IN FO R M A TIO N TO USERS This material was produced from a microfilm copy of the original document. While the most advanced technological means to photograph and reproduce this document have been used, the quality is heavily dependent upon the quality of the original submitted. The following explanation of techniques is provided to help you understand markings or patterns which may appear on this reproduction. 1.T h e sign or "target" for pages apparently lacking from the document photographed is "Missing Page(s)". If it was possible to obtain the missing page(s) or section, they are spliced into the film along with adjacent pages. This may have necessitated cutting thru an image and duplicating adjacent pages to insure you complete continuity. 2. When an image on the film is obliterated with a large round black mark, it is an indication that the photographer suspected that the copy may have moved during exposure and thus cause a blurred image. You will find a good image of the page in the adjacent frame. 3. When a map, drawing or chart, etc., was part of the material being photographed the photographer followed a definite method in "sectioning" the material. It is customary to begin photoing at the upper left hand corner of a large sheet and to continue photoing from left to right in equal sections with a small overlap. If necessary, sectioning is continued again — beginning below the first row and continuing on until complete. 4. The majority of users indicate that the textual content is of greatest value, however, a somewhat higher quality reproduction could be made from "photographs" if essential to the understanding of the dissertation. Silver prints of "photographs" may be ordered at additional charge by writing the Order Department, giving the catalog number, title, author and specific pages you wish reproduced. b, PLEASE NOTE: Some pages may have indistinct print. Filmed as received. Xerox University Microfilms 300 North Zeeb Road Ann Arbor, M ichigan 48106 I 74-13,937 MELTON, Ellis Eugene, Jr., 1943A STUDY OF DIFFERENCES AMONG VARIOUS GROUPS OF MICHIGAN STATE UNIVERSITY MUSIC STUDENTS. Michigan State University, Ph.D., 1973 Music University Microfilms, A XEROX Com pany, Ann Arbor, Michigan A STUDY OF DIFFERENCES AMONG VARIOUS GROUPS OF MICHIGAN STATE UNIVERSITY MUSIC STUDENTS By Ellis E. Melton A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements f o r .the degree of DOCTOR OF PHILOSOPHY Department of Music 1973 ABSTRACT A STUDY OF DIFFERENCES AMONG VARIOUS GROUPS OF MICHIGAN STATE UNIVERSITY MUSIC STUDENTS By Ellis E. Melton The purpose of this study was to discover whether certain variables differentiate among groups of college music majors. Of particular interest was whether any of the measures could be used in the selection process for sophomore music students entering the music education curriculum. Thirty-one variables were selected which of­ fered some promise of usefulness. These variables in­ cluded 3 scales of the Aliferis Music Achievement T e s t , 1 rhythmic score from the Drake Musical Aptitude T e s t s ,, the 15 scales of the Edwards Personal Preference Schedule (EPPS), and 2 scales from the SAT (SAT-V and SAT-Q). last ten variables were grade point averages (GPA's) The in both music courses and general college courses. Over 500 music majors at Michigan State University were involved in the study, either by their taking various tests or contributing GPA's. The students were divided into groups on the basis of six independent variables: class, Ellis E. Melton curriculum, s e x , enrollment status, transfer status, and major instrument. Enrollment status was determined by the students presence or absence in the music department 18 months after the study began. All of the information about the students was pre­ pared for data analysis by computer. Appropriate statisti­ cal analyses were made. Certain conclusions were drawn based on the findings •of this study: 1. Music students who leave the music department for any reason except graduation do not score differently on the D r a k e , Aliferis, E P P S , or S A T . 2. The only variable which differentiates students according to the music curriculum in which they are enrolled is SAT-V. None of the other thirty-one variables result in statistically significant differences. 3. Five variables differentiate between sexes. Males score significantly higher on Deference, Heterosexuality, and SAT-Q, while females score significantly higher on Aliferis Melodic and Endurance. 4. Two variables show significant differences among students grouped by major instrument: Harmony GPA and Ad­ vanced Harmony G P A . 5. to the class There are no differences among students according (Freshman, Sophomore, and Junior) on any of the thirty-one variables used in this study. Ellis E. Melton 6. Transfer students do not differ from native students on any of the nineteen test scores. 7. Michigan State University music students differ significantly from published college norms on ten of the fifteen variables of the E P P S . The significance is due largely to sample size, however, and the actual differences may not be considered meaningful. A correlation matrix of the thirty-one variables revealed certain interesting correlations: 1. Applied music grades had very low correlations with other GPA's and test scores. Apparently, grades in applied music are fairly independent. 2. Among the EPPS scores, four consistently corre­ lated highly with GPA's: Achievement had high positive correlations, and Affiliation, Abasement, and Nurturance had high negative correlations. 3. Aural Harmony GPA's had a strong positive rela­ tionship with SAT-V and academic GPA's, but a low (-.05) relationship with applied music. Although some significant differences were found, the actual differences among groups of music majors were not considered meaningful as a basis for music department selec­ tion procedures. Future researchers would be well advised to look at fewer paper-and-pencil tests and more strongly at structured interviews. cal situations Student behavior in actual musi­ (for instance, conducting class or some early student teaching situations) also warrants monitoring. To Kay 11 ACKNOWLEDGMENTS The author is grateful and deeply indebted to: committee members, Dr. Merrell Sherburn, Professor Richard Klausli and Professor Robert Unkefer, who helped make graduate study and research very e x ­ citing, both in the classroom and around the con­ ference table, committee chairman and advisor, Dr. Robert Sidnell, who is due special appreciation for his guidance of this project and of the author's overall graduate program, Kay, whose encouragement and sacrifice enabled this work to be attempted, the faculty and students of the Michigan State Uni­ versity Music Department who participated in the research, Him who is involved in every worthwhile endeavor. iii TABLE OF CONTENTS Page LIST OF T A B L E S ............................................ yii Chapter I. THE P R O B L E M ................................. Introduction .............................. Need for the S t u d y ....................... Purpose ..................................... H y p o t h e s e s ................................. D e f i n i t i o n s .............................. F r e s h m a n ................................. S o p h o m o r e .............................. J u n i o r ................................. Music Education Students ............. S A T ..................................... G P A ..................................... University College Courses ............. Cumulative G.P.A.......................... Transfer Student ....................... Scope and L i m i t a t i o n s .................... Overview of the T h e s i s .................... II. SURVEY OF THE RELATED LITERATURE . . . . The Prognosis of Success in Teaching . . The Prognosis of Success in Music T e a c h i n g .................................. Relevant Studies with Music Majors . . . EPPS (Edwards Personal Preference S c h e d u l e .............................. D r a k e ..................................... C o n c l u s i o n s .............................. III. DESIGN OF THE S T U D Y ....................... Population and Setting .................... Data Gathering Procedures ................. Independent Variables .................... iv 1 1 5 8 11 12 12 12 12 12 13 13 13 13 13 13 15 16 16 18 21 28 29 30 32 32 32 35 Chapter Page C u r r i c u l u m .............................. I n s t r u m e n t .............................. ....................... Transfer Status Enrollment Status ....................... S e x ..................................... Dependent Variables— Course Grades . . . Variable 1. H a r m o n y .................... Variable 2. Advanced Harmony . . . . Variables 3. and 4. Aural Harmony and Advanced Aural Harmony ............. Variable 5. Applied 1 ................. Variable 6. Applied 2 ................. .............. Variable 7. U College 1 .............. Variable 8. U College 2 Variable 9. Cumulative 1 ............. Variable 10. Cumulative 2 ............. Dependent Variables— Test Scores . . . Variables 11. and 12. SAT-V and SAT-Q . Variable 13. The Drake Musical Aptitude T e s t s ................................. Variables 14. 15. 16. A1 Mel, A1 Har, Al R h y ................................. Variables 16 - 31. E P P S ............. R e l i a b i l i t y .............................. Handling of Missing D a t a ................. A n a l y s i s ................................. Multivariate Tests ................. Alpha L e v e l s ........................... IV. PRESENTATION OF THE D A T A .................... Correlations .............................. S A T - V ..................................... SAT-Q Aural (Aural H a r m o n y ) ................. H a r m o n y ................................. U College 1 (University College Freshman C o u r s e s ) .............................. Applied 1 .............................. Cum. 1 (Freshman Cumulative GPA) . . . Al. Mel. (Aliferis Melodic Sub-test) . Al. Har. (Aliferis Harmonic Sub-test) . A l . Rhy. (Aliferis Rhythmic Sub-test) . D r a k e ..................................... EPPS (Edwards Personal Preference S c h e d u l e ) .............................. H y p o t h e s e s ................................. v 37 38 39 40 42 42 43 44 44 45 45 46 46 47 47 47 48 50 52 53 57 59 60 60 62 63 63 65 65 66 66 66 67 67 67 67 68 68 68 69 Chapter V. Page SUMMARY AND C O N C L U S I O N S .................... 87 S u m m a r y ..................................... C o n c l u s i o n s .............................. D i s c u s s i o n ................................. Implications .............................. Suggestions forFuture Research . . . . 87 88 89 96 97 BIBLIOGRAPHY ........................................ vi 99 LIST OF TABLES Table Page 1. Number of Students in Each C l a s s .............. 37 2. Number of Students in Each Level of C u r r i c u l u m ..................................... 38 Number of Students in Each Major Instrument C a t e g o r y ........................................ 39 Number of Students According to Transfer S t a t u s ........................................ 40 Number of Students in Each Level of Enroll­ ment S t a t u s 41 6. Numbers in Each Level of S e x ................. 42 7. Reported and Obtained Reliabilities 58 8. C o r r e l a t i o n s ................................. 64 9. Means and Obtained "t" Values for the SAT for Students Grouped by Enrollment Status . . . 69 Multivariate ANOVA for Nineteen Test Variables for Students Grouped by Enrollment Status . . 70 Multivariate ANOVA for the Two SAT Tests for ............. Students Grouped by Curriculum 71 Multivariate ANOVA for the Ten Grade Variables for Students Grouped by Curriculum . . . . 72 Multivariate ANOVA for the Nineteen Test Variables for Students Grouped by C u r r i c u l u m ..................................... 72 Univariate ANOVA results for SAT tests for ............. Students Grouped by Curriculum 73 3. 4. 5. 10. 11. 12. 13. 14. vii . . . . Table 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. Page Multivariate ANOVA for Ten Grade Variables for Students Grouped by S e x ................. 73 Multivariate ANOVA for Nineteen Test Variables for Students Grouped by S e x ................. 74 Multivariate ANOVA for the Two SAT Tests for Students Grouped by S e x ....................... 74 Univariate ANOVA Results when Students are Grouped by S e x ................................. 75 Multivariate ANOVA for Ten Grade Variables for Students Grouped by Major Instrument . 76 Multivariate ANOVA for Nineteen Test Variables for Students Grouped by Major I n s t r u m e n t ..................................... 76 Multivariate ANOVA for the Two SAT Tests for Students Grouped by Major Instrument . . . 76 Univariate ANOVA Results when Students are Grouped by Major Instrument ................. 77 Multivariate ANOVA for Nineteen Test Variables for Students Grouped by Class 78 . . Multivariate ANOVA for Nineteen Test Variables for Students Grouped by Transfer S t a t u s ........................................ 79 Means and Significant Differences forMusic Majors and College Norms .................... 79 26. Means of 31 Variables by Major Instrument . . 81 27. Means of 31 Variables by Curriculum Groups . 82 28. Means of 31 Variables by Transfer Status . 83 . 29. Means of 31 Variables by C l a s s .............. 84 30. Means of 31 Variables by S e x .................. 85 31. Means of 19 Test Variables by Enrollment Status after 18 months ..................... 86 viii CHAPTER I THE PROBLEM Introduction People who practice a particular profession must possess certain skills, information, and attributes which are not as concentrated throughout the general population. These individual attributes are not shared exclusively by a particular profession, but such a unique combination of at­ tributes can define a profession. Some of these character­ istics, particularly highly complex cognitive skills, are acquired through formal schooling. Some others, including advanced motor skills, are often acquired through individual practice and life experiences. A third category of character­ istics of professionals includes those which are often called aptitudes— that is, those held by individuals before advanced training. These include psychological needs, physical char­ acteristics and inborn inclinations. A definition of "profession" includes the statement: "an occupation or vocation requiring training in the liberal arts of the sciences and advanced study in a specialized field."1 Training in the liberal arts and sciences includes 1The American Heritage Dictionary of the English Language. (New York: Houghton Mifflin Company, 1971). 1 2 by tradition a common body of knowledge generally recommended for all educated citizens. It is in the "advanced study" portion of a professional's education that the basic skills, information, and attributes are acquired. A university pro­ vides both types of education for the professional. The fact that some individuals are better suited than others for a particular profession is well demonstrated by simply observing that some succeed while others fail. The failure of individuals can be observed at many points before and during a career. Dropping out of school, failing to find a professional position, or changing professions late in life might indicate that the choice of career was a poor one. On the other hand, a person who exhibits none of the above b e ­ havior is not necessarily successful. If he is actually prac­ ticing his chosen profession he is at least demonstrating that he possesses some of the traits which set him apart from the general population. Even a poor scientist knows much about science, just as a not-too-successful professional ath­ lete is probably better developed physically than the nonathletic person. Naturally, the ability to assist people with career decisions is important to counselors, so important that much research has been directed to that end. In spite of the ob­ vious differences in job descriptions among professions, no one has achieved a high degree of accuracy in predicting the most suitable occupation for a particular individual. This is especially true in the early stages of training, referred to above as the "liberal arts and sciences" portion of general education. Another part of the problem of prediction lies in dis­ covering which traits are really unique or at least highly developed among successful people in any occupational group. For instance, one might assume that all successful lawyers possess a basic knowledge of law, as well as oratorical tal­ ents. In fact, the common bond might be simply perseverance or verbal reasoning ability. Or perhaps the term "lawyer" is too heterogeneous and there are only common traits between "trial lawyers" or "trial lawyers from Harvard." Thus, the definition of "successful professionals" and the discovery of unique traits held in common by these successful professionals seem to be two primary problems in prognosis of professional success. Who is charged with the responsibility of screening people for various professions? The precedent for selection by colleges seems clearly established: "Colleges and universities are the primary institu­ tions in our society that affect and direct the flow of students into the professional occupations. Colleges not only provide an education for the professions and the specialized occupations, they also serve a critical function in channeling students into careers and in eliminating students who seem to lack the prerequisites for entry to a profession.2 In liberal arts and humanities, colleges have never promised graduates that they would be qualified for any particular occu­ pation. These fields have traditionally concentrated on devel­ oping a "well rounded" adult who can then adapt himself to a number of occupations with the basic tools he has acquired. Professional schools, on the other hand, are established as training grounds for particular occupations and have a respon­ sibility to graduate only those who have acquired entry skills for those occupations. Assuming that teaching is a profession according to the definition stated earlier, there is the possibility that teachers differ from the general population in one or more variables. Since the term "teacher" encompasses high school, college, elementary and private teachers in widely varied fields, it is not surprising that research has not been con­ clusive when defining "teaching characteristics." Perhaps a chemistry teacher holds more in common with laboratory scientists'tfian with teachers in other disciplines. Or perhaps college teaching requires a different set of traits than those needed for kindergarten teaching, the two groups having little in common to distinguish them from the general population. Isolating and measuring a homogeneous group of teachers appears to be a prerequisite for definitive research in the prognosis of teaching success. 2 John K. Folgen, Helen S. Astin, and Alan E. Bayer, Human Resources and Higher Education (New York: Russell Sage Foundation, 1970), p. 13. 5 This study will deal with music teachers in the public schools as they enter the university preparation portion of their careers. This is usually the earliest point at which these teachers make a commitment to their chosen field, and therefore a critical point in their lives. Need for the Study Universities are constantly faced with the task of evaluating students for various purposes. Such routine matters as college admissions, grades, course requirements and graduation criteria all require the institution to make a judgement about the student on the basis of a variety of collected data. Many times these data consist of some easily obtainable but statistically questionable scraps of informa­ tion such as test scores or previous course grades. All too often major decisions are based on date of application, as is sometimes the case in college admission procedures. This method of selecting college students may be expedient and even democratic, but it is not justified in terms of the best use of human potential. One particular area which needs precise evaluation is in the selection of prospective students for teacher educa­ tion programs. Typically, the student makes application after his first, second or third year of college and the responsibility for the screening rests with either the 6 professional education faculty or a committee from the subjectmatter field. Drayer summarizes the usual procedure: "Some objective evidence for evaluating the applica­ tion is immediately available. At the time he applies for the program, the candidate may be required to present a doctor's certificate . . . his past academic record . . . (and) the student has a personal interview, during which a general impression of his motivation, personality, and expression may be formed. After his application, and before his admission to the program, further information may be gathered from faculty members. If all the factors appear to be favorable, the applicant is admitted to the program."3 Drayer continues to state that the candidate is watched as he moves through the teacher education program and is sometimes dropped for academic or other reasons. For many years the net result of this rather loose procedure was that just about anyone who passed prescribed courses with a sufficient grade was graduated and allowed to enter the teach­ ing profession. In 196 4 Farr studied the tests which were being used in the process of screening future teachers in institutions which belonged to the American Association of Colleges of Teacher Education. Of 4 4 3 schools which responded, he found 445 distinct test titles.^ Obviously, a valid test has not 3Adam M. Drayer, The Teacher in a Democratic Society (Columbus, Ohio: Charles E. Merrill Publishing Company, 1970), p. 247. 4David Farr, "Evaluation and Selection Instruments in Teacher Education Problems," summary report of the Subcommittee on Testing in Teacher Education Committee on Studies, American Association of Colleges for Teacher Education, 1964, p. 3, cited by Drayer, 249. 7 been found, for if it existed, most institutions would adopt it. At the present time there is a particular need for research concerning procedures of evaluating music education students at Michigan State University. Because of the rapid growth in the enrollment of the College of Education and the declining job market for teachers, a quota system was established in the Spring term of 1971 whereby only a specific number of juniors would be allowed to enter the College of Education from each subject-matter field. In the case of the music department, approximately 40 of 130 sophomore music edu­ cation majors would not be allowed to continue in that curricu­ lum. In effect, students were to be screened and selected for the teaching profession at the end of the sophomore year. It was possible, of course, for a student to transfer to another institution and complete the music education degree. A reliable screening device is needed not only when a student makes official application to enter the teacher training curriculum, but at other points in his academic career. Many music majors arrive on campus with no idea of the standards and demands of a music department and eventually become dropouts. They have wasted the time and resources of the institution as well as their own. A testing and counsel­ ing program immediately preceding a student's first enrollment would alert the student to any weaknesses in his background so that he might develop other options for himself. In the hands of a skilled counselor, an ego-shattering confrontation with the facts of one's limitations can be turned into a posi­ tive rather than a negative experience. The counselor needs facts and statistics to make counseling sessions valid. What information does the college counselor need? There are basically three types: details of the curricula available; information about the individual student; and the ways a given student is likely to react to a curriculum. The first of these is rather easy for the counselor to acquire, as the curriculum is obvious and fairly constant. The infor­ mation concerning the individual student is also easy to acquire if the exact kinds of information needed are shown. The last of these— the way particular students will achieve in a curriculum and ultimately in a profession— is more com­ plex than the first two and requires more than casual observa­ tion to obtain. Purpose The purpose of this study is to investigate certain measurement tools to determine their usefulness in the selec­ tion of music education majors. This purpose is pursued by means of an extensive description of the music students enrolled during one term, using a number of variables. If these vari­ ables do not in any way discriminate among specific sub-groups 9 of the music student population, particularly between music education and non-music education students, it would seem that the administration should look at other variables. If at least one variable does indeed discriminate in some way, a long term longitudinal study would be in order to determine if it is also a predictor of success in teaching. Research literature indicates that the ultimate screen­ ing tool will not be a simple univariable test. For example, J. P. Guilford hypothesized that any attempt to measure gen­ eral intelligence and arrive at a single score, such as an "IQ," would be futile. He and his followers have proposed a three-dimensional model which includes 120 cubicles, each of which supposedly can be defined and measured. 5 If intelli­ gence is this complex, and native intelligence is one facet of success in a profession, it follows that the predictor of success will not be a simple one-dimensional attribute. It can be hypothesized that the traits necessary to become a successful music teacher, or even remain in college as a music student, form a complex pattern of interrelated characteristics. One might suspect, combination of perseverance, for instance, that some intelligence, training, person­ ality, physical characteristics and background affect college success and that slightly different combinations of these 5 J. P. Guilford, The Mature of Human Intelligence (New York: McGraw Hill Book Company, 1967). 10 characteristics would be found in the various disciplines. If this is true, simple SAT scores or IQ test scores would not seem inclusive enough. Perhaps a multivariate approach is mor.i applicable, whereby a number of variables contribute to make up a profile of each student. The problem is: what are the appropriate variables for the profile? The purpose of this study is not an exhaustive search for one specific, foolproof variable which will predict a student's potential as a music teacher. This type of finding appears impossible as long as humans have individual differ­ ences and educational measurements remains an inexact proced­ ure. One immediate possible outcome, however, is the elimina­ tion of consideration of variables which seem to have no bearing at all on the music students. The particular need for a study using more than one or two variables will be evident when the literature is re­ viewed in Chapter II. Many researchers have dealt with a small number of subjects who were measured on a few variables, whereas this study will approach the problem on a broader scale. In addition, this study is useful as a reassessment of the one-variable types of research, and will either help reinforce or refute those findings. The variables chosen for the study include two types. The first type are written tests: 11 1. 2. 3. 4. 5. 6. 7. SAT Verbal SAT Qualitative Aliferis Musical Achievement (Melodic) Aliferis Musical Achievement (Rhythmic) Aliferis Musical Achievement (Harmonic) Drake Musical Aptitude (Rhythmic) Edwards Personal Preference Schedule (Fifteen scores) The second category of data includes various grades for specific courses and grade point averages as follows: 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. Freshman University College grade average Sophomore University College grade average Basic Harmony grade average Advanced Harmony grade average Aural Harmony grade average Advanced Aural Harmony grade average Freshman applied music grade average Sophomore applied music grade average Freshman cumulative grade average Sophomore cumulative grade average. Hypotheses The hypotheses for this study are: Hypothesis 1: Music students who leave the music department for any reason except graduation will differ on one or more variables from those who remain. Hypothesis 2: Music students differ on one or more variables when grouped according to the specific music curriculum in which they are enrolled. Hypothesis 3 : Music students differ on one or more variables when they are grouped according to sex. Hypothesis 4: Music students differ on one or more variables when they are grouped according to their major instrument. Hypothesis 5: There will be no difference among music students on the Aliferis, Drake, or EPPS when the students are grouped by class. 12 Hypothesis 6 : Music students who transferred to Michigan State University from other institutions do not differ from students who began at the University as Freshmen on standardized apti­ tude and achievement tests. Hypothesis 7 : Music students do not differ from the general population of college students on any of the personality variables measured by the Edwards Personal Preference Schedule. Definitions Freshman At Michigan State University, a freshman is any stu­ dent who has earned less than forty term credits. Sophomore At Michigan State University, a sophomore is any stu­ dent who has earned forty to eighty-four term credits. Junior At Michigan State University, a junior is any student who has earned eighty-five to one hundred twenty-nine term credits. Music Education Students Music education students are any undergraduate stu­ dents who have declared themselves in the curriculum leading to certification as school music teachers, and who have been so designated in official university records. 13 SAT SAT is the Scholastic Aptitude Test.^ GPA GPA is the Grade Point Average as computed on a 4.0 scale, where 4.0 is equivalent of a grade of "A." University College Courses University College Courses are those basic courses required by the University for any Baccalaureate degree. Cumulative G.P.A. Cumulative G.P.A. is the grade point average on all courses taken at the University to date. For instance, a Junior would have acquired both a Freshman and Sophomore cumu­ lative G.P.A. Transfer Student A transfer student is any student who begins his college work in another institution or another department at Michigan State University before entering the music department. Scope and Limitations The study was undertaken at Michigan State University on the campus in East Lansing. The data concering course grades, major instrument, transfer status, and curriculum were ^Scholastic Aptitude Tests (Princeton, New Jersey: College Entrance Examination Board, 1961). 14 gathered by the researcher directly from the student records as maintained by the music department. All testing and data gathering was done between November 1, 1971, and April 30, 1972, among students who were classified as music majors as i of November 1, 1971. Although the study concerns students at one University 7 at one point in time, the Cornfield-Tukey argument allows others to generalize the results of this or any study if they believe that the conditions are similar to their own. For instance, a music department at a large state university in another state might decide that the students involved in this study in 1971-1972 are enough like theirs in 1975 that the re­ sults would be applicable. For this reason, the population and conditions will be carefully described in Chapter III. This is not a true experimental study which offers several treatments of different samples in the hope that the results may be inferred to larger populations. Many assumptions are necessary for that type of study, not the least of which is either randomization or random selection. The students about whom data were gathered were not randomly selected in any sense. In some cases, by refusing to attend the testing sessions, students chose not to be a part of the study. Since most of the testing took place in classes, some students were excluded. 7 J. Cornfield and J. W. Tukey, "Average Value.; ^ Mean Squares in Factorials," Annals of Mathematical Statistics XXVII, pp. 907-949. 15 The data will be handled through a variety of sta­ tistical processes including correlation, multivariate analy­ sis of variance, and "t" tests. Means and variances will be computed and reported as descriptive data. Overview of the Thesis In the following chapter, literature related to this study will be discussed. Some of the related literature deal­ ing with the uses of the individual variables will be reviewed, but the chapter will deal primarily with the problem of selec­ tion in music schools. In Chapter III the design of the study is discussed, along with the independent and dependent variables. sults of the study are presented in Chapter IV. The re­ The final summary and conclusions will be submitted in Chapter V, along with suggestions for further research. CHAPTER II SURVEY OF THE RELATED Three areas of review seem gation: LITERATURE pertinent to this investi­ the prognosis of success in teaching, the prognosis of success in music teaching, and other more directly related studies with music majors using a number of variables. The Prognosis of Success in Teaching Research on predicting teaching success can be found dating from 1905, with the largest bulk of the literature in the 1930's. Generally, the studies before 1950 were not very well done by modern standards, and the results were contra­ dictory. Advanced quantitative measurement techniques were not available before the computer age, and the research d e ­ signs tended to be faulty, causing the findings to be suspect. Beginning around 1950 the problem of prognosis of teaching success was attacked in earnest. One of the first conclusions was that there would be no easy answer. instance, IQ measures enjoyed wide popularity during the decade of the 1 9 5 0 's, but they turned out to be tors. For poor predic­ One viewpoint follows: It seems highly unlikely that future researchers using global measures of intelligence and conventional 16 17 criteria of teaching success will tell us much more than we now know . . .^ In fact, approximately 30% of the studies between 1927 and 1952 found negative correlations between IQ and teaching success. As Guilford pointed out with his structure of the intellect model, human intelligence is far too complicated to summarize with a single number. Another well researched area of the problem is the use of a personality profile as a predictor. findings have resulted, such as: Some very pedestrian good teachers are more kind, cheerful, and courteous than are poor teachers. This type of information is not useful because it is not stated quantita­ tively. Projective techniques, such as the Rorschach Ink Blot Test and the Thematic Apperception Test (TAT) showed much promise twenty years ago but have failed to distinguish effectively good teachers from bad ones. These tests are difficult to score and interpret and are therefore impractical for many u s e s . A number of fairly recent investigations have pertained to the prediction of general teaching success. Bach 2 found negligible correlations between student teaching grades 1J. W. Getzels and P. W. Jackson, "The Teacher's Personality and Characteristics," Handbook of Research on Teaching, ed. N. L. Gage (Chicago: Rand McNally and C o . , 1965), p. 572. 2 Jacob 0. Bach, "Practice Teaching Success m Relation to Other Measures of Teaching Ability," Journal of Experimental Education, XII (1952), 57-80. and success in the field. Rmgness 3 found certain attitudes predictive of teaching success, particularly interest in a 4 subject matter field. Charles reported the use of projective techniques in the selection of teachers. A great many psychological, intelligence and aptitude tests have been used to try to predict success in teaching. The primary problem seems to be that teachers are more closely allied with their subject matter fields than they are to other teachers in different fields. For instance, a music teacher may have more in common with a professional conductor than with a chemistry teacher. Therefore, any instrument which purports to measure a trait or produce a profile of "teachers" faces the problems of a heterogeneous population. It is for the above reason that the related literature reviewed herein deals primarily with music oriented studies. The fact that successful music teachers appear to be a much more homogeneous population than teachers in general offers hope that a more clear-cut profile will emerge from the data. The Prognosis of Success in Music Teaching The Music Supervisors National Conference passed the following Resolution at its 19 30 meeting: 3 Thomas A. Ringness, "Relationships Between Certain Attitudes Toward Teaching and Success," Journal of Experimantal Education, XXI (1952) 1-55. ^Harvey Charles, "The Use of a Selected Projective Technique in the Teacher Selection Process," Studies in Education, Abstracts of Theses, 1952 (Bloomington: Indiana University 1953). 19 Be it resolved, that institutions for the training of teachers and supervisors of school music be urged to exercise greater care in the selection of students who seek to undertake this training, by demanding not only that they have adequate previous musical study, but also the assurance that they possess possibilities of necessary future development.5 In 1932, Lowell Mason Tilson made the following state­ ment which laid the foundation for all music teacher prognosis studies since that time: If there are great individual differences in the native music endowment of the students who decide to enter upon courses leading to licenses for teaching and supervising music in the public school it is very important that instructors in charge of such courses know how to select students who are most likely to succeed, and to eliminate those who are almost sure to fail. This selection and elimination should not be attempted except upon the basis of carefully evaluated results of music talent tests.6 Although the prognosis of success in music teaching is not the primary purpose of this study, five of these types of investigations from the literature are relevant to this study and appear below. A University of Illinois researcher 7 using the Minnesota Multiphasic Personality Inventory found that very successful high school instrumental teachers were less neurotic, less 5 Minutes of the Music Supervisors National Conference, cited by Lowell Mason Tilson, "Music Talent Tests for TeacherTraining Purposes," Journal of Experimental Education, XVIII (1932), 26. 6Ibid. 7 Warren Lutz, "Personality Characteristics and Experi­ ence Backgrounds of Successful High School Instrumental Music Teachers" (Unpublished Doctoral Dissertation, University of Illinois 1963). 20 moody, worked harder, and were more satisfied with their profession than were less successful teachers. Q Using the same MMPI test, another Illinois study found that music teachers with the highest festival ratings were cool, aloof, optimistic and methodical. Directors with lower ratings tended to be gloomy, depressed, sensitive, r e ­ ligious and had difficulty with authority figures. Thirty instrumental music teachers were involved in the study. In 1961 a University of Southern California researcher q was unable to find significant differences between 156 selected music teachers and 160 randomly selected ones. The Thurstone Temperament Schedule and the Sixteen Personality Factors Test, were used. However, when the 316 teachers were redivided a c ­ cording to "job satisfaction" the tests discriminated very well. Anderson‘S found that music students who did better in student teaching differed from those who did less well on Q John Fosse, "The Prediction of Teacher Effective­ ness: An Investigation of Relationships Among High School Band Contest Ratings, Teacher Characteristics, and School Environmental Factors" (Unpublished Doctoral Dissertation, University of Illinois, 1965). 9 George Barth, "Some Personality and Temperament Characteristics of Selected Music Teachers" (Unpublished Doctoral Dissertation, University of Southern California, 1961). "^John Martin Anderson, "The Use of Musical Talent, Personality and Vocational Interest Factors in Predicting Success for Student Music Teachers" (Unpublished Doctoral Dissertation, University of Southern California, 1965). 21 certain personality and vocational interest tests. Specifi­ cally, the objectivity and masculinity scales from the Guil­ ford- Zimmerman Temperament Survey and the symbol reproduction scale from the Project Potential Creativity Battery showed promise in selecting and counseling future music student teachers. Anderson failed to find differences between nine­ teen successful student teachers and a group of ten master teachers. Turrentine"*"1 investigated the SAT, an IQ test, high school GPA, and various college GPA's as predictors of success in music student teaching. The best single predictor of success was the GPA for teacher training courses; however, the correlation was only .564. The second best predictor was high school GPA, and all other variables correlated rather low. Relevant Studies with Music Majors Since such populations as "students," "student teach­ ers," and "professionals" are extremely heterogeneous, en­ compassing many subject matter fields, it would seem necessary to look at research concerning college music students only. Perhaps music students represent a relatively homogeneous population for which definite criteria can be established. If this is true, music education students and practicing music "^Edgar Turrentine, "Predicting Success in Practice Teaching in Music" (Unpublished Doctoral Dissertation, University of Iowa, 1962). 22 educators may be even more homogeneous as a group than are all "musicians." An early attempt at setting levels for entrance into advanced courses in a music teaching training program was by Tilson. 12 . His purpose was the validation of the Seashore Musical Talent Tests and a standardized intelligence test for use in deciding which students should be permitted to enter courses intended for the training of music supervisors. Tilson found that those students scoring in the lowest quartile on the music talent tests should be discouraged from enrolling in freshman music courses, regardless of intelligence test scores. Students scoring in the lowest quartile on both tests should positively be eliminated from the music program. The Tilson study was one of the first attempts to study the relationships of music talent, course grades, and intelligence. He found correlations of .399 between music talent scores and grades in ear training, and .340 between intelligence and grades in ear training. These correlations are rather low and are of little value in predicting success in ear training. Another early study was by More, very much like that of this study. 12 13 whose purpose was She attempted to validate . Tilson, op. c i t . "*"^Grace van Dyke More, "Prognostic Testing in Music on the College Level: An Investigation Carried on at the North Carolina College for Women," Journal of Educational Research, XXVI (1932), 199-212. 23 a battery of tests for use in counseling freshmen music majors. The test battery included Seashore tests, two tests of melodic and harmonic sensitivity, a "silent reading" test, a test of relative pitch, and five tests designed by the researcher. The latter tests attempted to measure the students' ability to solve musical problems, primarily the detection of pitch and rhythmic errors. The correlations of these test scores with freshman music grades were very interesting. Although none of the tests had a higher correlation with grades than .563, the tests involving problem solving, such as error detection, correlated higher than those which measured only one factor of perception. The researcher suggested that future research consider tests of score reading, error detection, and other tests which involve musical problem solving. Stanton 14 used the Seashore Measures of Musical Talent and the Iowa Comprehension test, a non-musical test, as a test battery to be given to high school seniors as a proqnosis tool for college music achievement. On the basis of a combination of the two scores, plus first term theory grades and ratings by applied teachers, she assigned each student into one of five categories: Doubtful, and Discouraged. 14 Safe, Probable Possible, The Safe category included all Hazel M. Stanton, "Testing the Cumulative Key for Prognosis of Musical Achievement," Journal of Educational Research, XXVII (1934), 45-53. 24 those students considered a "Safe" academic risk. At the other end of the scale, the "Discouraged" category included those students whose test scores indicated greater odds against them to the extent that encouragement was not justi­ fied. The categories were established using teachers' ratings and grades, although the teachers were not aware of the cate­ gory for each student. The students were freshmen entering Eastman School of Music from 1920 through 1929. The records of the music students during their four years of college were used to validate the test battery. More than half of the "Safe" group continued into the fourth year of college, compared with two-fifths of the "Probable" group and none of the "Discouraged." From the percentage of those dismissed from the School of Music for academic reasons the increase was from four per cent for the "safe" group to sixty- four per cent for the "Discouraged." Further records show that those in the upper groups gave more recitals, won more scholarships, and earned higher grades than did those in the lower groups. Using data gathered over a five year span from 19 25 through 1929, Stan­ ton reports that sixty per cent of the "Safe" students were graduated, forty-two per cent of the "Possible," twenty-three per cent of the "Doubtful," and seventeen per cent of the "Discouraged." The results of this study appear significant, although it is important to note that teachers' ratings were 25 one of the variables. Certainly, the teachers' expectations of the students biased the results of the study, because other research has shown that students tend to achieve at the level that teachers expect. Another extensive early study was by E. M. Taylor at the College of Music of Cincinnati. 15 A battery of 22 tests was given to 150 freshmen entering from 1930 through 1935. Coefficients of contingency for scores of these tests and ratings for professional success in music in 1939 ranged from .21 to .62. Grades in sight singing and dictation courses and instructors ratings had coefficients ranging from .50 to .62. At Indiana University Peterson 16 studied 259 graduate music students who had been given the University of Indiana Test Battery, (which included the pitch, rhythm, and tonal memory sections of the Seashore Measures of Musical Talent, the musical memory sub-test of the Drake, and the Madison Test of Interval Discrimination). He concluded that although the battery will result in a slight improvement over chance selec­ tion, its use as a final arbiter is unwarranted. He also found only negligible differences among the various students grouped by curriculum. 1^ E. M. Taylor, "A Study in the Prognosis of Music Talent," Journal of Experimental Education, X (1941), 1-28. ■^Floyd J. Peterson, "A Study of the Relationship Between Music Aptitude and Academic Achievement of Graduate Music Students" (Unpublished Doctoral Dissertation, Indiana University, 1963). 26 A study by Snapp 17 attempted to develop an interest inventory for use in vocational guidance for musicians. He validated the inventory with active musicians in a number of fields including orchestral playing, college theory teaching, public school teaching, and applied teaching. Snapp con­ cluded that the interests of these groups differed signifi­ cantly but failed to report any longitudinal validation. His most striking finding was that public school music teach­ ers were the most dissimilar to the other groups of musicians in terms of interests. C. H. Taylor 18 gave a composite picture of a hypotheti­ cal student entering the Peabody Conservatory of Music and concluded that if the student failed it would probably be because of personality or emotional difficulties rather than for lack of necessary musical abilities. Bienstock 19 used some of the variables of the present study when she tried to discover a predictor of college music achievement. Among her findings were: (1) the Kwalwasser- Dykema tests were too unreliable to be used for the 17 Kenneth 0. Snapp, "Development of a Musician Inter­ est Inventory for Use in Vocational and Educational Guidance" (Unpublished Doctoral Dissertation, Indiana University, 1953). 19 C. H. Taylor, "Characteristics of First Year Conserva­ tory Students," Journal of Research in Music Education, I (1953), 105-118. "^Sylvia Bienstock, "A Predictive Study of Musical Achievement," Journal of Genetic Psychology, LXI (1942), 135-145. 27 prediction of individual success in music: (2) there was a positive correlation between the Kwalwasser-Dykema tests and success in music theory and applied music, but it was too low to be of practical value; (3) the most effective measures for the prediction of success in theory were the IQ scores and the age of the students, while the least contributive were the extent of prior music training and performance background. Gallagher 20 found that music students who scored high on various standardized tests of music aptitude and achieve­ ment had similar backgrounds, such as socio-economic status, home environment, and formal study of music. Stone 21 investigated high school grades, scholastic aptitude test scores, theory placement test scores, and a psychological adjustment test as predictors of Freshman GPA's. The findings were complex but include the following which are relevant to this study: (1) grades in applied music, music classroom courses, and non-music courses can be treated as discrete experimental variables; (2) high school grades were not found to be the best predictor of overall college GPA; and (3) theory placement test scores were the single best 20 Fulton D. Gallagher, "A Study of the Relationships Between the Gordon Musical Aptitude Profile, the Colwell Music Achievement Tests, and the Indiana-Oregon Music Discrim­ ination Test" (Unpublished Doctoral Dissertation, Indiana University, 1971). 21 Michael Horace Stone, "A Study of the Relationships Between Selected Variables and the Differential Academic Achievement of Freshmen in the University of Michigan School of Music" (Unpublished Doctoral Dissertation, University of Michigan, 1969). 28 predictor of achievement in applied music, all music class­ room courses, as well as non-music courses. This theory placement test was developed especially for use at the University of Michigan, where the study took place. EPPS (Edwards Personal Preference Schedule Although personality factors are generally thought to be an important influence on academic achievement and success in a vocation, only limited success has been achieved in using them as predictors. The present study begins with an assumption that the previous limitations of success have been a result of failing to delimit sufficiently the nature of the group being predicted. Subtle differences may have been hidden in the remaining random variance of the proced­ ures. Using music majors rather than a more heterogeneous population offered hope of some success with personality tests. The EPPS has been in constant use since its develop­ ment in 1954 and there is a wealth of related literature in­ cluding many validity and reliability studies. many uses for this ubiquitous instrument are: Among the prediction in teacher education; comparison of various groups of teachers according to needs; descriptions of literally hundreds of sub-groups of teachers; and comparisons of EPPS scores with other test batteries and observations. 29 Lunneborg 22 tried to relate EPPS patterns with academic achievement, but her population were counseling clients in a wide variety of subject matter fields. The moral would appear to be that to account for college grades the best predictors are traditional aptitude tests and high school grades unsupplemented by personality measures. Personality patterns have again failed to live up to the hopes of many that they repre­ sented the unpredictable variance in school achievement. Drake The literature concerning the use of the Drake Musical Aptitude Tests 23 is scarce indeed. Although many re­ viewers have been impressed with the validity and reliability as published in the manual and researched by Drake himself, few independent investigations have sought to confirm Drake's findings. Drake reports remarkably good validities using teacher ratings of "talent" as the external criteria. The validity coefficients reported in the manual range from .31 to .91, with a majority attaining a value greater than .58. 24 Concern­ ing predictive validity, Drake states that "scores are highly 22 Patricia W. Lunneborg, "EPPS Patterns and Academic Achievement in Counseling Clients" (Washington University Bureau of Testing, Seattle, April, 1969). 23 2nd ed. . Raleigh M. Drake, Drake Musical Aptitude T e s t s , (Chicago: Science Research Associates, inc.” TJ57) . 24 Raleigh M. Drake, Manual for the Drake Musical Aptitude Tests (Chicago: Science Research Associates-, T957) . 30 predictive of success in music courses and music schools." 25 Unfortunately, no data or references are given in the manual to support this claim, and evidence of any independent r e ­ search was not found by this writer. Conclusions Despite the critical importance of the problem and years of prodigious research effort, very little is known for certain about the prediction of success in music, music teaching, or teaching in general. self-evident Some of the findings have been ("teachers are friendly and sympathetic"); most of the rest have simply been inconclusive. Researchers have attempted to discover or devise means of predicting success in music schools through person­ ality tests. Findings apparently are not sufficient to yield a comprehensive predictive device. Many researchers found that, while certain single variables evidenced some significant correlations, predictions on this basis would be spurious. Little is known concerning the relationship of a person's personality and his association with the arts. No conclusive evidence has been presented which can support the rather common notion that people in the arts have unusual personalities. Studies dealing with the differences between success­ ful and unsuccessful music majors have also been inconclusive. 31 Specific locally made tests, such as theory tests for enter­ ing Freshmen, seem to hold more promise than standardized tests of musical aptitude and achievement. Batteries of tests have been used, but the literature is unclear concerning treatment of the data obtained from these test batteries. Little evidence of advanced statisti­ cal procedures such as vector analysis and analysis of vari­ ance can be found. Correlational studies are by far the most prominent. None of the studies located by the writer approach the problem with the combination of variables used herein. Although the EPPS, Drake, and Aliferis tests have all been widely used for many purposes, the literature concerning their validity for prognosis of college music students is not so plentiful as one might hope. The purpose of this study is to try a unique combination of variables with one of two possible outcomes: (1) a measuring device will be found which will aid in the counseling of music majors, or (2) these particular variables are invalid for counseling and prognosis purposes and therefore should not be used without further study. CHAPTER III DESIGN OF THE STUDY Population and Setting The students involved in the study were undergraduate students in the Michigan State University Department of Music during the Fall term of 1971. Approximately 475 undergraduate students were coded "music" by the University and constitute the group referred to as "music majors." All had been accepted by the music department to work toward a Bachelor of Music degree. There was no particular reason to believe that these students differed from the general population of University music students throughout the nation, or were unique when com­ pared to past students at the University. It is impossible to test the latter assumption due to lack of normative data gathered prior to this study. Data Gathering Procedures The data for this study were obtained from two sources: the students1 academic files and tests administered by the researcher. The Director of Students for the Music Department maintains an academic file for each student in the department. 32 33 The files contain such items as grade reports, audition r e ­ ports, a copy of the student's University admission applica­ tion, and other material pertaining to his academic progress. In accordance with University academic freedom policy, the student has complete access to his own file. The information obtained from the academic files in­ cluded the following: 1. 2. 3. 4. 5. 6. 7. 8. sex curriculum transfer status major instrument enrollment status SAT scores GPA's from previous institutions GPA's from various MSU courses and cumulative G P A 's . As the information about each student was collected, it was recorded according to a code number assigned to that student by the researcher. This procedure assured that data were used for research purposes only and the identities of the particular students were protected. The information from the files was collected during the month of October, 1971. All data in this study are to be considered correct through that month. Any information which was not in the academic files, such as a missing grade report, was treated as "missing data" as if the information had never existed. The statistical treatment of "missing data" is explained at the end of this chapter. GPA's taken from the files were computed on whatever information was available. For instance, if a student had 34 only taken two of the three terms of Advanced Aural Harmony, only those two terms were used in computing the Advanced Aural Harmony GPA. The second category of data came from various tests administered by the researcher. These tests were given during regular music classes from October 21, 1971, through April 12, 1972. Because individual instructors had to give permission for their classes to be used for testing, it was necessary to extend the testing period over approximately six months, reducing the amount of time students would miss from regular classroom work during any given term. Larger classes were chosen in which to give the tests in order to include the most students with the least duplica­ tion. The classes used for testing were Harmony, Advanced Harmony, Aural Harmony, Advanced Aural Harmony, and Conduct­ ing. The sizes of these classes ranged from fifty students to over two hundred. None of the classes was normally taken by seniors, although a few seniors were in the various classes being tested. Since the prediction and selection process for music majors was planned for students no later than the first term of their junior year, no effort was made to include seniors in the testing program. At each testing session the students were told that the tests were for research pruposes only and that their anonymity would be protected. In spite of these assurances, the researcher noted some suspicion, especially when students 35 were asked to write their names on the test answer sheets. They were assured that the names merely made it possible for their scores to be matched with their personal data, but no names would appear after the data processing had begun. The researcher also noted some animosity concerning the use of regular class sessions for testing. Several stu­ dents left the class as soon as they discovered that normal classroom work was being interrupted. A few left before com­ pleting the tests, verbally expressing to the researcher their resentment at losing a class session. Others completed the test but stated that their privacy had been invaded, especially by the Edwards Personal Preference Schedule. Attendance at classes is not required by the Univer­ sity, and in the large lecture sessions roll is often not checked. Consequently, attendance at the testing sessions did not approach 100% because normal class attendance figures are not that high. On the other hand, since the testing was not announced in advance, attendance was not lower than u s u a l . There was no reason to suspect any systematic bias in the study due to class attendance. Class absences appeared to be random. Independent Variables To test the hypotheses that any of the dependent variables discriminate among sub-groups of music students, six independent variables were established. The information 36 from the academic files was used in placing the students in the various levels of each variable. The independent variables were Class, Curriculum, Instrument, Transfer Status, Enrollment Status, and S e x . They are each discussed separately below. Class The class of each student was found by checking his official University status at the beginning of fall term, 1971. Since only undergraduates were considered, students were placed into four categories: juniors and seniors. freshmen, sophomores, The senior category was not considered to be a level of this independent variable for two reas o n s : (1) they were not readily available for testing, because no large classes were designed for seniors, and (2) the selection process for music majors is normally completed before the senior year; therefore, scores which seniors make on various tests were not pertinent to the study. The seniors were a part of the study only to the extent that their GPA and SAT scores were used in the various levels of the other independent variables. 37 TABLE 1.— Number of Students in Each Class . Total N ^ Freshman Sophomore Junior N used for analysis of 19 test variables^ 126 35 97 25 117 39 includes Drake, Aliferis, and EPPS explained on pages 49-56. SAT scores were not used by classes because many of those in the junior class were not required to furnish the University these scores when they were admitted in 1969. The ten GPA variables were not used because freshmen, of course, had no GPA's, because they had not yet completed a term. Curriculum Although the students working toward a Bachelor of Music degree are classified in one of eight curricula, some curricula contained too few students for any meaningful sta­ tistical analysis. year degree) Music therapy/music education {a five students were combined with music therapy stu­ dents, because their desire for a therapy degree of any type distinguished them as special. Theory majors numbered less than twenty and therefore were deleted from the study for this independent variable. String specialist majors were 38 combined with music education (instrumental). A fourth category were those students who were music/no major. had not yet declared a music curriculum. These They were deleted from the study for this independent variable. The numbers of students who had no missing data and were therefore used to test the various hypotheses are listed in Table 2. TABLE 2.— Numbers of Students in Each Level of Curriculum. N for 10 Test Var. Curriculum N for 19 GPA Var. N for 2 SAT Var. Applied Music 19 11 37 Music Therapy 31 21 31 Music Education (Choral) 26 21 33 Music Education (Instr.) 37 36 51 113 89 152 Totals Instrument The students were grouped according to the major in­ strument which they studied. Because of some very small numbers for certain instruments (for example, oboe, bassoon, tuba, and euphonium) the major instruments were grouped into families of instruments for the levels of this independent variable. Even after grouped, some families had an insuffi­ ciently large N for analysis. For the purpose of this study, 39 voice is considered an "instrument" and the harp is consid ered a member of the string family. TABLE 3.— Numbers of Students in Each Major Instrument Category. Instrumental Family N for 19 Test Var. N for 10 GPA Var. N for 2 SAT Var. Piano 20 25 45 Voice 20 21 47 Brass 21 10 49 Woodwind 23 24 31 * * 17 84 80 190 Strings Totals *N was too small for statistical analysis Transfer Status Each student's file indicated whether he entered the University as a Freshman or transferred from another college or university. The following three levels of this independent variable were used: (1) non-transfers, or those who entered the University as Freshmen; (2) transfers, or those who entered the University after attending another college or university; and (3) internal transfers, or those who trans­ ferred into the music department after beginning study at Michigan State University in some other discipline or as "no-preference" students. 40 For students in group two, information was also ob­ tained concerning their previous institutions. The numbers of each of these sub-levels were too small for meaningful comparison of dependent variables. TABLE 4.— Numbers of Students According to Transfer Status. Transfer Status ?estv“ . Non-transfers 74 Transfers 25 Total 99 SAT scores were not used for this independent varia­ ble because transfer students do not furnish these. GPA's were also not used, because grades earned at other institu­ tions are not necessarily comparable to those earned at Michigan State University. Enrollment Status Perhaps the most important of the independent varia­ bles concerns the students' continuance as music students. The criteria used was the enrollment status of each student during the Spring term of 1973. If a student who was in the study during the Fall of 1971 was still a music student in May of 1973, he was classed as Enrolled. If he was not 41 enrolled in the department for any reason other than having graduated, he was considered to be Non-enrolled. A total of eighty students were found to be Non-enrolled. The student may have been Non-enrolled during the Spring term of 1973 for a number of reasons. dropped out for personal or academic reasons. He may have He may have been suspended by the University for academic reasons . He may have changed his major to another University department or he may have transferred to another music department. Ac­ cording to an informal survey of selected non-enrolled stu­ dents, however, most of the students in this category trans­ ferred to other departments within the University. TABLE 5.— Numbers of Students in Each Level of Enrollment Status. N for 19 Test Var. Status N for 2 SAT Var. Enrolled 81 59 Non-enrolled 18 43 The G P A ’s were not considered for this independent variable because the Non-enrolled students all had incomplete GPA's. 42 Sex The final independent variable was sex. Table 6 gives the N's for each level of each independent variable. TABLE 6.— Numbers in Each Level of Sex. N for 19 Test Var. N for 10 GPA Var. N for 2 SAT Var. Male 41 41 82 Female 65 62 123 Dependent Variables- -Course Grades Thirty-one dependent variables were used in this study. They are divided into two categories: scores. course grades and test Ten of the dependent variables were grade averages (GPA's) for various combinations of courses. In cases where only some of the courses in a group had been taken before the term of data collection (Fall, 1971), the scores available in the group were averaged. When only one course in a group had been taken, that grade became the GPA. All grades were averaged to one decimal place. The grading scale at Michigan State University is a numerical scale from 0.0 (failing) to 4.0 (superior). complete list of numerical grades is as follows: 1.0, 1.5, 2.0, 2.5, 3.0, 3.5, and 4.0. The 0.0, 0.5, The University author­ ized a grade of 4.5 for exceptional work in a course, but 43 its use was limited. Fewer than 1% of music grades were 4.5. Students who had earned credit and received grades before Fall of 1968 had letter grades. There were converted by the researcher according to the following formula: F = 0.0? D = 1.0; C = 2.0; B = 3.0; and A = 4.0. Two more grades were encountered frequently during the data gathering process. The incomplete grade, "I," was considered missing data unless the transcript showed that the course had been completed. substituted for the "I." In that case, the new grade was The pass grade, "P," is used pri­ marily for performing ensembles and therefore did not enter into any GPA calculations. Listed below are the ten grade variables along with a brief description of the content of each course. Variable 1. Harmony Music students at the University, regardless of the specific music curriculum, take a sequence of three courses in the Freshman year called Harmony and Music 182). (Music 180, Music 181, The content is similar to many "Freshman Theory" courses at a great number of other music schools. The first few weeks are devoted to fundamentals of written music, including scales, intervals, triads, keys, clefs, and modes. The remainder of the academic year is spent with 44 traditional harmony, chromatic harmony, and a survey of post-Bach (Classic and Romantic) idioms and forms. The class meets in a large group lecture situation (usually larger than 150 students) three times a week. In addition, the students attend a small group "lab" one hour per week where they receive individual help, and study key­ board harmony. Variable 2. Harmony Advanced This three term sequence of courses (Music 280, Music 281, and Music 282) is what is often called "Sophomore Theory." The course content includes advanced chromatic harmony, im­ pressionism, and twentieth century harmony. The lecture-lab arrangement is similar to Harmony, except that the lecture group is usually fifty per cent smaller due to attrition from the Freshman to the Sophomore year. Variables 3. and 4 . Aural Harmony and Advanced Aural Harmony All music majors must take two full years of the Aural Harmony sequence, which includes Aural Harmony and Advanced Aural Harmony (Music 183, Music 184, Music 185, Music 283, Music 284, and Music 285). Each course is one credit total­ ing six credits during the Freshman and Sophomore years. The class is designed to run concurrently with Harmony and Advanced Harmony. Students meet twice each week in a 45 large group and also receive additional drill during the one hour lab, discussed earlier with Harmony. In addition, students spend much time with a program of tape recordings designed to supplement class drill. The course content in­ cludes melodic and harmonic dictation, sight singing, and aural recognition of intervals, scales, and chords. The two years of Aural Harmony and Advanced Aural Harmony are generally considered by the students and faculty to be among the major barriers facing music students, particu­ larly those judged as being "weak" musically. According to the advisors in the music academic advising office, a number of students drop out of the department on the basis of their inability to pass these courses. Variable 5 . Applied 1 Regardless of the specific music curriculum, each music major takes applied music lessons each term of the four years of the degree. For this dependent variable, all grades for the first three terms of study on the major in­ strument were averaged. Grades for class instruments and private study of minor instruments were not included. Variable 6 . Applied 2 For this dependent variable, all grades for the fourth, fifth, and sixth terms of private study on the major 46 instrument were averaged and recorded. Class instruments and minor instruments were not included. Variable 7 . u College 1 All students at the University, regardless of major or curriculum, are required to complete four series of courses in order to graduate. (Some students, those in the Honors program in particular, have certain courses waived and/or others substituted.) Two of these series are normally taken by music majors in the first year of college. Because the courses are administered by the University College, they are referred to in this paper as U College courses. The two series are: Natural Science ................. 12 credits (Three 4 credit courses) American Thought and Language . 9 credits (Three 3 credit courses) The grade point average for any or all of these six courses was computed for the value of the dependent variable U College 1 , regardless of the term taken. Variable 8. U College 2 Students in all of the music curricula at the Univer­ sity are advised to take the required University College course called Social Science in the Sophomore year. This series includes three courses, four credits each, for a total of 12 credits. The GPA for these three courses, or those 47 substituted for them, was computed for this dependent varia­ ble. The computation was made regardless of the year in which the courses were taken, although approximately 75% of the music students took these courses in the Sophomore year. Variable 9. Cumulative 1 This dependent variable was simply the overall GPA at the end of the student's last term as a Freshman, includ­ ing all courses taken up to that point. It includes, of course, certain courses for which separate GPA's were com­ puted for other dependent variables. In that sense it is not a completely new variable, but a combination of certain variables with some new information (i.e., courses) added. Variable 1 0 . Cumulative 2 This variable is similar to Cumulative l except that it includes all courses through the student's last term as a Sophomore. Both Cumulative 1 and Cumulative 2 appear on official University grade reports sent to the department each term to be filed in the academic files. Dependent Variables— Test Scores In addition to the ten grade variables, there were twenty-one test variables. Nineteen of these were scores from various tests administered by the researcher, and two 48 were from standardized academic aptitude tests. The twenty- one test score variables will be discussed below. Because the previously discussed course grade variables are also dependent variables, the numbering of dependent variables will continue from the previous list. Variables 11. and 12. SAT-V and SAT-Q Since its inception in 1926, the College Board Scholas­ tic Aptitude Test*- has been one of the most prestigious and popular tests of mental ability. While the College Entrance Examination Board is responsible for the SAT, most of the operations, development, updating, and statistical analyses are carried out by the Educational Testing Service, Prince­ ton, New Jersey. The extent of interest in the SAT as a predictive tool is shown by the bibliography in The Seventh Mental Measure2 ments Yearbook article on the S A T . Over 400 pieces of re­ search are reported and at least 30 0 of these concern its use as a predictor, usually of success in college work as measured by grades. The SAT is basically an aptitude test for college study, although it measures a certain amount of achievement in academic studies. It yields two scores: Verbal (SAT-V) Scholastic Aptitude Test (Princeton, New Jersey: College Entrance Examination Board, 1961). 2 Oscar Buros, The Seventh Mental Measurements YearBook (Highland Park, New Jersey: The Gryphon Press, 1972). 49 and mathematical, sometimes called quantitative (SAT-Q). The SAT-V is related to social, political, scientific, artistic, philosophical, and literary areas. The SAT-Q requires as background only the math taught in grades one through nine, but also measures reasoning ability, logical ability, and the 3 perception of mathematical relationships. The SAT is administered to high school students usually early in their senior year. The Educational Testing Service supervises the nationwide administration and sends the results to the colleges and universities designated by the student. Students wishing to enter Michigan State University as Fresh­ men are advised by the University to take the SAT and have the scores submitted. Transfer students generally do not furnish SAT scores. The norms for the SAT were established by the over 10,000 students who took the test in April of 1941. A score of 500 and a standard deviation of 100 was established for the norm for both the SAT-V and SAT-Q. the combined test (SAT-V + SAT-Q) The mean score for for students entering the University in a recent year was 1002, but this may vary slightly from year to year. Other institutions report com4 bined scores from less than 700 to more than 1400. ^I b i d . ^Alexander W. Astin, Predicting Academic Performance in College (New York: The Free Press, 1971), 54. 50 The University has been requiring SAT scores for enter­ ing Freshmen since Fall of 1969. For that reason, only 207 students in the study had these two scores. Transfer students and those entering as Freshmen before Fall of 1969 did not have the scores. Variable 13. The Drake Musical Apti­ tude Tests The Drake test was included in the test battery to provide some measure of music aptitude. According to the test manual, the Drake Tests "have been designed to provide consistent and valid measures of musical aptitudes, and to render sound evidence regarding an individual1s potential 5 for a successful career in music." The tests can be used with subjects of any age above eight years. It has been demonstrated that training has no significant effect on Drake scores, giving credence to the claim that it is a true g aptitude test. The Drake Tests are actually four t e s t s , including two equivalent forms measuring musical m e m o r y , and two non­ equivalent forms measuring rhythm. Of the four tests, only form B of the rhythm test was used in this study for two reasons: (1) the melodic portion of the Aliferis test Drake Manual, op. c i t . ^Edwin Gordon, "A Study to Determine the Effects of Training and Practice on Drake Musical Aptitude Test Scores" (Unpublished Doctoral Dissertation, University of Iowa, 1958). 51 (discussed on page 52) is similar to the two. musical memory tests of the D r a k e , and (2) the test manual suggests that musically trained subjects be allowed to skip form A of the rhythm tests since it is easier than form B. The actual test used (Rh y t h m , form B, hereafter called the Drake) consists of fifty items. Each item is exactly the same type, measuring the ability to maintain a steady beat in spite of distracting aural stimuli. In each item a particu­ lar beat is given, and the listener counts to himself at the same rate. Following cessation of the presented beat, the listener continues to count until told to stop. He then writes the number he has reached at that point. A distract­ ing beat at another tempo is heard following cessation of the original beat. The test manual gives the rationale for this type item: Drake's Rhythm Test has been constructed on the principle that a test of musical aptitude should ap­ proximate life-like musical situations as closely as possible. The author has taken the premise that a successful musical performer (1) must feel rhythm strongly, (2) must be able to maintain a set tempo despite distractions, and (3) must be able to main ­ tain accurately a set of tempo before he can take musical liberties with rhythm in the form of rubato, ^ accelerando or deviations from equally divided beats. Each subject records his answers on an answer sheet. The answers are hand scored by simply noting the difference 7 Drake M a n u a l , op. cit. 52 between the correct answer and the given answer. For example, if the correct answer is "eleven" and the student's answer is "nine," he received a score of "two" on that item. The total of all fifty item scores gives the overall score. A perfect score is zero, meaning that there were no differences between the correct answers and the given answers. In this study, all scores were subtracted from one hundred so that higher numbers would mean better scores. Variables 14. 15. 16. A1 Mel, Al Har, Al Rhy g The Aliferis Music Achievement Test, by James Aliferis, is a three section test, consisting of Melodic, Harmonic, and Rhythmic sections, each yielding one score. Section one deals with the recognition of melodic elements and idioms. The subject hears a melodic interval or four note melody and is asked to match what he hears with one of four choices given in musical notation. The second section of the test deals with harmonic elements and idioms. In the first part of the Harmonic sec­ tion the subject matches four-note piano chord with written notation. In the second section the problem is the matching of sequences of three chords with written notation. The third section, Rhythmic, measures the subjects' ability to match aural rhythmic elements and idioms with a o Manual James Aliferis, Aliferis Music Achievement Test (Minneapolis: The University of Minnesota Press, 1954). written pattern. The author defines a rhythmic element as a rhythmic figure of one beat duration. a combination of two elements. A rhythmic idiom is The examples played in either case consist of a C major scale in a variety of rhythms. Sixty-four test items are included, with twenty-six, eighteen, and twenty items in each section respectively. The test booklets contain brief but clear instructions and also the actual written musical examples for each test item. The booklet has a space for each answer and normally must be hand scored. In this study, machine scored mark sense sheets were adapted and seemed to work satisfactorily for the students, and also eliminated tedious hand scoring. Although the use of separate answer sheets is undoubtedly a slight inconvenience for the subject, there is no evidence that this would lower the reliability of the test. Students at the University regularly use machine scored answer sheets. The test items are furnished by the publisher on tape, recorded at 7 1/2 i p s . The use of this tape causes each test administration to be consistent and is much more satis­ factory than playing each item on a piano. High fidelity audio equipment was used for playing the tape. Variables 16-31. EPPS The Edwards Personal Preference Schedule (EPPS) was designed to provide convenient measures of a number of rela­ tively independent normal personality variables. Some other 54 instruments, such as the Minnesota Multiphasic Personality Inventory, are designed to measure abnormal personality traits and are useful in clinical settings. The EPPS, on the other hand, has been shown to be useful for research and counseling of relatively normal subjects. The EPPS provides measures of 15 manifest needs. The name of each variable along with a summary of the manifest needs associated with each, is listed below: !• Achievement, To do one's best, to be successful, to accomplish tasks requiring skill and effort, to be recognized authority, to accomplish some­ thing of great significance, to do a difficult job well, to solve difficult problems and puzzles, to be able to do things better than others, to write a great novel or play. 2. Deference.— To get suggestions from others, to find out what others think, to follow in­ structions and do what is expected, to praise others, to tell others that they have done a good job, to accept the leadership of others, to read about great men, to conform to custom and avoid the unconventioanl, to let others make decisions. 3. Order.— To have written work neat and organized, to make plans before starting on a difficult task, to have things organized, to keep things neat and orderly, to make advance plans v;hen taking a trip, to organize details of work, to keep letters and files according to some system, to have meals ^-_anized and a definite time for eating, to have things arranged so that they run smoothly without change. 4. Exhibition.— To say witty and clever things, to tell amusing jokes and stories, to talk about personal adventures and experiences, to have others notice and comment upon one's appearance, to say things just to see what effect it will have on others, to talk about personal achievement, to be the center of attention, to use words that others do not know the meaning of, to ask ques­ tions others cannot answer. 55 5. Autonomy.— To be able to come and go as desired, to say what one thinks about things, to be inde­ pendent of others in making decisions, to feel free to do what one w a n t s , to do things that are unconventional, to avoid responsibilities and obligations. 6. Affiliation.— To be loyal to friends, to participate in friendly groups, to do things for friends, to form new friendships, to make as many friends as possible, to share things with friends, to do things with friends rather than alone, to form strong attachments, to write letters to friends. 7. Intraception.— To analyze one's motives and feel­ ings , to observe ot h e r s , to understand how others feel about problems, to put one's self in an­ other's place, to judge people by why they do things rather than by what they do, to analyze the behavior of oth e r s , to analyze the motives of others, to predict how others will act. 8. Succorance.— To have others provide help when in trouble, to seek encouragement from others, to have others be kindly, to have others to be sym­ pathetic and understanding about personal problems, to receive a great deal of affection from others, to have others do favors cheerfully, to be helped by others when depressed, to have others feel sorry when one is sick, to have a fuss made over one when hurt. 9. Dominance.— To argue for one's point of view, to be a leader in groups to which one belongs, to be regarded by others as a leader, to be elected or appointed chairman of committees, to make group decisions, to settle arguments and disputes between others, to persuade and influ­ ence others to do what one w a n t s , to supervise and direct the actions of others, to tell others how to do their jobs. 10. Abasement.— To feel guilty when one does something wrong, to accept blame when things do not go right, to feel that personal pain and misery sufferred does more good than harm, to feel the need for punishment for wrong doing, to feel better when giving in and avoiding a fight than when having one's own way, to feel the need for con­ fession of errors, to feel depressed by inability 56 to handle situations, to feel timid in the presence of superiors, to feel inferior to others in most respects. 11. Nurturance.— To help others when they are in trouble, to assist others less fortunate, to treat others with kindness and sympathy, to forgive others, to do small favors for ot h e r s , to be generous with others, to sympathize with others who are hurt or sick, to show a great deal of affection toward ot h e r s , to have others confide in one about personal problems. 12. Change.— To do new and different things, to travel, to meet new people, to experience novelty and change in daily routine, to experi­ ment and try new thi n g s , to eat in new and different pla c e s , to participate in new fads and fashions. 13. Endurance.— To keep at a job until it is finished, to complete any job undertaken, to work hard at a task, to keep at a puzzle or problem until it is solved, to work at a single job before taking on others, to stay up late working in order to get a job done, to put in long hours of work without distraction, to stick at a problem even though it may seem as if no progress is being made, to., avoid being interrupted while at work. 14. Heterosexuality.— To go out with members of the opposite sex, to engage in social activities with the opposite sex, to be in love with some­ one of the opposite sex, to kiss those of the opposite sex, to be regarded as physically at­ tractive by those of the opposite sex, to parti­ cipate in discussions about sex, to read books and plays involving sex, to listen to or tell jokes involving sex, to become sexually excited. 15. Aggression.— To attack contrary points of view, to tell others what one thinks about them, to critize others publicly, to make fun of others, to tell others off when disagreeing with them, to get revenge for insults, to become angry, to blame others when things go wrong, to read newspaper accounts of v i o l e n c e . 9 9 A. L. Edwards, Edwards Personal Preference Schedule Manual (New York: Psychological Corporation, 1954T. 57 The EPPS is published as a booklet containing 225 pairs of statements and a separate answer sheet. The sub­ ject is asked to read each pair of statements and mark the answer sheet according to the statement which is more char­ acteristic of himself. A typical item is given below: a. I like to be the center of attention in a group. b. I like my friends to make a fuss over me when I am hurt or sick. If neither statement especially appeals to the subject he is asked to make a "forced choice/" since each item must be answered for the final scores to be meaningful. According to the test manual, most college students finish the EPPS in about forty minutes, although there is no time limit. The times for ranged from twenty-five to taking the EPPS in this study sixty-five minutes. The test may either be hand scored or machine scored, depending on the type of answer sheets used. For reasons of economy, the researcher chose to hand score the EPPS. Answer keys are provided with the test manual, and scoring inaccur­ acies are minimal. Reliability Split-half reliability coefficients, or coefficients of internal consistency, were computed for the nineteen test scores used as dependent variables. the Spearman-Brown formula. These were corrected by Both the reliabilities computed in this study and those reported by the publishers of the various tests are given in Table 7. 58 TABLE 7.— Reported and Obtained Reliabilities. Reported Reliability Obtained Reliability Aliferis Mel. .84 .81 Aliferis Har. .72 .47 Aliferis Rhy. .67 .78 Variable Drake .60 - .90 .88 EPPS Achievement .74 .90 Deference .60 .80 Order .74 .79 Exhibition .61 .77 Autonomy .76 .74 Affiliation .70 .80 Introspection .79 .91 Succorance .76 .92 Dominance .81 .90 Abasement .84 .89 Nurturance .78 .78 Change .79 .91 Endurance .81 .88 Heterosexuality .87 .89 Aggression .84 .92 59 Handling of Missing Data The lack of complete score profiles for each student resulted in missing data. For instance, if a student did not attend the testing session when the Drake was administered, and that student also had not taken Harmony, the Drake and the Harmony GPA were considered missing data for that student. Simple statistical procedures such as means, standard deviations, and correlations can be computed for the entire sample of students, regardless of missing data. The N for each variable changes according to the number of students for whom the value is present. Simple correlations can be computed by means of a special "Missing Data" computer pro­ gram which ignores a particular dependent variable when it is missing on either of the independent variables being correlated. No computer program was available which would handle missing data in a multivariate analysis of variance. For this statistical procedure it was necessary to use only those students who had no missing data on the variables under consideration. For instance, ninety-nine students had no missing data on the Alife r i s , D r a k e , and E P P S . Only these students were used for that particular multivariate analysis. Whenever missing data occurred, the student was not in that particular sample, although he might be included in 60 other samples. The sample of those who took the tests was affected by three factors: class were eliminated; (1) those who did not come to (2) those few who left and refused to complete the test were eliminated; and (3) those who were not taking a class in which testing took place were eliminated. Furthermore, those who had not taken certain courses or se­ quences of courses were eliminated from the samples using G P A 's . Analysis Raw data for the various dependent variables were punched onto computer cards, along with the student's coded identification number and numbers identifying the student in each independent variable level. used: Two computer programs were a Missing Data program supplied means, standard devia­ tions, and simple correlations for each of the levels of the independent variables. Finn Another program, written by Jeremy (1968), State University of New York at Buffalo, was used for multivariate analysis. Multivariate Tests A multivariate F test is capable of handling any number of dependent variables at one time. mean vectors It tests for equality of (ui = u2) taking into account any correlation among the variables. The primary advantage of the single mul­ tivariate test over 'a number of univariate F tests concerns 61 the alpha level. The large number of univariate tests would raise the alpha level, greatly increasing the possibility of finding significant differences merely by chance (a Type I error). The degrees of freedom for the multivariate tests re­ ported in this study do not conform to the usual pattern of univariate tests. In the univariate test, the total degrees of freedom equals N-l, and this total is divided into dfbetween and dferror* The de9rees of freedom reported on the tables in Chapter IV are approximations of the univariate degrees of freedom appropriate for each multivariate test. It is possible for the multivariate test to show significance while the various univariate tests do not. occurred only one time in this study, however. This In most cases one or more of the univariate tests are significant when multivariate significance is found. It is only with great caution, however, that one can place much faith in the signifi­ cance of the univariate tests, error is still present. for the chance of a Type I Any significance found in the uni­ variate tests merely points to an area for further research. It is also possible for the multivariate test, not to be significant while one or more of the associated univariate tests are highly significant. In that case, the common practice is to declare that significance was not found, re­ gardless of the univariate results. 62 Alpha Levels Since this was an exploratory study, it was important not to overlook any possible variable which might be promis­ ing. To fail to find any differences among the samples is called a Type XX error, and setting a high alpha level re­ duces the chance for a Type II error. For that reason, the alpha level was set a priori at the .05 level. Any more stringent level, such as .01 or .005, would increase the chances of a Type II error. CHAPTER IV PRESENTATION OF THE DATA This chapter consists of three sections. The first concerns the correlation matrix presented in Table 4.1. Each of the more interesting correlation coefficients is discussed and some possible explanations of unusual coefficients are given. The second section is a review of each of the six hypotheses and summaries of the appropriate statistical pro­ cedures used to test each hypothesis. The third and final portion of the chapter consists of tables and means for the various levels of each independent variable. These means were computed using the "Missing Data" computer program and have varying "N's" depending on data available. No means were included when N was less than ten. Because the multivariate F tests used to test the hypotheses required complete data on each student, the N's were smaller for the multivariate tests. The exact N's used were given in the Tables in Chapter III. Correlations Pearson Product Moment Correlations were computed on all thirty-one dependent variables using the Missing Data 63 TABLE 8.— Correlations. EAT V e r b a l 1 . CO S*T '. M a r t . .1 5 i . ;a .it . 35 .44 .'ii .ti .64 . 4. 2 . 19 . b6 I . no .0? .2 9 .2 * .2 1 .4« .S t .t 1 . ?'J .34 .4J j :t7 .OS .09 -.31 ,17 -.01 .11 . J& .14 J C aliche A pplied Cvo*. 2 2 2 i 'a r . ; . -.01 A c h i e v e r y r.t -IS . o7 1 .0 0 - .06 .Cl .25 . ) 1 .53 . 45 .41 .4 9 . 65 ,C6 .06 .20 .0 0 .17 . 18 . 10 .0 6 . 16 -,04 -.12 .u .19 .23 .04 .1 5 .1 1 .1 1 .24 . 16 . 30 -.04 -.11 -.0 4 .0 2 .07 .1 2 . 15 . 10 .2 4 .07 .07 ,n ,08 -.02 -.01 -.11 - .1 1 -13 - , Dft .02 .2 0 .06 . n . 14 .1 6 .*0 .01 -.0 5 .0 8 . 04 -.02 . 08 .04 -.1 1 -C.5 •..'4 * . 1B . -.28 -.27 -.26 -.2 4 -.37 -.15 -.4 1 -.44 .14 .13 .04 .CIO - . 35 -.o ; -. P -.35 -.2d .1 j .017 -.06 - . 10 . 02 -.07 --29 -.05 . 11 -.01 +.24 . 3h -.06 Osd er -.13 -.1 0 C 'h ifc it. -.11 -.06 V j-r.rvry .1 9 -.11 . 32 » .n -.0 3 .0 9 . 11 l.o o .2 2 1 .00 1.00 1 1 .0 0 1.00 l .00 .06 . L4 .V i .16 . 1I 1 - . 01 .1 2 .21 .1 0 ■OJ .09 -.12 .0- -.01 -.31 -.07 .0 1 .0 5 -.71 .04 -.04 -.m ••01 -.25 i ,oo - . .-4 .1.3 - .09 -.12 -.01 -.01 -.05 -.19 - . 14 -.13 .0 5 -.10 --1 J -.06 .06 -.10 .U» - .u - . £1* .22 .0 4 -.2 4 -.1 5 '.1 5 I . Jii -.14 -.C d - .10 -.18 -. 2J -.16 -.50 - . 34 - . J2 - , 39 -.17 -.02 -.15 ,09 .1 5 -.31 .1*. . 13 -.29 -.2 2 ,r . b - .02 -.01 *.2 6 i . j ' (j .21 .0 4 -.1J -.14 -.24 -.14 - . 10 -.32 -.08 - .2ft -.11 .1 2 . 1j -.02 . 13 - . 12 .0 4 - . 10 -.29 -.21 .(0 . - . 2 »■ . 14 -.0 6 .1 0 .1J ,o . '. n - .09 ,on . 70 i .no -.14 -.it -.01 -. ji -.0 5 .0 4 *.1 9 - .10 1 . Oi l E j 7 « n ^ . Abdterent -U .42 1. 0 3 -.- 7 i .00 .11 . 17 -.01 ..;7 t“ tro » K C ft. Kartutisnce . 19 .01 .41 -.21 LA iMr.ifi.re -.12 I.L1C - . 10 -.12 \ r f : 1u t icn 1.00 -.01 .0 9 fV ittri'nc t: .2 3 i .o o . 31 . !6 c . If. 1 ,.M) C h a ng e -.C b -.12 - .08 -.2 6 -.20 .01 * . 16 -.21 -.3 5 -.11 -.29 - . 47 .0 5 .OC -.05 -.0 6 - .20 . 01 -.O B r.n d jra rs r e -.01 . 01 - 14 .2 * .20 . 12 .23 . 31 .2 4 .27 .2 9 . J7 .0 9 .0 0 .03 .20 .1 3 .01 .29 -.17 rotex . .12 .00 -.11 -.0 ft -.11 . 14 -.13 .77 .13 -.a j -.05 -.04 -.1 8 -.03 .00 -.03 -.04 -.20 -.12 .12 . 19 - . 19 -.30 ,0 b -.01 -.18 -.2 5 -.0 1 - . 14 I .0 A ^ r e i s id n .02 . 00 -.02 -.14 -.02 .1 0 .0 6 -.10 -.08 -.26 .15 -.11 -.07 .04 .11 -.57 .0 5 - . 31 -.10 . 18 .25 -.31 -.28 .0 0 .22 -.14 -.11 -.71 - . 0 rf , 1 V c - Het« « J % >. :* h 5 •J H 5 j £ = < s S’ S 2 C o i. < E •j < - h> O u 9 = - < a u 3 r* ■p r a < c - z S 2 L a c — p = - x V 6 ■4 * < «; a i < £ u t) 5 £ 2 < < 3 " - £ i 3 ■5 3 b < j» 3 ?’ T -i c -3 -* i = /hc^icfttiun lU rrcr.y 65 computer program. This program uses whatever data are avail­ able in computing each value in the correlation matrix. When a subject has no data available for the variables being corre­ lated, he is ignored for that computation, but is included in the other computations. This procedure allows all avail­ able subject data to be used, greatly increasing the N and therefore the validity of each correlation in the matrix. Interesting or unusual correlations are discussed below. All correlations not discussed were considered to be "normal" or "expected." SAT-V Although high correlations might be expected with the more academic-type studies, such as U. College courses or Harmony, SAT-V also had a strong relationship with Advanced Aural (.62). The correlation with Applied 1 (-.05) lowest with a GPA. is the Apparently, SAT-V is much more closely related with "classroom" grades than with applied music g r a d es. SAT-Q This variable had positive but slightly lower corre­ lations with GPA's than did SAT-V. with Applied The very low correlation (.07) was also similar to that of the SAT-V. All other correlations with SAT-Q were considered normal. 66 Aural (Aural Harmony) The correlations are fairly high with Harmony and Cum 1 (.63). (.61) The lower correlations with Applied 1 (.28) indicate that Aural Harmony grades are more closely related to "academic" GPA's than to private lesson GPA's. The highest correlation is with Ad. Aural, indicating a tendency for students to continue making approximately the same grades in the second year of Aural Harmony as the first. The Aliferis variables correlated positively with A u r a l, with Al. M e l . (.66) being the highest of the three. Harmony The highest correlations were with Cum 1 (.79) and Cum 2(.74) indicating that Harmony grades give a good in­ dication of a student's overall academic potential. The low correlation with Al. H a r . may be due to the fact that Harmony class is not as aurally based as are the Aliferis tests. U College 1 (University College Fresh­ man courses) High positive correlations were found with Cum 1 (.84) and Cum 2 (.83). Moderate correlations were found with the other GPA's, Applied 1 (.23) being the lowest. All other Correlations with U College 2 were considered normal. 67 Applied 1 Very low correlations were found with other variables with the exception of Applied 2 (.67). Applied grades seem to be the most independent among all GPA's, suggesting that ability to make a high grade in a private lesson situation is unique when compared to other academic abilities. Cum. 1 (Freshman cumulative GPA) This variable is a combination of A u r a l , Harmony, U College 1 , and Applied 1 plus any other courses taken in the first year. For that reason, the high correlations with grades shown in Table 8 are to be expected. Al. Mel. (Aliferis Melodic Sub-test) A high positive correlation was found with Aural (.66) and lesser positive correlations were found with Ad. Aural (.44), Al. H a r . (.55) and A l . R h y . (.34). All others were considerably lower. Al. H a r . (Aliferis Harmonic Sub-test) The highest correlation was with Al. M e l . (.55). Per­ haps the reason so many low correlations occurred with this variable was its own low reliability (r = .47). 68 A l . R h y . (Aliferis Rhythmic Sub-test) The low negative correlation with the Drake (-12) in­ dicates that the two tests measure different aspects of rhythm. An examination of the two tests reveals that the Aliferis is closely bound to the matching of aural perceptions to written score, whereas the Drake measures the ability to maintain a steady beat without regard to written rhythm. Drake None of the correlations with the Drake was higher than .30 (Applied 2) and the rest were lower than .20. This indicates the uniqueness of the trait which the Drake is measuring. EPPS (Edwards Personal Preference Schedule) Among the EPPS variables, four consistently correlated highly with GPA's, but none had much relationship with any of the test variables. Achievement was positively correlated with all of the GPA's except Aural. Affiliation, Abasementf and Nurturance correlated negatively with all GPA's. Per­ haps these scales hold some promise in the search for pre­ dictors of academic success. 69 Hypotheses Hypothesis 1 Music students who leave the music department for any reason except graduation will differ on one or more variables from those who remain. Students were classified enrolled if they were still working for a music degree in April, 1973, 18 months after the study began. All those who had left for any reason ex­ cept graduation were non-enrolled. Separate "t" tests were performed on each SAT tests for the two groups. The results appear in Table 9. TABLE 9.— Means and Obtained "t" value for the SAT for students grouped by Enrollment Status. N enrolled SD mean N non-enrolled mean df SD t P SAT-V 59 513 89 43 520 106 110 .39 <.35 SAT-Q 59 499 102 43 514 123 110 .65 <.30 A multivariate F test was performed on nineteen test variables. These include the three Aliferis scores, the Drake scores, and fifteen scores from the EPPS. Table 10 contains the results of that multivariate analysis of variance. 70 TABLE 10.— Multivariate ANOVA for Nineteen Test Variablesa for Students Grouped by Enrollment Status. Source dfb Between 19 Within 79 Multivariate F 1.0072 P <.4627 aThe nineteen test variables include the Drake, Aliferis, and EPPS. The degrees of freedom for multivariate F tests are obtained from the computer and represent an approximation of the appropriate univariate degrees of freedom. (Rao*s approxi­ mation .) Both the "t" tests and the multivariate F tests indi­ cate a rejection of the hypothesis that enrolled students differ from non-enrolled students. No significant differences were found at the .05 level. The ten grade variables were not used to test this hypothesis because a larger number of the non-enrolled stu­ dents left before the end of one year and had too few grades for completed courses. Throughout this chapter actual "P" values, or proba­ bilities, will be given. This is more exact than merely stating, for example, that "the test is significant at the .05 level of confidence." 71 Hypothesis II Music students differ on one or more variables when grouped according to the specific music curriculum in which they are enrolled. The levels of the independent variable curriculum in­ cluded music education choral, music education instrumental, applied m u s i c , and music therapy. Other curricula categories contained too few students for analysis. All thirty-one dependent variables were used in testing this hypothesis. Analysis of variance was used to test the two SAT variables. The results appear in Table 11. TABLE 11.— Multivariate ANOVA for the Two SAT Tests for Students Grouped by Curriculum. Source df Between 6 Within Multivariate F 2.6354 P <.0167 294 A multivariate F test was performed using the ten grade variables. These included A u r a l , Harmony, Applied 1 , U College 1 , Cum 1 / Ad A u r a l , Ad Harmony, Applied 2 , U College 2, and Cum 2. The results appear in Table 12. 72 TABLE 12. — Multivariate ANOVA for the Ten Grade Variables3 for Students Grouped by Curriculum. Source df Multivariate F Between 20 1.1543 Within P <.3034 140 aThe ten grade variables are listed on pages 43-47, and on the preceding page. A multivariage F test was performed using the nine­ teen test variables and the four levels of the dependent variable curriculum. The results appear in Table 13. TABLE 13. — Multivariate ANOVA for the Nineteen Test Variables for Students Grouped by Curriculum. Source df Multivariate F Between 57 1.3154 Within P < .0904 179 Of the three multivariate tests, only the one dealing with the SAT scores showed significance (p = .0167). This was sufficient, however, to accept the hypothesis that stu­ dents differ on one or more variables when grouped by the specific music curriculum in which they are enrolled. The procedure after finding significance on the multivariate F test is to make a guarded examination of the univariate tests 73 of the variables which comprised the multivariate test. The results, as they appear in Table 14, suggest that the SAT-V scores contributed more heavily to the significance of the multivariate test than did the SAT-Q. TABLE 14. — Univariate ANOVA Results for SAT Tests for Students Grouped by Curriculum. Variable Univariate F P SAT-V 3.8923 <.0104 SAT-Q 2.2300 < .0872 Hypothesis III Music students differ on one or more variables when they are grouped according to sex. For the testing of this hypothesis all thirty-one dependent variables were analyzed using three multivariate F tests. The results of the three F tests appear in Tables 15, 16, and 17. TABLE 15.— Multivariate ANOVA for Ten Grade Variables for Students Grouped by Sex. Source df Multivariate F Between 10 .7886 Within 92 P <.6255 74 TABLE 16.— Multivariate ANOVA for Nineteen Test Variables for Students Grouped by Sex. Source df Multivariate F Between 19 2.0176 Within 86 P <.0153 TABLE 17. — Multivariate ANOVA for the Two SAT Tests for Students Grouped by Sex. df Source 2 Between Multivariate F 8.2362 P <.0004 202 Within An examination of the results of the three multivar­ iate tests reveals that two of them are significant at the .05 level. A closer look at the associated univariate Analysis of Variance F tests indicates that several were significant and probably contributed to the multivariate significance. The various univariate tests which are significant are summar­ ized in Table 18. 75 TABLE 18.- -Univariate ANOVA Results when Students are Grouped by Sex. Univariate F Variable P Level with Higher Value Al Mel 3.8377 <.05 Females Deference 4.7336 <.03 Males Endurance 3.7602 <.05 Females Het Sex 3.6734 <.05 Males SAT-Q 8.5465 <.01 Males Hypothesis IV Music students differ on one or more variables when grouped according to their major instrument. There were four levels of the independent variable instrument: piano, v o i c e , b r a s s , and woodwinds. families of instruments The other (organ, strings, and percussion) tained too few students for analysis. con­ All thirty-one depend­ ent variables were analyzed by using three multivariate F tests. The results of the three tests appear in Tables 19, 20, and 21. 76 TABLE 19.— Multivariate ANOVA for the Ten Grade Variables for Students 'Grouped by Major Instrument. df Source Between Within Multivariate F 2.1252 30 P <.0013 197.3342 TABLE 20. — Multivariate ANOVA for Nineteen Test Variables for Students Grouped by Major Instrument. Source df Multivariate F Between 57 1.2481 Within P <.1379 185 TABLE 21. — Multivariate ANOVA for the Two SAT Tests for Students Grouped by Major Instrument. Source Between Within df 8 Multivariate F 2.1490 P <.0308 366 On the basis of the three multivariate tests, the hypothesis that music students differ on one or more variables when grouped according to their major instrument is accepted. Two of the three multivariate tests showed significance at 77 the .05 level. An examination of the individual univariate tests reveals two variables which show promise. The univar­ iate results are summarized in Table 22. TABLE 22.— Univariate ANOVA Results when Students are Grouped by Major Instrument. Variable Univariate F P Harmony GPA 6.0031 <.0013 Ad. Harmony GPA 3.7452 <•0145 An examination of Table 26, page , reveals that string players probably account for the significance of the Harmony and A d . Harmony differences. Although the multivariate F test for the two SAT scores was significant (p = .0308), neither of the individual scores was significant alone. Hypothesis V There will be no differences among music students on the Aliferis, D r a k e , or EPPS when students are grouped by class. There were three levels of the independent variable class: Freshmen, Sophomores, and Juniors. There were a total of nineteen dependent variables which were scores and sub­ scores on the three tests. A multivariate F test was per­ formed and the results are in Table 23. 78 TABLE 23.-— Multivariate ANOVA for Nineteen Test Variables for Students Grouped by Class. Source df Multivariate F Between 38 1.0354 P <.4249 162 Within The F test failed to reject the hypothesis at the .05 level of significance. Therefore, there is no reason to believe that freshmen, sophomores, and juniors will score differently on the three paper-and-pencil tests used in this study. Hypothesis VI Music students who transferred to Michigan State University from other institutions do not differ from students who began at the University as Fresh­ men on standardized aptitude and achievement tests. There are two levels of the independent variable trans fer status: transfers and non-transfers. A multivariate F test was performed using the nineteen test scores as the de­ pendent variable and transfer status as the independent variable. The results appear in Table 24. 79 TABLE 24.— Multivariate ANOVA for Nineteen Test Variables for Students Grouped by Transfer Status. Source df Multivariate F Between 19 .0502 Within 79 P <.5160 The null hypothesis that there are no differences between transfers and non-transfers was not rejected at the .05 level. Hypothesis VII Music studentsi do not differ from the general population of college students on any of the personality variables measured by the Edwards Personal Preference Schedule. TABLE 25.— Means and Significant Differences for Music Majors and College Norms. EPPS Scale Achievement Deference Order Exhibition Autonomy Affiliation Introspection Succorance Dominance Abasement Nuturance Change Endurance Heterosex Aggression College Norms 14.4 11.8 10. 2 14.3 13.3 16.2 16.7 11.6 15 .8 13. 7 15.2 16.4 12.7 16.0 11.7 MSU Music Significant Majors at .05 N = 212 14.7 10.9 9.5 14.1 14.0 16 .4 17.7 12. 8 12.0 15.6 17.1 16. 3 12.5 15.2 10.8 * * * * * k k k k k 80 The null hypothesis that music majors will not differ from the established norms on the EPPS is rejected. Ten of the fifteen variables showed that significant differences do exist."*■ The remaining section of Chapter IV consists of Tables of means for each of the thirty-one variables when the students are divided according to the various levels of the independent variables. The N's given in each table represent the total number of students in each level of the independent variable, although the N's for each individual mean will be somewhat smaller because of missing data. No means are given when N is less than ten. Certain levels of the independent variables are not included in the tables because the number of students was too small for any meaningful calculations. As an example, “percussion" was omitted in the "major instrument" table be­ cause of low m unbers. As explained in Chapter III, however, the percussionists were put back into the data pool for the other independent variables. 1 X — A The "t" formula used was t = ------S //n where x X = mean of the sample, A = population mean (norms), S = standard deviation of the sample and n = number in the sample. 81 TABLE 26.-“Means of 31 Variables by Major Instrument. . ,, Variable SAT Verbal SAT Quant. Piano Voice Brass N=78 N=60 N=52 537 527 528 530 52 563 Wood. winds N=70 5 527 552 Strings N=2 6 5 44 532 Aural Harmony Harmony U. College 1 Applied 1 Cumulative 1 2.81 3.21 2.99 2.80 3.01 2.45 2.68 2.70 3.04 2.8 0 2.63 2.70 2.73 3.13 2.90 2.46 2.88 2.94 3.3 6 3.05 3.49 3.61 3.50 3.77 3.51 Adv. Aural H. Adv. Harmony U. College 2 Applied 2 Cumulative 2 2.89 2.77 2.58 3.05 3.06 2.23 2.22 2.45 3.26 2.86 3.17 2.46 2.51 3.33 3.04 2.60 2.95 2.83 3.56 3.71 3.55 2.87 3.14 3.92 3.48 Aliferis Mel. Aliferis Har. Aliferis Rhy. Drake Rhythm 17.9 10.8 15.8 70.0 15.2 8.9 14.0 70.1 13.5 8.6 16.9 68.3 15.7 8.7 16.2 68.4 17.3 8.9 16.3 69.0 14.2 10.8 9.5 13.6 13.7 17.0 17.2 13.1 11.6 15.9 3.8.0 16.6 12.8 14.7 10.1 14.3 10.3 8.9 14.6 13.3 15.8 18.9 13.7 11.7 14.9 18.6 17.0 10.3 17.0 11.7 14.9 10.9 8.4 15.0 14.5 15.5 18.5 12.0 13.4 15.2 16.2 14.2 11.9 17.1 11.9 15.0 11.0 9.7 13.6 14.1 16.5 17.6 12.3 11.9 15.8 17.4 16.6 13.0 15.1 10.6 18.5 11.3 11.7 14.8 14.9 15.1 18.2 13.1 10.5 13.8 14.2 16.1 15.1 11.7 10.3 EPPS Achievement Deference Order Exhibition Autonomy Affiliation Introspection Succorance Dominance Abasement Nurturance Change Endurance Heteroscx. Aggression 82 TABLE 27.— Means of 31 Variables by Curriculum Groups. Variable Applied N=72 SAT Verbal SAT Quant. 570 563 Instr. Mus.Ed. N=124 509 542 Choral Mus.Ed. N= 97 506 496 Mus. Therapy N=80 564 559 Mus. Theory N=21 595 584 Mus. No Major N=43 510 512 Aural Harmony Harmony U. College 1 Applied 1 Cumulative 1 2.91 3.05 3.02 3.31 3.13 2.37 2.84 2.80 3.23 2.98 2 .20 2.78 2.68 2.98 2.89 2.57 3.04 3.04 3.06 3.02 3.06 3.50 2.90 2.89 3.06 1.91 2.52 2.62 3.12 2.66 Adv. Aural H. A d v . Harmony U. College 2 Applied 2 Cumulative 2 2.78 2.57 2.59 3.53 3.13 2.48 2.73 2 .70 3.40 3.09 2.45 2.44 2.37 2.03 2.85 2.38 2.78 2.51 3.15 2.92 3.18 3.33 2.65 3.02 3.08 * * * * * Aliferis Mel. Aliferis Har. Aliferis Rhy. Drake Rhythm 16.8 9.6 14.7 72.3 14.7 8.5 16.9 69.4 15.1 9.2 15.0 71.2 17.0 9.7 15.5 66.0 15.8 11.0 15.5 72 .4 14.8 9.9 15.4 68.5 15.5 11.5 10.0 15.0 14 .8 16.0 17 .4 12.2 9.9 16.0 16.8 16.7 13 .6 14 .3 10.2 15.0 11.2 9.0 14.1 14.2 16.0 17.2 12.4 12.6 15.5 16. 7 14 .8 12.5 17.0 .11.3 14.0 11.0 10.2 13.9 14. 0 16.7 16 .7 12.4 11.1 15.4 17.5 18 .4 12.3 15 .6 10.3 13.7 10.3 9.5 13.8 13.7 16.9 .1.Ci. 4 13.4 12.1 15.7 17.6 17.3 12.0 14 .4 1.1 .1 14.9 11.2 9.8 14.6 14 .1 15.8 16.3 10. 2 10.2 14.7 14.3 15.5 10.2 13.5 13.4 15.2 15.9 16.2 12.7 13 .5 11.7 EPPS Achievement Deference Order Exhibition Autonomy Affiliation Introspec tion Succorance Dominance Abasement Nurturance Change Endurance Heteroscx. Aggression *Insufficient number for meaningful analysJs. t o r J.U . J 12.4 13.8 14 .8 18.4 15.1 12.1 13.8 10.3 83 TABLE 28.--Means of 31 Variables by Transfer Status Variable SAT Verbal SAT Quant. Non-Transfers N= 33 6 533 538 Aural Harmony Harmony U. College 1 Applied 1 Cumulative 1 2.44 2.90 2.83 3.14 2.97 Adv. Aural H. A d v . Harmony U. College 2 Applied 2 Cumulative 2 2.53 2.63 2.55 2.78 3 .00 Aliferis Mel. Aliferis Har. Aliferis R h y . Drake Rhythm Transfers N= 92 15 .6 9 .0 15 .8 6 9.6 15.5 9.6 15 .5 69.7 14 .7 10.8 9 .6 14 .0 14 .0 16 .6 17.5 13 .0 12 .1 15.5 17 .1 16.0 12 .7 15.4 10 .7 14 .76 11.2 9.2 14 .4 14 .0 15 .7 18 .1 12.3 11.8 16 .0 16.9 17.2 12.1 14 .7 11.2 EPPS Achievement Deference Order Exhibition Autonomy Affiliation Introspection Succorance Dominance Abasement Nurturance Change Endurance Heterosex. Aggression 84 TABLE 29.--Means of 31 Variables by Class. Variable SAT Verbal SAT Quant. Freshmen N=12 6 516 506 Aural Harmony Harmony U. College 1 Applied 1 Cumulative 1 Sophomore N=97 531 543 2.87 2.88 2.88 3 .12 2.92 Adv. Aural H. Adv. Harmony U. College 2 Applied 2 Cumulative 2 Aliferis M e l . Aliferis Har. Aliferis Rhy. Drake Rhythm Junior N=117 579 604 2.53 3.03 2.94 3.25 3.11 2.68 2.65 2.60 3.34 3 .08 14 .9 8.7 15.6 68.1 16 .6 9.6 16.2 70.0 15 .6 9.6 15 .7 70.4 15.3 10 .9 10.0 14 .7 13 .6 16 .4 16 .9 13 .3 11.8 15.6 16 .6 16 .0 13.0 14 .4 10.7 14 .1 10 .4 9.1 13.7 14 .8 16.5 18.9 12.9 11.5 15 .6 17 .6 16 .9 12.4 15.4 11 .0 14 .8 11.3 8.9 14 .0 14 .2 16,1 18.0 11.8 12 .9 15.3 17 .1 16.2 11.9 16.0 10 .9 EPPS Achievement Deference Order Exhibition Autonomy Affiliation Introspection Succorance Dominance Abasement Nurturance Change Endurance Heterosex. Aggression 85 TABLE 30.— Means of 31 Variables by Sex Variable SAT Verbal SAT Q ua nt. Males N=131 531 564 Females N=209 531 515 Aural Harmony Harmony U. College 1 Applied 1 Cumulative 1 2 .51 2.79 2.88 3.19 2.96 2.63 2.96 2. 91 3 .14 3 .00 Adv. Aural H. Adv. Harmony U. College 2 Applied 2 Cumulative 2 2 .66 2.57 2.46 3.40 3.05 2.58 2.58 2.61 3.25 3.02 Aliferis M e l . Aliferis Har. Aliferis Rhy. Drake Rhythm 15 .2 9.4 16 .3 69.4 15.9 9.1 15 .4 69.4 14 .8 11 .0 8 .R 14 . 3 14 .5 16.1 17 .1 12 .3 13.3 15.4 16.6 14 .8 11. 9 16 .7 11.7 14.8 10.8 9.7 14.1 14 .0 EPPS Achievement Deference Order Exhibition Autonomy Affiliation Introspection Succorance Dominance Abasement Nurturance Change Endurance Heterosex. Aggression 18 .2 12.9 11. 3 15 .5 17 .4 1.7 .3 13.8 14.4 10.4 86 TABLE 3 1 .--Means of 19 Test Variables by Enrollment Status after 18 Months. Variable Aliferis Mel. Aliferis Har. Aliferis Rhy. Drake Rhythm Enrolled Non-Enrolled 16.1 8.9 15.9 6 9.0 12.6 8.3 14.7 6 8.6 15.1 11.1 9.6 14.8 13.2 16.0 17.7 .13.1 12.1 15.1 17.3 16.1 13.0 15.5 10.7 15.6 10.7 10.0 13.8 14.1 15.9 18.9 13.1 12.7 15.7 16.2 15.6 lU.b 16.4 10.3 EPPS Achievement Deference Order Exhibition Autonomy Affiliation Introspection Succorance Dominance Abasement Nurturance Chanqe Endurance Heterosex. Aggression CHAPTER V SUMMARY AND CONCLUSIONS Summary The primary purpose of this study was to discover whether certain variables differentiate among groups of music majors. To that end, thirty-one variables were selected which offered some promise of usefulness based on a review of the literature. These variables included tests of aca­ demic aptitude, music aptitude, music achievement, and person­ ality. In addition, ten GPA's were selected as measures of academic work in the first two years of college. A large number of music majors at Michigan State Uni­ versity were involved in the study, either by their taking various tests, contributing GPA's, or both. The students were then subdivided into sub-groups on the basis of six in­ dependent variables. These independent variables were class, curriculum, s e x , enrollment s tatus, transfer status, and major instrument. All of the information about the students was prepared for data analysis by computer. A number of multivariate analyses of variance statistical tests were performed, and simple correlations, means, and variances were obtained. 87 88 Conclusions The conclusions drawn from this study apply only to the sample from which the data were obtained. That which is true for this sample may not be true for students in future years at Michigan State University or students at other Universities and Colleges. Based on the results of this in­ vestigation, the following conclusions can be admitted. 1. Music students who leave the music department for any reason except graduation do not score differently on the D r a k e , Aliferis, E P P S , or S A T . 2. The only variable which differentiates students according to the music curriculum in which they are en­ rolled is SAT-V. None of the other thirty-one variables results in statistically significant differences. 3. Five variables differentiate between sexes. Males score significantly higher on Deference, Heterosexuality, an(^ SAT-Q, while females score significantly higher on Aliferis Melodic and Endurance. 4. Two variables show significant differences among students grouped by major instrument: Harmony GPA and A d ­ vanced Harmony G P A . 5. There are no differences among students accord­ ing to the class (Freshman, Sophomore, and Junior) on any of the thirty-one variables used in this study. 89 6. Transfer students do not differ from native stu­ dents on any of the nineteen test scores. 7. Michigan State University music students differ significantly from published college norms on ten of the fifteen variables of the EPPS. The significance is due largely to sample size, however, and the actual differences may not be considered meaningful. Discussion This study came about because of a particular need at Michigan State University. The music department was faced with the dilemna of selecting music education majors for the third year of study, and no selection procedure was then in effect. The immediate reaction of the faculty was to try to discover what other music departments had done and whether or not they were successful. A check revealed that no selection process had been devised and validated which would meet the needs of the University. Course grades were immediately considered, but much literature supported the intuitive feeling LliaL much more than grades should be considered. The selection of other variables from among the vast array of tests, demographic variables, and other factors was made, partly based on a re­ view of the literature, and partly based on what could be readily obtained. For instance, SAT scores were used simply because they were already available and wore being used by 90 University admissions. It was hoped that this investigation would either support or help reject the use of these variables. If the findings were negative, then more "shotgun" style studies could investigate still more variables. The end re­ sult would either be the discovery or development of one definitive tool, or the conclusion that no such tool exists. This empirical approach to the problem was almost completely without theoretical framework. this was a handicap can only be guessed. Whether or not Certainly, many discoveries in the behavorial sciences have been made without regard for theories. The developers of the Ali fer is, D r a k e , SAT, and EPPS claimed to be operating under a theoretical framework, thereby lending those theories to this study. Any long term solution to the problem of measuring and predicting success for music majors is going to be diffi­ cult to find due to changing grade standards and the chang­ ing level of entering students. Between 1958 and 196 2, enter­ ing freshmen at the University maintained a constant GPA for their first year (2.21) while their scores on an entering College Qualifying Test rose from 120.3 to 135.9. In other words, a constant grading scale was being applied to essentially different student populations. just the reverse occurred. Between 1967 and 197 3, however, The students scored slightly lower on standardized tests while G P A 1s soared dramatically from 2.4 to 2.9.1 ■'■Arvo E. Juola, "Illustrative Problems in Collegelevel Grading," Personnel and Guidance Journal (1968) 29-33. 91 According to the Evaluation Services department at Michigan State University, SAT scores are becoming less re­ liable as predictors for students on the low end of the ability scale. One reason for this might be that students with low academic ability, and therefore low SAT scores, are often placed in special classes or given special help. The average and above average students simply continue to achieve at their own natural rate and the SAT is a reliable predictor of their academic potential. A major problem encountered was that of missing data. Any testing program of this type which involves using stu­ dents in classes for testing may expect to encounter these problems: First, faculty are reluctant to give up teaching time, especially more than once per term. This is, of course, understandable, especially in large lecture sessions where each class session must be planned long in advance, and where make-up of lost time is impossible. The second problem is the students’ resistance to any test which is a threat, either to their privacy or their self­ esteem. Students are consistently aware of their rights and are quicker to question any apparent invasion of these rights than ever before. The purpose of a tost is likely to be mi s­ understood by some students, and they are extremely apprehen­ sive about being "judged on the basis of a few test scores. Evan as students are convinced of the value of a test and of its non-threatening nature, getting them to attend 92 the testing session remains a third major problem. Even in a class which normally has a high attendance rate, many of the students do not feel obliged to attend if they know a "special" test is scheduled. An example of this occurred in this study, when a test was scheduled for a conducting class while the instructor was known to be away from ca m p u s . Although the instructor had made every effort to stress the importance of attending the testing session, attendance was approximately half of its normal rate. A word of caution is in order concerning the inter­ pretation of statistical significance. As sample size grows large, degress of freedom increase to such an extent that statistical significance becomes easier and easier to obtain. The ultimate extension of this would be when sample size reaches that of the population. at all would be "significant." In that case, In this study, any difference for example, a reduction of the sample size from several hundred to less than fifty would have resulted in fewer significant differ­ ences being found. Even after statistical significance is found in any research study, the question of meaningfulness must be ad­ dressed. For example, regardless of the level of signifi­ cance reached using a test score, if the difference in the two sample means is only two points on a scale of several hundred, the meaningfulness ol the results can be questioned. 93 In the case of this study, significant differences were found in a number of instances which related to several of the hypotheses. Upon closer examination of the means given in the tables at the end of Chapter IV, however, one finds that, in most cases, the differences are only a few score points or a fraction of a grade point. Particularly the EPPS scores were only two or three points apart, meaning that two or three different answers by an individual could have assoc­ iated him with either sample's mean. The fact that this study failed to find any variable which would distinguish "dropouts" from "survivors" or music education majors from other music students does not mean that such variables do not exist. Based on the literature on the subject, however, one can safely conclude that there will be no simple answer to the problem. Perhaps it is time to stop wasting valuable resources on studios involving paper-andpencil personality tests, standardized music aptitude tests, and grade point averages. As was stated in the opening chap­ ter, maybe it will be necessary to loo’: at hundreds of differ­ ent facets ol: the human mind and hew these facets interrelate with each other. Such a solution would certainly involve more sophisticated statistical techniques than are currently in common use by behavioral science researchers. Another possibility exists which was. mentioned in Chapter I. Perhaps "music majors" and ofnor broad categories 94 are simply too broad to describe. Certainly, the definition and prognosis for success for other professions have not been forthcoming. For example, no one has yet described the "typi­ cal physician," much less predicted which nineteen year old pre-med student will be the most successful physician. In Chapter III certain correlations from Table 8 were noted as being interesting, either because of their being high when not expected, or low when some relationship was ex­ pected by the writer. Some possible explanations for a few of these correlations are discussed below. SAT Verbal is a test of word analogies and verbal reasoning ability. The .62 correlation of SAT-V and A d­ vanced Aural Harmony is indeed interesting, since aural harmony is supposedly non-verbal in nature. The strong positive relationship between the two scores might indicate that some students who are weak academically (and therefore would score lower on the SAT-V) have dropped out of music before they completed Advanced Aural Harmony. the correlation of SAT-V with Aural Harmony course) The fact that (a first year is much lower tends to support that contention. Another possible answer to the unexpected high corre­ lation of SAT-V and Advanced Aural Harmony is that the latter requires not so much an aptitude for working with aural stimuli as it requires quick thinking and reasoning under stress (testing situations), abilities partially measured by 95 SAT-V. A third possible answer is simply that "good students" have worked hard in high school and score well on the SAT tests. These students continue to be industrious in college, and perhaps dilligence and hard work are primary prerequisites for a high grade in Aural Harmony. The very high correlation between Cumulative 1 and Cumulative 2 (.92) is expected, since first year grades are included in both. On the other hand, a counselor can safely predict that a student's overall GPA will not change much be­ tween the first and second year of college. The low correlations between applied music and all of the other variables indicate that the ability to make a grade in this area is a relatively independent trait. In a large University applied music teachers have little opportu­ nity to know what the students are doing in other academic areas. Since grading in applied music is partly subjective, it is possible that the teacher's knowledge of a student's work in other classes would alter the teacher's perception of the student's performance ability and therefore the ulti­ mate grade earned. This "expectancy o f feet" might enter into grading in smaller music schools, where applied teachers teach many other music courses. It seems to have little ef­ fect at Michigan State University. One unexpected benefit of this study was an apparent change in the attitude of some of the students toward the 96 music education program. This writer worked in the music academic advising office during the time of this study. As the study progressed it was noticed that more and more stu­ dents were interested in their academic progress, ways to improve their chances of remaining in the music education curriculum, and in alternate programs in the event the stu­ dents were eliminated from music education. Because of the testing program, students became aware that the music department was serious in its attempt to elim­ inate the weaker students from the music education program. The rumor that the College of Education was limiting the number of students admitted to the Junior year of their pro­ gram was being proven true. The advisors in the music aca­ demic advising office, as well as some faculty members, noted that music education was enjoying higher esteem because it was no longer a haven for those denied admission to the other music curricula. Implications H?xIC ^ 1. v on rn o i rt " I ic Th * 1n n n r\ ■f* f Hi d c;+- j j ^ \ y Students should not be selected for the music cation curriculum on the basis of any o n e o r a n y combination of the variables in this study. An examination of the tables of means in Chapter IV re;veals that music education majors differ only slightly front obiter music majors. Also, those who leave the department are not particularly different from those who remain. 97 2. In the counseling of college music students, the variables of this study should not be used to encourage or discourage a student's change in music curricula. Apparently, students in applied music, music theory, music education, music therapy, and even those with no declared music specialty are very much alike. Certain differences are indicated in the tables of Chapter IV, but they are not large enough for most counseling purposes. 3. Statements often made concerning the "personality of music majors" are not entirely justified. It is true that statistical significance was found between college norms and Michigan State University music majors, but the differ­ ences were not large enough to be meaningful. The EPPS uses a scale from one to twenty-eight, and the largest difference between means was less than two points. Suggestions for Further Research Three suggestions for further research seem the most promising to this researcher: 1. Other types of variables need to be measured. Traditional music and academic aptitude tests have not proven effective; therefore, such attributes as creativity, speech ability and self-concept might be researched. Although stand­ ardized tests are already in existence in these a r e a s , spe­ cial tests for music prognosis purposes should be considered. 98 2. A follow-up of this study might include the study of attrition at later points in a student's career. A longi­ tudinal study of five or more years is needed, using this data or other data. For instance, after five years it would be interesting to determine the characteristics of those who remained in music and those who left the profession. The raw data used in this study is on file in the music depart­ ment. 3. Devices other than paper-and-pencil tests need to be considered. One possibility would be a structured interview in which certain information is obtained which can be quantified. Another method for gathering data, about music education students in particular, would be in conjunc­ tion with a course early in the college experience. The course might involve visiting public schools, experiencing some small amount of student teaching, or some other activi­ ties closely related to actual music teaching. BIBLIOGRAPHY 99 BIBLIOGRAPHY Books Astin, Alexander W. Predicting Academic Performance in College. New York: The Free Press, 1971. Buros, Oscar. The Seventh Mental Measurements Yearbook. Highland Park, New Jersey: The Gryphon Press, 1972. Crawford, Albert B . , and Burnham, Paul S. Forecasting College Achievement. New Haven: Yale University Press, 1948. Drake, Raleigh. Drake Musical Aptitude Tests M a n u a l . Chicago: Science Research Associates, 1957. Drayer, Adam M. The Teacher in a Democratic Society. Columbus, Ohio: Charles E. Merrill Publishing Company, 1970. Edwards, Allen L. Edwards Personal Preference Schedule Manual. New York: Psychological Corp., 1959. Folgen, John K., and Astin, Helen S., and Bayer, Alan E. Human Resources and Higher Education. New York: Russell Sage Foundation, 1970. " ' Gekoski, Norman. Psychological Testing. Springfield, Illinois: Charles C. Thomas, 1964. GeLzels, J. W . , and Jackson, P. W. "The Te acher’s Person­ ality and Characteristics," Handbook of Research on Teaching. Edited by N. L~ G a g e . 'Chicago: Rand McNally and Co., 1965. Guilford, J. P. The Nature of Human Intelligence. York: McGraw-Hill Book Company, 1967. Guilford, J. P. Psychometric Methods. Hill, 1954. New York: New McGraw- Harris, Chester W . , ed. Encyclopedia of Educational Research. New York: The MacMillan Co., 1960. 100 101 Holstrum, Lars-Gunnar. Musicality and P r o g n o s i s . Svenska Bok for Laget, 1963. Lehman, Paul. Tests and Measurements in M u s i c . Cliffs: Prentice Hall, Inc., 1968. Sweden: Englewood Levin, David. The Prediction of Academic Performance. New York"! Ru'ssell Sage Foundation, 19 65. Siegel, Sidney. Nonparametric Statistics for the Behavioral Sciences. New York: McGraw-Hill Book Co., Inc., 1956. Stanley, Julian C . , and Hopkins, Kenneth D. Educational and Psychological Measurement and E v a l u a t i o n . Englewood Cliffs, New Jersey: Prentice-Hall, Inc., 1972. Stern, George G. "Measuring Noncognitive Variables in R e ­ search on Teaching," Handbook of Research on Tea c h i n g . Edited by N. L. Gage. Chicago: Rand McNally and Co., 1965. Stinnett, T. M . , ed. The Teacher D r o p o u t . Bloomington, Indiana: Phi Delta Kappa Inc., 1970. Super, Donald E. Appraising Vocational F i t n e s s . Harper & Brothers., 1949. New York: Tyler, Fred T. "The Prediction of Student Teaching Success from Personality Inventories," University of California Publications in Edu cation. XI Berkeley, California: University of California, 1954. Whybrew, William. Measurement and Evaluation in M u s i c . Dubuque, Iowa: William C. Brown C o . , 19 62. Periodicals Bach, Jacob O. "Practice Teaching Success in Relation to Other Measures of Teaching Ability," Journal of Experimental E duc ation, XXI (1952), 57-80. Barr, A. S. "Problems Associated with the Measurement and Prediction of Teacher Success," Journal of Ed u c a ­ tional Res ea r c h , LI (1958), 695-699. Barr, A. S., and Eustice, David., and Noe, Edward. "The Measurement and Prediction of Teacher Efficiency," Review of Educational Research, XXV (1955), 261-269. 102 Beinstock, Sylvia. "A Predictive Study of Musical Achieve­ ment," Journal of Genetic Psychology, LXI (1942), 135-45. : Charles, Harvey. "The Use of a Selected Projective Technique in the Teacher Selection Process," Studies in Education, Abstracts of Theses, 1 9 5 2 . Bloomington: Indiana University, 1953. Cornfield, J. and Tukey, J. W. "Average Values of Mean Squares in Factorials," Annals of Mathematical Statistics, XXVII, 907-949. Ernest, David J. "The Prediction of Academic Success of College Music Majors," Journal of Research in Music Education, (Fall, 1970), 273-276. Gordon, Edwin. "A Study to Determine the Efficacy of General Intelligence and Musical Aptitude Tests in Predict­ ing Achievement in Music," Council for Research in Music Education, XII (1968), 40-45. Gross, B . , and Seashore, R. H. "Psychological Characteristics of Student and Professional Musical Composers," Journal of Applied Psychology, XXV (1941), 159-170. Jackson, P. W . , and Guba, E. G. "The Need Structure of InService Teachers: An Occupational Analysis," School Review, LXV (1957), 176-192. Juola, Arvo E. "Illustrative Problems in College-Level Grading." Personnel and Guidance Journal. (1968) 29-33. More, Grace van Dyke. "Prognostic Testing in Music on the College Level: An Investigation Carried On at the North Carolina College for Women," Journal of E du­ cational Research, XXVI (1932). Ringness, Thomas A. "Relationships Between Certain Attitudes Toward Teaching and Success," Journal of Experimental Education, XXI (1952), 1-55. Roby,A. Richard. "A Study of the Correlation of Music Theory Grades with the Seashore Measures of Musical Talents and the Aliferis Music Aptitude Test," Journal of Research in Music Education, X (1962). Stanton, Hazel M. "Psychological Tests— A Factor in Admission to the Eastman School of Music," School and Society, XXIX (1929), 889-891. 103 Stecklein, John E., and Aliferis, James. "The Relationship of Instrument to Music Achievement Test Scores," Journal of Research in Music Education, V (1957), 3-15. Taylor, C. H. "Characteristics of First Year Conservatory Students." Journal of Research in Music Education, I (1953), 105-118. Taylor, E. M. "A Study in the Prognosis of Music Talent." Journal of Experimental Education, X (1941), 1-28. Tilson, Lowell Mason. "Music Talent Tests for TeachingTraining Purposes," Music Supervisors Journal, XVIII (1932), 26-27, 61. Vernon, P. E. "The Psychology of the Composer," Letters, XI (1930), 38-40. Music and White, Adolph. "The Aliferis Music Achievement Test as a Predictor of Success in Music Theory," Journal of Educational Research, LIV (1961), 315-317. Wilson, M. Emmett. "The Prognostic Value for Music Success of Several Types of Tests," Music Supervisors Journal (1932), 83-98. Unpublished Anderson, John Martin. "The Use of Musical Talent, Person­ ality and Vocational Interest Factors in Predicting Success for Student Music Teachers." Unpublished Doctoral Dissertation, University of Southern California, 1965. Barth, George. "Some Personality and Temperament Character­ istics of Selected Music Teachers." Unpublished Doctoral Dissertation, University of Southern Calif­ ornia, 1961. Beaver, Maxie. "Personal Characteristics of Successful Band Directors in North Carolina." Unpublished Doctoral Dissertation, University of North Carolina, 1973. Burton, Roger. "The Personality of the Contracted Studio Musician." Unpublished Masters Thesis, University of Southern California, 1956. 104 Cooley, John Christopher. "A Study of the Relation Between Certain Mental and Personality Traits and Ratings of Musical Abilities." Unpublished Doctoral Disser­ tation, Michigan State University, 1952. Davis, Samuel E. "Predicting Probable Failure in CollegeLevel Music Theory Courses." Unpublished Doctoral Dissertation, University of Montana, 1968. Duda, Walter. "The Prediction of Three Major Dimensions of Teacher Behavior for Student Teachers in Music Educa­ tion." Unpublished Doctoral Dissertation, University of Illinois, 1961. Ehlert, Jackson K. "Selection and Education of Public School Music Teachers." Unpublished Doctoral Dissertation, University of Colorado, 1949. Ferrell, John W. "A Validity Investigation of the Drake Musical Aptitude Tests." Unpublished Doctoral Dissertation, State University of Iowa, 1961. Fosse, John. "The Prediction of Teacher Effectiveness: An Investigation of Relationships Among High School Band Contest Ratings, Teacher Characteristics, and School Environmental Factors." Unpublished Doctoral Dissertation, University of Illinois, 1965. Gallagher, Fulton D. "A Study of the Relationships Between the Gordon Musical Aptitude Profile, the Colwell Music Achievement Tests, and the Indiana-Oregon Music Discrimination Test." Unpublished Doctoral Dissertation, Indiana University, 1971. George, Warren. "Significant Predictors for College Achieve­ ment in Specified Areas of Music Education and Identi­ fication of Potential Graduates." Unpublished Doctoral Dissertation, University of Kansas, 1969. Gordon, Edwin. "A Study to Determine the Effects of Training and Practice on Drake Musical Aptitude Test Scores." Unpublished Doctoral Dissertation, University of Iowa, 1958. Lutz, Warren. "Personality Characteristics and Experience Backgrounds of Successful High School Instrumental Music Teachers." Unpublished Doctoral Dissertation, University of Illinois, 1963. 105 Megginson, Jean Haydock. "The Relationship of the Edwards Personal Preference Schedule Variables to the Academic Achievement of Selected University of Mississippi Female Students within Five Ability Levels." Unpublished Doctoral Dissertation, The University of Mississippi, 1971. Peterson, Floyd J. "A Study of the Relationship Between Music Aptitude and Academic Achievement of Graduate Music Students." Unpublished Doctoral Dissertation, Indiana University, 1963. Rohrs, D. K. "Predicting Academic Success in Liberal Arts College Music Education Programs." Unpublished Doctoral Dissertation, University of Iowa, 196 2. Schleuter, Stanley. "An Investigation of the Interrelation of Personality Traits, Musical Aptitude and Musical Achievement." Unpublished Doctoral Dissertation, University of Iowa, 1970. Schnitzer, Joseph. "Assessment of Certain Personality Potentials in Relation to Student Teaching." Un­ published Doctoral Dissertation, University of Minnesota, 1962. Stone, Michael Horace. "A Study of the Relationships Between Selected Variables and the Differential Academic Achievement of Freshmen in the University of Michigan School of Music." Unpublished Doctoral Dissertation, University of Michigan, 1969. Turrentine, Edgar. "Predicting Success in Practice Teaching in Music." Unpublished Doctoral Dissertation, University of Iowa, 1962. Wink, Richard Lee. "The Relationship of Self-Concept and Selected Personality Variables to Achievement in Music Student Teaching." Unpublished Doctoral Dissertation, Ohio State University, 1967. Woerner, Gerald H. "An Analysis of Certain Criteria of Musicianship as Recommended by the Teacher Training Curriculum Committee of M. E. N. C." Unpublished Doctoral Dissertation, Pennsylvania State University, 1949. 106 Miscellaneous Aliferis, James. Aliferis Music Achievement T e s t . Minneapo­ lis: The University of Minnesota Press, 1954, with ma nua l. Drake, Raleigh M. Drake Musical Aptitude Tests. 2nd e d . Chicago: Science Research Associates, I n c . , T957. _________. Manual for the Drake Musical Aptitude T e s t s . Chicago: Science Research Associates, 1957. Edwards, A. L. Edwards Personal Preference Schedule. New York: Psychological Izorp., 1954, with manual. Finn, Jeremy. "Univariate and Multivariate Analysis of Variance and Covariance: A Fortran IV Program." Revised by Wright, David J. Occasional Paper No. 9, Office of Research Consultation, College of Education, Michigan State University, 1970. Lunneborg, Patricia W. "EPPS Patterns and Academic Achieve­ ment in Counseling Clients." Seattle: Washington University Bureau of Testing, April, 1969. Scholastic Aptitude T e s t . Princeton, New Jersey: Entrance Examination Board, 1961. College