Five listeners, three males and two females, were recruited from Williams's (1998) study so that results from this study, could be compared with those from the previous study. Only five of the original sixteen listeners were available to participate in the present study. All listeners had completed high school, were between 17 and 40 years of age, had normal hearing and spoke English as a first language. The mean age of listeners was 25 years and the range of ages was from 19 to 31 years of age. Each listener signed a consent form explaining the purpose and requirements of the study. Listeners were required to attend two, one-hour sessions, with the sessions being a minimum of three weeks apart in order to minimize any learning effects. After each session, listeners were compensated with 15 dollars. During the first session, listeners were asked to take a hearing test which included pure-tone air conduction audiometry, speech discrimination, and speech recognition testing. Passing criteria included: (1) pure-tone air conduction thresholds at or below 20 dBHL at 500, 1000, 2000, 4000, and 6000 Hz, (2) a speech recognition threshold at or below 20dBHz, and (3) a speech discrimination score over 88%. Each listener passed the eligibility criteria. Listener characteristics are outlined in Table 1.0 below: 64 Table 1.0: Characteristics of listeners Listener identification Age Education Occupation Gender L l 22 High School Snowboard instructor Male L2 27 B.Sc. Web Designer Female L3 19 High School Cashier Female L4 26 Diploma Audio-Visual Technician Male L5 31 B.FA Theatre Artist Male Research Questions 1. Do untrained listeners correctly identify more post-therapy words than pre- therapy words in the open-set word identification task? Null Hypothesis 1: There will be no difference in word identification accuracy between the pre- and post-therapy samples. 2. Do untrained listeners correctly identify more post-treatment sentences than pre-treatment sentences in the open-set sentence identification task? Null Hypothesis 2: There will be no difference in sentence identification scores between the pre- and post-treatment samples. 65 3. Do untrained listeners correctly identify more post-treatment target phonemes than pre-treatment target phonemes in the open-set word identification task? Null Hypothesis 3: There will be no difference in phoneme identification accuracy between the pre- and post-treatment samples. 4. Do untrained listeners correctly identify more post-treatment generalization target phonemes than pre-treatment generalization target phonemes in the open-set word identification task? Null Hypothesis 4: There will be no difference in phoneme identification accuracy between the pre- and post-treatment samples. 5. Do untrained listeners choose more post-therapy words than pre-therapy words as being correct in the closed-set identification task? Null Hypothesis 5: There will be no difference in word choice between the pre- and post-therapy samples. 6. Do untrained listeners correctly identify more post-therapy treatment phonemes than pre-therapy treatment phonemes in the closed-set word identification task? Null Hypothesis 6: There will be no difference in phoneme identification accuracy between the pre- and post-therapy samples. 66 7. Do untrained listeners correctly identify more post-treatment generalization target phonemes than pre-treatment generalization target phonemes in the closed-set word identification task? Null Hypothesis 7: There will be no difference in phoneme identification accuracy between the pre- and post-treatment samples. Experimental Set-Up for Listener Tasks Twenty words from the pre- and post-treatment assessments were selected randomly for digitizing (Bernhardt, 1990 Word List). In addition, five stimuli (three from the post-treatment assessment, T2, and two from the pre- treatment assessment, Tl) were selected as "warm-up." This set of stimuli was presented to each listener before presentation of the actual test stimuli. The purpose of the "warm-up" stimuli was to allow each listener time to become familiar with the speech of the speaker and the task itself. Twelve of the twenty words contained the goal target phonemes, Igl, /J/, /t$/ and M 3 / , and the remaining eight words contained the generalization target phonemes, Is/, Izl, and fkl. In addition, three T l and three T2 sentences from the Rainbow Passage were selected for digitizing. The same sentences were selected from the T l and T2 recordings. All stimuli were selected in a random fashion; however, the three selected sentences had to be similar in terms of the number of words contained within each sentence. In addition, because of Samantha's highly unintelligible, 67 slow, and effortful speech it was felt that recording sentences as opposed to conversation would be less stressful for Samantha. All stimuli were digitized onto a NeXT computer system sound-recording programme, Sound Works 3.0 Version 2. The sound files were recorded at a sampling rate of 22,050 Hz and stored on the hard disk of the computer. After recording, the gain of each sound file was increased to the maximum point where clipping did not occur. The sound files were then saved once again. Output of NeXT was rooted to an audiometer and delivered over headphones. Calibration of the Sound Level for the Words and Sentences A non-selected word from the post-treatment assessment word list was calibrated using an in-house calibration programme which calculates the root mean square (RMS) of the sound pressure level of a speech signal contained in a sound file (Liang, 1994). Once the RMS scale value of the speech signal has been calculated, a second sound file is created. This file contains a calibration tone of a user-specified frequency and a user-specified duration. The RMS scale value of the calibration tone equals the RMS scale value calculated for the speech signal. The calibration signal allows each separate recording to be presented at approximately equal levels. The calibration signals created for the selected stimuli in this study were ten-second 1000 Hz tones. 68 In order for the RMS programme to process a sound file, a one-second silent segment at the beginning of the sound file must be inserted, followed by a two-second sample of typical background noise. The one-second silent interval was generated using Sound Words 3.0 Version 2 and the two-second sample of background noise was copied into position from the end of the sound file. Once completed, the RMS programme instructs the user to specify an amplitude threshold and then a duration threshold. The RMS programme eliminates any segment from the calculation of the RMS scale value for the sound file if the segment has an RMS value below the amplitude threshold value for a duration greater than the duration threshold. The programme counts these segments as pauses, includes only the amplitude of the speech signal in the calculation of the RMS scale value, and excludes from the calculation the amplitude of the background noise that is present in pauses. The user must choose a value between 1 and 10 to specify the amplitude threshold level. The programme then calculates the RMS scale value for the background noise of the sound file based on the two-second inserted sample of typical background noise. The resulting amplitude threshold is equal to the RMS scale value for the two-second sample of the recording's background noise, multiplied by a factor of 1 to 10 as set by the user. To specify the duration threshold, the user sets a duration threshold in milliseconds. An amplitude threshold of 2 and a duration threshold of 500 msec was used in the present study. Thus, the programme identified as pauses all portions of the sound file in 69 which the amplitude value never exceeded two times the RMS scale value of the background noise for an interval of at least 500 msec. The sound file containing the calibration tone was always played before presenting the stimuli to the listener. The intensity of the calibration tone was used to adjust the level of the output of the audiometer. Each recording was presented to the listener at 70 dBSPL (50 dBHL on the audiometer dial), the average level of conversational speech (Davis, 1947). All stimuli were presented to the listeners in the sound-attenuating, double-walled IAC booth. Participants listened through binaural headphones on one side of the booth while the experimenter was positioned on the other side of the booth. Listener Tasks Each listener was asked to participate in two one-hour sessions three weeks apart so as to minimize practice effects. Table 1.1 below outlines the order of events during each session: 70 Table 1.1: Order of events during listener sessions one and two Listener Session One Listener Session Two 1. Hearing Test (if all criteria met, continue with the session). 1. Open-set word identification task. 2. Brief break. 2. Open-set sentence identification task. 3. Open-set word identification task. 3. Brief break. 4. Questionnaire part 1: Word Task 4. Closed-set word identification Task (Pre-assessment words, randomized). 5. Open-set sentence identification task. 5. Questionnaire part 3: Activity/Participation 6. Questionnaire part 2: Sentence Task. 6. Closed-set word identification task (Post-assessment words, randomized). Open-Set Word Identification Task The goal of the open-set word identification task was to investigate if the listeners correctly identified more post-therapy goal phonemes than pre-therapy goal phonemes. This task was presented before the closed-set word identification task because with the latter task, listeners were presented with a list of words containing the target word and four to five alternative words. If the closed-set word task was administered prior to the open-set word task, then 71 more of a learning effect might occur. Listeners could have remembered the words from the closed-set task and recorded these words for the open-set task. Thus, in order to avoid a learning effect the open-set task was administered before the closed-set task. Forty words, twenty of which came from the pre-assessment recording (Tl), and an identical twenty from the post-assessment recording (T2), were selected for this task. The selected words were chosen because they contained both generalization and target sounds in word-initial and word-final position, and there was a variety of both mono- and bi-syllabic words. During the listeners' first visit, each listener was presented with five warm-up words followed by twenty stimulus words of which ten were from T l and ten from the T2 recordings. The purpose of the warm-up words was to allow listeners time to become familiar with disordered speech (i.e., the speaker's speech) and the task itself. The selection of the words was limited by the therapy targets. Twenty stimulus words were used because this was the maximum number of words that could reasonably be presented in twenty minutes. The twenty words were presented in a randomized fashion in order to prevent a practice effect. Similarly, during the second visit, each listener was presented with the same five warm-up words followed by the remaining twenty words (remaining ten from T l and the remaining ten words T2). Listeners were asked to record by means of orthographic transcription the sounds or the word they thought they heard for each word presented. Each stimulus word was presented 72 twice in order to allow each listener a second chance to hear the word (similar to asking for repetition during conversation) and to lessen demands on auditory memory. (Monsen [1983] showed that if listeners heard two presentations, rather than only one of the test sentence, then intelligibility scores generally improved). There was a ten-second interval of silence between each presentation of the stimulus word in order to provide each listener with enough time to record his/her answer. Listeners were asked to record two answers for each word presented (one answer for each presentation) and the two answers were allowed to be identical or different. Since Samantha's speech was highly unintelligible even to a trained listener, listeners were encouraged to listen to the sounds within the word and or to the word Samantha was saying. It was felt by the experimenters that by providing the listeners with two options (i.e., sounds and or words) for recording their answers, the listeners would be more encouraged to record an answer as opposed to recording nothing. Furthermore, some listeners might find it easier to listen for sounds instead of the word and vice versa. The Discussion chapter of this thesis elaborates on this issue. In addition, each listener was asked to rate after each word how well s/he understood what the speaker said by circling one of the following options: understood very well, understood moderately well, understood hardly anything, did not understand anything. 73 Open-Set Sentence Identification Task Six sentences, three from the T l recording and the same three sentences from the T2 Rainbow Passage recordings, were chosen as stimuli. Three sentences were selected because that was the maximum that could be presented in ten minutes. These three selected sentences contained the fewest self- corrections, if any. The selected sentences were matched as closely as possible according to the mean length of utterance (TVILU). During the first visit, each listener was presented with three of the six sentences in a randomized fashion in order to avoid practice effects. During the second visit, each listener heard the remaining three sentences that were again presented in a randomized fashion. The procedure for the sentence identification task was the same as for the word identification task; however, listeners heard each sentence a total of four times in order to provide each listener with ample listening time. Lastly, as in the word identification task, each listener was asked to rate, after each sentence, how well s/he understood what the speaker said for each sentence. Closed-Set Word Identification Task The goal of this task was to determine if listeners correctly selected more post-therapy words than pre-therapy words when provided with a number of alternatives. Williams' study (1998) did not include a closed-set word 74 identification task. However, because research on closed-set word identification tasks indicated that closed-set tasks may provide more meaningful information about intelligibility with very unintelligible speakers (see discussion in introduction section), it was decided that a closed-set task should be conducted due to Samantha's low intelligibility (as reported in Osberger, 1992; Boothroyd, 1985; Monsen, 1981). This task was presented during the listeners' second visit after the second administration of the open-set word identification task. The same forty words from the open-set identification task were used for this task in order to compare results between the two different tasks. Listeners were presented with forty words twenty of which were from the pre-treatment assessment (Tl) recording, and twenty of which were from the post-treatment assessment (T2) recording (see Appendix A for words). The selected twenty T2 words were the same words selected from the T l recording. Each listener listened to the first twenty T l words through headphones in the IAC booth. Listeners were asked to select the word that they thought they heard from four to five alternatives. The chosen alternative words consisted of similar consonants and vowels to the stimulus word. The alternative words were made up by the experimenter and her supervisor. Listeners were then asked to complete a brief questionnaire before listening to the remaining twenty T2 words. Each stimulus word was presented a total of two times so as to mimic a request for repetition in conversational speech. Each presentation of the T l and T2 words was randomized in order to minimize practice effects. 75 MEASURE OF PARTICIPATION Research questions and their associated hypotheses are presented below followed by a discussion on the questionnaires administered to Samantha and the listeners. 1. Does Samantha indicate an improvement due to changes in her speech on her level of participation, as based on responses from questionnaires? Null Hypothesis 1: There will be no improvement indicated in Samantha's responses to questionnaires addressing level of participation. 2. How do listeners perceive their level of participation with Samantha based on their perceptions from the listeners' tasks? Null Hypothesis 2: There will be no evidence of participation with Samantha in contexts involving oral modes of communication. In order to examine Samantha's level of participation, she was administered two questionnaires approximately three months after her palatometry therapy sessions. The questionnaires were administered three months post-treatment because the experimenter wanted to provide Samantha with a considerable amount of time away from palatometry treatment in order for her to reflect upon the treatment and any participation changes, if any, the 76 treatment may have influenced. Three questionnaires were used in order to obtain a variety of responses with respect to Samantha's level of participation and her perceptions of her speech. First, a palatometry research questionnaire was administered which contained questions developed by the experimenter and by Samantha's former palatometry therapists/experimenters (Bernhardt, Muir, & Loyst, 1996). This questionnaire was previously administered twice during Samantha's original (1996-1997) palatometry therapy (four and eight months post-therapy). It was decided that Samantha would be given the same questionnaire during the present study in order to compare her answers from the previous two administrations and to determine if she indicated any change or improvement as result of therapy. Additional questions were asked during the present administration because the experimenter wanted to obtain a better understanding of Samantha's perceptions of her speech, her speaking abilities, and her level of participation within society. Questions focusing on level of participation were created by the experimenter and her supervisor because it was felt that the questions had to be specific to Samantha's age, her interests, and her likes and dislikes. All participation questions were asked in the form of "do you find you avoid any of these situations because of your hearing loss." Questions which focused on Samantha's perceptions of her speech and her speaking abilities were the same as those from a questionnaire created for speakers with dysarthria by Yorkston, Bombardier, and Hammen (in Till, 77 Yorkston, & Beukelman, 1994). Ten questions from this questionnaire were used because they were designed for speakers with dysarthria such as Samantha's. When answering all questions, Samantha circled one of the following answers from two options depending on the type of question: (1) a lot, a medium amount, a little bit, or not at all; (2) always, some of the time, hardly ever, or never. For example, when the question was time related such as "I can usually make strangers feel at ease with me" the options for answering included (2) always, some of the time, hardly ever, or never. If the question however was quantity related such as, "my speech will improve if I work hard" the options included, (1) a lot, a medium amount, a little bit, or not at all. Thus, the length of the questionnaire administered for this study was slightly longer than the previously administered questionnaires. Generally, the purpose of the palatometry questionnaire was to examine whether Samantha noticed any changes with respect to her speech as a result of the palatometry therapy, and to gather Samantha's opinion on her perception of her speech and her participation within society. Secondly, 34 questions from The Communication Profile for the Hearing Impaired (CPHI, Demorest & Erdman, 1986, 1987) and from Erickson's S-24 Scale (S-24 Scale, Erickson, 1969) were administered (see Appendix C for questionnaires and see Introduction Chapter for details on the focus of these questionnaires). Thirty-four questions was the maximum that could be administered in 30 minutes. Questions from the CPHI were selected for use with 78 Samantha because they focused on psychological functioning and level of participation within society. Similarly, questions from the S-24 Scale were selected because they focused on participation and they provided a measure of attitudinal dimensions. Even though Erickson's S-24 Scale was devised for use with stutterers, the experimenters felt that questions from the S-24 would be suitable for use with Samantha. Furthermore, because the S-24 Scale focuses on a person's self-concept as a speaker and on attitudinal changes in speech behaviour that may or may not have resulted because of therapy, it was thought to provide further insight into Samantha's level of participation within society. Rather than give Samantha the entire CPHI and the entire S-24 Scale, the experimenter and her supervisor selected questions that would be of relevance for Samantha and level of participation. For example, questions that did not focus on participation were not selected and questions that were particular to Samantha's family situation were not asked, because these questions were felt to be too personal. The rating scale for this 25 item questionnaire was similar to the one used in the CPHI. A value of 1 meant "rarely occurs," a 2 meant "sometimes/occasionally occurs," a 3 meant "occurs half the time," a 4 meant "sometimes frequently occurred," and a value of 5 represented "occurred almost always." This rating scale was used because it provided a range of variables and it was efficient to use. Prior to answering each questionnaire, Samantha was instructed to answer "any, all, or none of the questions." Thus, Samantha had the option to 7 9 avoid answering a question if she felt the question was demeaning, embarrassing, or hurtful. The second level of participation measure was the listeners' perceptions. The WHO's ICIDH-2 states that "participation restrictions are problems an individual may have in the manner or extent of involvement in life situations" (ICIDH-2 Beta-2, p. 83). To obtain an idea of some of the participation restrictions Samantha may encounter, the experimenter and her supervisor decided to ask each listener for his/her point of view. Since literature on level of participation and since questionnaires addressing participation for clients with communication disorders are sparse, the experimenter and her supervisor created the listeners' questionnaire. Furthermore, because the questions had to relate to Samantha's case, for example relate to her age and to her likes and dislikes, it was decided that a generic participation questionnaire might not best capture Samantha's participation restrictions. During the listener's first visit, only two questions were asked that examined level of participation: (1) If given the opportunity, where would you feel comfortable having a conversation with this speaker? and (2) What kinds of activities would you feel comfortable participating in with this speaker? Only two questions were asked at this time because of time limitations, and because the majority of the questions on the questionnaire were about the listener tasks (i.e., the word and sentence identification tasks). Each listener was instructed to "circle one, more than one, or none of the answers." For the first question above, the options were as follows: 80 (a) at a busy restaurant, (b) at a quiet coffee place, (c) at a night club, (d) at a library, (e) other: for example , (f) nowhere I can think of. For the second question above, the options included (a) a college project, (b) a game of doubles tennis where this speaker is your partner, (c) a game of chess, (d) shopping in a mall, (e) other , and (f) none I can think of. For option (e) of both questions, listener could write in their own answer. The types of options available to the listeners were chosen because they offered a variety of activities and places. During each listener's second visit, s/he was asked 12 questions that were asked in the form of, 'How often would you . . . . ' . For example, 'How often would you like to go to the movies with this speaker?' and 'How often would you like to email this speaker?' (see Appendix C for further examples). Twelve questions were asked because that was the maximum that could reasonably be asked within ten minutes. The rating scale for each of the 12 questions included four alternative answers: (1) all the time, (2) sometimes, (3) hardly ever, and (4) not at all. This scale was used because it provided a range of answers. Each listener was instructed to complete as much or as little of the questionnaire as s/he chose. Completion of the questionnaire occurred at the end of each session in the sound booth so as to provide each listener with a quiet and private environment. The questionnaire had to be completed at the end of the session after each listener had listened to Samantha's speech. 81 CHAPTER THREE RESULTS: BODY LEVEL The current study used narrow transcriptions of pre- and post-therapy word lists, narrow transcriptions of treatment target phonemes in therapy, and palatograms to assess the effects of palatometry therapy on body level (i.e., impairment) (World Health Organization [WHO], 1999). This section outlines the research questions which motivated the analysis at the level of the body, the rationales for the treatment targets, and the experimental results. Research Question #1 1. Do narrow transcriptions of the post-therapy single word list show improvement in terms of percent consonants correct and percent vowels correct, and in terms of feature matches with adult targets, compared to pre-therapy single word list? Null Hypothesis 1: There will be no difference between narrow transcriptions of the post-therapy single word list and the pre-therapy single word list in terms of percent consonants correct, percent vowels correct, and for feature matches with adult targets. 82 This question was addressed by analyzing both the pre-therapy and the post-therapy transcriptions of the 164 word list (Bernhardt, 1990 Word List). The pre-assessment results are discussed first as a baseline comparison. Treatment targets are identified and then post-assessment results are reported. Pre-Therapy Assessment Results The pre- therapy consonant inventory is shown below in Table 2.0, the vowel inventory in Table 2.1, and the consonant and vowel percentage match with adult targets is shown in Table 2.2. Table 2.0: Samantha's pre-assessment consonant inventory over all word positions1 Note: Infrequent speech sounds are in parentheses. Labial Alveolar Palatoalveolar Velar/ Uvular Glottal Nasal m (mj) (mb) (mp) n n: (nt) (nz) (ntsss) (ntsts) (nssokh) (nP D Stop p p: (PD Ph (pr) (bz) b (bl) (bwl) (br) t th ts (tr) d (dr) (dw) k kh (kj) (kl) (ks) (qh) (kx) (kw) (kae) (g) (gD Fricative f fff fff: (fr) (v) e Or) s s: (stwo) sssb (s:k) (skr) (sel) (sssel) (sm) sn (spw) (skw) (st) (ssst) (str) (sw) z (s:z) (sts) (!) ( H X O (stj) (tS<0 (tJ) Approximant (w) (wo) (w:) 1 0 0 Ok) r (rw) (rf) (rs) Laryngeal h (?) (2m) 1 sss or n X= extra frication, < = fronted ungrooved, T = voiced interdental, o = devoiced, () = stopping 84 Table 2.1: Pre-therapy vowel inventory Front Mid Back High i i>(barred i) u I 1 9 Mid ei £ 0 ou 9 0 A ou Low ae/a a ai au Table 2.2: Pre-therapy consonant and vowel percentage match with adult targets Category Pre-Therapy Assessment Consonants 25.66% Vowels 43.04% Pre-therapy consonant match with adult targets was at a severe level of impairment-25.66%. Approximately 43% of vowels matched adult targets (a severe level of impairment). The following summarizes the pre-treatment assessment results. 85 Manner of Articulation Stops and nasals Stops were the most frequent category, matching the adult target 46.2% of the time for manner. Occasionally a nasal was inserted immediately before a stop (e.g. eat [kli(n)t],_ truck [terjkh]) or stops were omitted (e.g., musicbox [m(j)usibatsss]). Voiced stops were frequently replaced with their voiceless counterpart (e.g. go [kou]. dollhouse [ toulhAutsss]]). For example, the target phoneme Igl was replaced with either [k] or [kh] 100% of the time in all word- positions. Word-initial / k h/ was accurate in terms of manner 100% of the time and fkl was 100% accurate in both word-medial and word-final positions. Nasals matched the adult target only 19.3% of the time. The nasal [n] in word initial and medial postions was produced with the most accuracy; word initial [m] however, was sometimes replaced by [w] or [b] and word final [n] and [rj] were frequently missing (e.g., wagon [w:e?a:], hanging [haici]). Fricatives and affricates Fricatives and affricates matched the adult target for manner 24.7% of the time. When inaccurate, often an instrusive stop appeared with the fricative as in sunglasses [ (t)sAna?klets ] and sweater [stweta]. In word-initial position, both Is/ and Izl were accurate for manner 100% of the time. In word- medial and word-final positions, Is/ was accurate 80% and 75% of the time 86 respectively. In word-medial position for example, a [t] replaced /s/ and in word- final position, [th] replaced /s/. In word-final position, Izl was accurate 100% of the time in terms of manner; however it was never produced correctly in all word- positions. It was either omitted or a voiceless affricate or [s] were substituted (e.g.. cherries [tjer2ri:] and zoo [tsu:]). In addition,/s/-clusters showed differences from adult targets. The second consonant in the cluster was often omitted as in snake [s:selkh] and snow [sou] or a schwa was inserted (e.g.. sleep [ ssl ip] and sleeping [ ssss l ip i ]. In terms of manner of articulation however, /s/-clusters were accurate 100% of the time. Affricates were often produced as fricatives in all word positions, as in jump [ J Amph ], shoes [ swu:tsss ], orange [?AUinsJ], and watching [waji]. The voiceless affricate target, /t$/, was accurate for manner 37.5% of the time in word-initial position, 50% in word- medial position, and 100% in word-final position. In word-final position however, 1/2 or 50% of the targets were accurate in terms of manner. The incorrect target had a stop replacing the affricate. For /d^/, manner was accurate 100% of the time in word-initial position, 50% of the time in word-medial position, and 67% of the time in word-final position. Inaccuracies included, stops being substituted for M 3 / (i.e., d3 •--> [kw] and [thi]). 87 Approximants Approximants matched the adult target only 15.9% of the time in terms of manner. The approximants, [1] and [j] were produced appropriately, however f\J in a cluster was produced with only a trace of an "1" (e.g., plum [p(T)92mhm] and glove [g(l)a?9v]). All word-initial [r] productions had either a [2(0)] or a [29] inserted before the [r] and in 25.0% of the word-initial [r] cases, [r] was replaced with [ 2 u w ] . Word-final [r] was frequently omitted as in, for example, roar [ 299Wo22(x)i] and chair [$S$eu]- The [r] in r-clusters was usually either omitted or substituted with [w] (e.g., dress [ tweitss], truck [ tsrjkh ]). Overall Accuracy Organized by Place Categories Labials Labials plus labial clusters matched the adult target 37.5% of the time in terms of place of articulation. Generally, velars were substituted in place for labials or an additional consonant was added to a labial cluster, as in, for example, brush [bwlgt $ ]. Labials alone matched the adult target 43.4% of the time. The labials [b], [p], [m], [f] and [w] were the most frequent labials produced. In terms of patterns by place, occassionally a labial replaced a labial, as in, for example, me [bi: ] and milk [bjukh ]. Similarly, voicing appeared to be sometimes difficult and often the voiceless labial would replace voiced labials 88 (e.g., bootie [punttai??] and boot [punt ]). There were two instances in which the velar [k] replaced the voiceless labial [p] (e.g., pie [khai22 ] and piggy [kei??gi22 ]). Generally however, labials did not replace other places of articulation such as coronal or velar. Coronals Coronals matched the adult target 18.9% of the time in terms of the feature [Coronal]. The alveolar coronals [t], [d], [n], and [s] were the most frequent consonants produced. The palatoalveolar affricates and the voiced alveolar coronal [z], matched adult targets the least out of all the coronals. Alveolar coronals and the palatoalveolar fricative [J] tended to replace palatoalveolar affricates as in, for example, judge [ sssASt h i]. In word-initial position, the generalization targets /s/ and Izl were 100% and 50% accurate in terms of place, respectively. For word-initial Izl, [d3] replaced Izl. In addition, Izl was accurate 100% of the time in terms of place in both word-medial and word- final positions. The voiceless cognate Is/ however, was accurate 80% and 75% of the time in word-medial and word-final positions respectively. Inaccuracies included /s/ being replaced with [t J]. The affricates I til and M 3 / were accurate for place 50% of the time in word-initial and word-medial position. In word-final position, /t$/ was accurate 100% of the time and /d3/ was accurate 33.3% of the time. Inaccuracies included M 3 / being replaced with [thi]. The palatoalveolar 89 fricative III was accurate in terms of place 0% of the time in word-initial position, 100% in word-medial position, and 67% of the time in word-final position. In word-initial position, III was replaced with [sw] and in word-final position, [nts] replaced III. Thus, specific coronal place (as alveolar or palatoalveolar) was generally accurate for stops Itl and Idl, and for the fricative Is/, but was limited for Izl and III in word-initial position and for the affricates. Velars The velar [k] and its aspirated cognate [kh] were produced most frequently. The velar [g] matched adult targets only 11.1% of the time for the feature [Velar]. The voiced velar was typically replaced by [k] , [kh ] or [ 2 ] (as in wagon above) in all word positions. In word-intial position, Igl was accurate 100% of the time in terms of place. In both word-medial and word-final positions, Igl was accurate 50% of the time for place. Inaccuracies included Igl being replaced with an alveolar (i.e., [nt]) in word-medial and word-final positions, and with [2] in word-medial position. The velar [k] however, matched the adult target 70.8% of the time in terms of place. Inaccuracies included word-final [k] being produced as [kle] or as [rjkh]. The aspirated fkl (i.e., /kh/) was accurate 91.6% of the time in terms of place in word-intial position. There was one instance where fkhl was replaced with [qj. In word-medial and word-final positions, fkl was accurate 100% of the time in terms of place of articulation. Word-final nasal velar /rj/ was 90 82% accurate in terms of place of articulation. Inaccuracies included /rj/ being replaced with a glottal stop or with [n]. Voicing Voicing of stops, fricatives, and affricates was poorly established. Voiced stops, in particular voiced velars, fricatives, and affricates were usually produced as their voiceless cognates or were omitted (across word positions). Voiceless stops and fricatives were generally accurate in all word positions; however, aspiration was inconsistently produced. Sometimes no aspiration occurred and other times it did occur. The target phoneme Igl was 11.1% accurate in terms of voicing. As mentioned previously, Igl was replaced with either [k] or [kh]. There was one instance in which word-medial Igl was appropriately voiced. The generalization target fkl was accurate for voicing in all word positions, except for one instance in word-final position in which fkl was replaced with [rjkh]. The voiced affricate Id^l was never accurate for voicing in either word-initial, word-medial, or word-final position. The target was replaced with either [tJ], [s], [$] or [thi]. The voiceless affricate was 100% accurate in terms of voicing in word-initial and word-medial positions. In word-final position, there was one instance in which Itl I was replaced with [nts]. The target phoneme /J/ was 100% accurate in terms of 91 voicing in both word-initial and word-medial position. Similar to /t$/, in word- final position, [nts] replaced III. The generalization target Is/ was 100% accurate for voicing in all word positions. The target Izl however, was never voiced in either word-initial, word-medial, or word-final position. Generally, the voiceless cognate [s] or [ts] replaced Izl. Palatometry Program Based on the pre-treatment assessment results treatment targets for the palatometry program were chosen. These included the voiced velar stop Igl, the palatoalveolar fricative III, and the palatoalveolar affricates /tJ/ and M3/. Rationales for the treatment targets follow. Velar Igl The velar fkl was significantly better established than Igl. Furthermore, the voiceless velar was substituted 56% of the time for the voiced velar. Thus, Samantha was showing evidence of velar place of articulation; however, her lack of the voiced velar indicated difficulties combining the place [Dorsal] with the feature [+ voice]. Since the velars represent a frequent place category in English 92 and since Samantha was able to produce [k], the Igl was chosen as a feature combination target. Generalization Target: It was hypothesized that Ikl and /rj/ would show gains in accuracy as a result of focusing on Igl. There might also be generalization to other voiced/voiceless obstruent contrasts (i.e., s/z, p/b, t/d, etc.). Fricative III Fricatives were one of Samantha's more weakly established manner categories. Since the palatometer does not provide lip information, the labiodental fricatives /f/ and /vl were not suitable targets. The alveolar Is/ was also not chosen as a target because III was produced less accurately than Is/ and thus it was felt by the experimenters that III needed more practice than I si. Pre- therapy palatograms of III further indicated that Samantha needed to establish a more consistent groove. Some of her palatograms had either no groove (similar to an I si articulation) or too wide a groove. Thus, she needed to learn the appropriate width of the groove for /$/. Analysis of the pre-therapy transcriptions also showed that Samantha was substituting [s] for III. This finding further supported the selection of III as a therapy target. 93 Generalization Target: Since Samantha would be learning to distinguish between the alveolar and the palatoalveolar placements, and to form an accurate tongue groove, it was hypothesized that there could be improvement with her voiceless alveolar, /si. There might also be generalization to the voiced alveolar, /z/asa result of focusing on manner and place, and on voicing for Igl (and M 3 / below). Palatoalveolar Affricates It 11 and /d3/ The palatoalveolar affricates were poorly established at initial assessment. Samantha used the fricatives [$] and [s] for Itl I, and [J], [s], and [tl] for M 3 / . Pre-therapy palatograms also indicated that Samantha was often not producing appropriate closure for the stop portion of the palatoalveolar affricates and many of the fricative portions of these sounds were produced with too wide a groove. Furthermore, by targeting affricates Samantha would receive additional practice with tongue grooving for the release portion of the affricates. All these factors led to the selection of It J/ and /d.3/. Generalization Targets: Again it was hypothesized that Is/ and Izl might show gains in accuracy as a result of focusing on placement, grooving, and voicing for Itl/ and /d3/. Post-Therapy Assessment Results 94 The post- therapy consonant and vowel inventories are shown below in Tables 2.3 and 2.4 respectively. Table 2.5 displays the consonant and vowel matches with adult targets for both the pre- and the post-therapy assessments. Table 2.3: Samantha's post-assessment consonant inventory over all word positions Labial Alveolar Palato- alveolar Velar/ Uvular Glottal Nasal m (mj) (mp) (mpw) (m?m) n n: (nt) (nz) (ntsss) (ntsts) (ns9 0k h) (nS) Stop P p: tts (tr) d (dr) (dw) (t$) k kh g (kj) (ks) (qh) (kx) (PD (tXSX) ((t)ssss) Ph (pr) ((t)ssz) (kX) (kae) (bz) b ((t)sss) (g) (kl) (krl) (bl) (pw) (tskr) (kr) (go) (go h0 (br) (goX) Fricative f fff fff: (9) Or) ((f)) (fl) ((v)) (T) (vw) (v:?v) (To) s s: (s:n) (stw0) sssb (s:k) (skr) (sal) (sssal) s:n (spw) (sssp) (skw) (st) (ssst) (str) z (s:z) (sts) (s9m) (sss:k) (s:ts) (sss:t) (ssstr:) (D ass) (stj) «<t) (tS>) C3) (d3) tX 95 (sss:w) (ss:>pw) (sss:t) (sss:dak) (ss:pw) (sssp) Approximant (w) (we) GO r (rf) Laryngeal h 2 (2m) (2J) Table 2.4: Post-therapy vowel inventory Front Mid Back High i i> (barred i) u u: I 1 9 Mid ei e 0 ou u 9 0 A ou Low ae a a ai au Table 2.5: Consonant and vowel percentage match with adult targets pre- and post-therapy assessments Category Pre-Therapy Assessment Post-Therapy Assessment % Gain Consonants 25.66% 38.84% 13.18% Vowels 43.04% 52.45% 9.41% Post-therapy consonant matches with adult targets increased by 13.18%, from 25.66% to 38.84%-a slightly less severe level of impairment post-therapy. Similarly, vowel matches with the adult target showed a gain of 9.41%, from 43.04% to 52.45%. These results however, still represent a severe level of 96 impairment for the consonants and a moderate-severe level of impairment for the vowels. The subsequent tables summarize and compare the pre- and post- treatment assessment results. Table 2.6a-j: Percentage match of Samantha's forms with adult targets at pre- and post-therapy assessments Table 2.6a: Consonants organized by manner: overall accuracy Category Pre-Therapy Assessment Post-Therapy Assessment % Gain Stops & Nasals 53/148 (35.8%) 41/87 (47.1%) 11.3% Stops only 42/91 (46.2%) 35/58 (60.3%) 14.1% Nasals only 11/57 (19.3%) 6/29 (20.7%) 1.4% Stops + Nasals + Stop-clusters + Nasal clusters 56/184 (30.4%) 44/116 (37.9%) 7.5% Fricatives & Affricates 23/93 (24.7%) 20/78 (25.6%) 0.9% Approximants 7/44 (15.9%). 7/26 (26.9%) 11.0% 97 Table 2.6b: Consonants organized by place: overall accuracy Category Pre-Therapy Assessment Post-Therapy Assessment % Gain Labials + Labial clusters 30/80 (37.5%) 13/40 (32.5%) - 5.0%2 Labials only 29/67 (43.3%) 12/30 (40.0%) - 3.3. % Coronals 31/165 (18.9%) 32/106 (30.2%) 11.3% Velars 18/59 (30.5%) 23/49 (46.9%) 16.4% Palatals 0/11 (0.0%) 1/8 (12.5%) 12.5% Table 2.6c: Consonants organized by voicing: overall accuracy Target Pre-Therapy Assessment Post-Therapy Assessment % Gain /g/ 1/9 (11.1%) 3/11 (27.2%) 16.1% /d 3 / 0/9 (0.0%) 1/8 (12.5%) 12.5% Izl 0/14 (0.0%) 2/15 (13.3%) 13.3% 2 Error types included the following: (1) manner errors (i.e., m --> b); mj --> bij, (2) place errors (i.e., pr --> pw; v --> v t stopped; f --> ftt stopped; fr --> fl), (3) voicing errors (i.e., b --> b0). 98 Table 2.6d: Percentage match of Samantha's therapy and generalization targets with adult targets Specific Target Pre-Therapy Assessment Post-Therapy Assessment % Gain /g/ 1/9 (11.1%) 3/11 (27.2%) 16.1% Ikl 17/24 (70.8%) 20/25 (80.0%) 9.2% /d 3 / 0/9 (0.0%) 1/8(12.5%) 12.5% ity 1/8 (12.5%) 4/9 (44.4%) 31.9% iv 1/6 (16.7%) 1/6 (16.7%) 0.0% Is/ 5/17 (29.4%) 7/16 (43.8%) 14.4% Isl + /s/-clusters 8/34 (23.5%) 14/33 (42.4%) 18.9% /s/-clusters 3/17 (17.6%) 7/17 (41.4%) 23.5% Izl 0/14 (0.0%) 0/15 (0.0%) 0.0% Table 2.6e: Place accuracy independent of other features for target consonants Pre- Assessment Palatoalveolar Place Pre- Assessment Velar Place Post- Assessment Palatoalveolar Place Post- Assessment Velar Place wiiy 0/1 (0%) 0/1 (0%) may 1/4 (25%) 1/4 (25%) WI/d3/ 2/4 (50%) 1/3 (33.3%) WI/g/ 3/3 (100%) 3/3 (100%) TOTAL FORWI 3/9 (33.3%) 3/3 (100%) 2/8(25%) 3/3 (100%) 3 WI= Word-Initial, W M = Word-Medial, WF = Word-Final. 99 Pre- Assessment Palatoalveolar Place Pre- Assessment Velar Place Post- Assessment Palatoalveolar Place Post- Assessment Velar Place WM ty 1/2 (50%) 2/2 (100%) WM/tJ/ 1/2 (50%) 2/2 (100%) WM/d3/ 1/2 (50%) 2/2 (100%) WM/g/ 2/4 (50%) 3/3 (100%) TOTAL FOR WM 3/6 (50%) 2/4 (50%) 6/6 (100%) 3/3 (100%) WF l$l 2/3 (66.6%) 0/3 (0%) WF /ty 1/2 (50%) 2/3 (66.6%) WF/d3/ 1/3 (33.3%) 2/3 (66.6%) WF/g/ 1/2 (50%) 5/5 (100%) TOTAL FOR WF 4/8 (50%) 1/2 (50%) 4/9 (44%) 5/5 (100%) 100 Table 2.6f: Place accuracy independent of other features for generalization consonants Word-initial position Pre- Assessment Alveolar Place Post- Assessment Alveolar Place Pre- Assessment Velar Place Post- Assessment Velar Place WI/s/ 8/8 (100%) 6/7 (85.7%) WI s-clusters 17/17 (100%) 17/17 (100%) WI/kh/ 11/12 (91.6%) 12/12 (100%) WI k-clusters 1/3 (33.3%) 1/3 (33.3%) WI/z/ 1/2 (50%) 2/2 (100%) TOTAL FOR WI 26/27 (96.3%) 22/23 (95.8%) 12/15 (80.0%) 13/15 (86.6%) Word-medial position Pre- Assessment Alveolar Place Post- Assessment Alveolar Place Pre- Assessment Velar Place Post- Assessment Velar Place WM/s/ 4/5 (80%) 6/6 (100%) WM s- clusters 0/0 0/0 WM/k/ 5/5 (100%). 5/5 (100%) WMk- clusters 2/2 (100%) 2/2 (100%) WM/z/ 5/5 (100%) 5/5 (100%) TOTAL FOR WM 9/10 (90%) 11/11 (100%) 7/7 (100%) 7/7 (100%) 101 Word-final position Pre- Assessment Alveolar Place Post- Assessment Alveolar Place Pre- Assessment Velar Place Post- Assessment Velar Place WF/s/ 3/4 (75%) 3/3 (100%) WF s-clusters 0/0 0/0 WF/k/ 7/7 (100%) 7/7 (100%) WF k-clusters 0/1 (0%) 1/1 (100%) WF/z/ 7/7 (100%) 6/8 (75%) TOTAL FOR WF 10/11 (90.9%) 9/11 (81.8%) 7/8 (87.5%) 8/8 (100%) Table 2.6g: Manner accuracy independent of other features for target consonants Pre- Assessment Affricates & Fricative Pre- Assessment Stop Post- Assessment Affricates & Fricative Post- Assessment Stop 1/1 (100%) 1/1 (100%) wi / ty 4/4 (100%) 4/4 (100%) WI/d3/ 4/4 (100%) 3/3 (100%) WI/g/ 3/3 (100%) 3/3 (100%) TOTAL FOR WI 8/9 (88.8%) 3/3 (100%) 8/8 (100%) 3/3 (100%) WM iy 2/2 (100%) 2/2 (100%) WM/tJ/ 2/2 (100%) 2/2 (100%) WM/d3/ 1/2 (50%) 2/2 (100%) WM/g/ 4/4 (100%) 3/3 (100%) 102 Pre- Assessment Affricates & Fricative Pre- Assessment Stop Post- Assessment Affricates & Fricative Post- Assessment Stop TOTAL FORWM 5/6 (83.3%) 4/4 (100%) 8/8 (100%) 3/3 (100%) WF iy 3/3 (100%) 3/3 (100%) WF ity 1/2 (50%) 3/3 (100%) WFM3/ 2/3 (66.6%) 2/3 (66.6%) WF/g/ 2/2 (100%) 5/5 (100%) TOTAL FOR WF 6/8 (75%) 2/2 (100%) 8/9 (88.8%) 5/5 (100%) Table 2.6h: Manner accuracy independent of other features for generalization consonants Word-initial position Pre- Assessment Fricatives Post- Assessment Fricatives Pre- Assessment Stop Post- Assessment Stop WI/s/ 8/8 (100%) 7/7 (100%) WI s-clusters 17/17 (100%) 17/17 (100%) WI/kh/ 12/12 (100%) 12/12 (100%) WI k-clusters 2/3 (66.6%) 1/3 (33.3%) WI/z/ 2/2 (100%) 2/2 (100%) TOTAL FOR WI 27/27 (100%) 23/23 (100%) 14/15 (93%) 13/15 (87%) 103 Word-medial position Pre- Assessment Fricatives Post- Assessment Fricatives Pre- Assessment Stop Post- Assessment Stop WM/s/ 4/5 (80%) 6/6 (100%) WM s-clusters 0/0 0/0 WM/k/ 5/5 (100%) 5/5 (100%) WM k-clusters 2/2 (100%) 2/2 (100%) WM/z/ 5/5 (100%) 5/5 (100%) TOTAL FOR WM 9/10 (90%) 11/11 (100%) 7/7 (100%) 7/7 (100%) Word-final position Pre- Assessment Fricatives Post- Assessment Fricatives Pre- Assessment Stop Post- Assessment Stop WF/s/ 3/4 (75%) 3/3 (100%) WF s-clusters 0/0 0/0 WF/k/ 7/7 (100%) 7/7 (100%) WF k-clusters 0/1 (0%) 1/1 (100%) WF Izl 7/7 (100%) 8/8 (100%) TOTAL FOR WF 10/11 (90.9%) 11/11 (100%) 7/8 (87.5%) 8/8 (100%) 104 Table 2.6i: Voicing accuracy independent of other features for target consonants3 [+ voice] [+ voice] [- voice] [- voice] Pre- Post- Pre- Post- Assessment Assessment Assessment Assessment WI/g/ 0/3 (0%) 0/3 (0%) WI /gl/ and /gr/ 0/4 (0%) 0/4 (0%) WI/J/ 1/1 (100%) 1/1 (100%) w i / t y 4/4 (100%) 4/4 (100%) WI/d3/ 0/4 (0%) 1/3 (33.3%) TOTAL FOR WI 0/11 (0%) 1/10 (10%) 5/5 (100%) 5/5 (100%) WM/g/ 1/4 (25%) 1/3 (33.3%) WM /gl/ and /gr/ 0/0 (0%) 0/1 (0%) W M / y 2/2 (100%) 2/2 (100%) WM/ty 2/2 (100%) 2/2 (100%) WM/d3/ 0/2 (0%) 1/2 (50%) TOTAL FOR WM 1/6 (16.6%) 2/6 (33.3%) 4/4 (100%) 4/4 (100%) WF/g/ 0/2 (0%) 4/5 (80%) WF /gl/ and /gr/ 1/4 (/gr/) 0/0 (0%) W F / y 2/3 (66.6%) 2/3 (66.6%) WF/tJ/ 1/2 (50%) 4/4 (100%) WF /d3/ 0/3 (0%) 1/3 (33.3%) TOTAL FOR WF 1/9 (11.1%) 5/8 (62.5%) 3/5 (60%) 5/7 (71.4%) 105 Table 2.6j: Voicing accuracy independent of other features for generalization consonants [+ voice] [+ voice] [- voice] [- voice] Pre- Post- Pre- Post- Assessment Assessment Assessment Assessment WI/s/ 8/8 (100%) 7/7 (100%) WI s-clusters 17/17 (100%) 17/17 (100%) WI/kh/ 0/0 12/12 (100%) WI k-clusters 3/3 (100%) 2/3 (66.6%) WI/z/ 0/2 (0%) 1/2 (50%) TOTAL FOR WI 0/2 (0%) 1/2 (50%) 28/28(100%) 38/39 (97.4%) WM/s/ 5/5 (100%) 6/6 (100%) WM s-clusters 0/0 0/0 WM/k/ 5/5 (100%) 6/6 (100%) WM k-clusters 2/2 (100%) 2/2 (100%) WM/z/ 0/5 (0%) 0/5 (0%) TOTAL FOR 0/5 (0%) 0/5 (0%) 17/17 (100%) 14/14 (100%) WM WF/s/ 4/4 (100%) 3/3 (100%) WF s-clusters 0/0 0/0 WF/k/ 6/7 (85.7%) 7/7 (100%) WF k-clusters 1/1 (100%) 1/1 (100%) WF/z/ 0/7 (0%) 1/8 (12.5%) TOTAL FOR 0/7 (0%) 1/8 (12.5%) 11/12 (91.6%) 11/11 (100%) WF The following summarizes the post-treatment assessment results. 106 Consonants Organized by Manner Categories Stops and nasals Stops and nasals matched the adult targets 47.1% of the time, showing a gain of 11.3% from the pre-therapy assessment. Stops and nasals generally replaced other stops and nasals, as in, for example, music [ (s)bi jut si: ], mask [baetskh ]. There were two tokens in which the nasal [m] was inserted prior to a [p] (e.g.. zipper [sijeamphu] and comb [khoum(p)]). There was one instance in which [pw] replaced [b] and one instance in which [t] replaced [d]. Lastly, there were two tokens in which [kh] replaced [rj]. Stops showed the largest gain in matched proportions with the adult targets, from 46.2% to 60.3%, a 14.1% gain. Nasals increased from 19.3% to 20.7%, a 1.4% gain. There were fewer place and voicing errors post-therapy for the stops. For example, the velar stop Igl was voiced more often post-therapy (27.2% of the time) than pre-therapy (11.1% of the time). Similarly, the voiced labial stop lb/ was substituted with the voiceless labial stop [p] less often post-therapy than pre-therapy (4/13 or 31% pre-therapy versus 1/6 or 17% post-therapy). The velar [k] had no placing errors post-therapy, although there had been placement errors evident pre-therapy (e.g., key [qhi:]). The alveolar stop Itl matched adult targets more often post-therapy than pre- therapy. For example, 5/11 matches (45%) were evident pre-therapy, whereas 4/5 matches (80%) were seen post-therapy. The nasal lx\l also showed improved 107 matches with adult targets post-therapy: 27% matches post-therapy compared with 18% matches pre-therapy. There were fewer placement errors post-therapy. Generally, the target phoneme Igl was accurate 100% of the time in terms of manner in all word-positions. This was also observed pre-therapy. The generalization target fkl and fkhf were also accurate 100% of the time in terms of manner. Improvement was noted however, for /kAclusters (i.e., [ks] and [kr]) post- therapy. Fricatives and affricates Post-therapy assessment results indicated a 25.6% match with adult targets, a 0.9% gain from 24.7% to 25.6%. The fricative fsf and /s/-clusters were the most frequent consonants. Similar to the findings from the pre-therapy assessment, generally fricatives and affricates replaced other fricatives and affricates. Occassionally however, a relatively short [t] was inserted prior to an [s], as in, for example, see [ (t)sssi: ]. The voiced fricative, [z] was always substituted with either an [s(s)(s)] or a [ts(s)(s)] in all word positions. Affricates were often replaced with alveolar and palatoalveolar fricatives (e.g., church [ tjeerts ], watch [ wa(e)(t) $ ], and judge [ SAdt 9 ]). The generalization targets fsf and Izl were accurate for manner 100% of the time in all word positions. This was an improvement from pre-therapy. Similarly, Ity, l&^l and iy were accurate in terms of manner 100%) of the time in all word positions except for word-final /d.3/. There was one instance in which word-final /d3/ was produced as [thi]. 108 Approximants Matched proportions with adult targets showed a gain of 11.0%, from 15.9% to 26.9%. There was one instance in which a short [n] was inserted immediately after an [1], and [1] was omitted in a cluster 4 out of 7 times (57% of the time). The approximant [w] replaced [r] when [r] was in a cluster 2/5 or 40% of the time (e.g., present [pwesssentis]). There was one token in which [1] replaced [r] (e.g.. frog [flakx]). Thus, approximants generally replaced other approximants. The gain of 11.0% was attributable to an increase in more correct [r] matches post-therapy. For example, 0/19 matches (0%) were correct pre- therapy; however, 5/6 matches (83%) were correct post-therapy. Additionally, there were fewer instances in which [w] was inserted for [r] in [spr] and [str] clusters. Consonants Organized by Place Categories Labials Labials plus labial clusters matched the adult target 32.5% of the time. This represented a decrease of 5.0% from the pre-therapy assessment match with the adult target. Singletons matched the adult target 40.0% of the time, a decrease of 3.3% from pre-therapy results. Labials tended to replace other labials, as in, for example, mask above and basket [pwaessskhith ]. Thus, place was not inaccurate; however, voicing and manner were often more 109 inaccurate, showing problems with combination of [Labial] and other features. Labial-/r/ clusters were altered to labial-/w/ clusters as discussed above. There was one instance in which [v] replaced [f] (e.g.. fish [svinj ]). The labial [m] matched the adult target 16.7% of the time post-therapy compared with 37.5% of the time pre-therapy. A decrease in matches was also evident for the labio- dentals, [f] and [v]. For example, [f] showed 7/10 matches (70%) pre-therapy and 1/3 matches (33.3%) post-therapy. For [v], there was 10% match with adult targets pre-therapy but 0% match post-therapy. However, there was a slight increase in [p] matches post-therapy (50% pre-therapy to 60% post-therapy). Coronals Post-therapy assessment results indicated a gain of 11.3% for matched proportions with adult targets. Coronals matched the adult targets 30.2% of the time post-therapy, and 18.9% of the time pre-therapy. Similar to the labial place of articulation, coronals generally replaced coronals, for example, dollhouse [ta(l)ehAUtsss ], and judge [ SAdte ]. However, manner and voicing were less accurate, as was specific place (i.e., alveolar versus palatoalveolar). In word- initial position, the generalization targets Is/ and Izl were 87.5% and 100% accurate in terms of place, respectively. This was a decrease for /s/ and an increase for Izl in terms of accuracy from pre-therapy. The voiceless cognate Is/ was accurate 100% of the time in word-medial and word-final positions. This was an improvement from pre-therapy. Even though place of articulation was 110 often correct the coronal Is/ was produced most of the time with excessive frication (across word positions). This was represented as [sss]. Similarly, [z] was replaced frequently with [sss] or [tsss]. The palatoalveolar fricative III did not improve from the pre-therapy assessment. This target was accurate in terms of place 0% of the time word-initially and finally and 100% of the time word- medially. In word-initial position, III was replaced with [sw] and in word-final position, [nts] replaced III. As in the pre-assessment, either an [n] or a [t] were inserted before the [J] or the [I] was produced with excessive frication ([111]). The affricates /tJ/ and M3/ were accurate for place 100% of the time in word- medial position. This was an improvement from pre-therapy. In word-final position, /tJ/ was accurate 100% of the time and /d3/ was accurate 75% of the time, an improvement from pre-therapy. In word-initial position however, Itl I was accurate for place 25% of the time and M3/ was accurate 33.3% of the time, a decrease in accuracy from pre-therapy. Overall, there were four correct instances of [tJ], an increase from the one correct instance in the pre-therapy assessment. Similarly, there was an increase of one correct instance of word-final [d3] from the pre-therapy assessment (e.g., pages [pe jad?s?sss]). Other coronals showed an increase in match proportion. For example, Itl showed a gain of 35% (45% pre- therapy to 80% post-therapy). Inaccurate productions, however, for both pre- and I l l post-therapy were of the correct place of articulation. Similarly, more post- therapy than pre-therapy productions of [d] were correct for place of articulation. For example, 7/14 or 50% of all pre-therapy productions were produced as [d]. Inaccurate productions were of the correct place of articulation, but were usually produced as the voiceless cognate, [t]. Post-therapy however, there was only one instance in which [t] was substituted for fdl. An increase in correct post-therapy productions was also evident for [1]/. Pre-therapy, 31% of all [1] productions were correct. Inaccuracies included incorrect places of articulation, for example [kae], [(x)6], and jjle]. A gain of 35% (31% to 66%) was noted post-therapy. Inaccurate productions included [l(n)] and [le]. Alveolar fnl also showed improvement from pre- to post-therapy. Pre-therapy 3/17 or 18% of all [n] productions matched adult targets. Inaccurate productions were of the correct place of articulation, for example [nt] and [nd]. Post-therapy 3/11 or 27% of all productions matched the adult target, however inaccurate productions were of the correct place of articulation. Velars Matched proportions with adult targets increased from 30.5% to 46.9%, a 16.4% gain from the pre-therapy assessment. For Igl, a gain of 16.1% was evident and for fkl, 9.2%. Generally, there were fewer tokens in which [k] replaced [g]. However, there were four instances in which [g] was replaced with either [kx] or [kx ]. This affrication was not found during the pre-therapy assessment transcription. Improvement was noted for both word-medial and word-final Igl productions. A gain of 50% was noted from pre- to post-therapy for place of articulation. Specifically, 100% of all word-medial and word-final [g] productions were accurate in terms of place of articulation. No significant changes were noted for [k] in all word-positions. However, 100% of all [kh] productions were accurate for place of articulation, an 8.4% gain from pre-therapy. No improvement was seen with /g/-clusters (0% matches pre- and post-therapy); however, [ks] and [kr] clusters both improved post-therapy (100% match with adult targets for both [ks] and [kr] post-therapy compared with 0% match pre-therapy). The [rj] showed 8/12 matches (67%) post-therapy compared with 18/24 (75%) pre-treatment. When inaccurate, [rj] was replaced by [kh], [2], [k h i ], or [ ta2 ], i.e., sometimes retaining place and sometimes not. Voicing Voicing of stops, fricatives, and affricates improved from pre-therapy. Significant improvement was noted for both word-medial and word-final Igl. For the former position, post-therapy productions were accurate 33.3% of the time, and for the latter position, 80% of the time. In comparison, pre-therapy word- medial Igl productions were accurate 25% of the time and word-final Igl productions were accurate 0% of the time. Thus, a gain of 8.3% and 80% was 113 evident for word-medial and word-final Igl productions, respectively in terms of voicing. Improvement was also noted for word-final fkl. Pre-therapy 85.7% of [k] productions were accurate for voicing; however, 100% of post-therapy [k] productions had accurate voicing. A gain of 100% was also noted for word-initial [kh]. The voiced affricate /dz>l showed improvement in all word-positions. A gain of 33.3% was noted for word-initial position, 50% for word-medial position, and 33.3% for word-final position. Generally, more /(I3/ targets were voiced post- therapy than pre-therapy. The voiceless affricate Itl I was 100% accurate in terms of voicing in all word-positions. This was an improvement because only 50% of word-final [tj] productions were accurate in terms of voice pre-therapy. The palatoalveolar fricative III did not show change in terms of voicing from pre- therapy. Similar to pre-therapy, inaccurate productions included [n$] and [2J]. The generalization target /s/ was 100% accurate for voicing in all word-positions. This was evident pre-therapy as well. Improvement was noted for Izl. In word- initial position, a gain of 50% was noted and in word-final position a gain of 12.5%. Inaccuracies included Izl being produced as [s], [ts], or [sss]. Improvement was also evident for non-target phonemes such as Idl, Ivl, and Ibl. More post-therapy than pre-therapy productions of Idl and Ivl were accurate for voicing. Specifically, 20% of pre-therapy [v] productions were voiced; however, 33.3% were voiced post-therapy. For Idl, 7/14 or 50% of pre-therapy productions 114 were produced as [t]. Post-therapy there was only one token in which [t] was substituted for /d/. Pre-therapy, 30% of Ibl productions were produced as [p]. Post-therapy, 50% of all productions were voiced incorrectly. Thus, improvement was noted for other non-target stops and fricatives. Research Question #2 2. Do narrow transcriptions of the goal target phonemes in palatometry therapy show improvement from goal target phonemes in non-palatometry therapy? Null Hypothesis 2: There will be no difference between the number of correctly produced goal target phonemes in palatometry therapy compared to the number produced in non-palatometry therapy. This question was addressed by transcribing the goal therapy phonemes. Narrow transcription was used for the goal phonemes and broad transcription was used for all vowels or non-target phonemes. Tables 1.7-2.0 below display the results for each target phoneme. Each therapy target phoneme was addressed twice over the eight treatment blocks. Therefore, results from the first block will be compared with results from the second block. Post-palatometry therapy recordings are marked with an asterisk; post-non-palatometry therapy recordings are not marked. See Appendix B for therapy target words. 115 Table 2.7: Pre- and post-therapy target Igl transcriptions for blocks one and two Note: tt=voiced interdental Block 1: Pre-Therapy Block 1: Post-Therapy Block 2: Pre-Therapy Block 2: Post-Therapy 1 [kho] [khou] [kho hawth] * [kho howth] 2 [kheim] [keim] [?ai k h aijy] * [?sei khaijy] 3 [khatsss] [keips] [tta kheim] * [tts kheim] 4 [khetsss] [khAumsss] [ttae kusss] * [ttaet kuss] 5 [khAm] [khAm] [hul futh] * [khud fud] 6 [khai] * [khei] [tta kul] [ttae khejt] 7 [khAnk] * [khem] [?ai kath] [?ai khasth] 8 [khunt] * [kheik] [k h sis l a i t h ] [khaitss laith] 9 [k h i j l ] * [khejw] [khan laith] [khan laith] 10 [khus] * [khausss] [k hAm mAim] [kham mAm] % Correct for Non- Palatometry Therapy (lb) 0/5 (0%) 0/5 (0%) % Correct for Palatometry Therapy (15) 0/5 (0%) 0/5 (0%) Total % Correct (110) 0/10 (0%) 0/10 (0%) 0/10 (0%) No improvement was noted between pre- and post-recordings, between palatometry or non-palatometry therapy, or between blocks one and two. 116 Table 2.8: Pre- and post-therapy target /(I3/ transcriptions for blocks one and two Note: C> = backed, C< = fronted, i = barred [i], tt = voiced interdental Block 1: Pre-Therapy Block 1: Post-Therapy Block 2: Pre-Therapy Block 2: Post-Therapy 1 [s>9m] [mjaeam] [2a t S e d h ] * [?ae jeith] 2 [s>im ] [mjim] [tjAT l A l t h ] * [fil l A i t h ] 3 [s>im ] bin] [ t f i p h p h 3 i t h ] * [ f i p h peith] 4 [s>u?u] [zu] [ t t a Jeim] * [tta j" asm] 5 [sow] [njow] [2a? f A t h ] * [?ae f e i t h ] 6 [S<eiw] * [ziej] [ t t a t h t S i m ] [tta fim] 7 Lf<it] * [zikh] [S Sow 2oth] [Jowi ?owth] 8 [s>ew] * [d3oa] [taph b u t h ] [tf tjamph buth] 9 [ t S < i m h ] * [z:ipkha] [tfin khinth] [tjin kinth] 10 [ t j e m ] * [z: aph] [tta t j e r ] [tta d3aj] % Correct for Non- Palatometry Therapy (lb) 0/5 (0%) 1/5 (20%) % Correct for Palatometry Therapy (lb) 1/5 (20%) 0/5 (0%) Total % Correct (/10) 0/10 (0%) 1/10 (10%) 0/10 (0%) 1/10 (10%) Improvement was noted between the first block pre- and post-therapy recordings. For the first block, 20% of the goal targets recorded immediately after palatometry therapy were perceived to be correct; however 0% of [d3] targets were perceived to be correct after non-palatometer therapy. For the second block, 0% 117 of the targets were correct during the pre-therapy recording, however 10% were correct after therapy. An improvement of 20% (or 1/5) was noted for the goal targets recorded immediately after non-palatometry therapy. None of the goal targets were perceived to be correct after palatometry therapy. Table 2.9: Pre- and post-therapy target /tj/ transcriptions for blocks one and two Note: C> = backed, C< = fronted, i = barred [i], t = voiced interdental, c> = ungrooved alveopalatal (between alveolar and palatoalveolar), ,= ejective. Block 1: Pre-Therapy Block 1: Post-Therapy Block 2: Pre-Therapy Block 2: Post-Therapy 1 [sjow] [s:>ow] [ 2se tf ,alvav9] * [?ae stS<aiv:] 2 [Jem] [s:>ein] [taph man] * [tfaeiph mein] 3 [sjow] [SOJ] [d 3 , i l l a i t h ] * [tjii laith] 4 [sjAmp] [sAmph ] [tta tjajph] * [tta tfahaejph] 5 [tceim] [s h£9m] [t ha9n net] * [ (t)Jayn neth] 6 [s>ein] * [tssn] [ttae tja+] [tt:aeth tj* tfaw] 7 [s>eu ] * [s:>il] [ t h i p h p hin] [tfimph pun] 8 [s>9S ] * [tSElts] [ha thamph] [?8B ju:mph] 9 [S>EJ ] * [tjeu] [theimph phinth] [tf<iph pint] 10 [c>iph] * [tr<imph] [ t t 9 thAmm] [tta fimp] % Correct for Non- Palatometry Therapy (lb) 0/5 (0%) 2/5 (40%) 118 % Correct for Palatometry Therapy (15) 1/5 (20%) 3/5 (60%) Total % Correct (/10) 0/10 (0%) 1/10 (10%) 2/10 (20%) 5/10 (50%) Improvement was noted after therapy during the first block. An increase of 10% was perceived. For session one, 20% or 1/5 [tj] targets were perceived to be correct immediately after palatometry therapy, however 0/5 or 0% were perceived to be correct immediately after non-palatometry therapy. Improvement was also noted for the second treatment block. Goal target productions during pre- and post-therapy recordings were 20% and 50% correct, respectively. Forty-percent or 2/5 of the [tj] productions were perceived to be correct immediately after non- palatometry and 60% or 3/5 were perceived to be correct immediately after palatometry therapy. 119 Table 2.10: Pre- and post-therapy target /J/ transcriptions for blocks one and two Note: O = backed, C< = fronted, i = barred [i], tt = voiced interdental, c> = ungrooved alveopalatal (between alveolar and palatoalveolar), , = ejective, Jj=extra frication. Block 1: Pre-Therapy Block 1: Post-Therapy Block 2: Pre-Therapy Block 2: Post-Therapy 1 [Saph] [J< amphs] cmiph] * [tta 9(t)Haph] 2 Lf< uth ] LTuth3] [fJ:Ath taith] * [fit taith] 3 Lf<ith] U< iphs] [?aeSUth] * [?a (t)Ut h] 4 Lf<awl] [S<owl] [|e+ H] * Lffe* H] 5 [Jeivsv] [$eivah] [tta (t)Juth] * [ttae (t)Suth>] 6 [f< ow] * [S<ow] [ttath Sow] [ttae i t S<ow] 7 LfAnth] * LfAth] [$<aiv bim] [jeiv vim] 8 [S<ain] * LT< am] [Jaeja*] [?« Jar] 9 [J< awph ] * [J< amph] [fain naith] [fain laith] 10 Kt)Xel] * [f< e+ha] [J<awph p h in] [f<2awph phin>] % Correct for Non- Palatometry Therapy (lb) 2/5 (40%) 3/5 (60%) % Correct for Palatometry Therapy (lb) 1/5 (20%) 1/5 (20%) Total % Correct (/10) 3/10 (30%) 3/10 (30%) 4/10 (40%) 4/10 (40%) 120 For the first block, two of five of the goal target productions were perceived to be correct immediately after non-palatometry therapy and one was perceived to be correct after palatometry therapy. Pre- and post-therapy recordings during the second block both had 40% correct perceived productions. Specifically, 60% or 3/5 of the therapy target productions were perceived to be correct immediately after non-palatometry therapy and 20% or one was perceived to be correct after palatometry therapy. Figures 1.1 and 1.2 summarize the transcription results for blocks one and two, respectively. Figure 1.1 displays the percentage of therapy phonemes that were correct after palatometry and non-palatometry therapy for block one. For this block, non-palatometry therapy was conducted first and palatometry therapy second. Figure 1.2 displays the percentage of therapy phonemes that were correct after palatometry and non-palatometry therapy for block two. During the latter block, palatometry was conducted prior to non-palatometry therapy. 121 Figure 1.1: Percent Correct After Palatometry and Non- Palatometry Therapy, Block One. 60% 50% 40% 30% 20% 10% 0% X t j d 3 Therapy Target g B Palatometry| Therapy • Non- Palatometryl Therapy 122 Figure 1.2: Percent Correct After Palatometry and Non-Palatometry Therapy, Block Two. 60% 50% 40% o o 30% 20% 10% 0% o Palatometry Therapy • Non- Palatcmetry Therapy S tS d 3 Therapy Target 9 123 Research Questions #3 3. Do post-therapy palatogram sound file recordings show a reduction in impairment compared to pre-therapy palatogram sound file recordings? Null Hypothesis 3: There will be no difference between post-therapy and pre- therapy palatogram sound recordings in terms of narrow phonetic transcriptions. A sample of 20% of the palatogram audio-recordings showed over 90% agreement between transcribers. Transcriptions of each of the therapy and generalization target tokens are displayed below in Tables 3a-k. Details on each transcription are provided. Table 3a: Pre- and post-assessment palatogram sound file recording transcriptions for word-initial Itl I tokens Palatometry Assessment Words Pre-Assessment Palatogram Sound File Transcription Post-Assessment Palatogram Sound File Transcription Degree of Improvement cheap [Jiph] [tS<ip] + + chop [Jamph] [t$amph] + + chew [lu] [Ju] Improvement was noted with both cheap and chop. The post-assessment recordings contained target phonemes that more closely resembled I til. For 124 cheap post-therapy, the target phoneme was produced slightly forward on the palate, somewhere between palatoalveolar and alveolar places of articulation. This was an improvement from pre-therapy since no stop closure was evident before therapy. No change was noted for chew. Table 3b: Pre- and post-assessment palatogram sound file recording transcriptions for word-final It 11 tokens Palatometry Assessment Words Pre-Assessment Palatogram Sound File Transcription Post-Assessment Palatogram Sound File Transcription Degree of Improvement each [?i(t)f<] [?i(t)J<] botch [batas] [batj<] + + pooch [phots>] [PAS] + + Improvement was noted for both botch and pooch. No change was noted from pre- to post-therapy for each. For botch, the target phoneme was perceived to have improved from an alveolar [s] to a fronted palatoalveolar [t$]. For pooch. the pre-therapy recording was perceived as a backed [s] and the post-therapy recording as a palatoalveolar fricative [$]. This was judged to be an improvement since the post-therapy production was produced with affrication and palatoalveolar place of articulation. 125 Table 3c: Pre- and post-assessment palatogram sound file recording transcriptions for word-initial /$/ tokens Palatometry Assessment Words Pre-Assessment Palatogram Sound File Transcription Post-Assessment Palatogram Sound File Transcription Degree of Improvement she [J<i] [Si:] + + shah [sah] [s>ah] + + shoe [Ju:] [s>ju] + - Both she and shah were perceived to have improved from pre- to post- therapy. For she, the pre-therapy production was judged to have been fronted (i.e, somewhere between palatoalvoelar and alveolar place of articulation). However, the post-therapy production was perceived to be correct. For shah, the pre- therapy production was judged to be an [s] and the post-therapy production a backed [s]. This change in place of articulation was noted to be an improvement since the latter production more closely matched the palatoalveolar place of articulation. Token shoe was judged to be worse post-therapy; however, the pre- therapy recording was judged to be correct. 126 Table 3d: Pre- and post-assessment palatogram sound file recording transcriptions for word-final /$/ tokens Palatometry Assessment Words Pre-Assessment Palatogram Sound File Transcription Post-Assessment Palatogram Sound File Transcription Degree of Improvement posh [phuj] [ba(t)X:] + - push [PM] [DAtS>] + - No improvement was noted for either posh or push. The pre-therapy recordings were perceived to be correct. However the post-therapy recordings were judged be less correct. For example, the post-therapy recording of posh contained a stop prior to [J], and the post-recording of push contained less affrication than the pre-therapy recording. Table 3e: Pre- and post-assessment palatogram sound file recording transcriptions for word-initial /d3/ tokens Palatometry Assessment Words Pre-Assessment Palatogram Sound File Transcription Post-Assessment Palatogram Sound File Transcription Degree of Improvement jaw [s>aw] + - jujube [ s>ef<emb'] [z>uz>umh] + + 127 Improvement was noted for jujube because the post-therapy recording contained voicing whereas the pre-therapy recording did not. The post-recording of jaw was worse because less affrication was noted and place of articulation was further forward than that perceived pre-therapy. Table 3f: Pre- and post-assessment palatogram sound file recording transcriptions for word-final /d3/ token Palatometry Assessment Words Pre-Assessment Palatogram Sound File Transcription Post-Assessment Palatogram Sound File Transcription Degree of Improvement hodge [hant9h] [ha(d)S<] + + The post-assessment recording was perceived to have been better than the pre-assessment recording. Pre-therapy the target phoneme was produced as a stop. Post therapy however a stop plus a fronted palatoalveolar fricative were perceived. The addition of the fricative and the voiced stop was an improvement from pre-therapy. 128 Table 3g: Pre- and post-assessment palatogram sound file recording transcriptions of the word-initial Igl tokens Palatometry Assessment Words Pre-Assessment Palatogram Sound File Transcription Post-Assessment Palatogram Sound File Transcription Degree of Improvement geese [k h is] [k hiss] gob [khob] [khabhe] goo [khu] [khu] All recordings were perceived to contain [kh] in word-initial position. Thus, no improvement was noted from pre- to post-therapy. Table 3h: Pre- and post-assessment palatogram sound file recording transcriptions of the word-final Igl tokens Palatometry Assessment Words Pre-Assessment Palatogram Sound File Transcription Post-Assessment Palatogram Sound File Transcription Degree of Improvement egg [?ik hs] [?ekh] rog [0kh9] [jakh9] POg [pakh] [bakh9] No improvement was noted from pre- to post-therapy. All recording were judged to contain [kh] for Igl. 129 Table 3i: Pre- and post-assessment palatogram sound file recording transcriptions for word-initial and word-final fkl tokens Palatometry Assessment Words Pre-Therapy Palatogram Sound File Transcription Post-Therapy Palatogram Sound File Transcription Degree of Improvement cop [khob] [khamph] - + eek [?ik ha] [2 i :k h 8] - + No change was noted from pre- to post-therapy and all recordings were perceived to be correct both pre- and post-therapy, except for the addition of [a] in eek. Table 3j: Pre- and post-assessment palatogram sound file recording transcriptions for word-initial and word-final /z/ tokens Palatometry Assessment Words Pre-Assessment Palatogram Sound File Transcription Post-Assessment Palatogram Sound File Transcription Degree of Improvement zoo r(t)su:] [z zu] + + bees [phis:] [mbi(t)z] + + Improvement was noted for both zoo and bees. For zoo, the pre-therapy generaliztion target phoneme was produced with an initial [t] followed by [s]. 130 However, the post-therapy recording was perceived to be correct for both place of articulation and voicing. For bees, improvement was noted for voicing although a stop was inserted just prior to [z]. Table 3k: Pre- and post-assessment palatogram sound file recording transcriptions for word-initial and word-final /s/ tokens. Palatometry Assessment Words Pre-Therapy Palatogram Sound File Transcription Post-Therapy Palatogram Sound File Transcription Degree of Improvement saw [saw] [saw] - + boss [ ba(t)s>] [pha(t)s] + + No change was noted for saw from pre- to post-therapy; however, change was not needed since both recordings were perceived to contain [s]. Improvement was noted for boss because the post-recording did not contain a backed [s] as did the pre-assessment recording. 131 Research Question #4 4. Do post-therapy palatograms show a reduction in impairment compared to pre-therapy palatograms? Null Hypothesis 4: There will be no difference between pre- and post-therapy palatograms in terms of contact patterns. Pre- and post-therapy assessment palatograms were collected for the four target phonemes, Ig, t J, J, d 3 / and for the three generalization targets, Is, z, kl. In terms of reliability, the experimenter and her supervisor were in agreement 96.29%. Disagreements included factors such as how to appropriately evaluate contact points that were clearly not contacted by Samantha but were contacted because of excess saliva, and whether the groove width for the release phase of some palatograms was too wide or appropriate in width. Palatometry assessment stimuli are listed in Appendix A and examples of palatograms for tokens chop, jujube, she, geese, cop, bees, and boss are displayed in Appendix D. Results for each treatment and generalization target are discussed, followed by a summary of the overall results. Therapy Target: lt$l 132 There were a total of 12 It ll palatograms of which six showed the stop closure phase and six displayed the fricative phase. Six of the 12 palatograms had It 11 in word-initial position (three showing the stop closure phase and the remaining three showing the fricative phase) and the remaining six palatograms displayed the affricate in word-final position (again, three showing stop closure phase and three showing the fricative phase). Word-initial & word-final Ity token palatogram results Tables 4a and 4b below show the three word-initial and three word-final token palatogram results for the stop closure and fricative portion phases respectively. A percentage is provided for each criterion based on the results from all six palatograms. 133 Table 4a: Word-initial and word-final stop closure phase results for It 11 Criteria + + + N + - - + 1. Stop closure 3/6 (50%) 2/6 (33%) 1/6 (17%) 2. Alveolar tip region contact 4/6 (67%) 2/6 (33%) Alveolar & palatal medial contact: 3. R-side 4. L-side 4/6 (67%) 3/6 (50%) 2/6 (33%) 3/6 (50%) Alveolar & palatal lateral contact 5. R-side 6. L-side 3/6 (50%) 1/6 (17%) 3/6 (50%) 2/6 (33%) 2/6 (33%) 1/6 (17%) Dorsal medial contact: 7. R-side 8. L-side 2/6 (33.3%) 1/6 (16.6%) 3/6 (50%) 1/6 (17%) 5/6 (83%) Dorsal lateral contact: 9. R-side 10. L-side 1/6 (16.6%) 2/6 (33.3%) 3/6 (50%) 6/6 (100%) 134 Table 4b: Word-initial and word-final fricative phase results for It 11 Criteria + + + N + - - + 1. Groove width 2/6 (33%) 0/6 (0%) 2.6 (33%) 0/6 (0%) 2/6 (33%) 2. Alveolar tip region contact 1/6 (17%) 0/6 (0%) 2/6 (33%) 2/6 (33%) 1/6 (1%) Alveolar & palatal medial contact: 3. R-side 4. L-side 4/6 (67%) 0/6 (0%) 0/6 (0%) 0/6 (0%) 2/6 (33%) 1/6 (0%) 0/6 (0%) 4/6 (67%) 0/6 (0%) 1/6 (17%) Alveolar & palatal lateral contact 5. R-side 6. L-side 2/6 (33%) 1/6 (17%) 0/6 (0%) 0/6 (0%) 0/6 (0%) 2/6 (33%) 2/6 (33%) 3/6 (50%) 2/6 (33%) 0/6 (0%) 7. Dorsal medial contact overall 6/6 (100%) Dorsal medial contact: 8. R-side 9. L-side 1/6 (17%) 1/6 (17%) 5/6 (83%) 5.6 (83%) Dorsal lateral contact: 10. R-side 11. L-side 1/6 (17%) 1/6 (17%) 3/6 (50%) 2/6 (33%) 5/6 (83%) Therapy Target /$/ There were a total of 10 /$/ palatograms of which five showed the onset phase and five displayed the release phase. Six of the ten palatograms had / J/ in word-initial position (three of the six showing the maximum contact phase and 135 the remaining three showing the release phase) and the remaining four palatograms displayed the fricative in word-final position (two palatograms showing the maximum contact phase and two showing the release phase). Word-initial & word-final l\l token palatogram results Tables 5a and 5b show the word-initial and word-final palatogram results for the maximum contact and release portion phases respectively. A percentage is provided for each criterion based on the results from all five tokens. Table 5a: Word-initial and word-final maximum closure phase results for /$/ Criteria + + + N + - - + 1. Groove width 3/5 (60%) 2/5 (40%) 2. Alveolar tip region contact .1/5 (20%) 3/5 (60%) 1/5 (20%) Alveolar & palatal medial contact: 3. R-side 4. L-side 3/5 (60%) 1/5 (20%) 1/5 (20%) 2/5 (40%) 3/5 (60%) Alveolar & palatal lateral contact 5. R-side 6. L-side 1/5 (20%) 2/5 (40%) 2/5 (40%) 2/5 (40%) 3/5 (60%) Dorsal medial contact: 7. R-side 8. L-side 4/5 (80%) 3/5 (60%) 1/5 (20%) 2/5 (40%) Dorsal lateral contact: 9. R-side 10. L-side 5/5 (100%) 5/5 (100%) 136 Table 5b: Word-initial and word-final release phase results for III Criteria + + + N + - - + 1 Groove width 2/5 (40%) 1/5 (20%) 2/5 (40%) 2. Alveolar tip 5/5 1/5 (20%) region contact (100%) Alveolar & palatal medial contact: 3. R-side 1/5 (20%) 1/5 (20%) 1/5 (20%) 1/5 (20%) 1/5 (20%) 4. L-side 1/5 (20%) 4/5 (80%) 1/5 (20%) Alveolar & palatal lateral contact 5. R-side 3/5 (60%) 1/5 (20%) 1/5 (20%) 6. L-side 1/5 (20%) 3/5 (60%) 1/5 (20%) 7. Dorsal 2/5 (40%) 2/5 (40%) 1/5 (20%) medial contact overall Dorsal medial contact: 8. R-side 2/5 (40%) 3/5 (60%) 9. L-side 5/5 (100%) Dorsal lateral contact: 10. R-side 5/5 (100%) 11. L-side 5/5 (100%) 137 Therapy Target: / d3 / There were a total of six /CI3/ palatograms of which three showed the stop closure phase and three displayed the fricative portion phase. Four of the six palatograms had /d 3 / in word-initial position (of which two showed the stop closure phase and two the fricative portion phase) and the remaining two palatograms displayed the affricate in word-final position (of which one showed the stop closure phase and one showed the fricative portion phase). Word-initial & word-final /d?/ token palatogram results Tables 6a and 6b display the two word-initial and one word-final palatogram token results for the stop closure and fricative portion phases respectively. A percentage is provided for each criterion based on the results from all three palatograms. 138 Table 6a: Word-initial and word-final stop closure phase results for Id^l Criteria + + + N + - - + 1. Stop closure 2/3 (67%) 1/3 (33%) 2. Alveolar tip region contact 1/3 (33%) 2/3 (67%) Alveolar & palatal medial contact: 3. R-side 4. L-side 2/3 (67%) 2/3 (67%) 1/3 (33%) 1/3 (33%) Alveolar & palatal lateral contact 5. R-side 6. L-side 2/3 (67%) 1/3 (33%) 1/3 (33%) 1/3 (33%) 1/3 (33%) Dorsal medial contact: 7. R-side 8. L-side 1/3 (33%) 2/3 (67%) 3/3 (100%) Dorsal lateral contact: 9. R-side 10. L-side 3/3 (100%) 3/3 (100%) 139 Table 6b: Word-initial and word-final fricative phase results for /d 3 / Criteria + + + N + - - + 1. Groove width 3/3 (100%) 2. Alveolar tip region contact 3/3 (100%) Alveolar & palatal medial contact: 3. R-side 4. L-side 3/3 (100%) 1/3 (33%) 2/3 (67%) Alveolar & palatal lateral contact 5. R-side 6. L-side 3/3 (100%) 2/3 (67%) 1/3 (33%) 7. Dorsal medial contact overall 3/3 (100%) Dorsal medial contact: 8. R-side 9. L-side 2/3 (67%) 1/3 (33%) 3/3 (100%) Dorsal lateral contact: 10. R-side 11. L-side 1/3 (33%) 2/3 (66%) 1/3 (33%) 1/3 (33%) 1/3 (33%) Therapy Target: /g/ There were a total of 12 /g/ palatograms of which six showed the maximum contact point and six displayed the release phase. Six of the 12 palatograms had 140 Igl in word-initial position (three of the six showing the maximum contact point and the remaining three showing the release phase) and the remaining six palatograms displayed the voiced velar stop in word-final position (again, three showing maximum contact point and three showing the release phase). Word-initial and word-final Igl token palatogram results Tables 7a and 7b below show the word-initial and word-final token palatogram results for the maximum contact and release phases respectively. A percentage is provided for each criterion based on the results from all six palatograms. Table 7a: Word-initial and word-final maximum contact point results for Igl Criteria + + + N + - - + 1. Symmetry 4/6 (67%) 2/6 (33%) Dorsal region medial contact: 2. R-side 3. L-side 2/6 (33%) 3/6 (50%) 3/6 (50%) 3/6 (50%) 1/6 (17%) Dorsal region lateral contact: 4. R-side 5. L-side 3/6 (50%) 3/6 (50%) 2/6 (33%) 2/6 (33%) 1/6 (17%) 1/6 (17%) 6. Back Row Contact 3/6 (50%) 3/6 (50%) 141 Table 7b: Word-initial and word-final release phase results for Igl Criteria + + + N + - - + 1. Symmetry 3/6 (50%) 3/6 (50%) Dorsal region medial contact: 2. R-side 3. L-side 2/6 (33%) 2/6 (33%) 2/6 (33%) 1/6 (17%) 1/6 (17%) 3/6 (50%) 1/6 (17%) Dorsal region lateral contact: 4. R-side 5. L-side 3/6 (50%) 3/6 (50%) 1/6 (17%) 2/6 (33%) 2/6 (33%) 1/6 (17%) 6. Back Row Contact 3/6 (50%) 3/6 (50%) Generalization Target: fkl There were a total of four fkl palatograms of which two showed the maximum contact point and two displayed the release phase. Two of the four palatograms had fkl in word-initial position (one of the two showing the maximum contact point and the remaining palatogram showing the release phase) and the remaining two palatograms displayed the voiceless velar stop in word-final position (again, one showing maximum contact point and one palatogram showing the release phase). 142 Word-initial and word-final palatogram results Tables 8a and 8b show the word-initial and word-final palatogram results for the maximum contact and release phases respectively. A percentage is provided for each criterion based on the results from the two palatograms. Table 8a: Word-initial and word-final fkl token maximum contact point results Criteria + + + N + - - + 1. Symmetry 1/2 (50%) 1/2 (50%) Dorsal region medial contact: 2. R-side 3. L-side 1/2 (50%) 1/2 (50%) 2/2 (100%) Dorsal region lateral contact: 4. R-side 5. L-side 2/2 (100%) 1/2 (50%) 1/2 (5%) 6. Back Row Contact 1/2 (50%) 1/2 (50%) 143 Table 8b: Word-initial and word-final fkl token release phase results Criteria + + + N + - - + 1. Symmetry 2/2 (100%) Dorsal region medial contact: 2. R-side 3. L-side 2/2 (100%) 2/2 (100%) Dorsal region lateral contact: 4. R-side 5. L-side 1/2 (50%) 2/2 (100%) 1/2 (50%) 6. Back Row Contact 1/2 (50%) 1/2.(50%) Generalization Target: Izl There were a total of four Izl palatograms of which two showed the maximum contact point and two displayed the release phase. Two of the four palatograms had Izl in word-initial position (one of the two showing the maximum contact point and the remaining palatogram showing the release phase) and the remaining two palatograms displayed the voiced alveolar fricative in word-final position (again, one showing maximum contact point and one palatogram showing the release phase). 144 Word-initial and word-final Izl token palatogram results Tables 9a and 9b below show the word-initial and word-final palatogram results for the maximum contact point and release phases respectively. Percentages are provided for each criterion based on the results from the two palatograms. Table 9a: Word-initial and word-final Izl token maximum contact point results Criteria + + + N + - - + 1. Groove width 2/2 (100%) 2. Alveolar region contact 1/2 (50%) 1/2 (50%) Alveolar & palatal medial contact: 3. R-side 4. L-side 1/2 (50%) 1/2 (50%) 1/2 (50%) 1/2 (50%) Alveolar & palatal lateral contact 5. R-side 6. L-side 1/2 (50%) 1/2 (50%) 1/2 (50%) 1/2 (50%) Dorsal medial contact: 7. R-side 8. L-side 2/2 (100%) 2/2 (100%) Dorsal lateral contact: 9. R-side 10. L-side 2/2 (100%) 2/2 (100%) 145 Table 9b: Word-initial and word-final Izl token release phase results Criteria + + + N + - - + 1. Groove width 2/2 (100%) 2. Alveolar region contact 1/2 (50%) 1/2 (50%) Alveolar & palatal medial contact: 3. R-side 4. L-side 2/2 (100%) 1/2 (50%) 1/2 (50%) Alveolar & palatal lateral contact 5. R-side 6. L-side 1/2 (50%) 1/2 (50%) 2/2 (100%) 7. Dorsal medial contact overall 2/2 (100%) Dorsal medial contact: 8. R-side 9. L-side 2/2 (100%) 1/2 (50%) 1/2 (50%) Dorsal lateral contact: 10. R-side 11. L-side 1/2 (50%) 2/2 (100%) 1/2 (50%) Generalization Target: Isl There were a total of four Isl palatograms of which two showed the maximum contact point and two displayed the release phase. Two of the four palatograms had Is/ in word-initial position (one of the two showing the maximum contact point and the remaining palatogram showing the release phase) and the remaining two palatograms displayed the voiceless alveolar fricative in word-final position (again, one showing maximum contact point and one palatogram showing the release phase). Word-initial and word-final Is/ token palatogram results Tables 10a and 10b display the word-initial and word-final palatogram results for the maximum contact point and release phases respectively. 147 Table 10a: Word-initial and word-final Is/ token maximum contact point results Criteria + + + N + - - + 1. Groove width 2/2 (100%) 2. Alveolar region contact 1/2 (50%) 1/2 (50%) Alveolar & palatal medial contact: 3. R-side 4. L-side 1/2 (50%) 1/2 (50%) 1/2 (50%) 1/2 (50%) Alveolar & palatal lateral contact 5. R-side 6. L-side 1/2 (50%) 1/2 (50%) 1/2 (50%) 1/2 (50%) Dorsal medial contact: 7. R-side 8. L-side 1/2 (50%) 1/2 (50%) 2/2 (100%) Dorsal lateral contact: 9. R-side 10. L-side 2/2 (100%) 2/2 (100%) 148 Table 10b: Word-initial and word-final Is/ token release phase results Criteria + + + N + - - + 1. Groove width 2/2 (100%) 2. Alveolar region contact 2/2 (100%) Alveolar & palatal medial contact: 3. R-side 4. L-side 2/2 (100%) 1/2 (50%) 1/2 (50%) Alveolar & palatal lateral contact 5. R-side 6. L-side 2/2 (100%) 1/2 (50%) 1/2 (50%) 7. Dorsal medial contact overall 1/2 (50%) 1/2 (50%) Dorsal medial contact: 8. R-side 9. L-side 1/2 (50%) 2/2 (100%) 1/2 (50%) Dorsal lateral contact: 10. R-side 11. L-side 2/2 (100%) 2/2 (100%) Tables lla-f show the percentage of criteria noted to have changed for each degree of change (e.g., + +, + N, + -, - +, and - -). The criteria are described in the Methods Chapter. 149 Table 11a: Percentage of criteria noted to have changed for the stop phase of the affricate treatment targets, /d3/ and Ity WORD + + + N + - - + chew 4/10 (40%) 1/10 (10%) 1/10 (10%) 4/10 (40%) 0/10 (0%) chop 5/10 (50%) 1/10 (10%) 0/10 (0%) 4/10 (40%) 0/10 (0%) cheap 4/10 (40%) 2/10 (20%) 1/10 (10%) 3/10 (30%) 0/10 (0%) each 4/10 (40%) 0/10 (0%) 1/10 (10%) 5/10 (50%) 0/10 (0%) botch 3/10 (30%) 0/10 (0%) 3/10 (30%) 3/10 (30%) 1/10 (10%) pooch 1/10 (10%) 1/10 (10%) 3/10 (30%) 5/10 (50%) 0/10 (0%) jaw 5/10 (50%) 0/10 (0%) 0/10 (0%) 5/10 (50%) 0/10 (0%) jujube 1/10 (10%) 1/10 (10%) 2/10 (20%) 3/10 (30%) 2/10 (20%) hodge 1/10 (10%) 1/10 (10%) 5/10 (50%) 3/10 (30%) 0/10 (0%) Total 28/90 (31.1%) 7/90 (7.7%) 16/90 (17.7%) 35/90 (38.8%) 3/90 (3.3%) Average % 31% 7.7% 17.7% 38.8% 3.3% Table lib: Percentage of criteria noted to have changed for each degree of change for the fricative phase of the affricates WORD + + + N + - - + each 4/11 0/11 (0%) 3/11 3/11 1/11 (36.4%) (27.2%) (27.2%) (9.0%) botch 3/11 1/11 (9.0%) 1/11 (9.0%) 3/11 3/11 (27.2%) (27.2%) (27.2%) pooch 4/11 1/11(9.0%) 2/11 4/11 0/11 (0%) (36.4%) (18.1%) (36.4%) chew 3/11 1/11 (9.0%) 2/11 5/11 0/11 (0%) (27.2%) (18.2%) (45.5%) chop 3/11 2/11 3/11 3/11 0/11 (0%) (27.2%) (18.2%) (27.2%) (27.2%) 150 WORD + + + N + - - + cheap 2/11 1/11 (9.0%) 1/11 (9.0%) 6/11 0/11 (0%) (18.1%) (54.5%) jaw 3/11 2/11 1/11 (9.0%) 5/11 0/11 (0%) (27.2%) (18.1%) (45.5%) jujube 6/11 1/11 (9.0%) 1/11 (9.0%) 3/11 0/11 (0%) (54.5%) (27.2%) hodge 8/11 0/11 1/11 2/11 0/11 (72.7%) (0%) (9.0%) (18.1%) (0%) Total 36/99 9/99 (9.0%) 15/99 34/99 4/99 (36.4%) (15.1%) (34.3%) (4.0%) Average % 36.6% 9.0% 15.1% 34.3% 4.0% Table 11c: Percentage of criteria noted to have changed for each degree of change for the maximum contact phase of fricative / J/ WORD + + + N + - - + she 6/10 (60%) 1/10 (10%) 0/10 (0%) 3/10 (30%) 0/10 (0%) shah 0/10 (0%) 2/10 (20%) 3/10 (30%) 5/10 (50%) 0/10 (0%) shoe 2/10 (20%) 3/10 (30%) 1/10 (10%) 4/10 (40%) 0/10 (0%) push 2/10 (20%) 1/10 (10%) 0/10 (0%) 7/10 (70%) 0/10 (0%) posh 1/10 (10%) 1/10 (10%) 3/10 (30%) 5/10 (50%) 0/10 (0%) Total 11/50 (22%) 8/50 (16%) 7/50 (14%) 24/50 (48%) 0/50 (0%) Average % 22% 16% 14% 48% 0% 151 Table lid: Percentage of criteria noted to have changed for each degree of change for the release phase of fricative /$/ WORD + + + N + - - + she 5/11 (45.5%) 1/11 (9.0%) 1/11 (9.0%) 4/11 (36.3%) 0/11 (0%) shah 2/11 (18.1%) 3/11 (27.3%) 1/11 (9.0%) 5/11 (45.5%) 0/11 (0%) shoe 4/11 (36.4%) 1/11 (9.0%) 3/11 (27.3%) 3/11 (27.3%) 0/11 (0%) posh 1/11 (9.0%) 1/11 (9.0%) 1/11 (9.0%) 7/11 (63.6%) 5/11 (45.5%) push 0/11 (0%) 0/11 (0%) 1/11 (9.0%) 5/11 (45.5%) 1/11 (9.0%) Total 12/55 (21.8%) 6/55 (10.9%) 7/55 (12.7%) 24/55 (43.6%) 6/55 (10.9%) Average % 21.8% 10.9% 12.7% 43.6% 10.9% Table lie: Percentage of criteria noted to have changed for each degree of change for the maximum contact phase of Igl WORD + + + N + - - + geese 6/6 (100%) 0% 0% 0% 0% goo 3/6 (50%) 1/6 (16.7%) 0% 0% 2/6 (33.3%) gob 5/6 (83.3%) 0% 1/6 (16.7%) 0% 0% egg 2/6 (33.3% 0% 4/6 (66.7%) 0% 0% rog 2/6 (33.3%) 0% 4/6 (66.7%) 0% 0% POg 0/6 (0%) 1/3 (33.3%) 5/6 (83.3%) 0% 0% Total 18/36 (50%) 2/36 (5.5%) 14/36 (38.8%) 0/36 (0%) 2/36 (5.5%) Average % 50% 5.5% 38.8% 0% 5.5% 152 Table llf: Percentage of criteria noted to have changed for each degree of change for the release phase of Igl WORD + + + N + - - + geese 6/6 (100%) 0% 0% 0% 0% gob 4/6 (66.7%) 0% 0% 0% 2/6 (33.3%) goo 3/6 (50%) 2/6 (33.3%) 1/6 (16.7%) 0% 0% egg 0% 0% 3/6 (50%) 3/6 (50%) 0% rog 3/6 (50%) 0% 3/6 (50%) 0% 0% POg 0% 1/6 (16.7%) 4/6 (66.7%) 0% 1/6 (16.7%) Total 16/36 (44.4%) 3/36 (8.3%) 11/36 (30.5%) 3/36 (8.3%) 3/36 (8.3%) Average % 44.4% 8.3% 30.5% 8.3% 8.3% Tables 12a and 12b display the percentage of criteria noted to have changed for each degree of change for the generalization target tokens. 153 Table 12a: Percentage of criteria noted to have changed for each degree of change for the maximum contact phase of the generalization target words WORD + + + N + - - + cop 3/6 (50%) 0% 2/6 (33.3%) 0% 3/6 (50%) eek 2/6 (33.3%) 0% 0% 0% 2/6 (33.3%) bees 3/10 (30%) 0% 0% 7/10 (70%) 0% zoo 4/10 (40%) 1/10 (10%) 0% 4/10 (40%) 0% saw 0/10 (0%) 0% 4/10 (40%) 4/10 (40%) 0% boss 3/10 (30%) 0% 0% 7/10 (70%) 2/10 (20%) Total 15/52 (29%) 1/52 (1.9%) 6/52 (12%) 22/52 (43%) 7/52 (14%) Average% 30.55% 1.6% 12.2% 36.6% 34.3% Table 12b: Percentage of criteria noted to have changed for each degree of change for the release phase of the generalization target words WORD + + + N + - - + cop 4/6 (66.7%) 0% 0% 0% 2/6 (33.3%) eek 2/6 (33.3%) 0% 0% 0% 4/6 (66.7%) bees 2/11 (18.2%) 1/11 (9.0%) 5/11 (45.5%) 2/11 (18.2%) 1/11 (9.0%) zoo 1/11 (9.0%) 2/11 (18.2%) 5/11 (45.5%) 2/11 (18.2%) 1/11 (9.0%) saw 2/11 (18.2%) 2/11 (18.2%) 4/11 (36.4%) 4/11 (36.4%) 0% boss 2/11 (18.2%) 1/11 (9.0%) 4/11 (36.4%) 5/11 . (45.5%) 0% Total 13/56 (23.2%) 6/56 (10.7%) 18/56 (32.1%) 13/56 (23.2%) 8/56 (14.3%) Average % 27.2% 9.1% 27.3% 19.7% 19.6% 154 Figures 1.3 shows the total number of + +, + N, + -, - +, and - - for all treatment and generalization targets in all phases, respectively. 155 Figure 1.3: Total Number of + +, + N, + -, - +, and - - For All Treatment and Generalization Targets in All Phases did it • Treatnen t Targets i Genera lizationl Targets + + + N + - - + - - Type of change from pre- to post-therapy 156 ACTIVITY LEVEL Intelligibility measures were used to assess the effects of palatometry therapy on activity level (WHO, 1999) as perceived by untrained listeners. Listeners completed open-set word and sentence identification tasks and a closed-set word identification task to assess intelligibility. This section outlines the results of each task. Research Questions 1. Do untrained listeners correctly identify more post-therapy words than pre- therapy words in the open-set word identification task? Null Hypothesis 1: There will be no difference in word identification accuracy between the pre- and post-therapy samples. Listeners completed orthographic transcriptions of the words presented in the open-set word identification task. Words were marked right if they were correctly identified and wrong if incorrectly identified. A percentage correct score out of twenty, pre- and post-therapy, was yielded. The scores for word identification, pre- and post-therapy, and the differences in the scores are summarized in tables 13 and 14: 157 Table 13: Number of words correctly identified pre-therapy and post-therapy in the open-set word identification task (out of 20) Listener (L) Pre-therapy Post-therapy 1 0 (0%) 1 (5%) 2 1 (5%) 1 (5%) 3 3 (15%) 1 (5%) 4 0 (0%) 2 (10%) 5 0 (0%) 1 (5%) Average % 4% 6% Table 14: Difference between the percentage of words correctly identified pre- and post-therapy in the open-set word identification task. Listener (L) Difference 1 5% 2 0% 3 -10% 4 10% 5 5% Average 2% 2. Do untrained listeners correctly identify more post-treatment sentences than pre-treatment sentences in the open-set sentence identification task? 158 Null Hypothesis 2: There will be no difference in sentence identification scores between the pre- and post-treatment samples. Listeners orthographically transcribed each sentence and the percentage of words correctly identified was calculated for each sentence. The sentences each listener heard at each visit are displayed in Appendix A. Table 15a: Percentage of words per sentence correctly identified pre-therapy in the sentence identification task. L 1 L2 L3 L4 L 5 Average % *Sentence A 67% 25% 75% 75% 42% 56.8% Sentence B 25% 63% 63% 38% 63% 50.4% Sentence C 33% 33% 33% 44% 44% 37.4% Average % 41.66% 40.3% 57.0% 52.33% 49.6% 48.19% Table 15b: Percentage of words per sentence correctly identified post-therapy in the sentence identification task L 1 L2 L3 L4 L 5 Average % Sentence A 58% 42% 58% 20% 50% 45.6% Sentence B 13% 25% 75% 50% 25% 37.6% Sentence C 22% 22% 11% 22% 22% 19.8% Average % 31.0% 29.66% 48% 30.66% 32.3% 34.32% 159 Table 15c: Difference between the average percentage of words per sentence correctly identified pre- and post-therapy in the sentence identification task. Listener Difference (%) 1 -10.66% 2 -10.64% 3 -9% 4 -21.67% 5 -17.33% Average % -13.86% 3. Do untrained listeners correctly identify more post-treatment target phonemes than pre-treatment target phonemes in the open-set word identification task? Null Hypothesis 3: There will be no difference in phoneme identification accuracy between the pre- and post-treatment samples. The percentage of treatment phonemes pre- and post-therapy correctly identified in the open-set word identification task was calculated. Tables 16 and 17 below, summarize the percentage of treatment targets correctly identified and the difference between the percentages pre- and post-therapy. 160 Table 16: Percentage of treatment targets correctly identified pre-therapy and post-therapy in the open-set word identification task. Pre-therapy: L 1 L2 L 3 L4 L 5 Average % 0 % 0% 33% 0 % 33% 13.2% HI 0% 0 % 33% 33% 0 % 13.2% /ts/ 33% 67% 66% 33% 33% 46.6% /d 3 / 0% 0% 33% 0% 0 % 6.6% Average % 8.25% 16.75% 41.25% 16.5% 16.5% 19.85% | Post-therapy: /g/ 0% 0% 0% 0% 0 % 0 % 0% 0 % 0% 0% 0 % 0 % /ts/ 0% 0% 0 % 33% 33% 13.2% M 3 / 33% 0% 0% 33% 0 % 13.2% Average % 8.25% 0% 0% 16.5% 8.25% 6.6% Table 17: Difference between the percentage of treatment targets correctly identified pre- and post-therapy in the open-set word identification task. L 1 L2 L3 L4 L5 Average % Ig/ 0% 0% -33% 0% -33% -13.2% III 0% 0% -33% -33% 0% -13.2% /ts/ -33% -67% -66% 0% 0% -33.2% /d 3 / 33% 0% -33% 33% 0% 6.6% Average 0% -16.75% -41.25% 0% -8.25% -13.25% 161 4. Do untrained listeners correctly identify more post-treatment generalization target phonemes than pre-treatment generalization target phonemes in the open-set word identification task? Null Hypothesis 4: There will be no difference in phoneme identification accuracy between the pre- and post-treatment samples. The percentage of generalization phonemes pre- and post-therapy correctly identified in the open-set word identification task was calculated. Tables 18 and 19 below, summarize the percentage of generalization targets correctly identified and the difference between the percentages pre- and post-therapy. Table 18: Percentage of generalization targets correctly transcribed pre-therapy and post-therapy in the open-set word identification task. Pre-therapy: L 1 L2 L3 L4 L 5 Average % fkl 100% 100% 100% 100% 100% 100% Isl 100% 67% 67% 67% 100% 80.2% Izl 0% 0% 0% 33% 33% 13.2% Average % 66.66% 55.66% 55.66% 66.66% 77.6% 64.46% Post-therapy: Ikl 100% 100% 67% 67% 67% 62.2% Isl 100% 100% 67% 67% 100% 86.8% Izl 33% 33% 67% 67% 67% 53.4% Average % 77.66% 77.66% 67.0% 67.0% 78.0% 67.4% 162 Table 19: Difference between the percentage of generalization targets correctly identified pre- and post-therapy in the open-set word identification task. L 1 L2 L 3 L4 L 5 Average % Ikl 0% 0% -33% -33% -33% -19.8% Isl 0% 33% 0% 0% 0% 6.6% Izl 33% 33% 67% 33% 33% 39.8% Average % 11.0% 22.0% 11.3% 0% 0% 8.86% 5. Do untrained listeners choose more post-therapy words than pre-therapy words as being correct in the closed-set identification task? Null Hypothesis 5: There will be no difference in word choice between the pre- and post-therapy samples. In analysis of the closed-set word identification data, the number of words correctly identified pre- and post-therapy was calculated. Scores for words identified and the differences between the scores pre- and post-therapy are summarized in Tables 20 and 21 below. 163 Table 20: Number of words correctly identified pre- and post- therapy in the closed-set word identification task (out of 20) Listener Pre-therapy Post-therapy 1 6 (30%) 12 (60%) 2 4 (20%) 9 (45%) 3 3 (15%) 12 (60%) 4 4 (20%) 10 (50%) 5 1 (5%) 9 (45%) Average % 18% 52% Table 21: Difference between the percentage of words correctly identified pre- and post-therapy in the closed-set word identification task. Listener Difference 1 30% 2 25% 3 45% 4 30% 5 40% Average 34% 6. Do untrained listeners correctly identify more post-therapy treatment phonemes than pre-therapy treatment phonemes in the closed-set word identification task? 164 Null Hypothesis 6: There will be no difference in phoneme identification accuracy between the pre- and post-therapy samples. The percentage of treatment phonemes pre- and post-therapy correctly identified in the closed-set word identification task was calculated. Tables 22 and 23 below, summarize the percentage of treatment targets correctly identified and the difference between the percentages pre- and post-therapy. Table 22: Percentage of treatment targets correctly identified pre-therapy and post-therapy in the closed-set word identification task. Pre-therapy: L 1 L2 L 3 L 4 L 5 Average % 33.3% 0% 0% 0% 33.3% 13.2% HI 33.3% 0% 0% 33.3% 33.3% 19.8% / t y 33.3% 33.3% 33.3% 33.3% 33.3% 33.3% /d 3 / 33.3% 33.3% 0% 0% 0% 13.2% Average % 33.3% 16.5% 8.25% 16.5% 24.7% 19.8% Post-therapy Igf 66.7% 33.3% 33.3% 66.7% 0% 39.88% IV 33.3% 0% 33.3% 0% 33.3% 19.8% Itv 100% 66.7% 100% 100% 100% 93.34% /d3/ 66.7% 33.3% 100% 33.3% 100% 66.54% Average % 66.55% 33.15% 66.5% 49.9% 58.3% 54.89% 165 Table 23: Difference between the percentage of treatment targets correctly identified pre- and post-therapy in the closed-set word identification task. L l L2 L 3 L4 L 5 Average % Ig/ 33.3% 33.3% 33.3% 66.7% -33.3% 26.66% III 0% 0% 33.3% -33.3% 0% 0% ity 66.7% 33.3% 66.7% 66.7% 66.7% 60.02% /d 3 / 33.3% 0% 100% 33.3% 100% 53.32% Average % 33.2% 16.5% 58.2% 33.35% 33.35% 35.0% 7. Do untrained listeners correctly identify more post-treatment generalization target phonemes than pre-treatment generalization target phonemes in the closed-set word identification task? Null Hypothesis 7: There will be no difference in phoneme identification accuracy between the pre- and post-treatment samples. The percentage of generalization phonemes pre- and post-therapy correctly identified in the closed-set word identification task was calculated. Tables 24 and 25 below, summarize the percentage of generalization targets correctly identified and the difference between the percentages pre- and post-therapy. 166 Table 24: Percentage of generalization targets correctly identified pre-therapy and post-therapy in the closed-set word identification task Pre-therapy: L l L2 L 3 L4 L 5 Average % Ikl 33.3% 0% 0% 0% 0% 6.66% Isl 33.3% 0% 0% 0% 0% 6.66% Izl 0% 66.7% 66.7% 33.3% 0% 33.34% Average % 22.2% 22.2% 22.2% 11.1% 0% 15.5% Post-therapy Ikl 66.7% 66.7% 100% 66.7% 33.3% 66.8% Isl 66.7% 66.7% 33.3% 33.3% 66.7% 53.34% Izl 33.3% 66.7% 33.3% 66.7% 0% 40.0% Average % 55.6% 66.7% 55.6% 55.6% 33.3% 53.38% Table 25: Difference between the percentage of generalization targets correctly identified pre- and post-therapy in the closed-set word identification task. L 1 L2 L3 L4 L 5 Average % Ikl 66.7% 66.7% 100% 66.7% 33.3% 66.68% Isl 33.3% 66.7% 33.3% 33.3% 66.7% 46.66% Izl 33.3% 0% -33.3% 33.3% 0% 6.66% Average % 44.5% 44.5% 33.3% 44.5% 33.3% 40.0% Tables 26a-c below display the words correctly identified by each listener in each task. Pre-therapy words are indicated as PRE and post-therapy words are indicated as POST. 167 Tables 26a-c: Words correctly identified by each listener for each task (a) Open-set word identification task L 1 L2 L 3 L4 L 5 PRE chicken snake, fishing, page POST jump shoes snake zoo, jump itchy (b) Closed-set word identification task L 1 L2 L 3 L 4 L 5 PRE shoes, pig, judge, zoo, zoo, zipper, shoes, shoes jump, snake, zipper, chicken pig> chicken chicken zipper, chicken POST ice, shoes, ice, shoes, judge, glasses, ice, shoes, glasses, cherries, cherries, pig, cherries, judge, judge, pig, bees, jump, calling, pig, jump, cherries, cherries, pig, jump, snake, itchy, snake, jump, jump, snake, snake, truck, page, itchy, snake, itchy, truck, itchy, chicken, zoo truck, itchy, chicken, zoo truck, zoo chicken, page, bees, zoo chicken 168 (c) Open-set sentence identification task: Sentence A L l L2 L 3 L 4 L 5 Sentence the, is, a, the, light, the, the, the, of, A: pre light, into, beautiful rainbow, is, rainbow, light, beautiful, a, of, light, is, a, of, beautiful, colours many, light, into, colours beautiful, beautiful, colours, colours Sentence the, is, a, the, is, of, the, the, the, is, a, A: post of, light, light, rainbow, is, rainbow, of, beautiful, beautiful a, is, a, light, beautiful, colours beautiful, into, colours colours beautiful, colours Sentence B L 1 L2 L 3 L 4 L 5 Sentence B: pre no, one people, look, but, no, one people, look, no, one, it but, no, one people, look, but, no, one Sentence B: post look look, no people, look, but, no, one, it people, look, no, it look, it Sentence C L 1 L2 L3 L4 L 5 Sentence C: pre when, for, something when, looks, for, some- thing, when, for, something when, looks, for, something when, looks, for, something Sentence C: post when, for looks, for for for, something looks, for 169 Qualitative Comments by Listeners After listening to each word and sentence in the open-set word and sentence identification tasks, listeners were asked to indicate how well they understood what the speaker had said. Listeners circled one of the following: (1) understood very well, (2) understood moderately well, (3) understood hardly anything, or (4) did not understand anything. Tables 27a,b and 28 summarize the listeners' responses to each presentation. Table 27a: Listeners' understanding ratings in the open-set word identification task for pre-therapy words. Target Word Listener 1 Listener 2 Listener 3 Listener 4 Listener 5 Most Frequent Responses glasses 3 2 2 2 2 2 gum 4 1 1 2 1 1 Pig 1 2 1 1 2 1 judge 4 2 1 2 3 2 jump 4 1 2 2 3 2 page 3 1 2 2 3 2 & 3 cherries 4 2 2 2 3 2 chicken 2 1 2 3 2 2 itchy 4 3 3 2 2 3 & 2 shoes 2 1 3 1 3 3 & 1 170 Target Word Listener 1 Listener 2 Listener 3 Listener 4 Listener 5 Most Frequent Responses brush 3 2 2 2 2 2 fishing 2 3 3 2 2 2 calling 3 2 2 2 2 2 truck 1 1 1 1 2 1 snake 3 2 2 2 2 2 soap 2 2 1 1 1 1 ice 2 2 2 2 3 2 zipper 4 4 3 2 3 3 & 4 zoo 4 1 1 . 2 3 1 buzz 3 3 2 3 4 3 Table 27b: Listeners' understanding ratings in the open-set word identification task for post-therapy words. Target Word Listener 1 Listener 2 Listener 3 Listener 4 Listener 5 Most Frequent Responses glasses 2 2 2 1 2 2 gum 2 1 2 1 2 2 Pig 3 1 1 3 2 1 & 3 judge 3 4 3 3 3 3 jump 3 1 1 3 2 1 & 3 page 2 4 3 4 2 2 & 4 cherries 1 2 2 1 1 1 chicken 4 4 2 4 2 4 itchy 3 4 4 3 2 3 & 4 171 Target Word Listener 1 Listener 2 Listener 3 Listener 4 Listener 5 Most Frequent Responses shoes 2 2 1 2 2 2 brush 2 3 2 4 2 2 fishing 3 4 3 3 3 3 calling 4 2 3 2 3 3 truck 3 2 2 3 2 2 snake 2 2 2 1 2 2 soap 2 1 2 1 2 2 ice 3 3 1 1 2 3 & 1 zipper 3 2 3 1 2 2 & 3 zoo 3 3 3 3 3 3 buzz 2 3 3 2 2 2 172 Table 28: Listeners' understanding ratings in the open-set sentence identification task for pre- and post-therapy sentences. Pre- Therapy Listener 1 Listener 2 Listener 3 Listener 4 Listener 5 Most Frequent Responses Sentence A 3 4 1 2 3 3 Sentence B 3 2 2 4 2 2 Sentence C 2 2 2 2 3 2 Post- Therapy Listener 1 Listener 2 Listener 3 Listener 4 Listener 5 Most Frequent Responses Sentence A 2 2 2 2 2 2 Sentence B 4 3 1 3 3 3 Sentence C 3 3 3 2 3 3 PARTICIPATION LEVEL Responses from questionnaires administered to Samantha and listeners' perceptions were used to assess the effects of palatometry therapy on level of participation (WHO, ICIDH-2, 1999). The World Health Organization's concept of participation is "an individual's involvement in life situations in relation to health conditions, body functions or structures, activities, and contextual factors. Participation restrictions are problems an individual may have in the manner or 173 extent of involvement in life situations" (ICIDH-2 Beta-2, 83). This section outlines the research questions and the results from each questionnaire. Results from questionnaires will be descriptive. Research Questions 1. Does Samantha indicate an improvement due to changes in her speech on her level of participation, as based on responses from questionnaires? Null Hypothesis 1: There will be no improvement indicated in Samantha's responses to questionnaires addressing level of participation. This question was addressed through questions selected from two different questionnaires each addressing level of participation. Samantha's responses were collected and analyzed to determine if she indicated an improvement with respect to her participation level. Responses from the first questionnaire were compared with Samantha's responses from the same questionnaire from the previous palatometry project. A descriptive analysis of Samantha's responses is provided below for each questionnaire. Questionnaire #1: Palatometry Research Project Samantha completed this questionnaire three times: (1) after four months of palatometry therapy in a previous study, (2) after eight months of palatometry 174 therapy, and (3) three-months post-palatometry treatment with the present experimenter. The latter administration of this questionnaire included novel questions that were not asked during the previous two administrations because more emphasis was placed on level of participation. The following description summarizes Samantha's responses for each administration of the questionnaire. (1) Questionnaire #1 (1996): Samantha's responses were very general and vague in nature. For example, when asked, "list the things you do now to improve your speech" Samantha responded, "It is too early to say." Additionally, she reported not making any changes in how she spoke as a result of the four months of palatometry therapy. In terms of communication strategies, if someone did understand her, then she would repeat herself twice or write her messages on paper to allow her listener to read her message. Samantha reported enjoying speaking more as a result of palatometry therapy and she stated that she was able to use the targeted sounds outside of the therapy session only when practising. With respect to family and friends, Samantha stated that they had not made any comments about her speech since starting the palatometry program. Furthermore, she reported that she still needed to repeat herself when talking to familiar speakers. 175 (2) Follow-up questionnaire (1996): Samantha reported that her palatometry goals at this time included Iri, Isl, fkl, and III. She strongly agreed that using the palatometer was a good way to provide speech therapy and she reported that her progress had been faster and better using the palatometer than it was without using the palatometer. Interestingly, Samantha's answers were less general or vague. For example, when asked, "list the things you do now to improve your speech" Samantha responded this time by saying that she uses her new speech sounds, Iri, Isl, fkl, and /$/, to improve her speech. She also provided a more detailed answer when asked what she does if someone does not understand her. Samantha responded that she would talk to the person twice, speak very slowly and clearly, and if the person still did not understand, then she would write the message on paper. Unlike her former answer, Samantha reported that she had made changes in how she spoke as a result of palatometry therapy. She reported that her speech had improved and that she felt better. Samantha did not provide more detail with respect to her latter comment. Similar to her previous answer, Samantha reported enjoying speaking more as a result of palatometry therapy. Interestingly, instead of reporting that she was able to use her new speech sounds outside of the therapy session only when practicing, this time Samantha reported that she was using them "occasionally in conversation." Furthermore, 176 she reported that this time, her family and friends had indicated that her speech was improving. Similarly, this time Samantha reported that she repeated herself less often when talking with familiar speakers; however, as previously reported, she still needed to repeat herself as often when talking with unfamiliar listeners. (3) Questionnaire: This Study Therapy goals for the present palatometry project included, /g/, /$/, /t$/, and M 3 / . Samantha reported that she uses these new sounds to improve her speech. Similar to her previous two answers, Samantha indicated that she repeats herself when someone does not understand her or she records her message on paper. She also reported that she has made changes in how she spoke as a result of this present palatometry project. Unlike her previous two answers however, this time she provided a more detailed explanation by stating that she produces Igl with "voice," and /d3/ with "a little bit of voice and air." These explanations indicate improved awareness of her productions. Samantha reported enjoying speaking more as a result of palatometry therapy, and her family and friends had stated that Samantha's speech was improving. Significant change was noticed from her last two answers when asked if she was able to use her new sounds outside of therapy. This time, Samantha 177 reported being able to use her new sounds all the time. This indicates improved awareness of her new productions and possibly an increase in her self-confidence because she was attempting to use her new sounds in many environments with both familiar and unfamiliar speakers. Similar to her last answer, she noticed that she was repeating herself less often when talking with familiar speakers. However she still reported that she needed to repeat herself when conversing with unfamiliar speakers. As previously mentioned, additional questions were asked that focused on Samantha's perceptions of her speech, her speaking abilities, and her level of participation within society. The following description summarizes Samantha's answers. With respect to Samantha's speech, she reported that her speech will improve "a lot" if she continues to work hard and she felt that she could make strangers feel at ease with her "some of the time." Interestingly, even though Samantha reported that she needed to repeat herself when conversing with unfamiliar speakers, she reported that her speech was difficult for strangers to understand only "a little bit" and "some of the time." These same answers were reported when asked if her family finds her difficult to understand. Additionally, Samantha reported "a little bit" when answering the following questions: if her speech was slow, fast, too loud, too soft, and if her voice sounded hoarse or harsh. Lastly, Samantha reported that her speaking "never" becomes poorer when she 178 fatigues and, interestingly, she reported that her speech does not have a nasal quality. Generally, Samantha reported that she avoids some situations "a little bit" or "a medium amount" because of her hearing loss. For example, situations that Samantha answered with "a little bit" included the following: going for coffee with a friend, at a restaurant with your friend, talking on a telephone, and going to church. For the latter situation, Samantha also recorded that she avoids going to church only "a little bit" because she can "hide her hearing aids behind her ears and cover them with her hair." Situations that she responded "a medium amount" to included: going to class and talking with teachers/professors. When asked if she avoided going to a movie because of her hearing loss, Samantha responded "not at all." The same response was given when asked if she avoided going shopping in a shopping mall. Questionnaire #2: Questions selected from The Communication Profile for the Hearing Impaired (CPHI. Demorest & Erdman, 1986. 1987") and from Erickson's S-24 Scale (S-24 Scale. Erickson. 1969). Generally, questions from the CPHI were asked that focused on one of the following three areas: (1) types of communication strategies used by Samantha, (2) Samantha's feelings and attitudes about her hearing loss, and (3) the kinds of experiences Samantha may have had when communicating with others. 179 Samantha's responses with respect to the types of communication strategies used by Samantha indicated that she uses a variety of strategies either "half the time," "frequently" or "almost always." Strategies that she reported using almost always include positioning herself where she will be able to hear, and asking for repetition when she doesn't understand what someone has said. Samantha reported using the following strategies frequently: paying close attention to her listener's face, and asking her family to get her attention before they speak to her. Strategies such as positioning herself where there is less distraction, avoiding conversing with others, staying in a well-lit area so she can see her speaker's face, explaining that she has a hearing loss when someone does not appear to understand her, and asking for repetition when she didn't hear, were reported by Samantha as being used "half the time." Samantha reported a range of responses for questions focusing on her feelings and attitudes about her hearing loss. For example, she reported that she sometimes did not enjoy going places with her friends because she has trouble hearing, and she sometimes finds it difficult to follow conversations at social gathering because of her hearing loss. Additionally, she reported that the difficulties she has with her hearing restrict her social and personal life some of the time and she reported that she felt discouraged sometimes when she was not able to understand. According to Samantha, she rarely feels embarrassed when she has to ask someone to repeat what s/he has said, and she reported that she almost always gives the impression of normal hearing. 180 Samantha reported that she sometimes communicates with others in a group situation that may be unfamiliar to her. However, she almost always has conversations with familiar speakers. She also reported that people speak frequently to her when they are not facing her and she sometimes finds it difficult to follow a conversation when others are talking nearby. Samantha's responses from Erickson's S-24 Scale showed little variation. Generally she responded with "sometimes." This response was chosen by Samantha when asked the following statements: she finds it easy to talk with almost anyone, a person who is her teacher or her boss is hard to talk to, some words are harder than others for her to say, she would rather not introduce herself to a stranger, she feels comfortable speaking before a group, her speaking voice is rather pleasant and easy to listen to, and even though she knew the right answer she has failed to give it because she was afraid to speak out. Additionally, Samantha reported that she "never" finds it easier to talk with a person who is younger than herself, and she reported being able to talk better than she writes "a medium amount." Lastly, she reported feeling most comfortable conversing with her family "all the time." Research Questions #2 2. How do listeners perceive their level of participation with Samantha based on their perceptions from the listeners' tasks? 181 Null Hypothesis 2: There will be no evidence of participation with Samantha in contexts involving oral modes of communication. This question was addressed by having each listener answer, if s/he chose to do so, a total of 14 questions (2 questions the first visit and 12 questions the second visit). Table 29 displays the questions. Table 30 summarizes the answers given by each listener. The initials NR represent no response, L represents listener, and Q represents question. Table 29: Questions asked to each listener 1. If given the opportunity, where would you feel comfortable having a conversation with this speaker? 2. What kinds of activities would you feel comfortable participating in with this speaker? 3. If given the opportunity, how often would you like to go to the movies with this speaker? 4. If given the opportunity, how often would you like to go for coffee with this speaker? 5. If given the opportunity, how often would you like to talk on the telephone with this speakers? 6. If given the opportunity, how often would you like to go shopping in a shopping mall with this speaker? 7. If given the opportunity, how often would you like to participate in a work project with this speaker? 8. For approximately how long would you like to have a conversation with this speaker? 9. If given the opportunity how often would you like to participate in a letter exchange (i.e., write letters) with this speaker? 10. If given the opportunity, how often would you like to participate in a team athletic event (i.e., soccer or volleyball) with this speaker? 11. If given the opportunity, how often would you like to email this speaker? 12. Would you feel comfortable conversing with this speaker? 13. What problems do you think this speaker may have, if any, in the following situations: (a) work, (b) school, (c) family (d) peer group. 14. Is there anything you would like to say about communication with this speaker or others with similar speech? 182 Table 30: Listeners' responses 1 1 Qi Q2 Q3 Q4 L l (f) nowhere I can think of I would need to know more about the person sometimes NR L2 NR NR sometimes sometimes L3 (b) at a quiet coffee place, (d) at a library (b) a game of doubles tennis where this speaker is your partner, (c) a game of chess, (d) shopping in a mall, (e) going to the movies. sometimes sometimes L4 (e) in a quiet place without too many people present (b) a game of doubles tennis where this speaker is your partner, (c) a game of chess sometimes hardly ever L5 (b) at a quiet coffee place, (e) anywhere quiet enough (a) a college project, (b) a game of doubles tennis where this speaker is your partner, (c) a game of chess, (d) shopping in mall NR NR 1 Q5 Q6 Q7 Q8 L l not at all NR depends on the work between 3-5 minutes L2 sometimes sometimes sometimes more than 5 minutes L3 hardly ever sometimes sometimes more than 5 minutes L4 not at all hardly ever sometimes, depends on the work NR L5 NR NR NR NR 183 1 Q9 Q10 Q U Q12 L l not at all not at all, I'd have to see them play NR slightly comfortable L2 sometimes sometimes all the time very comfortable L3 all the time all the time all the time slightly comfortable L4 sometimes sometimes sometimes slightly comfortable L5 NR NR NR NR 1 1 Q13 Q14 L l NR It can be frustrating and also mildly depressing. L2 a) presentations, (b) children teasing speaker causing speaker to be afraid to ask questions, thus hindering education, (c) none, (d) peers being impatient while they are speaking. It would be easier to speak in person to them than over the phone, because one has context. L3 (a) no problem, (b) when giving oral presentations, (c) no problem, (d) no problem no L4 (a) being understood, (b) asking questions NR L5 NR NR 184 CHAPTER 4 DISCUSSION The primary goal of this study was to evaluate the outcomes of palatometry therapy in relation to the World Health Organization's (1999) model of body (structure and function), activity, and participation. This chapter closely examines the results that were derived for each of the three levels of functioning: body, activity, and participation. In addition, the clinical implications of this study, its limitations, and directions for further research will be discussed. BODY LEVEL Data for this level of functioning included narrow transcriptions of pre- and post-therapy words lists, narrow transcriptions of treatment target phonemes produced in therapy, narrow transcriptions of palatometry assessment words, and palatogram results. This section discusses the results and compares the findings to previous research. 185 Narrow Transcriptions of Pre- and Post-Therapy Word Lists As indicated in Table 2.5 from the results section, consonant and vowel matches with adult targets showed improvement from pre- to post-therapy evaluations. Results showed a gain of 13.18% (from 25.66% to 38.84%) in consonant match proportions with the adult targets. Vowel match proportions with the adult target showed a gain of 9.41%, from 43.04% to 52.45%. Both of these results represent a reduction in impairment to a slightly less severe level. Specific therapy targets and generalization targets showed gains from pre- to post-therapy. For example, the velar Igl showed a 16.1% gain in match proportions with the adult target, from 11.1% to 27.2% accuracy. The generalization target, fkl showed a 9.2% gain in match proportions with the adult target, from 70.8% to 80.0%. A decrease in match proportions with adult targets from pre- to post-therapy, however, was noted for / J/ in word-final position, for word-initial Id^,/, and for labials and labial clusters. Even though fewer of Samantha's post-therapy word-initial Id^l productions were correct for place, more post-therapy than pre-therapy productions contained a stop. The fricative portions, however, were generally produced in the alveolar region which contributed to the decrease post-therapy. Since therapy focused only on / J/ in word-initial position the lack of improvement with word-final III, from pre- to post-therapy, was to be expected. Similarly, therapy did not target any labial 186 sounds therefore no gains were expected. These regressions will be discussed in more detail throughout this section. Both stops and nasals showed gains of 14.1% and 1.4% respectively. The improvement from pre- to post-therapy for stops was due to the gain in match proportions with the adult target for Igl and fkl. The therapy target Igl showed slight improvement both in place and voicing (e.g., [dorsal] + [voice]), with one of the post-therapy tokens containing word-final Igl being voiced and produced correctly, but with all other tokens in all word positions being either devoiced or produced as [k]. The generalization target fkl, showed improvement in place post-therapy. Thus, the improved changes in velar production suggest a positive effect of palatometry treatment. Both the treatment and the generalization target became better established. All other non-therapy target stops and nasals (e.g., fm, n, b, p, t, dl) either showed slight improvement from pre- to post-therapy, deteriorated, or remained the same pre- and post-therapy with the exception of Itl, which improved from 5/11 matches (45%) pre-therapy to 4/5 matches (80%) post-therapy. Match proportions with the adult target for the phoneme Idl were 7/14 pre-therapy and 1/2 post-therapy. The nasal Ird showed improved matches with adult targets post-therapy: 27% matches post-therapy compared with 18% matches pre- therapy. There were fewer placement errors post-therapy. The improvement with the alveolar stops and nasals was possibly due to the palatometry therapy on the stop portion of the affricates. Since the affricates are produced initially 187 with a stop in the alveolar/post-alveolar region the alveolar place was targeted in therapy. The nasal /rj/ had 18/24 (75%) matches with the adult target pre- therapy and 8/12 (67%) matches post-therapy. Thus, there was no generalization at the end of therapy in terms of place to the velar nasal /rj/. Most of Samantha's post-therapy /rj/ productions were produced as [kh]. This indicates that she had the correct place of articulation but incorrect voicing and manner. Since voicing showed some signs of improvement with the treatment target Igl, perhaps with more therapy, voicing would improve for the velar nasal as well. The labials and labial clusters were the only types of sounds that did not show improvement from pre- to post-therapy. In fact, a negative gain of -5.0% was seen in terms of match proportions with adult targets from pre- to post-therapy. No gain was expected at the start of this study. Therapy did not target any labial sounds and the palatometer does not provide visual feedback information in terms of labial production. Regression however, was in manner and voicing, not in labial place of articulation. This indicates that Samantha had the correct place of articulation for all labials, but had difficulty combining the following features: [- sonorant] plus [- continuant] plus [voice], and [+ continuant] plus [voice]. The coronal fricatives and affricates improved in terms of place of articulation and in manner of production. Results indicated a gain of 11.3% in match proportions with the adult targets in terms of place of production. For 188 example, the treatment target It 11 showed better place of articulation post- therapy than pre-therapy (12.5% pre-therapy versus 44.4% post-therapy). Generally, pre-therapy lt\l productions were produced as a tense /$/, sometimes with an initial [t] (e.g., [tH J]). Other incorrect productions included an initial [s] (e.g, [st I ]) or an initial [n] (e.g., [nts]). Post-therapy productions showed improved manner of production and place of articulation across all word positions. Four out of nine productions were produced correctly and each post- therapy Itl I production was noted to have better fricative portions. The improvement in frication indicates that Samantha improved her tongue grooving for this target, thereby suggesting an effect of visual feedback given by the palatometer. The same improvement was noted for the treatment target M 3 / . None of Samantha's pre-therapy productions for /d3/ were correct; however, one out of nine or 11.1% of her post-therapy productions were correct. Both place and voicing improved from pre- to post-therapy. For example, pre-therapy, there were six incorrect productions that were produced with no voicing; however, there were only four of such productions post-therapy. For one of these four nonmatching tokens Samantha inserted a schwa at the end of her production (e.g., [ 11&\). This indicates that Samantha was aware that voicing was needed; she added a vowel to compensate for her lack of voicing throughout the phoneme. There was one instance in which Samantha produced the velar [k] for /d3/ pre-therapy; 189 however, all post-therapy /d.3/ productions were correct for place of articulation. Other post-therapy inaccuracies included omission of either the alveolar stop or of the coronal fricative. For example, there was one instance in which M 3 / was produced as [3 ] and there was another instance in which / d3 / was produced as [dts]. Thus, both treatment targets / d3 / and Itl I showed improvement in terms of place of articulation and manner, and voicing of the voiced affricate showed improvement. The treatment target III did not show signs of improvement from pre- to post-therapy in terms of match proportions: 1/6 or 16.6% for both pre- and post- therapy. The one correct production of III was produced tensely (e.g., [SSS]) both pre- and post-therapy. In both cases, Samantha's [SSS] occurred word-medially in the following words: fishing (pre-therapy) and brushing (post-therapy). Samantha's production indicates that is aware of the correct place of articulation but more work is needed to better establish the III place of articulation, improve tongue grooving, and reduce tension. Post-therapy inaccuracies included the addition of a consonant in front of III or a III that was produced further forward in the mouth and was therefore ungrooved (e.g., between coronal and alveolar place). For example, three out of six tokens that ended with the III phoneme were produced as [2 S ], [ n S ], and [ t S ]. In addition, analysis of the pre-therapy 190 transcriptions indicated that Samantha was substituting [s] for III and therefore had too far forward a tongue position. Improvement however, in terms of tongue placement was evident post-therapy because there were no instances in which [s] replaced III. This improvement suggests an effect of visual feedback given by the palatometer. Furthermore, even though there was no increase in match proportions with the adult target from pre- to post-therapy for this treatment target, improvement was evident in terms of the types of errors Samantha was making. As discussed above, for example, tongue grooving and placement appeared to have improved from pre- to post-therapy. In comparison to the affricate treatment targets, Samantha had more difficulty producing [J] than she did [t$] and [d3]. This suggests that the oral stop portion of the affricates might have facilitated her productions in terms of place of articulation. EPG research has shown that the stop component of Itl I and /d3/ differs from Itl and Idl in terms of place of articulation. Tongue contact for Itl and Idl in the affricates is produced more in the postalveolar region than in the alveolar region (Fletcher, 1989; Gibbon, Hardcastle, & Dent, 1995). As a result, achieving correct linguapalatal placement for the sibilant portion of the affricates is facilitated because of the coinciding location of the stop constriction (Fletcher, 1989). In addition, it has been reported that producing a groove configuration is more complex than articulating Itl and Idl (Gibbon, Hardcastle, & Dent, 1995). Ill production requires immediate groove configuration in the 191 postalveolar region; however, the affricates require groove configuration after production of the stop component which also occurs in the postalveolar region. A positive change was also noted for the coronal generalization target /s/. Match proportions with adult targets were 5/17 (29%) pre-therapy and 7/17 (44%) post-therapy. Many of Samantha's pre-therapy productions of [s] were produced with affrication (e.g., [tch ], and [tl] for /s/) or were produced as an alveolar stop (e.g., [t] and [th] for /s/). Improvement however, was noted post- therapy. Fewer of Samantha's productions of [s] were affricated and there were no instances in which [t] replaced [s]. Since Samantha learned to distinguish between the alveolar and the palatoalveolar placements during therapy, it was assumed that improvement in terms of place and manner would be evident with /s/. Indeed, this was case; thus, generalization from the affricates and III to [s] appeared to have occurred. A gain of 23.5% was also seen for /s/-clusters such as /sn, st, and str/ in terms of place of articulation. One token each of [sn] and [str] and two tokens of [st] were produced correctly. The [sn], [str], and [st] clusters may also have generalized from training on the distinction between the alveolar and palatoalveolar place of articulation for the affricates. There was no increase from pre- to post-therapy in terms of match proportions with the adult targets for the generalization target Izl. Results for both pre- and post-therapy were 0/14 (0%) and 0/15 (0%) respectively. Similar 192 substitutions were made pre- and post-therapy. For example, [ts], [s], and [sss] frequently replaced word-final [z]. There was one instance, in the post-therapy sample, in which [z:?2dj] replaced word-initial [z] in zoo. This indicates that Samantha was aware of the correct place of articulation but she had difficulty producing the consonant and the vowel together. Generally, each of Samantha's substitutions had the correct place of articulation (e.g., coronal place) and manner; however, the feature combination [coronal] + [voice] + [continuant] was difficult for Samantha pre- and post-therapy in all word-positions. Positive changes were noted in terms of voicing and place of articulation. For example, in terms of voicing in word-initial position, a gain of 50% was noted and in word- final position a gain of 12.5%. Inaccuracies included Izl being produced as [s], [ts], or [sss]. Overall, a 13.3% gain was observed for voicing. Place of articulation improved from pre- to post-therapy in word-initial position only. Thus, voicing for Id^l and possibly for Igl appeared to have generalized and the place of articulation for the affricates may also have generalized. Non-targeted phonemes also showed improvement post-therapy. These included Iwl, Itl, clusters containing [w], and the glottal [2]. The improvement in [w] both in singletons and in clusters may reflect the training for the dorsal velar Igl. Since [w] contains both the [labial] and [dorsal] place features and since training did not focus on labial production, the training on place of articulation for the Igl treatment phoneme may have generalized to [w]. The lxl also showed a 193 gain in match proportions with the adult target from pre- to post-therapy. As discussed in the Results Chapter, there were fewer instances in which [w] was inserted for [r] in [spr] and [str] clusters. One of the treatment goals from the previous palatometry project that Samantha took part in was the Irl phoneme. Perhaps this present therapy project reminded Samantha of her previous palatometry treatment goals, specifically the Irl phoneme, and thus she attempted to improve her Irl productions even though they were not the focus of this therapy project. Furthermore, because some of the treatment targets focused on palatoalveolar place of articulation (e.g., the affricates), perhaps generalization to post-alveolar [r] occurred. The glottal [?] also showed improvement from pre- to post-therapy. Specifically, 2/14 (14.3%) and 5/7 (71%) match proportions with adult targets were correct pre- and post-therapy respectively. In summary, all therapy and generalization targets except for / J/ showed improvement from pre- to post-therapy in terms of place of articulation, manner, and voicing. The palatoalveolar fricative did, however, appear to have better tongue grooving post-therapy compared with pre-therapy. The generalization target Izl had improved voicing in word-initial and word-final position post- therapy compared with pre-therapy. Non-target stops and nasals, specifically Itl and InJ, also showed improvement. Nasal /rj/ however, did not show improvement post-therapy and thus it was concluded that generalization in terms of place and 194 voicing from Igl to /rj/ did not occur. Improvement was not evident for labials and labial clusters. The regression in match proportions with adult targets from pre- to post-therapy was in manner and voicing, not in labial place of articulation. Additionally, improvement was noted for /s/-clusters in terms of place of articulation and for /w/, /r/, clusters containing [w], and the glottal [?]. Many of the errors Samantha was making were similar to the errors made by a ten-year-old deaf girl named Sophie (Parsloe, 1998, as discussed in the Introduction). Sophie had difficulties distinguishing between the alveolar, palatoalveolar, and velar places of articulations and also had difficulties with voicing for the alveolar place of articulation (Parsloe, 1998). Samantha had fricative and velar errors and also had difficulty voicing the generalization target Izl and the treatment target Igl. Furthermore, some of Samantha's /tJ/ and M 3 / productions indicated that she was sometimes not distinguishing between the alveolar and palatoalveolar places of articulations. Further comparisons with the study by Parsloe will be presented in the palatometry discussion section below. Narrow Transcriptions of Treatment Target Phonemes in Therapy In the narrow transcriptions of treatment target phonemes produced in therapy, the palatoalveolar fricative and the affricates each showed improvement 195 from pre- to post-therapy. Treatment target Igl, however, was produced as [kh] pre- and post-therapy during blocks one and two. Generally more gains were noted after palatometry therapy for Itl I, and M 3 / than after non-palatometry therapy. For Igl, no differences between palatometry and non-palatometry treatment were noted. Interestingly, as noted in the above discussion, one of the word-final Igl targets from the post-therapy transcription was produced with correct voicing. Perhaps voicing for Igl was easier in the word-final position because of continuation of voicing from the preceding vowel. Based on these results and the results from the pre- and post-assessment word list transcriptions Samantha appeared to be having difficulty combining the following features: [dorsal] + [voice]. Block one pre-therapy recordings of/d3/, were produced as a backed [s] (i.e., [s>1, produced between alveolar and palatoalveolar regions). For the block two recording however, five out often M 3 / targets were produced as [t I]. In terms of features, the [t I] phoneme matched the treatment target M 3 / more closely than Samantha's backed [s>] production. For example, [t I ] differs from / d3 / only in terms of the feature [voice] (e.g., M 3 / is voiced and [t I ] not voiced). The backed [s>] however differs from M 3 / in terms of the following features: [continuant], [voice], and [spread glottis]. Thus, [t$] was more desirable since both place of articulation and affrication improved. Similar results were found for the pre- 196 and post-assessment word list transcriptions. The feature [voice] in combination with the place of articulation (e.g., [coronal]) however, remained a difficulty for Samantha even after therapy. As mentioned in the above discussions, voicing has been reported to be particularly difficulty for persons with severe to profound hearing impairments (Alpiner & McCarthy, 1987). The treatment target /tJ/ showed improvement between blocks and within blocks. Samantha's pre-therapy block one errors were similar to those reported in the above discussion on the pre- and post-assessment word list transcriptions. After block one, the only correctly produced token was chair. Interestingly, during the treatment block focusing on /d.3/, the only correctly produced token was jar. Both jar and chair may be produced with mid or low vowels. Thus, perhaps Samantha had more success producing the affricate targets when an open vowel followed the phoneme. Tongue to palate contact is absent with an open vowel and therefore articulation patterns between the initial consonant and the preceding vowel are minimized. It was also observed that Samantha was more aware of the need for stop closure during the second block. None of Samantha's productions during the first block were judged to have contained stop closure. Research has consistently reported that individuals with moderate to profound hearing losses have particular difficulty with sounds such as Itl that are produced within the oral cavity and that are less readily observed on the lips (Alpiner and 197 McCarthy, 1987). Perhaps the visual information on tongue to palate contacts offered by the palatometer aided Samantha in her [t] productions for It ll. Improvement was noted for III between the first and second therapy blocks and between palatometry and non-palatometry therapy for both blocks. Samantha appeared to have more success articulating III before mid or low vowels as she did for /tj/. Tokens containing high front and back vowels such as sheet and shoot respectively, were produced with a fronted [I]. Other pre-therapy inaccuracies included / J/ being produced with excess frication (e.g., [I $]), a fronted ll] (e-g. [S<])> a n d a III with slight affrication (e.g, [(t)J]. These errors were also noted in the pre-therapy word list transcriptions. Research on speech production of speakers with hearing impairments has shown that "some consonants are more conducive to air expenditure, especially the aspirants and sibilants" (Alpiner and McCarthy, 1987, 199). Fewer post-therapy productions contained excess frication. In the second block, the gains seen after non-palatometry therapy may have been due to the visual feedback information offered by the palatometer or to the information given during non-palatometry therapy. It is possible however, that the visual feedback information from palatometry therapy aided Samantha in her non-palatometry therapy productions. 198 Narrow Transcriptions of Pre- and Post-Therapy Palatometry Assessment Data and Pre- and Post-Therapy Palatograms The following section discusses the results for each treatment and generalization target. Results from the perceptual measure (e.g., transcriptions) and from the instrumental measure (e.g., palatometer) will be discussed and compared with each other and the results will be compared with results from other research studies. Treatment Target Ity (I) Perceptual Evaluation Of the three items containing word-initial Ity two of them were perceived to have changed from pre- to post-therapy and to have shown improvement. For example, chop was produced by Samantha with the correct target (e.g., [tj]) and the token cheap was produced with a fronted Ity (e.g., [t$<]). This latter production was judged by the experimenter and her supervisor, to be an improvement since the pre-assessment token of cheap was produced with [J ]. Two of three word-final Ity tokens showed change and improvement from pre- to post-therapy. 199 (II) Instrumental Evaluation: Stop Closure and Fricative Phases of Word-Initial Ity Tokens Overall, 40%-50% of all changes from pre- to post-palatograms for all three word-initial tokens were rated as an improvement (+ +). In terms of stop closure, all three words showed a desired increase in contact from pre- to post- assessment. These results were compatible with the results from the perceptual evaluation since the stop portion of the affricate was perceived for both chop and cheap (see above for transcription details). The treatment target in the token cheap was perceived to be produced with a fronted [t$<]. Analysis of the palatogram data for this token showed that more contact was noted in the alveolar and palatal medial region and in the alveolar tip region than was for example, chop. This excess contact probably resulted in the perception of fronting. The majority of fricative phase criteria, specifically 54.5%, were shown to have changed negatively (e.g., change was observed but improvement was not evident, + -). Results from the perceptual analysis for these tokens were compatible with the palatogram results. For example, the fronted [tj<] production for the token cheap may be explained in terms of the criterion, right side medial alveolar and palatal contact. Post-therapy, more contact was noted in this region than desired, giving rise to more fronted contact. For chew, both pre- and post-perceptual evaluations resulted in a [$] being produced. Analysis 200 of the groove width measure on the palatograms showed that groove width was too wide pre- and post-therapy. Thus, little alveolar contact was noted on the palatogram or judged perceptually. Interestingly, slightly more stop closure contact was noted for the stop closure phase of this phoneme. The extra contact, however, was not enough to improve the perceptual rating. Stop Closure Phase and Fricative Phase of Word-Final /t$/ Tokens The token each was produced pre- and post-therapy with a slightly affricated and fronted /tJ/ production (e.g., [(t)f<]). The tongue-palate contact seen on the pre- and post-palatograms supported the perceptual results. More contact was observed in the alveolar and palatal regions and stop closure was noted. The token botch was perceived to be produced with a fronted [t$<] post- therapy. In terms of contact seen on the palatograms, alveolar region contact for the fricative portion was judged to be too forward post-therapy. The treatment target in pooch was perceived to change from [ts>] to [J] pre- to post-therapy. In terms of the palatograms less stop closure was observed post-therapy, hence no stop contact would have been noted perceptually, as was the case. Summary In terms of the criteria used to assess the palatograms, the stop closure measurement and the alveolar tip region contact measurement appeared to have the most affect on the perceptual evaluation. For example, when less stop closure was observed, as was the case in botch above, stop closure was not 201 perceived. When more alveolar region contact was noted fronted productions were perceived. Thus, for this treatment target, treatment was aimed primarily at increasing Samantha's awareness of the distinction between the alveolar and palatoalveolar place of articulation and improving her lingual placement and configuration. These goals were identical to some of the goals in the study by Dent, Gibbon and Hardcastle (1995), as discussed in the Introduction of this thesis. In Dent et al.'s study, it was shown that palatometry therapy was successful for a group of young adults with motor speech impairments who needed to improve their tongue-palate contact patterns and to become more aware of lingual placement. In terms of the overall results from the stop closure phase palatograms for both the word-initial and word-final tokens, 21/60 criteria or 35% showed change and improvement from pre- to post-therapy (e.g, + +). Similarly, 19/66 (28.7%) criteria for the fricative phase showed change and improvement (e.g, + +). For both the stop closure and the fricative phases, more positive changes (e.g., + + and - +) than negative changes (e.g., - - and + -) were noted. The results from the palatograms indicate that Samantha improved her place of articulation for /t$/ and showed better distinction between the alveolar and palatoalveolar places. These results are consistent with the results from the pre- and post-therapy word list transcriptions. Significantly more post-therapy word list productions were correct across all word positions than were pre- 202 therapy word list productions. In addition, it was noted that each of Samantha's Itl I word list productions had better frication. This result was found for four out of six (67%) of the word-initial and word-final palatogram tokens. Samantha's case is similar to the participant's case in Parsloe's (1998) study, specifically with respect to the degree of hearing-impairment (i.e., profound hearing-impairment), the sounds targeted during therapy (i.e., palato- alveolar fricatives and affricates, velars, and alveolars), and the need to obtain distinct phonological contrasts. In Parsloe (1998), similar results were found immediately after and three months post-palatometry therapy. Palatograms of participant Sophie's Itl I recordings showed a significant improvement post- treatment in all word positions. In addition, improvement was noted in terms of her alveolar, palatoalveolar, and velar distinctions. Treatment Target Id* I (I) Perceptual Evaluation The only token containing the target M 3 / that was perceived to have been worse post-therapy was jaw. Both jujube and hodge. however, were perceived to have improved from pre- to post-therapy. For jaw, the post-therapy target was produced as a backed /s/ (e.g, [s>]), however the pre-therapy target was produced as [J:]. This latter production was more desirable because it contained better tongue grooving and frication. For jujube and hodge. post-therapy productions of 203 the treatment target were [z>] and [(d)$<] respectively. Each of these productions were judged to be an improvement from the pre- therapy productions. Pre-therapy, the treatment target in jujube was produced as [s>] and the target in hodge was produced as [tsh ]. The change from [s>] to [z>], for the token jujube was judged to be an improvement because the feature [voice] was evident post-therapy. For hodge. considerable improvement was perceived. Pre-therapy, Samantha produced a voiceless alveolar stop followed by an aspirated schwa (perhaps the addition of the schwa was her attempt at voicing). Post-therapy however, her production contained the voiced alveolar stop and frication. (II) Instrumental Evaluation: Stop Closure Phase and Fricative Phase of Word-Initial & Word-Final /d 3 / Tokens Results from palatograms supported the perceptual evaluation results. For example, for jaw the negative change from [$:] to [s>] was perceived due to more stop closure contact. In order to produce an alveolar, more contact along the alveolar tip region should be noted on the palatogram, as was the case. Furthermore, because the post-therapy production was perceived to be produced somewhere between the alveolar and palatoalveolar regions (e.g., [s>]), more alveolar region contact would need to have been noted on the post-therapy palatogram, as was the case. For iuiube. the only significant change from pre- to 204 post-therapy was the addition of voicing, as explained above. Since this feature cannot be assessed by the palatometer, analysis of the palatograms for this token did not indicate the change in voicing. However, palatogram analysis did support the perceived place of articulation (e.g, [coronal]). For example, fricative portion alveolar region contact was observed to be too forward pre- and post- therapy and the groove width was noted to have improved but was in fact very narrow. Both Is/ and /z/ are produced with a narrower groove width than l^l. Thus, the observation of the narrow groove width was compatible with the perception of [s] and [z]. Furthermore, the lack of stop closure was observed both perceptually and on the palatograms. The token hodge showed significant improvement pre- to post-therapy in terms of the perceptual and palatogram evaluations and results from palatograms supported the perceptual findings. Summary In terms of the results from the stop closure phase palatograms for both the word-initial and word-final tokens, 7/30 criteria (23.2%) showed change and improvement from pre- to post-therapy (e.g, + +). For the fricative phase, 17/33 (52%) criteria showed change and improvement (e.g, + +). Overall, more positive changes (e.g., + + and - +) than negative changes (e.g., - - and + -) were noted for both phases of the Id?,/ palatograms. The perceptual and instrumental results for this treatment target suggest that Samantha became more aware of the correct lingual placement and 205 configuration, and the distinction between the alveolar and palatoalveolar regions. Both the stop phase and the fricative phase showed improvement from pre- to post-therapy. Similar findings were reported by Dent, Gibbon, and Hardcastle (1995) as previously discussed. Treatment Target/W (I) Perceptual Evaluation Two out five tokens containing the treatment target /$/ were judged to have improved from pre- to post-therapy. For she, the pre-therapy production of /$/ was as [f<] (fronted); however, the post-therapy production was as [J]. This change represented improvement since the correct place of articulation was produced post-therapy. The token shah, also showed improvement. Pre-therapy Samantha produced III as [s] and post-therapy she produced the target phoneme as a backed Is/ (e.g., [s>]). The change from [s] to [s>] was judged to be an improvement since the post-therapy production had better place of articulation than the pre-therapy production. For the remaining tokens, change was noted; however, improvement was not perceived. For example, the pre-therapy production of shoe was as [J<] (a fronted l\f) and the post-therapy production as [ s>] (a backed Isl). The pre-therapy production was in fact more desirable because it contained better frication than the post-therapy production. For posh. 206 the pre-therapy production was as [$]--the intended target; however, the post- therapy production was produced with slight affrication, [(t)J:]. Similarly, the pre-therapy production of push was in fact the treatment target [j], but the post- therapy production was a slightly affricated, backed Is/ (e.g., [ts>]). These changes suggest that therapy was successful for only two out five iy tokens in terms of the perceptual evaluation. (II) Instrumental Evaluation: Maximum Contact and Release Phases of Word-Initial iy Tokens Improvement was noted for the token shah both perceptually and in terms of the palatograms. Post-therapy a backed Is/ was perceived (e.g., [s>]) which meant that the place of articulation had to be somewhere between the alveolar and palatoalveolar regions. Results from the palatograms supported the perceptual findings since less alveolar tip region contact was noted post-therapy on the maximum contact phase palatogram. Also noted for this phase, was less dorsal medial and less alveolar and palatal region medial contact. The reduction in medial contact was also perceived perceptually, since less tongue grooving was noted. Improvement was noted for the token she both perceptually and on the palatograms. As already mentioned, Samantha produced the intended target post-therapy. Results from palatograms supported the perceptual results. 2 0 7 Lateral contact in the alveolar and palatal regions was pulled back more post- therapy than was pre-therapy. This finding was compatible with the transcription results, since pre-therapy the treatment target was perceived to be produced as a fronted /J/. Tongue grooving was also noted to be better post- therapy both perceptually and in terms of the palatograms. The token shoe, was produced with a fronted [J] pre-therapy and with a backed /s/ post-therapy. In terms of the maximum contact phase, more alveolar and palatal medial contact was noted post-therapy. This finding is compatible with the perceptual results since the backed /s/ production would result in more contact in the post-alveolar and palatal regions. Groove width was also noted to be narrower post-therapy. Maximum Contact and Release Phases of Word-Final /£/ Tokens Both tokens were noted to have worse perceptual ratings post-therapy than pre-therapy. For posh, [j] was perceived pre-therapy but [(t)$:] post- therapy. Palatogram results supported the perceptual change. For example, post-therapy, more alveolar tip region contact was noted for the maximum contact phase. This finding explains why a stop was perceived post-therapy. The token push was also judged to be produced with the intended treatment target pre-therapy but was produced as [ts>] post-therapy. The palatogram results showed that change was minimal from pre- to post-therapy for both the maximum contact and release phases. Too much contact was noted for the alveolar tip region of the release phase. This finding provides support for the 208 perception of the stop post-therapy. In addition, contact was noted to be too forward for the alveolar and palatal medial and lateral regions for III. This explains why the backed Isl was perceived post-therapy. Dorsal contact, both medially and laterally was judged to be appropriate for both phases. Summary Similar to the treatment target /tj/, the stop closure measurement, the alveolar tip region contact measurement, and the alveolar and palatal region contact measurements, appeared to have the most affect on the perceptual evaluation of III. For example, when more alveolar tip region contact was observed, as was the case in push above, a stop was perceived. Furthermore, when more alveolar and palatal region contact was noted, especially medially, more fronted III productions or backed Isl productions were perceived. Other studies have reported that fricative production for persons with hearing impairments and motor speech disorders is often very poor in terms of place of articulation (Hardcastle, Gibbon, & Jones, 1991; Dagenais, 1992). For example, alveolar fricatives are sometimes produced as palatals and palatoalveolar fricatives are often fronted (Hardcastle, Gibbon, & Jones, 1991). Some of Samantha's III productions were fronted. In addition, it has been reported that "speakers with abnormal fricative production show an overall increase in the amount of tongue-palate contact" (1991, p. 65). This was true of 209 some of Samantha's productions, for example, when she added an alveolar stop in her [J ] production (e.g, [(t)$ ]. Generally, the only improvement for this treatment target was seen with the palatometry assessment. Perceptual results did not indicate improvement. For example, 16.6% of Samantha's III word list productions matched adult targets both pre- and post-therapy and only two out of five palatometry tokens showed improvement in terms of the perceptual measurement. This suggests that therapy was not successful in improving the perceptual results. Thus, the improved contact patterns as seen with the palatometer did not affect the perceptual ratings. Similar findings have been reported in the EPG literature (Gibbon, Dent, & Hardcastle, 1993; Scobbie, Gibbon, Hardcastle, & Fletcher, 1995). For example, Scobbie et al. (1995) report that with some children, speech sound contrasts exist covertly in their speech (i.e., are articulated) but are not heard perceptually (termed "covert contrasts"). Gibbon, Dent, and Hardcastle (1993) add that the presence of covert contrasts, also termed subphonemic contrasts, suggests that a speaker has a "productive phonological knowledge" of the adult category (1993, p. 248). Thus, prior to reaching the adult target, speakers may produce subphonemic contrasts, as was the case for Samantha's III productions. 210 Treatment Target Ig / (I) Perceptual Evaluation Each token, pre- and post-therapy, containing the treatment target Igl was judged to have been produced with [kh]. Thus, no improvement was noted from pre- to post-therapy; however, Samantha was producing the correct place of articulation (e.g, [dorsal]). (II) Instrumental Evaluation Maximum Contact and Release Phases for Word-Initial Igl Tokens Even though improvement was not noted perceptually for any token, improvement was observed for the pre- and post-palatograms. For example, all criteria for the token geese were judged to have improved from pre- to post- therapy for both the maximum contact and release phases. That is, each criteria was given a rating of change and improvement (+ +). More contact was noted during the maximum contact and release phases for the following criteria: symmetry, dorsal region medial contact (right and left sides), dorsal region lateral contact (right and left sides), and back row contact. For the token gob, change and improvement (+ +) was noted for 5/6 (83.3%) criteria for the maximum contact phase and 4/6 (66.6%) of the criteria were noted to have changed and shown improvement for the release phase. Both phases of the token goo generally showed less forward contact in the dorsal lateral region. This was 211 to be expected because Igl was followed by the back vowel IvJ and therefore less forward contact should occur. Thus, even though improvement was noted on each word-initial Igl palatogram, the improvement did not affect the perceptual judgments. Since the palatometer does not offer feedback in terms of voicing, these findings are not surprising. Maximum Contact and Release Phases for Word-Final Igl Tokens Generally, less contact was noted post-therapy for all word-final Igl tokens for both the maximum contact and release phases. More contact would have been desirable. For example, egg showed change from pre- to post-therapy but the change resulted in less contact for all criteria (e.g., less dorsal region medial contact, less dorsal region lateral contact, and less back row contact). Thus, symmetry was also worse due to the reduction in contact. Similar findings were found for the other two word-final Igl tokens, rog and pog. These results suggest that therapy was more successful when the treatment target Igl was in word- initial position (which was the targeted position during therapy) than in word- final position. Summary In terms of the overall results from the maximum contact phase palatograms for both the word-initial and word-final tokens, 18/36 criteria or 50% showed change and improvement from pre- to post-therapy (+ +). Generally, more positive (+ + and - +) than negative changes (- - and + -) were noted for this 212 phase. Similar results were found for the release phase; however, word-final Igl release phase palatograms showed more negative than positive changes. Therapy, however, focused only on word-initial Igl. Palatogram perceptual results did not indicate improvement but perceptual results from the word list transcription showed a 16.1% gain in match proportions with the adult target. This suggest that the pseudopalate may have negatively affected Samantha's Igl productions. She had more success with word-initial and final Igl when she was wearing her palatal lift than when she was wearing the pseudopalate. Since the palatal lift covers more of the dorsal region than the pseudopalate, velar articulations should be easier with the former device than with the latter. Generalization Target Ik I (I) Perceptual Evaluation Perceptually, each token containing fkl was judged to be correct both pre- and post-therapy. All fkl targets were produced as [kh]. (II) Instrumental Evaluation Maximum Contact and Release Phases of the Word-Initial and Word-Final fkl Tokens Interestingly, even though perceptually all tokens were judged to be correct, the results from the palatograms suggested that some of the results were 213 slightly worse post-therapy. For example, for the release phase of eek 4/6 or 66 .6% of the criteria were noted to not have changed but change was needed (- -). Similarly, dorsal medial region contact (right and left sides) and back row contact for the token cop was also given a rating of [- -]. More contact was needed in these regions for improvement. Improvement was noted for both tokens in terms of dorsal lateral region contact (right and left sides). Contact in this region was noted pre-therapy for both tokens but more was noted post-therapy. These results suggested that dorsal lateral region contact was important for the perception of [kh] and that the amount of medial contact in the dorsal region could vary without greatly affecting the perceptual results. Summary Since all of Samantha's tokens containing the treatment target /g/ were produced as [kh] the findings for the generalization target Ikl were not surprising. The phoneme fkl was well established. In terms of word-initial versus word-final Ikl, more improvement was noted for cop than for eek. This outcome was also seen with the Igl tokens. Thus, this finding provides further support for the suggestion that therapy was more successful for Igl in word-initial position than in word-final position, and this appeared to have generalized to the velar Ikl. Since therapy was only provided for word-initial Igl and not for Igl in word-final position, these results are as expected. 214 Generalization Target /z / (I) Perceptual Evaluation Each token containing Izl was judged to have improved from pre- to post- therapy. For zoo, the pre-therapy production of Izl was as [(t)s] and the post- therapy production was as the intended target, [z]. Pre-therapy, the Izl in bees was produced as [s:] but post-therapy the following was produced: [(t)z ]. Thus, since both post-therapy productions contained the intended target, improvement was noted. It was important to note however, that the post-therapy production of bees contained a stop prior to the [z]. (II) Instrumental Evaluation Maximum Contact and Release Phases of the Word-Initial and Word-Final Izl Tokens Generally more positive than negative changes were noted for both tokens for the maximum contact phase; however, for the release phase, more negative changes were noted than positive changes. For the token zoo, [z] was perceived post-therapy and [(t)s] pre-therapy. Analysis of the palatogram data for this token showed that less alveolar tip region contact was noted post-therapy and the groove width was wider post- therapy. These findings support the perceptual results, since less contact in the alveolar tip region would be needed for Izl than for Itl. Furthermore, with Itl there 215 is no groove width since stop closure is present. With Izl however, groove width must be present. As for the release phase, negative changes were noticed for the alveolar and palatal lateral contact regions because less contact was noted post- therapy than pre-therapy. The post-therapy production of bees, was [(t)z] and the pre-therapy production was [s:]. Since this change involves voicing the palatogram data did not provide information on this change. However, the trace of the stop [(t)] from Samantha's post-therapy production was confirmed by the palatogram data. The maximum contact phase of bees showed that the groove width was very narrow. Full closure however, was not noted. The lack of full closure supports the perceptual finding since only a trace of Itl was perceived. More negative changes in terms of the release phase were noted for this token than for zoo. For example, less alveolar and palatal medial contact was noted on the right and left sides and less left side dorsal lateral contact was noted post-therapy. These changes were judged to be negative since more contact was needed. Summary The findings for this generalization target were particularly interesting because the results for this target from the pre- and post-therapy word list transcriptions indicated that Izl was not perceived in any of the tokens containing this target pre- or post-therapy. The results from the palatometry pre- and post- assessments however, showed that Izl was perceived post-therapy for both bees and zoo. For example, zoo was produced as [z zu] post-therapy. This production 216 shows considerable improvement compared with Samantha's previous production for zoo above. In addition, the perceptual findings appeared to be compatible with the findings from the palatograms. This suggests that Samantha had more success with this target when producing it with the palatometer than without the palatometer. Perhaps the visually displayed tongue-palate contact points offered by the palatometer provided Samantha with appropriate feedback. Concomitantly, since the treatment target /d 3 / was practiced during therapy, both with and without the palatometer, Samantha may have generalized the voicing feature to Izl. These results differ from those found for the treatment target Igl. With Igl, it appeared as though Samantha had more success when producing Igl without the palatometer than with the palatometer (see above for discussion). Similar findings were reported in the study by Parsloe (1998). Pre-therapy palatometry results showed that 90% of Sophie's word-initial Idl productions had the correct place of articulation (e.g., [coronal] place, as does Izl) but 50% of her productions contained incorrect voicing. Immediate and three months post- palatometry intervention results showed that all of Sophie's Idl targets were produced with the correct voicing and place of articulation. Thus, for both Sophie and Samantha improvement was noted after palatometry therapy for [coronal] place of articulation in combination with voicing. 217 Generalization Target Is / (I) Perceptual Evaluation Both tokens containing Isl were produced as [s] pre- and post-therapy. The only significant difference between the two tokens, saw and boss, was that the latter token was produced as [(t)s>] pre-therapy and as [(t)s] post-therapy. This suggests that Samantha may have had difficulty with the word-final vowel- consonant sequence, [as]. Perhaps the addition of the [t] allowed Samantha to find the correct place of articulation for the [s] before producing excess airflow for the fricative (i.e., a coarticulatory effect). (II) Instrumental Evaluation Maximum Contact and Release Phases of the Word-Initial and Word-Final Isl Tokens For both tokens, generally more positive changes (+ + and - +) were noted than negative changes (- - and + -). For boss. [(t)s>] was perceived pre-therapy and [(t)s] post-therapy. Analysis of the data from the palatograms showed that more alveolar region contact was noted post-therapy than pre-therapy. This finding supports the perceptual results because post-therapy, less of a backed Isl was perceived; thus, more contact would have been needed in the alveolar contact region. For the intrusive stop [(t)], the maximum contact phase of both the pre- and the post- 218 palatograms showed almost complete closure along the alveolar tip region. Full closure was not noted but since only a trace of a [(t)] was perceived full closure was not needed. Thus, the results from the palatograms supported the perceptual results. Improvement was noticed for both phases, maximum contact and release phases, in the dorsal region. Samantha had appropriate contact both medially and laterally in the dorsal region pre- and post-therapy. The token saw was produced pre- and post-therapy with an [s]. Few changes were noted in terms of the palatograms. The only significant changes included less alveolar and palatal lateral and medial contact during the maximum contact phase and more alveolar and palatal lateral contact on the right side during the release phase. Summary The results from both boss and saw suggest that the groove width criterion and the alveolar tip region contact criterion are the two most significant criteria for producing change in terms of perception. Changes were noted for the token saw for the alveolar and palatal lateral and medial contact regions; however, these changes did not result in changes to the perceptual evaluation of this token. It is important to note however, that the changes that were made were not drastic; only slightly less or slightly more contact was noted (e.g, one to five contact points more). Similar findings were found for the treatment targets /$/, Ity, and/d3/. 219 The perceptual and the palatogram results reflect the results from the pre- and post-therapy word list transcriptions. For example, as mentioned above, there were some post-therapy word list tokens that contained a [t] immediately prior to an [s] (e.g, [(t)s] as in boss above). Summary & Significance of Findings to the Body Level (ICIDH-2) Narrow transcriptions of pre- and post-assessment word lists indicated a general reduction in impairment to a slightly less severe level. The affricate treatment targets and the velar treatment target Igl, each showed improvement from pre- to post-therapy in terms of place of articulation, voicing, and manner. The changes in velar and affricate production suggest an effect of therapy. The palatoalveolar.fricative did not show improvement in match proportions with the adult target; however, tongue grooving and placement appeared to have improved slightly. As for the generalization targets, both Isl and fkl showed improvement in terms of place and manner, and Izl showed improvement in terms of voicing. Thus, based on these results some generalization occurred between Igl and fkl, between the affricates and the alveolars. In terms of the transcriptions of the treatment targets during therapy, the palatoalveolar fricative and the affricates showed improvement from pre- to post-therapy. Generally more gains were noted after palatometry therapy for It J/ 220 and /d3/ than after non-palatometry therapy. This suggests that palatometry therapy resulted in improvement for both of the affricate treatment targets, thereby indicating a reduction at the level of the body. For /$/ more gains were noted after non-palatometry therapy. However, non-palatometry therapy followed palatometry therapy during the second block; thus, any gains seen with non-palatometry may have been due to the visual feedback information offered by the palatometer. The treatment target Igl showed no gains with palatometry nor non-palatometry therapy. This latter finding is interesting since gains were noted post-therapy for this target on the word list assessment done one week later. The gains may have occurred because Samantha was wearing her palatal lift which was noted to make velar articulations easier than when she was wearing the pseudopalate, and because she may have had more time to practice her therapy sounds at home, before the post-therapy assessment session. Pre- and post-transcriptions of the palatometry assessment tokens indicated that improvement was perceived for both the generalization and treatment targets. Perceptual results for each target were compatible with the production results as based on the information from the palatograms. More positive changes than negative changes were noted for the stop and fricative phases of the affricates. Similarly, more positive than negative changes were noted for the maximum contact and release phases for III. Slightly more positive than negative changes were noted for the maximum contact and release phases 221 for word-initial Igl. Gains were also noted for each of the generalization targets. Where regressions occurred, (e.g., + -), generally too little or too much contact was observed. As a result, perceptual ratings were also noted to be worse post- therapy. In terms of the criteria used for the palatometry assessment, the stop closure contact, the alveolar tip region contact, groove width, and the alveolar and palatal contact regions, appeared to have the most affect on the perceptual evaluations for the affricates and the palatoalveolar fricative. For Isl and Izl, the groove width criterion and the alveolar tip region contact criterion had the most affect. According to Fletcher (1989), the groove width criterion is "likely the primary articulatory parameter in sibilant production" (1989, p. 747). For Igl and fkl, the dorsal lateral region criterion appeared to be most important for the perceptual evaluation. Another interesting finding, was the presence of a covert contrast for /J"/. Improvement was observed with the palatometer but was not heard perceptually. As reported in the literature, covert contrasts suggest that the speaker is in a period of "active phonological discovery" (i.e., finding out how to produce the target sound within the constraints of the production mechanism) (Gibbon, 1990; Gibbon, et al., 1993; Scobbie, et al., 1995). Previous EPG studies have used more intensive short-term treatment sessions (i.e., daily therapy for three to four weeks) as opposed to less intense and more long-term sessions as used in this study (i.e., one hour per week for 222 eight weeks) (e.g., as reported in Dagenais, 1992; Dagenais, Critz-Crosby, Fletcher, & McCutcheon, 1994; Fletcher, Dagenais, Critz-Crosby, 1991). Perhaps with more intense short-term therapy, results from this study would show even greater improvement. In summary, the results from the palatograms suggest that therapy had a positive effect at the level of the body. ACTIVITY LEVEL Five untrained listeners completed three intelligibility measures: (1) an open-set word identification task, (2) an open-set sentence identification task, and (3) a closed-set word identification task. Goal phoneme identification was calculated within words and word identification was calculated within sentences. The following section compares the results of the open-set identification tasks and the closed-set identification tasks. In addition, the activity level results from this study will be discussed in relation to previous research findings such as those found in Williams (1998). Williams (1998) chose different therapy targets and sentences for Samantha than those reported in this study. Therefore, comparisons between Williams' (1998) study and this study can only be made in terms of the open-set word identification task. 223 Open-Set Word Identification and Closed-Set Word Identification Previous studies have suggested that closed-set identification tasks provide more meaningful intelligibility information for speakers with more severe hearing impairment compared with open set identification tasks (e.g., Osberger, 1992; Boothroyd, 1985; Yorkston & Beukelman, 1981b). Closed-set tasks provide more context, making identification easier for the listener. A comparison of the difference between the percentage of words correctly identified pre- and post-therapy in the open-set and closed-set word identification tasks (Tables 13 and 21 respectively) indicates that listeners found the closed-set word identification task easier. The average difference between the open-set and the closed-set in word identification for all listeners was 2% and 34% respectively. Furthermore, all listeners showed improvement from pre- to post-therapy with the closed-set word identification task. This was not the case with the open-set word identification task. One of the five listeners correctly identified fewer words post-therapy compared to pre-therapy with the open-set word identification task. Another listener correctly identified the same number of words pre- and post- therapy. The results from these two listeners highlights the difficulty each listener had in correctly identifying words in the open-set identification task. Some listeners found the pre- and the post-therapy words to be equally difficult to identify. Furthermore, the qualitative comments made by the listeners indicated that both the pre- and the post-therapy words were difficult to 224 understand. However, listener variables such as fatigue and level of linguistic competence of the listener should be considered when analyzing such results. These variables were discussed in more detail in the Introduction section of this thesis. Generally however, the better results from the closed-set word identification task suggest that the listeners benefited from the contextual information within the closed-set identification task as other researchers have suggested (Boothroyd, 1995). A comparison of the open-set task results from this study with Samantha's open-set task results in the study by Williams (1998), shows that the same five listeners had an average difference in word identification of 2%- the same result as reported in this study. It should be noted however, that different words were used in the present study than were used by Williams (1998). The words in the present study contained the phonemes which Samantha had the most difficulty producing. In Williams' (1998) study, the words contained phonemes which were not as challenging for Samantha (e.g., /s/, lxl, and fkl). Williams (1998) did not include a closed-set word identification task; thus, no comparison can be made. Goal Phoneme Identification in Open- and Closed-Set Word Identification Tasks Overall, goal phoneme identification for the closed-set word identification task from pre- to post-therapy was better than for the open-set word 225 identification task. The average difference between the percentage of treatment targets correctly identified pre- and post-therapy in the closed-set word identification task was about 35%. In the open-set word task, the average difference was 9%. All treatment targets with the closed-set task improved from pre- to post-therapy. The only exception to this was the treatment target III which remained at the same average (19.8%) pre- and post-therapy. These results are compatible with the results from the body level evaluation, since pre- and post-therapy results for III at the body level did not change. For the open-set word identification task however, the average percent of treatment phoneme identification improved for Id^l only. The other targets, Igl, Itl I, and III, showed a decrease in phoneme identification from pre- to post-therapy in terms of the average percent correct. These findings suggest that improvements in the open- set word identification task from pre- to post-therapy were partially due to the improvement in the Ids/ phoneme from pre- to post-therapy. Indeed, two of the five listeners correctly identified the word jump post-therapy; this word contains the /d 3 / phoneme. These same two listeners did not identify this word or any other word containing this phoneme pre-therapy. Thus, this suggests that therapy was successful for /d3/ phoneme identification by untrained listeners. Results for the III treatment target in the open-set identification task indicated that identification was worse post-therapy compared with pre-therapy. The 226 average percent across listeners pre-therapy was 13.2% and post-therapy it was 0%. For the closed-set task, the average percent of III identification across listeners was 19.8% both pre- and post-therapy. Since the results did not decline post-therapy as they had in the open-set task, this indicates that the listeners needed the contextual information within the closed-set task in order to identify correctly the /$/ target. Generally, these results suggest that therapy was not successful for this phoneme. Likewise, results from the body level discussion suggested that very few gains were observed for this treatment target. The only notable gains were with the lingual-palate contact patterns observed on the pre- and post-palatograms. Identification of the treatment target I til was worse post-therapy with the open-set task, but significantly better post-therapy with the closed-set task. In the former task, the average percent across listeners pre- and post-therapy was 46.6% and 13.2% respectively. In the latter task, the average percent across listeners pre- and post-therapy was 33.3% and 93.3% respectively. Thus, the listeners benefited from the context within the closed-set task. In the open-set task, only two of the listeners correctly identified words containing the treatment target Itl I. For example, one listener correctly identified the word chicken pre- therapy, while another listener identified the word itchy post-therapy. However, I til phoneme identification in the open-set task was generally better for all 227 listeners than word identification. One listener, for example, wrote "clinching" pre-therapy for the target word, itchy. Open-set task phoneme identification for the generalization targets Is/ and Izl showed improvement from pre- to post-therapy. The average percent across listeners for the Isl phoneme was 80.2% pre-therapy and 86.8% post-therapy. Significant improvement was noted for the Izl phoneme. The average percent across listeners was 13.2% pre-therapy but 53.4% post-therapy. For the generalization target fkl, more listeners correctly identified this target pre- therapy than post-therapy. The average percent across listeners pre-therapy was 100% but post-therapy it was 62.2%. The reduction in fkl phoneme identification can be explained by the fact that three of the five listeners recorded "pokey" for the token calling. Perhaps Samantha aspirated prior to the fkl in calling but the listeners perceived the aspiration as "po." In terms of the closed-set task, phoneme identification of each generalization target improved from pre- to post-therapy. This suggests that the contextual information provided by the closed-set task was useful for correct phoneme identification of the generalization targets. Word and Sentence Identification Tasks A comparison of the pre-therapy identification of single words in the open- set task (Table 13), of the pre-therapy identification of words in the closed-set 228 task (Table 21), and of the pre-therapy identification of words in sentences (Table 15a) shows that intelligibility scores for the sentence identification task tend to be higher than intelligibility scores for word identification tasks. Yorkston and Beukelman (1981b) determined that sentence identification tasks usually yield higher intelligibility scores for word identification than do word identification tasks. In terms of post-therapy results, the sentence identification task results were higher than the open-set word identification task results but were not higher than the closed-set word identification task results. As discussed above, the closed-set word task results showed significant improvement over the open-set word task results. The context provided within the closed-set task was essential for word identification by the listeners. Furthermore, unlike the sentence identification task, the closed-set task required listeners to listen to only one word. For the sentence identification task however, listeners had to listen to either eight, nine, or twelve words at once, depending on the sentence. Thus, even though more context was available with the sentences, the demands on the listeners were slightly greater. Sitler, Schiavetti, and Metz (1983) noticed that for speakers with significantly unintelligible speech, contextual cues were no longer helpful in identification. For Samantha, post- therapy sentence identification results were worse than pre-therapy sentence identification results. The average difference between the pre- and the post- therapy results for all listeners was -13.86%. All of the listeners thus found the post-therapy sentences to be more difficult to identify than the pre-therapy 229 sentences. Listeners' understanding ratings showed that the post-therapy sentences were in fact more difficult to understand. That is, more listeners rated the pre-therapy sentences as being easier to understand compared with the post- therapy sentences. For sentence A (e.g., The rainbow is a division of white light into many beautiful colours), the only words that were not correctly identified pre- or post- therapy, by any of the listeners, were division and white. In addition, the word into was only identified by one of the listeners. Since Samantha had never received therapy on the phonemes contained within the words division and white, the lack of identification by the listeners for these words was not surprising. Two of the five listeners had better word identification for sentence A post-therapy; however, one of the listeners had significantly worse results for this sentence post-therapy. As a result, the average score for all listeners for sentence A declined post-therapy. This latter listener was the same listener who had worse results on the post-therapy open-set word identification task. Since this listener performed significantly better on the pre-therapy results, it was concluded that listener variables, specifically fatigue, affected this listener's results. Sentence B identification results were better post-therapy for two of the five listeners. None of the listeners correctly identified the words finds or ever. Since these words contain phonemes for which Samantha had never received therapy, the lack of identification was not surprising. In addition, the post- therapy recording of this sentence was noticed to contain many self-corrections by 230 Samantha. For example, the word look was said once as lo and then was followed by look. Perhaps the added self-corrections decreased the amount of context available within the sentence, thereby making word identification far more difficult for the listeners. Many self-corrections were also noted in Samantha's post-therapy production of sentence C. For example, the words look, something, and reach were each repeated twice. These self-corrections perhaps made word identification more complicated for the listeners. In fact, each listener identified fewer words post-therapy compared with pre-therapy. The remaining Rainbow Passage sentences that were recorded post- therapy by Samantha each contained many more self-corrections than the three sentences chosen for the listener tasks (sentences A, B, and C). Thus, Samantha's post-therapy recording was, in general, very poor compared with her pre-therapy recording. In retrospect, it may have been better to have recorded a conversation between Samantha and the experimenter rather than to have had Samantha read sentences. While Samantha was reading the passage post- therapy, she appeared to be very stressed and nervous. She took a longer time reading the passage post-therapy compared with pre-therapy, hence the many self-corrections. According to Yorkston and Beukelman (1981b), when speakers with dysarthria produce longer utterances, often fatigue and muscle weakness sets in which negatively affects articulation. Since Samantha appeared more stressed, her oral and facial muscles were probably more tense and therefore her 231 -•v. articulation may have been somewhat distorted compared to normal. Single word reading appeared to be easier for Samantha because of less opportunity for fatigue. In summary, the results from the sentence and word identification tasks suggest that the word identification task results may be a better measure of intelligibility for this study. Since the sentence identification recordings were not reflective of Samantha's best speech due to probable stress factors, the results for this task may not reflect Samantha's true abilities. PARTICIPATION LEVEL The World Health Organization's concept of participation is "an individual's involvement in life situations in relation to health conditions, body functions or structures, activities, and contextual factors." Participation restrictions are "problems an individual may have in the manner or extent of involvement in life situations" (ICIDH-2 Beta-2, p. 83). Responses from questionnaires administered to Samantha and listeners' perceptions were used to assess the effects of palatometry therapy on level of participation (WHO, ICIDH-2, 1999). This section discusses the responses from the questionnaires and compares the findings to previous research. 232 Level of Participation: Samantha's Responses Samantha's responses from two questionnaires, each addressing level of participation, were collected and analyzed. The first questionnaire was a Palatometry Research Project questionnaire and the second questionnaire included questions selected from The Communication Profile for the Hearing Impaired (CPHI, Demorest & Erdman, 1986, 1987) and from Erickson's S-24 Scale (S-24 Scale, Erickson, 1969). Responses from the first questionnaire were compared with responses from the same questionnaire administered to Samantha in a previous study (Bernhardt, Muir & Loyst, 1996). Samantha's responses from the previous Palatometry Research Project questionnaires were general and vague. As discussed in the Results Chapter, Samantha did not provide much detail in her answers nor did she show much awareness of changes resulting from palatometry therapy. A comparison of her earlier responses with her present responses, however, shows improved awareness of palatometry therapy goals. For example, instead of reporting, like she had previously, that she uses her new speech sounds such as Irl, Isl, fkl, and 111, to improve her speech, Samantha added more detail to her present answer. She stated that in addition to using her new speech sounds Igl, It^l, /d 3/, and / J/, she also produces Igl with "voice," and /d 3 / with "a little bit of voice and air." As well, in her previous answers, she reported being able to use her new speech 233 sounds only when practising and occasionally in conversation. This time, however, Samantha stated that she uses her new sounds all the time. Samantha's responses to the participation situation questions were interesting because they revealed a distinction between her self-confidence and self-awareness as a speaker with a motor speech disorder, and as a person with a hearing impairment. For example, when asked if she avoids going to church, Samantha responded with "a little bit" because she can "hide her hearing aids behind her ears and cover them with her hair." This latter response shows that she might be self-conscious about her hearing loss, specifically about wearing hearing aids. According to Maxon and Brackett (1992, p. 133-134), wearing amplification such as hearing aids is one of the most common "problem areas" school-aged children and adolescents experience. When asked if she avoided going to a movie because of her hearing loss, Samantha reported "not at all." Further analysis of Samantha's responses from questions focusing on her speech supported the above conclusion;-i.e., that Samantha may be more self-confident with her speech but less self-aware of her speaking abilities compared to her hearing loss. For example, Samantha reported that her speech was difficult for both familiar and unfamiliar speakers to understand only "a little bit." In addition, she reported "a little bit" when asked if she felt her speech was slow, fast, too loud, too soft, and if her voice sounded harsh or hoarse. Her responses show a lack of self-awareness with regard to her speaking abilities since results from the body level and activity level evaluations showed that Samantha 234 presents with a severe speech sound production disorder. Furthermore, she reported that her speaking abilities "never" become poorer when she fatigues. This response is interesting because, as discussed previously, results from the open-set sentence identification task indicated that she appeared to have more difficulty articulating sentences than single words. Yorkston and Beukelman reported that speakers with dysarthria often fatigue and have muscle weakness when producing longer utterances (1981b). In summary, Samantha's responses from the Palatometry Research Project questionnaires indicated improved awareness of the goals targeted during palatometry therapy. In addition, responses to the participation situation questions and questions focusing on her speech, revealed a difference between her self-awareness of her speaking abilities and self-awareness of her hearing abilities. Her responses with regard to her speech showed better self-confidence but a lack of awareness as to the severity of her speech disorder. Responses with regard to her hearing, on the other hand, showed a greater awareness of her hearing impairment. Samantha's responses to questions focusing on her feelings and attitudes about her hearing loss on the CPHI revealed that her participation within society is restricted some of the time because of her hearing loss. In addition, she reported not enjoying going to places with friends because she has trouble hearing, and that she almost always tries to give the impression of having normal hearing. According to Samantha, when someone does not understand her 235 she sometimes feels discouraged. Similar findings were reported by Maxon and Brackett (1992), who suggest that difficult listening conditions may lead to communication breakdowns which, in turn, may cause the hearing-impaired person to feel discouraged and remove him/herself from the communication situation. Responses from Erickson's S-24 Scale also revealed that Samantha's participation within society is restricted some of the time because of her difficulties producing intelligible speech. For example, according to Samantha, sometimes she fails to provide an answer in a group situation because she is afraid to speak out. Additionally, she reported that she sometimes would rather avoid introducing herself to a stranger. Furthermore, her lack of awareness of her speaking abilities was evident when she reported that her speaking voice is sometimes easy to listen to. A person with a severe speech sound production disorder is usually very difficult to understand and hence challenging to converse with. Summary Responses from the CPHI and from Erickson's S-24 Scale supported the conclusion that differences exist between Samantha's self-awareness of her speaking abilities and hearing abilities. According to her responses, Samantha demonstrates greater self-confidence but poorer self-awareness as to the severity of her speech disorder, and less self-confidence but greater awareness of her hearing impairment. Furthermore, her responses reveal that her social and 236 personal life are sometimes restricted because of both her hearing impairment and her speech sound production disorder. Level of Participation: Listeners' Perceptions In order to obtain an idea of some of the participation restrictions Samantha may encounter, each listener was asked to complete a questionnaire developed by the experimenter and her supervisor. Participation restrictions are "problems an individual may have in the manner or extent of involvement in life situations" (ICIDH-2 Beta-2, p. 83). Examination of the listeners' responses (all approximate age peers) shows that Samantha's involvement in life situations may be restricted. For example, three out of the five listeners responded that they would feel comfortable having a conversation with Samantha at a quiet place as opposed to a loud place. One of the listeners however, reported not feeling comfortable conversing with Samantha at any place s/he could think of and another listener did not provide a response. When asked how often one would like to talk on the telephone with Samantha, two listeners responded "not at all," one responded "sometimes," one "hardly ever," and the fifth listener did not provide a response. Clearly, Samantha would encounter a problem if she wished to talk with one of these listeners over the telephone. Participation restrictions were also reported by the listeners for the following situations: going shopping in a shopping mall, going to 237 the movies, and participating in a work project with Samantha. For each of these situations, listeners generally responded with "hardly ever" or "sometimes." Interestingly, situations that did not involve oral conversing with Samantha were rated by the listeners as being the least restrictive in terms of participation. For example, two listeners responded that s/he would e-mail Samantha "all the time" and one listener reported that s/he would e-mail "sometimes." Similar responses were given when asked how often one would like to participate in a letter exchange (i.e., write letters) with Samantha and participate in a team athletic event such as soccer. One of the listeners did not provide a response to 12 of the 14 questions. This listener wrote at the end of his questionnaire that s/he did not feel comfortable answering some of the questions because s/he did not know the speaker. In summary, responses from listeners suggest that Samantha would encounter more participation restrictions with situations involving speech than with situations involving non-oral modes of communication. APPLICATION AND SIGNIFICANCE OF FINDINGS TO THE WORLD HEALTH ORGANIZATION MODEL (1999) In order to obtain a more complete evaluation of outcomes, the effects of palatometry therapy were investigated beyond the body (impairment) level. 238 Activity and participation level information provided more functional outcome measures and they allowed the relationships between the impairments (i.e., level of body), the daily life activities, and the sociocultural environment of Samantha to be examined. At the level of the body, Samantha made improvements from the pre- to the post-therapy assessments and with the therapy stimuli. Body level measures focused exclusively on speech as a function of the oral motor structures. Results from palatograms showed that Samantha improved all her therapy target articulations {Igl, HI, /tj/ and /d3/) which were judged to be closer to "normal." In addition, improvement was noted for the generalization targets, Isl, Izl, and fkl. Perceptual results for each therapy target and generalization target were compatible with the palatogram results. Regressions were outweighed by improvement overall. These results, in conjunction with the pre- and post- assessment word list results, indicated a reduction in impairment to a slightly less severe level. Furthermore, they suggest that after therapy, Samantha's oral structures were better able to produce the targeted speech sounds. In order to investigate whether the improvements generalized to Samantha's day to day activities and how her speech disorder interacted with contextual factors such as the listener and the physical environment, untrained listeners completed several intelligibility measures. 239 Intelligibility measures were used to assess how well Samantha was understood by a listener. Such measures focus more on the level of communication rather than speech. If Samantha is better understood post- therapy compared with pre-therapy, then activity limitations ought to be decreased. Overall, results from the five untrained listeners indicated improvement from pre- to post-therapy in all measures except for the sentence identification task. This latter measure was not reflective of Samantha's best speech and therefore was not a reliable measure. The improvement, however, suggests that untrained listeners were better able to understand Samantha's speech post-therapy and further implies that therapy had a positive effect both at the level of the body, and at the level of activity. Thus, activity limitation was reduced post-therapy. Lastly, it is important to note that intelligibility was measured in the most stringent context. Visual cues and topic cues were not available to the listeners. Such cues, however, would be available in everyday conversation. Since results from the closed-set word identification task were superior to those from the open-set word identification task in which no context was provided, it was evident that the addition of contextual cues benefited the listeners' understanding. Therapy should also include strategies for maximizing contextual supports for listeners. Participation level measures were used to assess the extent to which Samantha viewed her own participation restrictions and to obtain an idea of some of the participation restrictions Samantha may encounter according to 240 listeners. If improvement is evident at the conversational level (i.e., the activity level), then perhaps a person's level of participation within society may increase; hence, participation restrictions will be decreased. Self-reports from Samantha indicated improved self-awareness of the palatometry therapy goals from a previous palatometry study to the end of the present study. Additional answers from questionnaires suggested that she demonstrates poor self-awareness as to the severity of her speech disorder but less self-confidence and greater awareness of her hearing impairment. Furthermore, as reported by Samantha, her social and personal life are sometimes restricted because of her speech impairment, but are very much restricted because of her hearing impairment. Responses from listeners also suggested that Samantha would encounter more participation restrictions with situations that involved speech as the primary mode of communication (i.e., going for coffee or class presentation) than with situations involving primarily non-oral modes of communication (i.e., tennis match or soccer game). Thus, even though improvement was evident at the activity level, responses from Samantha and from the listeners suggest that participation restrictions exist; however, according to Samantha, she feels more confident with her speech and with making herself understood. Hence, Samantha's activity limitations should decrease and participation within society should be facilitated, if she continues to receive treatment and improve at the body level. In conclusion, improvement was noted for Samantha's speech production abilities at the oral motor level, at the level of communication, and in terms of 241 Samantha's self-confidence as a speaker. Increased intelligibility indicated that listeners not familiar with disordered speech were able to understand more of Samantha's articulations and the information she was transmitting through her speech. An increase in the amount of information conveyed fosters a more successful communicative interaction between Samantha and her listeners. It is important to note that the listeners were engaged in tasks that demanded full attention and cooperation which might not be forthcoming in the real world. Lastly, examining all levels of functioning provides better insight as to the relationship between the levels and the effect therapy had at each level. LIMITATIONS OF THIS STUDY AND DIRECTIONS FOR FUTURE RESEARCH The following section discusses limitations of this study and provides directions for future research. The present study investigated the effects of palatometry therapy with only one speaker. If more speakers were used for this study, then speaker variables such as severity, disorder, and age could be analyzed. Furthermore, by increasing the number of speakers, more information could be obtained in terms of the frequency and types of activity and participation restrictions. The restrictions discussed in this study are only significant to the speaker who participated in the present study. 2 4 2 Secondly, untrained listeners as opposed to listeners familiar with disordered speech were asked to participate in this study. While responses from untrained listeners tend to be more representative of listeners in general, responses from trained listeners may provide more detailed information. Ideally, results from both untrained and trained listeners should be compared. It would also be useful to use listeners who have social roles with the speaker (e.g., family, teacher, or friend). For this study, software was not available for quantifying pre- and post- therapy palatograms. As a result, the experimenter and her supervisor created criteria that represented each contact region on the pseudopalate. Analysis of the palatogram data was time consuming; thus, there may be more efficient and more systematic methods to analyze palatograms. With less time required for analysis, more tokens could be analyzed perceptually and instrumentally, increasing reliability of results. The fourth limitation concerns the selection of sentence level material. For this study, utterances from the Rainbow Passage were recorded and selected. During the post-therapy recording, it was noted by the clinician that Samantha appeared very stressed and fatigued, and that her recording contained many self- corrections. It may have been more ideal to record a conversation rather than have Samantha read a passage containing long utterances. During a conversation, Samantha would be able to control the length of her utterances and therefore would fatigue less. 243 The open-set word identification task was presented prior to the closed-set word identification task. Even though some of the listeners claimed that the closed-set task words were "new" words, a learning effect may have occurred from the open-set to the closed-set task. Finally, the participation level results in this study were gathered at the end of Samantha's therapy and during each listener's second visit. It may be more beneficial to collect the speaker's and the listeners' perceptions both pre- and post-therapy. This way responses could be compared over time and the effects of therapy on level of participation could be better analyzed. Some EPG studies have measured both the production and perception abilities of hearing-impaired speakers. This current study focused only on the former measure; however, the relationship between perception and EPG production training for individuals with hearing impairments is an interesting and presumably worthwhile area for future research. Finally, the qualitative results from Samantha suggest that her participation restrictions continue to decrease as her speech continues to improve. A follow-up questionnaire or some other qualitative measure administered to Samantha six months post-therapy may provide further insight as to her perception of her speech and hearing impairment, her participation restrictions, and her needs and aspirations. Responses from this follow-up questionnaire could then be compared with her responses from this study to determine if participation restrictions have decreased or remain the same. 244 CONCLUSION Some of the key findings for each of the three levels of functioning are summarized below. Body Level: Narrow transcriptions of pre- and post-assessment word lists indicated a reduction in impairment to a slightly less severe level. In general, more gains were noted after palatometry therapy than after non-palatometry therapy; pre- and post-transcriptions of palatometry assessment tokens indicated that improvement was perceived for both the generalization and treatment targets; perceptual results for each target were compatible with the production results as based on the information from the palatograms; and gross results from the palatograms indicated that more positive changes (+ + and - +) than negative changes (- - and + -) were noted for all target phonemes from pre- to post-therapy. The phonemes /d3/, /t$/, and Igl had better success than III. Activity Level: Word and phoneme identification with the closed-set task was better than with the open-set word identification task. The context provided in the closed-set task facilitated the listeners' responses. Phoneme identification for the palatoalveolar fricative, however, did not show improvement from pre- to post-therapy and this result was compatible with the results from the body level 245 evaluation. Word identification results were a better measure of intelligibility than were sentence identification results because of production demands. Participation Level: Responses from questionnaires indicated that Samantha demonstrates improved awareness of the goals targeted during palatometry therapy, less self-awareness as to the severity of her speech disorder, and less self-confidence but greater awareness of her hearing impairment. Additionally, responses from listeners suggested that Samantha would encounter participation restrictions particularly with situations involving speech. This study found that intervention using the palatometer had positive effects for a twenty-two year old woman with a bilateral, severe-to-profound hearing loss, a severe speech sound production disorder, and oral musculature weakness. It was determined that palatometry therapy not only resulted in gains at the level of the body, but also at the levels of activity and participation. Qualitative data from untrained listeners suggested that Samantha may continue to experience participation restrictions without further intervention programs. As researchers and clinicians it is of upmost importance that we consider levels of functioning beyond the body (impairment) level. Phonological therapy must focus on making changes that are relevant to the client as a speaker, as a communicator, and as a participant within society. Intelligibility measures and self-report questionnaires provide ways of measuring activity and participation. Responses from questionnaires and results from intelligibility measures can help guide therapy goals. In summary, addressing all levels of functioning is particularly useful for the clinician and the client in order to derive a more functional and purposeful therapy program and to allow for client-focused outcome measures. Lastly, in order for participation and activity restrictions to decrease therapy should also focus on training listeners on how to facilitate interactions between the speaker and the listener and on creating more awareness of deviant speech and hearing. REFERENCES 247 Alpiner, J. and McCarthy, P. (1987). Rehabilitative Audiologv: Children and Adults. Baltimore, Maryland: Williams & Wilkins. Andrews, G. & Culter, J. (1974). Stuttering therapy: the relation between changes in symptom level and attitudes. Journal of Speech and Hearing Disorders. XXXIX. 3. 312-319. Barry, R. M., (1993). Measuring segmental timing in pathological speech using electropalatography. Clinical Linguistics and Phonetics. 7. 4. 275-283. Bernhardt, B. (1990). Application of nonlinear phonological theory to intervention with six phonologically disorder children. Unpublished doctoral dissertation, University of British Columbia, Canada. Bernhardt, B., Muir, S., & Loyst, D.. Palatometry pilot project: Outcomes for a variety of speech disorders. Presented in Seattle. ASHA Abstracts. November, 1996, p. 72. 248 Bernhardt, B., Pichora-Fuller, K., and Williams, R. (1999). Enhancing speech production outcomes of cochlear implantation with palatometry. Unpublished manuscript. Boothroyd, A. (1985). Evaluation of speech production of the hearing impaired: Some benefits of forced-choice testing. Journal of Speech and Hearing Research. 28. 185-196. Byrd, D. (1994). Palatogram reading as a phonetic skill: a short tutorial. Journal of the International Phonetic Association. 24. 1. 21-35. Carter, C. R., Yorkston, K. M., Strand, E. A., & Hammen, V. (1996). The effects of semantic and syntactic content on the actual and estimated sentence intelligibility of dysarthric speakers. In Garcia, J. M. & Dagenais, P. (1998). Dysarthric sentence intelligibility: contribution of iconic gestures and message predictiveness. Journal of Speech. Language, and Hearing Research. 41, 1282-1293. Cowie, R. and Douglas-Cowie, E. (1992). Postlingually Acquired Deafness: Speech Deterioration and the Wider Consequences. Berlin, Germany: Walter de Gruyter & Co. 249 Creaghead, N., P. Newman, & Secord, W. (1989). Assessment and Remediation of Articulatory and Phonological Disorders. Second Edition. New York, New York: Macmillan Publishing Company. Dagenais, PA. (1992). Speech training with glossometry and palatometry for profoundly hearing-impaired children. The Volta Review. 94. 261-282. Dagenais, P. A. (1995). Electropalatography in the treatment of articulation/phonological disorders. Journal of Communication Disorders. 28. 303-329. Dagenais, P. A. (date unknown). Clinical Applications of the Palatometer: Prepared for Kay Elemetrics Corporation. University of South Alabama. Dagenais, P. A. & Critz-Crosby, P. (1991). Consonant lingual-palatal contacts produced by normal-hearing and hearing impaired children. Journal of Speech and Hearing Research. 34. 1423-1435. Dagenais, P. A., Critz-Crosby, P., Flectcher, S., and McCutcheon, M. (1994). Comparing abilities of children with profound hearing impairments to learn consonants using electropalatography or traditional aural-oral techniques. Journal of Speech and Hearing Research. 37. 687-699. 250 Davis, H. (1947). Acoustics and psychoacoustics. In H. Davis and S.R. Silverman (Eds)., Hearing and Deafness. HY: Holt, Rinehart and Winston. Demorest, M. E., & Erdman, S. A. (1986). Scale composition and item analysis of the Communication Profile for the Hearing Impaired. Journal of Speech and Hearing Research. 29. 515-535. Demorest, M. E., & Erdman, S. A. (1987). Development of the Communication Profile for the Hearing Impaired. Journal of Speech and Hearing Disorders. 52, 129-143. Dent, H., Gibbon, F., & Hardcastle, W. (1995). The application of electropalatography (EPG) to the remediation of speech disorders in school- aged children and young adults. European Journal of Disorders of Communication. 30. 264-277. Dworkin, J. & Culatta, R. (1996). Oral Mechanism Examination and Treatment System. Nicholasville, Kentucky: Edgewood Press, INC. Erdman, S.A. (1994). Self-assessment: From research focus to research tool. In Metz, D.E., Caccamise, F. & Gustafson, M. (1997). Criterion validity of the 251 language background questionnaire: a self-assessment instrument. Journal of Communication Disorders. 30. 23-32. Erikson, R. L. (1969). Assessing communication attitudes among stutterers. Journal of Speech and Hearing Research. 12. 711-724 in Andrews, G. & Culter J. (1974). Stuttering therapy: The relation between changes in symptom level and attitudes. Journal of Speech and Hearing Disorders. 39. 312-319. Esposito, S., Mitsumoto, H., & Shanks, M. (2000). Use of palatal lift and palatal augmentation prostheses to improve dysarthria in patients with amytrophic lateral sclerosis: A case series. Journal of Prosthetic Dentistry. 83. 1. 90-98. Fairbanks, G. (1960). Voice and Articulation Handbook. New York: NY, Harper & Row. Fletcher, S. (1989). Palatometric specification of stop, affricate, and sibilant sounds. Journal of Speech and Hearing Research. 32. 736-738. Fletcher, S.G., Dagenais, PA., & Critz-Crosby, P. (1991). Teaching consonants to profoundly hearing-impaired speakers using palatometry. Journal of Speech and Hearing Research. 34. 929-942. 252 Frattali, C. M. (1998). Measuring Outcomes in Speech-Language Pathology. New York: NY, Thieme. Garcia, J. M. & Cannito, M. P. (1996a). Top down influences on the intelligibility of a dysarthric speaker: Addition of natural gestures and situational context. In D. Robin, K. Yorkston, & D. Beukelman (Eds.), Disorders of Motor Speech: Assessment. Treatment, and Clinical Characterization, (pps. 89-103). Baltimore: Paul H. Brookes. Gibbon, F. (1990). Lingual activity in two speech-disordered children's attempts to produce velar and alveolar stop consonants: Evidence from electropalatographic (EPG) data. British Journal of Disorders of Communication. 25. 329-340. Gibbon, F., Dent, H., & Hardcastle, W. (1993). Diagnosis and therapy of abnormal alveolar stops in a speech-disordered child using electropalatography. Clinical Linguistics and Phonetics. 7. 247-267. Gibbon, F., Hardcastle, B., & Dent, H. (1995). A study of obstruent sounds in school-age children with speech disorders using electropalatography. European Journal of Disorders of Communication. 30. 213-225. 253 Gibbon, F., Hardcastle, W.J., Dent, H. & Nixon, F. (1996). Types of deviant sibilant production in a group of school-aged children, and their response to treatment using EPG. In M.J. Ball and M. Duckworth (Eds.), Advances in Clinical Phonetics, (pp. 115-149). Amsterdam: John Benjamins Company. Hardcastle, W. J. (1972) The use of electropalatography in phonetic research. Phonetics. 25. 197-215. Hardcastle, W.J., Gibbon, F. E., & Jones, W. (1991). Visual display of tongue- palate contact: Electropalatography in the assessment and remediation of speech disorders. British Journal of Disorders of Communication. 26. 41-74. Hardcastle, W. J., Gibbon, F. E., & Nicolaidis, K. (1991). EPG data reduction methods and their implications for studies of lingual coarticulation. Journal of Phonetics. 19. 251-266. Howard, S. & Varley, R. (1995). Using electropalatography to treat severe acquired apraxia of speech. European Journal of Disorders of Communication. 30, 246-255. 254 International Phonetics Association. (1989). The principles of the International Phonetic Association (Handbook of the IPA). London: London University College. Liang, B. (1994). RMS Value: Calibration Tone (version 1). In-house computer programme, School of Audiology and Speech Sciences, UBC. Ling, D.L. (1976). Speech and the hearing-impaired child: Theory and practice. Washington, DC: A G . Bell Association fof the Deaf. Markides, A. (1970). The speech of deaf and partially hearing children with special reference to factors affecting intelligibility. British Journal of Audiology. 11. 51-58. Maxon, A. & Brackett, D., (1992) The Hearing Impaired Child: Infancy through High- School Years. Maryland: Butterworth-Heinemann. McGarr, N. (1983). The intelligibility of deaf speech to experienced and inexperienced listeners. Journal of Speech and Hearing Research. 26. 451- 458. 255 Michi, K , Yamashita, Y., Imai, S., Suzuki, N., & Yoshida, H. (1993). Role of visual feedback treatment for defective /s/ sounds in patients with cleft palate. Journal of Speech and Hearing Research. 36. 277-285. Monsen, R. B. (1978). Toward measuring how well hearing-impaired children speak. Journal of Speech and Hearing Research. 21. 197-219. Monsen, R. B. (1981). A usable test for the speech intelligibility of deaf talkers. Journal of Speech and Hearing Research. 21. 197-219. Monsen, R. B. (1983). The oral speech intelligibility of hearing-impaired talkers. Journal of Speech and Hearing Disorders. 48. 286-296. Monsen, R.B., Moog, J.S., and Geers, A E . (1987). CID Picture Speech Intelligibility Evaluation (SPINE). St. Louis, MO: Central Institute for the Deaf. Northern, J. and Downs, M., (1991). Hearing in Children. 4th edition, Williams & Wilkins: Balitmore, MD. Osberger, M. J. (1992). Speech intelligibility in the hearing impaired: Research and clinical implications. In R. D. Kent (Ed.), Intelligibility in 256 Speech Disorders: Theory. Measurement and Management (pp. 233-264). Amsterdam/Philadelphia: John Benjamins Publishing Company. Parsloe, R., (1998). Use of the speech pattern audiometer and the electropalatograph to explore the speech production/perception relationship in a profoundly deaf child. International Journal of Language and Communication Disorders. 33. 1, 109-121. Rammage, L. (1996) Vocalizing With Ease: A Self-Improvement Guide. Vancouver, Canada: Broadway Printers. Scobbie, J. M., Gibbon, F., Hardcastle, W., & Fletcher, P. (1995) Covert contrast and the acquisition of phonetics and phonology. In Ziegler, W. & Deger, K. (Eds.). Clinical Phonetics and Linguistics, (pp. 147-156), London: Whurr. Schiavetti, N. (1992). Scaling procedures for the measurement of speech intelligibility. In R. D. Kent (Ed.), Intelligibility in speech disorders: Theory, measurement and management (pp. 233-264). Amsterdam/Philadelphia: John Benjamins Publishing Company. 257 Sitler, R., Schiavetti, N., & Metz, D. (1983). Contextual effects in the measurement of hearing impaired speakers' intelligibility. Journal of Speech and Hearing Research. 26. 30-35. Smith, C. R. (1975). Residual hearing and speech production in deaf children. Journal of Speech and Hearing Research. 18. 795-811. Stephens, D. and Hetu, R., (1991). Impairment, disability, and handicap in audiology: towards a consensus. Audiologv. 30. 4, 185-200. Till, J., K. Yorkston & D. R. Beukelman. (1994). Motor Speech Disorders: Advances in Assessment and Treatment. Baltimore, Maryland: Paul H. Brookes Publishing Co., Inc. Whitehill, T., Stokes, S., Hardcastle, B. and Gibbon, F. (1995). II: EPG in the description of normal and disordered speech production electropalatographic and perceptual analysis of the speech of Cantonese children with cleft palate. European Journal of Disorders of Communication. 30. 193-202. Williams, R. (1998). Outcomes of Palatometry Therapy as Perceived By Untrained Listeners. Unpublished Masters Thesis, University of British Columbia, Canada. 258 World Health Organization. (1999). ICIDH-2: International Classification of Functioning and Disability. Beta-2 Draft. Short Version. Geneva, Switzerland: Author. Yorkston, K. M., & Beukelman, D. R. (1981a). Assessment of Intelligibility of Dysarthric Speech. Tigard, OR: C. C. Publication. Yorkston, K. M., & Beukelman, D. R. (1981b). Communication efficiency of dysarthric speakers as measured by sentence intelligibility and speaking rate. Journal of Speech and Hearing Disorders. 46. 296-301. Yorkston, K., Strand, E., and Kennedy, M. (1996). Comprehensibility of dysarthric speech: Implications for assessment and treatment planning. American Journal of Speech-Language Pathology. 5. 1. 55-66. APPENDIX A 259 EXPERIMENTAL STIMULI Stimuli Chosen for the Open-Set and Closed-Set Word Identification Tasks (Goal and generalization target phonemes are italicized). Pre-Assessment (Tl) Post-Assessment (T2) 1 glasses glasses 2 gum gum 3 PW Pi£ 4 judge judge 5 jump jump 6 page page 7 cherries cherries 8 chicken chicken 9 itchy itchy 10 shoe shoe 11 brush brush 12 fishing fishing 13 calling calling 14 truck truck 15 snake snake 16 soap soap 17 ice ice 18 zipper zipper 19 200 200 20 DU22 b u 2 2 260 Sentences from the Rainbow Passage Pre-Assessment (Tl) Post-Assessment (T2) 1 The rainbow is a division of light into many beautiful colours. The rainbow is a division of light into many beautiful colours. 2 People look but no one ever finds it. People look but no one ever finds it. 3 When a man looks for something beyond his reach, When a man looks for something beyond his reach, Sentence Presentation Per Listener Per Visit. Listener Visit 1 Visit 2 1 1. People look but no one ever finds it (pre-therapy, sentence B). 2. When a man looks for something beyond his reach (post-therapy, sentence C). 3. The rainbow is a division of white light into many beautiful colours (pre-therapy, sentence A). 1. The rainbow is a division of white light into many beautiful colours (post-therapy, A). 2. People look but no one ever finds it (post-therapy, B). 3. When a man looks for something beyond his reach (pre-therapy, C). 2 1. People look but no one ever finds it (post-therapy, B). 2. The rainbow is a division of white light into many beautiful colours (pre-therapy, A). 3. When a man looks for something beyond his reach (pre- therapy, C). 1. When a man looks for something beyond his reach (post-therapy, C). 2. People look but no one ever finds it (pre-therapy, B). 3. The rainbow is a division of white light into many beautiful colours (post-therapy, A). 3 1. People look but no one ever finds it (pre-therapy, B). 2. The rainbow is a division of white light into many beautiful colours (post-therapy, A). 3. When a man looks for something beyond his reach (post-therapy, C). 1. When a man looks for something beyond his reach (pre-therapy, C). 2. People look but no one ever finds it (post-therapy, B). 3. The rainbow is a division of white light into many beautiful colours (pre-therapy, A). 4 1. People look but no one ever finds it (pre-therapy, B). 2. When a man looks for something beyond his reach 1. The rainbow is a division of white light into many beautiful colours (post-therapy, A). 2. People look but no one ever 261 (post-therapy, B). 3. The rainbow is a division of white light into many beautiful colours (pre-therapy, A). finds it (post-therapy, B). 3. When a man looks for something beyond his reach (pre-therapy, C). 5 1. People look but no one ever finds it (post-therapy, B). 2. When a man looks for something beyond his reach (pre- therapy, C). 3. The rainbow is a division of white light into many beautiful colours (post-therapy, A). 1. The rainbow is a division of white light into many beautiful colours (pre-therapy, A). 2. People look but no one ever finds it (pre-therapy, B). 3. When a man looks for something beyond his reach (post-therapy, C). Palatometry Pre- & Post-Assessment Stimuli Igl HI ity /dz/ Ikl Isl Izl egg posh each hodge eek peace bees POg push botch jujube mock boss Oz rog shoe pooch jaw koop moose ooze goo Shah chew cop sue zoo gob She chop key saw zee geese cheap see APPENDIX B THERAPY WORDS Treatment Cycle 1: Session 1: isl Session 2: / t s / Session 3: /dz/ Session 4: IV 1 go cho jam shop 2 game chan jim shoot 3 got chore jean sheet 4 gas chomp jude shawl 5 gum chum joe shave 6 guy chain gel show 7 gone chill jut shut 8 good chess jar shine 9 girl chair jeep sharp 10 goose cheap jab shell Treatment Cycle 2: Session 1: Session 2: Session 3: Session 4: fty /dz/ iy /g/ 1 a chive a jade the shop go out 2 chap man gel light shut tight a guy 3 chill light jeep pit a sheet the game 4 the chirp the jam shell lee that goose 5 chain net a jut the shoot good food 6 that chore that jim that show the girl 7 cheap pin joe oat shave vim I got 8 a chomp jab boot a shawl gas light 9 chip pit jean knit shine night gone night 10 the chum the jar sharp pin gum mom 263 APPENDIX C QUESTIONNAIRES Original Palatometry Research Questionnaire (administered to Samantha in previous studies) 1. a) List the things you do now to improve your speech. b) When someone doesn't understand you, what do you do to make yourself understood? 2. Have you made any changes in how you speak as a result of this therapy? yes no. If yes, describe the changes. 3. Do you find that you enjoy speaking more as a result of this therapy? yes no. 4. a) Are you able to use your new sounds outside of the therapy session? (circle one) never when practising—-occasionally in conversation—frequently in conversation always 264 b) Is it easier or harder to imitate the speech therapist now? 5. Have any of your family or friends made comments about your speech since starting this program? yes no. If yes, what did they say? 6. a) Have you noticed that you repeat yourself less often when talking to familiar speakers (friends or family you see often)? yes no. b) when talking to unfamiliar listeners? yes no. 7. Have you noticed any other changes? 8. Is the palate lift helping your speech? If yes, describe how. Questionnaire #1: Palatometry Research Questionnaire (administered to Samantha during the present study) Questions 1-8 were the same as above. Additional questions, 9-20 were included. These questions focused on her speech and participation within society. 9. My speech will improve if I work hard: A LOT A MEDIUM AMOUNT A LITTLE BIT NOT AT ALL 265 10. I can usually make strangers feel at ease with me: ALWAYS SOME OF THE TIME HARDLY EVER NEVER 11. My speech is difficult for strangers to understand: (circle one) A LOT A MEDIUM AMOUNT A LITTLE BIT NOT AT ALL (circle one) ALWAYS SOME OF THE TIME HARDLY EVER NEVER 12. My family finds me difficult to understand: (circle one) A LOT A MEDIUM AMOUNT A LITTLE BIT NOT AT ALL (circle one) ALWAYS SOME OF THE TIME HARDLY EVER NEVER 13. My speech is slow: A LOT A MEDIUM AMOUNT A LITTLE BIT NOT AT ALL 14. My speech is fast A LOT A MEDIUM AMOUNT A LITTLE BIT NOT AT ALL 15. My speech is too loud: A LOT A MEDIUM AMOUNT A LITTLE BIT NOT AT ALL 16. My speech is too soft: A LOT A MEDIUM AMOUNT A LITTLE BIT NOT AT ALL 266 17. My speaking is poorer when I am tired: ALWAYS SOME OF THE TIME HARDLY EVER NEVER 18. My voice sounds hoarse or harsh: A LOT A MEDIUM AMOUNT A LITTLE BIT NOT AT ALL 19. My speech has a nasal quality: A LOT A MEDIUM AMOUNT A LITTLE BIT NOT AT ALL 20. Do you find you avoid any of these situations because of your hearing loss? (a) Going for coffee with a friend: A LOT A MEDIUM AMOUNT A LITTLE BIT NOT AT ALL (b) Going to class: A LOT A MEDIUM AMOUNT A LITTLE BIT NOT AT ALL (c) Talking with teachers/professors: A LOT A MEDIUM AMOUNT A LITTLE BIT NOT AT ALL (d) At a restaurant with a friend: A LOT A MEDIUM AMOUNT A LITTLE BIT NOT AT ALL (e) Talking on the telephone: A LOT A MEDIUM AMOUNT A LITTLE BIT NOT AT ALL 267 (f) Going to a movie: A LOT A MEDIUM AMOUNT A LITTLE BIT NOT AT ALL (g) Going to church: A LOT A MEDIUM AMOUNT A LITTLE BIT NOT AT ALL (h) Going shopping in a shopping mall: A LOT A MEDIUM AMOUNT A LITTLE BIT NOT AT ALL Questionnaire #2: Participation Research Questionnaire Questions selected from The Communication Profile for the Hearing Impaired (CPHI, Demorest & Erdman, 1986, 1987). The rating scale was as follows: (1) rarely, (2) sometimes or occasionally, (3) half the time, (4) sometimes frequently, (5) almost always. 1. When I have trouble understanding someone, I pay close attention to his or her face. 2. Since I have trouble hearing, I don't enjoy going places with friends as much. 3. When I must listen in a group, I try to sit where I'll be able to hear better. 4. Others become impatient because I don't always understand. 5.1 try to hide my hearing problems. 6. When there's background noise I position myself so that it's less distracting. 268 7. Sometimes it's hard for me to understand what's being said in meetings, conferences, or other large groups. 8. At social gatherings I sometimes find it hard to follow conversations. 9. The difficulties I have with my hearing restrict my social and personal life. 10.1 avoid conversing with others because of my hearing loss. 11. Members of my family speak to me when they're not facing me. 12. When I don't understand what someone has said, I ask them to repeat it. 13. Sometimes it's difficult for me to follow a conversation when others are talking nearby. 14. At parties or other social gatherings I try to stay in a well-lit area so I can see the speaker's face. 15. Not being able to understand is very discouraging. 16. When I don't understand what someone has said, I explain that I have a hearing loss. 17.1 have to communicate with others in a group situation. 18.1 interrupt others when listening to them is difficult. 19. I've asked my family to get my attention before speaking to me. 20. If I hear part of what someone has said, I only ask them to repeat what I didn't hear. 21.1 feel embarrassed when I have to ask someone to repeat what they've said. 22. In difficult listening situations, I try to position myself so that I can hear as well as possible. 269 23.1 have conversations with others when I'm home. 24.1 try to give the impression of normal hearing. 25. If someone repeats what they've said and I still don't understand, I ask them to repeat again. Questions selected from Erickson's S-24 Scale 1.1 find it easy to talk with almost anyone: ALWAYS SOME OF THE TIME HARDLY EVER NEVER 2. A person who is my teacher or my boss is hard to talk to: ALWAYS SOME OF THE TIME HARDLY EVER NEVER 3. Some words are harder than others for me to say: ALWAYS SOME OF THE TIME HARDLY EVER NEVER 4.1 would rather not introduce myself to a stranger: ALWAYS SOME OF THE TIME HARDLY EVER NEVER 5.1 feel comfortable speaking before a group: ALWAYS SOME OF THE TIME HARDLY EVER NEVER 6.1 find it easiest to talk with persons younger than me: ALWAYS SOME OF THE TIME HARDLY EVER NEVER 7. My speaking voice is rather pleasant and easy to listen to: ALWAYS SOME OF THE TIME HARDLY EVER NEVER 8.1 face most speaking situation with complete confidence: ALWAYS SOME OF THE TIME HARDLY EVER NEVER 270 9. There are A LOT OF / SOME / FEW/ NO people I can talk with easily. 10.1 talk better than I write: A LOT A MEDIUM AMOUNT A LITTLE BIT NOT AT ALL 11. Even though I knew the right answer I have often failed to give it because I was afraid to speak out: A LOT A MEDIUM AMOUNT A LITTLE BIT NOT AT ALL 12.1 feel most comfortable talking with my family: ALWAYS SOME OF THE TIME HARDLY EVER NEVER Questionnaires Administered to Listeners Visit #1 Read 1 & 2 below: 1) You may fill out as much or as little of this questionnaire as you choose. 2) Your answers will not identify you by name in any work or in any data files. Task 1: Word Identification 1. How clifficult did you find this task (circle one)? very difficult moderately difficult not at all difficult 271 2. How much effort did this task require? 3. What did you base your decisions on during this task? 4. Describe the speech of this speaker's words: Task 2: Sentences 1. How difficult did you find this task (circle one)? very difficult moderately difficult not at all difficult 2. How much effort did this task require and WHY? 3. What did you base your decisions on during this task? 4. Describe the speech of this speaker's sentences: 5. If given the opportunity, where would you feel comfortable having a conversation with this speaker (circle one or more below)? a) at a busy restaurant b) at a quiet coffee place c) at a night club d) at a library e) other: for example f) nowhere I can think of 6. What kinds of activities would you feel comfortable participating in with this' speaker (circle one or more below)? 272 a) a college project b) a game of doubles tennis where this speaker is your partner c) a game of chess d) shopping in a mall e) other : f) none I can think of 7. Comments? Visit #2 Read 1 & 2 below: 1) You may fill out as much or as little of this questionnaire as you choose. 2) Your answers will not identify you by name in any work or in any data files. Word Choice Task: 1. How difficult did you find this task (circle one)? very difficult moderately difficult not at all difficult 2. Was this task easier, more difficult or just as difficult or easy than the previous task (i.e., task #1 in which you had to write the sounds/words you thought you heard)? 273 Circle One: easier more difficult just as difficult just as easy Why? 3. How would you describe this speaker's speech (circle one)? severely impaired moderately impaired mildly impaired not impaired For the following questions, circle one of the options: 1. If given the opportunity, how often would you like to go to the movies with this speaker? all the time sometimes hardly ever not at all 2. If given the opportunity, how often would you like to go for coffee with this speaker? all the time sometimes hardly ever not at all 3. If given the opportunity, how often would you like to talk on the telephone with this speaker? all the time sometimes hardly ever not at all 4. If given the opportunity, how often would you like to go shopping in a shopping mall with this speaker? all the time sometimes hardly ever not at all 5. If given the opportunity, how often would you like to participate in a work project with this speaker? all the time sometimes hardly ever not at all 274 6. For approximately how long would you like to have a conversation with this speaker? more than 5 minutes between 3-5 minutes under 3 minutes not at all 7. If given the opportunity, how often would you like to participate in a letter exchange (i.e., write letters) with this speaker? all the time sometimes hardly ever not at all 8. If given the opportunity, how often would you like to participate in a team athletic event (i.e., soccer or volleyball) with this speaker? all the time sometimes hardly ever not at all 9. If given the opportunity, how often would you like to email this speaker? all the time sometimes hardly ever not at all 10. Would you feel comfortable conversing with this speaker? very comfortable slightly comfortable not at all comfortable 11. What problems do you think this speaker may have, if any, in the following situations: (a) work (b) school (c) family (d) peer group 12. Is there anything else you would like to say about communication with this speaker or others with similar speech? 275 APPENDIX D PALATOGRAM RESULTS CHOP (stop portion) + + + N (neutral) + - - + stop closure * (narrower) alveolar region contact * (more) alveolar & palatal "medial" contact: R-side * (more) L-side * (more) alveolar & palatal "lateral" contact: R-side * L-side * dorsal "medial" contact: R-side * (more) L-side * dorsal "lateral" contact: R-side * (more) L-side * 276 Pre-Assessment Stop Portion Phase of chop 0.71 0.72 0.73 0.74 0.75 0.76 0.77 0.78 0.79 0.80 277 Post-Assessment Stop Portion Phase of chop 8.81 0.02 0.03 0.04 0.05 278 CHOP (fricative portion) + + + N (neutral) + - - + groove width * (wider) alveolar region contact * alveolar & palatal "medial" contact: R-side L-side * (more) * (less) alveolar & palatal "lateral" contact: R-side L-side * (less) * dorsal "medial" contact overall * (more) dorsal "medial" contact: R-side L-side * (more) * dorsal "lateral" contact: R-side L-side * (more) * (less) 279 Pre- and Post-Assessment Fricative Phases of chop Channel 1 represents Pre-Assessment Channel 2 represents Post-Assessment 280 JUJUBE (stop portion) + + + N (neutral) + - - + stop closure * alveolar region contact * (less) alveolar & palatal "medial" contact: R-side L-side * (more) * (could have more) alveolar & palatal "lateral" contact: R-side L-side * (less) * (could have more) dorsal "medial" contact: R-side L-side * (more) * dorsal "lateral" contact: R-side L-side * * Pre-Assessment Stop Portion Phase of jujube 281 1.46 1.47 1.48 1.49 1.58 - s * ^ 282 Post-Assessment Stop Portion Phase of iuiube 1.4G 1.47 1.48 1.49 1.58 + + + * + . + + + + + • * 283 JUJUBE (fricative portion) + + + N (neutral) + - - + groove width * (narrower) alveolar region contact * (too forward) alveolar & palatal "medial" contact: R-side L-side * (more) * alveolar & palatal "lateral" contact: R-side L-side * (more) * (more) dorsal "medial" contact overall * (more R- side medially) dorsal "medial" contact: R-side L-side * (more) * dorsal "lateral" contact: R-side L-side * (more) * 284 Pre- and Post-Assessment Fricative Phases of jujube 285 SHE (maximum contact) + + + N (neutral) + - - + groove width * (wider) alveolar region contact * (less) alveolar & palatal "medial" contact: R-side L-side * (less) * (less) alveolar & palatal "lateral" contact: R-side L-side * (pulled back) * (pulled back) dorsal "medial" contact: R-side L-side * (more) * dorsal "lateral" contact: R-side L-side * * 286 Pre-Assessment Maximum Contact Phase of she 287 Post-Assessment Maximum Contact Phase of she 1.5G 1.57 1.58 1.59 1.G0 288 SHE (release) + + + N (neutral) + - - + groove width * (wider) alveolar region contact * (less) alveolar & palatal "medial" contact: R-side L-side * (more, less front) * alveolar & palatal "lateral" contact: R-side L-side * (less) * (less) dorsal "medial" contact overall * (less R- side) dorsal "medial" contact: R-side L-side * (less) * dorsal "lateral" contact: R-side L-side * * 289 Pre- and Post-Assessment Release Phases of she 2 9 0 GEESE (maximum contact) + + + N + - - + Symmetry * (more) Dorsal region medial contact: R-side L-side * (more) * (more) Dorsal region lateral contact: R-side L-side * (more) * (more) Back row contact * (more) GEESE (release) + + + N + - - + Symmetry * (more) Dorsal region medial contact: R-side L-side * (more) * (more) Dorsal region lateral contact: R-side L-side * (more) * (more) Back row contact * (more) 291 Pre- and Post-Assessment Maximum Contact Phases of geese 292 Pre- and Post-Assessment Release Phases of geese 293 COP (maximum contact) + + + N + - - + Symmetry * (more) Dorsal region medial contact: R-side L-side * * Dorsal region lateral contact: R-side L-side * (more) * (more) Back row contact * COP (release) + + + N + - - + Symmetry * (more) Dorsal region medial contact: R-side L-side * * Dorsal region lateral contact: R-side L-side * (more) * (more) Back row contact * (more) 2 9 4 Pre- and Post-Assessment Maximum Contact Phases of cop 2 9 5 Pre- and Post-Assessment Release Phases of cop 296 BEES (maximum contact) + + + N (neutral) + - - + groove width * (not full closure) alveolar region contact * (too back) alveolar & palatal "medial" contact: R-side L-side * (less for groove) * alveolar & palatal "lateral" contact: R-side L-side * (too back) * (too back) dorsal "medial" contact: R-side L-side * (more) * dorsal "lateral" contact: R-side L-side * * Pre-Assessment Maximum Contact Phase of bees 297 Post-Assessment Maximum Contact Phase of bees 298 2.GG 2.G7 2.G8 2.69 2.78 299 Pre- and Post-Assessment Release Phases of bees 300 BEES (release) + + + N (neutral) + - - + groove width * (wider) alveolar region contact * alveolar & palatal "medial" contact: R-side L-side * (less) * (less) alveolar & palatal "lateral" contact: R-side L-side * * dorsal "medial" contact overall * (more symmetrical) dorsal "medial" contact: R-side L-side * (slightly less, yet acceptable) * dorsal "lateral" contact: R-side L-side * (less) * 301 BOSS (maximum contact) + + + N (neutral) + - - + groove width * alveolar region contact * (more) alveolar & palatal "medial" contact: R-side L-side * * alveolar & palatal "lateral" contact: R-side L-side * (more) * (more) dorsal "medial" contact: R-side L-side * * dorsal "lateral" contact: R-side L-side * * 302 Pre-Assessment Maximum Contact Phase of boss 0.67 0.68 0.69 303 Post-Assessment Maximum Contact Phase of boss 2.26 _ 2.27 2.28 2.29 2.30 304 305 BOSS (release) + + + N (neutral) + - - + groove width * alveolar region contact * (less) alveolar & palatal "medial" contact: R-side L-side * (less) * (less) alveolar & palatal "lateral" contact: R-side L-side * (more) * (more) dorsal "medial" contact overall * dorsal "medial" contact: R-side L-side * * dorsal "lateral" contact: R-side L-side * *

Cite

Citation Scheme:

    

Usage Statistics

Country Views Downloads
China 7 25
United States 1 3
City Views Downloads
Guangzhou 3 0
Fuzhou 2 0
Shenzhen 1 25
Beijing 1 0
Ashburn 1 0

{[{ mDataHeader[type] }]} {[{ month[type] }]} {[{ tData[type] }]}

Share

Share to:

Comment

Related Items