You are on page 1of 11

Computers in Human Behavior, Vol. 4, pp. 311-321, 1988 0747-5632/88 $3.00 + .

00
Printed in the U.S.A. All rights reserved. Copyright 1989 Pergamon Press plc

The Effects of Computer Mediated


Communication on an
Individual's Judgment:
A Study Based on the Methods of Asch's
Social Influence Experiment

Michael Smilowitz
The Ohio University

D. Chad Compton
University of Missouri

Lyle Flint
Ball State University

Abstract -- Because of the importance placed on social influence in the group process, this study inves-
tigates how the exclusion of the contextual cues provided by face-to-face interaction influences individual
judgments in computer mediated contexts. The study employs a computerized revision of the classic proce-
dures of Asch's "majority against a minority of one" investigation. The results of the study indicate that
computer mediated communication is likely to diminish the effects of social pressures to conform to majority
judgments.

The rapid technological changes in the ways people communicate are an appeal-
ing area for social science investigation. As technology changes the ways that people
interact there may develop fundamental differences in the processes by which peo-
ple accomplish their communicative tasks.
Computer mediated communication is an area of particular interest because of
the rapid rate by which organizations are adopting the technology. A n u m b e r of
descriptive studies (Danowski, 1982; Fergeson, 1977; Hihz, Turoff & Johnson,
1985; Johansen & DeGrase, 1979; Phillips, 1983; Rice, 1982, 1987) suggest that

Requests for reprints should be addressed to Michael Smilowitz, School of Interpersonal Commu-
nication, The Ohio University, Athens, OH 45701.
311
312 Smilowitz, Compton, and Flint

computer mediated conferencing ( C M C ) is a satisfactory alternative to face-to-face


communication for a variety of organizational communicative activities. In par-
ticular, C M C is regarded as a useful device for group problem solving. C M C is
attractive to organizations because of its cost and time saving advantages. As C M C
becomes increasingly part of the communication practices of organizations there
is increasing need to understand how the medium alters group interaction pro-
cesses, especially decision making processes.
The absence of the contextual cues available in face-to-face interaction is ex-
pected to have some impact on C M C group communication (Edinger & Patter-
son, 1983). The computer terminal simply does not provide the same type of
information that physical proximity provides for the participants of a group (Kor-
zenny, 1978; Short, Williams, & Christie, 1976). The differences in the type and
amount of potentially communicative information should be important (Weick,
1985). If not, textbooks on group decision making have for too long given unneces-
sary attention to nonverbal interaction as a significant contributor to the commu-
nication process.
The precise ways in which computer mediated communication can be expected
to differ from face-to-face interaction remain inconclusive primarily because
research in the area is still sparse. This study examines one particular area of group
processes which can be expected to differ in computer mediated contexts. The pres-
ence of others is recognized to have definite influence on individual judgments.
Because of the importance placed on social influence in the group process, this
study investigates how the exclusion of the contextual cues provided by face-to-face
interaction influences individual judgments in computer mediated contexts.

REVIEW OF THE LITERATURE

A body of literature on the effects of C M C on individual judgments is beginning


to accumulate. A consistent claim in the literature is that since the physical pres-
ence of others is excluded, individuals may be expected to be less inhibited and
more task oriented in face-to-face situations (Hiltz, Johnson, & Turoff, 1986).
One of the expected consequences of the task orientation of C M C is that par-
ticipants feel less group belongingness. The reasoning behind this proposition
assumes that the impersonal nature of C M C group processes results in participants
identifying less with the group than they would in face-to-face situations. The
assumption is supported by the literature that suggests that C M C is less friendly
or less socially oriented than face-to-face interaction (Champanis, 1975; Danowski,
1982; Hiemstra, 1982; Hiltz, 1975; Hiltz, Johnson, & Turoff, 1986; Phillips,
1983).
The literature on C M C also proposes that it is much easier for a group mem-
ber to hold out against the other members of the group. There is less certain evi-
dence for this claim because the research has only tangentially examined the social
pressures of C M C . Hiltz, Johnson, and Mathews (1978) allude to the absence of
nonverbal behaviors as a basis for expecting positions to be held more strongly.
Kiesler, Siegel, and McGuire (1984) argue for less social pressure in C M C because
of the increased occurrences of uninhibited behaviors. For Kiesler et al., unin-
hibited behavior is defined as the frequency of flaming comments; that is, utter-
ances in which there is swearing, insults, name calling, or other statements that
CMC effects on individual judgments 313

appear as hostile. Kiesler et al. conclude that the increases in uninhibited comments
indicate decreases in social pressure and thus individual members can be expected
to make more independent judgments.
Flaming comments, however, are not a direct nor adequate measure of the will-
ingness of group members to succumb to group pressures. Hiltz, Turoff, and John-
son (1985) also examined flaming comments. Their research reported a very low
level of disinhibited behaviors in computer conferences of middle-managers, but
nevertheless a high degree of independence.
The difference between the findings of Kiesler's et al. and findings of Hiltz et
al. suggest that "flaming comments" have different consequences for different
groups. The differences result from each groups' own normative structure. The
subjects of the Kiesler et al. study may have had cultural norms that sanctioned
flaming comments when communicating via computers, whereas Hiltz et al. stud-
ied individuals who may have had cultural norms against such behaviors in any
context. Different groups develop their own different rules by which hostile com-
ments and deviant behaviors are used and understood by the members (Fisher,
1980). If the group has norms which approve the presence of such comments, then
their presence does little to affect the selection of solutions. If group norms disap-
prove of such behaviors, then their occurrence can be expected to be minimal. Indi-
viduals who persist with their flaming comments may be regarded as deviant and
excluded from the discussion. Both conditions, therefore, result in flaming com-
ments having little impact on the group's decision processes.
There remains a need to more completely test the C M C environment for its
effects individuals' judgments. The presence of a group exerting social influence
is recognized as able to affect an individual's critical judgments. If, as Champanis
(1975) claims, C M C is more egalitarian than face-to-face interaction, it follows that
in C M C there would be less pressure to conform since participation is more equal-
ized among all group members.
Further reason to anticipate less pressure to conform comes from Hiltz (1975)
hypothesis that the remoteness and impersonality of C M C would allow participants
to feel free to express disagreement or suggest potentially unpopular ideas. Hiltz's
position may account for Furgeson's (1977) observation that C M C group mem-
bers are able to express their ideas more completely. Both positions indicate the
role of social influence on making individual judgments.
A comprehensive review and critique of social influence theories is provided by
Smith (1982). In brief, it was the psychologists and social psychologists of the times
who identified the effects others have on the formation and maintenance of an indi-
vidual's judgments. Festinger (1954) claimed that individuals have a drive to estab-
lish support for their standards and beliefs. Accordingly, individuals are influenced
by the judgments of others, and through a process of evaluation, internalize beliefs
and positions which appear as socially valid.
One of the classic studies in social influence was that of Asch (1956). In a
remarkably elegant experimental design, Asch demonstrated the influence of
majority opinion on an individual's opinion. In the experiment, subjects was asked
to make comparisons regarding the length of lines placed on cardboard cards. To
manipulate social influence, Asch had confederates make erroneous comparisons.
During the experiment, the confederates were instructed to periodically unani-
mously pick the wrong line. The confederates announced their decisions and sub-
jects sometimes agreed with the obviously incorrect judgments of the confederates.
314 Smilowitz, Compton, and Flint

Since the fifties, the study of social influence has been greatly refined. But, for
the purposes of this study, Asch's experimental design remains a viable procedure
to investigate the influence of group opinion on individual judgments. Asch's
method easily lends itself to a replication by programming a computerized pseudo-
conference. By communicating the judgments of the majority only through C M C ,
it is possible to have a more direct measure of the majority's social influence on
an individual's judgments. The following study was therefore designed to answer
the question:

Q I : Will individuals confronted with majority opinions which are in error make in-
correct judgments about the length of lines displayed on a computer terminal?

METHOD

Subjects
The 45 subjects for this study were drawn from the fundamental public speaking
courses at a midwestern university. Included in the sample were 19 males and 23
females. Nineteen point three was the mean age of the subjects, with 57 % fresh-
men and 33.3 % sophomores. Eighty-three point three percent of the subjects had
spent no time with computers on a weekly basis.

Procedures
The procedures for this experiment used the methods of Asch's original study as
closely as possible. In this study, however, a computer program substituted for the
physical presence of the experimenter and the confederates.
Five subjects at a time were brought into a microcomputer laboratory and asked
to sit in front of a terminal. The terminals were so arranged that the subjects could
not see each other's screens and the terminals interfered with the subjects' abili-
ties to look at one another for clues regarding their choices.
The subjects were told that they were participating in a study about computers
and that all the necessary instructions would be presented on their screens once
they were all connected to the "network." The "lab attendant" went to each sta-
tion and entered "instructions" to "hardware" the terminal to tl~e "central network."
Then each subject was asked to enter her or his name. The "network" was a fab-
rication. The machines were not connected. Instead, each machine was pro-
grammed to simulate a computer network. Prior to the subjects entering the lab,
the lab attendant went to each station and entered the names of the five subjects.
W h e n the lab attendant was finished making the "connection" the subjects then
saw the screen depicted in Figure 1.
The subjects were informed, through the screen depicted in Figure 2, that the
"purpose" of the study was to investigate their judgments about the length of lines
displayed on terminals. To encourage the subjects to accept the ruse, they were
asked to input a three digit number so that their station number could be randomly
assigned. It made no difference what number they entered, all subjects were
assigned station number 5 (see Figure 3).
Asch's original investigation presented the subject with 18 trials. Asch placed his
lines on cardboard signs and held them in front of the group. In six trials, the con-
federates responded correctly. In six trials the confederates made moderate errors,
CMC effects on individual judgments 315

Welcome, and t h a n k you for participating with this research.


You will be inter-acting with four other people to make discriminations about the
length of lines. Your computer is networked with four other computers through a
central computer located in X X X X X X Hall. The network looks something like
this:

FRED JOHN

\ J
CENTRAL COMPUTER ~ KATHY

/
JILL JIM

The central computer will store the individual responses so that you m a y work at
your own pace.

Press space bar to continue.

Figure 1. Screen describing the "network."

We are interested in your judgments about the length of lines displayed on this
terminal.

You will be shown eighteen different screens. Each screen will have four lines
displayed. The lines are identified as line (1), (2), (3), and (4).

At the bottom of the screen, you will be asked to identify which line (2, 3, or 4) is
identical to line 1.

Press space bar to continue.

Figure 2. Screen describing the "purpose" of the study.

a n d the other six, e x t r e m e errors. For this replication, Asch's original lengths
n e e d e d to be scaled down; it wasn't possible to have 10 inch lines across the ter-
minal. T h e line lengths were, however, kept in proportion to Asch's specifications.
Since the subjects were unable to see the other terminals, and each of the subjects
316 Smilowitz, Cornpton, and Flint

Because you are part of a network each of you will be randomly assigned a
station number.

Please enter any three digit number and the computer will randomly assign you a
station number.
Please enter a three digit number -> 561

Station Number -> 5

Your station number is 5 .

-> Please remember that you are station number 5.

Press space bar to continue.


. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Figure 3. Screen seen by all subjects-assigning their "station number."

were told that each was station number 5, each was presented the 18 trials which
had the first four answers already programmed but attributed to the other subjects
in the lab (see Figure 4).
At the end of the trials, subjects were asked to input their ages, year in school,
and gender, and the number of hours they typically spend in front of a computer
each week. Subjects were then asked if they believed they were actually networked
with the other stations. Three subjects reported that they did not, and were there-
fore excluded from the analysis. Finally, the subjects were asked to enter a two line
description of the purpose of the study. All of the subjects' answers were stored on
floppy disks and then analyzed with SPSSX.

RESULTS

According to Asch (1956), "While the majority effect was considerable, it was by
no means complete, or even the strongest force at work. The preponderance of esti-
mates was, in each o f the experimental groups, correct or independent of the
majority . . . . " (p. 10). In the original study, most of the subjects trusted their
own opinions most of the time.
Similar are the results of this study, except that the simulated majority had an
even smaller effect on the judgments of the subjects than what was reported by
Asch. Table 1 compares the results in the same format that Asch employed. Fig-
ure 5 presents the results in graphic form.
In Asch's study, the n u m b e r of subjects made no errors was less than 25 %. In
CMC effects o n individual judgments 317

Case #3

FRED: Please enter choice for case 3 -> 2 Thank You.


JOHN: Please enter choice for case 3 -> 2 Thank You.
KATHY: Please enter choice for case 3 -> 2 Thank You.
JILL: Please enter choice for case 3 -> 2 Thank You.
JIM: Please enter choice for case 3 ->

Figure 4. Example of trials presented to subjects.

70

60

= 5o
u
e

~ 40

e 30
o
g
ID
u
L

10

1 1 I
0 ,4 5 6 7 8 g 10 11 12

N u ~ of errors
n CMC Sul:)jec-t.G + Asch's Subjects;

Figure 5. Comparison of error frequencies.


318 Smilowitz, Compton, and Flint

Table 1. Number of Error's In CMC Context and in Asch's Study

CMC Context Asch's Study


N = 42 N = 123
Number of
Errors Frequency Percentage Frequency Percentage

0 29 69.0 29 23.6
1 6 14.3 8 6.5
2 4 9.5 10 8.1
3 1 2.4 17 13.8
4 0 0.0 6 4.9
5 0 0.0 7 5.7
6 0 0.0 7 5.7
7 1 2.4 4 3.3
8 1 2.4 13 10.8
9 0 0.0 6 4.9
10 0 0.0 6 4.9
11 0 0.0 4 3.3
12 0 0.0 6 4.9
Mean .7619 4.41
Median 0.0 3.00
Mean percent 0.17 36.80

the present study, 69% of the subjects are error free. Considering the cumulative
percentages, only 52 % of Asch's subjects committed fewer than 3 errors, as com-
pared to 95 % of the subjects in the computer simulation. The mean number of
errors for Asch's subjects was 4.41. The mean for the replication subjects was less
than one, at . 7619. Half of Asch's subjects made at least 3 errors. The median for
the replication subjects is zero.
Contrasting the results with regards to the severity of the discrepancy between
the correct judgment and the majority opinion, the few errors made by the repli-
cation subjects were with moderate discrepancies. Only one subject made an error
in favor of the majority when the discrepancy was extreme. In Asch's study, the
frequency of errors was about the same on moderate and extreme trials.
The graph in Figure 5 makes the differences evident. In Asch's study, the con-
federates exerted more influence on subjects than did the simulated confederates
in the computer network. The line representing the present study remains below
the line of the Asch study after two errors.

DISCUSSION

Care must be taken in comparing the results of the present study to the results
reported by Asch. As the subjects in this study did not also participate in a sec-
ond and exact replication of Asch's study--that is, in the physical presence of the
experimenter and confederates--other explanations for the differences are possi-
ble. It might be that since Asch's study is nearly 25 years old, the population itself
has changed and individuals have, in general, become less susceptible to group
pressures. A true experimental design was necessary to provide the control for
direct comparisons.
However, the replication of Asch's procedures does allow for an assessment of
CMC effects on individual judgments 319

the social pressures expected of C M C . Subjects were apparently willing to trust


their own judgments and behave independent of the influence of the simulated net-
work members.
One explanation for the results is suggested by Kiesler, Siegel and M c Q u i r e
(1984). The physical absence of others allows individuals to be more responsive
to the immediate textual cues. Gone from C M C are the various cues people give
one another to encourage or discourage choices. Instead, individuals only have the
information which is presented on the screen. The narrow bandwidth of C M C
seems to allow individuals to critically evaluate the information they are receiv-
ing. There are at least three explanations to account for this more deliberate
evaluation.
The first explanation is that physical cues from others may be used to justify a
short circuiting of the critical evaluation process. In the conditions of this study,
the subjects were not able to directly influence one another through their physi-
cal presence. The subjects, instead, were able only to read the judgments of the
"others." The subjects frequently appeared to be experiencing some doubt as they
watched their screens during the responses of the other network "members." Several
times they looked at the lab attendant with puzzlement. The lab attendant, how-
ever, remained noncommittal. With no encouragement to discount their own judg-
ment, the subjects were in a better position to trust their own perceptions than were
the subjects in Asch's design.
The second explanation depends upon the asynchronous nature of C M C . Since
C M C users are by themselves when communicating with others, they can take
more time to examine the information being presented. In conversation, the ex-
pectation is for the listener to "be on the ball" and hesitations before responses are
discouraged. The C M C environment alleviates the responsibility for the quick
response, and sanctions taking the time to think about what is being communi-
cated. In that the user is alone with only the text, the C M C environment may
allow for more critical evaluation of the message and thereby more assertiveness
for one's own positions.
The third explanation became evident during the debriefing. When asked if they
had heard about experiments like the one they were just in, the subjects all said
no, even though many had recently taken an introductory psychology course. Dur-
ing this time, a few subjects reported that there was a problem with the computers.
Rather than mistrust the judgments of their fellow classmates, these subjects
thought that the computer network was making some mistake in the reporting.
These comments suggest that the mediating effects of the computer were regarded
as interfering with the actual communication of the "other network members." Per-
haps out of a need to rationalize their choice to ignore majority opinion, these sub-
jects preferred to consider the computer at fault.

CONCLUSIONS

This study indicates that removing the physical presence of others diminishes the
influence a unanimous majority has on the opinions of an individual. The results
further imply that in the C M C environment, subjects may be more critical and
more willing to evaluate the information they are receiving.
The most obvious implication for C M C group communication echoes Hiltz'
320 Smilowitz, Compton, and Flint

(1975) position: It is easier for a deviant to persist in the C M C environment. Since


the effect of the majority opinion is diminished, individuals with deviant opinions
are more likely to hold out rather than succumb. As pointed out by Fisher (1980)
the deviant is important to group functioning in a n u m b e r of ways, and therefore
fewer constraints on deviants might be a positive outcome of using C M C groups.
T h e findings of this study also suggest that individuals m a y be more critical of
the information they receive in the C M C environment. T h e opportunity to focus
only on the textual message as well as the opportunity to take one's time m a y allow
subjects to involve more of their critical abilities. However, the critical evaluation
of messages which are too discrepant m a y result in attributions of a failure in the
C M C network, or miss-typed information, rather than the messages being per-
ceived as actually representing the intent of the sender.
It would be wrong to think that these apparent features of C M C necessarily
improve group decision m a k i n g without further study. Because of the limited
capacity of C M C to effectively incorporate the richness of traditional group pro-
cesses, it m a y be that C M C is only minimally useful to group decision making,
and perhaps a potential interference. Factors such as social validation, commitment
to the choices m a d e by the group, and substantive conflict m a y depend u p o n the
quality of interaction possible only in more immediate contexts. T h e mode of com-
m u n i c a t i o n of C M C is, after all, simply writing. A n d , most C M C users are not
likely to be the literary giants who are capable of adequately expressing in writ-
ing the social dimension of the group process.
This study, however crude as it might be, suggests one method for further work
in investigating C M C group processes. T h e p r o g r a m m i n g for this investigation
might be expanded to simulate group members' "actual discussion" rather than just
choices about the length of lines. As the messages of the "network m e m b e r s " are
completely controlled by the program, investigators can examine the effects of dif-
ferent messages on individual j u d g m e n t s as well as the messages produced by the
subjects. It m a y even happen that C M C technology proves useful for informing
about face-to-face interaction.

REFERENCES

Asch, S.E. (1956). Studies of independence and conformity: I. A minority of one against a unani-
mous majority. PsychologicalMonographs, 70, 9, 1-70.
Champanis, A. (1975). Interactive human communication. Scientific American, 232, 36-42.
Danowski, J.A. (1982). Computer-mediated communication: A network-based content analysis using
CBBS conference. In M. Burgoon (Ed.), Communicationyearbook VI, (pp. 905-924). Beverly Hills,
CA: Sage.
Edinger, J.A., & Paterson, M.L. (1983). Non-verbal involvement and social control. Psychological
Bulletin, 93, 30-56.
Festinger, L. (1954). A theory of social comparison processes. Human Relations, 7, 117-140.
Fisher, B.A. (1980). Small group decision making. New York: McGraw-Hill.
Furgeson, J.A. (1977). Planet: A computer conferencing system and its evaluation through a case
study. Behavioral Research Methods and Instrumentation, 9, 92-95.
Hiemstra, G. (1982). Teleconferencing, concern for face, and organizational culture. In M. Bur-
goon (Ed.), Communication yearbook 111, (pp. 874-903). Beverly Hills, CA: Sage.
Hiltz, S.R. (1975). Communications and group decision making: Experimental evidence on the potential impact
of computer conferencing (Research Report No. 2). Newark, N.J. Computerized Conferencing and
Communications Center, New Jersey Institute of Technology.
CMC effects on individual judgments 321

Hiltz, S.R. (1978). Controlled experiments with computerized conferencing: Results of a pilot study.
Bulletin of the American Society of Information Science, 4, 11-12.
Hiltz, S.R. (1982). Experiments and experiences with computerized conferencing. In R. Landau,
J.H. Bair, & J. H. Siegman (Eds.), Emerging Office Systems. Norwood, N J: Ablex.
Hiltz, S.R., Turoff, M., & Johnson, K. (1985). Mode of communication and the "risky shift"." A controlled
experiment with computerized conferencing and anonymity in a large corporation (Research Rep. 21). New-
ark: New Jersey Institute of Technology, Computerized Conferencing and Communications
Center.
Hiltz, S.R., Johnson, K., & Mathews, G. (1978). Replicating Bale's problem solving experiments on a com-
puterized conference:A pilot study (Research Report No. 8). Newark: New Jersey Institute of Tech-
nology, Computerized Conferencing and Communications Center.
Hiltz, S.R., Johnson, K., & Turoff, M. (1986). Experiments in group decision making: commu-
nication process and outcome in face-to-face versus computerized conferences. Human Communi-
cation Research, 13, 225-252.
Johansen, R., & DeGrase, R. (1979). Computer-based teleconferencing: Effects on working pat-
terns. Journal of Communication, 29, 30-41.
Kiesler, S., Siegel, J., & McGuire, T.W. (1984). Social psychological aspects of computer-mediated
communication. American Psychologist, 39, 1123-1134.
Korzenny, F. (1978). A theory of electronic propinquity: Mediated communications in organiza-
tions. Communication Research, 5, 3-24.
Phillips, A.F. (1983). Computer conferences: Success or failure? In R.N. Bostrom (Ed.), Commu-
nicationyearbook VII, (pp. 837-856). Beverly Hills, CA: Sage.
Rice, R.E. (1982). Communication networking in computer conferencing systems: A longitudinal
study of group roles and system structure. In Burgoon, M. (Ed.), Communication yearbook VI.
Beverly Hills, CA: Sage.
Rice, R.E., & Associates (1984). The new media: Communication research and technology. Beverly Hills,
CA: Sage.
Short, J., Williams, E., & Christie, B. (1976). The social psychology of telecommunications. London: John
Wiley & Sons.
Smith, M.J. (1982). Persuasion and human action: A review and critique of social influence theories. Belmont,
CA: Wadsworth.
Weick, K.E. (1985). Cosmos versus chaos: Sense and nonsense in electronic contexts. Organizational
Dynamics, 14, 51-65.

You might also like