Chapter 1 The CAI Survey

THE CAI SURVEY

PURPOSE OF THE CAI SURVEY
The survey was designed to provide general information about how ear-training CAI software was integrated into undergraduate music theory instruction during the 1998-99 academic year. The questions asked pertained to which programs were used, how CAI use was integrated, how CAI was graded, and how the instructor rated the software.

METHODS
The method chosen was a convenience sample using a ten-question e-mail survey that required approximately four minutes to complete. Eight of the ten questions were multiple choice, and one question (regarding the respondent's rating of the software) was optional. In addition, optional comments were requested at the end of the survey. After pretests of the survey instrument were completed in the fall of 1998 it was decided to use three e-mail lists to obtain a sampling of undergraduate institutions. Two of the lists chosen were the SMT (Society for Music Theory) list, and the ATMI (Association for Technology in Music Instruction) list. The third source, referred to as the MG sample, was a list of e-mail addresses which were graciously provided by Ann Blombach, a professor at Ohio State University and author of the ear-training software, MacGAMUT. The SMT list contained 853 addresses, the ATMI list contained 282 addresses, and the MG list contained 103 addresses. The instrument, which is shown in Appendix A, was sent to the SMT and ATMI lists on February 2, 1999 and again on February 12, 1999. It was sent to the MG sample on February 15, 1999, February 23, 1999, and finally on March 6, 1999.

The SMT and ATMI samples were intended to provide a random sampling of undergraduate institutions; they generated 134 responses. The MG sample was intended to provide a closer look at institutions using MacGAMUT software; it generated seventy responses. Five additional responses arose from contact with various people while research for the thesis was being conducted; these included software developers, instructors who were consulted via e-mail, and responses to a copy of the survey posted on my webpage between February 2, 1999, and March 15, 1999. These five responses were counted towards the SMT/ATMI sample bringing the number for the SMT/ATMI sample up to 139 responses.

THE SAMPLE SIZE AND RESPONSE RATE
The actual sample size represented by the SMT and ATMI lists was difficult to determine. Membership lists of the SMT and ATMI lists were obtained on March 4, 1999. The SMT list had 847 subscribers with 843 of these listed as unconcealed e-mail addresses. The ATMI list had 282 subscribers. Although this indicates a population of about 1125 subscribers, the number of institutions represented in the sample is considerably lower-possibly as low as 450 undergraduate institutions.

The 1125 available e-mail addresses were combined into a database to determine a closer approximation of the sample size. This number was reduced to 1066 addresses simply by discounting for duplication between the two lists. This number was further reduced to 1030 by discounting thirty-six addresses that indicated publishers or news organizations. Often, there were ten or more people from the same academic institution subscribed to one of the lists. Because only one response from each institution was counted in the survey results, multiple subscribers were removed so that each institution was represented only once by an academic e-mail address. There were 456 duplicate institutional addresses which further reduced the sample size to 574 possible institutions. Using the SMT research profiles database it was determined that an additional twenty-four of these addresses had no academic affiliation. Of the 550 possible academic institutions remaining in the SMT/ATMI sample, 191 of these addresses lacked any indication of academic affiliation; for instance, there were sixty-four AOL (America Online) subscribers. Of the 550 possible institutional addresses, 134 responded. This indicates a response rate of about twenty-four percent. Even if 100 of the 191 non-academic address are discounted, the response rate is still very low at about thirty percent. One possible reason for such a low response rate may be that the use of the word "survey" in the subject line of the e-mail may have prompted many subscribers to skip the message. In addition, many persons subscribed to the SMT or ATMI lists were undoubtably students or instructors not directly associated with any undergraduate aural skills classes.

The response rate of the MG sample was sixty-eight percent. Although the MG sample contained more than 150 addresses representing 145 institutions, a few of the addresses were of students or secretaries at music schools, and twenty of the institutions were already represented in the random sample. The instrument was sent to instructors at 123 institutions; however, eighteen were immediately returned as undeliverable. Two working addresses represented instructors no longer on the faculty at the indicated institutions. The sample size represented 103 institutions, and generated seventy responses. This higher response rate of sixty-eight percent may be due to the fact that the instrument was e-mailed privately to each individual. MacGAMUT users may also have been more inclined to fill out surveys and more likely to speak favorably about the ear-training program.

DEMOGRAPHICS
The survey represents seven countries in addition to the United States; forty-three of the fifty states are represented, as are Puerto Rico and the District of Columbia. The survey includes responses from 194 four-year colleges and fifteen two-year colleges. Ten of the responding institutions were not listed in the College Music Society Directory of Music Schools; however, the survey represents fifteen percent of the four-year undergraduate institutions listed in the 1999 CMS Directory-or 184 of the 1213 listed four-year schools. Appendix B lists alphabetically the 209 institutions represented in the survey. Appendix C lists the geographic distribution for the 209 survey responses along with the subset of seventy responses generated by the MG sample. The geographic distribution of the1817 institutions represented in the CMS directory are also listed. In order to maintain the confidentiality of the responses, the numbers for the seventy institutions in the MG sample are indicated only as a subset of the 209 responses. In cases where the survey represents only one institution from a geographic region in the U.S., the listing for the MG sample is indicated as not applicable.

The SMT/ATMI sample tends to represent larger institutions with fifty-nine percent of the responses coming from institutions with 100 or more music students. This may be due to a bias in the survey method that would favor larger schools. The MG sample tends to represent smaller schools, and serves well in complementing the SMT/ATMI sample. The size distribution of undergraduate institutions represented in the survey is show in Table 1 below and indicates that institutions of all sizes are well represented. The smaller number (139) represents results from the SMT/ATMI sample only, while the larger number (209) includes results from the MG sample.

TABLE 1 SIZE DISTRIBUTION OF INSTITUTIONS

Number of Music Majors	SMT/ATMI n=139		MG + SMT/ATMI n = 209
0 to 19	21	15%	33	16%
20 to 49	16	12%	32	15%
50 to 99	19	14%	33	16%
100 to 199	24	17%	42	20%
200 or more	59	42%	69	33%

BIASES OF THE SURVEY METHOD
The SMT and ATMI lists provided a convenience sample that represents a random sampling of schools of differing sizes, but there is a pronounced technology bias in the type of individual who was able to respond to the survey. Only instructors regularly using e-mail and subscribing to one of two e-mail lists were even likely to see the survey. There may also have been a reluctance on the part of instructors to respond at all if their institution was not using ear-training CAI. Institutions not using ear-training CAI are therefore likely to be under-represented in the survey results.

There is also a bias toward the type of undergraduate institution, whether four-year or two-year, that was likely to respond to the survey. While the CMS directory lists 1213 four-year schools and 506 two-year schools, the 209 responses represent only fifteen two-year schools. One possible reason may be the phrasing of the second question in the survey which asked for an indication of the number of undergraduate "music majors". A second possibility is that the above mentioned technology bias may be even more pronounced with regard to smaller two-year institutions. The technology bias, as well as the low response rate of about 24%, prevent the SMT/ATMI sample from being a truly random sample which can be used to make generalizations regarding the state of ear-training; however, a descriptive analysis of the survey results follows.

SURVEY RESULTS
Many of the tables used below will provide results from the SMT/ATMI sample followed by the results from the SMT/ATMI and MG samples combined. This is done to provide as much information as possible and to allow for comparison between the responses from the various samples. Table 2 shows that twelve of the 139 responses from the SMT/ATMI sample reported that they did not use ear-training CAI. Of these, one respondent reported having used CAI in the past but had since discontinued its use. Others mentioned that they were currently looking into CAI for ear-training. Considering the technology bias of the survey, one could infer that the actual percentage of undergraduate institutions using ear-training CAI is significantly less than the 91% indicated below. One response from the MG sample indicated that ear-training CAI was no longer being used.

TABLE 2 PERCENTAGE OF INSTITUTIONS USING CAI

CAI usage	SMT/ATMI n=139		MG + SMT/ATMI n=209
None	12	9%	13	6%
CAI used	127	91%	196	94%

Appendix D lists more than twenty-five commercial programs along with a number of the 22 "homegrown" CAI programs used in the responding institutions. The number of institutions that reported using each program is also indicated. Many schools reported using two or three different software programs. While the use of multiple ear-training programs may imply a search for variety, it may also indicate a level of dissatisfaction with the software currently being used. Conversely, the use of a single program may indicate a greater level of satisfaction with the CAI software. The SMT/ATMI sample indicates that forty-six percent of institutions using CAI use two or more programs. Table 3 shows the rates at which multiple CAI programs are used in the responding institutions.
TABLE 3 NUMBER OF PROGRAMS USED

Number of programs used	SMT/ATMI n=127		MG + SMT/ATMI n=196
1 program	69	54%	113	58%
2 programs	32	25%	47	24%
3 programs	16	13%	21	11%
4 or more programs	10	8%	15	8%

The Macintosh platform was by far the most widely used platform for ear-training CAI. At least eighty-seven percent of the institutions from the SMT/ATMI sample reported using Macintosh computers for ear-training. IBM-compatible computers were used for ear-training at about sixteen percent of the institutions from the SMT/ATMI sample. Table 4 shows the computer platforms used for ear-training CAI.
TABLE 4 PLATFORMS USED FOR CAI

Platforms used for CAI	SMT/ATMI n=127		MG + SMT/ATMI n=196
Macintosh	105	83%	173	88%
IBM-compatible	16	12%	16	8%
Both Mac and IBM	5	4%	6	3%
NeXT	1	1%	1	1%

The percentage of course grade determined by CAI use ranges from nothing to more than eighty percent. The SMT/ATMI sample shows that three percent of the schools reported using CAI for fifty percent or more of the grade, but that forty percent of the schools used CAI for ten percent or more of the grade. Similarly, forty percent of the schools used CAI only for ungraded practice. The MG sample indicates that institutions using MacGAMUT are more likely to grade the use of ear-training CAI. Table 5 shows the percentage of the grade based on CAI usage in the responding institutions.
TABLE 5 PERCENTAGES OF GRADE BASED ON CAI USE

Grade evaluation based on CAI use	SMT/ATMI n=127		MG + SMT/ATMI n=196
Ungraded Practice	51	40%	66	34%
Extra credit	8	6%	11	6%
1% to 9% of the grade	15	12%	25	13%
10% to 19% of the grade	30	24%	54	28%
20% to 29% of the grade	12	9%	18	9%
30% to 39% of the grade	2	2%	10	5%
40% to 49% of the grade	3	2%	3	1%
50% or more of the grade	4	3%	4	2%
Other	2	2%	5	3%

Many different methods of integrating the CAI software were reported. The most common use for CAI in both samples was some form of graded practice. The SMT/ATMI sample indicated that thirty-four percent of the responding institutions used CAI only for ungraded practice. Grades were based on passing levels or completing tests at twenty-one percent of the institutions. Recording of practice time along with completing levels accounted for CAI use in another twenty-one percent of the institutions. Approximately eleven percent of the institutions included CAI as lab work during part of a class period, and nine percent of the institutions based the grade only on the amount of time spent with CAI. Table 6 shows the various methods of integrating CAI into undergraduate aural skills classes.
TABLE 6 METHODS OF INTEGRATING CAI

Integration	SMT/ATMI n=127		MG + SMT/ATMI n=196
Ungraded practice	43	34%	54	27%
Testing (Passing Levels) Only	27	21%	40	20%
Testing + Time	27	21%	40	20%
Testing + Lab work	3	2%	5	3%
Testing + Time + Lab work	6	5%	17	9%
Time Only	11	9%	22	11%
Time + Lab work	3	2%	5	3%
Lab work	2	2%	5	3%
Extra credit practice	5	4%	8	4%

The most common method for students to access the CAI was in a single computer lab. Nearly eighty percent of schools indicated the use of a primary ear-training lab. About eleven percent added that the CAI could be accessed through a campus network. The limited access to ear-training software in computer labs was mentioned often in the optional instructor comments. TABLE 7 ACCESS TO CAI SOFTWARE

Software Access	SMT/ATMI n=127		MG + SMT/ATMI n=196
One Lab	102	80%	147	75%
Many Labs	23	18%	37	19%
Personal Copies (at least one lab)	2	2%	12	6%
Through Campus Network	(14)	(11%)	(16)	(8%)

The limited number of computers for students to do CAI work was also a frequent comment in the survey responses. Many schools required the students to purchase personal copies of CAI which were not dependent on the use of a single computer lab. Other schools used a campus network to address the accessibility problem. Table 1 previously showed that more than half of the 209 schools in the survey have 100 or more students. Table 8 below indicates that forty-six percent of the schools have fewer than nine computers in a music lab that can access ear-training CAI. Only twenty percent of the institutions have twenty or more computers available in a music lab.
TABLE 8 NUMBER OF COMPUTERS AVAILABLE FOR CAI

Computers available in music lab(s)	SMT/ATMI n=127		MG + SMT/ATMI n=196
1 to 9	59	46%	90	46%
10 to 19	43	34%	69	35%
20 to 29	16	13%	22	11%
30 or more	9	7%	15	8%

Nearly sixty-six percent of the institutions using CAI reported that their ear-training computer labs were connected to the internet. Three instructors from the SMT/ATMI sample responded that the computer labs were intentionally not connected to the internet so that students would not waste time surfing the web. Although a few instructors in the SMT/ATMI sample did not respond to this question, Table 9 shows the breakdown of internet accessibility of the labs in the responding institutions.
TABLE 9 INTERNET ACCESSIBILITY OF COMPUTER LABS

Internet Accessibility	SMT/ATMI n=127		MG + SMT/ATMI n=196
Yes	80	63%	129	66%
No	39	31%	59	30%
Not available	8	6%	8	4%

In order to identify some of the more useful CAI programs, the institutions reporting the use of one CAI program will be further examined. The SMT/ATMI sample contained sixty-nine institutions which reported using a single CAI program. Practica Musica led the list and was reported at forty-three percent of these institutions. MacGAMUT was second and was reported at thirty-two percent of these institutions. Most of the programs listed had been available for five to ten years; however, Auralia, which was first released in 1998, posted a relatively strong showing despite its IBM platform and recent release date. Four commercial programs are represented by a single responding institution. These four programs are: Computerkolleg Musik, Guido, teoría, and Musique. The three "homegrown" software programs include: Audio Challenger written by Anthony Holland, a professor at Skidmore College; Harmonic Idioms written by Edward Chudacoff, a professor at the University of Michigan, and; a set of custom MIDI sequences used for melodic dictations. Table 10 lists the programs used by sixty-nine institutions reporting only a single CAI package.
TABLE 10 SOFTWARE AT SCHOOLS USING ONLY ONE PROGRAM

Programs reported	SMT/ATMI n=69
Practica Musica	30	43%
MacGAMUT	22	32%
Music Lab Melody	4	6%
Auralia	2	3%
C.A.T.	2	3%
ETDrill	2	3%
Other "Homegrown" programs	3	4%
Other commercially available programs	4	6%

Of the sixty-nine institutions reporting only one CAI program, twenty-six reported using CAI for ten percent or more of the grade. Table 11 lists the programs used at institutions integrating CAI as ten percent or more of the grade and using only one CAI program. Practica Musica again tops the list, but MacGAMUT follows as a close second.
TABLE 11 PROGRAMS USED FOR 10% OR MORE OF GRADE

Software titles used for 10% or more of grade	SMT/ATMI n=26
Practica Musica	10	38%
MacGAMUT	9	35%
Music Lab Melody	3	12%
Curriculum for Aural Training (C.A.T.)	2	8%
Computerkolleg Musik	1	4%
Musique	1	4%

Nearly sixty percent of the instructors rated the various ear-training programs as good. About 20% of the instructors rated the software as only fair or poor, and two instructors discontinued using ear-training CAI altogether. A number of instructors did not rate the software. In the few cases where an instructor indicated a rating between two categories, the lower of the two categories was counted. Table 12 shows the approximate instructor ratings of various CAI packages.
TABLE 12 INSTRUCTOR RATINGS OF VARIOUS CAI SOFTWARE

Rating categories	SMT/ATMI n=127		MG + SMT/ATMI n=196
Excellent	19	15%	35	18%
Good	76	60%	114	58%
Fair	24	19%	34	17%
Poor	3	2%	4	2%
Not available	5	4%	9	5%

The final question on the survey instrument asked whether the students seemed to find the CAI helpful. More than seventy percent of the instructors responded that students did find the programs helpful. Some instructors at institutions where CAI was used for ungraded practice commented that although students found the CAI helpful, only a few students actually used the programs.
TABLE 13 STUDENT RATINGS OF CAI HELPFULNESS

Did students find the CAI helpful?	SMT/ATMI n=127		MG + SMT/ATMI n=196
Yes	97	72%	155	77%
Indifferent	15	12%	21	11%
Varies	8	6%	11	6%
No	3	2%	5	3%
Not available	5	4%	5	3%

BRIEF SUMMARY OF SURVEY RESULTS
Of the 209 survey responses, fifty-three percent of the schools had 100 or more music majors. The following comments refer to the 196 institutions that used ear-training CAI. Approximately forty-five percent of the institutions used two or more ear-training programs. The Macintosh platform was used at well over eighty-five percent of the institutions. CAI use was evaluated as part of the course grade at more than fifty-two percent of the institutions. At nearly fifty percent of the institutions, the most common method of integrating CAI software included testing or the passing of levels. Eighty percent of the institutions reported using only one computer lab for the ear-training CAI, and forty-six percent of the institutions had fewer than nine computers in music labs for use with ear-training software. Nearly seventy-five percent of instructors rated the software as good, and seventy-seven percent said that students seemed to find the software helpful. There were sixty-nine institutions from the SMT/ATMI sample that used only one ear-training CAI program. Practica Musica and MacGAMUT appear as the most used programs in this category and when combined were used at seventy-five percent of these institutions. These two programs were also used in seventy-four percent of the 127 institutions that used CAI and were from the SMT/ATMI sample. While thirty of these institutions reported using both programs, Practica Musica was used at seventy-two institutions, and MacGAMUT was used at fifty-two institutions.