University of Southern California Dissertations and Theses

Construction validation of the Vocal Behavior Inventory

(USC Thesis Other)

Construction validation of the Vocal Behavior Inventory

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content CONSTRUCT VALIDATION OF THE VOCAL BEHAVIOR INVENTORY by Susan Gillen Gray A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (Communicative Disorders) June 1979 Copyright Susan Gillen Gray 1979 UMI Number: DP22515 All rights reserved INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted. In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if material had to be removed, a note will indicate the deletion. UMI DP22515 Published by ProQuest LLC (2014). Copyright in the Dissertation held by the Author. Dissertation Publishing Microform Edition © ProQuest LLC. All rights reserved. This work is protected against unauthorized copying under Title 17, United States Code ProQuest LLC. 789 East Eisenhower Parkway P.O. Box 1346 Ann Arbor, Ml 48106-1346 U N IVE R S ITY O F S O U T H E R N C A L IF O R N IA T H E G R A D U A TE SCHO O L U N IV E R S IT Y PARK LOS A N G ELES. C A L IF O R N IA 9 0 0 0 7 This dissertation, written by ..Susan. Gillen G r a y ................................. under the direction of he.V..... Dissertation Com mittee, and approved by all its members, has been presented to and accepted by The Graduate School, in partial fulfillm ent of requirements of the degree of D O C T O R O F P H I L O S O P H Y Dean _____ DISSERTATION O 'hair man ‘ Ph.D. C ytvcL /lcj &vdx TABLE OF CONTENTS Chapter I. II. INTRODUCTION .......... REVIEW OF THE LITERATURE Vocal Abuse ................... Assessment of Vocal Behavior . . Review of Methodological Problems Involved in the Assessment of Vocal Behavior ............... III. METHODS AND PROCEDURES Design Overview ................. Construction of the Vocal Behavior Inventory ..................... Pre-test of the Vocal Behavior Inventory ..................... Subjects--Phase 1 ............... Subjects--Phase 2 ............... Procedure--Phase 1 and Phase 2 . . Data Analysis--Phase 1 .......... Data Analysis--Phase 2 .......... VI RESULTS Phase 1 .... Phase 2 .... Post Hoc Analysis , V. DISCUSSION ................... |APPENDIX A. Vocal Behavior Inventory APPENDIX B. Tables ............... . Table A: Table B: Variable Names and Labels of the 49 VBI Items ................... Phase 1. Final Factor Analysis: All Factor Score Coefficients Page 1 4 12 15 21 22 23 28 29 30 30i 32| 34j 35} 35; 39; 41' 431 5 1 i 621 63 65 ii Table C Table D Table E Table F Table G Table H Table I Table J Table K Table L Table M Table N Table 0 Table P Phase 1. Final Varimax Rotated ' - Factor Matrix (N = ,53)........ .. Phase 1. Final Factor Analysis: A Summary of Factor Statistics and Item Communalities . . .......... Phase 1. Discriminant Analysis of Factor-Predictors (Method-Direct) . . : Phase 1. Case Classification Results of Discriminant Analysis of Factor- Predictors (Method-Direct) .......... Phase 1. Discriminant Analysis of Factor-Predictors (MethOd-Wilks) . . . Phase 1. Case Classification Results of Discriminant Analysis of Factor-Predictors (Method-Wilks) . . . Phase 2. Discriminant Analysis of Factor-Predictors (Method-Direct) . . Phase 2. Case Classification Results of Discriminant Analysis of Factor-Predictors (Method-Direct) . . Phase 2. Discriminant Analysis of Factor-Predictors (Method-Wilks) . . . Phase 2. Case Classification Results of Discriminant Analysis of Factor-Predictors (Method-Wilks) . . . Phase 2. Summary Table of Discriminant Analysis of 49 VBI Items ............. Phase 2. R.esults of Discriminant Analysis of 49 VBI Items ............. Phase 2. Case Classification Results for Discriminant Analysis of 49 VBI Items ......................... Case Classification Results for Phase 2 Subjects Based Upon Discriminant Analysis of 49 VBI Items for Phase 1 Subjects ............. ............... Table Q Phase 2. Percent of Correct and Incorrect Predictions Compared to Actual Group Membership . . REFERENCES CHAPTER I INTRODUCTION The assumption advanced in the literature is that vocal nodules develop as a result of vocal abuse. However, which vocal behaviors produce laryngeal tissue reactions viewed as signs of abuse have not been established. Also, the levels of frequency of such behaviors necessary to produce laryngeal tissue reaction seen as abuse remain uncertain. Vocal nodules are small, non-neoplastic swellings typically found at the juncture of the anterior j and middle third of the true vocal cords, the point of j ' I greatest force during vocal fold vibratory contact. Nodulesj are usually bilateral. Vocal abuse has been described as j Imechanical trauma to vocal fold tissue due to faulty or j excessive vocal use (Luchsinger 6c Arnold, 1965). Synonyms j i ;for vocal nodules such as "screamer's nodes," "preacher's j nodes," "singer's nodes," and "teacher's nodes" are noted by ; I Perkins (1977) to be descriptive of some of the vocal 1 behaviors considered to be abusive. In addition to scream- , ing, prolonged talking and singing, activities thought to be' 'abusive include talking in noise (Brown Kelly 6c Craik, 1952; Cooper, 1973; Ferguson, 1955; Greene, 1972), yelling and ■cheering (Withers, 1961), throat-clearing (Brewer, 1975; Brodnitz, 1971; Cooper, 1973; Greene, 1972; Luchsinger & Arnold, 1965; New & Erich, 1938) and coughing (Ballenger, 1969; Brewer, 1975; Greene, 1972; Jackson & Jackson, 1959). Reports in the literature implicate a number of phonatory symptoms which appear to relate to vocal abuse and development of nodules. The symptoms most frequently associated with disturbance of the vocal mechanism is hoarseness (Ash & Schwartz, 1943-44; New & Erich, 1938; Ryan et al., 1970). Perceptions of strain and tension (Brown Kelly & Craik, 1952; Cooper, 1973; Froeschels, 1943; Greene, 1972; Jackson & Jackson, 1959; Perkins, 1971a; jVan Riper & Irwin, 1958; Wilson, 1972), and fatigue and weakness (Brewer, 19 75; Brodnitz, 1971; Brodnitz & Froeschels, 1954; Cooper, 1973; Ferguson, 1955; Greene, ; i 1972) in the voice and throat are also discussed as being j isymptomatic of vocal abuse and nodules. i | Brown Kelly and Craik (1952) pointed out that for j jevery individual who develops vocal nodules, there may be | I ! [others who appear to engage in similar vocal behavior and . I | | •do not develop nodules. Variables thought to be related to ; vocal nodule development are: | I 1. Constitutional characteristics such as pyknic body i * f ,type, narrow asymetric larynges with "physiologically" lerossed arytenoids, allergic sensitivity and endocrine problems (Arnold, 1962), _ - - - - ^ 2. Psychological elements such as hostility, garru lousness, immature personality type (Ash 6c Schwartz, 1943-44), as well as vociferous and aggressive personalities (Arnold, 1962). 3. Adults who develop vocal nodules may use their voices differently, that is abusively, from those who do not develop nodules. Such vocal use may be more properly seen as vocal abuse. It is this latter aspect of vocal nodule development which has been selected for study in the present inves tigation. i Though there is a general assumption put forth in the jliterature that vocal abuse is related to the development of I i jvocal nodules, there has been no systematic investigation ofj i vocal behavior and its potential abusiveness which would j corroborate or clarify this thesis. The importance of this : 1 I jstudy lies in the need to provide empirical data that Iconfirm or refute the idea that certain types of vocal ; ibehavior are abusive in that they are strongly related to | the development of vocal nodules. Information gathered from’ this study can be utilized in more efficacious planning and limplementation of vocal rehabilitative procedures in adults ! j with vocal nodules. 1 3 CHAPTER II REVIEW OF THE LITERATURE The focus of this survey has been limited to publica tions which deal with concepts of vocal abuse, symptoms which may be related to vocal abuse, and the possibility of detecting pathology of the vocal cords based on assessment of vocal usage and vocal abuse symptomatology. This section has, therefore, been divided into three major parts: Vocal Abuse, Assessment of Vocal Usage and Symptomatology, and Review of Methodological Problems Involved in the i Assessment of Vocal Behavior. Within part one, ten sub- , i ' i ’ headings have been established to provide a meaningful j t [framework for discussion of the development of the Vocal ; 1 | jBehavior Inventory. References cited represent the thrust ( of work written in this area. No new perspectives involving1 i I 'these concepts have been recently published. i Vocal Abuse 1 Behaviorally engendered mechanical trauma to the vocal cords, described as vocal abuse, is cited as the most [probable cause of vocal nodules. Categories of vocal abuse l suspected by New and Erich (1938) as precipitators of vocal nodules are overuse or misuse of the phonatory mechanism. Misuse and excessive use of the voice are cited by Ash and Schwartz (1943-44) as etiological elements in nodule development. In 1948, Harris noted that nodules appear in individuals who almost always abuse their voice, but he does not describe the abusive components of voice usage. The special categorization of nodules as visible organic changes in the vocal cords due to functional misuse or abuse of the voice was noted by Brodnitz and Froeschels (1954). Withers (1961) wrote that using the voice for overly-long periods, and voice use at high pitches or during inflamation of the vocal cords constitutes vocal abuse. Arnold (1962) and Luchsinger and Arnold (1965) stated that faulty and excessive use of the voice is abusive. Ballenger (1969) classified vocal abuse under Autogenous Laryngeal Injuries i jand considered prolonged and strained use of the voice to Iconstitute such an abuse. Excessive or constant voice use < I is classified as vocal abuse by Ryan et al. (1970). j i •According to Greene (1972), vocal abuse has to be persistent y Jin order to cause definitive vocal cord changes. However, i 1 ! Wilson (19 72) pointed out that vocal abuse could occur as ! |the result of a single incident as well as from persistently, ^ maladaptive voice use. Aronson (1973) stated that high j pitch levels during talking, screaming or singing appeared X j to be the main element in causing phonation to be abusive. j 'Vocal abuse is described by Cooper (1973) as mistreatment of! the vocal cords, laryngeal and pharyngeal musculature > secondary to shouting, screaming or talking in a noisy environment. Abusive vocal use was ascribed by Perkins (1977) to phonation in high pitches and screaming. He pointed out that tense vocal muscles are heavily implicated in this type of vocal cord trauma. The foregoing impressions concerning the nature of vocal abuse and its relationship to vocal nodules remain speculative. Ferguson (1955) and Rubin (1961) indicated that elements which comprise voice abuse and their association with vocal nodule development continue to be unknown. Specific vocal activities considered to cause mechanical trauma to the vocal cords and consequently to be abusive are: excessive talking, shouting, screaming, yelling, cheering, talking in noise, singing, laughing, : yt I I (throat-clearing, and coughing. Damage to the vocal cords is J presumably increased if any of these activities are associated with laryngitis or extreme hoarseness. Symptoms ] j frequently associated with vocal abuse and nodules include j i , 1 'hoarseness, strain, tension, weakness, and fatigue of the j j i (voice and throat. Each of these aspects will now be ; I > jdiscussed in more detail. J i I Talking. Several authors consider talking to be a ; rprimary etiological factor in the development of vocal j nodules. New and Erich (1938) stated that "speaking out of , one's normal vocal register or attempting to produce a i 6 volume of tone beyond one's normal capacity" may cause vocal nodules. Brown Kelly and Craik (1952) reported that a primary cause of vocal nodules is vocal abuse in persons who must talk for long periods due to professional or personal demands. Froeschels (1943) writes "if a person has to speak very much, as in the case of a salesman, teacher, or lawyer, after a certain time symptoms and signs of voice trouble may appear." They stated that "those who spend hours conversing are often afflicted with vocal nodules." Arnold (1962) cited at risk for vocal nodules those persons in "talking occupations." Adults who use their voices professionally are frequently seen to have vocal nodules according to Brodnitz and Froeschels (1954). Persons who juse their voices "a great deal" are considered to show a j"high prevalence" of vocal nodules by De Weese and Saunders 1(1964). According to Brodnitz (1971) vocal nodules "occur j ! i in every group of persons who make extensive use of their j f (voice." Aronson (1973) stated that women who "use their j I i t Jspeaking voices excessively consist of those who teach or j ;are in other high voice use occupations and those who do | |much talking as a way of life" show a "high incidence" of j i ;vocal nodule development. : I 1 1 Talking in noise. An important contribution to ; 'development of vocal nodules is loud talking in the presence^ iof environmental noise. Brown Kelly and Craik (1952) found ( 1\ that several of their vocal nodule cases "were workers in machine shops who carried on their daily conversation above the noise of machinery.” Ferguson (1955) said that "vocal nodules occur most frequently in workers who must speak above noise." Cooper (1971) stated that "competition with industrial noise leads to voice problems" and "any area in which environmental noise or sound is excessive or abnormal for normal conversation" leads to vocal abuse by "demanding an altered vocal pattern." Greene (1972) found that vocal nodules might develop in "people who have to speak in noisy surroundings may also strain the voice working in shops and factories." Shouting and s creaming. Shouting and screaming are thought to be responsible for vocal cord trauma leading to ivocal nodules. New and Erich (1938) reported several ! 'patients who developed nodules "after a period of shouting." Episodes of "acute screeching" are cited as contributing to j vocal nodule development by Ash and Schwartz (1943-44). i j : Vocal nodules can be produced, according to Brown Kelly and !Craik (1952) by "shouting or screaming." "Shouting, screaming, and excessive use of the voice" are common forms 1 i ' I : iof laryngeal trauma leading to vocal nodules according to ; i De Weese and Saunders (1964). Jackson and Jackson (1959) ' 1 Y 'write that nodules of the vocal cords may result from a m isingle instance of shouting. Shouting and screaming are cited by Perkins (1971b), Greene (1972), and Cooper (1973) as sources of vocal abuse leading to vocal nodule develop ment . Cheering and yelling. Withers (1961) reported that cheerleaders are often subject to vocal nodules. Singing. New and Erich (1938) stated that vocal nodules can develop as a result of "singing out of one's normal vocal register or attempting to produce a volume of ^ tone beyond one's normal capacity." Singing by untrained persons is considered by Froeschels (1943) to be potentially more harmful to the vocal cords than speaking "even if used only for pleasure and relatively seldom." In singing, "straining to produce a forceful note" is thought to contribute to vocal nodule development by Jackson and j Jackson (1959). Arnold (1962) noted that even professional ] ; Throat-clearing and coughing. Cited as representing (either causes and/or symptoms of vocal abuse are throat- singers may develop vocal nodules after "much singing." jBrodnitz and Froeschels (1954) and De Weese and Saunders |(1964) added that singing above one's natural range contributes to vocal nodule development. i. i ’ clearing and coughing. Brewer (1975), Brodnitz (19 71), 'Cooper (1971), and New and Erich (1938) cited throat- Iclearing as a symptom often associated with vocal abuse and particularly with talking in a noisy environment. Cooper (19 71) reported that "throat-clearing may be contributory to the initial and continued irritation of the vocal folds." Luchsinger and Arnold (1965) and Greene (1972) included vocal cord action involved in throat-clearing as one cause of vocal abuse. Cough was mentioned as a cause of vocal abuse by Jackson and Jackson (1959, Ballenger (1969), Cooper (1971), and Greene (1972). Brewer (1975) stated that Inonproductive coughing is suggestive of laryngeal irritation I jand may be due to vocal abuse. ! | Laryngitis. Withers (1961) and Rubin (1961) concurred jthat talking or singing during acute inflamation of the ivocal cords may result in vocal abuse that is related to | | Jdevelopment of vocal nodules. De Weese and Saunders (1964) jstated that using the voice during a period of laryngitis ' i 1 jmay produce vocal nodules. j ! ! i Hoarseness. Rubin (1961) reported that vocal nodules I * I cause varying degrees of hoarseness in the speaking voice 1 i and was joined in his observations by New and Erich (1938), | Ash and Schwartz (1943-44) , Harris (.1948) , Brodnitz and | 1 I Froeschels (1954), Ryan et al. (1970), Brodnitz (1971), and iCooper (1971). i 1 ! Strain and tension. Feelings of strain and tension in the throat are reported by individuals who develop vocal nodules. Froeschels (1943) and Greene (1972) Indicated that speaking for long periods of time or in a noisy environment can cause vocal strain and lead to vocal nodules Froeschels (1943) mentioned the vocally abusive nature of overstraining the singing voice. Brown Kelly and Craik (1952) stated that 54 percent of their patients with vocal nodules gave a "definite history of overstraining the voice." They cited shouting or screaming as sources of strain which are abusive to the vocal mechanism. Van Riper and Irwin (1958) concurred with Tarneaud's early work indicating that vocal nodules are caused by the action of highly tensed vocal cords producing a condition of (excessive strain during phonation. Straining to produce a l forceful note in singing is considered to be traumatic to j :the larynx by Jackson and Jackson (1959). The association between vocal nodules and high pitches and screaming points |to tense laryngeal muscles as a major factor in vocal j itrauma, according to Perkins (1971b). Excessive vocal volume! ' ! and vocal misuse are reported by Cooper (1971) as straining ■ the vocal musculature. Wilson (1972) states that tension ^ ! A j ■may appear in various muscles of the larynx, pharynx and j (neck and which may result in vocal abuse. Greene (19 72) and; ■Perkins (1971a) discuss the relationship between hyperkinetic functioning of the vocal cords and excess laryngeal tension i and possible consequent vocal nodules. Greene (1972) stated^ that vocal strain due to excessive laryngeal muscle tension ' is the most common form of voice disorder. She reported that this condition is particularly prevalent in professional individuals who use their voices extensively. Cooper (1973) wrote that excessive laryngeal and pharyngeal tension is often created by the need for individuals to talk above noise. This can progress to a feeling of marked tens ion in the throat. Fatigue and weakness. Fatigue and weakness of the vocal musculature have been implicated in the development of vocal nodules. Vocal fatigue is considered by Brodnitz and Froeschels (1954) and Brewer (1975) to be associated with vocal abuse and/or vocal pathology. Ferguson (1955) considered musculature tension in the laryngeal area to be responsible for feelings of fatigue in the voice and in the the voice becomes increasingly fatigued as vocal nodules I the vocal cords can cause "chronic laryngitis and weakness Authors alluding to the possibility of detecting individuals with vocal cord nodules based on voice usage patterns and symptoms include Brodnitz (1971), Wilson (1972), throat area. Brodnitz (1971) and Cooper (1973) stated that jdevelop. Greene (1972) reported that hyperkinetic use of or tiredness of the muscles and laryngeal joints" involved in phonation. -Assessment of Vocal Behavior Laguaite (1972), and Brewer (1975). Brodnitz (1971) recommended comprehensive questioning concerning the emotional and vocal behavior of a person with defective voice. He stated that this was very important in the differential diagnosis of laryngeal disorders. No specific directions for such questioning, or correlation of patient responses to types of laryngeal dysfunction were provided. In 1972, Wilson reported a Vocal Abuse Rating Scale he devised for use on children with nodules. The scale appeared to be designed primarily as a descriptive tool j noting the absence or presence of vocal behaviors consideredj by him to be abusive. The behaviors found were then rated j on each of two 3-point scales. The first scale scored the { I f frequency of the observed behavior and the second scale i described the severity of each behavior. No information ! i was provided as to the response patterns or scores of ! jchildren with nodules on this Vocal Abuse Rating Scale. ! ! ! j Laguaite (1972) attempted to detect the presence of ( jearly laryngeal disease in people participating in a multi- ; Iphasic health screening program. She evaluated her subjectSi in two ways. First, samples of the vocal quality of the i subjects were tape recorded and judged for deviancy by the !examiner. Second, each participant completed a question- i naire designed to elicit information about voice usage and history of voice or throat ailments. Individuals whose i voice quality was judged to be deviant, or who gave a positive history of voice difficulty, were referred for laryngeal examination. Analysis of the type and frequency of voice usage in the referred group was not compared to type and frequency of voice usage in the non-referred group. Consequently, the relationship between responses to vocal usage questions and the presence or absence or vocal pathology was left unexplored. Brewer (19 75) listed groups of physical and phonatory signs and symptoms which he stated could direct professional attention to the larynx as a possible site of early disease. The paper contained no data concerning the number of subjects involved nor did it state what the correlations were among the signs, symptoms and instances of laryngeal j i 'pathology. However, he did indicate that symptoms | j I 'frequently considered to be "emotional" were, on more j I I jcareful examination of the larynx, often related to organic J ' i ilaryngeal disease. ! i ! j How vocal use behaviors and symptoms may differ in . , \ persons with and without vocal nodules is at present i i lundetermined. The relationship of vocal use behavior to the' 1 i ^development of vocal nodules is unclear at the present time., I ; i These papers seem to indicate that laryngeal pathology : might be identified earlier by the evaluation of a patient's1 vocal use and symptoms. 14 Review of Methodological Problems Involved in the Assessment: of Vocal Behavior Two general problem areas related to the methodological approaches of the aforementioned studies addressing vocal assessment became apparent. First, each of the studies considered different types and extents of vocal behavior. Second, each of the studies demonstrated certain limitations in the measurement and analysis of subject responses. A more detailed discussion of these issues as they related to I the studies in question follows. That Brodnitz (1971) did not specify which questions concerning vocal behavior he developed over his many years as a distinguished laryngologist represents an unfortunate (loss. It was not possible to determine, therefore, what ! vocal activities or conditions he may have considered particularly relevant to the development of vocal nodules. j ! I Furthermore, no mention was made of any analysis being ; ! j [performed on the information Brodnitz obtained from his j jquestions. Current investigators of vocal pathology cannot,^ ;then, replicate any aspect of Brodnitz1s technique which he ! reported enhanced the differential diagnosis of voice I disorders. ! ; While Wilson (1972) listed what he considered to be eleven of the most common vocal activities assumed to be abusive, he did not explore any vocal conditions. Both , clinical literature and experience indicate that the I 15 presence of vocal conditions as well as the continuation of vocal activity in the presence of these conditions are implicated in the phenomenon of vocal abuse. Omission of these potential aspects of vocal abuse may be related in part to the method of measurement of behavior used by Wilson He utilized observers to rate subjects on two scales he devised to measure the vocal abuse activities he considered important. Ratings could only be made on directly observable activities, and not conditions. In addition, observers using the test varied and were not trained in any specific observation techniques. The first of Wilson's 3-point scale was for rating the amount of vocal activity observed. Its points represented (a) Little Used, (b) Frequently Used, and (c) Excessively i Used. The distance between each of the three points does J not appear to represent a reasonable progression of j ! intervals or space. The designations Little Used and j I j Excessively Used were made upon individual judgments on the ' !part of each observer-rater as to their concept of "little" j i i i iand "excessive" in terms of vocal activity. Furthermore, 1 I ! I the designation Frequently Used does not appear to represent, 'a mid-point between "little" and "excessive" (Kerlinger, > ,1964). This apparent unequal approximation of space ! i ibetween points 1 and 2 would seem to predispose ratings of vocal activity to be placed in the high frequency of t 'occurrence range of the scale. i Wilson's second scale was designed for observers to rate the degree of severity of each activity. This requires a tremendous amount of inference on the part of the rater. How "severe" an activity may be in terms of precipitating laryngeal tissue reaction, however, cannot be known from witnessing a subject's outward manifestation of vigor in performing an activity. The three points on this second scale purporting to measure severity are arranged over more equal-appearing distances from one another and are designated (a) Mild, (b) Moderate, and (c) Severe. Both scales contain too few points to obtain much variability in response to the items. The limitations of the two rating scales used to measure selected behaviors extended further into issues of validity related to observer ratings. Problems (Kerlinger, 1964; Webb et al., 1966) inherent in i the use of observers to rate data include: 1. The observer may not see a truly representative sample of subjects' vocal behavior. 2. The observer's presence can bias the behavior of subjects during observation. ! 3. Observation of individuals known to have vocal ■nodules can be biased by possible alterations in their jvocal behavior secondary to the presence of the voice !disorder itself. \ ; The subjective reaction of individuals to their vocal i jactivity, and not the activity itself, could well be a component of vocal abuse. This type of reaction is not visible to an observer's perception of the performance of an activity per se. In addition to aspects of vocal abuse not examined and limitations in rating scales used, Wilson does not mention how many subjects were evaluated or what pattern of results was obtained. Consequently, there is no way to determine the efficacy of his examination of vocal abuse or to replicate his approach. In Laguaite's (1972) study addressing detection of vocal pathology using analysis of vocal quality and a Voice j Questionnaire she developed for that purpose, several issues! emerged. First, only six vocal activities and five vocal j ! I jconditions she thought to be related to vocal abuse were 1 I i |evaluated. This number of questions appeared to be too i * t » Jlimited to effectively assess an adequate array of vocal j behavior. The fact that no case of suspected vocal ! ► pathology was detected from responses to the questionnaire | I • lalone (Laguaite, 1976) tends to support this conclusion. j 1 jSelected items on Laguaite's Voice Questionnaire were i ] ^permitted a response range of Frequently, Occasionally, and ; ! Never. As in Wilson's (1972) scale of amount of vocal •activity, Laguaite's scale points appear to represent unequal distances between them. The end point Never is not ■ counterbalanced by an opposite end point of Always, and Occasionally does not logically fall as a mid-point between 18, Never and Frequently. These scale points would tend to have subjects place themselves at either a high or low end of a continuum of frequency with no designation available for more moderate alternative responses. This study, too, appears to examine an insufficient number of vocal behaviors to an adequate degree for the purpose of detecting vocal pathology based on questionnaire results alone. Analysis of subjects brought to laryngeal examination revealed a rela tively small number of heterogeneous cases of laryngeal jpathology in the sample. Detection occurred through t analysis of vocal quality and was not substantial. This approach format did not produce sufficient results to warrant attempts at replication of the study. ! Brewer (1975) listed 33 laryngeal symptoms reported by i jhis patients, and 54 laryngeal signs that he observed as I j i ; possible indicators of early laryngeal disease. A reason- j able array of vocal conditions was queried. An analysis of J i vocal quality constituted the predominant method of assess- | ing phonatory behavior. He omitted questioning the ' jperformance of various types of vocal activity and the ! * frequency of those activities. Nevertheless, Brewer's ! statement that a correlation of signs and symptoms to detection of early laryngeal disease was suggestive. It was : unfortunate that he did not state the size of his, subject sample, the exact correlation he found or its level of ; f 1 significance. No explicit information about his results was 19: provided and, therefore, replication of his work is not possible. In summary, several methodological issues emerged from a review of the above-mentioned studies. First, a determin ation of an appropriate selection of vocal behavior adequate in terms of illustrating possible differences between normal and abusive vocal activity would have to be made. Second, an effective way to measure vocal behavior so that comparisons could be made between adults with and without vocal nodules would have to be developed. The above studies each contributed to a concept of identifying vocal jpathology and aberrant vocal behavior through evaluation of jan individual's vocal performance. The variations and 'limitations in the I iindicates the need i I ; i examination and measurement techniques for further investigation in this area. 20, CHAPTER III METHODS AND PROCEDURES The purpose of this study was to develop an instrument, the Vocal Behavior Inventory (VBI) to measure, for adults, those vocal behaviors which were thought to be vocally abusive. The measure was to be administered to two groups of subjects: (a) adults medically diagnosed as having vocal nodules, and (b) adults with no evidence of vocal nodules. The vocal nodule group would be considered representative of individuals who had used their voices abusively. Subjects in the non-nodule group were considered! i representative of individuals who used their voice but did ; i jnot experience subsequent development of vocal nodules. It \ jwas hoped that the VBI could demonstrate differences in I jvocal behavior between these two groups. If such differ- i jences could be measured, the VBI could be used to determine |which behaviors would most clearly distinguish between ( Inodule and non-nodule subjects. The ability of the VBI to i make such distinctions would be an indicator of a t theoretical construct of vocal abuse. If an indicator of I such a construct could be obtained, then it would be [valuable to test the validity of that construct. 21 Design Overview A Vocal Behavior Inventory was constructed for the purpose of identifying and measuring vocal behavior thought to contribute to vocal abuse. This instrument was then administered to two groups of subjects: adults with vocal nodules and adults considered to have normal larynges. Administration of the VBI was conducted in two phases. Phase 1 involved obtaining responses on the VBI from adults representing each of the two groups, that is, subjects with and without vocal nodules for the purpose of factor analysis Following collection of this sample of data, Phase 2 was initiated. Collection of responses to the VBI from a ; similar sample of adults with and without nodules as in ! I ! ;Phase 1 was conducted and designated as Phase 2. The I I !purpose of Phase 2 was to test the construct validity of ! I i i the VBI. | Factor analysis was performed on the responses to the VBI for the Phase 1 subjects. The factor structure of 1 iresponses was then examined to determine whether interpret- i |able dimensions of vocally abusive behavior could be found. Factor coefficients for each subject were used to compute :factor scores to be used in a predictive analysis in Phase 2. ‘ i ; Discriminate function analysis was performed on the I !factor scores resulting from administration of the VBI. This was to determine which test items best discriminated ^between the two groups. Then, the discriminant function scores were used in an attempt to correctly classify Phase 1 subjects as to their actual group membership. Discriminant function scores were also computed for Phase 2 subjects and an attempt was made to correctly classify them Jas to their known group membership. A high degree of success in correctly classifying Phase 2 subjects, on the basis of construct criteria derived from Phase 1 subjects, would indicate strong support for, and help to establish the construct validity of, vocal abuse as measured by the VBI. I Construction of the Vocal Behavior Inventory' , A review of the literature yielded what appeared to be ( t Jan appropriate selection of vocal behavior historically ' considered to be related to the development of vocal nodulesi jStatements describing these behaviors were formulated and | ;fell into three general categories: (a) vocal activities, j l i(b) debilitating vocal conditions (analogous to previously ! Imentioned "symptoms"), and (c) continuation of vocal j i ■activity in the presence of any debilitating condition. The' i ivocal activities represented in the Inventory were talking, ! ■'talking in noise, shouting and screaming, yelling and cheering, singing, and throat-clearing and coughing. The debilitating vocal conditions included for investigation , iwere hoarseness, tense-strained throat and tired-weak throat. i These conditions were examined in terms of their occurrence for each of the six vocal activities. Also examined was the frequency of laryngitis experienced by the subjects. This item in the Inventory was classified as a debilitating condition which represented a disease process and so was distinct from the other conditions which were viewed as possible direct sequelae of vocal abuse. Consequently, its occurrence for each of the six vocal activities was not queried, but continuation of any of the activities in the presence of laryngitis was examined. Then, continuation of any of the six vocal activities in the presence of any of the four debilitating vocal conditions was probed in the third category. Thus, 49 separate questions (items) formed j Part I of the VBI. All further discussion of the VBI, unless otherwise specified, refers to these 49 questions. j f Directions about administration of the VBI were I ! - I Iprinted at the beginning of the Inventory. Subjects with jvocal nodules were requested to indicate their average vocal [behavior during the period prior to the onset of their voice! i j ^problem. This time frame was specified in the VBI for two I I [reasons. First, to record the vocal behavior which preceded) .and presumably precipitated the development of nodules. 'Second, to exclude any modifications in vocal usage which ,may have resulted directly or indirectly from the laryngeal < ! .pathology itself. Directions to normal subjects simply | asked for their estimate of their average vocal behavior. , 1 I 24^ Respondents rated each of the 49 items on a 7-point scale. Each subject was asked to offer an opinion about the occurrence and frequency of vocal activity and conditions he experienced. The wording which designated the points on the opinion scale was adapted from wording on similar frequency scales presented in Shaw and Wright (1967), and Berdie and Anderson (1974), which discussed a bipolar model. Each end of the scale continuum is anchored by opposite and extreme concepts of frequency, using the terms Always and Never. The discussion of Berdie and Anderson (1974) regarding the importance of a defined mid-point for a scale suggested l About as Often as Not as a mid-point for the VBI 7-point response scale. Additional points were placed between the mid-point and each of the end points to create seven points i on a frequency continuum, which provided a range of response (alternatives to the items. The points on the scale ! Jprovided subjects with easily describable concepts of jfrequency represented by (a) None of the Time, (b) Very I , l ,'Seldom, (c) Seldom, (d) About as Often as Not, (e) Often, ! , ; (f) Very Often, and (g) All of the Time. Since the sense of' I i jthe wording designating the frequency points on the result- I ! j ing 7-point opinion scale did not differ markedly from j other previously tested scales, there was no reason to . • 'think that the equal-appearing intervals on the scale used i ' i for this study differed from the equal-appearing intervals !found on those other scales. In addition to the 49 items directly addressing aspects of vocal behavior, supplemental information was obtained from subjects. It was thought that this informa tion regarding aspects of the subjects' lives could be examined in an attempt to determine the existence of unique or confounding variables in the event of problems with the analysis of the 49 VBI items. The following supplemental information was gathered: 1. The age, sex and occupation of the respondents was the only demographic data asked. The respondents' occupa tion was coded into one of nine categories, professional, < managerial, technical, clerical, service, assembly, craftsman, homemaker, and not applicable. The occupations were also coded, -if possible, as to whether or not the occupational environment was noisy, and as to whether or not an integral aspect of the occupation included voice usage. ! [Since certain occupations require frequent and/or loud I i I I ivoice usage, it was thought that this information might be ! iuseful in the interpretation of later results. 1 1 2. Section D, Comparison, on the VBI was designed to lallow respondents to describe whether or not there had ever been a period prior to onset of the vocal difficulty when [their voice was optimum. Subjects giving an affirmative iresponse were asked to indicate the chronological time frame of their optimum vocal ability. They were also asked to (compare the frequency of occurrence of their vocal I 26. activities and debilitating conditions in that time frame to the period prior to developing nodules. Comparison of frequency of activities and conditions was made on a 3-point scale whose intervals represented Less, Same, and More. This information would allow examination of an inferred chronology of progressive laryngeal debility, ranging from a period of optimal vocal use through decreased vocal function to the presence of a recognizable J problem. ! 3. Section E, Comments, elicited information from subjects about the onset of their voice problem and about their impressions of any stressful vocal situations or activities. They were also asked to specify any difficulty j they may have had in interpreting or answering questions. ! i This feedback would be considered in any future modifica tions of questions on the VBI. 4. A Physician's Form was provided to each cooperating1 ! Jlaryngologist to complete. This provided a case history ; ] ' iwhich allowed comparison of the subjects' report of the 1 (Onset and course of their vocal problem. Description of the 'vocal nodules in terms of appearance and location, a ; [severity rating of the problem, and any other comments the j i , laryngologist wished to make were requested on this form. j |Such information might offer suggestions regarding problems with the data analysis. Pre-test of the Vocal Behavior Inventory In order to pre-test the VBI, a pilot study was conducted. Ten adults with vocal nodules and ten adults without nodules were administered the Inventory. At this stage, the VBI was administered in a structured interview format and the interview was tape recorded. Following the interview, subjects were asked whether they had experienced any difficulty with any aspect of the VBI. Any reports by subjects indicating ambiguity or lack of clarity in a question resulted in revision of the question. Modification of the Inventory continued until it was felt that a series jof clearly worded, workable questions had evolved. jAdministration time of the VBI at this stage ranged from i ;40 to 60 minutes. I t j An examination of these data showed that most adults j jwithout nodules demonstrated low frequencies of vocal ) jactivities and debilitating conditions, and virtually no j 1 occurrence of continuance of activities in the presence of j ! !conditions. In contrast, most adults with nodules exhibited, i ! jhigher frequencies of activities, debilitating conditions 1 land continuance of activities in the presence of these j i ^ \ conditions. Due to the small size of the pre-test sample, j however, no statistical tests could be performed on the ' data. It was the researcher's impression, however, from ■ I examination of cross tabulations, that the VBI was able to i register differing frequencies of vocal behavior among the subjects studied. Subjects--Phase 1 For Phase 1 of the study, 32 subjects with a medical diagnosis of vocal nodules, and 24 subjects without vocal nodules participated in the investigation. The subjects with vocal nodules included 7 males and 25 females. Their mean age was 35.2 years with a standard deviation of 12.75 years. These subjects were selected from available patients who had received treatment from various laryngological and speech pathology practices. The diagnosis of vocal nodules I for each subject had to meet the following criteria: (a) j i I the diagnosis had to be made by a laryngologist, and (b) each subject's diagnosis had to be approved by one of three I I icooperating laryngologists of high academic standing. ; ' I J Subjects without vocal nodules totaled 24 adults. ! | 1 'Twelve males and 12 females with a mean age of 38.2 years i |and a standard deviation of 9.61 years were represented. ; .The criteria for selection of non-nodule subjects included: | 1 : (a) normal vocal quality at the time of VBI administration, > ■this was judged by the researcher, an experienced speech pathologist, and (b) report by the subject of a negative 1 ^history for a voice problem of any type. This group was obtained from hospital personnel and acquaintances outside I of the hospital and appeared to represent a broad range of occupational categories. ; ' 29 Subjects--Phase 2 For Phase 2 of the study, 29 subjects with vocal nodules and 25 subjects without nodules were selected in a manner identical to that employed in Phase 1. There were 4 males and 25 females in the vocal nodules group. Their mean age was 35.7 years with a standard deviation of 13.64 years. In the non-nodules group there were 6 males and 19 females. Their mean age was 30.5 years with a standard deviation of 9.00 years. i Procedure--Phase 1 and Phase 2 j The procedure used for Phase 1 and Phase 2 was identical. Each subject with vocal nodules was asked to t read the directions prefacing the VBI and to complete the i ! | entire Inventory. The laryngologists who examined these . subjects completed the Physician's Form section of the VBI. ■ I I [Subjects without vocal nodules were asked by the researcher \ I j jto complete only the identification section and the 49-item 1 ;section of the VBI. Administration time of the VBI ranged ( i i from 10 to 20 minutes. Completed Inventories and Forms were [returned to the investigator and coded as soon after [completion as possible. This allowed for detection and remediation of any omission of identifying data on any ! ^subject. Consideration was given to the problem of missing data. Whenever a subject responded that he did not engage (had not engaged) at all in one or more of the six vocal behavior activities being investigated, information for that subject regarding that item was missing from the analysis. One course of action would have been to assign a "missing value" to later questions dealing with any debilitating conditions associated with that activity, as well as to questions dealing with the frequency of continuing that activity in the presence of any debilitating conditions. In addition, any response of "none of the time" to questions of experiencing a debilitating condition after a "missing value" activity implied that a "missing value" should also be assigned to those questions regarding continuing the "missing value" activity in the presence of a debilitating I condition. After due consideration of the implications, thej I decision was made to assign a score of "1" (None of the j I ' I Time) to the "debilitating condition" and "continuing jactivity with debilitating condition" items for any jactivity which a subject had responded to with "1” (None of jthe Time). This meant, for example, that if a subject j !answered that he yelled or cheered "none of the time," he * 1 I iwas assigned scores of "1" to indicate that his voice was | | | ihoarse, or tired-weak, or strained-tense "none of the time" j > \ 'after yelling or cheering, and scores of "1" to indicate he ;continued to yell and cheer "none of the time" if his voice or throat were tired-weak, strained-tense or if he had i laryngitis. In a similar manner, if a respondent answered 1 that he never experienced a debilitating condition as a result of an activity, he was assigned a score of ”1” to indicate he never continued the activity while experiencing the debilitating condition. This solution does result in obscuring the differences, for example, between an individual who shouts but does not get hoarse, and another individual who does not shout and, therefore, does not get hoarse from shouting. Nevertheless, it allows for a response pattern of "none of the time" which is reasonable as regards interpretation of results for this exploratory analysis. It also had the result of providing "valid" data for almost the entire sample. This made it jpossible to compute factor scores easily for virtually the i j jentire sample population. The few cases with missing data j (after implementation of the technique described were ! | i (Subjects who failed to respond to some questions in any way i i i jthat would permit the researcher to rationally assign jscores to replace the missing data. i 1 t I ; ; Data Analysis--Phase 1 j The first step of the analysis in Phase 1 was to jfactor-analyze the data to look for patterns of vocal k ‘ behavior. This was done using the Factor subroutine of the ( Statistical Package for the Social Sciences (SPSS) (Nie, 1975). All 49 variables (Table A, Appendix B) of the VBI ’ were entered into a principal components analysis with 32 iteration and the factors derived subjected to a varimax orthogonal rotation to aid in the interpretation of the factors. The second step of the analysis in Phase 1 used those items that seemed to best describe the five factors, and a factor score for each factor for each respondent (from Phase 1) was computed. Each factor score was calculated from its correlated items by summing the products of the factor score coefficients (derived from the Phase 1 analysis--see Table B, Appendix B) and the a-score transfor mations of the items. The factor scores were then entered into a Discriminant Function analysis, also an SPSS subroutine. The criterion measure, or grouping variable, for the discriminant function analysis was the identifica- i |tion of a subject as a nodule patient or a normal subject. ! Two approaches to this analysis were considered: The | jfirst entered all five factors as a group of "predictor" j i [variables into the analysis (Method-Direct); the second I l | [allowed for a stepwise procedure (Method-Wilks) by means of j i [which each factor (predictor) is judged by Wilks' lambda, a t jcriterion serving to judge each variable according to its j jability to maximize the differences among centroids of the j jgroups being analyzed. A group centroid is the group mean > score of the discriminant function computed. It is important to note here that interpretation of the discrim inant function derived is restricted in two ways: first, 33 the variables entered as predictors are themselves reduced "constructs" from the factor analysis; and second, only one discriminant function can be extracted from a two-group analysis. Data Analysis--Phase 2 | At this point data from Phase 2 subjects were analyzed. Factor scores for each subject were computed using the same formulae for Phase 1 subjects. Because of the small eigenvalues and canonical correlations in the Phase 1 analyses, it was decided to not attempt to predict Phase 2 group membership with respect to the Phase lj population. Instead, using only Phase 2 subjects the five j ! I I factor-predictors were computed and entered into j jDiscriminant Function analyses exactly like those for J [Phase 1. This would allow examination of the predictors' | i i \ jability to discriminate between groups (i.e., nodule or | formal) with respect only to the Phase 2 sample. ! i ! t ! i i i i i 34- CHAPTER IV RESULTS Phase 1 Examination of the final varimax rotated factor matrix i (Table C, Appendix B), and the item communalities and factor summary statistics In Table D, showed the presence of five reasonably clear, Interpretable factors. The first I Jfactor had an eigenvalue of 20.97284 and accounted for 59.3 tpercent of the total variance among the items. It was jhighly correlated (> 0.5, the customary criterion indicating | the item shares 25 percent or more variance in common with j I the factor) with 16 of the 19 Items concerned with i i 'experiencing debilitating conditions and with 7 of the 8 ; I I ! items (three items being debilitating condition items also) ! !concerned with coughing and throat-clearing. Two other ; items were also correlated with the factor but not in any [ t way that aided in interpretation. Since the eighth coughing1 ‘item and another condition item was just below the arbitrary! 0.5 correlation criterion, it seemed apparent that Factor 1 , could be described accurately as a "debilitating conditions/1 coughing and throat-clearing" factor. | Factor 2 had an eigenvalue of 4.93873 and accounted for 14.0 percent of the variance. It was correlated with all four of the items regarding continuing to shout and scream in the presence of any of the four debilitating conditions (hoarse, strained-tense, tired-weak, laryngitis). It was also correlated with five of the six items concerning the continuance of an activity while experiencing laryngitis. (One of these items was included in both groupings.) Only two other variables met the 0.5 criterion and offered no assistance to the interpretation. Factor 2 seemed to be a "continuing activity with debilitating condition" factor that was limited to "continuing to do any activity (except sing) in the presence of laryngitis" and to "continuing to shout and scream in the presence of any condition." , Though Factor 3 accounted for only 9.3 percent of the I variance and had an eigenvalue of 3.29262, it had an obvious structure. It was correlated highly--and only--with i ! all eight items about "singing." It was clearly a "singing"1 I factor. Similarly, Factor 4 accounted for only 7.4 percent ( of the variance, with a 2.62792 eigenvalue, but had a clear j ‘ i structure of all items related to "talking in noise." | ; ! jFactor 5 also had a clear structure being correlated only j i with seven of eight "yell or cheer" items. It had a very •low eigenvalue of 2.00003 and accounted for 5.7 percent of ; I the total variance. The sixth factor was correlated with no t item or set of items, and was clearly only picking up 'residual variance. Together, the five interpretable factors accounted for 95.7 percent of the total variance among the 49 items entered into the analysis. The clarity of the five' orthogonal factors and the total amount of explained variance they circumscribed allowed the next stage of the analysis to proceed. The results of the Phase 1, second step, first analysis (Method-Direct) considering the five factors as a predictive group are shown in Table E, Appendix B. The Wilks' lambda was large, 0.8009864, though it had an associated chi-square* of 11.428 (5 df) which was significant at p = 0.435. However, the eigenvalue was extremely small, 0.24846, and the canonical correlation was only 0.44. The canonical I correlation can be considered as the correlation between the i i I discriminant function and the grouping variable, and thus j I the square of the canonical correlation can be seen as the j proportion of variance explained by the function. A J jcanonical correlation of 0.44, then, accounted for only I i ! ■ 19.9 percent of the total variance. ; Examination of the standardized canonical discriminant 1 I i j function coefficients (.Table E, Appendix B) , which are j :analogous to beta weights in multiple regression, showed i [ i [that Factor 2 was of primary importance in defining the j i i [function (-0.82892), followed by Factor 1 (0.50252), with j k Factors 3, 4 and 5 contributing little (0.22675, -0.17558, and 0.22141, respectively). The absolute values of these 37 coefficients indicated the relative contribution of the predictor to the extracted function; the sign indicated merely whether the contribution was positive or negative. But, despite the small proportion of variance explained by the function, classification of cases was quite high (Table F, Appendix B). The prior probability for nodule patients was 0.571, but 75.0 percent of the nodule patients were correctly classified. The prior probability for normal subjects was 0.429, but 70.8 percent were correctly classified. Overall, 73.21 percent of the subjects were correctly assigned to their respective groups. This percentage may be somewhat misleading, inasmuch as the prediction was made for the same sample data from which the discriminant function was extracted. One would expect a [smaller "percent correctly classified" when using the [coefficients from these data upon a different sample. i j The Phase 1, second step, second analysis, the stepwise [procedure, resulted in only two of the five factor- i jpredictors being retained for classification purposes. As [would be expected from consideration of the first analysis, i the two variables were Factor 2 and Factor 1 (see Table G, ! (Appendix B). With just these two variables, the results were almost the same as with the first analysis. Wilks' lambda was 0.823896, though its associated chi-square of 10.267 (2 df) was significant at p = 0.0059. The eigenvalue was somewhat smaller, 0.21375, as is the canonical 38 correlation, 0.41965, which indicated that the function accounted for only 17.6 percent of the variance. But, still the function is able to classify cases with a surprising degree of accuracy, 73.21 percent (see Table H, Appendix B). Again, Factor 2 makes the primary contribution as indicated by the coefficient of 0.81715, while Factor 1 contributes a coefficient of -0.56953 to the function. i Phase 2 As can be seen from Table I, Appendix B, entering all five predictors as a group (Method-Direct) produced results comparable to the identical Phase 1 analysis. The Wilks' j I lambda was smaller, 0.76146, and its associated chi-square j i Iwas 13.490 (5 df), with a p = 0.0192 significance level. | t i jThe eigenvalue was higher, though still low, 0.31327, and | { I [the canonical correlation of 0.4884 accounted for 20.1 i j percent of the variance. Examination of the standardized , {canonical discriminant function coefficients showed a difference from the comparable Phase 1 analysis, however. j In this analysis, Factor 2 was still the primary contributor, i j 'but a little stronger (-1.00654, compared to -0.82892). , Factor 1, however, made virtually no contribution at all in I this analysis, after having been the second strongest contributor in the Phase 1 analysis (-0.00781, compared to ; 0.50252). As a consequence, Factors 3, 4 and 5 are somewhat1 stronger (0.27146, -0.28141, and 0.30843, respectively), and' retained the same respective contributions to the discriminant functions. Surprisingly, the ability of the function to correctly Jclassify cases is comparable to Phase 1 (see Table J, Appendix B), correctly classifying 65.5 percent of nodule cases, 88.0 percent of normal cases--75.93 percent overall. The second Phase 2 analysis, the stepwise procedure, retained only two predictors as fulfilling the criteria, but they were Factors 2 and 3, in contrast to Phase 1 where the retained predictors were Factors 2 and 1. Table K, Appendix B, shows a Wilks1 lambda of 0.78316, with an i associated chi-square of 12.466 (2 df), significant at the { p = 0.002 level. But, again, the eigenvalue was low, j | I i0.27688, and the canonical correlation of 0.46566 accounted 1 1 ! • for only 21.7 percent of the variance. The standardized I I • ! coefficients showed Factor 2 again to be the primary j contributing variable (0.91486), with Factor 3 at -0.3091. I i | Unlike the first analysis of Phase 2, the discriminant ; I i function was less able to correctly classify cases. Only 1 i 1 58.6 percent of the nodule patients were correctly identified, in contrast to 76.0 percent of the normal subjects. This resulted in an overall 66.67 percent 'correctly "grouped" cases (see Table L, Appendix B). 40 Post Hoc Analysis Further thought was given to. the unusual aspects of the results. Though the reductions in Wilks' lambda was statistically significant, the eigenvalues and canonical correlations were small. Yet, the extracted function was able to discriminate group membership relatively well. It seemed that the discriminatory power of the function was being inhibited somehow. It was conjectured that, by being subsumed within a created factor score, any single variable was less able to differentiate between groups. It was, therefore, decided to enter the 49 VBI items as predictor variables into a Discriminant analysis using a (stepwise procedure. This allowed each variable the j I opportunity to make a unique contribution to the discrimi- l inant function. Table M, Appendix B, shows the Summary Table I - i |for the 21 variables that met the minimum criteria necessaryj |to produce a statistically significant reduction in Wilks' j i j (lambda and were thus retained as predictors. j I Table N, Appendix B, contains the results concerning j ! ;the extracted function. The 21 variables retained produced a Wilks' lambda of 0.05539 which was highly significant. ; ■ I i The associated chi-square was 117.18 (21 df), significant atj p < 0.0001. The function then had an eigenvalue of j 17.05305 and an astonishing 0.97191 canonical correlation, , indicating the function was almost perfectly correlated withr the criterion grouping variable, that is, whether a subject was a nodule or a non-nodule respondent. This canonical correlation indicated that the function accounted for 94.46 percent of the total variance. As can be seen in Table 0, Appendix B, the function has high discriminatory power for Phase 1 subjects, correctly classifying 98.15 percent. By including the Phase 2 i subjects as "ungrouped cases" in the classification portion of the analysis, a comparison was able to be made between the predicted group membership with what was known to be actual group membership. Of the 54 Phase 2 subjects, 21 (72.4 percent) nodule patients were classified correctly as i were 21 (84.0 percent) normal subjects (Table P, Appendix B)j Overall, 42 of the 54 subjects, 77.8 percent, were correctly! I predicted to belong to their respective group. j Another way of looking at the data results is by ! ; I jexamining the percentage of correct predictions, rather than ■the percentage of people correctly classified (Table Q, | ! .Appendix B). Of the 25 predictions to be nodule patients, I : 21 were correct (84.0 percent), and 21 of the 29 predictions! to be normal subjects were correct (72.4 percent). These figures indicate strong support for the validity of the VBI as a measure of vocal abuse. 42 CHAPTER V DISCUSSION It had been intended to use the factor analysis to obtain factors which could be thought of as theoretical constructs of vocal abuse, that is, patterns of vocal abuse behavior and consequences. This was to have been done on a population of nodule patients. Because there was an insufficient number of nodule patients available for a j factor analysis, normal subjects were obtained and added to the sample to increase the N. By performing a factor ! analysis on these two groups of subjects, however, the j derived factors were less factors of vocal abuse than they j ! were of vocal use. Thus, while nodule patients scored in ; ways which might seem to be indicative of abuse, their j i j"scores" are more correctly seen as indications of greater 1 i jfrequencies of vocal use. Such "greater frequencies" may be 'abusive vocal use, but they should be seen as abusive levels' ! I lof the vocal usages comprising the extracted factors. Such j ia perspective is helpful with regard to understanding and interpreting the factors. ; Factor 1 was highly correlated with most of the "debilitating conditions" one could experience as a result of various vocal usages, and with the items regarding ; coughing and throat-clearing. Coughing and throat-clearing are viewed in the literature both as potentially abusive activities and as responses to laryngeal irritation. From the latter perspective coughing and throat-clearing may be viewed as "debilitating conditions" and their appearance in the conditions factor appears reasonable. Thus, Factor 1 could be seen as a "debilitating conditions" dimension that says the frequency of experiencing a condition after any one vocal activity is related to the frequency of experiencing a condition after any other vocal activity. Persons who, for example, experienced hoarseness after talking in noise, were more likely to experience feelings of vocal tension, strain, weakness and fatigue after any other vocal activity. That is, persons who experienced any one debilitating condition were more likely to experience all other debilitating J jconditions, and persons who did not experience any one I | | |debilitating condition were less likely to experience any j I | |other condition. , i Factor 2 was composed of "continuing to shout and ; i j l 'scream in the presence of conditions” items and "continuing ; jail activities (except singing) with laryngitis" items. It | I i [shows that levels of usage are related among these items. I i I |If a person will shout or scream when a condition is ' < present, it makes no difference what condition it is--he will continue to engage in the activity. As well, such a person will continue using his voice in all activities, except singing, even with laryngitis. It is of some interest, perhaps, that "continuing to sing with laryngitis" does not load oh the factor. Of the activities represented, singing is the only one which requires fine control of the voice, which is almost impossible with laryngitis. Factors 3, 4 and 5 were composed of all the items for each of three vocal activities, singing, talking in noise, jand yelling and cheering, respectively. For each of these, it indicated merely that the frequency of engaging in the j activity was related to the frequency of experiencing a j i 'debilitating condition and to the frequency of continuing toj jengage in the activity. For example, if a person reported j jthat he sang, experienced hoarseness after singing, and , i 'continued to sing while hoarse, his frequency range for each! J ! of these categories would be similar. That is to say, there! i ! |is a pattern to the usage. People experienced activities, : 'debilitating conditions and continuance of an activity with ! ' ‘ ? a condition in either a consistently low, or consistently j i middle, or consistently high range. ; Though there seems to be reasonable clarity to the ; interpretation of the factors resulting from the factor ' .analysis, it is important to realize that there may be limitations to extending this interpretation beyond this sample. While there was no evidence to indicate that the sample groups were somehow different from other members of their respective general populations, one still must restrict any interpretation to this sample. It should be kept in mind that the effective sample sized is only 53 Phase 1 subjects who had complete data sets, while the I desirable "minimum" for a factor analysis of 49 items would be approximately five hundred subjects. It is to be expected that another sample might well yield different i factors. It is the above point that may explain the inefficiency of the factor scores in the later discriminant function l analyses. Because the eigenvalues and canonical correla- j tions were small for the discriminant function analyses j i using factor scores as predictors, any attempt to interpret I the results would be inappropriate. In order to have any j • j !confidence in a function, one would need an eigenvalue j I ' !greater than 1.0, while the eigenvalues for the two derived j !functions were less than 0.25. These results, then, were I j jnot supportive of the constructs of vocal abuse suggested by; :the factor analysis despite the greater than chance correct : case classification by the function. This, perhaps, was i i best illustrated by the relationship of the factors for the 1 Phase 1 analysis as compared to the Phase 2 analysis. Though Factor 1 was the primary factor resulting from the i ' ■ factor analysis and was also the primary discriminator for ! 46; the Phase 1 discriminant analysis, it was of no use in differentiating between nodule and non-nodule subjects in Phase 2. Yet, the Phase 2 discriminant analysis correctly classified a substantial number of cases based on Factors 2 and 3, which constituted only about 25 percent of the variance among the VBI items (Table L, Appendix B). It seems, then, that differences do exist between nodule and j non-nodule subjects, but that the factors, as derived, are not capable of pointing to how they are different. This may well be due to the small sample size; a larger sample would provide more stable means and inter-item correlations, and i might produce a more definitive factor analysis. Another possibility is that the large number of items loading on Factor 1, which is a debilitating conditions factor, might 1 be found to load on a few more specific correlated factors. | JAn oblique factor analysis could be performed to investigate; I \ jthis possibility. Both of these questions remain for ; |resolution by future research. i : I 1 j When all 49 VBI items were entered into discriminant !analysis, the results obtained were significant. The first , i ~ | four items of the function were more than twice as important 'as the other 17 items. These first four items all :represented debilitating vocal conditions. This lends support to the impression that the large number of items in iFactor 1 were obscuring some of the discriminatory ability 'of some of the items. It also suggests that the construct of vocal abuse is composed first and foremost of the responses experienced as a result of a vocal activity. This sheds some light, perhaps, on the inability of other studies to isolate indicators of vocal abuse by looking only at the type or frequency of vocal activities performed by a person. What was insufficiently considered was the person's reaction to an activity. This lends support to the as yet untested theories that for some persons, the occurrence of "vocal abuse" may be influenced by physical characteristics and not by vocal activity alone. The VBI was developed to investigate a construct of vocal abuse. The results of the 49-item discriminant analysis indicated that it may serve as a useful diagnostic [ itool as well. The unique contribution of this instrument inj j determining the presence of debilitating vocal conditions, I i and the indications that these correlate highly with the | presence of "vocal abuse," would provide a useful tool for ! i detecting persons incurring vocal abuse. Differentiation of1 I i |vocal abusers from vocal non-abusers would allow for early : |intervention procedures which might eliminate the abuse aspects of their vocal behavior. This might prevent the | 'development of handicapping laryngeal pathology such as i 'vocal nodules. i | Additional issues suggested for further research include: I i 4 8 : 1. The differences between adults who have high vocal abuse scores and nodules, adults who may have high vocal abuse scores and no nodules, and adults who may have low vocal abuse scores and yet do have nodules need to be examined. Considerations such as genetic and constitutional characteristics, and subtle differences in laryngeal physiology and anatomy would need to be evaluated in such a s tudy. 2. Administration of the Vocal Behavior Inventory to a large sample of adults without vocal nodules to determine the range of scores obtained by "normal" individuals should jbe explored. Those normals who score on the high end of the J jnormal range might be examined medically for subtle j i |laryngeal changes which may signal the presence of early i (laryngeal pathology, such as vocal nodules. i | j 3. Vocal Behavior Inventory score profiles could be ! ! Jobtained on groups of individuals with contact ulcers and J 1 ' --V ’ f I ^polyps of the vocal cords. There is a suggestion in the j 1 literature, less strong than the assumed association between: i I jvocal abuse and nodules, that these disorders may be related ,in some way to vocal abuse. It would be important to see ifi I these people do score differently than vocal nodule subjects; 4. The Vocal Behavior Inventory might be used to ; investigate possible differences between subjects who reported a specific vocal trauma prior to the onset of their 49 voice disorder, and subjects who did not report such an incident. 5. Consideration should be given to modification of the Vocal Behavior Inventory for use with children. It would be important to examine whether the behavior related to vocal abuse in adults is also true for children. In summary, the construct validity of the Vocal Behavior Inventory was not supported by discriminant analysis using factor scores as predictors. Nevertheless, the ability of the derived functions to correctly classify cases according to their respective groups at a greater than chance percentage was suggestive. Discriminant analysis with all 49 VBI items entered as predictors resulted in a J ( I |function that was almost perfectly correlated with the i ]grouping variable and was capable of correctly classifying ! I j ja high percentage of both Phase 1 and Phase 2 subjects. If j • I ( i ithe potential of this instrument is to be realized through j J I jfactor analysis, further investigation with much larger j | j isubject samples will need to be conducted. i APPENDIX A VOCAL BEHAVIOR INVENTORY APPENDIX A VOCAL BEHAVIOR INVENTORY Directions This inventory examines different vocal activities, or ways in which people may use their voice. You are being asked to answer these questions because you have i been diagnosed by a doctor as having a voice problem. Your answers will provide valuable information to medical researchers studying this increasingly common condition. The inventory is divided into sections with questions that concern specific types of vocal activity, how often you may have engaged in that activity, and problems you may have j ! i experienced. j ! Please try very carefully to remember the period of time just i I ' :before you began to have, or notice, problems with your voice. Read i i each question keeping this period of time in mind. Then circle the ; ! i inumber of the phrase that best describes your opinion of your experience i I during this period. ; j Because it is essential that we know your answer to each question, ♦ 1 please circle the one number that best applies even if you have a i problem understanding any question. There is room at the end of the ; ’ form for your comments regarding such problems or anything else you wish' ! i to say. \ Copyright by Susan G. Gray 19 79 L . . 52, APPENDIX A (continued) You will find some questions may seem very similar. Please read each one carefully before you give an answer. Remember, you will be responding about the period of time before you began to have, or notice, problems with your voice. We appreciate your helping us to understand better how voice problems develop. AGE: ___________ SEX: | ! OCCUPATION: [PHYSICIAN: APPENDIX A (continued) ■ u o a C f l C L ) c d s •H g H 0) +J 0) < U 6 4-1 a O o a •H 4-i T3 c u EH i —1 w •U M-i a > c d 4-1 C U O cn & O -C o •U G -M 0) 'O 5 5 a ) >> a 1 —1 o 4J i —1 o 0) < 1 ) 0) i —! 5 2 5 > G O < o > <d A. Vocal Activity 1. During an average week, how often did you: a. Talk.......................................1 2 3 4 5 6 7 b. Talk in noise (for example, in crowded places, around machinery, in heavy traffic)..........1 2 3 4 5 6 7 c. Shout or s c r e a m ......................... 1 2 3 4 5 6 7 d. Yell or c h e e r ........................... 1 2 3 4 5 6 7 e. Sing...................................... 1 2 3.4 5 6 7 f. Cough or clear your throat 1 2 3 4 5 6 7 2. If you sang: a. Did you sing in your upper range, or use high pitches?.................... 1 2 3 4 5 6 7 b. How is your voice classified? ________________________________ c. Did you sing professionally?............1 2 3 4 5 6 7 B. Vocal Conditions 1, On the average, how often did you feel or sound hoarse after you had been: a. T alking..................................1 2 3 4 5 6 7 b. Talking in noise (for example, in crowded places, around machinery, in heavy traffic) ................ 1 2 3 4 5 6 7 c. Shouting or screaming ........... . , 1 2 3 4 5 6 7 ■ d. Yelling or cheering ............. . . 1 2 3 4 5 6 7 ' e. Singing ........................... 1 2 3 4 5 6 7 f. Coughing or clearing your throat . . . 1 2 3 4 5 6 7 ' 54 APPENDIX A (continued) o is t o 0 ) c t i a •H P S H C D 4- 4 c u C U 6 < 4 - 1 S .C o O c s •ri 4 - 1 T3 rH 03 0 ) 4 -1 E - 4 < 4 - 1 < U c O < 4 4 0) o C O e o vd o 44 a 4 -4 0 ) T) 0 C U e i— 3 o 44 u ! — 1 o C L ) 0 ) < 4 4 Q J I— i £5 > C O <1 o > c 2. On the average, how often did your throat feel tense or strained after you had been: a. Talking ........................... ... 1 2 3 4 5 6 7 b. Talking in noise ................ ... 1 2 3 4 5 6 7 c. Shouting or screaming ........... ... 1 2 3 4 5 6 7 d. Yelling or cheering .............. ... 1 2 3 4 5 6 7 e. Singing . ......................... ... 1 2 3 4 5 6 7 f. Coughing or clearing your throat ... 1 2 3 4 5 6 7 3. On the average, how often did your throat feel tired or weak after you had been: a. Talking ........................... ... 1 2 3 4 5 6 7 b. Talking in noise ................ ... 1 2 3 4 5 6 7 c. Shouting or screaming ........... 2 3 4 5 6 7 d. Yelling or cheering .............. ... 1 2 3 4 5 6 7 e. Singing ........................... ... 1 2 3 4 5 6 7 f. Coughing or clearing your throat ... 1 2 3 4 5 6 7 4. On the average, how often did you have laryngitis?.................................... 1 2 3 4 5 6 7 C. Vocal Activity with Vocal Condition 1. When you felt or sounded hoarse, how often did you continue: a. Talking ......................... .... 1 2 3 4 5 6 7 b. Talking in noise .............. .... 1 2 3 4 5 6 7 c. Shouting or screaming ......... .... 1 2 3 4 5 6 7 d. Yelling or cheering . ......... .... 1 2 3 4 5 6 7 55 APPENDIX A (continued) o E 3 m < u c f l .H C O -U m O J a s m < u o o o s o jC o JJ p ! J-i qj > , t 3 3 ai c S - I rH O +j M iH o aj a j m h is > c o < d Q > < e. Singing ................................ 1 2 3 4 5 6 7 f. Coughing or clearing your throat . . 1 2 3 4 5 6 7 If your throat felt tense or strained, how often did you continue: a. Talking ................................ 1 2 3 4 5 6 7 b. Talking in noise .................... 1 2 3 4 5 6 7 c. Shouting or screaming ................ 1 2 3 4 5 6 7 d. Yelling or cheering .................. 1 2 3 4 5 6 7 e. Singing ......... .................... 1 2 3 4 5 6 7 f. Coughing or clearing your throat . . 1 2 3 4 5 6 7 If your throat felt tired or weak, how often did you continue: a. Talking ................................ 1 2 3 4 5 6 7 b. Talking in noise .................... 1 2 3 4 5 6 7 c. Shouting or screaming ................ 1 2 3 4 5 6 7 d. Yelling or cheering .................. 1 2 3 4 5 6 7 e. Singing ................................ 1 2 3 4 5 6 7 f. Coughing or clearing your throat . . 1 2 3 4 5 6 7 If you had laryngitis, how often did you continue: a. Talking ................................ 1 2 3 4 5 6 7 b. Talking in noise .................... 1 2 3 4 5 6 7 c. Shouting or screaming ................ 1 2 3 4 5 6 7 d. Yelling or cheering .................. 1 2 3 4 5 6 7 56 APPENDIX A (continued) ■u o S3 C O C U n ) B *H G H 0 ) 4 - 1 0 1 C U a m a JC o o G •H ■U XI C U EH t—1 t o 4J > 4 - 1 a ) n j M - 4 C U o ca a o rC o + j G c u X) 3 0 J > - > c V I r _ j O 4J V i T “I o 0 1 Q ) uQ < 4 - 1 a ) rH 13 > CO < o > < 1 2 3 4 5 6 7 1 2 3 4 5 6 7 e. Singing ................................ f. Coughing or clearing your throat . • D. Comparison 1. In comparison with the period before you developed a voice problem, that is, the period about which you have just answered questions, was there a time prior to that period when your voice was best, that is, stronger, clearer, smoother, and more effortless to produce? No (If you answered No_, please turn to page , Section E) Yes (If you answered Yes, please respond to the following questions): a. When was the last time you can remember your voice being best, that is, stronger, clearer, smoother, and effortless to produce?______________________________________________________ ! b. Comparing when my voice was best with the period before my j voice problem started: (check one per activity) More Often Same Less Often' a. I talked b . I talked in noise c. I shouted or screamed d. I yelled or cheered e. I sang f . I coughed or cleared my throat [ 57' APPENDIX A (continued) c. Comparing when my voice was best with the period before my voice problem started, I was hoarse: (check one per activity) More Often Same Less Often a. after talking ____ ____ ____ b. after talking in noise ____ ____ ____ c. after shouting or screaming ____ ____ ____ d. after yelling or cheering ____ ____ ____ e. after singing ____ ____ ____ f. after coughing or clearing my throat__________________________ ____ ____ d. Comparing when my voice was best with the period before my voice problem started, I had feelings of tension and strain in my throat: (check one per activity) More Often Same Less Often a. after talking ____ ____ ____ b. after talking in noise ____ ____ ____ c. after shouting or screaming ____ j d. after yelling or cheering ____ ' e. after singing j f. after coughing or clearing I my throat__________________________ ____ ____ e. Comparing when my voice was best with the period before my voice problem started, when my voice was best I had feelings; of weakness and fatigue in my throat: (check one per . activity) More Often Same Less Often, a. after talking ’ b. after talking in noise ____ ____ ____ c. after shouting and screaming ____ ! d. after yelling and cheering ____ _ e. after singing ____ ____ ____ f. after coughing or clearing my throat ____ ____ ____ 58 APPENDIX A (continued) f. Comparing when my voice was best with the period before I developed a voice problem, I had laryngitis: (check one) More Often Same Less Often E. Commen ts 1. Can you remember any time when you over-used or strained your voice just before you noticed a voice problem? No ____ Yes ____ If you answered yes, briefly describe what happened: f j 2 . Do you feel that any of the voice activities mentioned have j contributed to your voice problem? j No ____ J Yes ____ j If you answered yes, please list these activities: * In the parentheses at the right, please rank them from the most i important (#1) to the least important, as you feel they relate j to your voice problem. 3. About how long has it been since you first noticed a problem with your voice? (onset date) ___________________________________ 4. About how long was the period between the time you first noticed a problem with your throat (or voice) and the time you saw a doctor about the problem? __________________________________ I 59 APPENDIX A (continued) Did you have a problem understanding any question? If so, which ones and why? Please explain. Do you have any comments regarding this inventory? Please explain. ______________________________________ APPENDIX A (continued) PHYSICIAN'S FORM Patient's Name: Date of First Visit with you: Had patient received treatment elsewhere before coming to you? Voice Problem: Onset Date of voice problem: Date of your diagnosis of vocal nodules: Description of nodule(s): (Rate severity of problem: 1 2 I Not I Serious Comments you may care to make about case: 5 Very Serious Physician 61 APPENDIX TABLES APPENDIX B Table A Variable Names and Labels of the 49 VBI Items Variables Labels Al-A Activity: Talk Al-B Activity: Talk in noise Al-C Activity: Shout or scream Al-D Activity: Yell or cheer Al-E Activity: Sing Al-F Activity: Cough or clear throat Bl-A Condition: Hoarse: Talk Bl-B Condition: Hoarse: Talk in noise Bl-C Condition: Hoarse: Shout or scream Bl-D Condition: Hoarse: Yell or cheer Bl-E Condition: Hoarse: Sing Bl-F Condition: Hoarse: Cough/clear throat B2-A Condition: Tense: Talk B2-B Condition: Tense: Talk in noise B2-C Condition: Tense: Shout or scream B2-D Condition: Tense: Yell or cheer B2-E Condition: Tense: Sing B2-F Condition: Tense: Cough or clear throat B3-A Condition: Tired: Talk B3-B Condition: Tired: Talk in noise B3-C Condition: Tired: Shout or scream B3-D Condition: Tired: Yell or cheer B3-E Condition: Tired: Sing B3-F Condition: Tired: Cough or clear throat B4 Condition: Frequently laryngitis Cl-A Activity with condition: Hoarse: Talk Cl-B Activity with condition: Hoarse: Talk in noise Cl-C Activity with condition: Hoarse: Shout or scream Cl-D Activity with condition: Hoarse: Yell or cheer Cl-E Activity with condition: Hoarse: Sing Cl-F Activity with condition: Cough/clear throat C2-A Activity with condition: Tense: Talk C2-B Activity with condition: Tense: Talk in noise C2-C Activity with condition: Tense: Shout or scream C2-D Activity, with condition: Tense: Yell or cheer C2-E Activity with condition: Tense: Sing C2-F Activity with condition: Tense: Cough/clear throat C3-A Activity with. condition: Tired: Talk C3-B Activity with Condition: Tired: Talk in noise 63 APPENDIX B (continued) Variables Labels C3-C Activity with condition: Tired: Shout or scream C3-D Activity with condition: Tired: Yell or cheer C3-E Activity with condition: Tired: Sing C3-F Activity with condition: Tired: Cough/clear throat C4-A Activity with condition: Laryngitis: Talk C4-B Activity with condition: Laryngitis: Talk in noise C4-C Activity with condition: Laryngitis: Shout or scream C4-D Activity with condition: Laryngitis: Yell or cheer C4-E Activity with condition: Laryngitis: Sing C4-F Activity with condition: Laryngitis: Cough/clear throat i j ] | j i l ! j ! i i I i 64 APPENDIX B (continued) Table B Phase 1. Final Factor Analysis: All Factor Score Coefficients iriables Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 Factor 6 Al-A -0.31367 0.26333 -0.82338 1.05119 -0.57524 -0.25267 Al-B 0.74167 -1.29884 0.37610 -1.70588 1.53639 -0.27842 Al-C 3.00159 -4.28706 -3.39108 -4.34378 3.34223 -0.39468 Al-D -3.14321 3.66399 1.02625 5.62564 -5.04794 1.09854 Al-E 3.50638 -3.80814 0.44001 -6.70587 5.34392 -1.60057 Al-F -0.93083 1.59095 0.14072 1.62144 -0.97433 0.23285 Bl-A 15.51749 -23.67314 -9.95096 -26.10304 24.23737 -3.24899 Bl-B 0.25782 1.77730 -1.46282 2.67171 -1.93033 0.61146 Bl-C -12.09125 14.11546 5.64262 21.26346 -19.30885 -2.32919 Bl-D 1.86893 -1.10562 0.17356 -4.81490 4.04187 3.27855 Bl-E -15.49654 28.02980 19.67070 20.00414 -22.43169 8.89563 Bl-F 10.48976 -19.85014 -14.78581 -13.40844 15.37873 -3.48105 B2-A -6.90337 11.30775 11.49188 9.13580 -12.54323 2.67137 B2-B -6.65634 7.00726 -2.68968 14.46620 -9.59264 -2.11726 B2-C 19.98872 -30.14223 -8.71097 -36.54652 31.58434 0.33578 B2-D 7.12506 -15.43534 -4.84139 -11.99336 11.57602 -6.69409 B2-E 13.65167 -26.34698 -17.16664 -20.27278 19.74008 -5.36792 B2-F -1.22328 2.88933 1.73962 0.37576 -1.91544 2.27337 B3-A -7.82066 11.53738 2.33589 13.67092 -10.04830 -0.81107 B3-B 12.76520 -22.94325 -1.46571 -27.38440 19.41574 2.53320 B3-C -13.11653 25.07068 4.74954 27.74146 -21.84294 -0.18988 B3-D -6.13165 13.02993 4.12008 8.21935 -8.26683 4.34135 B3-E -2.62856 5.18104 3.47203 3.75107 -3.45217 2.58186 B3-F -5.55521 14.33293 6.61050 10.60690 -8.52458 -0.34521 B4 -3.12321 4.36995 1.55227 5.09940 -3.79729 -0.12580 Cl-A -6.67074 10.45695 3.09547 10.64607 -10.46166 2.67227 Cl—B -7.19119 11.74364 9.30208 11.50449 -8.45866 0.55629 Cl-C 3.81061 -6.53117 -4.93599 -5.24586 5.67795 -0.76341 Cl-D -6.32704 9.60497 3.03788 10.61599 -6.45232 0.70171 Cl-E 9.90764 -17.00539 -13.33073 -12.38849 13.14043 -4.36913 Cl—F -7.76660 12.80042 8.64868 10.82222 -10.95237 1.43713 C2-A 3.42332 -5.76523 -2.80052 -4.89979 5.86158 -1.95824 C2-B 1.63862 -0.79944 1.02635 -4.14388 1.84979 1.44720 C2-C -2.41783 9.10208 5.01261 4.06521 -3.10835 1.18018 C2-D 0.49704 1.40366 -1.88374 -0.07264 -1.94310 2.64382 C2-E -56.52026 77.21338 31.64388 100.56808 -90.56830 4.42285 C2-F 4.32591 -11.40770 -5.61920 -8.41556 7.54184 -1.67007 65 APPENDIX B (continued) Variables Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 Factor 6 C3-A 1.90871 -2.29160 -0.66729 -1.86373 1.76373 -0.35799 C3-B -1.74995 4.66515 -1.99048 3.24194 -3.18975 1.51505 C3-C 5.63185 -14.48886 -4.13836 -12.05145 9.70674 -0.12454 C3-D -2.59087 4.23726 2.37904 5.25151 -4.07395 -2.57628 C3-E 50.82275 -70.21558 -24.54770 -92.86610 83.80995 -4.54951 C3-F 2.10378 -2.87755 -0.55838 -1.02075 0.09519 -0.94826 C4-A -3.06999 4.04912 5.38397 4.38091 -4.74378 -1.24683 C4-B 7.05013 -11.07470 -7.59992 -8.63659 3.96155 -3.58058 C4-C -3.90021 7.05235 4.45204 1.89415 -3.06765 4.64699 C4-D 4.36282 -9.97897 -3.85123 -4.55016 4.89224 -3.97612 C4-E -4.10256 7.25180 0.66948 8.66495 -6.58806 1.09075 C4-F 3.31821 -3.78955 -3.08864 -6.70221 4.79003 2.01526 APPENDIX B Table C Phase 1. Final Varimax Rotated Factor Matrix (N = 53) ariables Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 Factor 6 Al-A -0.05014 0.02017 0.20591 0.43590 0.15964 -0.13262 Al-B 0.06390 -0.00402 0.20265 0.65028 0.11390 0.03488 Al-C 0.11309 0.29014 0.24685 0.42082 0.21078 0.21657 Al-D -0.05795 0.11628 0.09282 0.02664 0.69722 0.01929 Al-E 0.12416 -0.00535 0.76857 0.12865 0.11955 0.16529 Al-F 0.68351 -0.02235 0.17847 0.12365 0.09293 -0.07026 Bl-A 0.57772 0.16641 0.24165 0.21084 0.08095 0.53794 Bl-B 0.53622 -0.05053 0.13915 0.55907 -0.00915 0.44621 Bl-C 0.57669 0.17871 0.42845 0.07318 0.16653 0.34861 Bl-D 0.35459 0.01810 0.21259 0.04423 0.85931 0.20438 Bl-E 0.39680 0.11548 0.75911 0.11491 -0.02555 0.36472 Bl-F 0.74993 0.21701 0.19409 0.06596 -0.10267 0.20373 B2-A 0.70185 0.10723 0.25327 0.16007 0.18597 0.26159 B2-B 0.69307 -0.06233 0.21053 0.53685 -0.00495 0.12223 B2-C 0.69715 0.22130 0.41404 0.04879 0.31537 0.14637 B2-D 0.44553 -0.00611 0.27246 0.10497 0.71970 0.03791 B2-E 0.56645 0.03495 0.63669 0.12147 0.24759 0.02666 B2-F 0.86873 0.23365 0.21200 0.11235 -0.02520 0.06673 B3-A 0.78865 0.12616 0.22292 0.07298 0.07322 0.11051 B3-B 0.77656 -0.05903 0.16404 0.38287 0.05428 -0.03570 B3-C 0.79649 0.17167 0.32818 -0.07609 0.20804 0.06469 B3-D 0.61515 0.04539 0.27969 0.03192 0.57206 -0.02554 B3-E 0.59736 0.00997 0.64468 0.04735 0.16583 -0.01217 B3-F 0.88875 0.21964 0.20441 0.12815 0.04722 -0.04067 B4 0.58399 0.27289 0.08681 0.12735 0.00424 0.21658 Cl-A 0.34974 0.27885 0.22033 0.43883 0.23368 0.32235 Cl—B 0.23647 0.29479 0.13214 0.79183 0.01910 0.16773 Cl-C 0.20072 0.58151 0.19288 0.31235 0.01386 0.11004 Cl—D -0.13788 0.45891 0.04770 0.21979 0.76533 9.00576 Cl-E 0.26322 0.17121 0.74164 0.21875 0.10276 0.09771 Cl-F 0.75917 0.33734 0.12756 0.21466 -0.00720 -0.12522 C2-A 0.40834 0.27404 0.08007 0.42716 0.21699 0.10861 C2-B 0.26401 0.30013 -0.04517 0.72313 -0.00157 -0.06939 C2-C 0.21094 0.75968 0.12680 0.23799 0.16073 -0.22277 C2-D -0.05969 0.53558 0.06440 0.22214 0.58371 -0.18912 C2-E 0.40111 0.18644 0.73684 0.23281 0.19512 -0.22497 C2-F 0.65249 0.50455 0.10098 0.28495 0.03251 -0.03720 C3-A 0.64900 0.30666 0.17467 0.19547 0.25181 -0.00465 C3-B 0.53046 0.24994 0.04309 0.65883 0.02762 -0.17956 67 APPENDIX B (continued) Variables Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 Factor 6 C3-C 0.25796 0.61714 0.09588 0.09358 0.24669 -0.24450 C3-D 0.18067 0.48742 -0.02062 0.10774 0.54351 -0.31406 C3-E 0.42623 0.15419 0.75326 0.28572 0.17607 -0.24049 C3-F 0.79947 0.34761 0.12525 0.23337 0.08117 -0.17727 C4-A 0.22613 0.79137 0.10466 0.15509 0.07387 0.18053 C4-B 0.03035 0.64195 0.00398 0.45368 0.01658 0.02035 C4-C 0.08488 0.89374 0.06376 0.02779 0.08827 0.15797 C4-D 0.15280 0.78893 0.07326 -0.10211 0.22253 0.02555 C4-E 0.37589 0.28636 0.54829 -0.00758 0.05394 -0.03474 C4-F 0.48598 0.73295 0.10653 0.04594 -0.00982 0.14977 I i APPENDIX B (continued) Table D Phase 1. Final Factor Analysis: A Summary of Factor Statistics and Item Communalities Percentage Cumulative Factor Eigenvalue of Variables Percentage 1 20.97284 59.3 59.3 2 4.93873 14.0 73.3 3 3.29262 9.3 82.6 4 2.62792 7.4 90.0 5 2.00003 5.7 95.7 6 1.51881 4.3 100.0 } Variable Communality Variable Communality Al-A 0.32444 Cl-A 0.59972 Al-B 0.48223 Cl-B 0.81578 Al-C 0.42632 Cl-C 0.52551 Al-D 0.51270 Cl-D 0.86048 Al-E 0.64843 Cl-E 0.71658 Al-F 0.52840 Cl-F 0.76822 Bl-A 0.76023 C2-A 0.48960 Bl-B 0.82120 C2-B 0.68956 Bl-C 0.70270 C2-C 0.76979 Bl-D 0.95341 C2-D 0.72039 Bl-E 0.89391 C2-E 0.93212 Bl-F 0.70357 C2-F 0.77416 B2-A 0.60688 C3-A 0.64779 B2-B 0.83174 C3-B 0.81277 B2-C 0.82970 C3-C 0.58599 B2-D 0.80320 C3-D 0.67630 B2-E 0.80424 C3-E 0.94332 B2-F 0.87217 C3-F 0.86815 B3-A 0.71048 C4-A 0.75046 B3-B 0.78425 C4-B 0.61956 B3-C 0.82482 C4-C 0.84356 B3-D 0.78762 C4-D 0.71173 B3-E 0.80244 C4-E 0.52809 B3-F 0.90022 C4-F 0.80948 B4 0.48620 69 APPENDIX B (continued) Table E Phase 1. Discriminant Analysis of Factor-Predictors (Method-Direct) Function Eigenvalue Canonical Wilks' Correlation Lambda chi-squared DF Significance 1* 0.24846 0.4461094 0.8009864 11.428 5 0.0435 Standardized Canonical Discriminant Function Coefficients Factors Function 1 1 2 3 4 5 i *Marks the 1 canonical discriminant function to be used in the remaining analysis. i i 1 Table F ] Phase 1. Case Classification Results | of Discriminant Analysis of j Factor-Predictors (Method-Direct) I _________________________________________________________________________ ) Predicted Group Membership Number of — — — ..— ——----------- Actual Group Cases 1 2 Group 1 32 24 8 Nodules Phase 1 75.0% 25.0% Group 2 24 7 17 Normals Phase 1 29.2% 70.8% Note. Percent of "grouped" cases correctly classified: 73.21%. _______________________ 7CL 0.50252 0.82892 0.22675 0.17558 0.22141 APPENDIX B (continued) Table G Phase 1. Discriminant Analysis of Factor-Predictors (Method-Wilks) Canonical Discriminant Functions Canonical Wilks' Function Eigenvalue Correlation Lambda chi-squared DF Significance 1* 0.21375 0.4196474 0.8238961 10.267 0.0059 Standardized Canonical Discriminant Function Coefficients Factors Function 1 1 2 -0.569-53 0.81715 *Marks the 1 canonical discriminant function to be used in the remaining analysis. I Table H Phase 1. Case Classification Results of Discriminant Analysis of Factor-Predictors (Method-Wilks) Predicted Group Membership Actual Group Number of Cases Group 1 Nodules Phase 1 Group 2 Normals Phase 1 32 24 24 75.0% 7 : 29.2% 8 25.0% 17 70.8% Note. Percent of "grouped" cases correctly classified: 73.21%, 71 APPENDIX B (continued) Table I Phase 2. Discriminant Analysis of Factor-Predictors (Method-Direct) Function Eigenvalue Canonical Wilks' Correlation Lambda chi-squared DF Significance 1* 0.31327 0.4884083 0.7614574 13.490 5 0.0192 Standardized Canonical Discriminant Function Coefficients Factors Function 1 1 -0.00781 2 -1.00654 3 0.27146 4 -0.28141 5 0.30843 *Marks the 1 canonical discriminant function to be used in the iremaining analysis. i I I Table J Phase 2. Case Classification Results of Discriminant Analysis of Factor-Predictors (Method-Direct) i Number of Cases Predicted Group Membership Actual Group 1 2 Group 4 29 19 ■ 10 Nodules Phase 2 65.5% 34.5% Group 5 25 3 22 Normals Phase 2 12.0% 88.0% Note. Percent of "grouped" cases correctly classified: 75.93%. 72 APPENDIX B (continued) Table K Phase 2. Discriminant Analysis of Factor-Predictors (Method-Wilks) Canonical Wilks' Function Eigenvalue Correlation Lambda chi-squared DF Significance 1* 0.27688 0.4656645 0.7831566 12.466 2 0.0020 Standardized Canonical Discriminant Function Coefficients Factors Function 1 2 0.91486 3 -0.30910 *Marks the 1 canonical discriminant function to be used in the remaining analysis. Table L < 1 Phase 2. Case Classification Results of Discriminant I Analysis of Factor-Predictors (Method-Wilks) j Predicted Group Membership j Number of ----------------------------- j Actual Group Cases 4 5 ( _______________________ I Group 4 29 17 12 Nodules Phase 2 58.6% 41.4% Group 5 25 6 19 Normals Phase 2 24.0% 76.0% Note. Percent of "grouped" cases correctly classified: 66.67%. APPENDIX B (continued) "j Phase 2. Table M Summary Table of Discriminant Analysis of 49 VBI Items Step Action Entered Removed Variables In. Wilks' Lambda Significance Label 1 Bl-A 1 0.434422 0.0 Condition: Hoarse: Talk 2 B3-E 2 0.292826 0.0 Condition: Tired: Sing 3 C4-A 3 0.239063 0.0000 Activity with condition: Laryngitis: Talk 4 C3-F 4 0.200613 0.0000 Activity with condition: Tired: Cough/clear 5 Al-A 5 0.188880 0.0000 throat Activity: Talk 6 C4-F 6 0.176106 0.0000 Activity with condition: Laryngitis: Sing 7 C2-F 7 0.165850 0.0000 Activity with condition: Tense: Cough/ clear 8 B3-A 8 0.151596 0.0000 throat Condition: Tired: Talk 9 C4-D 9 0.141672 0.0000 Activity with condition: Laryngitis: Yell or 10 C3-A 10 0.130712 0.0000 cheer Activity with condition: Tired: Talk 11 C2-B 11 0.116113 0.0000 Activity with condition: Tense: Talk in 12 C4-B 12 0.105760 0.0000 noise Activity with condition: Laryngitis: Talk in 13 Cl-A 13 0.097683 0.0000 noise Activity with condition: Hoarse: Talk 14 Al-C 14 0.087899 0.0000 Activity: Shout or scream 15 Cl-C 15 0.082931 0.0000 Activity with condition: Hoarse: Shout or 16 B3-A 14 0.084122 0.0000 scream Condition: Tired: Talk APPENDIX B (continued) Action Variables Wilks' Step Entered Removed In. Lambda Significance Label 17 C4-A 13 0.085389 0.0000 Activity with condition: Laryngitis: Talk 18 Al-A 12 0.086975 0.0000 Activity: Talk 19 B2-E 13 0.082300 0.0000 Condition: Tense: Sing 20 Bl-B 14 0.078819 0.0000 Condition: Hoarse: Talk in noise 21 B3-F 15 0.075308 0.0000 Condition: Tired: Cough or clear throat 22 C3-C 16 0.072507 0.0000 Activity with condition: Tired: Shout or scream 23 C2-C 17 0.067909 0.0000 Activity with condition: Tense: Shout or scream 24 Al-E 18 0.065679 0.0000 Activity: Sing 25 C3-E 17 0.066424 0.0000 Activity with condition: Laryngitis: Sing 26 Cl-D 18 0.063643 0.0000 Activity with condition: Hoarse: Yell or scream 27 Bl-E 19 0.061148 0.0000 Condition: Hoarse: Sing 28 C4-C 20 0.059142 0.0000 Activity with condition: Laryngitis: Shout or scream 29 B2-A 21 0.056757 0.0000 Condition: Tense: Talk 30 B3-F 20 0.058095 0.0000 Condition: Tired: Cough or clear throat 31 C4-E 21 0.055392 0.0000 Activity with condition: Laryngitis: Sing APPENDIX B (continued) Table N Phase 2. Results of Discriminant Analysis of 49 VBI Items Classification Function Coefficients (Fisher's Linear Discriminant Functions) 1 2 Nodules Normals Patient-Group Phase 1 Phase 1 Al-C 7.356250 1.8818260 Al-E -2.449459 1.2855610 Bl-A 7.051602 2.2704880 Bl-B -2.498340 -0.4555918 Bl-E 5.478228 -0.6525979 B2-A 4.718330 0.6912881 B2-E -11.621040 -1.1240000 B3-E 19.407970 2.4784190 Cl-A -2.053965 -0.2004012 Cl-C -4.280266 -0.9009125 Cl-D 1.067568 -0.2911567 C2-B -2.870588 -0.1795353 C2-C 5.652195 0.4639186 C2-F 16.020190 1.4866280 C3-A 7.149727 0.5342258 C3-C -6.315562 -0.2666366 C3-F -19.052300 -1.8837950 C3-B 5.987922 1.5522310 C4-C 2.648609 -0.2985017 C4-D -7.623087 0.2792118 C4-E -2.697072 -0.8019723 (constant) -54.802430 -6.7546300 Canonical Discriminant Functions I Canonical Wilks' 1 I I jFunction Eigenvalue Correlation Lambda chi-squared DF Significance i | 1* 17.05305 0.9719093 0.0553923 117.18 21 0.0000 *Marks the 1 canonical discriminant function to be used in the remaining analysis. 76 APPENDIX B (continued) Table 0 Phase 2. Case Classification Results for Discriminant Analysis of 49 VBI Items Predicted Group Membership Number of ----------------------- Actual Group Cases 1 2 Group 1 31 31 0 Nodules Phase 1 100.0% 0.0% Group 2 23 1 22 Normals Phase 1 4.3% 95.7% Note. Percent of "grouped" cases correctly classified: 98.15%. Table P Case Classification Results for Phase 2 Subjects Based Upon Discriminant Analysis of 49 VBI Items for Phase 1 Subjects Predicted j 1 Nodules Normals 1 Nodules 21 8 29 72.4% 27.6% 100% Normals 4 21 25 16.0% 84.0% 100% 77 APPENDIX B (continued) Table Q Phase 2. Percent of Correct and Incorrect Predictions Compared to Actual Group Membership Predictions to be Nodule Predictions to be Normal Correct Incorrect Correct Incorrect Nodule 21 8 (N = 29) 84% 27.6% Normal 4 21 (N = 25) 16% 72.4% 25 29 100% 100% REFERENCES Aronson, A. Psychogenic voice disorders: An interdisciplinary approach to detection, diagnosis and therapy"! Philadelphia: W. B. Saunders Company, 19 73. Arnold, G. E. Vocal nodules and polyps: Laryngeal tissue reaction to habitual hyperkinetic dysphonia. Journal of Speech and Hearing Disorders, 1962, 2_7, 205-217. Ash, J. E., & Schwartz, L. The laryngeal (vocal cord) node. Transactions of the American Academy of Opthalmology and Otolaryngology, 1943-44, 48, 323-332. Ballenger, J. J. Diseases of the nose, throat and ear. Philadelphia: Lea and Febiger, 1969. Berdie, D. R., 6c Anderson, J. F. Questionnaires: Design and use. Metuchen, N.J.: The Scarecrow Press, Inc., T977T Brewer, D. W. Early diagnostic signs and symptoms of laryngeal disease. Laryngoscope, 1975, 85_, 499-515. Brodnitz, F. S. Vocal rehabilitation. Rochester: American! Academy of Opthalmology and Otolaryngology, 1971. j Brodnitz, F. S., 6c Froeschels, E. Treatment of nodules of vocal cords by chewing method. Archives of Otolargyngology, 1954, .59, 560-565. j I i |Brown Kelly, H. D. , 6c Craik, J. E. Laryngeal nodes and the * so-called amyloid tumor of the cords. The Journal of I I Laryngology and Otology, 1952, 66, 339-358. j I (Cohen, J., 6c Cohen, P. Applied multiple regression/ | ■ correlation analysis for the behavior sciences. New York:1 j John Wiley and Sons, 1975. I (Cooper, M. Modern techniques of vocal rehabilitation for j i functional and organic dysphonias. In L. E. Travis (Ed.), 1 Handbook of speech pathology and audiology. New York: j Appleton-Century-Crofts, 1971. j I iCooper, M. Modern techniques of vocal rehabilitation. [ [ Springfield: Charles C. Thomas, Publisher, 1973. j I i (DeWeese, D. D. , 6c Saunders, W. H. Textbook of ] otolaryngology (2nd ed.). Saint Louis: The C. V. Mosby Company, 1964. Ferguson, G. B. Organic lesions of the larynx produced by mis-use of the voice. Laryngoscope, 1955, 6_5, 327-336. 80. Froeschels, E. Hygiene of the voice. Archives of Otolaryngology, 1943, _38, 122-130. Greene, M. C. L. The voice and its disorders (3rd ed.). Philadelphia: T~. W. Lippencott, Company, 1972. Harris, H. H. Benign lesion of the true vocal cords. Annals of Otology, Rhinology and Laryngology, 1948, 57, 189-196. Horst, P. Matrix algebra for social scientists. New York: Holt, Rinehart and Winston, Inc., 1963. Jackson Chevalier, & Chevalier Jackson. Diseases of the nose throat and ear (2nd ed.). Philadelphia: W. B7 Saunders Company, 1959. Kerlinger, F. Foundations of behavioral research (2nd ed.) New York: Holt, Rinehart and Winston, Inc., 1973. Laguaite, J. K. Adult voice screening. Journal of Speech and Hearing Disorders, 1972, 3_7, 147-151. Laguaite, J. K. Personal communication, November 10, 1976. iLuchsinger, R., & Arnold, G. E. Voice-speech-language. ! Clinical communicology: Its physiology and pathology, j Belmont: Wadsworth Publishing Company, Inc., 1965. [New, G. B., & Erich, J. B. Benign tumors of the larynx. Archives of Otolaryngology, 1938, 28y 841-910. I Nie, N. H. et al. Statistical package for the social ! sciences (2nd ed.). New York: McGraw Hill Book Company,' 1970” , lOppenheim, A. N. Questionnaire design and attitude ; ! measurement. New York: Basic Books, Inc., Publisher, j -------- I ‘ Perkins, W. H. Speech pathology--An applied behavioral ! science. Saint Louis: The C. V. Mosby Company, 1971. (a)> i Perkins, W. H. Vocal function: Assessment and therapy. In! L. E. Travis (Ed.), Handbook of speech pathology and ! audio logy. New York"! Appleton-Century-Crof ts , 1971. (b)| Perkins, W. H. Speech pathology— An applied behavioral ! science.(2nd ed.). Saint Louis: The C. V. Mosby | Company, 1977. 81 Rubin, H. J. Vocal nodules--A high speed photographic analysis: Notes on treatment. California Medicine, 1961, 95, 374-377. Ryan, R. E. et al. Synopsis of ear, nose and throat diseases. Saint Louis: The C. V. Mosby Company, 1970. Shaw, M. E., & Wright, J. M. Scales for the measurement of attitudes. Hew York: McGraw-Hill Book Company, 1967. Tatsuoka, M. M. Multivariate analysis--Techniques for educational and psychological research. New York: John Wiley 6c Sons , Inc. , 19717 Van de Geer, J. P. Introduction to multivariate analysis for the social sciences. San Francisco: W. H. Freeman and Company, 1971. Van Riper, C., & Irwin, J. V. Voice and articulation. Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1958. Webb, E. J. et al. Unobtrusive measures: Nonreactive research in the social sciences. Chicago: Rand McNally College Publishing Company, 1966. Wilson, D. K. Voice Problems in children. Baltimore: The Williams and Wilkins Company, 1972. Withers, B. T. Vocal nodules. The Eye, Ear, Nose and Throat Monthly, 1961, 40, 35-38. 82

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

An exploratory investigation of responses of young children to nonverbal auditory cues in messages

PDF

Pulsated voice production as a function of vocal constriction in adult males

PDF

Differential Effects Of Contingent Positive And Negative Listener Response On The Percentage Of Syllable Disfluency Of Preschool Boys

PDF

A Study Of The Influence Of Wife-As-Therapist On Language Progress With Adult Male Dysphasics

PDF

The Effects Of Harmonic Distortion On Sentence Intelligibility Of Hypacusic Adults

PDF

Effect Of Intensity Of Presentation On Pitch-Matching Behavior In Unilateral Meniere'S Disorder And Normals

PDF

Effects Of Audition And Vision On The Perception Of Selected Consonant-Vowel Syllables Under Conditions Simulating Hearing Loss

PDF

Effect Of Types Of Rehearsal On Frequency Of Stuttering

PDF

Speech And Language Characteristics Of Children With Developmental Dyslexia

PDF

Auditory Retention Abilities Of Young Children

PDF

Listener Identification Of 'Stuttering' And 'Stutterer' As A Function Of Variation In Speech Patterns

PDF

Effects Of Deglutition And Speech Training On Dental Overjet

PDF

Frequency Of Stuttering As A Function Of Connotative Word Meaning

PDF

An Analysis And Comparison Of The Figure-Ground And Visual Closure Abilities Of Two Preschool Groups Of Children With A Deprived Or Advantaged Language Background

PDF

An Experimental Study Of The Effects Of Fluency Feedback And Stuttering Feedback On The Subsequent Frequency Of Stuttering Utterance Duration, And Latency Of Response

PDF

Autonomic correlates of stuttering, fluency, and threat-of-shock

PDF

Development Of A Test Of Auditory Word Discrimination In Spanish

PDF

The Discrimination Of Directionality And Its Relation To Laterality And Measures Associated With Language Proficiency

PDF

The Effects Of Stuttering On Systolic Blood Pressure

PDF

Effect Of Selected Combinations Of Stimuli On Communicative Responses Of Children With Downs Syndrome

Asset Metadata

Creator Gray, Susan Gillen (author)

Core Title Construction validation of the Vocal Behavior Inventory

Degree Doctor of Philosophy

Degree Program Communicative Disorders

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag health sciences, speech pathology,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Advisor [illegible] (committee chair), [illegible] (committee member), Smith, Robert A. (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c17-741624

Unique identifier UC11345331

Identifier DP22515.pdf (filename),usctheses-c17-741624 (legacy record id)

Legacy Identifier DP22515.pdf

Dmrecord 741624

Document Type Dissertation

Rights Gray, Susan Gillen

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA