Computer-Mediated Communication Effects On Disclos PDF

See
discussions, stats, and author profiles for this publication at:

https://www.researchgate.net/publication/232558393
Computer-Mediated
Communication Effects on
Disclosure, Impressions, and
Interpersonal Evaluations:
Getting to...
Article in Human Communication Research · July 2002

DOI: 10.1111/j.1468-2958.2002.tb00811.x
CITATIONS READS
550 2,107
2 authors, including:
Joseph B. Walther
Nanyang Technological University
115 PUBLICATIONS 11,827 CITATIONS
SEE PROFILE
All content following this page was uploaded by Joseph B. Walther on 17 February 2014.
The user has requested enhancement of the downloaded file.

Tidwell, Walther / KNOWLEDGE STRATEGIES IN CMC 317
Computer-Mediated Communication Effects

on Disclosure, Impressions, and Interpersonal
Evaluations
Getting to Know One Another a Bit at a Time
LISA COLLINS TIDWELL

Telecheck Services, Inc.
JOSEPH B. WALTHER
Cornell Universiy
This investigation examined how computer-mediated communication (CMC) partners ex-

change personal information in initial interactions, focusing on the effects of communication
channels on self-disclosure, question-asking, and uncertainty reduction. Unacquainted indi-
viduals (N = 158) met either face-to-face or via CMC. Computer-mediated interactants ex-
hibited a greater proportion of more direct and intimate uncertainty reduction behaviors than
unmediated participants did, and demonstrated significantly greater gains in attributional
confidence over the course of the conversations. The use of direct strategies by mediated
interactants resulted in judgments of greater conversational effectiveness by partners. Re-
sults illuminate some microstructures previously asserted but unverified within social infor-
mation processing theory (Walther, 1992), and extend uncertainty reduction theory (Berger
& Calabrese, 1975) to CMC interaction.
T
he advent of computer-mediated communication (CMC) and its
penetration into so many people’s lives provides communication
theorists with an interesting lens on human behavior. Do individu-
als adapt to the restrictions of the medium, and if so, when and how? Can
we learn about human communication in general from these adaptations,
the ways that communication with new tools subtly differs from tradi-
tional modes, and how they affect our fundamental relations with one
Lisa Collins Tidwell (Ph.D., Northwestern University, 1998) is director of market strategy at
Telecheck Services, Inc., in Houston, Texas. Joseph B. Walther (Ph.D., University of Arizona,
1990) is an associate professor of communication at Cornell University. This research was
partially funded through a grant provided by the Ameritech Consortium for Research on
Telecommunication Strategy and Policy (#0965340E257). Results of this research were pre-
viously reported at the 7th International Conference on Language and Social Psychology,
Cardiff, Wales, July 2000, and at the Association of Internet Researchers, Lawrence, KS, Sep-
tember 2000. The authors wish to express their gratitude to Alison Foster for her extensive
assistance in coding, and to the anonymous reviewers whose comments aided in the im-
provement of the article. Correspondence about this article should be addressed to Joseph
B. Walther, Department of Communication, 336 Kennedy Hall, Ithaca, NY 14853-4203.
Human Communication Research, Vol. 28 No. 3, July 2002 317–348

© 2002 International Communication Association
318 HUMAN COMMUNICATION RESEARCH / July 2002
another? As important as the burgeoning of the Internet is the growth of

research on its impacts. Yet despite many studies examining the perceived
similarities and differences in communication outcomes across media, in
many areas, fundamental questions about the processes that lead to such
outcomes lag behind.
For instance, while it is now known that the Internet provides venues
for the initiation of relationships, from friendly to romantic (e.g., Baker,
1998; Parks & Floyd, 1996; Parks & Roberts, 1998), such studies do not
describe at a theoretical or behavioral level how these relationships come
to be. Research provides little at the level of particulars regarding changes
in the basic process of acquaintance development and interpersonal knowl-
edge acquisition in online settings. These processes not only deserve at-
tention in their own right, but examination of these building blocks in-
forms other theories, traditional and newer, about communication with
and without media. They also reflect public interest in questions such as:
How can people get a sense of one another without face-to-face commu-
nication and the observation of nonverbal cues?
This study sought to determine the means by which impression devel-
opment takes place in CMC, and the extent to which impression forma-
tion is accomplished, relative to face-to-face (FtF) communication. Given
the differential availability of cues and interaction contexts between me-
diated and FtF communication, this research focuses on (a) the potential
redistribution of impression-seeking questions and self-disclosure pro-
cesses within CMC and FtF channels, which affect impressions and
attributional confidence, and (b) the evaluative consequences of poten-
tially different patterns. Does one’s self come shining through regardless
of the interaction conduit? If so, how?
CMC Theories and Evidence

It is well known that text-based CMC such as email, computer
conferencing, and chat systems differ from FtF communication in several
respects, the most apparent of which is that the written medium precludes
the exchange of nonverbal cues that accompany FtF speech. This differ-
ential has led some to suggest that impressions and relational develop-
ment might be thwarted in CMC, rendering it a relatively impersonal
medium (e.g., Kiesler, Siegel, & McGuire, 1984; Siegel, Dubrovsky, Kiesler,
& McGuire, 1986). Other perspectives suggest that users rely on alterna-
tive mechanisms to accomplish these functions (see for review Walther,
1996). Empirical research on the formation of relationships online has
produced mixed conclusions, especially with regard to users’ impression
formation processes. Early research suggested that people using CMC
were prevented from gaining impressions due to the lack of nonverbal
cues in the medium (e.g., Kiesler, 1986; see also Kiesler et al., 1984; Siegel,
et al., 1986). Other research shows that online impressions and relation-
ships are developmental, operating within different temporal frames than
FtF communication (Walther, 1993; Walther & Burgoon, 1992) or that they
are subject to the effects of salient group identities (Spears & Lea, 1992;
Walther, 1997). The social information processing (SIP) theory of CMC
(Walther, 1992) argues that without nonverbal cues, communicators adapt
their relational behaviors to the remaining cues available in CMC such as
content and linguistic strategies, as well as chronemic (Walther & Tidwell,
1995) and typographic cues (Walther & D’Addario, 2001). Despite these
formulations and studies, it remains to be seen precisely whether the re-
spective relational states are achieved through the specified mechanisms.
That is, while there must be some means by which CMC partners trans-
late social information into the text-based medium of CMC, few studies
have examined by what specific means users adapt to the medium in
order to seek and to exhibit uncertainty-reducing, impression-bearing cues.
Research supporting the paradigm of impersonal CMC effects was gen-
erally strong on content analysis of conversational behavior, but did not
as a rule assess interpersonal evaluations (e.g., Hiltz, Johnson, & Turoff,
1986). In contrast, research supporting alternative paradigms has not, for
the most part, explicitly examined the alternative behavioral processes
that allegedly occur in the theoretical chain of events they specify. The
communication behaviors nominated to correspond to the outcomes speci-
fied in these theories either (a) have not been investigated, or (b) have
been analyzed but not related to outcomes. Social information processing
research, for example, generally involved assessing CMC/FtF differences,
in combination with temporal variations, on outcome measures includ-
ing self-reported impressions of partners and relational communication
(Walther, 1992; Walther & Burgoon, 1992). In a study on impression de-
velopment, Walther (1993) concluded that “CMC users formed increas-
ingly developed impressions over time, presumably from the decoding
of text-based cues” (p. 393). In that study, impressions and time were ex-
amined, while the text-based cues were presumed to have occurred. Thus
these empirical tests have focused on the conditions and outcomes that
are consistent with the framework, rather than on the specific
microbehaviors that are presumed to facilitate them. Using a more micro-
level focus, Utz (2000) found a relationship between the use of textual
paralinguistic devices and the development of relationships in a CMC
setting, consistent with SIP theory. That study did not isolate the mediat-
ing role of impression development, however, and the frequency of
paralinguistics was assessed through self-report rather than content, leav-
ing some room for bias, as Utz acknowledged.
A related approach, the hyperpersonal perspective of CMC, argues that
the absence of nonverbal cues, as well as editing capabilities, identity cues,
and temporal characteristics may prompt CMC users to engage in selec-

tive self-presentation and partner idealization, enacting exchanges more
intimate than those of FtF counterparts. Tests of this framework have in-
cluded manipulation of salient identities, time frames, and the presence
of photographs (in combinations of CMC/FtF settings), on dependent
measures of self-reported assessments of partners’ relational communi-
cation and attractiveness (Walther, 1997; Walther, Slovacek, & Tidwell,
2001). Results supported and extended the framework. Recent research
specifically examining online impressions has found that CMC partners
form deeper, but not broader, impressions during task-focused communi-
cation (Hancock & Dunham, 2001). That is, in CMC versus FtF settings, part-
ners rated one another on a smaller number of attributes, but with more
extreme scores, suggesting more selective and yet exaggerated social in-
formation sharing online, consistent with the hyperpersonal approach. In
all these cases, however, the specific cues potentially leading to the self-
reported effects were presented anecdotally but not analyzed formally.
Similarly, the social identity and deindividuation (SIDE) theory (Lea &
Spears, 1992) argues that the lack of nonverbal cues in online interaction
prompts users to form impressions that are based on social categories of
communicators, not interpersonal cues. Tests of this theory have involved
salient personal, group, or national identities, and the presence of primed
norms, or photos, in combination with CMC. The effects of these inde-
pendent variables on self-reported ratings of partners’ social attraction,
similarity, and other evaluations were studied (see for review Postmes,
Spears, & Lee, 1998), but not the behaviors that prompted them. In a re-
cent exception, one test of SIDE theory examined the content and style of
users’ online comments, finding convergence in message attributes to-
ward subgroup norms over time (Postmes, Spears, & Lea, 2000). How-
ever, these convergences were not shown to relate to affective outcomes
or impressions. Similarly, Joinson’s research (2001, Study 1) found more
self-disclosure in CMC than parallel FtF discussions, but the connection
between disclosure and participants’ goals and outcomes was not specified.
Each of these approaches has advanced our knowledge about media
and communication in the development of impressions and relationships,
through the kinds of sensible first steps of research in any area. After all,
if there are no gross effects there is little reason to look for particulars. But
an important second step is to see whether and how the intervening pro-
cesses occur as specified.
Uncertainty Reduction: Strategies and Limitations

Although CMC research has been relatively silent at the level of par-
ticulars, an approach exists which focuses on behavioral strategies lead-
ing to interpersonal impressions and evaluations, and which is associ-
ated with accompanying methodologies for examining the behavioral

strategies it specifies. Although it has not been applied to CMC, uncer-
tainty reduction theory (URT; Berger & Calabrese, 1975) is among the most
heuristic approaches to impression formation and initial relational devel-
opment in FtF settings. The ways that its dynamics manifest themselves
in CMC settings, however, may not be the same. An unstated boundary
condition in URT is the presence of FtF interaction; URT presumes that
partners encounter each other physically when they interact. This aspect
is not the central focus of this theory (or others on impressions and attrac-
tion) and might be dismissed as theoretically irrelevant (Lea & Spears,
1995); even though URT explicitly includes nonverbal affiliative behav-
iors in its calculus, these are substitutable with verbal behaviors. How-
ever, tests of URT have been conducted using FtF conversation, necessar-
ily involving the conventionally obvious aspect of seeing one’s partner
before speaking and potentially gleaning data from physical appearances
and other nonverbal indicators (see Douglas, 1990). In the age of the
Internet, the FtF boundary is an anachronism and the FtF research set-
ting, a potential confound. Since the role of physical presence in URT has
not been examined, its applicability to situations in which there is no physi-
cal presence remains unknown. The constraints on uncertainty reduction
strategy selection, usage, and interpretations demanded by virtual inter-
action deserve exploration both to inform URT as well as to unpack com-
munication processes suggested in the CMC theories described above.
According to URT, uncertainty reduction is the exchange and collec-
tion of information that allows one to predict another’s attitudes and be-
haviors (Berger & Calabrese, 1975). In reducing uncertainty people create
impressions—mental models that help them to make sense of people and
situations (Srull & Wyer, 1989). The more information one gets about one’s
partner, the less uncertainty he or she has. The mechanisms by which
they exchange information, however, may be limited in CMC settings.
URT predicts that individuals use passive, active, and interactive strate-
gies as a means of reducing uncertainty. Berger (1979; Berger, Gardner,
Parks, Schulman, & Miller, 1976) identified three passive strategies, two
active strategies, and three interactive strategies, that people utilize in
order to acquire information about others. Which of these strategies are
employed, and when, depends in part on politeness and situational ap-
propriateness requirements.
Passive strategies are those in which the information seeker collects
information about the target without affecting the target’s or other actors’
behavior. The first passive strategy identified by Berger (1979) is the reac-
tivity search, the unobtrusive observation of a target person in a situation
in which the target interacts with, and reacts to others. The second is so-
cial comparison, observing a target as he or she interacts with others who
are known to the observer, and drawing comparisons based on similarity

of the target to the others. The third passive strategy is disinhibition
searches or deviation testing (Berger et al., 1976). This strategy seeks to
identify the target’s candid behaviors by looking at them in situations
where targets’ inhibitions are likely to be less impacted by social norms
and rules.
The two active, yet indirect strategies, differ from the passive strate-
gies in that they involve proactive efforts to gain knowledge about an-
other person (Berger, 1979). However, they are indirect as they do not
confront the target. The first of these two consists of asking others about
the target individual. The second, environmental restructuring, entails
the creation of a set of circumstances followed by unobtrusive obser-
vation of the target in that environment, in order to view the effect of
the manipulation.
Interactive strategies for uncertainty reduction require direct and ob-
trusive exchanges with targets. The first, deception detection, involves
identification of falsifications, distortions, or omissions of information.
Second, verbal interrogation refers to question asking. Finally, self-dis-
closure is that which a person knowingly communicates to another about
him- or herself, which is not publicly known (Worthy, Gary, & Kahn, 1969). Dis-
closure is bound by the norm of reciprocity: self-disclosure begets disclosure
from a target person, thus its utility as an information-seeking behavior.
Although any of these strategies might be chosen in FtF settings, their
availability may be much more limited in CMC settings. For instance,
passive strategies require public settings, rather than private ones. In CMC
these may be confined to unusual contexts such as Usenet groups, Multi-
User Dungeons/Object Oriented (MUDs or MOOs), and mailing lists.
However, the most common form of CMC is electronic mail (Kraut,
Mukhopadhyay, Szczypula, Kiesler, & Scherlis, 1999). Given the predomi-
nance of email for communicating online, unless one is operating within
one of these other special settings, passive strategies are largely unavail-
able (cf. Ramirez, Walther, Burgoon, & Sunnafrank, 2002).
Active strategies, too, can be more difficult to deploy in CMC than FtF,
as they require certain resources that may be less available online. Spe-
cifically, in the case of asking third parties for information, CMC relation-
ships offer a more limited network of common acquaintances than in FtF
associations (Parks & Floyd, 1996). Similarly, environmental restructur-
ing requires advance planning and the involvement of third parties. These
two strategies might be adaptable to ongoing CMC environments such as
MUDs or lists, where it may be possible to observe the target unobtru-
sively, but in general, the opportunities to implement passive and active
uncertainty reduction strategies are very limited in CMC, particularly in
initial interaction settings.
All of the interactive strategies suggested by Berger and colleagues

may be utilized in both FtF and CMC. Deception detection, however, tends
to be considered unreliable in CMC (see Donath, 1999; Van Gelder, 1985).
Whether reliable or not, people attempt to detect deception FtF through
observation of individuals for nonverbal “signs” of deception (Zuckerman,
DePaulo, & Rosenthal, 1981). Unless target individuals contradict either
themselves or known facts, deception detection is less likely employed in
a CMC environment. This leaves verbal interrogation (i.e., question ask-
ing) and self-disclosure as the uncertainty reduction strategies most avail-
able in CMC. Questions and disclosures are both readily deployed in CMC.
Personal questions can easily be posed in text. Self-disclosures may also
be effective in CMC settings and operate in patterns similar to FtF: They
not only provide impression-bearing information, but the process of dis-
closing creates a demand, so that the recipient feels obligated to respond
in kind, typically generating return disclosures from the target individual
(Jourard, 1971).
While these formulations about strategies available in CMC seem rather
straightforward, the literature has not considered the implications. In one
exception, Joinson (2001, Studies 2 and 3) found greater self-disclosure in
CMC chat than FtF meetings, although these findings were largely in re-
sponse to experimental variations in public and private self-awareness or
videoconferencing; question-asking, as a means of interpersonal knowl-
edge acquisition, was not regarded. In a sense, the relative unavailability
of strategies in common CMC settings provides a scope condition to URT:
The interactive strategies of questioning and self-disclosure should be
those most relied upon for acquiring knowledge in CMC, whereas a vari-
ety of strategies may be used FtF. The question remains whether CMC
users will use more interactive strategies as an accommodation to the
lack of other means, as suggested by SIP theory, or whether they will
simply mirror the same proportion of interactive uncertainty reduction
in their conversations as do FtF communicators (although other theories
question whether interpersonal impressions are even sought through
CMC, much less how).
H1: In initial interactions, computer-mediated interactants deploy a greater

proportion of interactive uncertainty reduction strategies (questions and
self-disclosures) than do FtF counterparts.
Reconciling Efficiency With Politeness

Other interesting questions about the different natures of FtF and CMC
are raised by considering the implications of such a shift in strategy use.
As interactive strategies are considered riskier in FtF than are other strat-
egies, how might such a difference in usage affect users’ reactions to one
another in CMC? It is a rule in FtF communication that in order to appear

polite, interactants must balance the efficiency of uncertainty reduction
strategies with social appropriateness (Berger & Kellermann, 1983;
Kellermann & Berger, 1984). Although they are also more time- and re-
source-consuming to use, passive and indirect strategies for uncertainty
reduction are less face-threatening than interactive strategies, and there-
fore more polite (Brown & Levinson, 1987). The desire to demonstrate
both effective and appropriate behavior may hinder more direct and effi-
cient uncertainty reduction activities in a FtF context. However, appro-
priateness and effectiveness concerns may take shape differently in CMC
contexts. Given the lack of alternative strategies for uncertainty reduc-
tion available in CMC, the otherwise face-threatening interactive strate-
gies may be forgiven in this context due to their greater relative normalcy.
While question-asking and disclosure might be less polite FtF, evaluation
of these strategies may be more favorable in CMC.
H2: Media type interacts with interactive uncertainty reduction strategies such
that when CMC interactants use more interactive strategies, they are judged
to be more appropriate by their partners; whereas, when FtF interactants
use more interactive strategies they are judged to be less appropriate by
their partners.
Although questions are efficient, purposeful, and easily understood in

both CMC and FtF settings, these attributes may be particularly valued
in CMC. Efficiency is important since CMC simply takes more time and
effort than FtF communication (Olaniran, 1996; Sproull & Kiesler, 1991).
Questions are also perceived to be honest in nature. In the absence of
most nonverbal cue systems, expressions of sincerity are particularly im-
portant online (see Donath, 1999). Proactive self-disclosing is also valued
for the same reasons. Disclosures commonly occur when the other is per-
ceived to be trustworthy or when the “discloser expects he will benefit in
some way if he permits the other person to know him as he is” (Jourard,
1971, p. 65). Due to the norm of reciprocity, mediated interactants may
realize that proactive sharing of information assists in the mutual acqui-
sition of information and reduction of uncertainty. Since CMC interactants
are assumed to share the same desire to reduce uncertainty as FtF partici-
pants (Walther, 1992), this may be one of the few practical means of achiev-
ing that goal. If mediated interactants understand that certain (passive)
strategies in FtF settings may not be possible in CMC, then those who
adapt to CMC and use direct, interactive uncertainty reduction strategies
will most likely be judged to be more effective.
H3: Media type interacts with interactive uncertainty reduction strategies such
that when mediated interactants use more interactive strategies, interactants
are judged to be more effective by their partners; whereas, when FtF

interactants use more interactive strategies, they are judged to be less effec-
tive by their partners.
CMC may alter not only the proportion and the evaluation of self-dis-
closure and interrogation, but also expand the depth. Altman and Taylor
(1973) explain in social penetration theory that as a relationship devel-
ops, the individuals become willing to disclose more aspects of their self,
providing depth of information on a breadth of topics. Initial social ac-
quaintanceship conversations in a FtF setting do not provide the proper
forum for allowing truly deep revelations: Prior research suggests that at
least the first minute of FtF conversation is consumed primarily with rela-
tively low-intimacy demographic information. The CMC channel, as we
have argued above, provides fewer efficient methods to reduce uncer-
tainty, and questions of greater depth may provide another efficient twist
on otherwise less personal exchanges. At the same time, CMC provides
some sheltering effects not offered in FtF conversation, and mediated
interactants may feel more comfortable disclosing more intimate infor-
mation, with less self-consciousness, than in FtF settings.
H4: CMC interactants engage in deeper self-disclosures than do FtF

counterparts.
In addition to leaving open the role of impression-facilitating behav-

iors in CMC, previous research has not attempted to assess directly the
degree of mediated interactants’ attributional confidence regarding their
partners. As mentioned above, Hancock and Dunham (2001) found that
in a short time interval CMC partners rated each other on a smaller num-
ber of attributes than FtF counterparts. Walther (1993) found that, over
extended interactions and similar amounts of message exchange, CMC
and FtF partners reached similar levels on the number of traits they could
attribute to their partners. However, the degree of uncertainty reduction
has not been directly assessed. Although both FtF and CMC interactants
are predicted to reduce uncertainty, they may not do so to equal extents.
FtF interactants have access to a wider range of cues than are available in
CMC and deliver nonverbal cues simultaneously with verbal cues. Nu-
merous studies comparing FtF and CMC, as well as the exchange of physi-
cal and aural cues via audio- and video-conferencing systems, predomi-
nantly report greater feelings of “social presence” when more cues are
available, suggesting that uncertainty reduction is itself highly depen-
dent on the exchange of multiple cues (e.g., Gale, 1991; Whittaker &
O’Conaill, 1997; cf. Burgoon et al., 2000; Walther et al., 2001). SIP (Walther,
1992) argues that impressions and relational communication improve over
time in CMC to parallel that of FtF communication, precisely due to the
development of personal knowledge structures of one’s partners. In con-

trast, the hyperpersonal perspective (Walther, 1996) describes how CMC
partners develop impressions and relational states that are as positive or
better than those of FtF partners, a perspective premised on the exaggera-
tion and overprocessing of partial and fragmented information. It is thus
unclear whether impression development is due to actual uncertainty
reduction, or to biased processing and relationally-positive yet relatively
uncertain states. Given these conflicting positions, a research question is posed.
RQ1: Do FtF interactants have greater attributional confidence than their

CMC counterparts?
The Nature of Cues in Impression Formation Online

As alluded to above, SIP predicts that personal knowledge structures
accrue during CMC interaction. If so, certain cues should serve as the
basis for attributions, which interactants may recall. Alternatively, the SIDE
model predicts that it is not the individual-level information that stimu-
lates impressions of others in CMC, but rather the social identity shared
by members of an online conversation (Lea & Spears, 1992; Spears & Lea,
1992, 1994). SIDE theory equates nonverbal cues with individuating im-
pression stimuli. Since such cues are absent in CMC, one’s social, or group,
identity may be more salient than personal identity. Context cues related
to social categories become more important, providing a primary basis
for attributions, within which users’ “paralinguistic content of messages
in CMC can provide social cues that assist with impression formation”
(Lea & Spears, 1992, p. 328). These dynamics promote attraction within a
group as they lead members to attribute similarities among themselves
and others. If SIDE is correct, a few prominent social identity cues impact
impressions through global heuristics, providing social category indica-
tors that allow for social comparisons.
At this time, it is unclear which elements in CMC lead to uncertainty
reduction: a variety of verbal and paralinguistic cues as predicted by SIP,
over-reliance on salient social identity cues as predicted by SIDE, or reli-
ance on some other facet of the interaction. Are mediated interactants
truly forming personal schemata of their conversational partners, or are
they simply deriving images from categorical heuristics, imbued with
misplaced confidence? Research in this area is both contradictory and
sketchy (see Jacobson, 1999). For that reason the following research ques-
tion was formulated in order to direct inquiry into the means by which
impression formation occurs in a mediated environment.
RQ2: On what cues do CMC and FtF interactants rely in order to reduce
uncertainty?
METHOD
Participants (N = 158), half male and half female, were recruited from
communication courses at a moderately-sized, private American univer-
sity where they received extra course credit for their participation. Ages
ranged from 17 to 24 with a modal age of 19. Participants were instructed
to sign up for a time slot with someone of the opposite sex with whom
they were not already acquainted (i.e., a stranger), and told that they would
engage in one or more conversations taking place across one or more com-
munication channels. All participants had email accounts through the
university, and were required to know how to use email in order to par-
ticipate. They rated their experience level using email very high on a 7-
interval scale (M = 2.09, SD = 1.54, where 1 = very experienced and 7 = not
at all experienced), and they used email frequently on both a daily (M =
6.06, SD = 1.48) and weekly bases (M = 5.03, SD = 2.05), as shown on 7-
point scales (1 = not at all and 7 = all the time).
Partners were instructed to report to different laboratory rooms in or-
der that they would remain unacquainted prior to their experimental in-
teraction. A research administrator presented a research consent form and
specific instructions in writing and aloud detailing the channels of com-
munication to be used. Half of the dyads then met in a small meeting
room for a FtF discussion. Each member from the other half of the dyads
went to an individual room to communicate with his or her partner over
a semi-synchronous CMC system. Using email accounts established
specifically for this research, they used the email system to send mes-
sages to one another and to read responses as soon as the messages were
completed. The system was similar in form to instant messaging systems,
but the email protocol facilitated generating copies of the messages for
retrieval and analysis of individuals’ responses.
Participants were instructed either to get to know one another, as is
common in URT research, or to work on a solution to a decision-making
problem, as is common in CMC research. Whereas this topic variable had
some main effects on disclosure behaviors (to be reported elsewhere), it
did not present disordinal interactions with the independent variables in
this report, and the analyses to follow collapsed across these conditions.1
Different time periods were allocated for FtF and CMC conditions.
Altman and Taylor (1973, p. 189) assert that “the social penetration pro-
cess is a time-bound one,” and SIP research also argues strongly that con-
strained or equal time periods in CMC and CMC/FtF research have dra-
matic impacts on communicator behaviors and their comparisons (see
Walther, Anderson, & Park, 1994). Therefore, it was important to make
interaction opportunities equivalent across the two communication con-
ditions. FtF subjects were told that they could converse for up to fifteen
minutes or until they signaled the administrator that they were finished,
but since 15 minutes in a FtF setting is not equivalent to 15 minutes in

CMC (typing alone accounts for differences in the amount of remarks
generated per minute by FtF and CMC participants; Siegel et al., 1986),
CMC interactants were allocated up to 60 minutes, four times the amount
of time provided for FtF interactants. A subsequent analysis of the con-
versations determined that the FtF meetings lasted from 57 seconds to
25.82 minutes (M = 15.1 mins, SD = 8.83 mins), while the CMC interac-
tions lasted from 8.1 minutes to 61.78 minutes (M = 43.1, SD = 17.01). The
FtF interactions ranged from 8 to 230 utterances (M = 87.32, SD = 64.04);
CMC interactions ranged from 4 to 150 utterances (M = 42.97, SD = 29.02).
Despite attempts to equalize talk time, approximately twice as many ut-
terances were contained in the FtF meetings as in the mediated interac-
tions. Since this discrepancy in the baseline number of messages appeared
between conditions, several hypothesis tests were calculated using pro-
portions of conversation elements rather than frequencies.
Following the interaction, participants were separated and asked to
complete a number of measures relating to their impressions of their part-
ners and their own behaviors. Measures featuring 7-interval Likert-type
scales, with demonstrated reliability and validity, assessed participants’
perceptions of conversational effectiveness, appropriateness, uncertainty,
and self-disclosure. Additionally, a small number of open-ended ques-
tions were utilized to identify conversation cues that aided or hindered
interactants’ uncertainty reduction efforts.
Conversational effectiveness was assessed using a five-item subset of
the conversational effectiveness scale (Canary & Spitzberg, 1987). Per-
ceptions of partners’ conversational appropriateness were determined
using a three-item subset of Canary and Spitzberg’s (1987) conversational
appropriateness scale. Both yielded acceptable Cronbach reliabilities (α =
.73 for each).
Postconversation uncertainty about the partner was assessed using a
five-item subset of Clatterbuck’s CL7 measure of global uncertainty
(Clatterbuck, 1979; Douglas, 1990). High scores represented greater un-
certainty, whereas low scores indicated greater levels of attributional con-
fidence (i.e., a high degree of confidence regarding knowledge of another
and ability to predict and understand another’s behavior), α = .90. Par-
ticipants also completed this measure with reference to their preinteraction
uncertainty in order to provide a retrospective pretest baseline. Partici-
pants were asked, “Remind yourself how you felt before the conversa-
tion began. How well did you ‘know’ your partner; that is, how well did
you know his or her beliefs, values, attitudes, goals, or experiences? Se-
lect a number on the scale from 0 to 100 that best indicates how well you
knew your partner before your conversation began.” This same scale was
used to measure terminal attributional confidence. The use of this mea-
sure retrospectively is somewhat unorthodox, yet methodological con-

straints, and other research precedents, suggested this approach. While
an actual pretest measure of attributional confidence would in some ways
be preferable, introducing one would have been likely to heighten artifi-
cially participants’ responsiveness to the information-seeking focus of the
research (as noted by Douglas, 1990), especially in the decision-making
context in which half of the dyads operated. However, posttest retrospec-
tive ratings have been employed successfully in interpersonal evaluation
research (e.g., Hale, Lundy, & Mongeau, 1989) as well as in communica-
tion technology research (e.g., Rice & Contractor, 1990) in which both pre-
and posttest ratings were gathered after the stimulus period. In order to
check whether the use of these scales revealed theoretically expected pat-
terns of uncertainty reduction over time (see Douglas, 1990), CMC/FtF
conditions were collapsed and tested together. Retrospective ratings of
initial attributional confidence were compared to terminal levels, which
differed significantly, t(156) = -4.16, p < .001: Interactants were more con-
fident at the end of the conversation (M = 47.3, SD = 25.69, n = 158) than
they believed they had been at the beginning (M = 6.15, SD = 14.98, n =
158), as would be expected. It thus appeared that the retrospective rating
operated as it was intended, although readers should interpret the results
of these measures cautiously in the context of their dual administration.
The cues that individuals utilized to reduce uncertainty, the focus of
RQ2, were collected by asking the open-ended questions, “What sorts of
things affected your ability to get to know your partner? That is, what
sorts of things made it easier to get to know him/her?” Three judges each
evaluated all the open-ended responses (using a procedure later repli-
cated with the conversation transcripts; complete unitizing and coding
rules are available upon request). First, by placing vertical lines around
each utterance, the judges designated utterances using Holsti’s (1969, p.
116) definition of an utterance as “a single assertion about some subject.”
Each judge initially identified utterances independently, after which re-
sponses were compared across judges. When two or more judges agreed
upon a judgment, it was accepted. When necessary, differences among
judges were resolved through discussion. This procedure yielded 301 ut-
terances. Intercoder reliability was calculated between judges A/B, B/C,
and A/C. The proportion of consensual answers (agreements divided by
the total number of utterances) generated interjudge agreements in the
range of .92 to .94. Next, a composite judgment reliability was calculated
using Guetzkow’s (1950) procedure, which yielded a proportion of .97.
Judges then identified and labeled uncertainty-reducing conversation
cues, by assigning each utterance to one of nine uncertainty-reducing cue
categories that best represented the participant’s meaning when taken in
context. The nine categories were taken from Tidwell, Douglas, and Banski
(1995), and are detailed in the results section. Interjudge reliabilities were
calculated based on content agreement. The proportion of correct responses
was calculated for each coding set by dividing the number of consensual
judgments by the total number of utterances. The proportion of agreements
ranged from .66 to .72, with a Guetzkow’s composite reliability of .88.
Participants were also asked to identify those conversation elements
that increased or enhanced uncertainty, with the question, “What sorts of
things affected your ability to get to know your partner? That is, what
sorts of things made it less easy to get to know him/her?” Judges evalu-
ated the content of the utterances once again using Tidwell et al.’s (1995)
cue categories, with interjudge agreement between .59 to .68, with a
Guetzkow’s composite reliability of .85.
Participants’ assessments about the amount and depth of their self-
disclosures were assessed using Wheeless’s (1978) instrument, adapted
to past tense to refer to the conversation the subject had just completed.
High scores reflected greater levels of self-disclosure (5 items, α = .77)
and control over depth of self-disclosure (5 items, α = .82).
Disclosures and questions. All audiotapes of the FtF meetings were tran-
scribed and electronic mail messages were printed. Verbal records were
coded by two judges in order to assess the quantity and quality of verbal
disclosure and interpersonal questions. The judges first identified all the-
matic utterances. Using the same coding procedure mentioned previously,
coding responses were compared and disagreements were resolved
through discussion. The conversations contained a total of 10,515 utterances.
Interjudge utterance reliability was calculated as a proportion of agree-
ments divided by the total number of utterances, with a proportion of .95.
Next, judges assigned utterances to one of three content categories:
question, self-disclosure, or other. Questions were operationally defined
as “an expression of inquiry that invites or calls for a reply; an interroga-
tive sentence, phrase, or gesture” (Morris, 1976, p. 1070). Self-disclosures
were operationalized as messages that reveal personal information about
the sender, as derived from Dindia (1983) and Chelune (1975, p. 133, as
cited in Tardy, 1988) as “a verbal response (thought unit) which describes
the subject in some way, tells something about the subject, or refers to
some affect the subject experiences.” Utterances that were neither a ques-
tion nor a self-disclosure (e.g., exclamations and imperatives) were coded
as “other.” Coding disagreements were resolved, and the interjudge reli-
ability proportion was .94 for the content of the utterances.
Breadth and depth of questions and self-disclosures. Next, two additional
judgments were performed on each self-disclosure and question utter-
ance. Judges identified the topic and degree of intimacy of each question
and disclosure. While several classification schemes were examined, only
one had actually been used to code utterances with reliability: Taylor and
Altman’s (1966). In the present effort, some topic categories from the Tay-
lor and Altman self-disclosure scheme were merged, while other distinctly
different topic constructs were added.2 Using the modified classification
scheme, the two judges evaluated each question, self-disclosure, and other
expression in order to assign a topic rating. A reliability proportion was
calculated by taking the initial number of agreements and dividing that
by the total number of utterances. Interjudge reliability was .88 for topi-
cal content. Disagreements were resolved through discussion.
The depth of each utterance was assessed using Altman and Taylor’s
(1973) three-layer categorization scheme consisting of the peripheral, in-
termediate, and core layers. The peripheral layer is concerned with bio-
graphic data. The intermediate layer deals with attitudes, values, and
opinions. The core layer is comprised of personal beliefs, needs, fears,
and values. Judges’ intercoder reliability for utterance depth was .95, and
disagreements were resolved through discussion. Once the utterance con-
tent, topic, and intimacy judgments were performed, frequencies were
calculated. This was done by assessing the number of questions, disclo-
sures, topic categories (breadth), and intimacy ratings (depth) utilized by
each individual.
RESULTS
Uncertainty Reduction Strategies in CMC

The first three hypotheses dealt with uncertainty, and the impact of
the medium on the strategies that were employed in order to reduce un-
certainty. 3 The first hypothesis predicted that computer-mediated
interactants would engage in more interactive uncertainty reduction strat-
egies in the form of increased questions and self-disclosures. The second
hypothesis predicted that the increased use of interactive uncertainty re-
duction strategies in CMC would lead to judgments of greater appropri-
ateness, and according to the third hypothesis, to greater effectiveness
also, compared to judgments of these behaviors FtF.
The first hypothesis was assessed using t-tests, comparing the arc sine
transformations of the square roots for proportions of questions, and for
those of self-disclosures in the FtF condition, to the transformed propor-
tions of each found in CMC (see Snedecor & Cochran, 1989). As predicted,
CMC and FtF interactants differed in both behaviors. CMC interactants
(n = 74) utilized a significantly higher proportion of questions (MCMC pro-
portion = .18, SD = .11) than their FtF counterparts (n = 84; MFtF propor-
tion = .13, SD = .08), t(156) = -3.20, p = .002. CMC interactants also pro-
duced significantly higher proportions of self-disclosures (MCMC = .69, SD
= .12) than did those in FtF interactions (MFtF = .59, SD = .13), t(156) =
-4.98, p < .001. FtF interactants, in contrast, displayed a significantly greater

proportion of other expressions (MFtF = .28, SD = .11) than did CMC par-
ticipants (MCMC = .13, SD = .08), t(156) = 9.28, p < .001. Other expressions
included conversation elements such as statements of fact that were
nonpersonal in nature, statements about third parties, exclamations, im-
peratives, preview and summary statements, greetings, backchanneling
elements, and other filler items that were clearly not questions or
self-disclosures.
Thus, CMC interactants used a greater proportion of direct and inter-
active uncertainty reduction strategies than did their FtF counterparts.
FtF interactants diluted their use of interactive strategies with other con-
versation elements. These findings support Hypothesis 1: Interactive un-
certainty reduction strategies are used with greater proportion in CMC
than they are in FtF.4
Hypotheses 2 and 3 concerned the differential effect of questions and
self-disclosures on conversation attributions. They predicted that increased
use of interactive uncertainty reduction strategies by CMC interactants
would raise judgments of appropriateness (H2) and effectiveness (H3) by
conversational partners, but lower appropriateness and effectiveness rat-
ings when FtF partners exhibited the same behaviors.
Correlations were computed between the frequency of questions, dis-
closures, and other utterances—and the proportions of these elements to
total utterances—with judgments of appropriateness and effectiveness
within each respective communication condition (see Table 1). In the FtF
condition, conversational effectiveness showed a negative zero-order cor-
relation with proportion of questions, r = -.28, p = .05 (two-tailed), and a
positive correlation with the proportion of “other” comments, r = .33, p <
.001. No other zero-order correlations between appropriateness or effec-
tiveness and utterances were significant.5 Hypothesis tests, however, in-
volved z-tests of the differences among (Fisher z‘ transformations of) the
correlations within CMC against those within FtF (the results of which
are also displayed in Table 1).
No differences in the relationship of appropriateness ratings and the
quantity or proportion of self-disclosures was obtained due to communi-
cation conditions. Similarly, the relationships among appropriateness and
the amount or proportion of questions did not differ between conditions.
H2 was not supported.
With regard to conversational effectiveness ratings, the correlations with
proportion of questions were significantly different due to communica-
tion channel, z = -2.56, p < .01 (one-tailed). The relationship for CMC
interactants was mildly positive (r = .12) whereas the FtF interactants’
was significantly negative (r = -.28). No effects were obtained with regard
to the correlations between effectiveness and proportion of self-disclo-
TABLE 1
Correlation Matrix of Interactive Strategies, Appropriateness, and Effectiveness
Perceived Proportion of Proportion of Proportion Total self- Total Total

effectiveness self-disclosures questions other disclosures questions other
Perceived FtF .14 .12 -.05 -.11 .12 .10 .01

appropriateness CMC .32** .08 -.04 -.04 .07 -.01 -.10
Perceived FtF -.12 -.28* .33** -.05 -.12 .09

effectiveness CMC -.13 .12 .07 -.04 .05 .01
Proportion of FtF -.54** -.79** .20 -.31** -.39**

self-disclosures CMC -.77** -.43** .33** -.31* -.17
Proportion FtF -.06 -.00 .56** .07

of questions CMC -.22 -.41** .33** -.24*
Proportion FtF -.17 -.00 .46**

other CMC .13 .15 .58**
Total FtF .63** .69**

self-disclosures CMC .61** .67**
Total questions FtF .62**

CMC .54**
NOTE: Proportion refers to the arc sine transformation of the square roots of the original proportions. Correlations for FtF, n = 84; CMC,
n = 74. Bold pairs indicate that the coefficients were significantly different from one another, p < 01.
*p ≤ .05, **p ≤ .01 for two-tailed correlations.
sures (nor with quantities of either questions or disclosures). Interestingly,

the difference between the proportion of “other” comments and effec-
tiveness ratings approached significance, z = 1.71, p = .09 (two-tailed),
favoring the FtF communicators, who were considered more effective by
their partners the more they used comments other than questions or dis-
closures. This finding may also reflect the notion that there are different
expectations for effective communication in CMC and FtF communica-
tion. Questions and disclosures may be superfluous or inefficient in FtF
compared to the bounty of nonverbal cues and superficial banter. One is
reminded of the customary chat room greeting, “R U MorF?” (“are you
male or female?”; Van Gelder, 1985) as an extreme example of an effective
online, but likely ineffective FtF, question. Overall, with respect to Hy-
pothesis 3, the results were mixed. Mediated interactants who used pro-
portionately more questions were perceived to be more effective than when
FtF partners did so, and FtF partners may be more effective when they
veer away from such direct strategies.
Accelerated Intimacy in CMC

The fourth hypothesis predicted that CMC partners would engage in
deeper questions and disclosures than would those in ongoing FtF rela-
tionships. Using the coders’ ratings of question and disclosure depth (pe-
ripheral, intermediate, or core), proportions of questions and self-disclo-
sures were calculated for each of these categories, within each individual’s
total utterances.
Analyses assessed transformed proportions of peripheral and inter-
mediate questions and self-disclosures between channel conditions, which
revealed contrasting patterns. FtF interactants communicated higher pro-
portions of peripheral questions (MFtF = .54, SD = .35) than did CMC part-
ners (MCMC = .36, SD = .32); t(156) = 3.32, p = .001. However, the pattern
was reversed for intermediate questions: CMC interactants conveyed a
higher proportion of intermediate questions (MCMC = .65, SD = .33) than
did FtF partners (MFtF = .45, SD = .35); t(156) = - 3.70, p < .001. No signifi-
cant differences were identified for proportions of self-disclosure levels.
These results suggest that CMC partners relied on relatively deeper probes
than the questions used by the FtF dyads. CMC dyads compensated for
the limitations of the channel by making their questions more intimate
than those exhibited FtF.
This effect may be threatened by the interdependency of the propor-
tions. In order for participants to demonstrate more peripheral questions,
they were likely to exhibit proportionally lower amounts of something
else. Due to the interdependency, it was helpful to examine the frequen-
cies of questions and self-disclosures in each layer directly, and then com-
pare them across the channel conditions. FtF interactants voiced a signifi-
cantly higher amount of peripheral questions (MFtF = 8.04, SD = 8.87; MCMC

= 3.27, SD = 4.32); t(156) = 4.25, p < .001. FtF partners also expressed more
peripheral self-disclosures (MFtF = 21.56, SD = 23.72; MCMC = 11.43, SD =
11.73), t(156) = 3.33, p < .001; and intermediate self-disclosures (MFtF =
28.31, SD = 18.79; MCMC = 18.41, SD = 12.42), t(156) = 3.85, p < .001.
However, the frequency was not significantly different for intermedi-
ate questions, (MFtF = 4.07, SD = 3.86; MCMC = 3.59, SD = 2.60), t(156) = .90,
p = .37, or core questions (MFtF = .02, SD = .15; MCMC = .008, SD = .40),
t(156) = .94, p = .35. While a lack of differences does not establish that
CMC partners used the same amount of intermediate or core questions
as FtF partners, it is nevertheless surprising given the greater number of
FtF comments overall. It demonstrates that on average the bulk of FtF
interactants’ exchanges were focused on more superficial questions and
disclosures. Finally, CMC interactants did not pose any core disclosures
and FtF levels were almost nonexistent as well (MFtF = .01, SD = .11).
The results with respect to Hypothesis 4 are generally supportive. As
opposed to FtF, CMC led to higher proportions of more intimate ques-
tions and lower proportions of peripheral questions, as well as fewer in-
stances of peripheral disclosures.
Uncertainty Reduction Levels in CMC

RQ1 addressed whether FtF interactants have greater attributional con-
fidence than CMC counterparts. In previous research (Walther, 1993), the
willingness of participants to make attributions about their partners was
affected by both time and media. In order to be sensitive to such effects
on attributional confidence, subjects’ postconversation (terminal) assess-
ments of attributional confidence were gathered, along with a retrospec-
tive assessment of initial, preconversation attributional confidence. Re-
search Question 1 asked whether FtF interactants achieve higher levels of
attributional confidence than their mediated counterparts, due to greater
access to conversation cues.
Initial and terminal attributional confidence scores were combined
within each respective communication condition. Comparison of the two
conditions, on the pooled attributional confidence scores, approached sig-
nificance favoring FtF, t (156) = 1.77, .05 > p > .10, two-tailed (MFtF = 8.11,
SD = 18.36, n = 84; MCMC = 3.92, SD = 9.48, n = 74). The terminal scores
alone, however, suggested a possible reversal of this direction favoring
CMC (MFtF = 44.67, SD = 26.71; MCMC = 50.30, SD = 24.31), although the
comparison of terminal scores alone did not approach significance, t(156)
= -1.39, p > .10. With respect to RQ1, these findings suggest that while FtF
interactants had some initial advantage in attributional confidence, per-
haps due to their ability to see their partners and make attributions based
on physical appearance, this advantage disappeared as the conversation
evolved. CMC participants ended conversations feeling just as confident

as FtF interactants, perhaps by making relatively greater gains in
attributional confidence.
A more direct method for assessing this possibility was conducted by
computing scores reflecting the change in attributional confidence over
the course of the conversation, by subtracting initial confidence scores
from terminal confidence scores. Results of the comparison between
change scores showed that the change in attributional confidence tended
to be significantly greater in the CMC condition (M = 46.38, SD = 24.52, n
= 74) than in the FtF condition (M = 36.56, SD = 24.37, n = 84), t(156) =
-2.52, p < .025, two-tailed. Not only did CMC interactants manage to gain
knowledge at the same level as FtF interactants, they made up for initial
deficiencies. The next research question explores the microprocesses un-
derlying such development in greater detail.
Attribution Stimuli
The second research question sought to identify the cues or strategies
upon which CMC and FtF interactants relied in order to reduce uncer-
tainty. In order to address this question, coders categorized individuals’
responses to an open-ended question into ten major categories of conver-
sation elements identified by Tidwell et al. (1995): (1) an attribute of the
partner, (2) an attribute of self, (3) the partner’s conversational perfor-
mance, (4) one’s own conversational performance, (5) the partner’s non-
verbal or paralinguistic expressions, (6) one’s own nonverbal or
paralinguistic expressions, (7) common ground, (8) context, or (9) noth-
ing. The tenth category contained “other” cues that did not fit into one of
the previously mentioned categories.
The conversational performance categories (3 and 4) were further sub-
divided into eight subcategories: (a) asked questions, (b) did not ask ques-
tions, (c) engaged in self-disclosure, (d) did not engage in self-disclosure,
(e) answered questions, (f) did not answer questions, (g) listened, or (h)
performed some other behavior. Common ground (7) was also subdivided
into (a) recognition of the presence of common ground and (b) realization
of the absence of common ground. Context (8) was further segmented
into aspects related to (a) the physical setting, (b) the time, and (c) the
topic/task/goal. A complete listing of these elements and the number of
respondents who identified various cue categories as a means of reduc-
ing uncertainty is shown in Table 2.
Multiple regression analyses were conducted to determine if specific
conversation elements were associated with reduced uncertainty. Each
cue category served as a separate predictor variable, while terminal mea-
sures of attributional confidence served as the outcome measures.
TABLE 2
Frequencies of Uncertainty Reducing and Enhancing Conversation Cues
Uncertainty type
Conversation cue Reducing Enhancing
Attribute of the partner 25 10
Attribute of self 25 9
Partner’s conversational behavior 51 13
Asked questions 8 0
Did not ask questions 0 1
Engaged in self-disclosure 21 1
Did not self-disclose 3 1
Asked questions and disclosed 8 0
Did not ask questions or disclose 0 2
Listened 1 0
Performed some other behavior 10 8
Own conversational behavior 31 12
Asked questions 14 1
Did not ask questions 1 0
Engaged in self-disclosure 5 1
Did not self-disclose 1 1
Asked questions and disclosed 4 0
Did not ask questions or disclose 0 1
Listened 1 0
Performed some other behavior 5 8
Partner’s nonverbal expressions 12 1
Own nonverbal expressions 1 1
Common ground 48 9
Recognition of presence 48 3
Recognition of absence 0 6
Context 52 143
Physical setting 19 80
Time 1 27
Topic/task/goal 32 36
Other 42 56
Nothing 1 4
Missing 1 0
When all major conversation elements were entered in a stepwise analy-

sis on both FtF and CMC participants’ scores, only partner’s lack of self-
disclosure had a significant impact on terminal attributional confidence,
F(1, 156) = 3.96, p < .05, β = -.16. A similar procedure was performed on
participants’ overall attributional confidence, on which physical setting
showed a significant impact, F(1, 156) = 4.89, p < .05, β = -.16. Both cues
had negative effects.
Not surprisingly, analyses of cues by channel condition indicated some
significant differences. FtF interactants reported nonverbal cues (M = .14,
SD = .35) whereas CMC interactants did not identify any paralinguistic
cues. CMC interactants were more likely to note the impact of the situa-
tion (MCMC = .18, SD = .38) than were FtF participants (MFtF = .07, SD =
.26), t(156) = -2.02, p < .05. Finally, FtF and CMC interactants differed in
their assessments of the impact of an attribute of the partner, t(156) = 2.07,
p < .05, or an attribute of oneself, t(156) = 2.07, p < .05, on their decreases
in uncertainty: FtF interactants noted more of an impact of their partner’s
attributes (MFtF = .21, SD = .41; MCMC = .09, SD = .29) and their own at-
tributes (MFtF = .21, SD = .41; MCMC = .09, SD = .29).
Evaluation of cues that were reported to reduce confidence revealed a
significant difference for the role of the physical setting. CMC partici-
pants felt the setting impaired their ability to get to know their partner
(MCMC = .64, SD = .48) more so than did FtF interactants (MFtF= .39, SD =
.49), t(156) = -3.11, p < .01.
With regard to RQ2, participants tended to identify similar cues, with
some slight variations by channel. The finding that CMC interactants felt
the situation impeded their attributional confidence is unsurprising on
one hand, yet notable on the other: One would have expected interactants
to find CMC somewhat arduous given the mechanical requirements of
the medium, yet CMC interactants were able to make greater gains in
attributional confidence. These patterns suggest loose support for the
hyperpersonal effect: Individuals attempted to compensate for perceived
limitations of the channel by engaging in more personalized conversations.
DISCUSSION
CMC has joined FtF communication as a common means by which

individuals communicate with others. The use of email and other forms
of electronic communication raise interesting theoretical questions about
interpersonal processes, not the least of which is how these channels may
cause people to alter or adapt their communication behaviors, methods
of forming acquaintances, processes of forming attributions, and ways of
relating to one another. This investigation sought to illustrate the manner
in which the communication channel impacts micro- and macrolevel pro-
cesses of acquaintanceship development in CMC.
Our findings indicate that the restrictions of CMC prompt users’ ad-
aptation to the medium through modification of uncertainty reduction
behaviors. Bereft of most nonverbal cues, CMC partners forgo the pe-
ripheral questions and answers that mark the normal, superficial ex-
changes among new acquaintances in FtF encounters. Instead, CMC
interactants appeared to employ a greater proportion of more direct, in-
teractive uncertainty reduction strategies—intermediate questioning and
disclosing with their partners—than did their FtF counterparts. The probes
and replies they exchanged were more intimate and led to levels of
attributional confidence similar to their offline counterparts’. According
to their partners, CMC interactants are more effective when engaging these
more intimate exchanges, compared to FtF communicators who act simi-
larly. It seems likely that the increased intimacy of these microlevel be-
haviors may lead to perceptions of extraordinarily affectionate relations,
or hyperpersonal states, as seen in recent studies among long-term CMC
partnerships; while such a theoretical connection would mirror the dy-
namics of FtF relations (Sunnafrank, 1985), that specific connection awaits
direct demonstration.
Theoretical Implications
Several theories are addressed as a result of this investigation. Berger
and Calabrese’s (1975) URT was found to account not only for behavior
in FtF settings, as it originally pertained, but in electronic settings as well.
Both FtF and CMC interactants exhibited efforts to become acquainted
with their conversational partners, and interactants exhibited classic un-
certainty-reducing strategies as they asked questions and performed self-
disclosures with their partners. Demonstrating the effect of the commu-
nication channel on these processes, and the application of the channel’s
characteristics to the politeness/effectiveness trade-offs associated with
strategy selection, results differed between CMC and FtF settings. It is
worthwhile to retest theories as new communication modes present new
theoretical boundary conditions, and as the characteristics of CMC place
certain limits on uncertainty reduction strategies, these results present a
modification of the derivations of URT, attesting to its heuristic value.
CMC partners employed a greater proportion of questions, disclosures,
and more intimate questions than their FtF counterparts, attesting to SIP
theory’s contention (Walther, 1992) that email users adapt their relational
needs and goals into the linguistic codes of CMC.
Early research in CMC indicated that it was a limited channel of com-
munication, ill suited for tasks such as getting to know one another (see
Rice, 1993, for review). Such a position is once again challenged by these
findings. More recent theories have suggested that either CMC interactants
come to know each other individually over time (Walther, 1992) or on the
heuristic basis of assumed ingroup similarities (Lea & Spears, 1995). The
extent to which these impressions are based on personal cues, even highly
selective ones, or based on social cognitive heuristics has become a mat-
ter of theoretical contention, and the present research adds new ballast to
both sides of the question. SIDE theory suggests that in the cue-limited
CMC environment, only social category cues remain, and the assumed
similarities among users’ social category memberships are subject to viv-
idness and overattribution effects (Spears & Lea, 1992). In this study, the
relatively dramatic gains in attributional confidence by CMC partners
might be taken as greater overattribution based on lesser personal infor-
mation, especially since, in one sense, CMC interactants generated and
received much less conversation (i.e., no physical cues and fewer utter-
ances). Given the lesser amount of cues present in CMC, but the same
level of terminal attributional confidence as FtF interactions, the
attributional confidence data appear to support this interpretation.
However, the main contribution of this study—its focus on the content
of CMC and FtF uncertainty reduction strategies—suggests a different
interpretation. SIDE theory emphasizes social cues rather than interper-
sonal cues based on the assumption that discreet, individuating informa-
tion is absent in CMC (Lea & Spears, 1995). Yet the relatively greater pro-
portions of questions and disclosures in CMC, and particularly the greater
depth of CMC questions indicate that while CMC interactants are mak-
ing more attributions from fewer absolute cues, they are efficiently ac-
commodating to the medium by deploying more intimately personal cues.
It seems contrary to SIDE that personal questions and self-disclosures,
offering potentially individuating information, reinforce the presence of
social, and the lack of individual, identity. Such efforts suggest instead
support for a SIP/hyperpersonal view of acquaintance development and
personal knowledge development through information exchange in CMC.
At the same time, the reciprocation of disclosure may function as a group
(dyadic) norm within these associations, and the adherence to norms is con-
sidered a potent process within SIDE. Overall, these results reinforce the
difficulty in untangling hyperpersonal from SIDE effects (see Walther, 1997).
Although the results reinforce SIP’s contentions about the adaptation
of CMC users to the linguistic and content mechanism of the channel
(Walther, 1992), the findings also challenge a dimension of the theory. It
appears that sometimes quality is more important than quantity: One
caveat in SIP is that adequate time must be provided in order to equalize
communication transmissions and that equalized message exchange is
necessary for relational development to be parallel in CMC and FtF con-
texts. Equal message exchanges did not occur across the two channels in
this investigation, with FtF conversations containing many more utter-
ances than the CMC conversations, at least the way they were coded.
Nevertheless, the mediated interactants did not seem to be disadvantaged.
In fact, they were shown to attain significantly greater increases in
attributional confidence during a single conversation than those in the
FtF condition, where multiple cues were present. Such variations were
also noted by Roberts, Smith, and Pollock (1996) in online settings where
socializing and developing relationships were the primary goals. When
people want to get to know one another, they overcome the limitations of
the medium and do so.
Limitations
One problem with the present research pertains to the unequal vol-
ume of utterances between CMC and FtF settings, and the use of propor-
tions rather than frequency data with which to test hypotheses. Although
attempts were made to provide time enough for equal accumulations
across the two channel conditions, FtF interactants generated about twice
as many utterances as their CMC counterparts. This prompted some troubling
issues. First, it suggests that CMC may require even more than four to five
times the amount of time spent in a FtF interaction in order to convey the
same amount of utterances. This level of disparity has not been seen in
previous studies: In time-unlimited tests, Dubrovsky, Kiesler, and Sethna
(1991) and Weisband (1992) found that CMC groups took four to five times
as long, respectively, as FtF groups. In other experiments that provided
unequal time limits in order to equalize communication opportunities,
Siegel et al. (1986) gave FtF groups 10 minutes and CMC, 20; Lebie,
Rhoades, and McGrath (1995/1996) gave FtF groups 15 minutes and CMC,
20. This underlying ratio may be at the heart of many studies showing
differences between CMC and FtF in experiments that imposed equal
amounts of time on both conditions, and that may now be considered
anomalous (see Walther et al., 1994). Alternatively, the present ratio could
indicate that these participants were particularly slow in their use of email
(which seems unlikely, given their experience levels), or that the units
that coders identified as single comments contained more verbiage when
they had been created by typing, and therefore appeared to be fewer. Re-
gardless of the cause, the inability to equalize communication efforts made it
difficult to assess differences in questions and self-disclosures. Use of pro-
portions of conversation elements controlled for volume differences, but
this practice established interdependencies that made detection of differ-
ences in use of strategies difficult. In order for a change in one area to be
detected, it required a corresponding decrease in one or more other areas.
Use of frequencies would have been more desirable, so that changes in
one conversation element would be independent of other changes. Hope-
fully, future research can equalize the two channels more successfully in
order to delineate the similarities and differences at the utterance level.
Another concern emanating from this disparity is the possibility that
CMC participants felt rushed or time-pressured, which is one of the very
problems for which the inequality in time periods was implemented. With-
out ample time, one could surmise that CMC partners felt rushed and
therefore employed more intimate and effective uncertainty reduction
strategies in order to compensate for the relative lack of time within the
medium, obfuscating our ability to determine whether the observed inti-
macy was actually the result of time rather than medium, or an interac-
tion effect of the two. While such a potential confound is an interesting
possibility, it would contradict a growing amount of empirical research
showing that time limits and time pressure in CMC lead to less affinity,
more confrontation, and less social orientation than when time is unlim-
ited (e.g., Reid, Ball, Morley, & Evans, 1997; Reid, Malinek, Stott, & Evans,
1996; Walther et al., 1994), and the current results, as predicted in several
hypotheses, indicate that such affinity was not hampered (at least termi-
nally); in other words, the results appear robust with respect to the prob-
lem of message inequality. While we cannot rule out that the possibility
that the extra time fatigued the CMC participants, it is unclear how fa-
tigue would have led to the intimacy effects they produced; neither can
we rule out or confirm the possibility that they were enjoying themselves.
CONCLUSIONS
One of the goals of this research was to investigate the impact of a

number of communication phenomena in computer-mediated settings.
The results demonstrated that interactants use standard means of uncer-
tainty reduction with slight modifications to accommodate channel dif-
ferences. Therefore, many common tenets of uncertainty reduction theory
may be extended to CMC, with the understanding that at least one stimu-
lus for uncertainty reduction is similar in CMC and FtF communication.
However, the means of reduction differ somewhat.
This investigation also provided further support for the notion that
mediated interactants can and will develop personalized relationships
characterized by intimacy, as suggested by SIP (Walther, 1992) and that
individuals’ reactions to the limitations of the media prompt them to over-
compensate (cf. Utz, 2000). Individuals attempted to compensate for the
limitations of CMC by hyperpersonalizing their interactions, presumably
leading to accelerated gains in attributional confidence over the course of
their brief conversations. Although numerous reports depict users’ projec-
tive impressions of CMC as an inadequate or inefficient means of getting to
know someone, CMC does not appear to hinder uncertainty reduction,
merely perceptions of ability to do so. As a result of its use, more effective
communicators exchange more intimate questions and disclosures than they
would in similar FtF contexts, and acquaintanceship develops in CMC as it
does FtF. CMC does not suffer in this comparison. CMC interactants work
to overcome limitations of the channel as they get to know one another.
NOTES
1. In the original design of this research, hypotheses were considered regarding possible
effects of anticipated future interaction on uncertainty reduction strategies and outcomes
(following Walther, 1994). Accordingly, some dyads were induced to believe that the con-
versation would be the first of several with a particular partner (even though only one
meeting actually took place), whereas others were encouraged to believe that this would be
their only interaction together. A single-item manipulation check was used to assess partici-
pants’ perceptions of the probability of future interaction with their partner: “What is the
likelihood that you will talk with your conversational partner again? Indicate the number
as a percentage from 0% to 100%.” The mean rating regarding a future meeting was 38.5 for
the one-shot interaction condition and 35.6 for the anticipated future interaction condition.
Anticipation did not differ within the FtF dyads, t (82) = .09, p = .93 or the CMC dyads, t (72)
= .68, p = .50. Similar to previous research concerning anticipated future interaction (Dou-
glas, 1987; Kellermann, 1986; Walther, 1994), this investigation was unable to manipulate
anticipated future interaction successfully, and as a result, this variable was dropped from
further analysis.
2. A number of disclosure rating schemes were considered, conflicts among which led to
modifications in the coding procedure. For instance, topic categorization schemes for self-
disclosure, such as those used by Jourard (1971) or Rubin and Shenker (1978) have only
been used in the development of questionnaire items rather than as content analytic ap-
proaches. A comparison of Taylor and Altman’s (1966) system with Jourard’s (1971) and
Rubin and Shenker’s (1978) revealed a number of blurred classification boundaries, which
were modified here so that functional and mutually exclusive categories could be created.
The first problem is that attitudes, values, and opinions together form a single topic cat-
egory on each list. However, these items do not constitute topic categories, but cut across all
topics rather than forming a single topic, i.e. attitudes are held on topics, such as family,
appearance, money, or current events, but are not a topic in and of themselves. Altman and
Taylor (1973) recognize this when they later use attitudes and opinions as a layer depicting
intermediate intimacy rather than a topic category. Thus categories relating to attitudes,
values, or opinions were removed from the Taylor and Altman scheme. A similar concern
arises from the categorization of feelings and emotions, which again does not meet the topic
construct, since emotions and feelings are possessed relative to specific topics such as self,
others, events, etc. This category was subsumed into all other categories as well.
Next, love and sex are grouped together by Taylor and Altman (1966) and Jourard (1971),
but Rubin and Shenker (1978) separate sex from emotion. Love is defined as an emotion or
way of relating to another person. So there is some conceptual overlap with categories that
deal with relationships with others and categories that deal with feelings and emotions.
This resulted in sex being placed in its own category. While Jourard (1971) places sexual
performance under the body category, he places facts about one’s present sex life in the
personality category, which is conceptually confusing. In our approach, sex and love were
separated, with love in a single category concerned with feelings toward others and utter-
ances about one’s sexual performance or sexual history contained in a single other category.
Distinctions among biography, current, and historical events must also be clarified. Re-
ligion is treated as a demographic/biographic factor, despite Taylor and Altman (1966) con-
sidering religion as one category and biography as another. Views on religion, as they relate
to one’s own religion, are best dealt with via demographic/biographic information. In con-
trast, views on religious persecution would best be categorized under current events. Cur-
rent events quickly become historical when a “new” issue preempts them. So it makes sense
to expand current events to include both current and historical events that one is not per-
sonally relating to self or to one’s own experience (that would be biography).
Further, regarding biography, Taylor and Altman (1966) separated one’s own family and
one’s parental family from biography. However, since (a) one’s biography does include the
topics of one’s own family and parental family, and (b) family relationships also dictate a
way of relating to others, information about the composition of one’s family was consid-
ered biographic, whereas information about the quality of the family ties would be included
under relationships with others.
Finally, while self-concept is typically considered to be a part of one’s biography, Rubin
and Shenker (1978) attempted to combine self-concept with interpersonal relationships. While
one’s self-concept can be shaped by, or drive behaviors and expectations within a relationship,
there is mixed opinion on its independence from relationships. Conversationally it is topi-
cally separate from relationships, however, and was treated within the biographic topic category.
3. A check for nonindependence of dyad scores was conducted on the results. Kenny
(1995) has demonstrated that in data from both members of a dyad, the scores are likely to
be correlated and correlations of greater than .3 result in biased p values in significance
testing. Dyadic scores in the present inquiry revealed that dyadic correlations between un-
certainty and proportion of questions and disclosures were correlated at levels less than .3,
and analyses continued using between-subjects designs. Technical problems interfered with
the recording and subsequent analysis of some CMC dyads, which accounts for the un-
equal n in the following analyses.
4. T-tests employing frequency data did not reveal these patterns. Indeed, significant
differences were obtained in the opposite direction when examining the number of ques-
tions, MFtF = 12.10, SD = 10.69, MCMC = 6.88, SD = 5.45; and the number of self-disclosures,
MFtF = 51.15, SD = 38.53, MCMC = 30.19, SD = 21.29; while the difference in other comments
still favored FtF to a great degree, MFtF = 24.94, SD = 20.84, MCMC = 5.86, SD = 5.40. All of
these patterns mirrored the baseline difference in the mean number of comments between the
communication conditions (87:43), and demonstrate the value of the proportional analysis.
5. Berger and Kellermann (1983) found that question-asking was associated with appro-
priateness, but that the relationship of question-asking to efficiency was moderated by the
extent to which participants were perceived to be high or low in information-seeking. In
order to detect curvilinear relationships for disclosure and questions with evaluations of
appropriateness or effectiveness, quadratic terms for these utterance types, and their pro-
portions, were computed and entered into regression analyses. No significant relationships
were obtained for any analyses of the CMC data. Within the FtF condition, the quadratic
transformation for the proportion of questions showed a significant negative zero-order
relationship with conversational effectiveness, r = -.31, p = .004; there was also a significant
zero-order relationship between the curvilinear representation of effectiveness and the pro-
portion of other comments, r = .31, p = .005. However, multiple regression analyses of the
change in R2 for effectiveness when the quadratic terms were added to the linear terms
indicated that the deviation from linearity was not significant: Fchange (1, 80) = 2.31, p = .13 for
questions, and Fchange (1, 80) = .289, p = .59 for other comments. Hypothesis tests continued
by comparing linear correlation coefficients between FtF and CMC conditions.
REFERENCES
Altman, I., & Taylor, D. A. (1973). Social penetration: The development of interpersonal relation-
ships. New York: Holt, Rinehart, and Winston.
Baker, A. (1998, July). Cyberspace couples finding romance online then meeting for the first
time in real life. CMC Magazine. Retrieved December 1, 2000, from http://
www.december.com/cmc/mag/1998/jul/baker.html
Bales, R. F. (1950). A set of categories for the analysis of small group interaction. American
Sociological Review, 15, 257–263.
Berger, C. R. (1979). Beyond initial interaction: Uncertainty, understanding, and the devel-
opment of interpersonal relationships. In H. Giles & R. St. Clair (Eds.), Language and
social psychology (pp. 122–144). Oxford, UK: Basil Blackwell.
Berger, C. R., & Bradac, J. J. (1982). Language and social knowledge: Uncertainty in interpersonal
relationships. London: Edward Arnold.
Berger, C. R., & Calabrese, R. J. (1975). Some explorations in initial interaction and beyond:
Toward a developmental theory of interpersonal communication. Human Communica-
tion Research, 1, 99–112.
Berger, C. R., & Douglas, W. (1981). Studies in interpersonal epistemology: Anticipated in-
teraction, self-monitoring, and observational context selection. Communication Mono-
graphs, 48, 183–196.
Berger, C. R., Gardner, R. R., Parks, M. R., Schulman, L., & Miller, G. R. (1976). Interpersonal
epistemology and interpersonal communication. In G. R. Miller (Ed.), Explorations in
interpersonal communication (pp. 149–171). Beverly Hills, CA: Sage.
Berger, C. R., & Kellermann, K. A. (1983). To ask or not to ask: Is that a question? In R. N.
Bostrom (Ed.), Communication yearbook 7 (pp. 342–368). Beverly Hills, CA: Sage.
Brown, P., & Levinson, S. C. (1987). Politeness: Some universals in language usage (Rev. ed.).
Cambridge, UK: Cambridge University Press.
Burgoon, J. K., Bonito, J. A., Bengtsson, B., Cederberg, C., Lundeberg, M., & Allspach, L.
(2000). Interactivity in human-computer interaction: A study of credibility, understand-
ing, and influence. Computers in Human Behavior, 16, 553–574.
Burgoon, J. K., & Hale, J. L. (1984). The fundamental topoi of relational communication.
Communication Monographs, 51, 193–214.
Burgoon, J. K., & Hale, J. L. (1987). Validation and measurement of the fundamental themes
of relational communication. Communication Monographs, 54, 19–41.
Calabrese, R. J. (1975). The effects of privacy and probability of future interaction on
initial interaction patterns. Unpublished doctoral dissertation, Northwestern
University, Evanston, IL.
Canary, D. J., & Spitzberg, B. H. (1987). Appropriateness and effectiveness perceptions of
conflict strategies. Human Communication Research, 14, 93–118.
Chelune, G. J. (1975). Studies in behavioral and self-report assessment of self-disclosure. (Doc-
toral dissertation, University of Nevada, 1975). Dissertation Abstracts International, 37, 453b.
Clatterbuck, G. W. (1979). Attributional confidence and uncertainty in initial interaction.
Human Communication Research, 5, 147–157.
Culnan, M. J., & Markus, M. L. (1987). Information technologies. In F. M. Jablin, L. L. Putnam,
K. H. Roberts, & L. W. Porter (Eds.), Handbook of organizational communication: An inter-
disciplinary perspective (pp. 420–443). Newbury Park, CA: Sage.
Dindia, K. (1983). Reciprocity of self-disclosure: A sequential analysis. In M. Burgoon (Ed.),
Communication yearbook 6 (pp. 506–528). Beverly Hills, CA: Sage.
Donath, J. (1999). Identity and deception in the virtual community. In M. A. Smith & P.
Kollock (Eds.), Communities in cyberspace (pp. 29–59). New York: Routledge.
Douglas, W. (1987). Question-asking in same- and opposite-sex initial interactions: The ef-
fects of anticipated future interaction. Human Communication Research, 14, 230–245.
Douglas, W. (1990). Uncertainty, information seeking, and liking during initial interaction.
Western Journal of Speech Communication, 54, 66–81.
Dubrovsky, V. J., Kiesler, S., & Sethna, B. N. (1991). The equalization phenomenon: Status
effects in computer-mediated and face-to-face decision-making groups. Human-Computer
Interaction, 6, 119–146.
Gale, S. (1991). Adding audio and video to an office environment. In J. M. Bowers & S. D.
Benford (Eds.), Studies in computer-supported collaborative work (pp. 49–62). New York:
Elsevier Science Publishers.
Grice, P. (1975). Logic and conversation. In P. Cole & J. Morgan (Eds.), Syntax and semantics:
Speech acts (Vol. 3, pp. 41–58). New York: Academic Press.
Guetzkow, H. (1950). Unitizing and categorizing problems in coding qualitative data.
Journal of Clinical Psychology, 6, 47–58.
Hale, J. L., Lundy, J. C., & Mongeau, P. A. (1989). Perceived relational intimacy and rela-
tional message content. Communication Research Reports, 6, 94–99.
Hancock, J. T., & Dunham, P. J. (2001). Impression formation in computer-mediated com-
munication revisited: An analysis of the breadth and intensity of impressions. Commu-
nication Research, 28, 325–347.
Hecht, M. L. (1978). The conceptualization and measurement of interpersonal communica-

tion satisfaction. Human Communication Research, 4, 253–264.
Hiltz, S. R., Johnson, K., & Turoff, M. (1986). Experiments in group decision making: Com-
munication process and outcome in face-to-face versus computerized conferences. Hu-
man Communication Research, 13, 225–252.
Holsti, O. R. (1969). Content analysis for the social sciences and humanities. Reading, MA:
Addison-Wesley.
Jacobson, D. (1999). Impression formation in cyberspace: Online expectations and offline experi-
ences in text-based virtual communities. Journal of Computer-Mediated Communication, 5(1). Re-
trieved May 18, 2000, from http://www.ascusc.org/jcmc/vol5/issue1/jacobson.html
Joinson, A. N. (2001). Self-disclosure in computer-mediated communication: The role of
self-awareness and visual anonymity. European Journal of Social Psychology, 31, 177–192.
Jourard, S. M. (1971). The transparent self (Rev. ed.). New York: Van Nostrand Rheinhold.
Kellermann, K. (1986). Anticipation of future interaction and information exchange in ini-
tial interaction. Human Communication Research, 13, 41–75.
Kellermann, K. A., & Berger, C. R. (1984). Affect and social information acquisition: Sit back,
relax, and tell me about yourself. In R. N. Bostrom (Ed.), Communication yearbook 8 (pp.
412–445). Beverly Hills, CA: Sage.
Kenny, D. A. (1995). The effect of nonindependence on significance testing in dyadic re-
search. Personal Relationships, 2, 67–75.
Kiesler, S. (1986, January-February). The hidden messages in computer networks. Harvard
Business Review, pp. 46–54, 58–60.
Kiesler, S., Siegel, J., & McGuire, T. (1984). Social psychological aspects of computer-medi-
ated communication. American Psychologist, 39, 1123–1134.
Kraut, R., Mukhopadhyay, T., Szczypula, J., Kiesler, S., & Scherlis, B. (1999). Information
and communication: Alternative uses of the Internet in households. Information Sys-
tems Research, 10, 287–303.
Lea, M., & Spears, R. (1992). Paralanguage and social perception in computer-mediated
communication. Journal of Organizational Computing, 2, 321–341.
Lea, M., & Spears, R. (1995). Love at first byte? Building personal relationships over com-
puter networks. In J. T. Wood & S. Duck (Eds.), Under-studied relationships: Off the beaten
track (pp. 197–233). Thousand Oaks, CA: Sage.
Lebie, L., Rhoades, J. A., & McGrath, J. E. (1995/1996). Interaction processes in computer-
mediated and face-to-face groups. Computer Supported Cooperative Work, 4, 127–152.
McCroskey, J. C., & McCain, T. A. (1974). The measurement of interpersonal attraction. Speech
Monographs, 41, 261–266.
Morris, W. (Ed.). (1976). The American heritage dictionary of the English language. Boston:
Houghton Mifflin.
Olaniran, B. A. (1996). A model of group satisfaction in computer-mediated communication
and face-to-face meetings. Behaviour and Information Technology, 15, 24–36.
Parks, M. R., & Floyd, K. (1996). Making friends in cyberspace. Journal of Communication,
46(1), 80–97.
Parks, M. R., & Roberts, L. D. (1998). Making MOOsic: The development of personal rela-
tionships on line and a comparison to their off-line counterparts. Journal of Social and
Personal Relationships, 15, 517–537.
Postmes, T., Spears, R., & Lea, M. (1998). Breaching or building social boundaries? SIDE-
effects of computer-mediated communication. Communication Research, 25, 689–715.
Postmes, T., Spears, R., & Lea, M. (2000). The formation of group norms in computer-medi-
ated communication. Human Communication Research, 26, 341–371.
Ramirez, A., Jr., Walther, J. B., Burgoon, J. K., & Sunnafrank, M. (2002). Information seek-
ing strategies, uncertainty, and computer-mediated communication: Toward a con-
ceptual model. Human Communication Research, 28, 213–228.
Reid, F. J. M., Ball, L. J., Morley, A. M., & Evans, J. S. B. T. (1997). Styles of group discussion
in computer-mediated decision making. British Journal of Social Psychology, 36, 241–262.
Reid, F. J. M., Malinek, V., Stott, C., & Evans, J. S. B. T. (1996). The messaging threshold in
computer-mediated communication. Ergonomics, 39, 1017–1037.
Rice, R. E. (1993). Media appropriateness: Using social presence theory to compare tradi-
tional and new organizational media. Human Communication Research, 19, 451–484.
Rice, R. E., & Contractor, N. S. (1990). Conceptualizing effects of office information systems:
A methodology and application for the study of alpha, beta, and gamma changes. Deci-
sion Sciences, 21, 301–317.
Roberts, L. D., Smith, L. M., & Pollock, C. (1996, September). A model of social interaction via
computer-mediated communication in real-time text-based virtual environments. Paper pre-
sented at the meeting of the Australian Psychological Society, Sydney.
Ross, L. (1977). The intuitive psychologist and his shortcomings: Distortions in the attribu-
tion process. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 10, pp.
173–220). New York: Academic Press.
Rubin, Z., & Shenker, S. (1978). Friendship, proximity, and self-disclosure. Journal of
Personality, 46, 1–22.
Siegel, J., Dubrovsky, V., Kiesler, S., & McGuire, T. W. (1986). Group processes in computer-
mediated communication. Organizational Behavior and Human Decision Processes, 37, 157–187.
Snedecor, G. W., & Cochran, W. G. (1989). Statistical methods (8th ed.). Ames: Iowa
State University.
Spears, R., & Lea, M. (1992). Social influence and the influence of the “social” in computer-
mediated communication. In M. Lea (Ed.), Contexts of computer-mediated communication
(pp. 30-65). London: Harvester-Wheatsheaf.
Spears, R., & Lea, M. (1994). Panacea or panopticon? The hidden power in computer-medi-
ated communication. Communication Research, 21, 427–459.
Sproull, L., & Kiesler, S. (1991). Connections: New ways of working in the networked organiza-
tion. Cambridge, MA: MIT Press.
Srull, T. K., & Wyer, R. S., Jr. (1989). Person memory and judgment. Psychological Review, 96, 58–83.
Sunnafrank, M. (1985). Attitude similarity and interpersonal attraction during early com-
municative relationships: A research note on the generalizability of findings to opposite-
sex relationships. Western Journal of Speech Communication, 49, 73–80.
Tardy, C. H. (Ed.). (1988). A handbook for the study of communication: Methods and instruments
for observing, measuring, and assessing communication processes. Norwood, NJ: Ablex.
Taylor, D. A., & Altman, I. (1966). Intimacy-scaled stimuli for use in studies of interpersonal
relations. Psychological Reports, 19, 729–730.
Tidwell, L. C., Douglas, W., & Banski, M. (1995, May). Uncertainty reduction during initial
interaction: How do people get to know each other? Paper presented at the annual meeting of
the International Communication Association, Albuquerque, NM.
Utz, S. (2000). Social information processing in MUDs: The development of friendships in
virtual worlds. Journal of Online Behavior, 1(1). Retrieved April 7, 2000, from http://
www.behavior.net/JOB/v1n1/utz.html
Van Gelder, L. (1985, October). The strange case of the electronic lover. Ms. Magazine. Re-
ported in C. Dunlop & R. Kling (Eds.) (1991), Computerization and controversy: Value con-
flicts and social choices (pp. 364–375). Boston: Academic Press.
Walther, J. B. (1992). Interpersonal effects in computer-mediated interaction: A relational
perspective. Communication Research, 19, 52–90.
Walther, J. B. (1993). Impression development in computer-mediated interaction. Western
Journal of Communication, 57, 381–398.
Walther, J. B. (1994). Anticipated ongoing interaction versus channel effects on relational
communication in computer-mediated interaction. Human Communication Re-
search, 20, 473–501.
Walther, J. B. (1996). Computer-mediated communication: Impersonal, interpersonal, and
hyperpersonal interaction. Communication Research, 23, 1–43.
Walther, J. B. (1997). Group and interpersonal effects in international computer-mediated
collaboration. Human Communication Research, 23, 342–369.
Walther, J. B., Anderson, J. F., & Park, D. W. (1994). Relational tone in computer-mediated
communication: A meta-analysis of social and anti-social communication. Communi-
cation Research, 21, 460–487.
Walther, J. B., & Burgoon, J. K. (1992). Relational communication in computer-mediated
interaction. Human Communication Research, 19, 50–88.
Walther, J. B., & D’Addario, K. P. (2001). The impacts of emoticons on message interpreta-
tion in computer-mediated communication. Social Science Computer Review, 19, 323–345.
Walther, J. B., Slovacek, C., & Tidwell, L. C. (2001). Is a picture worth a thousand words?
Photographic images in long term and short term virtual teams. Communication Research,
28, 105–134.
Walther, J. B., & Tidwell, L. C. (1995). Nonverbal cues in computer-mediated communica-
tion, and the effect of chronemics on relational communication. Journal of Organizational
Computing, 5, 355–378.
Weisband, S. P. (1992). Group discussion and first advocacy effects in computer-mediated
and face-to-face decision making groups. Organizational Behavior and Human Decision
Processes, 53, 352–380.
Wheeless, L. R. (1978). A follow-up study of the relationships among trust, disclosure, and
interpersonal solidarity. Human Communication Research, 4, 143–157.
Whittaker, S., & O’Conaill, B. (1997). The role of vision on face-to-face and mediated com-
munication. In K. E. Finn, A. J. Sellen, & S. B. Wilbur (Eds.), Video-mediated communication
(pp. 23–49). Mahwah, NJ: Erlbaum.
Worthy, M., Gary, A. L., & Kahn, G. M. (1969). Self-disclosure as an exchange process. Jour-
nal of Personality and Social Psychology, 13, 59–63.
Zuckerman, M., DePaulo, B. M., & Rosenthal, R. (1981). Verbal and nonverbal communica-
tion of deception. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol.
14, pp. 1–59). New York: Academic Press.
View publication stats

Computer-Mediated Communication Effects On Disclos PDF

Uploaded by

Copyright:

Available Formats

Computer-Mediated Communication Effects On Disclos PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Computer-Mediated Communication Effects On Disclos PDF

Uploaded by

Copyright:

Available Formats

See

discussions, stats, and author profiles for this publication at:

Article in Human Communication Research · July 2002

The user has requested enhancement of the downloaded file.

Computer-Mediated Communication Effects

LISA COLLINS TIDWELL

This investigation examined how computer-mediated communication (CMC) partners ex-

Human Communication Research, Vol. 28 No. 3, July 2002 317–348

another? As important as the burgeoning of the Internet is the growth of

CMC Theories and Evidence

and temporal characteristics may prompt CMC users to engage in selec-

Uncertainty Reduction: Strategies and Limitations

ated with accompanying methodologies for examining the behavioral

are known to the observer, and drawing comparisons based on similarity

All of the interactive strategies suggested by Berger and colleagues

H1: In initial interactions, computer-mediated interactants deploy a greater

Reconciling Efficiency With Politeness

another in CMC? It is a rule in FtF communication that in order to appear

Although questions are efficient, purposeful, and easily understood in

are judged to be more effective by their partners; whereas, when FtF

H4: CMC interactants engage in deeper self-disclosures than do FtF

In addition to leaving open the role of impression-facilitating behav-

development of personal knowledge structures of one’s partners. In con-

RQ1: Do FtF interactants have greater attributional confidence than their

The Nature of Cues in Impression Formation Online

but since 15 minutes in a FtF setting is not equivalent to 15 minutes in

sure retrospectively is somewhat unorthodox, yet methodological con-

Uncertainty Reduction Strategies in CMC

-4.98, p < .001. FtF interactants, in contrast, displayed a significantly greater

Perceived Proportion of Proportion of Proportion Total self- Total Total

Perceived FtF .14 .12 -.05 -.11 .12 .10 .01

Perceived FtF -.12 -.28* .33** -.05 -.12 .09

Proportion of FtF -.54** -.79** .20 -.31** -.39**

Proportion FtF -.06 -.00 .56** .07

Proportion FtF -.17 -.00 .46**

Total FtF .63** .69**

Total questions FtF .62**

sures (nor with quantities of either questions or disclosures). Interestingly,

Accelerated Intimacy in CMC

cantly higher amount of peripheral questions (MFtF = 8.04, SD = 8.87; MCMC

Uncertainty Reduction Levels in CMC

evolved. CMC participants ended conversations feeling just as confident

When all major conversation elements were entered in a stepwise analy-

CMC has joined FtF communication as a common means by which

One of the goals of this research was to investigate the impact of a

Hecht, M. L. (1978). The conceptualization and measurement of interpersonal communica-

View publication stats

You might also like

Proportion of FtF -.54 -.79 .20 -.31 -.39

Total FtF .63 .69