Research Applications in Music CAI
This article was co-authored by Professor Killam and her colleagues at North Texas State University: Philip Baczewski, Antoinette Corbet, Paul Edward Dworak, Jana Kubitza, Michael Morgan, and Lawrence Woodruff.
Computer-assisted instruction (CAI) in music involves a variety of research applications. Development of a viable system requires extensive research in theory pedagogy, computer hardware and software, and listening skills. Any CAI system must undergo extensive research and testing to reach maximum effectiveness for student users.1 Once the system is in use, continuing research is necessary to measure its educational value to students and to determine areas where it needs to be improved or expanded. A CAI system provides the means through which much additional research, such as investigations of auditory pattern perception and development of new pedagogical techniques, can be accomplished. In short, ongoing research is essential in many content areas and from multiple viewpoints to produce effective CAI. Before implementing a CAI system the limits of content to be presented as well as the appropriate pedagogical stance must be determined, and a detailed outline of curriculum must be planned. A complete curriculum outline requires considerable research, but is essential if student answer formats are to be clear, discrete, and noncontradictory. These factors will also play a crucial role in the necessary research into appropriate hardware and software. A profile of the potential student population should be prepared, so that formats for student interaction may be planned. To illustrate, complex responses on an alphanumeric keyboard should not be planned if a large proportion of the student body does not know how to type!
RESEARCH IN MODELING THE STUDENT
Modeling an image of the student user is probably the most important preliminary research activity.2 Design of curriculum, choice of hardware and software, and study of student data are all dependent on the establishment of accurate delineation of students for whom the CAI system is intended. Modeling the student image continues throughout the development and use of any CAI system. The effectiveness of any instructional strategy is dependent on development of a clear and accurate model of the learner. For music CAI this model should include physiological as well as psychological attributes. Unfortunately, a number of music educators have tended to rely on intuitive notions concerning the nature of these attributes; however, some areas of student image have been researched and can form the basis for good quality music CAI.
Music CAI facilities provide laboratories for the investigation of the music student's learning processes. The use of CAI for research activities includes the following advantages:
- Computer use requires a clear specification of the question (programing allows little tolerance for thinking processes which are poorly defined).
- A stimulus is precisely and infinitely repeatable, insuring the protection of results from an experimenter's bias or fatigue.
- Results may be efficiently collected, retrieved, and tabulated.
A functional model of the music student should include specifics of hearing, learning, and environment. Research by audiologists has established the structure and function of the auditory system and neural network to and from the brain. The music educator needs to know the effects of values of frequency, duration, amplitude, and timbre upon accuracy in perception and identification. Will certain values for these elements aid the learner's development? Where are errors in perception likely to occur? Are any of these errors physiologically based? Music researchers need to explore all of these questions.
Psychological questions in need of exploration are numerous; a few important ones are listed here. The first questions relate closely to those above, concerning theories of perception. The highly specific nature of musical notation (a shorthand for discrete pitch, duration, and temporal information) provides a tool for further investigation. Collaborations between musicians and psychologists can help verify and enlarge existing theories.
Concepts of memory, attention, and motivation are other relevant areas of pursuit.3 Musicians distinguish between hearing and listening. The latter implies a focus of attention and the exercise of cognitive powers. Can listening skills be taught and improved on? The intuitive answer is yes, yet the psychological mechanism of attention is only crudely understood. Clearly, further evolution of music CAI requires study in this area.
Additional areas of research have been suggested regarding peripheral factors such as time of day or barometric pressure. If a morning class is used as the experimental group, and an afternoon class as the control group, will the results be valid and replicable at any time of day? These questions cannot be ignored.
Environmental factors constitute a final area to be considered in modeling the learner. These factors include the sum of all the sociological, educational, and cultural situations which characterize a given student population. CAI strategies for a university will differ from those for a community college. Each individual student population may be described by an equation which includes such factors as socio-economic status, previous education, level of motivation, and attitude.4
Currently, research is being done to model the student's creative processes. Paul E. Dworak is developing, under a grant from the National Institute of Education, a microcomputer-based CAI system employing a dynamic memory model of creativity.5 This system will aid the student in composing melodies by reflecting decisions of the student in the selection of his melodic materials. This example of direct computer application to modeling the student image is a concept expandable to all levels of research.
RESEARCH IN CURRICULUM DESIGN
In preparation for aural courseware development a review of current literature concerning auditory perception, learning psychology and music theory is important. Consideration of fundamental pedagogical philosophy is inherent in curriculum development and is easy to overlook. However, one of the preliminary goals of curriculum design should be an examination of alternative teaching strategies. Although a number of valid theoretical approaches may present themselves, the adoption of a central point of view will be helpful for the sake of consistency within a set of lessons. Questions such as the following should be considered:
|1.||Is the overriding philosophy that of Schenker? Hindemith? Comprehensive Musicianship? "Traditional" theorists?|
|2.||Is the approach melodic or harmonic?|
|3.||Will scale step numbers or solfège be used?|
|a.||Moveable or fixed do?|
|4.||What is the role of the keyboard in theory instruction?|
|5.||Should the sound source vary in timbre for dictation?|
|6.||In melodic dictation, are pitch and rhythm equally important?|
|a.||Which should the student write first?|
|7.||In harmonic dictation:|
|a.||Are all four voices equally important or are outer voices more important?|
|b.||How important is Roman numeral identification, as opposed to harmonic function (tonic-dominant-subdominant)?|
|8.||Is the emphasis on speed or accuracy?|
|9.||What are acceptable competency levels?|
Ultimately, the form and substance of each lesson will reflect the nature of the developer's answers to these and other questions. After these choices are made, the content of each semester should be outlined in detail.
During the early planning stages the role of CAI within the learning environment must be considered. Do not assume that the computer is a replacement for classroom instruction, or that it is merely another dispensable option. CAI encompasses a variety of teaching and learning strategies, including competitive manipulation (gaming), simulation of the learning environment, presentation of new material in a tutorial format, and drill and practice of concepts already presented in the classroom.
When the role of the computer has been defined, lessons may be tailored to meet the needs of the music program for which CAI is being designed. Much of the lesson format is determined by functional considerations. A lesson which substitutes for classroom experience (e.g., to "make up" a class absence) requires a more lengthy tutorial than one which serves to supplement or reinforce. Be sure that answer formats reinforce valid relationships. Consider carefully how many elements are to be tested at a time and how students are to enter them into the machine. Text should be consistent in style and should include a clear explanation of answer format, lesson purpose, sample questions and answers, and appropriate response cues. Avoid succumbing to the temptation to be cute or clever; this approach rarely holds up under long-term use.
The effectiveness of curriculum is dependent upon careful study before lessons are designed and after they are in use. CAI is an ongoing process in which planning, design, testing, use, reevaluation and rewriting must take place. Research is the tool which illuminates this process and is the means by which the most efficient learning occurs.6
RESEARCH IN HARDWARE AND SOFTWARE
If the designer of a CAI system has a clear idea of what the system ought to do, he or she may theoretically implement it on almost any commercially-available computer system. Realistically, not any computer can be used for CAI. Very large machines may be overcommitted in large university settings, and may not be able to allocate enough processor time to drive the sound-generating hardware which is needed for ear training. On the other hand very small computers may have insufficient memory for music CAI.
In the selection of hardware for CAI much research must be devoted to how or why the particular configuration works. Most hardware configurations are not designed by musicians and fit only a small set of specifications. Moreover, very few components can be added to an incompatible machine by a musician without a background in computer hardware. The greatest disaster that can befall a CAI developer is that which comes from searching for any available components, without a thorough understanding of how they will connect with one another.
Research is necessary into all available sound-generating hardware and the computers with which it will operate. Three sources are particularly helpful for preliminary research:
- technical journals
- computer dealers who can specify what hardware is compatible with their machines; (and be careful! hardware supplied without complete software support is totally useless!)
- sales seminars and minicourses.
Of course the CAI developer must be concerned about the quality of sound generated by the hardware. One must remember that the sound generator is a musical instrument, to which serious investigation of quality must be given. Details of sound synthesis systems are reported in the literature.7
RESEARCH IN DATA SAVING AND ANALYSIS
The nature of CAI allows for extensive accumulation of data. Statistics can be saved which will generate research in areas such as auditory perception, memory, learning psychology, and the creative process. Information saved will make an essential contribution to the reevaluation of a system's effectiveness. Changes indicated by these data may involve lessons to be inserted, question or answer format, speed and/or timbre of musical examples, number of repetitions allowed, effectiveness of individual lessons or groups of lessons, revision of instruction, and so forth. At the inception of a CAI system an orderly approach to gathering and evaluating statistics should be devised.8
One of the most important tasks in music CAI research applications is the gathering of data. The results of a statistical analysis of these data can show the strengths and weaknesses of current curricula and their presentation. These results determine how the curricula will be modified for future use. The statistics on student use of the NTSU ear training CAI laboratory indicate peak laboratory use and help in planning laboratory and monitor schedules.
At North Texas State University the ear training CAI system is implemented on a Hewlett-Packard 2000 timeshared computer.9 Approximately 400 music students enrolled on the system in the spring semester of 1980. They registered during the first week of classes by entering their names, student identification number, ear training class and section, sex, and primary instrument on the CRT terminals. This information was saved in an index file for future reference and data processing.
The system software updates a statistics file with each individual lesson use, storing such information as student identification number, lesson name, date, beginning and ending time of lesson use, total number of questions asked, number of correct answers on first trial, number of correct answers by final trial, and number of replays of musical examples requested by students. Another file, containing individual student responses, is kept. As can well be imagined, much space is consumed storing these files. Even more computer space is required to conduct statistical analyses of the data. Due to the space requirements these data files are moved annually from the Hewlett-Packard to NTSU's National Advanced System 5000 computer, which has a four megabyte capacity. Through software packages such as SPSS (Statistical Package for the Social Sciences) and SAS (Statistical Analysis System) statistical research is carried out.
Statistics indicate that out of nearly 5,000 lesson uses during the spring semester of 1980, midmorning and midafternoon hours were preferred, with 33% of all lessons taken during the 10:00-11:00 a.m. and 3:00-4:00 p.m. hours. These data assist in laboratory management and planning.
Statistical analysis also indicates that, except for the anticipated increase of student usage prior to final exams, a fairly constant access of the lessons occurred throughout the semester, with mean use of 337 lessons per week. Research investigating student competence vs. time of day, sex, primary instrument, and other significant factors is in process; results should prove to be enlightening for further development of CAI in music.
SUMMARY AND CONCLUSIONS
To conclude, research is an integral part of effective music CAI. Extensive research must be conducted prior to and during system development. The research areas of modeling the student's musical processes, instructional design, system configuration, and data saving and analysis are all basic to CAI. Once a system is operational continued research is needed to evaluate the system's effectiveness and to plan future modification and expansion. Finally, music CAI supports a variety of research applications to inquiries which have long engaged musicians.
1. Robert W. Ottman, Rosemary N. Killam, Robert M. Adams, W. Kenton Bales, Steven W. Bertsche, Leslie C. Gay, Donald B. Marshall, Daniel A. Peak, and Douglas Ray, "How to Develop a Concept-centered Ear Training CAI System," National Consortium of Computer-Based Music Yearbook, Journal of Computer-Based Instruction, 1980.
2. Paul Lorton, Jr. and Rosemary N. Killam, "Modeling the Student and the Discipline in CAI Drill and Practice," Proceedings, SIGCUE-SIGSIE Conference, Anaheim, California, 1976.
3. Paul Edward Dworak, "Employing a Dynamic Model of Creativity in Written Music Theory CAI," National Society for Performance and Instruction, Eighteenth Annual Conference, Dallas, Texas, April 2, 1980.
4. Paul Edward Dworak, "Employing Schenkerian Structural Ideals and a Dynamic Model of Creativity in Music CAI," Symposium 80, International Conference on the Teaching of the Arts at Advanced Levels, Montreal, Quebec, August 18, 1980.
5. Philip Baczewski, "Experience and Evaluation: Ear training CAI in Action," Proceedings, Texas Music Educators Association Conference, San Antonio, Texas, 1980.
6. Rosemary N. Killam and W. Kenton Bales, "Music CAI: A Model for Epistemological Research Applications," Proceedings, Third Canadian Symposium on Instructional Technology, Vancouver, B.C., 1980.
7. Alice C. Parker, Richard D. Blum, and Paul E. Dworak, "The Carnegie Mellon Computer Music System Digital Hardware," 1977 International Computer Music Conference, October 27, 1977.
8. Alice C. Parker and Paul E. Dworak, "Envelope Control with an Optical Keyboard," 1977 International Computer Music Conference, October 29, 1977.
9. Richard L. Hamilton and Rosemary N. Killam, "An Integrated Data Management System for Computer-assisted Instruction," Proceedings, Association for the Development of Computer-Based Instructional Systems, Washington, D.C., 1980.
10. Rosemary N. Killam, W. Kenton Bales, Richard L. Hamilton and Dan W. Scott, "AMUS: The Computer in Music Instruction," Proceedings, Texas Music Educators Association Conference, Fort Worth, Texas, 1979.
1Robert W. Ottman, Rosemary N. Killam, Robert M. Adams, W. Kenton Bales, Steven V. Bertsche, Leslie C. Gay, Donald B. Marshall, Daniel A. Peak, and Douglas Ray. "How to Develop a Concept-centered Ear Training CAI System," National Consortium of Computer-Based Music Yearbook, Journal of Computer-Based Instruction, Vol. 6 (1980).
2Paul Lorton, Jr. and Rosemary N. Killam, "Modeling the Student and the Discipline in CAI Drill and Practice," Proceedings of the SIGCUE-SIGSIE Conference (1976).
3Paul Edward Dworak, "Employing a Dynamic Model of Creativity in Written Music Theory CAI," paper presented at the National Society for Performance and Instruction, Dallas, Texas, April 2, 1980.
4Paul Edward Dworak, "Employing Schenkerian Structural Ideals and a Dynamic Model of Creativity in Music CAI," paper presented at the International Conference on the Teaching of the Arts at Advanced Levels, Montreal, Quebec, August 18, 1980.
5Philip Baczewski, "Experience and Evaluation: Ear training CAI in Action," Proceedings of the Texas Music Educators Association Conference (1980).
6Rosemary N. Killam and W. Kenton Bales, "Music CAI: A Model for Epistemological Research Applications," Proceedings of the Third Canadian Symposium on Instructional Technology (1980).
7Alice C. Parker, Richard D. Blum, and Paul E. Dworak, "The Carnegie Mellon Computer Music System Digital Hardware," paper presented at the 1977 International Computer Music Conference, October 29, 1977; and Alice C. Parker and Paul E. Dworak, "Envelope Control with an Optical Keyboard," paper presented at the 1977 International Computer Music Conference, October 29, 1977.
8Richard L. Hamilton and Rosemary N. Killam, "An Integrated Data Management System for Computer-Assisted Instruction," Proceedings of the Association for the Development of Computer-Based Instructional Systems (1980).
9Rosemary N. Killam, W. Kenton Bales, Richard L. Hamilton, and Dan W. Scott, "AMUS: The Computer in Music Instruction," Proceedings of the Texas Music Educators Association Conference (1979).