George J. Hess, Jr., Alabama State University

In the past twenty-five years, numerous studies have concluded that CAI provides better overall results than traditional instruction. (Alessi & Trollip, 1991; Kulik & Kulik, 1991). Most studies concluded that computer-assisted drill and practice was more efficient for both teacher and student, was less judgmental and provided greater motivation. In music, similar results have been reported in studies on the development of aural skills. Students using CAI were found to attain higher initial aural skills scores with greater long-term retention, while at the same time developing a better attitude towards ear training and computer-assisted instruction (c.f. Hofstetter, 1975; Taylor, 1982). These benefits were primarily attributed to the computer's ability to provide individualized instruction and immediate evaluation and feedback ( Hofstetter, 1981).

It is therefore not surprising that the use of CAI for ear-training has increased over the years. By 1990, 71% of college music theory programs had some computer-assisted instruction available to students (Pembrook & Riggins, 1990). The inclusion of a music technology requirement for NASM accreditation ensures that that number is increasing. Still, we have not yet seen evidence that the problems posed by ear training instruction have been solved. One would be hard pressed to find many ear training instructors who are satisfied with the progress their students have made; and as for the students, ear training can best be described as a "dreaded, necessary evil" (Covington, 1992, p. 5). The question, then, is why has the great promise of computer-based ear training instruction not been realized.

An examination of previous studies reveals a number of commonalities. First, the courseware used in each study was, by necessity, developed specifically for its respective study and therefore was closely correlated to the curriculum. Second, in all of studies the use of the software by the students was required. Third, virtually all of the programs were drills, with most using a variation of the mastery learning model, with or without adaptive control or remediation. Finally, all of the software maintained extensive records regarding student use and progress.

It is proposed that computer-assisted instruction remains the area of greatest potential for improving aural discrimination skills. A well-designed program, that is thoroughly integrated with the curriculum, both pedagogically and strategically, can supplement and in fact, replace much classroom instruction.

The intent of the study was to determine whether computer assisted-instruction could be used to replace both the in-class instruction and drill in the areas of intervals and harmony. Additionally, it was expected that information would be gained regarding strategies for instructional design and methods of incorporating CAI with the curriculum. Specifically, questions to be answered included: Can CAI replace in-class drill and practice of dictation? Do students use CAI software when not required to do so? Does requiring the use of software cause an increase in frustration on the students part? Are instructional models other than drill and practice valuable? How do students react when given control over the instructional sequence?

The study was conducted during Fall 1993 in the School of Music of the University of Northern Colorado using two sections of the freshman music theory class. The class met four times a week for fifty minutes with both ear training and written skills covered in the same course. Both classes received similar instruction regarding written skills, melodic dictation and sightsinging. One section, functioning as a control group received instruction and in-class drills in interval and harmonic dictation and was encouraged to supplement that by using commercial ear training software on a voluntary basis. The experimental section received no in-class instruction in these areas. All instruction and drill related to intervals and harmony were provided by a computer program developed specifically for this course. The group was required to use the software for a minimum of two hours per week, with a weekly, time-based grade to be averaged with the homework grade for the course.

The courseware used by the experimental group was developed by the researcher to complement the freshman music theory program at the University of Northern Colorado. The instructional design is a combination tutorial and drill and is based on a cognitive apprenticeship model. Students are presented concept and strategies developed through conversations with students and teachers who have experienced success in ear training. They are then guided through the process of solving the problems in ungraded practice sessions. The student may then move to the graded drills and receive feedback and suggestions for further study and practice. The locus of control remains primarily with the student with occasional intervention by the instructor. The program maintains extensive data regarding student effort, progress and use patterns.

A quasi-experimental, pretest-posttest control group design using intact groups was chosen. In addition to the tests evaluating aural skills, a Likert-type attitude instrument was also administered both before and at the conclusion of the study. Formal and informal observation was also conducted, based on data stored by the courseware and conversations with students.

To determine whether either method produced better results, the mean scores for both groups were calculated for the pretest, posttest and final grade. The experimental group had slightly higher mean scores for both tests. It was also be noted that the standard deviation decreased from pretest (SD = 16.85) to posttest (SD = 13.87) for the experimental group while remaining virtually constant for the control group (SD = 11.88). The scores of both groups were compared using an ANCOVA with the pretest as the covariate and the posttest as dependent variable. The results indicated the results achieved by the two groups were statistically equivalent (MS = 0.026), F = 0.00, p < 0.99.

The attitude instrument asked students to rate their attitude towards the computer, ear training and CAI and the amount of time and effort they spent on ear training. No significant differences were found in the attitude of the students. Regarding effort, the experimental group (M = 3.56) felt they had spent considerably more overall effort on ear training than did the control group (M = 2.73), t(17), p < .004. Further, while both groups indicated in the pre-treatment survey that they expected to spend a similar amount of time practicing ear training outside of class, the difference in the perceived amount spent was found to be significant at the .05 level t(16), p < .022. One final question was administered to both groups on the post-treatment survey only: How much of your improvement do you attribute to using the computer? A t test comparing the two groups found that the experimental group (M = 3.824) attributed significantly more of their improvement to the use of the computer than did the control group (M = 2.77), t(16) = 2.960, p < 0.01.

In evaluating the software, students rated the program above average (M = 4.06). The drills were, by far, the preferred mode of instruction, (88.2%) with interval (47%) and soprano, type and bass (23.5%) rated the most valuable subjects.

The audit trail showed that the students-use patterns generally conformed to their perceptions of the program. On average students spent 85% of their time in the program using the drills section. However, the most popular subject was chord quality identification (35%) followed by intervals (30%) and interval rows (19%).

The results of the study indicate that CAI can be used to replace in-class dictation instruction and drill. Students using the custom software did slightly, but not significantly, better than students using traditional instruction. These results are consistent with results reported by previous researchers (Alessi, 1991; Kulik, 1991). The decrease in the standard deviation from pretest to posttest for the experimental group indicates the computer program provided more consistent instruction. It was also noted that both groups achieved considerably higher ear training scores than expected. Further, nearly half of the experimental group were remedial students, while the control group included only one.

The issue which stands out is the amount of time and effort put forth by the experimental group. Hofstetter (1978) found a correlation existed between the number of times a question was asked and the percentage of correct answers. The experimental group spent significantly more time on ear training than did the control group. At the same time, no increase in frustration was found. It can therefore be assumed that the motivation provided was effective.

Motivation was achieved extrinsically through the course requirement. Lepper (cited in Alessi, 1991) has suggested that extrinsic motivation such as a reward system detracts from learning. However, most of the studies reporting successful progress by students did use extrinsic motivation by requiring the use of the programs (Hofstetter, 1975; Taylor, 1980). Intrinsic motivation was provided both by the student and the computer software. Four factors--maintenance of attention, relevance of the material, student confidence and student satisfaction--are required to provide motivation (Keller ,cited in Alessi, 1991). The students' evaluation of the software and its effectiveness indicate that these criteria had been met. The lack of increased frustration, supported by the student-use patterns observed in the audit trail, appear to confirm the effectiveness of providing learner control with advisement.

On the other hand, the need for relevance and efficiency led the students to, generally, ignore the tutorial sections of the program. The program was designed with three sections, two of which--Lessons and Practice Sessions--embodied constructivist principles, while the remaining section, Drills, was clearly objectivist-based. By a large margin, students preferred the objectivist method. For the goal-oriented student, drills were clearly perceived as the more efficient method of practice. Feedback in the drills was more concrete, being more subjective in the other areas. This does not suggest that the lessons and practice sessions would not have benefited many of the students. But with the use pattern exhibited, no conclusions may be drawn as to the effectiveness of this type of instruction when it is used.

Based on previous research and the results of the present study, some issues must be addressed in the effective use of computer-assisted instruction for ear training. As with all educational software, effective computer-assisted music instruction must be based on sound principles of instructional design.

Interface. The overall interface, which includes screen design, sound, methods of interaction, as well as the general operation of the program, must be consistent, attractive and easy to understand. The quality of sound in ear training programs is critical. All programs must now include at least the option of MIDI control (Pople, 1992). The availability of low-cost 16-bit digital audio for both the PC and Macintosh offers another alternative. The program should maintain accurate records of student use and progress. Feedback should be immediate and positive.

Pedagogy. There are two distinct facets to the application of pedagogy to computer-based training: curriculum and design. The issues of curriculum are measured by how well the computer program adheres to the instructional philosophy of the curriculum it supports (Gross, 1983). The sequence and content of instruction must conform to that of the curriculum it supports.

The pedagogy of design refers to the sequence and manner in which the material is presented. Traditionally, computer-assisted instruction, particularly for music, has incorporated an objectivist design. Based on behaviorist principles, objectivist instruction is goal-oriented, focused on the accomplishment of a measurable task (Wilson, Teslow, & Osman-Jouchoux, 1993). Knowledge is considered an object to be transmitted from teacher to student; an end unto itself. The instructional model preferred by most students seems to be drills. For students who understand the process of solving aural skills problems, whether cognitively or intuitively, this model will prove efficient.

In contrast, constructivism promotes learning as both personal and experiential, with knowledge treated as a tool rather than an end. One constructivist-based model, known as cognitive apprenticeship (Collins, Brown & Holum, 1991) asserts that learning is best accomplished in a realistic situation imitating the master-apprentice relationship. This model calls for providing diverse types of knowledge through demonstrations and practice, along with methods to promote the development of expertise and the ordering of learning activities. Many struggling students lack an understanding of the fundamental process for solving aural skills problems. For these students, this model is appropriate and necessary as a prelude to drills.

Locus of control. There are four possible options for control: learner control, program control, adaptive control and learner control with advisement. Earlier studies primarily used a competency-based model and most early commercial software reflected this. However, Hofstetter (1978) reported that this design also increased frustration. Tennyson (1981) has found that learner control with advisement can be as effective as program control, but without the accompanying frustration. In general, motivated students will choose the method they perceive to be most efficient. While those decisions prove to be fairly accurate for students who are progressing satisfactorily, students who are struggling appear to make poorer choices. More forcible suggestions regarding sequence, coupled with hyperlinks, should be incorporated in areas likely to be used by students who are having difficulty.

Motivation. The successful music student is reasonably self-motivated--those who aren't will require greater intervention than a computer program. Even so, the evidence suggests they require some small amount of extrinsic motivation. Simply requiring the use of any software will not suffice. The motivation to use the software must lie primarily within the software. Relevance, effectiveness and efficiency are the primary requirements of most students. Students have little time to waste, particularly music students. Most studies have been conducted with software that did meet the criteria. As the courseware was most often designed by instructors to be used with the specific curriculum, students were confident that the program would accomplish the goals of the curriculum. Nonetheless, some extrinsic motivation is required. Students must make choices as to the use of their time. Activities which are not required and have no mechanism for accountability, regardless of their merit, are often left undone.

Instructional Strategy. Most authors have stressed that dictation drills are intended as a supplement to classroom instruction, not a replacement (Foltz, 1980). However, a number of the more successful programs have used the computer as the primary method of delivering dictation (cf. Hofstetter, 1975). The results of this study do not suggest that using the computer exclusively is the best method, but that it can and should be relied on to be more than a voluntary supplement.

For computer-based ear training instruction to be successful it must be a required part of the curriculum. The software must be well-designed, with a clean interface, maintain accurate records and provide appropriate feedback. The more closely the software conforms to the curriculum it supports, the greater the intrinsic motivation. Drills remain the preferred method for most students, but for students with less developed skills a model which focuses on cognitive development is more appropriate. The student must be given control of the instruction with both instructor and program advisement. This advice should be more prevalent and direct in the cognitive model and when remediation is required.

Alessi, S. M. & Trollip, S. R. (1992). Computer-based instruction: Methods and development. Englewood Cliffs NJ: Prentice Hall.

Collins, A., Brown, J. S., & Holum, A. (1991). Cognitive apprenticeship: Making thinking visible. American Educator, 15:3, 6-46

Covington, K. (1992). An alternative approach to ear training. Journal of Music Theory Pedagogy, 6, 5-18.

Foltz, R. E. & Gross, D. S. (1980). Integration of computer-assisted instruction into a music program. Journal of Computer-Based Instruction, 6 (3), 72-76.

Gross, D. S., & Foltz, R. E. (1983). Ideas on the implementation and evaluation of a CAI project. Computer Music Symposium, 21(2), 22-26.

Hess, Jr., G. J. (1994) Dictation Tutor: The effectiveness of a curriculum-specific tutorial in the acquisition of aural discrimination students for college music students. Ann Arbor MI: UMI.

Hofstetter, F. T. (1981). Applications of the GUIDO system to aural skills research, 1975-80. College Music Symposium, 21 (2), 46-53.

Hofstetter, F. T. (1980). Computer-based recognition of perceptual patterns in chord quality dictation exercises. Journal of Research in Music Education, 28 , 83-91.

Hofstetter, F. T. (1978). Instructional design and curricular impact of computer-based music instruction. Educational Technology, 18, 50-53.

Hofstetter, F. T. (1975). GUIDO: An interactive computer-based system for improvement of instruction and research in ear training. Journal of Computer-Based Instruction, 1 (4), 100-106.

Kulik, J. A., & Kulik, C. (1991). Effectiveness of computer-based instruction: An updated analysis. Computers in Human Behavior, 7 (1&2), 75-94.

Pembrook, R. G. & Riggins H. L. (1990). "Send help!": Aural skills instruction in U.S. colleges and universities. Journal of Music Theory Pedagogy, 4 (2), 231-242.

Pople, A. (1992). Computer music and computer-based musicology. Computers and Education, 19 (1 & 2), 173-182.

Taylor, J. A. (1982). The MEDICI melodic dictation computer program: Its design, management and effectiveness as compared to classroom melodic dictation. Journal of Computer-Based Instruction, 9 (2), 64-73.

Tennyson, R. D. (1981). The use of adaptive information for advisement in learning concepts and rules using computer-assisted instruction. American Educational Research Journal, 18, 425-38.

Wilson, B., Teslow, J., & Osman-Jouchoux, R. (1993). What does a constructivist ID model look like? A makeover of traditional ID based on constructivist and postmodern ideas. In E. E. Smith & J. G. Smith (Eds.), Proceedings of the Building Partnerships '93: Bridging Technology, Performance, and Development Conference. (pp. 64-74), Breckenridge, CO: Building Partnerships.