During the Fall semester of 1997, ten senior students at Miryang University, South Korea, left school after midterm exams to start work in other cities. These seniors were studying English Conversation with the author, who was unwilling to allow them to finish their studies prior to the end of the semester.

My decision to teach these students via telephone was made without reference to published research. Following the announcement of the teleconferencing sessions, I surveyed the literature for telephonic foreign language instruction, and found no information. Research of this sort continued through the months following the course. In addition, queries were posted on electronic discussion lists and an extensive internet search was conducted throughout the instructional period and beyond. A response from my inquiry to Stan Zenor, Executive Director of the Association for Educational Communications and Technology, was revealing. "Sounds like you have been doing distance learning at almost it's most basic level. I really don't know of anyone else who has conducted a class in this manner, although it must have happened some place at some time. I just don't have any idea where to suggest you look."

Studies have been conducted supporting teleconferencing in various fields of study, often concurrent with various forms of "live" video presentation: none were found specific to the efficacy of communicative-language instruction. Numerous leads were tracked, yet no current instructional programs in foreign languages via telephone were found. Those who had been involved in telephonic language instruction were unable to provide data, or no longer had documentation for programs long-since terminated. Others had done only teletutoring courses, not group instruction.

The telephone group continued with the same course workbook used with all English Conversation courses. This course book was designed in previous semesters by me to maximize student-talking by emphasizing listening skills and information-sharing in initial sessions, and requiring pair-work and small group-discussions for task completion. One coursebook unit is designed for one contact-hour. Anticipating some trouble-spots, a few unit substitutions were made, and supplemental materials mailed to students during the fourth week of teleconferencing; e.g. color photocopies of photographs used as discussion points were sent to all students, as well as a simplified "Locations and Directions" unit.

The telephone instruction lasted for five and a half weeks, into the examination week, three nights per week. Concurrently, the traditional group maintained the original teaching schedule, meeting two course-hours per week. Most of the telephone group participants missed the semester's ninth and tenth weeks of instruction, while I arranged the technical aspects of teleconferencing, as they had already left the area. The telephone group had two class sessions during final exams week (while the traditional group had one) for a total of 15 telephone class sessions. By the end of the course, the telephone group had had the same number of sessions with the instructor as had the traditional group, though these contacts were of shorter duration.

A typical teleconferencing class session consisted of three components: the group session, pairwork, and instructor follow-up. Students were to be available from 11:00 until 11:55pm. The operator would contact them at 11:00, or shortly thereafter. It was anticipated that group instruction and discussion would continue until approximately 11:35, at which time the conference call would terminate, and students would call their partner via ordinary telephone service for 10 minutes of pairwork. Partners changed each session. The instructor would then telephone a few students between 11:45 and 11:55 via ordinary telephone service to check their work and do one-on-one remedial work. That plan was realized for the most part; depending on the unit, some group sessions ran longer, some shorter.

These students had studied together throughout their university careers, and were an effervescent group in the classroom. After four or five teleconferencing sessions, a similar mood prevailed over the telephone, except when the material was too difficult for the teleconference environment. At that point the students became silent, they would not participate, and it became a teacher-centered lecture-format course for the remainder of the evening.

Three exams are available for consideration: the midterm exam, the final listening comprehension (L/C) exam, and the (final) Interview. In the absence of a semester pre-test, the midterm exam was used for benchmarking purposes. The midterm exam and final Listening Comprehension (L/C) exam are substantially similar in content and focus. The L/C exam includes materials from throughout the semester, but more importantly, the midterm includes the student's written responses to recorded oral questions. The grading is subjective on this "writing" portion: spelling is not graded, and grammar/syntax is one point of the three points available on each of the 5 questions. (The maximum score on the midterm is 30, 15 can be achieved through the written responses.) A sample question is "I don't like swimming." A correct response would be "Neither do I" or "But I do." Other appropriate responses can also receive full marks. The students are advised that this form of written response is in preparation for their final interview. The small size of the sample together with a high diffusion of scores, makes statistical reporting somewhat suspect, so data clusters from each exam are considered as well as the arithmetic means and medians. The mean for the first (low) quintile ("The Bottom 20%") and the high quintile ("The Top 20%") are reported in Table 1.

The quintiles for the traditional group are composed of three scores, except the third (middle) quintile has four, the telephone group has only two scores in each quintile. The separation between the scores of the telephone group and the traditional group is displayed in the third column as the quotient of the telephone group's score, divided by the traditional groups'.

Figure 1 Midterm Grade Distribution Figure 2 - Listening Comprehension Grade Distribution

The listening comprehension exam (L/C) indicates a strength of the teleconferencing environment. The arithmetic mean and median for the two groups were roughly the same, and the telephone group did much better in the first quintile, outscoring the traditional group, and reaching .819 of the coursewide median score. The separation between the groups in the high quintile decreased from the midterm's .477 to only .882. Figure 2 shows that in the third quintile, the scores overlapped.

Figure 3 overlays these plots (midterm scores x .667 to equate scores).

Figure 3 Midterm / L/C Plot Comparison Figure 4 displays that the telephonic language-learning environment is not a total success. In the interview exam, the lowest two scores of the two groups were identical. All scores in both high quintiles were the maximum score of 20. It is the third and fourth quintiles where there is apparent divergence between the plots of Figures 2 and 4, Figure 5 overlays these plots (midterm scores x .667 to equate scores).

Figure 4 Interview Grade Distribution

Figure 5 Midterm / Interview Plot Comparison

From the students' perspective, the "telephone class" was a success: 87.5% of the respondents voiced approval in terms of usefulness, 85.7% said they enjoyed the course, and 75% thought the course should be repeated for other students. An important consideration when analyzing the test data is whether the telephone group students were more motivated under a quasi-tutorial environment. The special circumstances of this small group (generally 6-9 students each session) allowed a different type of instructor-student interaction during the instructional period. On the other hand, there was virtually no student-instructor contact before or after the instructional period each day. This question is left unanswered in this study, but it would seem a likely benefit of smaller-group instruction under the teleconferencing environment. Of course, were the traditional classroom to contain a similar number of students, those students' motivational levels might increase comparably. More advanced planning would allow a significant change in instructional materials to improve the quality of discussions. This instructor generally prefers not to use a bilingual approach for vocabulary issues, utilizing instead an active visual presentation model; more attention is required to minimize lexical challenge. More coursebook illustrations would assist greatly in this area. In addition, each discussion item, each sentence, needs to be identifiable, so that students can find it easily. Numbering everything would be one solution.

Problems of confidence are exacerbated over the telephone, where students can't just lean over a desk to help each other. Simplified activities need to be developed for troublesome content areas, to reflect this weakened student-support system.

Teacher-centered activities with simple and complete materials for the students, such as "Past-tense Bingo," were generally successful. Group discussion activities were generally not very effective, due to the students' lack of consideration to the technological limitations of basic telephone service: only one voice can be transmitted at a time. A teleconferencing guide should be made available to students prior to the course, to facilitate more active group discussions. Call-waiting and other advanced telephone services are a hindrance. It is frustrating that the instructor can only do one-on-one work at the end of the session, so that students have little chance to practice their "repaired" language/pronunciation. A university with a high-tech communications system could arrange for students to call in for a teleconference course, with the university operator or instructor then matching up calls for pair-work, which the instructor could monitor, and the class could be reconvened. Only in such an environment could a teleconferencing English Conversation class be considered as a full-fledged replacement for a traditional course. One promising application would be in continuing education of English conversation for public school teachers. A large number of the reports published in the field of distance-learning were based on continuing education for professionals.

Many Korean public school teachers have a strong knowledge-base in English, but little opportunity to practice speaking and listening. Small groups of teachers could conference, and discuss topics of interest, with or without the involvement of a "tutor" who is more proficient in spoken English. This would be effective in groups who have existing ties.