This study investigated the differential effects of computer conferencing on expository writing for students of seven intelligence types. Students were assigned to treatment groups that provided controlled exposure to a topic: unstructured exposure; computer conferencing; face-to-face discussion; and computer conferencing and face-to-face discussion. All students were classified according to their intelligence type(s) and then wrote an essay on the topic. MANOVA indicated that participation in computer conferences about a specific topic did not significantly improve scores on essays about that topic. However, interactions between treatments and intelligence type were significant. Intelligence type is an important consideration when implementing computer conferences.
Background and Theoretical Perspective
Computer conferencing involves synchronous and asynchronous text-based electronic messaging, permitting members of groups to communicate with each other at the convenience of each of the members. It allows for self-paced learning and reflection, thorough reading of one's own and others’ entries and responses, easy storage of online text conversations, and participation in discussions without having to travel to a classroom. Messages are available to all conference members whenever they log on to a conference. Conferences are usually structured around specific discussion topics (Eastmond, 1992, Vrasidas & Mahesh, 1999).
The application of computer conferencing as a teaching/learning tool is increasing dramatically as students in educational institutions gain access to the Internet and as online course delivery becomes more common. Several studies indicate that computer conferencing can be a powerful force for facilitating discussion and for encouraging writing (Rafaeli, Sudweeks, & McLaughlin, 1995; Tagg & Dickinson, 1995). In such studies, researchers have found that when computer conferencing extends over time, groups develop positive relational communication (Walther, 1997) and that moderators need to overtly establish rapport with group members in order to facilitate positive communication (Murphy & Collins, 1997). Also, research indicates that a task-oriented condition in computer conferencing does not create an atmosphere of greater work effort; but a socially-oriented condition in computer conferencing results in more work effort (Walther, 1997). Other findings are that it encourages both reflection and interaction (Harasim, 1990) and that when compared to face-to-face groups, computer conferencing participants are less satisfied with communications processes, but generate ideas as well or better (Hollingshead, McGrath, & O'Connor, 1993; Straus & McGrath, 1994). In one study, discourse analysis revealed that computer conferencing contributed to preservice teaching apprenticeship and learning of educational psychology by providing for positive communication during early field experiences: “students were heavily involved in electronic learning … and teachers electronically scaffolded or apprenticed learning …without giving away answers” (Bonk, Malikowski, Angeli, & East, 1997).
Although a great deal of research has been conducted to analyze computer conferencing processes, our knowledge of external outcomes of computer conferencing and differential effects on different types of students is limited. Research on computer conferencing has been based largely on analyses of conference transcripts and results of attitude surveys. McIsaac and Gunawardena (1996) reported that more than 23 percent of the literature they reviewed focused on issues related to technology and the role of the distance educator rather than the role of the learner. Hara and Kling (1999) charged that anecdotal reports, which typically focus on the positive aspects of Web-based coursework, predominate over analytical research. Anglin and Morrison's (2000) review of distance education research included in The American Journal of Distance Education between 1987 and 1999 and Distance Education between 1991 and 1999, categorized 62 percent of the articles as conceptual, while only 38 percent of the articles were primary research studies. They concluded that “A variety of research designs are needed to answer questions on distance education” (p. 193). They recommended aptitude-treatment interaction, longitudinal, developmental, and media replication studies.
In this aptitude-treatment interaction study, the authors investigated the effects of computer conferencing on expository writing after participation in a computer conference. Our assumption that computer conferences would positively affect learning as reflected in expository writing was based upon constructivist learning theory, which indicates that providing learners with opportunities for written reflection, interaction, shared perspectives, and mentorship contributes positively to knowledge construction (Driscoll, 1994, Jonassen, Peck, & Wilson, 1999). In computer conferencing, participants grapple with their ideas and have time to think through their entries before carefully constructing messages and replies to others’ messages. They read each other's ideas and responses to their entries. On-line dialog provides students with the opportunity to test and refine their understandings in the ongoing process of social negotiation of meaning (Vygotsky, 1978).
Computer Conferencing and Writing
Several research studies have indicated that having online audiences for one's messages in computer conferences provides students with a purpose to write well; that is, to convey messages clearly and even eloquently (Duin & Hansen, 1994; Gallini & Helman, 1995; Garner & Gillingham, 1996; Riel, 1995; Weston, 1997). The online environment allows learners to shape understanding or socially construct meaning based upon feedback that they get from other students and/or moderators in the computer conference. Spaulding and Lake (1991) suggested that freeform online discussions provide for such shaping of understanding and that, therefore, specific writing tasks should be de-emphasized. Fabos and Young (1999) raised the interesting question regarding the content of freeform discussions: “Will students be innately curious or brave to raise questions and clarify meaning online, or will only certain students be?” (p. 225). They acknowledge the possibility that some learners may not benefit from telecommunications exchanges, and telecommunications exchanges may either help or exacerbate problems in education.
Another body of research explores the impact and effects of informal writing on learners’ abilities to write well. E-mail and computer conferencing exchanges tend to be informal in tone. Some researchers suggest that students relax in those informal environments and that the state of relaxation facilitates better writing (Batson, 1988; Wright, 1992), although others suggest that informal environments impede quality writing and foster sloppiness and off-task activity (Hawisher, 1992; Dyrli & Kinnaman, 1996). As part of this debate, we need more evidence regarding the impact that writing in computer conferences has on formal, expository writing.
One merit of telecommunications is that they can facilitate communication for people with various capabilities (Jonassen, et al., 1999). Moallem (1999) noted that:
Communication via CC tends to be more balanced or equitable [than in face-to-face classrooms]. Members with status and gender differences participate equally in CC discussions and it is less likely that one individual would dominate the discussion in the CC environment. (p. 4)
In addition, some researchers describe members of a computer conference as less inhibited than their face-to-face counterparts (Bordia, 1997). The asynchronous quality of computer conferencing allows learners time to consider an idea and formulate a response, thereby eliminating the problem of having to think on one's feet as in a face-to-face discussion. On the other hand, some suggest that, just as in face-to-face discussions, a few students dominate and many students contribute at a minimum (Gillingham & Topper, 1999). Such findings indicate that individual differences may play an important role in how active learners are in computer conferences.
In addition to investigating the comparative effects of face-to-face discussion, computer conference, and a combination of computer conference and face-to-face discussion, we included a second facet in our study: an investigation of the interactive effects of computer conferencing and participants’ multiple intelligences on students’ expository writing (Gardner, 1993).
Multiple Intelligences
Gardner's (1993) theory of multiple intelligences indicates that all individuals “have a repertoire of skills for solving different kinds of problems” (Armstrong, 1994, p. 26). In other words, each person might have an array of intelligences. Intelligences are both innate to learners and can be developed. Instructors can tailor activities to individual differences in students’ intelligences or design instruction meant to teach specific intelligences. Instructors can facilitate development of multiple intelligences by portraying topics of study using multiple representations and encouraging performance of varied types (Gardner, 1999). Initially, Gardner identified seven distinct and independent intelligences: linguistic, logical-mathematical, spatial, bodily/kinesthetic, musical, interpersonal, and intrapersonal. In 1998, he identified an eighth intelligence, that of the naturalist. Because a validated instrument was not available that included this intelligence at the time of this study, it was not included in our data.
Implications of multiple intelligences for education lie in the fact that the theory was “developed as an account of human cognition that can be subjected to empirical tests” (Armstrong, 1994, p. 27). Core abilities in each of the intelligences are detected from a person's first year of life by observing raw patterning ability. Later on, they are identified through symbol and notational systems, and finally through vocational and avocational pursuits. As Gardner (1993) noted, “…all humans possess certain core abilities in each of the intelligences.” However, some people are “at promise” while others are “at risk” in a given intelligence. In the absence of special aids, those at risk in an intelligence will be most likely to fail tasks involving that intelligence. Conversely, those at promise will be most likely to succeed. However, when “at promise” students are effectively integrated into a group, the overall knowledge of the group may be advanced in all intelligence types (Gardner, 1993). Gardner's research found that when students work in one of their stronger domains, they may undergo a strong positive affective reaction, and that young adult students may benefit from explicit instruction in the notational system of a type of intelligence. For instance, students can learn linguistic intelligence through experience with reading, writing, speaking, and debating, notational systems of the linguistic type of intelligence.
Computer Conferencing and Multiple Intelligences
Computer conferencing is a linguistic and interpersonal strategy involving reading, writing, and corresponding. The impact of learner differences such as multiple intelligences is an issue that needs to be addressed when investigating computer conferencing, because learner differences regarding cognitive and social barriers as well as attitudes may affect successful use of computer conferencing for learning (Moallem, 1999). Studies have shown that individuals’ attitudes toward computer conferencing differ (Burge, 1994; Eastmond, 1994; Robinson, 1998). Variable attitudes toward computer conferencing may reflect different intelligence types in learners. For instance, computer conferencing may facilitate learning best for linguistic learners because their notational system is being used. In addition, attitude and/or intelligence type may provide barriers to learning in computer conferencing for nonlinguistic learners. Computer conferencing might impact intrapersonal learners negatively because such learners dislike having to interact with others as part of their learning experience. Because of the specific representation modes of computer conferencing, we would expect to find differential effects across students of different intelligence types.
In a study similar to the current investigation, Brenner and Hill (1997) explored the comparative effects of asynchronous distance learning on achievement for field-dependent versus -independent learners and found no interaction. However, we expected to find a trait-treatment interaction in this study because multiple intelligences theory indicates that, for students to learn, each student's individual intelligence types must be addressed through varied representation modes.
We explored whether or not an individual's repertoire of skills, or intelligences, are related to a student's ability to learn through computer conferencing. We suspected that some types of intelligence might help students grasp concepts through the computer conferencing process and, in turn, clearly express what they learn, and that other types of intelligence might interfere with one's abilities to use computer conferencing effectively. Thus, we conducted this investigation of the interactive effects of computer conferencing and the multiple intelligences. If interactions exist between type of intelligence and instructional media applications, then those interactions need to be identified so that alternative media and teaching strategies can be applied to learners according to their type(s) of intelligence or so that explicit instruction in the notational system of a type of intelligence can be provided prior to computer conferences.
Research Questions
We investigated the effects of computer conferencing and face-to-face discussions on a specific topic on student expository writing about that specific topic. We asked whether or not students learned about the specific topic from their on-line discussions and face-to-face discussions, above and beyond unstructured study. Specifically, did scores on essays about a particular topic improve as a result of participation in a face-to-face discussion on that topic, participation in a computer conference on that topic, or a combination of a face-to-face discussion and a computer conference on that topic? In addition, we asked, did computer conferences or face-to-face discussions differentially affect writing for students of different intelligence types?
Based upon constructivist learning theory, we expected participation in computer conferences and/or face-to-face discussions to positively affect scores on a composition test above and beyond the effects of unstructured study or face-to-face discussion alone. In addition, multiple intelligences theory indicated that unstructured study time, computer conferencing, and face-to-face discussion would have differential effects on learning as reflected in expository writing by students with different intelligences. We expected multiple intelligences to influence the association between participation in a computer conference or face-to-face discussion and students’ composition scores. Because computer conferencing can be described as narrational and social, we expected that students with linguistic and interpersonal intelligence types would learn from computer conferencing, while students with logical-mathematical, spatial, musical, bodily-kinesthetic, and intrapersonal intelligence types would be unaffected by computer conferencing or that it would interfere with their learning.
Method
Ninety-nine undergraduate university students enrolled in an introductory educational technology course were divided into four 11-day computer conferences in which all discussed different topics on educational technology. Participants were preparing to either teach mathematics, science, English, social studies, or agriculture at the secondary level. Thirty-one students were male and 68 were female. All participants had previously used their university e-mail accounts in university courses and had previously participated in Web-boards. For this study, all used First-Class™ software to conduct computer conferences with an instructor/moderator who also led classes in their face-to-face discussions of the topics.
Study Design
A “posttest-only control group design” with four treatment groups was used to explore the comparative effectiveness of unstructured study time (control group), face-to-face discussion only, computer conference discussion only, and both face-to-face and computer conference discussion for preparing students for expository writing on a specific topic. In addition to exploring the differential effects of treatments, MANOVA was conducted to identify interactive effects among treatment groups and multiple intelligences (aptitude-treatment interaction). Independent variables under investigation were treatment groups and multiple intelligences. Dependent variables were content, organization, and overall composition scores on an essay about the specific writing-prompt that experimental treatment groups had prepared for in their computer conferences and face-to-face discussions.
Participants were not randomly assigned to groups. Instead, they received treatment according to their designated lab sections, and lab sections were randomly assigned to groups. However, no demographic factor was identified that distinguished one lab group from another. Groups were each heterogeneous by gender, race, and major.
All groups, including the control group, received the same exposure to the experimental writing-prompt through readings, lecture, assignments, and unstructured study time, and were informed that course tests would involve essay writing regarding topics studied in the course. The experimental writing-prompt computer conferencing discussion topic was “What impact might the combination of instructional design, media, and computing have on learning?” An alternate discussion topic was used in computer conferences and face-to-face discussions for treatments that did not include computer conferencing or face-to-face discussion of the writing-prompt so that those students could spend equal time with the instructor and equal time learning how to use computer conferences (one of the goals of the course). The alternate writing-topic was “What classroom management strategies might you use for integrating technology in the curriculum?” Lab groups were randomly assigned to one of four treatment groups (see Table 1):
Design of Computer Conferences and Face-to-Face Discussions
When designing the computer conferences, we considered issues associated with grading, grouping, collaboration, relevance, and learner control (Cifuentes, Murphy, Segur, & Kodali, 1998). Quality and quantity of entries in the computer conferences accounted for 20 percent of each student's grade. In order to keep the conferences intimate and manageable, students were grouped with approximately 14 conference members. We addressed the issue of relevance through the grading; through group composition based upon class section, which allowed for face-to-face discussion; and by regularly providing examples of applications of computer conferencing in the K-12 classroom. The researchers identified conference topics following consultation with students from a previous semester in the class.
The same graduate assistant served as chief moderator for each of the conferences, and student moderators collaborated with the graduate assistant to facilitate computer conferencing discussions. Several researchers have indicated the importance of shifting from teacher-centered control to student-centered control in computer conferencing (Berge, 1995; Cifuentes, et.al., 1998; Harasim, 1986; Murphy, Cifuentes, Yakimovicz, Segur, Mahoney, & Kodali, 1996; Tagg, 1994). Therefore, once the conference topic was entered by the instructor, control of the discussion was placed primarily in the students’ hands under the supervision of the graduate assistant. All students in the class were trained in conference moderation using Winograd's guidelines in draft form (Winograd, 2002). They learned how to prepare for conferences, address technical problems, create a social space, focus, summarize, and weave discussions prior to the treatments in this study.
In the computer conferences, students read a discussion topic that was entered by the instructor, entered a response to that topic in the computer conferences, read others’ responses, and replied to those responses. They were required to make three entries per week. In contrast to experiences in face-to-face discussions, participants had opportunities to reflect upon their ideas. Therefore, the instructor told the students that responses were expected to demonstrate deeper understanding and be less spontaneous, more deliberate, and more clearly stated than responses in the face-to-face discussions (Gallini & Helman, 1995; Garner & Gillingham, 1996; Riel, 1995).
The instructor shared some of the responsibility for conference moderation with the participants. As part of the course material, computer conferencing participants were trained briefly in the role of a computer conferencing moderator. The moderator's role in computer conferences was to provide structure, keep students on topic, and weave threads of the discussion (Berge, 1995; Eastmond, 1992; Winograd, 2000). The computer conferencing structure was such that the instructor entered the writing-prompt question in First-Class for the computer conferencing and computer conferencing/face-to-face group, made brief comments, and wove the discussion in each of the groups. Students had free reign to discuss whatever topics emerged. The instructor attempted to control for extended tangential discussions by creating new conference areas for those discussions. The instructor entered the alternate writing-prompt in the control group and in the face-to-face group.
For face-to-face discussions, the writing prompt was projected for the face-to-face group and the computer conferencing/face-to-face group. The alternate discussion topic was projected for the computer conferencing-only group and for the control group to read. Students were given 20 minutes to discuss the topic. The instructor attempted to keep students on the topic, mirrored ideas expressed, and contributed a summary of ideas to finalize each discussion.
Data Sources
The data sources included scores on essays written after computer conferences and face-to-face discussions, results on the Multiple Intelligences Inventory, and attitude surveys. The average number of entries in the computer conferences and number of sittings to make entries were determined for each participant in each treatment group. In addition, we characterized the computer conference's contents to ascertain the extent to which students in each group stayed on or strayed from the writing-prompt topic or alternate topic.
The essay writing following treatment was administered in a regular one-hour class session. It required students to respond to the experimental writing-prompt to which they had had differential exposure. The essays from the various treatment groups were collected by the instructor, number coded for anonymity, and scrambled. Benchmark papers were selected by the investigators for rater training.
Raters evaluated the written posttest with a holistic/analytic scoring procedure using the Composition Profile (Hughey & Wormuth, 1985, r = .98). The Composition Profile uses a 66-point scale, with a low score of 34 and a high score of 100 and has consistently shown score reliability of .84 over years of administration to thousands of students, and interrater reliability ranging from .85 to.93. Using the Spearman Rank Correlation Coefficient, intercorrelations of its components range from .57 to .81 with an average of .76; and correlation of the components with the total scores ranging from .73 to .89.
Two raters were trained with the benchmark papers and the scoring rubric and then read each paper independently. Each essay was rated by at least two people, with discrepant scores of more than 5 points being resolved by a third rater. The third rater resolution was made by selecting the two scores that were closest to each other and within the 5-point range. Reliability was assessed using Pearson Product Moment formula. Reader reliability on the content, organization, and overall composition were .91, .92, and .78 respectively. The overall discrepancy rate for this reading was 7 percent.
The above analytic scoring procedure provided a number of ratings on various qualities of the writing. However, since this study was concerned primarily with the written expression of the concepts students had learned as a result of the treatment—rather than with the vocabulary, language use, and mechanics of their writing—we focused on the qualities of content and organization. Thus, evaluation of students’ writing was based on the rubric's first two components, using a truncated 30-point version of the scale, with a low scale of 20 and a high of 50. The combined scores for contents and organization were used to establish a score for overall composition. The dependent variables, then, were scores on the quality of the contents, organization, and overall composition of essays on the writing-prompt discussion topic.
To inform ourselves as to the potential intelligences of the subjects participating in the study, we administered Armstrong's (1994, pp. 18-20) Multiple Intelligences Inventory for Adults to each participating student. The inventory consisted of a list of ten items for each of seven intelligences. The inventory was administered in the following way: in a regular class session, participants read the inventory and checked those items that they believed described them. An additional open-ended item with each intelligence allowed students to describe other strengths they might have in each area. We defined 6 or more selections in any of the intelligences as an indicator of strength in that particular intelligence as described by Armstrong. For groups, we defined a mean of 5.5 in any of the intelligences as an indicator of group strength.
In order to better understand the groups’ dynamics, we counted each student's entries, discarding all non-substantive entries. We examined the contents of the computer conferences and highlighted all language that strayed from the assigned topic. In addition, we administered a survey of students’ attitudes toward computer conferencing in order to determine how students felt about their experiences in computer conferencing, and to determine whether or not groups differed in their attitudes toward computer conferencing.
Aptitude-treatment interaction was explored through MANOVA. The effects of the group experiences on essay scores, and the interactive effects of treatment group and each intelligence type on essay scores were tested across the three dependent variables: content of essay, organization of essay, and overall composition of essay. For those interactions that were found to be significant, follow-up correlation coefficients were calculated to determine which treatment group was affected by intelligence type and if the impact was positive or negative. Effect sizes were calculated to estimate the percentage of variance explained by each interaction.
Results
Contrary to our expectations, treatment groups did not differ significantly on expository writing. Across groups, scores on essays’ contents, organization, and overall composition for the control, the face-to-face, the computer conferencing, and the computer conferencing and face-to-face groups were similar. Students who discussed the writing-prompt in their computer conferencing and/or face-to-face discussion did not perform significantly better on the three measures than those who did not discuss the writing-prompt in a computer conferencing and/or face-to-face discussion (see Table 2).
However, MANOVA identified interactions between treatment group and multiple intelligences. Computer conferences and face-to-face discussions differentially affected learning for students with different intelligence types (see Table 3).
Correlations revealed that students with interpersonal intelligence who participated in face-to-face discussions but did not computer conference on the writing-prompt benefited from such treatment. They out-performed other students in writing content, organization, and overall composition. Students with intrapersonal intelligence were negatively affected by the combination of computer conferencing and face-to-face discussion on both writing content and overall composition. However, computer conferencing alone and face-to-face alone did not negatively affect intrapersonal intelligence types’ writing. Students with bodily-kinesthetic intelligence were negatively affected on overall composition by computer conferencing without face-to-face discussion. However, when these students participated in both the computer conferencing and face-to-face discussions, they were not negatively affected (see Table 4).
Participants in each group had each of the types of intelligence. However, the groups differed in degree of reported intelligences. The control group was the only group with strength in linguistic and spatial intelligences and they were stronger overall in the intelligences. The control group also had strengths in bodily/kinesthetic, musical, and interpersonal intelligences. The face-to-face discussion group, the computer conferencing group, and the computer conferencing and the face-to-face discussion group each were strong in bodily kinesthetic, musical, and interpersonal intelligences and weak in linguistic, logical/mathematical, spatial, and intrapersonal intelligences (see Table 5).
Groups participated in the computer conferences to a similar degree (see Table 6). However, contents of computer conferences revealed that differential group dynamics may have affected the outcomes of the study. The control group, which was given the alternate topic for their computer conference, inadvertently found their way to the discussion of the experimental writing topic, whereas the experimental groups tended to drift away from the assigned discussion of the writing-prompt. This tendency of each of the computer conferencing groups to drift from the assigned topic supports previous studies that have found that learners strive to gain control over topic selection (Romiszowski & DeHaas, 1989; Romiszowski & Jost, 1989). Loss of control of the conference topic meant that the control group inadvertently had exposure to the experimental writing-prompt topic, while experimental groups had less exposure than we anticipated.
Attitudes toward computer conferencing differed across groups. The control group had a generally positive attitude toward computer conferencing. Two students offered that they liked “learning what others thought.” Only one student in that group complained that computer conferencing took too much time. The face-to-face discussion group unanimously liked computer conferencing. However, the computer conferencing discussion group generally expressed dislike for computer conferencing, claiming that it took too much time, was too difficult to access, and that topics were too “nebulous.” Only one student in the computer conferencing discussion group claimed to like the interactivity provided by computer conferencing. The computer conferencing/face-to-face group also generally disliked computer conferencing, claiming that it “didn't contribute to the course,” and that they “preferred to talk to people face-to-face.” Most in this group said that given the choice, they would not participate in a computer conference again. One intrapersonal learner in the group said that although computer conferencing might “help people who don't like to speak up in class to contribute,” he would not participate again and he asked that we not require it in the course.
Summary and Discussion
In summary, computer conferences and face-to-face discussions did not create differences between groups’ scores on expository writing. Face-to-face discussions positively affected writing for learners of interpersonal intelligence, while computer conferences did not. Computer conferencing and face-to-face discussions combined negatively affected intrapersonal learners and computer conferences alone negatively affected bodily/kinesthetic learners. The control group was stronger in the intelligences than were the other groups and was the only group with strength in linguistic and spatial intelligence. Participants in each of the computer conferences tended to stray from their assigned topics. The control group had a positive attitude toward computer conferencing, as would be expected, given its linguistic strength, and the face-to-face group liked their computer conference. However, the computer conferencing and the computer conferencing/face-to-face groups disliked their computer conferences.
Differences between groups’ intelligences and attitudes toward computer conferences may provide an explanation for the lack of difference between the control group and the other groups when writing about the prompt. The control group displayed broader intelligence than did the other groups. Therefore, we can hypothesize that the control group did well on their content, organization, and overall composition because of their strong multiple intelligence and because they had less need of computer conferencing and face-to-face discussion to support their learning of the writing prompt during unstructured study-time. The control group had linguistic, spatial, bodily-kinesthetic, musical, and interpersonal intelligences. The other three treatments were not strong linguistically or spatially. Given the control group's linguistic intelligence, we would expect the group to out-perform the other groups on an essay prior to treatment. Given the three experimental groups’ profiles of intelligences, (strengths in bodily/kinesthetic, musical, and interpersonal learning) we would expect their learning processes to be best served by physical experience, musical or lyrical experience, and human interaction. Multiple intelligences theory indicates that face-to-face discussions meet the needs of those types of learners, as was the case in this study.
Also, our expectation of low-to-moderate composition scores for those of low linguistic intelligence was met. In fact, the composition scores were higher for the three experimental groups than we would expect without treatment. That is to say, they were comparable to the control group that had linguistic strength. This suggests that the treatment had a positive effect because it brought these three groups up to a composition score equivalent to the linguistically strong control group's composition score (see Table 2).
Given the weaknesses in intelligences and attitudes of the computer conferencing group and the computer conferencing/face-to-face group, we would not expect the mean score on the composition profile to be as high as that of the control group. However, content, organization, and composition scores of this group were equivalent to scores of the other groups. While not significantly higher than scores for the other groups, we think that the students may have overcome weaknesses through participation in computer conferencing and face-to-face discussions.
Computer conferences did not facilitate learning for interpersonally strong learners. Perhaps they experienced the computer conferences as being impersonal. That intrapersonal learners were negatively affected by the combination of computer conferencing and face-to-face discussion, but not by either of those strategies alone, indicates that they may have been overwhelmed by too much interpersonal experience. That bodily-kinesthetic learners were negatively affected by computer conferencing without face-to-face discussion may reflect a need for physical human presence rather than the distance interaction provided by computer conferences.
A second possible explanation for the lack of difference between groups is that the writing prompt may have been too easy or familiar for participants to address during unstructured study of the text and lecture notes. A follow-up study using a writing-prompt on a topic to which no participants have been exposed might provide better understanding of the effects of computer conferencing on writing. Also, unexpectedly, the writing-prompt topic and alternate topic elicited entries that overlapped in concept so that the computer conferencing groups that were not assigned to discuss the writing-prompt topic still explored it in their alternate topic computer conferencing. Similarly, the groups that were assigned to discuss the writing-prompt topic in computer conferencing wandered from the topic in their discussions. We see loss of topic control as a research problem, but continue to embrace the value of learner control of the topic for learning (Cifuentes, et al., 1998).
A third explanation for lack of difference in student writing among groups is that perhaps students in the experimental groups did not use the computer conferencing effectively to grapple with meaning and construct new knowledge. We would expect those less linguistically talented students to be less able to construct meaning in the context of verbal discussions. Also, as the number of sittings per student and the amount of time spent in the computer conferences indicates, many students in each of the groups entered the computer conferencing to meet the minimum course requirement and may or may not have taken time to read other students’ comments. Therefore, we would not expect them to increase their understanding of the writing prompt through computer conferencing. These negative attitudes could well have influenced the power of the computer conferences to positively affect learning.
Educational Significance and Implications
Computer conferences are often used in distance education to facilitate interaction in online learning communities. Findings in this research provide evidence that multiple intelligences are important variables to consider when incorporating computer conferences in courses. Gardner (1999) recommends tailoring instruction to individual differences in students’ intelligences and providing special aids for those at risk in an intelligence. He offers six discrete “entry points” that instructors can design for appealing to individuals’ intelligences: narrational, quantitative/numerical, foundational/existential, aesthetic, hands-on, and social. Computer conferences can be both narrational and social entry points to understanding and therefore appeal most to linguistic and interpersonal learners. In addition, our findings indicate that experience in computer conferences may serve as the special aids recommended by Gardner for students who are at risk linguistically and/or interpersonally. They also indicate that special aids to support computer conferencing experiences prior to computer conferencing are most appropriate for bodily/kinesthetic and intrapersonal learners.
Our findings support Gardner's theory of multiple intelligences and indicate that online instruction should be designed using multiple modes of representation and that instructional designers of distance learning must resist “the temptation to represent the topic in one ‘optimal’ mode” (Gardner, 1999, p. 71). The linguistic, interpersonal nature of computer conferences indicates that they are not optimal entry points of instruction for all learners, even when those learners are as sophisticated in their learning as undergraduate students at a university. As computer conferences are commonly the major source of interaction in online courses, instructors should consider including alternatives to computer conferences for constructing understandings that provide for introspection and tactile experience, such as student-generated development of graphic organizers, multimedia presentations, or trigger videos (Cyrs, 1997).
Our findings indicate that computer conferences on a specific topic do not affect expository writing about that topic above and beyond face-to-face discussions or unstructured study. Students did not appear to learn what we intended for them to learn through computer conferencing and face-to-face discussion. When a specific writing topic was addressed for as long as eleven days, the computer conferencing students who participated in that computer conferencing did not out-perform those who did not participate when asked to write about the topic. However, each of the computer conferencing groups were weak linguistically and it may be that the computer conferencing compensated for this weakness and brought those learners up to the level of the linguistic learners in the control group. Therefore, we think it is premature to conclude that computer conferences have no impact on expository writing. Rather, studies that control for multiple intelligences among learners are indicated.
Although we had a specific objective in mind for learning in the computer conferencing, students took tangents from this objective as they discussed content of special interest to them. They collaboratively constructed meaning about topics of their choice. Collaborative, social learning environments such as computer conferences may not provide an atmosphere for learning specific objectives. This finding supports Spaulding and Lake's (1991) suggestion to provide freeform online discussions. As Ahern (1998) noted, “Conferencing software is … suited for ill-defined goals that require more interaction with a higher group interdependence.” Further research should be conducted to identify what is learned in computer conferencing above and beyond specific objectives. A follow-up study involving content analysis would be appropriate for identifying what was learned in the computer conferences under study. It seems likely that constraining computer conferences to a specific objective limits learning.
Our findings indicate that computer conferences should be used primarily to support constructivist environments as described by Jonassen, et al. (1999):
… rather than forcing students to conform to prepackaged instructional requirements, emphasis should be placed on the social and cognitive contributions of a group of learners to each other, with students collaborating and supporting each other toward commonly accepted learning goals. (p. 119)
Such environments do not prescribe what is to be learned in the environment but instead provide discourse and knowledge-building communities. Our findings support Walther's (1997) conclusion that task oriented discussion is less productive than socially oriented discussion in computer conferences.
Although we did not succeed in creating computer conferences that supported learning of a specific objective, we recommend further investigation to explore design strategies for developing computer conferences that help students attain specific objectives when mastery of certain curricular or disciplinary materials is critical to the course. Multiple intelligences theory (Gardner, 1999) indicates that valued course content should be accessed by students through multiple representations and that students understand concepts when they are able to represent the concept in more than one way. Computer conferencing and expository writing each provide narrational modes of representation and, therefore, should not be considered the end-all for learning or assessing any one course objective. Other modes of representation should be provided in courses, and students should be assigned to create alternate representations. In addition, that linguistically talented students did not excel in the objective through the computer conferencing indicates that those types of learners should use computer conferences constructively rather than to meet specific objectives.
Our findings indicate that computer conferences might negatively affect writing about a specific writing prompt for students with bodily-kinesthetic and intrapersonal abilities. Computer conferencing may interfere with those students’ abilities to write about the topic discussed in the computer conference. These findings support Fabos and Young's suggestion (1999) that some learners may not benefit from telecommunications exchanges and that, in fact, telecommunications exchanges may either help or exacerbate problems in education. Palloff and Pratt (1999) shared this concern and offered that “It is unrealistic to expect that all students will do well [in online courses]…. This should not be considered a failure but simply a poor fit” (p. 8). In addition, our findings indicate that the repertoire of skills needed for students to use face-to-face discussions to help them grasp concepts and, in turn, clearly express what they learned in an essay were the interpersonal skills (Gardner, 1999).
Students might be heterogeneously grouped in computer conferences so that learners with weaknesses in specific intelligences can benefit from those who are strong in those intelligences (Gardner, 1993). In addition, because Gardner indicates that intelligences can be developed or learned, special efforts could be made to prepare learners for the interpersonal skills that this research indicates are necessary for using face-to-face discussions effectively.
Our findings expand the research base related to the impact of computer conferencing on student performance and call for future investigation into specific facilitative strategies that might be employed within computer conferences in order to support constructivist or objectivist learning and to facilitate learning for different types of learners. Our study leads to the conclusion that the impact of multiple intelligences on distance learning within specific media should receive more attention than it has to date. We determined that intelligence types and computer conferencing interact to create a differential effect. Such findings speak to instructional designers regarding strategies to provide student-centered, individualized, as well as collaborative learning environments.
