26
The relationship between social presence and online privacy Chih-Hsiung Tu * Educational Technology Leadership, Department of Educational Leadership, Graduate School of Education and Human Development, 2134 G. St., NW, Washington, DC 20052, USA Received 3 May 2002; received in revised form 19 August 2002 Abstract Online privacy may critically impact social presence in an online learning environment. This study examined how online privacy affects social presence in online learning environments and whether e- mail, bulletin board, and real-time discussion affect online privacy. Mixed methods were used to examine the relationship between social presence and privacy. The participants rated computer- mediated communication (CMC) with a high degree of social presence, but the quantitative correlation between social presence and privacy failed to reach significance. Participants shared personal information on CMC knowing that it was risky because the medium lacked security despite the perceived high levels of social presence. This contradictional phenomenon can be explained as ‘‘risk- taking’’ behavior. Among three CMC systems, e-mail was ranked as the most private and followed by one-to-one real-time discussion, then many-to-many real-time discussion. Bulletin board was considered to afford the least privacy. D 2002 Elsevier Science Inc. All rights reserved. Keywords: Online learning; Computer-mediated communication; Social presence; Online privacy; Social environment 1096-7516/02/$ – see front matter D 2002 Elsevier Science Inc. All rights reserved. PII:S1096-7516(02)00134-3 * Tel.: +1-202-994-2676; fax: +1-202-994-2145. E-mail address: [email protected] (C.-H. Tu). Internet and Higher Education 5 (2002) 293 – 318

The relationship between social presence and online privacy

  • Upload
    nau

  • View
    0

  • Download
    0

Embed Size (px)

Citation preview

The relationship between social presence and online privacy

Chih-Hsiung Tu*

Educational Technology Leadership, Department of Educational Leadership,

Graduate School of Education and Human Development, 2134 G. St., NW, Washington, DC 20052, USA

Received 3 May 2002; received in revised form 19 August 2002

Abstract

Online privacy may critically impact social presence in an online learning environment. This study

examined how online privacy affects social presence in online learning environments and whether e-

mail, bulletin board, and real-time discussion affect online privacy. Mixed methods were used to

examine the relationship between social presence and privacy. The participants rated computer-

mediated communication (CMC) with a high degree of social presence, but the quantitative correlation

between social presence and privacy failed to reach significance. Participants shared personal

information on CMC knowing that it was risky because the medium lacked security despite the

perceived high levels of social presence. This contradictional phenomenon can be explained as ‘‘risk-

taking’’ behavior. Among three CMC systems, e-mail was ranked as the most private and followed by

one-to-one real-time discussion, then many-to-many real-time discussion. Bulletin board was

considered to afford the least privacy.

D 2002 Elsevier Science Inc. All rights reserved.

Keywords: Online learning; Computer-mediated communication; Social presence; Online privacy; Social

environment

1096-7516/02/$ – see front matter D 2002 Elsevier Science Inc. All rights reserved.

PII: S1096 -7516 (02 )00134 -3

* Tel.: +1-202-994-2676; fax: +1-202-994-2145.

E-mail address: [email protected] (C.-H. Tu).

Internet and Higher Education

5 (2002) 293–318

1. Introduction

Social presence is the degree of person-to-person awareness, which occurs in a mediated

environment.Multiple publications (McBride&Bazley, 1997; Rice, 1993; Spears&Lea, 1992;

Towell & Towell, 1997) have emphasized that social presence is an important construct for

future study. Recent studies have shown that social presence impacts online learners’ online

interaction and learning (Gunawardena & Zittle, 1997; Tu & McIsaac, 2002). Short, Williams,

and Christie (1976) indicate that social presence is the most important perception that occurs in

social context and is an important key to understanding person-to-person telecommunication.

Later, Rice (1993) concluded that social presence appears to provide a useful, consistent,

meaningful, stable, and discriminating way to characterize computer-mediated communication

(CMC). Gunawardena (1995) argued that social presence is necessary to enhance and improve

effective instruction in online learning environments. When the level of social presence is low,

interaction is also low (Garramone, Harris, &Anderson, 1986). In fact, a lack of social presence

may lead to a high level of frustration, a critical attitude toward the instructor’s effectiveness,

and a lower level of affective learning (Rifkind, 1992).

When one perceives an online learning environment to be less private, or they are unable to

maintain their privacy online, they would naturally be less interactive in his/her learning

process. In other words, privacy may be a critical factor influencing the level of social

presence. Therefore, a need exists to examine the relationship between social presence and

privacy in an online learning environment.

1.1. Social presence

Social presence is defined as the degree of awareness of another person in an interaction

and the consequent appreciation of an interpersonal relationship (Rice, 1993; Walther, 1992;

Walther & Burgoon, 1992). Biocca (1997) declared that, ‘‘The amount of social presence is

the degree to which a user feels access to the intelligence, intentions, and sensory impressions

of another.’’ Tu and McIsaac (2002) redefine it for online learning environments: ‘‘Social

presence is the degree of feeling, perception, and reaction of being connected on CMC to

another intellectual entity.’’ Factors that contribute to an appreciable degree of online social

presence are social relationship, trust, learners’ characteristics, learners’ perceptions on online

environments, attributes of communication media, learners’ computer literacy, use of

emoticons and paralanguage, communication styles, task types, privacy, etc. Several studies

(Leh, 2001; Polhemus, Shih, & Swan, 2001; Swan, Polhemus, & Shih, 2002) have shown

that perception of the degree of social presence in an interaction will vary among users.

Therefore, social presence should be viewed as a subjective quality that depends upon the

objective quality of the medium (Walther, 1992).

1.2. Social presence concepts in social psychology

Two concepts in social psychology grounded in face-to-face settings are related to

social presence are intimacy (Argyle & Dean, 1965) and immediacy (Wiener &

C.-H. Tu / Internet and Higher Education 5 (2002) 293–318294

Mehrabian, 1968). Traditionally, these two concepts are difficult convey in an online

learning environment.

Intimacy is a function of eye contact, physical proximity, topic of conversation, etc.

Changes in one will produce compensatory changes in the others (Short et al., 1976). A

communication with maintained eye contact, close proximity, body leaning forward, and

smiling conveys intimacy (Burgoon, Buller, Hale, & deTurck, 1984).

Equilibrium theory (Short et al., 1976) explains how humans balance the degree of intimacy.

In a face-to-face setting, people tend to avoid maintained eye contact and they increase physical

distance if personal topics or topics with which a person is uncomfortable are to be discussed.

People try to maintain an optimum level of intimacy. That is, when an uncomfortable degree of

intimacy is encountered, the participants of the conversation will attempt to alter their behavior

to maintain the degree of intimacy at an optimal comfort level. The interaction is unpleasant if

behavior cannot be altered to allow an optimal degree of intimacy.

Equilibrium theory is applied very differently in online situations. In online environments,

learners are authorized to master their online communications to determine when they would

like to communicate and about what via certain ideal CMC forms. The attributes of CMC

allow learners to manipulate the level of intimacy as they wish, a phenomenon that may lead

learners to react to online intimacy with extreme behaviors, embarrassing, flaming, dropout,

or lurking (eavesdropping in silence) etc.

The second psychological concept involved in social presence is immediacy. This is the

psychological distance a communicator places between himself or herself and the recipient of

the communication. It includes eye contact, smiling, vocal expressiveness, physical prox-

imity, appropriate touching, leaning toward a person, gesturing, using overall body move-

ments, being relaxed, and spending time with someone.

Online immediacy becomes very difficult to deliver because CMC lacks social cues and

nonverbal cues, however, this does not negate online immediacy or its importance. Human

are social beings and immediacy is still necessary for social contact among online learners. In

fact, immediacy in online learning environments is even more critical to affect interaction

than in face-to-face learning environments.

1.3. Degree of social presence

Social presence is a dynamic variable. The degree of social presence is based upon the

characteristics of the medium and the user’s perception. People discern different amounts of

social presence in various types of media. Goffman (1959) contended that humans construct

their self-presentations and carry them off in front of others either intentionally or

unintentionally. Hence, social presence is the internal image the perceiver evokes of a

moving, expressive body. Normally, the users are asked to assess the degree of social

presence. Short et al. (1976) measured social presence through the semantic differential

technique with a series of bipolar scales with a seven-point assessment, sociable/unsociable,

personal/impersonal, sensitive/insensitive, and warm/cold. A higher level of presence in a

medium confers the attributes of being more sociable, more personal, more sensitive, and

warmer.

C.-H. Tu / Internet and Higher Education 5 (2002) 293–318 295

Online learners can cultivate social presence by facilitating introductions of the CMC

communicators to each other in the initial learning sessions (Johansen, Vallee, & Spangler,

1988). This allows the participants the opportunity to become better acquainted and to

develop a trust relationship early in the course. Also, it can be used by the leader to encourage

participation by all of the members of the discussion. Gunawardena’s (1995) study suggested

the student’s perception of social presence is impacted by the instructor’s skilled use of

interaction techniques in initiating online conversations with introductions and salutations

that will impact. Consequently, instructors, or moderators, should develop interaction skills

that create a sense of social presence.

1.4. Three dimensions of social presence

Tu and McIsaac (2002) define social presence into three dimensions: social context, online

communication, and interactivity. These three dimensions function as a theoretical framework

for research in online social presence.

1.4.1. Social context

Social context is constructed from the CMC users’ characteristics and their perception of

the CMC environment. Social contexts, such as task orientation (Steinfield, 1986), users’

characteristics and perception on online environment (Steinfield, 1986), recipients/social

relationships (Walther, 1992; Williams & Rice, 1983), trust, availability of CMC, CMC

access locations, and social process (Walther, 1992) etc., contribute to the degree of social

presence. If the participants are unfamiliar with each other and the conversation is task

oriented and more public the degree of social presence will degrade. Walther (1992) proposed

that different social processes, settings, and purposes are components of social context and

affect social presence.

1.4.2. Online communication

Online communication is concerned with the attributes of the language used online and the

applications of online language, such as attributes of CMC, computer literacy skills, online

immediacy, and online language skills. Because of the technology and its text-based format,

CMC requires that users possess some level of computer communication literacy such as

typing, reading, and writing. People without these skills develop communication anxiety

(Gunawardena, 1991) when text-based communication is required. Therefore, it is suggested

that text-based communications should be initiated with some light or casual topics, like

introductions. Training students to use the medium and making them comfortable using it is

crucial to the success of collaborative learning. Garramone et al. (1986) and Perse, Burton,

Kovner, Lears, and Sen (1992) examined students’ perceptions of social presence and

concluded that the degree of social presence on computer bulletin boards was perceived as

higher for users who were more interactive than for those who were not. Perse et al. (1992)

found a positive relationship between social presence and the student’s perception of their

own computer expertise. Paralanguage and emoticons used to compensate lacks of social/

nonverbal cues have impacts on social presence. Tu (2001) concluded that students who

C.-H. Tu / Internet and Higher Education 5 (2002) 293–318296

appreciated uses of paralanguage and emoticons perceive higher level of social presence,

even they do not utilize them.

1.4.3. Interactivity

Interactivity includes the activities in which CMC users engage and the communication

styles they use, such as responsive time, communication styles, task types, topics (Argyle &

Dean, 1965; Walther, 1992), and size of groups. The potential for feedback from the other

also contributes to the degree of salience of the other person in the interaction. Immediate

response is another component of interactivity. In asynchronous CMC response occurs at a

different time, so it takes longer to obtain a response from the other party. When an immediate

response is expected and is not received a feeling of low interactivity is created, thus,

diminishing the level of social presence. However, Garramone et al. (1986) found that

interactivity, allowing for feedback, contributes to the social presence of an electronic bulletin

board. Gunawardena (1995) differentiates interactivity and social presence, arguing that

social presence requires user’s to add one more step to awareness of interactivity; in short,

when users notice and appreciate it, there is social presence. In other words, interactivity is

the design and strategy to stimulate social presence. When learners appreciate it, ideal social

presence is perceived and has a great deal potential to generate interactive learning.

1.5. Privacy

Privacy affects the degree of social presence. Research has shown that privacy has an

impact on human interaction in media-based communications (Champness, 1973; Christie &

Holloway, 1975; Steinfield, 1986; Tu, 2000; Weisband & Reinig, 1995).

If a medium is perceived more public, a sense of less privacy will be generated and vice

versa. Therefore, the level of privacy is determined by the users’ perceptions in addition to the

actual quality of security. Witmer (1997) identified two factors, feeling of privacy and system

privacy, which affect level of privacy. Researchers have been examining the causes of

perceived difference in levels of privacy. Illusion of privacy (Neumann, 1995), self-awareness

(Archer, Hormuth, & Berg, 1982), nonchalant attitude (Friedman, 1990; Tu, 2001), social

norms (Markus, 1994), self-disclosure (Weisband & Reinig, 1995), and risk-taking behaviors

(Tu & McIsaac, 2002; Witmer, 1997) have been applied to explain online users’ different

levels of perceived privacy.

Weisband and Reinig (1995) examined how e-mail affects online users’ privacy and

concluded three majors factors: (a) technology and users’ knowledge and experience; (b)

management policies in an organizational context; and (c) the psychological effects of e-mail

that encourage self-disclosure, development of interpersonal ties, and new norms of social

behavior. These three factors correspond to feeling of privacy, system privacy, and social norms.

1.6. Feeling of privacy

Feeling of privacy refers to online users’ perception of privacy psychologically, mentally,

culturally, or conditionally rather than the actual security. Generally speaking, online users

C.-H. Tu / Internet and Higher Education 5 (2002) 293–318 297

perceive different communication media with different levels of privacy in different circum-

stances. A less private setting results in a decreased perception of social presence by users. In

a videoconferencing environment, Champness (1972) reported that users felt more public and

perceived less social presence. By and large, a camera in operation may be seen as intrusive.

Attitudes toward the use of TV in public broadcasting may carry over into the laboratory

causing concerns of electronic eavesdropping and producing negative reactions (Ryan, 1976).

Steinfield (1986) examined the social presence of e-mail in organizational settings, and

reported that users were reluctant to employ e-mail for confidential matters.

Feeling of privacy may be an unstable and dynamic factor because it is subjective and

socially constructed in online communication messages. Weisband and Reinig (1995) argued

that online communications could result in two different areas, individual and organization

levels. However, Finholt and Sproull (1990) concluded that it was difficult to distinguish

work-related message from nonwork-related ones because users mixed subjects during

communications. Although these two studies resulted in different arguments, both studies

agreed with that CMC allows online users to modify the social constraints by reducing the

social context cues, such as information about social hierarchy, social differences, social

relationships, and personal meaning and implications of interaction (Sproull & Kiesler, 1986).

Online users tend to lose the sense of self-awareness and perceive they are invisible and

anonymous because of the absence of social context cues and visualization; therefore, they

felt free to express themselves and allow self-disclosure.

1.7. System privacy

System privacy refers to the actual security of CMC technologies and considers the

likelihood that someone may read, send, or resend a message to or from you. Some users are

not familiar with CMC system security and think that CMC technologies are private because

in some cases user’s name and password are required to conduct online communication, such

as e-mail communications. In Kerr and Hiltz’s (1982) study, it was found that more than a

third of the online users agreed with the statements that ‘‘information can come into the

wrong hands’’ and ‘‘outsiders can see private messages.’’ Generally, online users who have

better knowledge of computer systems and CMC technologies will perceive low privacy

because of the systems not being secure.

It appears that online privacy cannot be predicted because of the many complicated human

psychological and social behaviors that are difficult to explain. A few activities have been

examined by researchers and are discussed below to facilitate a better understanding of the

relationship between online users’ behaviors and privacy.

1.8. Illusion of privacy

Convenience may override privacy risk because users are unable to visualize the negative

impacts although less private environments may decrease user’s tendency to online inter-

action. This is called Illusion of Privacy (Neumann, 1995). Oftentimes less communication

cues are presented; therefore, users lose the sense of who else is in the virtual environment

C.-H. Tu / Internet and Higher Education 5 (2002) 293–318298

and the size of the audience, such as lurkers (people who only observe activities and never

participate in them). Users feel psychologically secure in their communication. It is difficult

for some users to consider negative consequences that they can’t actually see in their minds,

and the sharing of information about them is hard to visualize. Therefore, people do not ask

themselves what negative things could happen to them as a result of people possessing

knowledge about them. Consequently, the illusion created is that users think their commu-

nications are much more private than they really are. In other words, certain users think that

they are invisible if they participate in the online activities. Online users continue to say

things that they wouldn’t say in regular communication settings, even though they are warned

(Denning & Lin, 1994).

1.9. Social norms

CMC technologies change rapidly; therefore, few social norms are speculated for different

circumstances and environments. CMC applications are influenced by social norms in

organizations (Markus, 1994) and the adoption of CMC policies (Weisband & Reinig,

1995). If organizations fail to articulate organizational CMC policy to employee, it may lead

CMC users to believe that they are allowed to express whatever they want to. Additionally,

CMC users observe online communication of what seems to be private disclosure of others or

the free expression permitted online, they perceive that such ‘‘open’’ communication is the

norm, not the exception. Adopting appropriate social norms is necessary and perhaps as users

become more acquainted with the computer systems and CMC technologies, social norms

will stabilize and management policies will be more effective.

1.10. Self-disclosure

Some users express personal and sensitive information online because it is easier to

disclose when no one is present to respond to the communication. Archer et al. (1982)

concluded that it is more unpleasant for subjects disclosing in the presence of a large

mirror than disclosing without a mirror in a psychological experimental study. This

experimental result can be explained as reduced self-awareness online due to the absence

of social context cues; as a result, users feel safe to express private matters and personal

information. In other words, other users have potential opportunities to alter/foster self-

disclosure benefits such as through esteem support, informational support, instrumental

support, and motivational support (Derlega, Metts, Petronio, Hendrick, & Margulis,

1993).

1.11. Nonchalant attitude

Some educational online users express nonchalant attitudes towards online privacy

because they think that all class-related communications are class work, nothing personal

(Friedman, 1990) or confidential. Online students take for granted that no one would be

interested in their personal information, even though they do deliver it online. In fact, by

C.-H. Tu / Internet and Higher Education 5 (2002) 293–318 299

collaborating and examining each other’s work, some instructors thought that students

learned from each other, teacher actively structured students’ computer use so that files were

largely public (Friedman, 1990), such as applications of online student publishing and peer

evaluation (Tu & Blocher, 2000).

1.12. Risk-taking behavior

Online users possess conflicting perceptions of online privacy. This phenomenon is

explained as risk-taking behavior (Witmer, 1997). It was found that online users feel

personally and technically secure in CMC, and felt that they had little or nothing to lose if

their activities were discovered by unintended others. This, then, indicates that the perceived

risk is low among users who engage in risky CMC behaviors in these newsgroups. In fact,

some users may think, ‘‘it won’t be me if it does occur.’’

1.13. More private, the better?

Is a more private online environment always better? It seems that more private envir-

onment is always the desired level of social presence for a learning environment. However,

some instructional strategies may disagree with it. Online student publishing and peer

evaluation that situate students in more public and less private situation have positive

impacts on students’ online learning (Tu & Blocher, 2000). In fact, in an extremely private

online learning environment, there may be little interaction because of the lack of social

interaction between learners and between teachers and students. In other words, the degree of

online privacy is a dynamic variable; online learners may feel and need various levels of

online privacy under different circumstances.

2. Research on impacts of social presence

Recent studies have found that social relationship (Tu, 2002a), task types (Tu, 2002a),

attributes of CMC, and confidence, choice, and involvement (Blocher, 1997) have impacts on

the degree of social presence while social presence has impacts on online interaction (Tu &

McIsaac, 2002), user satisfaction (Gunawardena & Zittle, 1997), depth of online discussions

(Polhemus et al., 2001), online language learning (Leh, 2001), critical thinking (Tu & Corry,

2002), and Chinese students’ online learning interaction (Tu, 2001).

Different social relationships and task types demonstrated both positive and negative

impacts on the levels of social presence. Love, information, familiarity, and trust social

relationships exert a positive impact on social presence while service, status, assertive/

acquiescent, and conflict relationships exert a negative impact (Tu, 2002a). Task types,

generate, choose, and social tasks appear to exert a positive impact on social presence while

negotiating/conflict tasks exert a negative impact (Tu, 2002a).

Online learners perceive different levels of social presence in various CMC systems, e-

mail, bulletin board, and real-time discussion. Tu (in press) concluded that students perceived

C.-H. Tu / Internet and Higher Education 5 (2002) 293–318300

one-to-one e-mail with the highest social presence followed by one-to-one real-time

discussion, one-to-many e-mail, many-to-many real-time discussion, and one-to-many

bulletin board.

From a learner-centered aspect, Blocher (1997) concluded that confidence, choice, and

involvement had impacts on the levels of social presence. When learners feel more confident,

are able to make learning choices, and actively are engaged in learning activities, it

demonstrated higher level of social presence.

While researchers have identified different various variables that affected the level of social

presence, they also examined how social presence affect various online learning effects, such

as online interaction, satisfactions, language learning, critical thinking and Chinese online

interaction, etc.

Tu and McIsaac (2002) studied the relations of social presence and online interaction

and concluded that three dimensions of social presence (social context, online commun-

ication, and interactivity) have impacts on online interaction. Additionally, it was

concluded that social presence had influences on Chinese students’ online interaction

(Tu, 2001).

Social presence is a significant predictor of the user’s satisfaction of CMC (Gunawardena

& Zittle, 1997). CMC provides the students with the ability to use ‘‘emoticons’’ to create

socioemotional experiences although some learners do not use them. It is suggested that the

teacher/moderator must create a sense of social presence, which greatly impacts the user’s

satisfaction of the medium used in the classroom.

Polhemus et al. (2001) studies the relations of social presence and complexity of online

discussions and the findings revealed that postings with a high degree of social presence

were likely to initiate more complex discussions than postings with a low degree of social

presence. To foster high-level social presence to promote complexity of online discussion,

researchers suggest that online instructors should develop a rapport with learners by

creating a trustworthy community-like environment with use of high amount of affective

languages. These strategies will encourage and motivate learners to model similar online

communication.

Social presence additionally was found as a critical factor that affects online language

learning. Leh (2001) examined Spanish learning via pen-pal design with native Spanish

speaking learners. She concluded that higher social presence would enhance students to learn

Spanish via various CMC forms.

In preliminary analysis, Tu and Corry (2002) found that there was a relation between social

presence and critical thinking. Three dimensions of social presence (Tu & McIsaac, 2002)

related to five factors of critical thinking (Bullen, 1998). However, the exact relations were

not reported in this study. Further examinations are indicated to examine what the relations

between social presence and critical thinking are and whether social presence is a critical

factor affecting learners’ critical thinking.

Social presence has impacts on non-English speaking learners in English speaking

classroom. Chinese students responded that social presence has impacts on their online

interaction; even they were unable to gather nonverbal cues, like in traditional commun-

ication (Tu, 2001).

C.-H. Tu / Internet and Higher Education 5 (2002) 293–318 301

2.1. Research questions

By examining the learner’s perception of social presence and privacy in three CMC

systems, e-mail, bulletin board, and real-time discussion, the following questions were asked:

1. What are the relations between social presence and online privacy?

2. Do e-mail, bulletin board, and real-time discussion affect online privacy?

3. Method

Both quantitative and qualitative methods were used to gain a better understanding of the

relations of between social presence, privacy, and text-based CMC (e-mail, bulletin boards,

and real-time discussion). Fifty-one students enrolled in a graduate level course at a 4-year

university in Southwest of the US were the subjects. The course was offered in two formats,

one was televised and the other was face-to-face. Both classes were taught by the same

instructor with exactly the same course content, lectures, assignments, and class require-

ments.

3.1. Qualitative method

Participant observation was used to understand the issues social presence, privacy in three

CMC forms from the student’s point of view. FirstClass, a computer conferencing system,

was used for class communication among the instructor, teaching assistants, and students. It

provides e-mail, bulletin board, and real-time discussion functions. Each student was assigned

a username and a password to access FirstClass. Pseudo identity was not allowed.

Participant observation was the primary procedure used to capture the students’ percep-

tions of social presence and privacy and as a way to determine how they make sense out of

the class activities in which they participate via different CMC systems.

The data were collected through casual conversation, in-depth interview, direct observation,

and document analysis. The casual conversation was conducted between the researcher and the

subjects in different settings, the researcher’s office, the classroom, or any convenient location.

The questions were casual, free flowing, and unencumbered by preconceptions of how the

topics ‘‘should’’ be discussed. Observations were conducted in the classroom, the computer

laboratory, and through online asynchronous and synchronous class discussions.

At the 12th week of the semester, eight semistructured in-depth interviews were conducted

with participants to explore particular concepts in social presence and privacy to collect

elaborate and comprehensive details.

Document analysis included all messages delivered on FirstClass and outside e-mail

received by the instructor and the teaching assistant. The analysis began after some of the

data was acquired which gave the researcher a better idea of where to focus further data

collection. This helped to develop interview questions and to decide which students to

interview.

C.-H. Tu / Internet and Higher Education 5 (2002) 293–318302

3.2. Quantitative method

Quantitative methods were used to examine the relationship between social presence and

online privacy and impacts of three CMC systems on privacy.

At the 12th week of the semester, 51 participants were asked to answer the CMC

Questionnaire (Tu, 2002b). This questionnaire, evaluating e-mail, bulletin board, and real-

time discussion, contains 17 social presence items and 13 privacy items each with a five-point

Likert scale, and 12 demographic identities. Participation in this survey was voluntary. Forty-

three responses were returned with 84.31% return rate.

Bartlett’s test of sphericity (Bartlett, 1950) was applied to increase the validity because of

the small number of participants. This tested whether the correlation was statistically different

from zero by comparing the correlation matrix (R) and identity matrix (I). If R 6¼ I, the

correlation was significant then factor analysis could follow (Pedhazur & Schmelkin, 1991).

The power of the Bartlett’s test of sphericity is that it is sensitive to the sample size (Knapp &

Swoyer, 1967). Therefore, if the zero correlation is rejected by a small sample, there is greater

validity.

Exploratory factor analysis was utilized to examine the dimensions of social presence and

privacy. Items producing eigenvalues greater than 1.00 were considered to be the significant

factors. Scree testing with visual inspections was used to support and determine the number

of factors/clusters to be extracted. Exploratory factor analysis was utilized. Humphrey-llgen

Parallel Analysis was applied, two data matrices are analyzed simultaneously, and their

Eigenvalues were plotted. Additionally, Cattell’s scree test and Kaiser’s criterion were used to

determine the number of factors to extract. Pearson correlation was computed to explain the

relationship of privacy and the social presence.

3.3. Triangulation

Triangulation was not used as a validation process in this study. Rather, it was considered a

means to achieve a better understanding about the students’ perceptions of social presence.

The rationale for using different methods of triangulation is that the flaws of one method are

often the strengths of another; by combining methods observers can achieve the best of each

while overcoming their unique deficiencies (Denzin, 1989). It was not expected that the

findings generated by different methods would fall into a coherent picture (Lincoln & Guba,

1985; Patton, 1980).

4. Results

Forty-three of 51 subjects responded to the online questionnaire. More than half of the

subjects 28 (65.12%) were female, 15 (34.88%) males comprised the remainder of the study

group. The Ethnic mix was composed of 31 Caucasian (72.09%), 4 Latinos (9.30%), 4

African–Americans (9.30%), and 4 Asians and Pacific Islanders (9.30%). Subjects estimated

their computer expertise as intermediate (29, 67.44%), novice (9, 20.93%), and expert (5,

C.-H. Tu / Internet and Higher Education 5 (2002) 293–318 303

11.63%). In other words, more than half of subjects felt that their computer expertise was

intermediate or higher. Of the 43 subjects, most accessed computers at home (40, 93.02%),

followed by computer laboratories (30, 69.77%), classrooms (19, 44.19%), office (18,

41.86%), and library or media center (14, 32.56%).

Subjects had been using e-mail longer than bulletin board and real-time discussion.

Slightly less than 75% of the students had been using e-mail from 1 to 6 years, while more

than a half of them had less than 1-year experience in bulletin board and real-time

discussions. Subjects’ e-mail experiences were 1–2 (27.91%), 5–6 (25.58%), and 3–4 years

(20.93%), respectively. More than the half of students had been using bulletin board (55.81%)

and real-time discussion (60.47%) less than 1 year. The majority of the subjects had less than

1 year’s experience using the bulletin board (55.81%) followed by 1–2 (20.93%), 3–4

(11.63%), 5–6 (4.65%), and 7–10 years (6.98%), respectively; while experience in real-time

discussion was, predominately, less than 1 year (60.47%) followed by 1–2 (25.58%), 3–4,

5–6, and 7–10 years (4.65%), respectively. This group of subjects had more experience with

e-mail than with bulletin board and real-time discussion; but their experience on bulletin

board and real-time discussion was equivalent.

4.1. Quantitative results

Because of the small number of participants (n = 43), it was necessary to conduct Bartlett’s

test of sphericity to examine the validity of the results, c2 = 774.90 with df= 44. The

hypothesis that the correlation matrix was an identity matrix was rejected at the .01 of a level.

The correlation matrix produced a significant chi-square by this test; therefore, factor analysis

proceeded.

An exploratory factor analysis was performed on 30 questionnaire items concerning social

presence (social context, online communication, and interactivity) and computer privacy

(system privacy and perception of privacy). These five factors accounted for 76.74% of the

variance. The five factors were extracted using varimax rotation. With a cutoff of .45, three

items were removed from the loading, item numbers 15, 17, and 29, respectively. These five

factors were social context, online communication, interactivity, system privacy, and feeling

of privacy. Cronbach’s coefficient a for these five factors were .82, .88, .73, .76, and .71,

respectively.

The five variables that loaded on the first factor, social context, were: CMC as a social

form, CMC as an informal and casual way to communicate, CMC is personal, CMC as a

sensitive means of communication, and comfort with familiar persons. The second factor,

system privacy, included system operator and someone may repost messages sent to or from

you, someone may accidentally send and receive messages to and from individuals other than

the intended recipients, CMC as technically reliable, possibility of embarrassment, and

identity concerns. The third factor, interactivity, consisted of five variables: CMC as pleasant,

immediate, responsive, and comfortable with familiar topics. The fourth factor, online

communication, included four variables: CMC conveys feeling and emotion, CMC as

stimulating, expressive, meaningful, and easily understood. The fifth factor, feeling of

privacy, included feeling of confidentiality, feeling of privacy, perception of privacy,

C.-H. Tu / Internet and Higher Education 5 (2002) 293–318304

importance of privacy, level of security/secret, risk of sharing personal topics. The items of

user relationship, and being unfamiliar with persons and topics did not load on any of the

factors (smaller than .45).

4.1.1. Relationship of social presence and privacy

Correlations between social presence (social context, online communication, and inter-

activity) and privacy (system privacy and perception of privacy) were computed, and the

result was insignificant with r=.286. In this study group, there was no significant relationship

between social presence and privacy.

4.1.2. Differences of privacy in CMC systems

One-way repeated measures, ANOVA, were computed for three CMC systems, e-mail,

bulletin board, and real-time discussion on the level of privacy. The result indicated a

significant difference in the level of privacy among these three CMC systems, Wilks’

l= 0.67, F(2,41) = 10.32, P < .05 (see Table 1). E-mail received the highest rate on the level

of social presence (M = 3.15, S.D. = 0.58), followed by the real-time discussion (M = 3.13,

S.D. = 0.49) and bulletin board (M = 2.97, S.D. = 0.48). E-mail was perceived as more private

medium, while bulletin board was perceived as less private.

Because the ANOVA overall test yielded a significant result, three pairwise comparisons

among e-mail, bulletin board, and real-time discussion were conducted to assess which means

differed from each other (see Table 2). Two of the three pairwise comparisons were

Table 1

Multivariate tests among three CMC systems for privacy factora

CMC effect Value F df Error df Significance h2 Noncent.

parameter

Observed

powerb

Pillai’s trace 0.34 10.32c 2 41 .00 .34 20.63 0.98

Wilks’ l 0.67 10.32c 2 41 .00 .34 20.63 0.98

Hotelling’s trace 0.50 10.32c 2 41 .00 .34 20.63 0.98

Roy’s largest root 0.50 10.32c 2 41 .00 .34 20.63 0.98

Within subjects design: privacy factor.a Design: intercept.b Computed using a=.05.c Exact statistic.

Table 2

Paired samples test among three CMC systems for privacy factor

Paired differences t df Significance

M S.D. S.E.M. 95% confidence interval of the difference

Lower Upper

E-mail–board 0.17 0.37 0.06 0.06 0.29 3.01 42 .00 *

E-mail–discussion 0.01 0.46 0.07 � 0.13 0.16 0.212 42 .83

Board–discussion � 0.16 0.28 0.04 � 0.24 � 0.07 � 3.61 42 .00 *

* P < .05, two-tailed.

C.-H. Tu / Internet and Higher Education 5 (2002) 293–318 305

significant controlling for familywise error rate across the three tests at the .05 level, using the

Holm’s sequential Bonferroni procedure.

The smallest P-value was for the comparison of bulletin board and real-time discussion,

and its value of .001 was less than the P-value of .05/3=.0167 and, therefore, the differences

between the means for these two systems were significant. The next smallest P-value is for

the comparison of e-mail and bulletin board, and its value of .004 was less than the P-value of

.05/2=.025 and, therefore, the differences between the means for these two systems were

significant. The last comparison, between e-mail and real-time discussion, was not significant

at .83.

4.1.3. System privacy and three CMC systems

One-way repeated measures, ANOVA, were computed for three CMC systems, e-mail,

bulletin board, and real-time discussion on the level of system privacy. The result indicated a

significant difference in the level of system privacy among these three CMC systems, Wilks’

l = 0.82, F(2,41) = 4.44, P< .05 (see Table 3). Real-time discussion received the highest rate on

the level of system privacy (M = 3.30, S.D. = 0.66), followed by the bulletin board (M = 3.14,

S.D. = 0.67) and e-mail (M = 3.13, S.D. = 0.70). Real-time discussion was perceived as a

medium with more system privacy, while e-mail was perceived as having less system privacy.

Because the ANOVA overall test yielded a significant result, three pairwise comparisons

among e-mail, bulletin board, and real-time discussion were conducted to assess, which

means differed from each other (see Table 4). Two of the three pairwise comparisons were

Table 3

Multivariate tests among three CMC systems for system privacya

CMC effect Value F df Error df Significance h2 Noncent.

parameter

Observed

powerb

Pillai’s trace 0.18 4.44c 2 41 .00 .02 8.89 0.73

Wilks’ l 0.82 4.44c 2 41 .00 .02 8.89 0.73

Hotelling’s trace 0.22 4.44c 2 41 .00 .02 8.89 0.73

Roy’s largest root 0.22 4.44c 2 41 .00 .02 8.89 0.73

Within subjects design: system privacy.a Design: intercept.b Computed using a=.05.c Exact statistic.

Table 4

Paired samples test among three CMC systems for system privacy

Paired differences t df Significance

M S.D. S.E.M. 95% confidence interval of the difference

Lower Upper

E-mail–board 0.01 0.34 0.05 � 0.11 0.10 � 0.17 42 .87

E-mail–discussion � 0.17 0.45 0.07 � 0.31 � 0.03 � 0.24 42 .02 *

Board–discussion � 0.16 0.36 0.05 � 0.27 � 0.05 � 2.95 42 .01 *

* P < .05, two-tailed.

C.-H. Tu / Internet and Higher Education 5 (2002) 293–318306

significant controlling for familywise error rate across the three tests at the .05 level, using the

Holm’s sequential Bonferroni procedure.

The smallest P-value was for the comparison of bulletin board and real-time discussion, and

its value of .005 was less than the P-value of .05/3=.0167 and, therefore, the differences bet-

ween the means for these two systems were significant. The next smallest P-value was for the

comparison of e-mail and real-time discussion, and its value of .019 was less than the P-value of

.05/2=.025 and, therefore, the differences between the means for these two systems were

significant. The last comparison, between e-mail and bulletin board, was not significant at .87.

4.1.4. Feeling of privacy and three CMC systems

One-way repeated measures, ANOVA, were computed for three CMC systems, e-mail,

bulletin board, and real-time discussion on the level of feeling of privacy. The result indicated

a significant difference in the level of feeling of privacy among these three CMC systems,

Wilks’ l= 0.48, F(2,41) = 22.09, P < .05 (see Table 5). E-mail received the highest rate on the

level of feeling of privacy (M = 3.06, S.D. = 0.68), followed by the real-time discussion

(M = 2.46, S.D. = 0.58) and bulletin board (M = 2.23, S.D. = 0.59). E-mail was perceived as a

medium with a greater feeling of privacy while bulletin board perceived as a medium with

less feeling of privacy.

Because the ANOVA overall test yielded a significant result, three pairwise comparisons

among e-mail, bulletin board, and real-time discussion were conducted to assess which means

differed from each other (see Table 6). All three pairwise comparisons were significant

Table 5

Multivariate tests among three CMC feeling of privacya

CMC effect Value F df Error df Significance h2 Noncent.

parameter

Observed

powerb

Pillai’s trace 0.52 22.09c 2 41 .00 .52 44.17 1.00

Wilks’ l 0.48 22.09c 2 41 .00 .52 44.17 1.00

Hotelling’s trace 1.08 22.09c 2 41 .00 .52 44.17 1.00

Roy’s largest root 1.08 22.09c 2 41 .00 .52 44.17 1.00

Within subjects design: privacy factor.a Design: intercept.b Computed using a=.05.c Exact statistic.

Table 6

Paired samples test among three CMC systems for feeling of privacy

Paired differences t df Significance

M S.D. S.E.M. 95% confidence interval of the difference

Lower Upper

E-mail–board 0.82 0.80 0.12 0.58 1.07 6.73 42 .00 *

E-mail–discussion 0.60 0.69 0.10 0.38 0.81 5.68 42 .00 *

Board–discussion � 0.23 0.42 0.06 � 0.36 � 0.10 � 3.56 42 .00 *

* P< .05, two-tailed.

C.-H. Tu / Internet and Higher Education 5 (2002) 293–318 307

controlling for familywise error rate across the three tests at the .05 level, using the Holm’s

sequential Bonferroni procedure.

The smallest P-value was for the comparison of e-mail-bulletin board, and e-mail-real-time

discussion, and its value of .000 were less than the P-value of .05/3=.0167 and, therefore, the

differences between the means for these two systems were significant. The next smallest P-

value was for the comparison of bulletin board and real-time discussion, and its value of .001

was less than the P-value of .05/3=.017 and, therefore, the differences between the means for

these two systems were significant.

4.2. Qualitative results

The results of qualitative explained that why the correlation between social presence and

privacy was insignificant and refined the how different CMC forms influencing online

privacy.

4.2.1. Privacy

The privacy factor included system privacy and feeling of privacy. Students’ responses to

CMC privacy were inconsistent and were subject to individual differences. Many students

were not familiar with the preservation of privacy in the CMC environment. Some students

thought that there was no difference in privacy among three CMC systems, while others felt

there were differences. It appeared that e-mail is the most private and followed by one-to-one

real-time discussion, then many-to-many real-time discussion, and the bulletin board was

considered to be the least private.

4.2.2. System privacy

System privacy referred to the security of CMC technologies regarding the likelihood that

someone may read, send, or resend a message to or from you. Students were aware that CMC

systems are not private. However, some were more cautious than the others. Some students

were aware of it but they thought that they were ‘‘fine’’ with the degree of privacy. Salena, a

Hispanic female, expressed her feeling about the privacy: ‘‘E-mail. I am pretty sure. . . I knowthat it is not very secure. But, I don’t think about that. I think it’s fine.’’ Students experienced

less concern about CMC privacy on FirstClass because a password was required for access. In

fact, one student used someone else’s e-mail account to send the TA an e-mail message.

Obviously, students were not that much concerned about privacy. Cheng, female Chinese

student, expressed about her feeling about privacy: ‘‘I don’t have problems with the privacy

issue at all. I believe that message I sent only goes to the person I intend to.’’

Students were concerned that the messages they posted might appear in public despite being

posted in a more private format. They felt that the recipients could very easily print the

messages, pass them around, or repost them to more public areas. Judy expressed her concerns:

It doesn’t matter how secure the lines are you should never say anything that could be

you know printed out by someone and then misinterpret it, pass around. . . and. . . I didn’tsee any difference between e-mail, real-time discussion, and bulletin board.

C.-H. Tu / Internet and Higher Education 5 (2002) 293–318308

4.2.3. Feeling of privacy

Feeling of privacy referred to student’s perception of the extent of privacy on CMC media.

When students perceived a medium to be more private, they would express more openly.

Adversely, when students perceived a medium to be more public, they would be more

cautious in expressing their feelings.

4.2.4. Public feeling

Bulletin board was considered by students to be a less private (more public) medium, even

more public than a many-to-many real-time discussion. Students thought the bulletin board

messages were permanent after they were posted, although FirstClass allowed users to

remove their own bulletin board messages. Additionally, accessing the messages was easy for

someone who had access to the system. Sven, a Swedish male in his early thirties, expressed

his feeling about the bulletin board: ‘‘No I think the bulletin board is even less secure because

the bulletin board if I post something today it may stay for like a week or a month or until

whoever decides to take it down. It may be there for a long while.’’

Messages on the bulletin board were very task-oriented and less personal information was

exchanged because it was perceived as being more public. Salena described her feeling about

the bulletin board:

Because everybody is going to see (the bulletin board messages) I don’t talk something

more personal. . .

Students responded that when many people are involved in the real-time discussions (more

than two persons in a real-time discussion) they were more public. Some students were aware

that real-time discussion messages could be permanent and less private. Some students felt

that the real-time discussion messages were deleted after the real-time discussion unless they

were saved.

When students developed the sense that the medium was more public, the message was

less personal. Therefore, if a personal message appeared in a more public arena, it could be

extremely embarrassing and created discomfort for either the sender or the recipient. In other

words, learners are unable to modify their online intimacy to comfortable level. A dispute in

one of the groups in the study resulted in a letter of apology, which was addressed

personally to the other party involved and copied to the other group members and to group

bulletin board. This could create extreme embarrassment and pressure for this particular

student. One team member posted her dissatisfaction with team project on the class bulletin

board in another instance. Another team member felt that it should not appear on a public

bulletin board because it was a more private message. In this instance, that message caused

others to feel uncomfortable. Ling, a female Chinese student, described her feeling about the

message:

She said she has difficulties to get in touch with us. She was not very happy. She

complained that the subject was not familiar with her. So she just does it perfunctory.

The group forming is not formed in organizationally. I feel very strange!! She posted

this message on the bulletin board.

C.-H. Tu / Internet and Higher Education 5 (2002) 293–318 309

The locations where students access CMC exert major influences on private feeling. When

the location is more public, students felt less private, even though some computer laboratories

were designed to enhance privacy. Several students, whose only access to CMC was in

computer laboratory, expressed major concerns about their privacy in a public computer

workstation. However, the observation of students in the classroom laboratory revealed that

they were not much concerned about privacy when they were engaged in CMC conversations

despite the fact that its design allowed very limited privacy. The layout was in straight rows,

allowing an easy view of the computer monitor at the station in front of any particular station.

One student felt that no one would be interested in her CMCmessages since she was not a well-

know public figure. She did not acknowledge privacy as a problem in public CMC laboratory.

4.2.5. Private feeling

Students felt that a message was more private when it was posted to a one-to-one format,

i.e., one-to-one e-mail or one-to-one real-time discussion. It was more comfortable for

students to share personal information in a one-to-one communication because they believed

that no one would read their messages, or that no one would be interested in their personal

conversation. One-to-one e-mail and real-time discussion are perceived as the most private

within all three CMC systems.

It can be embarrassing if a more personal message appears in public area. However, this

depends upon the contents of the personal messages. If the contents are more positive, such as

an expression of admiration, the subject of the commentary may feel comfortable having it

posted in a more public area.

4.2.6. Importance of privacy

Students reported that online privacy was important to them; but were unconcerned about

this issue in conversations about their class because they felt that no one would be interested

in their messages. Even so, it would take someone’s time to break into the system to retrieve

the messages. Risk-taking clearly emerged in the online environment. Cheng expressed her

feeling about privacy: ‘‘I think that no one is interested in my personal e-mail.’’ Sven had

similar thought:

But you feel more secure sending an e-mail because it takes an effort to find your e-

mail. . . who’s going to care about my e-mail to classmates. Nobody!! I mean who’s

going to care to dig that up.

Students perceived that the privacy of each CMC system was different. Therefore, students

viewed the different CMC systems as providing different solutions for divergent needs with

different levels of privacy.

5. Discussion

In this study, the results revealed that the correlation between social presence and privacy

was insignificant although participants perceived text-based CMC systems with high social

C.-H. Tu / Internet and Higher Education 5 (2002) 293–318310

presence. In factor analysis, system privacy and feeling of privacy emerged. Based on

qualitative data analysis, three CMC systems appeared as not the appropriate way to measure

social presence and privacy because the uses of different CMC systems are more critical than

types of CMC. It is suggested that it should be measuring as one-to-one e-mail, one-to-one

real-time discussion, one-to-many bulletin board, and many-to-many real-time discussion

formats rather than forms of e-mail, bulletin board, and real-time discussions.

5.1. Relationship of social presence and privacy

Social presence and privacy factor do not correlate clearly in the findings. The correlation

results between social presence and privacy showed insignificance. This result matched the

qualitative results that students perceived that the level of social presence was inconsistent

with the perceived level of online privacy. In the qualitative data analysis, students were

aware of the insecurity of online communication, and students showed their concerns about

online privacy in both the quantitative data analysis. Although students perceived CMC

systems with high levels of social presence, they responded that it was ‘‘risky’’ to share

personal information on CMC. The qualitative data analysis from the interviews, and review

of the online messages did not agree with the observations at the computer laboratory and the

classroom laboratory and the online postings. Students appeared very open in online learning

environment although they addressed their concerns on privacy and security. This reaction

can be explained as ‘‘risk-taking’’ behavior (Witmer, 1997).

Participants’ perceptions on privacy are not stable. Online learners perceive the different

levels of privacy in different online situations. It is very hard to determine when online users

would perceive what degree of privacy. It is clear that, in some situations, levels of social

presence have impacts on privacy.

5.2. Privacy

System privacy and a perception of privacy emerged as major variables from evaluation of

data from both methods. System privacy included a system operator and someone who may

repost messages sent to or from you, someone who may accidentally send and receive

messages to and from individuals other than the intended recipients, CMC as technically

reliable, possibility of embarrassment, and identity concerns. The perception of privacy

included importance of privacy, level of security/secret, risk of sharing personal topics.

Quantitative data analysis provided what was associated with the privacy factors, while

qualitative data analysis addressed how students perceived online privacy issues.

Among three CMC systems, e-mail was ranked as the most private and followed by one-

to-one real-time discussion, then many-to-many real-time discussion, and bulletin board was

considered to be the least private. These findings were supported by both quantitative and

qualitative data analysis. In quantitative data analysis, it appeared that there were significant

differences between e-mail and bulletin, and bulletin board and real-time discussion in the

level of privacy. There was no significant difference between e-mail and real-time discussion.

This can be explained by the fact that bulletin board was considered to be more public

C.-H. Tu / Internet and Higher Education 5 (2002) 293–318 311

because more people have access to it, unlike e-mail and real-time discussion which seemed

to occur in a more private setting, especially in one-to-one e-mail and in one-to-one real-time

discussion.

The qualitative data revealed that it was necessary to divide e-mail and real-time

discussions into two different categories, one-to-one e-mail and one-to-many e-mail; and

one-to-one discussion and many-to-many discussion, because the style of the discussions

imposed a very different level of privacy. It was concluded that one-to-one discussion was a

more personal form of communication, like e-mail; therefore, it was more private. Many-to-

many discussions have a more public access; therefore, it was perceived as less private. This

conclusion was not shared by the quantitative method because the different real-time

discussions were not separated.

5.3. System privacy

System privacy referred to the security of CMC technologies regarding the likelihood that

someone may read, send, or resend a message to or from you. Most students thought that no

privacy existed in computer systems. They reported that CMC was not secure because they

have been told that it was not secure. Students believed that a system administrator or

someone might break into the system and post/resend messages from/to you. But, on the other

hand, the students responded that no one would be interested in their messages because it was

just class work. It was difficult to determine if students really knew about the privacy and

security of CMC systems from this study. Therefore, students’ perceptions on the security of

CMC appeared to be a critical issue and an important area for investigation.

Real-time discussion was perceived as possessing the highest degree of system privacy

while e-mail had the least. A one-way ANOVA indicated that there were significant differences

among the three CMC systems. The differences between each two systems were examined and

demonstrated significant differences between e-mail and real-time discussion, and bulletin

board and real-time discussion. In other words, e-mail and bulletin board appeared to have a

similar level of system privacy. The explanation was that real-time discussion occurred all at

one time and all of the messages disappeared when the discussion ended, unlike bulletin board

and e-mail messages that were more permanent. Also, all of the participant’s names were listed

when they entered the discussion room and which made it easier to determine who was present

at any given time. Therefore, students perceived that it was less likely for anyone to resend/

repost messages from/to someone during a real-time discussion.

5.4. Feeling of privacy

Feeling of privacy referred to the student’s perception of how private they felt while using

the CMC media. When students perceived the medium to be more private, they would

express themselves more openly. The opposite was true when the medium was perceived as

more public.

E-mail was rated as the most private medium and the bulletin board was the least private.

This result was the same as overall privacy. The one-way ANOVA indicated that there were

C.-H. Tu / Internet and Higher Education 5 (2002) 293–318312

significant differences among the three CMC systems; each two systems were compared to

each other. It appeared that all three pairs of comparisons showed significant differences. In

other words, all three CMC systems demonstrated significant differences from each other in

the level of feeling of privacy.

5.5. One-to-many communication

Bulletin board, a one-to-many medium was seen as less private, more public, than many-

to-many real-time discussion and e-mail. Students were aware that the bulletin board

messages were permanent after being posted, although FirstClass allowed users to remove

their own bulletin board messages. Additionally, viewing the messages was easy for

someone who had access to the system. The messages on the bulletin board were very

task-oriented and there were less personal information shared because of the more public

nature.

5.6. Many-to-many communication

Students responded that many to many real-time discussions (more than two people in a

discussion) were more public. This was not examined in the quantitative data analysis

because the CMC questionnaire (Tu, 2002b) did not separate one-to-one real-time discussions

from many-to-many discussions. Some of the students were aware that discussion messages

could be permanent, creating less privacy. Some students felt that after the discussion, the

discussion messages would be deleted unless someone saved them. However, it was still

considered more private than bulletin board because messages were not necessarily

permanent.

5.7. One-to-one communication

It was more private when the conversations were conducted in a one-to-one communica-

tion medium, as one-to-one e-mail and one-to-one real-time discussion. People normally felt

more private in a one-to-one communication, because the conversation could only be

accessed by these two people, particularly in a one-to-one e-mail conversation. Therefore,

one-to-one e-mail communication was perceived as the most private medium. Although

students possessed a certain level of understanding about CMC, none of the students reported

that one-to-one e-mail conversations could be public. In fact, one-to-one e-mail messages

could be considered to be in a one-to-many form. In other words, e-mail can have multiple

recipients whose names may not appear on the e-mail messages. None of students have

reported this concern. In most e-mail applications, one is allowed to utilize ‘‘CC’’ and

‘‘BCC’’ functions to send e-mail to multiple recipients. If the message is sent with the CC.

format, the all recipients’ names or e-mail addresses will appear on the e-mail; while BCC

format does not show all of the recipients instead only one recipient is shown. The latter

format creates a more private feeling for that particular recipient. In fact, most recipients think

that the message is sent only to them and there were no other recipients.

C.-H. Tu / Internet and Higher Education 5 (2002) 293–318 313

E-mail messages with a long recipient list that preceded the body of the message were

perceived as less private; and, further, as impolite because e-mail was supposed to be a more

private communication medium. The long recipient list made the individual recipient feel less

important. Therefore, it is recommended that when one-to-many e-mail messages are used,

that the BCC function be applied to increase the recipient’s private and personal feeling.

5.8. Personal message in public

When students perceived a certain medium as more public, the messages tended to be less

personal. Therefore, if a personal message appears on a more public area, it would be

extremely embarrassing and made either the sender or the recipient embarrassed and

uncomfortable. However, a negative feeling was not always produced if a personal message

appeared in public. If the public, personal message was giving praise to someone, it ‘‘may’’

generate a more positive feeling for the recipient. Therefore, one must examine the context

and consider how the recipient will feel about the message if it appeared in public, then select

a more appropriate medium to deliver the personal message.

5.9. Locations

The location where students must access the CMC had major influences on their feelings

of privacy. When the location was more public, as in some computer laboratories, the students

perceived less privacy. The computer laboratories and classroom laboratories have very

different layout designs. Computer laboratories were designed to provide a more private

feeling because one does not easily see the other’s monitor. The class laboratories appear to

provide a more public feeling because the computers were laid out in rows. One can see the

other’s monitor in the front clearly. Analysis of the interview data and observations at both

computer laboratories and classroom laboratories are at odds. Students reported in their

interviews that there was less privacy in public areas, but students were observed conducting

e-mail in the classroom laboratory. This conflicting finding could be interpreted as ‘‘risk-

taking’’ behavior. In fact, students felt privacy on CMC was fairly important to them. Some

students felt that no one would be interested in their personal messages regarding class work,

so they were not concerned. This kind of human behavior exemplifies the expression; ‘‘it

won’t happen to me.’’ For example, everyone who drives an automobile is aware that if they

speed they may get caught and receive a traffic violation, but they still do it because, ‘‘It

won’t happen to me.’’ These were clear examples of ‘‘risk-taking’’ behavior.

5.10. Circulating messages

All CMC messages should be considered public (Witmer, 1997). Students’ familiarity to

public/private feeling varied. Students were aware that one might save the discussion

messages and send them to others or post it in public area. Some students do not consider

CMC as a private media, even one-to-one discussions and one-to-one e-mail because

messages may be passed around. In fact, CMC messages can be very easily saved and

C.-H. Tu / Internet and Higher Education 5 (2002) 293–318314

passed around or even posted in a public area. One does not have much control over this. It

seems students have developed a feeling that when one medium was more private others

possessed the same level of privacy. It is recommended that one should consider all CMC

messages as public.

6. Recommendations

Recommendations for practitioners and researchers are made based upon the findings of

this study. Although the privacy did not correlate to the level of social presence, one must

take privacy issues into account when they integrate CMC into their instruction. Fostering a

more private interaction environment is key to increasing interactivity, therefore, one must

understand the level of privacy provided by each of the CMC systems and the learner’s

perceptions of the privacy provided by each. It is clear that online environment that one

designs should be sensitive to various learning situation and students’ perception on privacy.

A sound learning environment will allow learners to adjust to the ideal levels of privacy and

give students more secure and more comfortable environments to increase their social

presence to enhance online interaction.

The format of CMC systems, e-mail and real-time discussion should be examined in two

different formats when evaluation privacy: one-to-one e-mail, one-to-many e-mail, one-to-

one real-time discussion, and many-to-many real-time discussion. Listservs should be

included in the future study. Though e-mail, listserv, and bulletin board systems may be

asynchronous forms of communication, they are quite different in their operation and

function.

7. Conclusion

Privacy issues appear as important factors in examining social presence although it does

not clearly correlate with the degree of social presence. ‘‘Risk-taking’’ is the explanation for

the absence of a correlation in this particular situation. Although a relationship between

privacy factor and social presence was not demonstrated, it must not be omitted when social

presence in an online environment is examined. A sound online learning environment should

allow learners to adjust their ideal levels of privacy to increase their social presence and

enhance online interaction. Privacy is a dynamic factor. To foster an interactive online

learning environment, it is not an issue of maximizing or minimizing the level of privacy. It

lies on whether a designed online learning environment is able to accommodate learners’

perceptions of privacy.

References

Archer, R. L., Hormuth, S. E., & Berg, J. H. (1982). Avoidance of self-disclosure: an experiment under conditions

of self-awareness. Personality and Social Psychology Bulletin, 8(1), 122–128.

C.-H. Tu / Internet and Higher Education 5 (2002) 293–318 315

Argyle, M., & Dean, J. (1965). Eye contact, distance and affiliation. Sociometry, 28(3), 289–304.

Bartlett, M. S. (1950). Tests of significance in factor analysis. British Journal of Psychology, Statistical Section,

3(2), 77–85.

Biocca, F. (1997). Cyborg’s dilemma: embodiment in virtual environments. Journal of Computer-Mediated Com-

munication, 3(2) (Retrieved November 11, 1998, from http://www.ascusc.org/jcmc/vol3/issue2/biocca2.html).

Blocher, J. M. (1997). Self-regulation of strategies and motivation to enhance interaction and social presence in

computer mediated communication. Unpublished doctoral dissertation. Tempe, AZ: Arizona State University.

Bullen, M. (1998). Participation and critical thinking in online university distance education. Journal of Distance

Education, 13(2) (Retrieved June 1, 2001, from http://cade.athabascau.ca/vol13.2/bullen.html).

Burgoon, J. K., Buller, D. B., Hale, J. L., & deTurck, M. (1984). Relational messages associated with nonverbal

behaviors. Human Communication Research, 10(3), 351–378.

Champness, B. G. (1972). Attitudes towards person–person communications media. University College, London:

Unpublished Communications Studies Group Paper No. E/72011/CH.

Champness, B. G. (1973). Attitudes toward person–person communications media. Human Factors, 15, 437–448.

Christie, B., & Holloway, S. (1975). Factors affecting the use of telecommunications by management. Journal of

Occupation Psychology, 48, 3–9.

Denning, D., & Lin, H. S. (1994). Rights and responsibilities in networked communities. Washington, DC:

National Academy Press.

Denzin, N. K. (1989). The research act: a theoretical introduction to sociological methods (3rd ed.). Englewood

Cliffs, NJ: Prentice Hall.

Derlega, V. J., Metts, S., Petronio, S., Hendrick, C., & Margulis, S. T. (1993). Self-disclosure. Newbury Park, CA:

Sage.

Finholt, T., & Sproull, L. S. (1990). Electronic groups at work. Organization Science, 1(1), 41–64.

Friedman, B. (1990). Societal issues and school practices: an ethnographic investigation of the social context of

school computer use. Paper presented at the Annual Meeting of the American Educational Research Association.

Garramone, G. M., Harris, A. C., & Anderson, R. (1986). Uses of political computer bulletin boards. Journal of

Broadcasting and Electronic Media, 30(3), 325–339.

Goffman, E. (1959). The presentation of self everyday life. New York: Doubleday.

Gunawardena, C. N. (1991). Collaborative learning and group dynamics in computer-mediated communication

networks. Research Monograph No. 9 of the Second American Symposium on Research in Distance Education

(pp. 14–24). University Park, PA: Pennsylvania State University, American Center for the Study of Distance

Education.

Gunawardena, C. N. (1995). Social presence theory and implications for interaction and collaborative learning in

computer conferences. International Journal of Educational Telecommunications, 1(2/3), 147–166.

Gunawardena, C. N., & Zittle, F. J. (1997). Social presence as a predictor of satisfaction within a computer-

mediated conferencing environment. American Journal of Distance Education, 11(3), 8–26.

Johansen, R., Vallee, J., & Spangler, K. (1988). Teleconferencing: electronic group communication. In R. S.

Cathcart, & L. A. Samovar (Eds.), Small group communication: a reader (5th ed.) (pp. 140–154). Menlo Park,

CA: Institute for the Future.

Kerr, E. B., & Hiltz, S. R. (1982). Computer-mediated communication systems: status and evaluation. New York:

Academic Press.

Knapp, T. R., & Swoyer, V. H. (1967). Some empirical results concerning the power of Bartlett’s test of the

significance of a correlation matrix. American Educational Research Journal, 4(1), 13–17.

Leh, A. S. C. (2001). Computer-mediated communication and social presence in a distance learning environment.

International Journal of Educational Telecommunications, 7(2), 109–128.

Lincoln, Y. S., & Guba, E. G. (1985). Naturalistic inquiry. Beverly Hills, CA: Sage.

Markus, M. L. (1994). Finding a happy medium: explaining the negative effects of electronic communication on

social life at work. ACM Transactions on Information Systems, 12(2), 119–149.

McBride, N. K., & Bazley, M. (1997). Threads of conversation: the life of a public e-mail conference. Paper

presented at the 5th European Conference on Information Systems.

C.-H. Tu / Internet and Higher Education 5 (2002) 293–318316

Neumann, P. (1995). Computer-related risks. New York: ACM Press.

Patton, M. Q. (1980). Qualitative evaluation methods. Beverly Hills, CA: Sage.

Pedhazur, E. J., & Schmelkin, L. (1991). Measurement, design, and analysis: an integrated approach. Hillsdale,

NJ: Lawrence Erlbaum Associates.

Perse, E. I., Burton, P., Kovner, E., Lears, M. E., & Sen, R. J. (1992). Predicting computer-mediated communi-

cation in a college class. Communication Research Reports, 9(2), 161–170.

Polhemus, L., Shih, L. F., & Swan, K. (2001). Virtual interactivity: the representation of social presence in

an online discussion. Paper presented at the Annual Conference of American Educational Research

Association.

Rice, R. E. (1993). Media appropriateness: using social presence theory to compare traditional and new organ-

ization media. Human Communication Research, 19(4), 451–484.

Rifkind, L. J. (1992). Immediacy as a predictor of teacher effectiveness in the instructional television. Journal of

Interactive Television, 1(1), 31–38.

Ryan, M. G. (1976). The influence of teleconferencing medium and status on participants’ perception of the

aestheticism, evaluation, privacy, potency, and activity of the medium. Human Communication Research, 2(3),

255–261.

Short, J. A., Williams, E., & Christie, B. (1976). The social psychology of telecommunications. London: Wiley.

Spears, R., & Lea, M. (1992). Social influence and the influence of the ‘social’ in computer-mediated commu-

nication. In M. Lea (Ed.), Contexts of computer-mediated communication (pp. 30–65). New York: Harvester

Wheatsheaf.

Sproull, L. S., & Kiesler, S. (1986). Reducing social context cues: electronic mail in organizational communica-

tion. Management Science, 32(11), 1492–1512.

Steinfield, C. W. (1986). Computer-mediated communication in an organizational setting: explaining task-related

and socioemotional uses. In M. L. McLaughlin (Ed.), Communication yearbook 9 (pp. 777–804). Newbury

Park, CA: Sage.

Swan, K., Polhemus, L., & Shih, V. (2002). An investigation of the development of social presence in asynchro-

nous online course discussions. Paper presented at the Annual Conference of American Educational Research

Association (AERA).

Towell, J., & Towell, E. (1997). Presence in text-based networked virtual environments or ‘‘MUDS’’. Presence,

6(5), 590–595.

Tu, C. H. (2000). Critical examination of factors affecting interaction on CMC. Journal of Network and Computer

Applications, 23(1), 39–58.

Tu, C. H. (2001). How Chinese perceive social presence: an examination of in online learning environment.

Educational Media International, 38(1), 45–60.

Tu, C. H. (2002a). The impact of online tasks and social relationships on social presence in online knowledge

construction learning community. Unpublished Manuscript. Washington, DC: The George Washington Uni-

versity.

Tu, C. H. (2002b). The measurement of social presence in an online learning environment. International Journal

on E-Learning, 1(2), 34–45.

Tu, C. H. (in press). The impacts of text-based computer-mediated communication on online social presence.

Journal of Interactive Online Learning.

Tu, C. H., & Blocher, J. M. (2000). Integrating student publishing and peer evaluation on the Internet using a

database-driven design. Paper presented at the Annual Conference of American Educational Research Asso-

ciation. Retrieved April 3, 2002, from http://home.gwu.edu/~ctu/publication/2002/sp3cmc.pdf.

Tu, C. H., & Corry, M. (2002). Social presence and critical thinking for online learning. Paper presented at

Annual Meeting of American Educational Research Association.

Tu, C. H., & McIsaac, M. S. (2002). An examination of social presence to increase interaction in online

classes. American Journal of Distance Education, 16(3), 131–150.

Walther, J. B. (1992). Interpersonal effects in computer-mediated interaction: a relational perspective. Communi-

cation Research, 19(1), 52–90.

C.-H. Tu / Internet and Higher Education 5 (2002) 293–318 317

Walther, J. B., & Burgoon, J. K. (1992). Relational communication in computer-mediated interaction. Human

Communication Research, 19, 50–88.

Weisband, S. P., & Reinig, B. A. (1995). Managing user perceptions of email privacy. Communications of the

ACM, 38(12), 40–47.

Wiener, M., & Mehrabian, A. (1968). Language within language: immediacy, a channel in verbal communication.

New York: Appleton-Century-Crofts.

Williams, F., & Rice, R. E. (1983). Communication research and the new media technologies. In R. N. Bostrom

(Ed.), Communication yearbook 7 (pp. 200–224). Beverly Hills, CA: Sage.

Witmer, D. F. (1997). Risky business: why people feel safe in sexually explicit on-line communication. Journal of

Computer Mediated Communication, 2(4).

C.-H. Tu / Internet and Higher Education 5 (2002) 293–318318