An important choice for the instructor is the amount and type of structure to provide for the list. As with participation options, decisions about structure have to do with goals for the course and electronic discussion, type of student, and content area. Decisions need to be made regarding such issues as links between the electronic discussion and the rest of the course, discussion format, participation expectations, and grading plan.

Links to the Course. Once the goals of the electronic discussion have been identified, they must be communicated. For example, an instructor might include a statement in the syllabus that says, "Classes will be mainly devoted to providing examples and clarifying the textbook readings. Because there is a great deal of content to this course, there will generally not be time for questions and comments. This is the role of the electronic discussion. Use this vehicle for asking questions that did not get asked in class, reading the responses to others' questions, making personal applications of the concepts and information, and discussing topics that we do not have the time to treat in class, such as the ethics of experimentation in some studies we mention."

In addition to being explicit regarding the rationale for the use of electronic discussion, it is also important to refer to the electronic discussion during class sessions (or in instructor communication, if the course is a distance education course). Beginning class with an issue that surfaced in the electronic discussion underscores the importance of the electronic discussion and the instructor's attentiveness to student learning in this format.

Discussion Format. The discussion format can be very structured, such as the French class mentioned above, which used one basic student-generated question per week to which all students sent one response. It can also be open-ended, a "Just talk about what's on your mind" format. Most students and faculty in the Ohio State study expressed a preference for a semi-structured format. A difficulty with a highly structured format is that spontaneity is suppressed and often cross-student dialogue does not occur. Students post their "duty" message and leave it at that. The opposite problem occurs with very open-ended structures. Participants may generate discussion points in no apparent order and may respond to old messages with no clear cues as to the context, thus resulting in confusion and unmanageable complexity. One student might be discussing the current topic while another is referring to something that was being discussed two weeks ago. The topics will range from being on task (about course content) to focusing on the weekend's football game or joking or teasing. While spontaneity is apparent, the instructional goals may get lost, unless they are primarily social in nature.

Instructors choosing a semi-structured approach outline the "playing field" by giving the discussion some broad boundaries, such as "The electronic discussion is to focus on critical reactions to the course readings" or "Use the electronic discussion to generate ideas for term paper topics with each other." They then outline some expectations for substantive messages, for example, "Avoid paraphrasing the reading. Assume that we have all read it and go from there, stating your opinion and responses."

Participation Requirements. Expectations for participation are likely to relate to choice of structure. Some instructors require all students to post a certain number of messages, others give guidelines, and still others leave it entirely open to the students. The experience of instructors in the Ohio State study is that some kind of requirement is necessary, although students sometimes expressed resentment with this approach, feeling that they should not have to participate if they have nothing worthwhile to say. The lists in this study that were most successful in terms of quality and quantity of messages, however, all had some participation requirement.

If requirements are used, the instructor must be careful that the amount of interaction that will ensue is manageable, not only from a reading perspective, but also from the perspective of the memory of the computers. Most mail programs have a limit on volume and students using computers in laboratories may have a time limit as well. If there are 60 students in the class, for example, requiring that each post two messages per week means that on the average, each student will receive 17 or so messages per day, assuming weekends are also used. If students check their mail every third day, they will have more than 50 messages to read at a time. For instructors who want very active participation in a large class, dividing the group into different "virtual discussion sections" would be a wise choice to reduce the volume of mail.

Latitude with respect to the timing of messages may also help to alleviate students' resentment at having to respond when they do not have anything to say. For example, requiring a certain number of messages across the term, rather than a weekly quota, is one option. For students who need external pacing, however, the danger is that the discussion will be very slow until the end of the term when students are scurrying to make up for lost time.

Grading. Whether or not requirements are used, instructors also have to make choices about the role of participation in electronic discussion in the course grade. Some instructors do not count participation for the grade at all; some use it as extra credit or consideration in making a decision on a borderline call, such as the difference between a high C and a low B; and others establish a clear point value and criteria for participation, such as "8 to 10 food labels with discussion of ingredient function are needed for an A; the electronic discussion grade will be 20% of the course grade" or "Submission of two drafts of your paper outline and response to one other person's outline is required. This assignment will account for up to 50 points of the total of 200 possible course points." Instructors in the study expressed a preference for a percentage that was more than a nominal 5% or so and most settled on a 15% figure. They also indicated that since participation is recorded in this kind of discussion and thus could be reviewed and analyzed for the grade, they felt more comfortable assigning grades for this form of participation than in-class discussion.