A Field Study on Distance Education and Communication: Experiences of a Virtual Tutor

Authors

  • Dr. Karin Schweizer,

    Corresponding author
    1. Research Assistant at the University of the Federal Armed Forces in Munich (Germany). She received her Ph.D. (1996) and graduated (1993) in Psychology and Information Theory from the University of Mannheim (Germany). She has been a GAAC (German American Academic Council Foundation) Fellow since 1988. With a extensive background in General and Cognitive Psychology and Psycholinguistics, her current research focuses on Distance Education and CMC.
    Search for more papers by this author
  • Dr. Manuela Paechter,

    Corresponding author
    1. Research Assistant at the University of the Federal Armed Forces in Munich (Germany). She received her Ph.D. (1996) in Education and Psychology at the Technical University of Braunschweig (Germany) and her M.Sc. (1989) in Industrial Psychology at the University of Hull (UK). She graduated in Psychology at the Technical University of Darmstadt (Germany) in 1990. Her current research interests lie in Learning with Media, Distance Education, and Computer-mediated Learning in Groups. Together with Prof. Dr. Bernd Weidenmann she is the director of a DFG-project on the social presence of virtual tutors.
    Search for more papers by this author
  • Prof. Dr. Bernd Weidenmann

    Corresponding author
    1. Professor at the University of the Federal Armed Forces in Munich since 1987; he received a Ph.D. in Psychology in 1978. He is co-director of a DFG project on the social presence of virtual tutors. His research concerns include learning with pictures, psychology of media, and learning with the Internet.
    Search for more papers by this author

Address: University of the Federal Armed Forces in Munich, Faculty of Social Sciences, Werner-Heisenberg-Weg 39, D-85577 Neubiberg, Germany.

Address: University of the Federal Armed Forces in Munich, Faculty of Social Sciences, Werner-Heisenberg-Weg 39, D-85577 Neubiberg, Germany.

Address: University of the Federal Armed Forces in Munich, Faculty of Social Sciences, Werner-Heisenberg-Weg 39, D-85577 Neubiberg, Germany.

Abstract

In a field study on distance education and communication we varied the social presence of a tutor in four degrees: a tutor mediated by verbal, written information (condition 1), the same tutor mediated by written information and various personal views (condition 2), the same tutor mediated by written and spoken information (condition 3), and the same tutor mediated by text, views and spoken language (condition 4). Three hypotheses derived from cues-filtered-out (e.g. Short, Williams, & Christie, 1976; Spears & Lea, 1992) and adaptation theories (e.g. Clark & Brennan, 1996; Walther, 1992) were tested: (1) To experience the tutor with less social presence leads to extremely emotional evaluations as well as more task oriented, informal, and tense reactions compared to conditions in which the tutor can be experienced with greater social presence. (2) Adaptation to the medium takes place via the use of typographical sideways symbols. (3) Time is an important factor in adaptation: with passing time, differences between groups converge.

We recorded data from 98 German male students who participated for 9 weeks in an off-campus online seminar on certain topics of General Psychology. Instruction took place via 6 virtual rooms (Web pages) on the Internet (library, virtual classroom etc.). The analyses of students's online activities and their communication style are based on a large amount of data: Altogether, students logged in 3608 times, read 1240 mails, and composed 160 mails. The communication style observed in the mails partly confirms hypotheses (1) and (2). We also noticed significant changes in the communication style with progressing time. The data of the investigated sample, however, could not fully support hypothesis (3). Here, further research seems to be necessary.

Distance education and communication: Social presence and nonverbal cues

As Benbunan-Fich and Hiltz (1999) point out, computer-mediated technologies play a more and more prominent role in education, a fact which is also reflected by a large and growing number of articles in the academic literature (e.g. Collis, 1996; Hiltz & Wellman, 1997; Spitzer, 1998; Stadtlander, 1998; Webster & Hackley, 1997). Mediated courses, however, are (like every educational situation) part of a communication situation, too. Communication in face-to-face courses relies to a large degree on spoken verbal information and communication. Such verbal communication is usually accompanied by numerous nonverbal and paralinguistic signs: e.g, communication partners maintain an individually and situationally appropriate body space and they employ mimicry and gesture as well as prosodic information to render additional information about the speech act (Austin, 1962; Searle, 1969; see also Clark & Brennan, 1996). Every face-to-face communication reveals something about the speaker's current condition, his or her emotional and cognitive state. These communicative means seem to determine the speaker's social presence, the degree to which a person's distinct characteristics and modes of expression are perceivable in a communication situation.

Theories about social presence and CMC

In the late 70s, theories of social presence were developed by the research group of Short, Williams, and Christie (1976). These researchers describe social presence as “the feeling that other actors are jointly involved in communicative interaction” (Short et al., 1976). On the one hand, they emphasize that perceiving other actors as persons is determined by the availability of visual, acoustic, and probably even haptic information. Consequently, social presence is characterized as an attribute of the communication medium which directly depends on the number of information channels made available by that medium.

On the other hand, Sproull and Kiesler (1989) accentuate the lack of cues referring to a person's social context, her or his personality, status, etc. The consequences are, for example, that to a certain degree group processes might be impeded. This might result in depersonalization and de-individuation problems as well as difficulties in turn-taking and disorganization (Kiesler, Siegel & McGuire, 1984). Hiemstra (1982, p. 883) concluded: “As bandwidth narrows from face-to face interaction, the communication is likely to be experienced as less friendly, emotional, and personal, and more serious, businesslike, depersonalized, and task-oriented.”

Spears and Lea (1992) combine these two views. In their SIDE model they take interpersonal cues as well as social information into account. They state that in some forms of CMC nonverbal and paralinguistic signs cannot be transmitted and hence important cues about the interpersonal identity are filtered out. Nevertheless, CMC users may know in which social context they communicate. Therefore, in CMC one may either feel like a member of the same group as one's communication partners or like an individual. Hence, social behavior can be classified into two important modes: acting as an individual and acting as a group member. Behavior in communication situations can be determined by both kinds of identity and which part dominates depends on the specific situation. Like other cues-filtered-out perspectives (see Culnan & Markus, 1987), the SIDE model assumes that an overbalance of personal identity and the simultaneous lack of nonverbal and paralinguistic cues leads to a higher task-orientation, informality, and a breach of manners.

Theories on affiliation and adaptation processes

Researchers like Walther (1992; see also Walther & Burgoon, 1992) investigate CMC from a different point of view. Walther and his co-workers emphasize the temporal development of forming impressions about the personality of communication partners. Communication partners are driven by a strong need for affiliation and for building up an impression of the other actors. Social information theory (Walther, Slovacek & Tidwell, 1999; Walther & Tidwell, 1995) acknowledges that in CMC the exchange of information is slower than in face-to-face interaction. Therefore, in comparison to face-to-face communication the construction of interpersonal knowledge is slower. In the beginning, CMC is more task-oriented than person-oriented and less powerful. In this phase, the communication style in computer-mediated situations is regarded as more informal but one is also likely to find cues for tension and extremely emotional evaluations of the communication partner or the communicational setting. When CMC participants receive enough opportunities to communicate the differences according to the categories listed above converge. In this context, Walther and Tidwell (1995) also emphasize the role of typographical sideways symbols (e.g. emoticons like;-) or :-( ) or so-called disclaimers (e.g. *lol*, *yuk yuk*, *rotfl*) as surrogates for nonverbal social cues (see also Carey, 1980; Sanderson, 1994; Turkle, 1995). The use of such metacommunicatively effective signs demonstrates that users adapt the possibilities a medium offers to the requirements of face-to-face communication situations.

Clark and Brennan (1996) assume a similar process of adaptation in CMC. They focus on the co-ordination of every behavior in which at least two people are involved. If one regards communication as just another form of behavior (see also Herrmann, 1983) coordination might be identified as a crucial aspect. Thus, coordination in communication means updating one's common ground. The process of grounding diverges for different situations and therefore varies with the medium used in a certain communication situation. Clark and Brennan (1996) enumerate eight constraints on grounding which are associated to different media: copresence, visibility, audibility, cotemporality, simultaneity, sequentiality, reviewability, and revisability.

Predictions on communicativel behavior in distance education

Up to now, most studies on social presence (Short et al., 1976; see also Boos, Jonas, & Sassenberg, 2000; Kiesler et al., 1984; Sproull & Kiesler 1986; Walter, 2000) or other theoretical concepts within the cues-filtered-out perspective investigated communication in organizational structures. We would expect that communication in educational situations also depends on a number of nonverbal personal and social cues. Therefore, we designed an off-campus online seminar in which the communication situation between the tutor and tutees was varied with regard to the mediated information. We contrasted four degrees of mediated interpersonal cues which we call four degrees of social presence.

Our assumptions can be characterized by three hypotheses.

Hypothesis 1: We assume that tutees who form a perception of the tutor by written text only (i.e. without any nonverbal clues) (condition 1) will react with more task-orientation, informality, and greater tension than those tutees who form a perception of the tutor by spoken language and a meaningful personal view (condition 4). Furthermore, in the reduced communication situation (condition 1) students should evaluate the tutor and the setting extremely emotionally.

Contrary to this notion, adaptation theories and the social information processing perspective (Walther, 1992) predict that with ongoing time communication processes are adjusted to the conditions imposed by the respective medium. From this assumption two hypotheses can be drawn:

Hypothesis 2: The tutees in condition (1) adapt their communication behavior to a stronger degree than the tutees in condition (4). Hence, this group should use typographical sideways symbols more often than group 4.

Hypothesis 3: Furthermore, we expect that with ongoing time and consequently with more opportunities for communication group differences in task orientation, informal language, and so on will disappear.

To testify hypotheses (2) and (3) we conducted an online seminar as a longitudinal study for a period of nine weeks.

An off-campus online seminar

Research design and independent variables

We designed and conducted an off-campus online seminar at the University of the Federal Armed Forces, Munich. For nine weeks, i.e., one term, our students participated in four different groups (conditions 1 to 4). Each group obtained admittance to the seminar via a web page (figure 1) and the intranet of the university. Students were instructed via six rooms, respectively, six different communication services (see also figure 1):

  • A virtual classroom, in which we presented the learning material in the form of a computer-based training. The material was realized as a tutorial with a description of learning objectives and other forms of learning guidance. For example, we offered an overview over the course structure, we tried to stimulate and to motivate our students through special instructional messages, we provided feedback and guided their attention. These instructional events (Gagné and Briggs, 1979; see also Aronson & Briggs, 1983) were provided with the course material.  (Figure 2 shows one of those instructional events.) The learning material consisted of four learning units explaining certain topics of cognitive psychology, namely knowledge structures, learning with text, learning with pictures, and problem solving. Students could either download the material or receive it off-line.
  • A virtual library offered additional learning material, e.g. texts, short computer-based training material, or software referring to the learning contents.
  • A notice board contained news and general issues concerning all students.
  • A chat room offered the opportunity for the students to participate in discussions.
  • A consultation room, where the tutor gave personal feedback and answered questions about the learning material.
  • An examination room, where the students participated in online tests.

As mentioned above we offered admittance to these services via the intranet of the university and a web server (Sambar 4.0). After having called the web address of the server, a web page of a central building of the University of the Federal Armed Forces in Munich (the background in figure 1) with links to the services illustrated beforehand appeared. With respect to a room metaphor we offered the various services in different windows of the building. The services themselves as well as the texts and questionnaires in the different locations had been programmed with HTML. The students' answers and the dynamic appearance of the pages were managed by CGI-protocols in the programming language PERL. The learning material in the virtual classroom had been programmed with the authorware TOOLBOOK 4.0 (CBT-edition).

At the beginning of the study, the virtual tutor for the students was a female with a professional and personal background unknown to them. Therefore, in terms of Sproull and Kiesler (1989) all participants received numerous hints for the social context of the tutor. Yet, participants' opportunities to form a perception of the tutor's appearance varied according to the above enumerated conditions (see table 1). The independent variable “communication of the tutor” was varied in the mails the tutor sent to the students via the consultation room, in the news at the notice board, and in the instructional events which were part of the learning material in the virtual classroom. That is, whenever the tutor commented on the learning material, gave a personal feedback, or responded to an inquiry she was presented by an appropriate personal view (condition 2 and condition 4) and/or an audio-file of the text (condition 3 and condition 4). Twenty-four different personal views were created as .jpg files and classified according to six different instructional situations: explaining, showing empathy, being interested, gaining attention, providing feedback (postive or negative) (Figure 3 shows six examples of the categories). For the same instructional event, the same personal views and the same audio-files were assigned to the respective conditions. (Participants in condition 4 received the same personal view as participants in condition 2 and the same audio-file as participants in condition 3). The tutor's answers to students' e-mails were also classified according to the six instructional events and corresponding personal views were assigned.

Table 1.  Four degrees of social presence.
Condition 1:a tutor mediated through verbal, written information
Condition 2:a tutor mediated through written information and various personal views
Condition 3:a tutor mediated through written and spoken information
Condition 4:a tutor mediated through text, views and spoken language

Our second independent variable was the duration of the seminar. Every second week students were to finish one learning unit of the computer-based training material. After completion of the learning material they had to complete tests on the learning material and evaluations of the tutor and the course. Within one week they received feedback on their tests. Hence, data were obtained over the course of nine weeks at four points in time. (One additional assessment was required at the end of the course.)

Dependent variables

In this paper our analyses will be focused on students' logins and their communication style during the different periods of time plus a final evaluation of the online seminar. (For performance measurements gathered at the online testing and the evaluations of the tutor and the setting see Paechter, Schweizer, & Weidenmann, in press.)

To investigate the assumed differences in students' communication behavior (hypothesis 1 and 3) we developed categories according to Walther and Burgoon (1992; see above). Our classification scheme comprised the attributes “task orientation vs. personal orientation”, “formal vs. informal communication style”, “tension and extremely emotional evaluations of the communication partner or the communicational setting”. Table 2 illustrates the classification with some examples.

Table 2.  Communication categories.
Categories for evaluating the communication styleExamples in students' mails to the tutor
Task orientation vs. personal orientation (ranging from 0 to 1)Task orientation (-1): “When will the first test be available?” Personal orientation (1): “I find this topic really interesting…”
Formal vs. informal communication style (ranging from 0 to 1)Formal style (1): saluations like “Dear tutor…” Informal style (-1): missing salutation, missing signatures
Tension (ranging from 0 to 1)Tension (1): usage of insulting utterances or codes expressing negative attitudes like capital letters “…much too RIGID programming.”
Extremely emotional evaluations of the communication partner (tutor) (ranging from extremely negative, -1, to extremely positive, 1)Extreme behavior: utterances with extreme evaluations of the tutor, her person, or her professional competence
Extremely emotional evaluations of the communication setting (ranging from extremely negative, -1, to extremely positive, 1)Extremely negative (-1): “…that we receive such a poor learning setting…” Extremely positive (1): “Everything looks terrific, the system works without faults.”

In addition to the communication categories investigated by a content analysis we also looked for expressive components and disclaimers (hypothesis 2).

Sample

In total, 101 German male students participated for nine weeks. They were all in their second term at the University of the Federal Armed Forces, Munich studying different subjects (engineering, educational sciences, computing etc.). Most participants were between 19 and 24 years old. Students in each course were randomly assigned to one of the four conditions so that the participants in every condition were comparable with respect to their study subjects.

At the end of the seminar our participants had to pass a written exam on the learning material which they needed for further advancement in their studies. In the first term all students had received a thorough introduction to Internet services such as chats, newsgroups, e-mail etc. and to the learning environment employed.

We also tried to control for the comparability of our students' technical equipment and their communicative interaction. Since nearly all of the participants lived on the campus we only had to exclude three students from the online seminar due to a different technical equipment. To prevent an interaction between students outside of the virtual classroom we asked them to sign a contract about the conditions of the investigation.

Results

First we analysed the frequencies of the logins.1 During the 9 weeks of virtual teaching our students visited the online seminar 3608 times (see figure 4). There were 727 contacts resulting from logins in the examination room, 515 from logins in the library, and 966 from logins in the notice board. The e-mails (an important basis for the derivation of our dependent variables) were received and sent via the consultation room. This communication service was utilized 1400 times. Of this number 1140 contacts were used for reading the tutor's e-mails and 160 logins were used to send e-mails to the tutor. The tutor herself composed 570 e-mails to the students. According to our experimental conditions, approximately 50% of the messages had to be illustrated with appropriate personal views, half of them with photographs of the tutor and recorded speech, and another 25 % of the messages with speech only.

The 570 e-mails to the students consisted of 404 feedback responses to the results of the online testing and of 166 responses to the 160 e-mails of the tutees. That means that every student got 4 feedback responses. (The distribution of the resisting e-mails will be illustrated with hypothesis 1.)

At first, we analyzed the logins with respect to our experimental conditions and tested the influence of our first independent variable, the social presence of the tutor mediated through communication richness. We tried to control for divergent frequencies of use and therefore different grades of adaptation. As expected we did not find any differences between groups.

Then we looked at the final evaluations of our online seminar. One of our questions of interest was how students would describe the communicative situation when asked to identify advantages and disadvantages of the course in their own words. Over all experimental conditions 68.46% of our participants mentioned the lack of communication to others (tutor or fellow students) during the virtual seminar as a clear obstacle, 11.2% mentioned one's anonymity, and 22.5% mentioned the high demands on technical resources. The students, however, listed also clear advantages of virtual seminars: 71.4% of our students appreciated the self-organization of one's own time schedule and 37.8% the possibility to determine one's own speed of learning.

Hypothesis 1

The e-mails we received consisted of a total of 13,293 words (M = 86.88, SD = 75.92). One hundred and ten of the messages referred to events in a certain phase of the virtual seminar. They were sent by 54 students. This means that every second student did not feel stimulated to write at all to the tutor. On an average, the “communicative” students mailed (and therefore also received responses from the tutor) 2.83 times (SD = 2.61). With the exception of the feedback e-mails on their test results, the “non-communicative” students did not receive e-mails from the tutor. The number of students who sent e-mails did not differ according to the experienced social presence.

We analyzed the students' e-mails with respect to the communication categories enumerated above (see table 2). The students' statements were coded by two raters with an interrater reliability of R = 1.0. In general, we found the following percentage of utterances in each category (see table 3):

Table 3.  Percentage of utterances according to communication categories.
Task orientation vs. personal orientationFormal vs. informal communication styleTensionExtremely emotional evaluations of the communication partner (tutor)Extremely emotional evaluations of the communication setting
task orientation: 53.6% neutral: 29.1% personal orientation: 17.2%formal: 19.6% neutral: 22.9% informal: 57.5%no tension: 84.3% extreme tension: 15.7%extremely positive: 0.7% neutral: 96.1% extremely negative: 3.3%extremely positive: 4.6% neutral: 73.2% extremely negative: 22.2%

The number of students who sent mails at all did not differ according to the experienced social presence. We, did, however, find differences according to the experienced social presence when analyzing the length of the mails (number of words) and the expressive components and disclaimers. Participants who experienced the tutor only through written information or written information and personal views used more typographical sideways symbols (see also table 1). Fewer words were written when the participants encountered the tutor without personal views. Both differences were significant (χ2= 16.89, df = 3, p < .001 and χ2= 13.33, df = 3, p < .005). Obviously, there is a distinguishable but not consistent difference between the varying degrees of social presence.

Table 1.  Number of words per e-mail and number of e-mails with typographical sideways symbols, by social presence condition.
 Social Presence
 Written informationWritten information and personal viewsWritten and spoken informationWritten and spoken information and personal views
Number of words3050 (M=84.72, SD=73.41)4405 (M=112.95, SD=111.44)1796 (M=57.94, SD=44.19)4042 (M=86.00, SD=47.63)
Number of e-mails with typographical sideways symbols11903

According to hypothesis (1) we found differences with regard to, first, the formality of the communication style and, second, extremely emotional evaluations of the communication setting. Participants who perceived the tutor with the highest degree of social presence (condition 4) communicated less informally than all other participants (H = 16.37, df = 3, p < .005). Figure 5 illustrates the average of categorized messages per degree of social presence on a scale ranging from -1 (informal) to 1 (formal).

These findings are consistent with the evaluation of the communication setting. Participants with less intensive experiences with the tutor's social presence (condition 1) evaluated the setting more emotionally and extremely. The difference lacks significance but can be regarded as a tendency (H = 6.21, df = 3, p = .102; see also figure 6).

At this point we would like to give two examples (translated messages) of divergent classified messages. (We rated message no. 140 with an +1 and message no. 251 with an -1 in both categories.)

Hypothesis 2

Our second hypothesis states that the use of typographical sideways symbols should occur more frequently in condition (1) (reduced social presence) than in condition (4) (highest degree of social presence). Therefore, we analyzed the number of e-mails with expressive components and disclaimers. As we had assumed, participants who experienced the tutor only through written information or written information and personal views used more typographical sideways symbols (see also table 4). This difference is significant (H = 16.89, df = 3, p < .001). Obviously, there is a distinguishable difference between the varying degrees of social presence. We regard this finding as a support for adaptation processes.

Table 4.  Number of e-mails with typographical sideways symbols according to the varied social presence.
 Social Presence
 Written informationWritten information and personal viewsWritten and spoken informationWritten and spoken information and personal views
Number of e-mails with typographical sideways symbols11903

Hypothesis 3

Finally, we analyzed the e-mails according to the supposed adaptation with respect to ongoing time. To test our assumption we computed Friedman tests for the above mentioned categories (see table 2 and 3) per point in time. Again, we obtained significant findings with regard to the number of informal communication utterances for each experimental group (with respect to the tutor's social presence) over time (textual information only: Q = 42.534, df = 3, p < .001; textual information and personal views: Q = 47.862, df = 3, p < .001; textual and spoken information: Q = 50.127, df = 3, p < .001; textual and spoken information and personal views: Q = 51.592, df = 3, p < .001; see also figure 7).

The significance of these differences, however, seems to be due to an increase of informal utterances during the second period of time followed by a decrease during the third and the fourth section.

Discussion

In an experimental study on distance education and communication we tested the following three assumptions: (1) To experience the tutor with less social presence leads to extremely emotional evaluations as well as more task oriented, informal, and tense reactions, compared to conditions in which the tutor can be experienced with greater social presence. (2) Adaptation to the medium is instantiated in the use of typographical sideways symbols. (3) Ongoing time also plays a role in adaptation: with ongoing time differences between groups converge. The data we recorded confirm the first hypothesis with respect to the ratio of informal messages for the tutor and extreme evaluations of the communication setting. In both categories the social presence of the tutor varied through the presence or absence of certain nonverbal cues led to a significant difference. Nevertheless, it should be emphasized that in general German students tend to behave much more formally than American students and students of the University of the Federal Armed Forces, all pursuing a military career, should behave even more formally and stick to a higher degree to conventional behavior. Certainly, the percentage of informal messages (57.5%) is rather high and the second example (message no. 251) illustrates a breach of manners. We already emphasized that the cues-filtered-out perspectives primarily focused on organizational structures. Our results accentuate that the reduction of a person's social presence (realized by the reduction of nonverbal cues in communication settings, is reflected in communication behavior in educational situations as well. This assumption is also supported by students' views on disadvantages of online courses, especially the missed opportunities for communication, anonymity, and the high demands on resources. Our findings underline that the form of communication between the tutor and his or her students is a crucial and important aspect in virtual seminars. This holds true even though supplementary (written) social information about the tutor was given to the tutees.

With hypotheses (2) and (3) we analyzed whether adaptation to the reduced situation takes place or not. However, we cannot answer this question conclusively. On the one hand, the use of typographical sideways symbols supports the notion of adaptation processes. Students who experienced the tutor without nonverbal cues like prosodic information also employed more typographical sideways symbols than the others. On the other hand, there is no clear evidence that the number of informal utterances or other communication clues usually indicating restrictions on social presence declines with ongoing time. In fact, we find an increased number of informal statements at the second point in time. At later points in time, however, the number of statements decreases to the level at the beginning of the course. This means that at least with respect to formality the communication style changes with ongoing time. A possible explanation for this result might be that our students did not receive sufficient opportunities to communicate with the tutor which means that possibly social situations demanding communication with the tutor were missing. As stated above only every second student sent e-mails to the tutor. Students were not obliged to communicate with the tutor. Therefore, it might be possible that the “non-communicative” students also adapted their behavior, only in another way.

Another concern refers to the overall length of the study and the temporal characteristics. Our results seem to suggest that we think of adaptation as a process fluctuating up and down during a certain phase of time. We think that further research should be addressed to this observation.

The results of the present study provide useful hints for the design of virtual seminars, yet the study might be limited in its generalizability. Since our sample consisted of German male students (officers of the Federal Armed Forces) it is not clear whether other demographic configurations (in other countries, with different age populations, etc.) would have provided other results.

Footnotes

  • 1

    In this paper we only report the results with respect to the login frequencies, the final evaluation of the seminar, and communication clues. For the performance results and the evaluation of the seminar please see Paechter, Weidenmann, and Schweizer (in press).

Ancillary