§§§ WNWWUWIHIIHHIWMIHWWIWIM

   

THESL

2 it 8“

‘llllllllllllllllll

 

 

LIBRARY
Michigan State
University

 

 

This is to certify that the

thesis entitled

The Reduction of Uncertainty in Initial
Multiple User Dimension (MUD) Interactions

presented by

Chad Martin Harms

has been accepted towards fulfillment
of the requirements for

Masters degree in Communication

Mew Mmlsdfih

0 Major professor

 

Date June 26, 2000

 

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

“ 2, DATE DUE

DATE DUE

 

. R
DEC 095? 3001

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1100 C‘JCIFIClDatDDUOpG5-p. 14

THE REDUCTION OF UNCERTAINTY IN INITIAL MULTIPLE USER
DIMENSION (MUD) INTERACTIONS

By

Chad Martin Harms

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF ARTS

Department of Communication

2000

T. .Iu.

ABSTRACT

THE REDUCTION OF UNCERTAINTY IN INITIAL MULTIPLE USER
DIMENSION (MUD) INTERACTIONS

By

Chad Martin Harms

This study examined uncertainty reduction in a Multiple User Dimension (MUD).
Subjects participated in an initial interaction with a stranger in a MUD and
completed a questionnaire about the interaction. The interactions were recorded
and coded for question type and sex. Results indicated that, in keeping with
URT, more demographic than value/opinion questions were asked and sex was
established early in the interaction. Furthermore, men reported more
attributional confidence about the interaction than women, and men and women
did not differ in confidence of affective qualities. Implications for technology and

communication are discussed.

Copyright by
Chad Martin Harms
2000

This thesis is dedicated to my coach, Terry Dowd. Knowing you has only made
me a better, stronger human being.

ACKNOWLEDGMENTS

I would like to thank my advisor, Dr. Kelly Morrison, and Dr. Sandi Smith,
my committee member. Both were instrumental in the completion of this project
and are deserving of the respect that l have for them. Also, thank you to Dr.
Frank Biocca, without whom this project could not have not been completed.

I would also like to specifically thank Dr. Dean Krikorian. Without his
inspiration and friendship, I would not have enjoyed and learned as much as I
have during the past years. Brainstorming for ideas with Dean will always be
worth any time or mileage.

Additional thanks go to Marge Barkman for too many things to count;
Jennifer Butler Ellis for her friendship and support; and Kimberly Stoker for her
kindness and love.

Most of all I would like to thank my family. Jerry, Laurie, Jere, Matt, Eric,
Zachary, and Becca...l love you all. Without your love and support I would not

be who I am today. Thank you.

TABLE OF CONTENTS

LIST OF TABLES ................................................................................ vii
LIST OF ABBREVIATIONS ................................................................... viii
INTRODUCTION .................................................................................. 1

CHAPTER1

LITERATURE REVIEW .......................................................................... 4
CHAPTER 2

METHODS .......................................................................................... 17
CHAPTER 3

RESULTS .......................................................................................... 21

CHAPTER 4

DISCUSSION ..................................................................................... 25
CHAPTER 5

CONCLUSION .................................................................................... 28
APPENDIX A ....................................................................................... 31
APPENDIX B ....................................................................................... 34
LIST OF REFERENCES ........................................................................ 37

vi

LIST OF TABLES

Table 1 - The Proportion of Demographic and Value/Opinion
Questions by Thirds of the Conversation ................................................. 23

vii

LIST OF ABBREVIATIONS

URT .............................................................. Uncertainty Reduction Theory
MUD .................................................................... Multiple User Dimension
FI'F .................................................................................... Face To Face
CMC ..................................................... Computer Mediated Communication
BRB ................................................................................... Be Right Back
AFK ......................................................................... Away From Keyboard
IMHO ....................................................................... In My Humble Opinion
LTNS ........................................................................... Long Time No See
LOL ........................................................................... Laughing Out Loud
ROFL ................................................................ Rolling On Floor Laughing
LMAO ....................................................................... Laughing My Ass Off
CLUES ......................................................... Attributional Confidence Scale
CL7 ............................................. Seven-Item Attributional Confidence Scale
HC ..................................................................................... High-Context
LC ....................................................................................... Low-Context

viii

INTRODUCTION

“Don’t talk to strangers!” Parents have been telling their children this
message for years, hoping to protect and keep them from harm. Through the use
of computers, the immediate danger involved with talking to strangers may seem
reduced. Today people are talking to others in France or in Japan, or whoever
happens to be on their computer when they turn on the power.

Imagine a 46-year-old widow in Tulsa, Oklahoma. This woman has not
had a physical relationship with a man in seven years. For the last seven
months however She has been chatting with a male high school teacher who lost
his wife to breast cancer. They met in a chat room set up for social support of
those grieving for loved ones taken by cancer. For the last seven months these
two individuals have laughed and cried together without even hearing the other’s
voice. Every day they spend hours with each other discussing current events,
telling each other about their day, and revealing personal information about each
other that they had never even told their spouses. They were in love before they
met face to face. They have been married for three of the most wonderful years
of their lives.

Consider another pair who met over the computer. A 45-year-old man
has been talking to several people when he meets a 12-year-old girl. The man is
kind and befriends the gin. The girl sees no harm in talking to this man and
continues to converse with him on-line. She is interested in all that he knows
and all he has done. He tells her about the secret world of adults and she in turn

becomes enthralled with him. The man starts to send the young girl presents,

about which She of course cannot tell her parents. He also starts to send her
pictures. The man puts the girl through a series of tests to determine the level of
trust they have built. At the right time he constructs a lie for the gir1, invites her to
meet him, and sends her a plane ticket. This man is a pedophile and this girl is a
potential victim.

If you do not think these events can happen, you are mistaken. As our
technology continues to advance, interactions with others via computers will
increase on both business and personal levels. Understanding these
interactions will be necessary in order to advance the effectiveness of
communication and prevent possible hazards that may arise due to computer
mediated communication.

Additionally, issues that are obvious or taken for granted in face to face
(FTF) encounters become salient in computer mediated environments. For
example, sex is readily apparent in FI'F interactions, but not in computer
mediated communication (CMC). Thus, how users reduce uncertainty about sex
in CMC needs to be examined. Furthermore, computer mediated interactions do
not involve as many nonverbal cues as FTF communication, and this lack of
nonverbal cues may influence attributional confidence. Researchers must
continue to explore CMC in order to ascertain the degree to which social science
research can be applied to these virtual settings.

The present study examines how uncertainty is reduced, as well as the
influence of sex and attributional confidence in initial interactions between

strangers in a Multiple User Dimension. First, I will discuss MUDs, including

their general properties and the types of communication involved. Next,
Uncertainty Reduction Theory (Berger and Calabrese, 1975), sex, and
attributional confidence will be reviewed. Then, the methods will be

described, followed by the results and discussion.

Chapter 1

LITERATURE REVIEW

Multiple User Dimensions (MUDS)
Gardner (1983) discussed one way in which we can understand the “self,”
or a person’s identity. He stated:

“Rather than a central ‘core self’ which organizes one’sThoughts,
behavior, and goals, the person is better thought of as a collection of
relatively diverse masks, none of which takes precedence over the others,
and each of which is simply called into services as needed.” (Gardner,

1983,p.23)

This concept of self is what Multi User Dimensions (MUDS) and other types of
computer mediated communication allow. It is an environment in which an
individual can be whoever he or she chooses to be without fear, shame, or
embarrassment. It allows disabled individuals who normally would not be able to
interact with many, the opportunity to meet all kinds of people. Those with
disfigurements, who previously felt the need to live a sheltered life, have found a
new life in which no one judges them by their disfigurement.

MUD is the acronym for Multi-User Dimension or Dungeon. MUDS of
today vary from the role playing games of the past, such as Dungeons and
Dragons, where individuals formed relationships with other characters, fought

against enemies and completed a quest. Current MUDS are much more

sophisticated, and include virtual social environments like a shopping mall, with a
variety of shops and strangers. Bennahum (1994) described MUDS as follows:
“T he first MUDS were built around 1978. At the time there was a

popular text-based adventure game called, appropriately Adventure,
which ran on a university computer. At first, players played alone, against
the machine, going from room to room collecting treasure, killing
monsters, until they eventually won the game. After a while it got
boring... Then, something unexpected happened. People no longer
played to win the game; they played to be with other people. A virtual

community began to form.” (Bennahum, 1994, p. 22)

This notion of “community” has its skeptics and proponents. Some
question the idea of a community existing within cyberspace. The skeptics
believe that rather than an actual community, a sense of interdependence
occurs, which creates an illusion of community, (Berry, 1993; Stoll, 1995).
However, there has been ovenivhelming support that strong relationships do exist
between individuals in cyberspace (Parks & Floyd, 1998; Rheingold, 1993). The
relationships formed in cyberspace exhibit similar characteristics to relationships .
formed in real world face-to-face interactions such as self disclosure and
interdependence. They also have strong potential to develop. The most
powerful evidence, however, comes from the users themselves. Users become

committed to their online relationships. They feel interdependent with one

another and help each other with problems by listening and offering advice
(Parks & Floyd, 1998).

MUDS come in a variety of forms. Recent MUDS, although primarily text
based in terms of the communication between individuals, increase sensory cues
by incorporating sounds, visuals, and nonverbal cues. The Palace
(h_ttp://www.thepalace.com) is a user-friendly MUD that offers its members a
number of options not available in other environments. Original MUDS gave text
descriptions of the rooms, corridors, and creatures within them. Now, each
different Palace room offers a visual backdrop created for that particular room.
Rooms vary in theme from bowling allies to deserted islands, all of which can be
accessed by simply clicking a mouse button. When an individual enters a room,
they are immediately given an avatar. An avatar is a visual representation of that
person’s online Palace entity. Avatars can be anything one’s imagination can
create and can be changed at any time. The avatar that everyone starts with is a
smiley face, but the variety of avatars found in the Palace is countless. A user
also chooses a name; this again can be changed at any point in time. From here
on out the user is free to travel to almost any room they would like. They can
use hot keys, or keyboard commands, that allow jumping from room to room or
locating a friend and transporting their avatar directly to that room.

While communication primarily is text based, interactions associated with
face-to-face communication (e.g., turn taking, spatial proximity) reflect an
interactional model of communication. In this text-based environment, one

person speaks, the other then reads the message sent and types a reply. John

SUIer, a psychology professor at Rider University, has studied cyberspace

communities for several years and explained the communication in this way:

“Text communication in a chat room or channel often assumes a
staccato style. Most of the time people express what they have to say in a
brief sentence or two, or in sentence fragments. ...The staccato style of
speaking is very apparent in a chat room or channel where people are
meeting each other for the first time. Because none of the visual cues of
face-to-face encounters are available, people feel the need to quickly test
the waters to determine the qualities of the users around them and whom
they want to engage. Questions that would be considered less than tactful
in face-to-face encounters are a bit more socially acceptable here. Terse
inquiries tossed out to a fellow user, or the entire room, might include
‘Age?’, ‘M/F?’, ‘Married?’

Staccato speak also includes a wide range of acronyms, like BRB
(be right back), AFK (away from keyboard), IMHO (in my humble opinion),
LTNS (long time no see), and LOL (laughing out loud). Different acronyms
evolve in different online cultures, but the LOL is ubiquitous. So important
is the need to express pleasure and laughter that a graduated series of
such expressions have evolved. “Hehe” or "hehehe' indicates a giggle or
a polite/obligatory chuckle: the user finds something humorous, but not
humorous enough to deserve a LOL, ROFL (rolling on floor laughing) or
LMAO (laughing my ass off).”

|http://www1 .rider.egU/~suler/stcyber/texttJalk.htmll

Nonverbal communication also is present, in limited forms, in these text-
based environments. There are a range of sound keys that allow a user to
create the sounds of applause, a burp, or even laughter. The use of these
sounds is infrequent, but spontaneously a group of users may decide to produce
sounds for minutes on end. Additionally, the position of the avatar is a form of
nonverbal communication. By positioning one’s avatar near another, it is made
apparent that the “hi” coming from the avatar is directed toward that person.
Also, the avatar itself communicates something about its user. Avatars often
come in the form of male or female characters and may indicate a person’s
association with a particular group. At times, the age of a user can be estimated
by their choice of avatar. Nothing is necessarily as it seems, though. Poster
(1990) discussed how computer mediated communication allows individuals the
ability to explore other identities without sex cues. An individual can play out a
role in a virtual environment without worries of social acceptability, due to
expected anonymity. Poster also suggested that a loss of actual identity, the
understanding of one’s self in the physical world, could be potentially harmful as
individuals lose themselves in a virtual world. Individuals have the option of
taking on any personality they choose to create. An older woman can be a
teenager in junior high school. A young man can be a mature woman. Creativity

is the only restraint.

Now that MUD’S have been defined and the verbal and nonverbal features
have been discussed, I will discuss Berger and Calabrese’s Uncertainty

Reduction Theory (URT) and relevant research.

Uncertainty Redgction Theorv
The study of initial interactions between two individuals has been the

focus of much research in human communication. Initial interactions have been
examined in a wide range of contexts from the business interview (Jablin and
Miller, 1990) to more personal romantic relationships (Knapp & Vangelisti, 1992).
Regardless of the area of study or the desired end result of the interaction, two
outcomes typically occur. First, a person’s uncertainty about the other individual
typically is reduced and second, an impression is formed of that individual.

Berger and Calabrese (1975), articulated how uncertainty is reduced between

strangers in initial interactions. They suggested that:

“...interpersonal communication behavior plays at least two
different roles. First, we must attempt to develop predictions about and
explanations for our own and. other’s communication behavior; that is,
communication behavior itself is something which we endeavor to predict
and explain. Second, communication behavior is one vehicle through
which such predictions and explanations are themselves formulated.”

(Berger and Calabrese, 1975, p. 101)

Thus, during initial interactions individuals are mainly concerned with predicting
behavior, and we look to communication behaviors specifically to facilitate this
process. Uncertainty reduction, according to Berger and Calabrese, consists of
tvvo processes. Proactive prediction refers to formulating and defining alternative
actions an individual may choose in an upcoming interaction. Thus, proactive
prediction refers to predicting what might occur. Alternatively, retroactive
explanation refers to explaining behavior which already has occurred. Their work
advanced 7 axioms and 21 theorems and has since been extended, with other
researchers examining the relationships between uncertainty reduction and
attributional confidence, culture, and context.

Clatterbuck (1979) extended URT by suggesting that the process of
uncertainty reduction was not simply calculated by correctly choosing which
alternative will occur or which explanation is correct. Rather, he explained it as a
function of the quality of information gathered about the other individual.
Clatterbuck operationalized uncertainty reduction through the attributional
confidence scale (CLUES). A seven-item instrument (CL7) was developed from
the axioms and theorems Berger and Calabrese originally forwarded.
Attributional confidence is the certainty an individual has in their understanding of
another’s characteristics and behaviors. Clatterbuck found positive correlations
between retroactive and proactive attributional confidence. A positive
relationship between time known and attributional confidence was also evident.
Weak support was found between similarity of individuals in a dyadic interaction

and the attributional confidence each reported in predicting the others behavior.

10

Building upon Clatterbuck’s (1979) attributional confidence scale,
Gudykunst, Yang, and Nishida (1985) examined relationships beyond initial
interaction, such as acquaintances, friends, and dating relationships, in three
different countries. They combined related research (i.e. self-disclosure, self-
monitoring, attraction, social anxiety, etc.) to better understand uncertainty
reduction from a multicultural perspective. Their study supported the
generalization of Uncertainty Reduction Theory across relationships and
cultures. However, they found that although Clatterbuck’s (1979) CL7 scale was
valid and reliable, it did not represent a cross-cultural approach to the study of
communication. Gudykunst and Nishida (1986) used URT to examine
differences between cultures that primarily used low-context (i.e., explicit)
communication and those who used high-context communication. Hall (1976)
differentiated cultures according to the communication used by that particular

culture.

A high-context (HC) communication or message is one in which

more of the information is either in the physical context or internalized in

the person, while very little is in the coded, explicit part of the message. A

low-context (LC) communication is just the opposite; i.e., the mass of

information is vested in the explicit code. (Hall, 1976, p 98)

Based on previous research (Gudykunst, Nishida, Koike, and Shiino,
1986), it was found that many forms of uncertainty relevant to other cultures
were not present in working attributional confidence scales. Differences in

communication were used to theoretically explain cultural differences. With this

11

differentiation of culture, Gudykunst and Nishida expanded the Clatterbuck
(1979) CLUES 7-item attributional confidence scale to include items related to
communication context, and in turn to culture.

Culture was not the only extension of the Uncertainty Reduction Theory;
context has been examined as well. In Berger and Calabrese’s (1975) original
discussion of how context plays a role in uncertainty reduction they suggested
that the context can provide uncertainty reducing information. For example, if an
individual attends a Democratic political rally, the context reduces uncertainty by
providing information about the agenda, values, and people. Rubin (1977)
studied how various contexts affected both the number and types of questions
asked in an initial interaction. The context varied from ambiguous to specific,
and the data indicated that encounters in ambiguous contexts lasted longer and
generated more demographic questions than in specific contexts. In an
ambiguous context, where uncertainty is high, much of the information we base
our impression of another person on tends to be demographic information.
Using this limited information, persons often rely on stereotypes to complete an
impression of the other. In specific contexts, more opinion and experience
questions were used. In another study, Rubin (1979) found that demographic
questions dominated the first third of conversations. This can be linked to
Berger and Calabrese’s (1975) third axiom, which stated, “High levels of
uncertainty cause increases in information seeking behavior. As uncertainty
levels decline, information seeking behavior decreased.” (Berger and Calabrese,

1975, p. 103) Rubin’s (1979) study supported Berger’s (1973) earlier findings.

12

He foUnd that demographic questions dominated the first few minutes of a
conversation and opinion questions increased over time. This occurred when
individuals were instructed to, “get to know,” another individual.

The sex of an individual also can influence uncertainty reduction. This
influence is complicated by the fact that the sex an individual takes on in a MUD
can be whatever that individual chooses. Though no empirical research exists
as to the numbers or frequency of sex misrepresentation, it certainly occurs
often. According to Turkel (1995), Habitat, a Japanese MUD, has 1.5 million
users. The member registration Shows a 4 tol ratio of males to females. Inside
the MUD the ratio falls to 3 males to every female. These numbers indicate (on
average) that 5% of the male members are females on Habitat. As a result, sex
becomes a salient issue in MUDS, as what is taken for granted in face-to-face
communication becomes less apparent in these environments. As Curtis (1996)
reported, the sex an individual presents affects their experience within a MUD.
For example, females, because they are fewer in number, may experience
special treatment and even harassment. People may not be who they define
themselves to be in MUDS. These points reiterate Suler’s (1996) observation
that sex is an important factor used to reduce uncertainty in MUDS.

Hypotheses
This section builds upon the context (MUDS) and theory (URT) presented

earlier, and proposes four hypotheses.

13

The use of technology allows individuals to interact within a context
(MUDS) that offers little factual information about the participants. Furthermore,
MUDS are an ambiguous context where uncertainty initially is high, therefore:

H1: The proportion of demographic questions will be greater than the
proportion of value/opinion questions in the first third of the interaction.

Furthermore, as previously articulated in Suler‘s description of staccato
Speak, individuals inquired about the other’s sex very early on in the
conversation. Wood (1997) suggested that we live in a gendered society where
people want to know the sex of other people they encounter. According to West
and Zimmerman (1991), individuals are ovenivelmed with gender on a daily
basis. Individuals display their gender through their accessories, dress and
behaviors in order to maintain and establish relationships. An individual’s
gender, their self-perceived masculine or feminine status, may differ from their
biological sex. Sex is a classification of biological characteristics, whereas
gender is acquired through interaction in a social world, and can change over
time (Wood, 1997). For this reason the term sex is used as opposed to the term
gender. Displays of sex in face to face interactions provide uncertainty reducing
information about an individual that is not readily available to those interacting in
a computer mediated environment. Therefore,

H2: The sex of both individuals will be established in the first third of the
interaction.

How women and men reduce uncertainty has rarely been examined.
Sanders, Wiseman, and Matz (1995) had nearly 900 students fill out

questionnaires in which they thought of an individual and answered questions

14

about that individual. In regards to attributional confidence, no difference in sex
was found for how the other behaved, what the other’s values were, and what
the others attitudes were. Clatterbuck (1979) found sex differences in that
females expressed significantly higher attributional confidence than males.
Wyer, Swan, and Gruenfeld (1995) gave a possible reason for this, showing
several studies that found women seek a greater feeling of “connectedness“ in
conversations than males. Wood (1998) described the communication of males
in terms of establishing their position with regard to others. Women, however,
use communication to learn about others (Johnson, 1996). Therefore,

H3: Women will rate their attributional confidence significantly higher than
men.

Other research has suggested that females report higher levels of
confidence as to perceptions of another’s affect (i.e., “how the other feels about
me, what the other’s feelings and emotions are, and how the other feels about
him/herself.”) (Sanders et al, 1995). This follows earlier work by Baird (1976),
who concluded from a thorough review of studies on sex differences in
communication, that females were more perceptive of other’s emotional states.
Baird also found that females relied more upon and were more sensitive to
nonverbal communication than were males. This is consistent with the finding of
Hall (1978), who found women better at decoding nonverbal cues than men.
However, these studies have been done in face-to-face environments allowing
for extensive nonverbal communication. Mehrabian (1971) described several
aspects of nonverbal communication that signify liking. The other’s voice quality,

proximity, touching and eye contact are all nonverbal cues used to determine

15

affective qualities of others. Due to the decreased amount of nonverbal
communication in MUD’S, especially voice quality, touching and eye contact,
effective perception of affective qualities will be limited. Therefore,

H4: Women and men will not significantly differ in confidence of affective
qualities of uncertainty.

It should be noted that previous studies are limited with respect to the
representation of sex using Uncertainty Reduction Theory. How and to what
extent men and women are able to reduce their uncertainty about individuals
becomes very important, especially in Situations when this is not explicit in their
interactional contexts (i.e., MUDS). The following section explains the methods

used to examine the Uncertainty Reduction Theory in Multiple User Dimensions.

16

Chapter 2

METHODS

Participants and Design it?

Participants consisted of 190 university students. Ninety-five students
were from a large midwestem university and 95 students were from a large
eastern university. Participants were offered extra credit for their participation in
the study. The students had not met before and were classified as strangers.
Procedure

Upon arrival, participants were greeted by the experimenter and were told
that development of an online class was in progress and that communication
options were being tested in order to determine a possible medium. Participants
were informed that they were to ”get to know“ another participant. Participants
were then asked to Sign a consent form stating that they understood any
conversation online would be recorded and would be kept confidential. After the
participant completed the consent form, they went though a basic tutorial on how
to use the computer software utilized in the particular study. Participants were
also informed that if at any point in the study they had questions the
experimenter would be of help. Next the participants were seated at a computer
and entered a private ZD virtual environment (MUD). Simultaneously, at the
other university the same procedure was completed and another subject also
logged on to the same private 20 virtual environment (MUD). The participants

were matched with others according to their sex in order to obtain equity among

17

male/male, female/female and female/male dyads. The participants were again
reminded that the purpose of the test was to simply, "get to know" the other
participants. The two participants were given 7 minutes to interact. After seven
minutes participants were told to log-off the system. During the interaction the
computer automatically recorded a log of the conversation. Participants were
then asked to complete a questionnaire after their interaction. The questionnaire
consisted of the Clatterbuck (1979) CL7 scale and the five questions added in
the Gudykunst et al (1986) study. Computer experience questions were
combined with demographic questions. Finally, questions as to the effectiveness
and efficiency of online communication were asked. After the study was
completed participants were fully debriefed via e-mail.

Coordination between both sites was imperative. The first author
controlled the beginning and ending of all interactions. Assisting at the second
site, an individual was always available via phone or within the virtual
environment. Participants were rotated on a 20-minute cycle. Research
assistants greeted and checked in participants at :00. At :05 participants were
seated and began their training which lasted until :10. The computer screen was
described as a virtual 2-D environment. The smiley face labeled as
“UNVEFISITY1” was explained to be an avatar, a representation of one’s virtual
entity. It was explained that individuals could communicate by typing and then
pressing Enter on the keyboard. Then participants were encouraged to interact
with others in the virtual training room. While participants practiced interacting

with each other, the first author and first assistant determined which rooms (of

18

the four identical rooms that existed) were to be used and which participants
would be sent to which room. At the :10 point, individuals entered into their
particular virtual room. Participants interacted for approximately seven minutes
and were then instructed to say goodbye to their conversational partner.
Participants were then escorted to a different room where they completed a
questionnaire about their interaction. The final :03 were used to reset the
computers to the virtual training room and the log option was reset to record the
next interaction. The phone was used to coordinate the majority of the
interactions, however, when the phone was not available, communicating via the
virtual room was sufficient (see Appendix A for facilitator script).

Varflales of Interest

There were two independent variables examined in this research. The
first independent variable was the sex of the participant, either male or female.
The second independent variable was the period of the conversation. Similar to
Rubin (1979), periods were divided into the first third, second third and third third
of the conversation. Thirds were created by dividing the total number of interacts
in the conversation by three.

Demographic and value/opinion questions, as well as the interacts, were
coded from transcripts of the conversations. Coders distinguished the questions
asked between three separate categories; demographic, value opinion, and
extra. Demographic questions included questions of sex, age, where an
individual was from, what an individual studied and other information that did not

involve a perceptual judgment or opinion (e.g. ”Where are you from? M/F? Do

19

you have any brothers or sisters?”). Value/opinion questions were those that
asked for perceptual judgments about a particular person, place or thing as well
as direct questions about feelings (e.g. “How do you like it there? Why did you
pick that major?”) Questions that were coded as extra included greetings (e.g.
What’s up? How’s it going?) and questions that could not be established as
demographic or value/opinion.

The first dependent variable of interest was the proportion of demographic
questions in the conversation. The proportion was obtained by dividing the
number of demographic questions per third by the number of interacts per third.
The second dependent variable, value/opinion questions, was obtained in the
same manner. The next dependent variable was the third in which the sex of the
individuals was established. Sex was considered established by either the
exchange of names or by direct claims by the individuals. Establishment of sex
also required that the sex of both individuals be given. Attributional confidence
(Clatterbuck, 1979) was the final dependent variable. Following Sanders,
Wiseman, and Matz (1995), the scale was divided into two separate subscales.
The first subscale dealt with affective qualities of uncertainty and was comprised
of three items (i.e. “How certain are you that he/she likes you? How well can
you predict his/her feelings and emotions? How much can you empathize with
(share) the way he/she feels about himself/herself?”). The second subscale
consisted of the remaining nine items of attributional confidence (see Appendix B

for the complete questionnaire).

20

Chapter 3

RESULTS

lntercoder Reliability

The first author trained multiple coders and then any discrepancies were
resolved by discussion. lntercoder reliability was established by comparing the
individually coded conversations and questionnaires. Reliability of conversation
questions coded was found to be acceptable at (or = .88,_Q< .01) for demographic
and (or = .90, p< .01) for value/opinion. Of the interactions in which sex of both
individuals was established, reliability was acceptable at (or = .98, Q< .01 ).
Scale Reliabilitv

Confirrnatory Factor Analysis (Hunter) was used to assess the reliability

of the scales. Items were analyzed for internal consistency and parallelism. The
attributional confidence subscale was found to be reliable with an alpha of .88
and consisted of 9 items (see Appendix B). Affective qualities of uncertainty
were reliable with an alpha of .71.
Hypothesis 1

The first hypothesis posited that there would be significantly more
demographic questions than value/opinion questions in the first third of the
conversations. In order to test this the interactions were first divided into thirds
according to the number of interacts in the conversation. A t-test was conducted
between the proportion of demographic questions (M = .689, SD = .348) and

value/opinion questions (M = .189, SD = .195) in the first third of the

21

conversation. Data were consistent with hypothesis one in that participants
asked significantly more demographic questions than value/opinion questions (1
= 9.62, Q< .001).
HVDotheSis 2

The second hypothesis stated that the sex of both individuals would be
established in the first third of the interaction. The proportion established in the.
first third (M = .642) was significantly higher than the proportion established in
the second third (M = .095)(1 = 8.026, p< .001). In addition, , the proportion
established in the first third (M = .642) was significantly higher than the last third
M = .053)(t = 9.698, p< .001 ). Sex was not established for both individuals in
21.06% of the conversations. The data were consistent with the hypothesis that
sex would be established in the first third of the conversation. Table 1 shows the

proportion of demographic and value/opinion questions by conversation thirds.

22

 

TABLE 1

ma PROPORTION OF DEMOGRAPHIC AND VALUE/OPINION QUESTIQNS
_BY THIRDS OF THE CONVERSATION

 

Conversation Demographic Value/Opinion
Thirds . Questions Questions
1 .69 .1 9
2 .65 .33

3 .36 .21

 

23

Hypothesis 3

The third hypothesis that claimed that women would rate their attributional
confidence significantly higher than males was not supported. This hypothesis
was tested by comparing the means between male and female scores on the
attributional confidence subscale. T—test results indicated that males average
attributional confidence (M = 335.26, _S_D = 196.61) differed from females
average attributional confidence (M = 282.12, SD = 163.86) significantly (1 =
3.75, p< .05), but in an opposite direction than that hypothesized. The data were
not consistent with the hypothesis.

Hypothesis 4

Hypothesis four stated that women and men would not significantly differ
in confidence of affective qualities of uncertainty. Testing this hypothesis was
done comparing the averages of male and female scores on the affective
attributional confidence questions. T-test results indicated the data were
consistent with hypothesis four in that on average, males (M = 113.06, SD =
72.44) and females (M = 103.19, SD = 65.25) did not differ significantly (t = .729,

Q = .394) .

 

24

Chapter 4
Dicussion

The results suggest that demographic information about another individual
is the first concern in an interaction. This type of information seems to be the
most useful in reducing the uncertainty an individual has when initially interacting
with a stranger. Uncertainty is increased by the ambiguity of the context in which
the interaction occurs. The results of this study are consistent with earlier
findings by Berger (1975) and Rubin (1979), who found that in face-to-face initial
interactions between strangers, demographic information was sought first. The
tendency to share equal information (reciprocity of information exchange) also
stimulates the exchange of demographic questions.

The results of this study indicated that establishing the sex of the other
individual at the beginning of the conversation, the first third, is considered
important, if not a necessity. Specifically, the sex of the other is extremely
important. The sex of the individual to whom one is conversing with can
influence the conversational topics (McCroskey, Richmond, and Stewart, 1986).
For example, a female may choose to discuss her relationships with another
female, whereas this topic may not be suitable for a female/male interaction.
Females are also more positively rated in social interactions than males (Reis,
1986; Wheeler, Reis, and Nezlek, 1983). If a male finds he is speaking with a
female, he may feel that she will empathize with his situation more than a male
might. According to Wood (1998), males use communication as a way to display

control, independence, and status. This differs from females, who use

25

communication to establish and maintain relationships. These differences
indicate that from the beginning of the interaction, the sex of the other individual
has an impact on the communication between the two individuals. Additionally, it
suggests the degree to which sex is socially construed and reliant Upon visual
and auditory cues. Therefore, West and Zimmerman’s (1991) conception of
“doing gender" is specific to FTF interactions, and future research should further
explore how people “do gender” in CMC environments.

The third hypothesis posited that women would rate their attributional
confidence higher than males. The data were not consistent with this
hypothesis. Males actually rated their attributional confidence significantly higher
than females. While Clatterbuck (1979) found that females had higher
attributional confidence than males, the context in which these studies were
conducted was extremely different in terms of the information provided.
Specifically, visual cues to an individual’s age, sex, or background as well as
voice or auditory cues allowing for expression and emotion were not present in
these contexts. As shown earlier, research has found that females utilize
nonverbal communication more than males (Baird, 1976; Hall, 1978). Perhaps
the lack of nonverbal cues available in CMC suppressed the female participant’s
perceptions of attributional confidence. Possibly, due to the highly unbalanced
ratio of males to females on-Iine, females felt more at risk or vulnerable in this
environment, thus influencing confidence levels. The amount of time that the
interaction lasted also may influence the confidence levels. Berger and

Calabrese (1975) propose that the longer individuals interact, the lower their

26

uncertainty of the other will be due to more information being exchanged.

In particular, the lack of nonverbal cues may have impacted the affective
quality of information exchanged in the interactions. Although the written word
has substantial power in the emotion it can evoke, the staccato quality to the
conversation may reduce the likelihood of such affect. The effect of sex in
utilizing types of information may influence one’s ability to confidently attribute
affect to another individual.

Although the lack of nonverbal cues may have affected the attributional
confidence of women and men, other aspects of the interaction should also
impact their perceptions. A question exists that, although impossible to answer
from the obtained data, needs to be addressed. Is it that women have lesser
attributional confidence in the current context, MUD’S, or is it that males have
greater attributional confidence specifically in these contexts? The answer is
unclear, and further research in MUD’S is needed to answer this. Caldwell and
Peplau (1982) found that women were more intimate in relationships than men.
It has also been found that women tend to self-disclose more information than
males (Sanders et al,1995). It seems logical that the seven minute interactions
between participants lacked intimacy and self-disclosure. Perhaps the limited
disclosure that did occur affects males and females differently, and in turn

affects their attributional confidence.

 

27

Chapter 5

CONCLUSION

Mmitations of the StLMy

Participants were told at seven minutes, to conclude their interaction.
However, the conclusions varied in length. This means that some interactions
lasted seven minutes and 10 seconds, and others may have lasted up to eight
and a half minutes. The coordination between the two universities could not be
so precise as to allow for strict identical timings. The error due to the difference
in times is an important consideration.

The difference in setting also gave room for error. The setting at one
university differed from that of the other university. Although this difference may
not have influenced the perceptions of the participants in any significant way, it is
necessary to consider all possible variations in participant perception.

Yet another limitation of this study is the generalizability of the findings.
The virtual world is ever changing. The technology of tomorrow will be
developed with reference to the shortcomings of today’s technology. Slater and
Wilbur (1997) discuss the concept of presence and its affect on perception within
the virtual realm. As nonverbal and tactic cues are developed and incorporated
into the virtual environment, the sense of presence an individual feels in that
environment will also change.

Despite the limitations to the study, the multi-university research

conducted was done with surprisingly few problems. Through proper and

28

lengthy forethought, potential problems were eliminated and the fluidity in which
the large sample of participants was coordinated effectively. The technology
used provided a look into future applications of multi-university research.
Epture Research

This research has added another piece of evidence through which we
may better understand and appreciate human communication. More importantly,

it has raised new and exciting questions that may be expanded and examined.

For instance, what is the process by which individuals are socialized into
virtual communities? The socialization of individuals into a new culture or
organization may provide greater understanding as to the change in
communication. Specifically, how quickly does a new member of the community
acquire the terminology of the community? How do new members to a virtual
community develop networks with the community? Doing a network analysis will
be useful in determining how such communities develop as well as the type of
communication between individuals.

Differences in the communication between males and females are small,
and according to Canary and Hause (1993) far less significant than other factors.
The current study differs from actual chatrooms or virtual environments due to
the participants in the study. The majority of participants were novices in their
use of chatrooms. This became apparent through the conversations between
individuals. In a chatroom of experienced users, interacts are, as SUIer (1996)

explained, more staccato in form. Experienced users will know and use more

29

shorthand expressions to convey the same thing that novice users would write
out completely. An example of this would be the question of another‘s sex. An
experienced user would ask for this distinction by Simply typing, “m/f?,” whereas
a novice user might type, “Are you a male or female?” First, this difference is
important in that it constricts the amount of information that can be exchanged in
a given amount of time. An experienced user can obtain much more information
about another individual in less time. Second, this distinction raises interesting
questions as to how an individual learns to communicate in a virtual
environment. How long does it take an individual to acquire this knowledge?
How is a novice individual perceived differently than an experience individual?
The relationships that develop on-line are another very interesting area of
research that needs to be better understood. For example, is attraction to
another, whom an individual has not met or possibly seen, different than
attraction has been conceptualized in face-to-face interactions? This research is
new and holds a vast range of questions, some of which have begun to be
explored (Parks and Floyd, 1998). A particular question raised by this research
is the extent to which the sex of an individual determines the conversational
topics in an interaction. This could be further developed through an experiment
and content analysis structured to determine conversational topics according to
varying dyadic sex combinations. Hopefully, these questions will grow new

questions and further expand social science research.

 

30

APPENDICES

31

APPENDIX A

Facilitator script

This script will be used to facilitate research studying interaction in a Multi-User
Domain. A MUD is an advanced chat room on the lntemet. The MUD being
utilized in this study is a user-friendly 2D environment in which participants can
interact. Participants will sign up for times to come into the lab and complete the
study. Upon arriving at the designated room, participants will be asked to sign a
consent form explaining their rights. Participants will then be escorted to a
computer by a research assistant who will explain basic functions of The Palace
and allow participants to perform such basic functions. The script for the
research assistants has been developed using instructions offered in The
Practice Palace, an introductory Palace offering basic function instruction. The
script is as follows:

At the end to the previous sample, reset log in the Intro Palace
Avatar should appear in the room and should be green/red in color.

TO PARTICIPANT> what you will be doing today is talking with another
individual on the computer and simply getting to know them. We are interested
in examining impression formation of others over the lntemet. I will go through
some basic functions of this program step by step with you. After we go through
the introduction I will put you on a different computer that is connected to the
lntemet and you will be able to talk with the other person at another Eastern
university. If you have questions at any point I am here to help.

Seat participant at the introduction computer.

TO PARTICIPANT> This is your Avatar, which is your face when chatting
with people in a room.

TO PARTICIPANT> You talk to other people by typing a message and
pressing enter. Go ahead and try it now.

TO PARTICIPANT> You can move around the screen by clicking. Try
clicking anywhere in the screen to see your avatar move.

TO PARTICIPANT> A log is kept so you are able to read messages that
you may have missed while typing.

TO PARTICIPANT> Do you have any questions at this time? Okay, it will
just take a minute to make sure the other computer is ready. I’ ll be right back.

32

Look at other computer to see if Cornell is in place and ready.
TO PARTICIPANT> If you would come with me.
TO PARTICIPANT> Once again, you are represented by the

colored avatar. Please spend 7 minutes getting to knowthis other individual.
After seven minutes I will come back and have you Iogoff.

7 minutes

TO PARTICIPANT> 7 minutes has passed, please go ahead and end
your conversation with the other person and come with me.

Take participant to next room to fill out questionnaire.
TO PARTICIPANT> Now to finish up, I have a short questionnaire for you

to fill out. Please make sure to read to the instmctions carefully and take your
time.

33

APPENDIX B

COMPLETE QUESTIONNAIRE

In the first part of this questionnaire, please think of the interaction you have just
experienced. Take time to read the instructions and answer the following
quesflons.

Instructions: Using the following scale, for each question please indicate how
confident you are in your abilities to answer, or predict, the correct response to
the question based on the computer interaction you just had. On these
questions, the answers should be written as a percentage, anywhere from 0% to
100%. For example, if you are totally confident that you know a particular fact,
you write 100%. If you were slightly less confident, you might put a number like
93%. On the other hand, if you were not at all confident you might place a very
low percentage, like 5% in the answer blank. If you absolutely are unable to
answer a question, and the answer would be a guess for which you had no basis
at all, you would write 0%. Remember, you may use any evidence as a basis for
your guess, even if the person has not explicitly told you the answer. We are
interested in your ability to form first impressions of others.

 

 

0% 50% 100%

unable to moderately completely

answer confident confident

1. How confident are you of your general ability to predict how he/she will
behave?

2. How certain are you that he/she likes you?

3. How accurate are you at predicting the values he/she holds?

4. How accurate are you at predicting his/her attitudes?

5. How well can you predict his/her feelings and emotions?

6. How much can you empathize with (share) the way he/she feels about

himself/herself?

7.

10.

11.

12.

How well do you know him/her?

How certain are you of his/her background?

How certain are you that he/she will behave in a socially appropriate
way when this is important?

How certain are you that he/she can understand your feelings when
you do not verbally express them?

How certain are you that you understand what this person means
when you communicate?

How confident are you that this person will be patient with you when
you communicate?

This last part of the questionnaire asks you a few questions about yourself.
Please provide the requested information in the space provided or circle all
that apply.

1. Age?

2. Sex:

a.
b.

Male
Female

3. Do you own a personal computer?

a.
b.

yes
no

35

4. What types of communication operations do you perform on a computer”?
(circle all that apply)

e-mail
on-Iine shopping

surfing the web

chat rooms

on-Iine magazines

virtual reality environments
on-line games

news groups

other?

"FER-*SDP-PP'P

 

5. How often do you access your e-mail?

never
seldom

once a week

more than once a week, but not daily
daily

multiple times a day

’PRPRP

6. In a typical week, how much time do you spend surfing the lntemet?

a 0 hours a week

b. less than 2 hour per week

c more than 2 hour per week but less than 5 hours
d more than 5 hours per week

36

BIBLIOGRAPHY

Bibliography

Baird, J. E., Jr. (1976). Sex differences in group communication: A review
of relevant research. Quarterly Journal of Speech, 62, 179-192.

Bennahum, D. (1994). Fly me to the moo: Adventures in textual reality.
Linguafranca, pp. 22.

Berry, W. (1993). Sex, economy, freedom, and community. New York, NY:
Pantheon.

Berger, C. R. (1973). The acquaintance process revisited: Explorations in
initial interaction. Paper presented at the annual convention of the Speech
Communication Association, New York, NY.

Berger, C. R. (1975). Proactive and retroactive attribution processes.
Human Communication Research, 2, 33-50.

Berger, C. R. & Calabrese, R. J. (1975). Some Explorations in Initial
Interaction and Beyond: Toward a Developmental Theory of Interpersonal
Communication. Human Communication Research, 1(2), 99-112.

Caldwell, M. A. and Peplau, L. A. (1982). Sex differences in same-sex
friendship. Sex Roles, 23, 713-725.

Clatterbuck, G. W. (1979). Attributional Confidence and Uncertainty in
Initial Interactions. Human Communication Research, 5(2), 147-157.

Curtis, P. (1996). Mudding: Social phenomena in text-based virtual
realities. In M. Stefik (Ed.) lntemet dreams: Archetypes, myths, & metaphors.
Cambridge, MA: The MIT Press.

Gardner, H. (1983). Frames of mind. New York, NY: Basic Books.

Gudykunst, W. B., Yang, S. and Nishida, T. (1985). A cross-cultural test of
Uncertainty Reduction Theory: Comparisons of aquaintances, friends, and dating
relationships in Japan, Korea, and the United States. Human Communication
Research, 11(3), 407-454.

Gudykunst, W. B. & Nishida, T. (1986) Attributional Confidence in Low-
and High-Context Cultures. Human Communication Research, 12(4), 525-549.

38

 

Gudykunst, W. B. & Nishida, T., Koike, H., and Shiino, N. (1986). The
influence of language on uncertainty reduction: An exploratory study of
Japanese-Japanese and Japanese-North American interactions. In M.
McLaughlin (Ed.), Communication Yearbook 9 (pp. 555-575). Beverly Hills,
CA:Sage.

Hall, E. T. (1976). Beyond culture. New York, NY: Doubleday.

Hall, E. T. (1978). Gender effects in decoding nonverbal cues.
Psychological Bulletin, 85, 845-857.

Jablin, F. M. and Miller, V. D. (1990). Interviewer and applicant
questioning behavior in employment interviews. Management Communication
Quarterly, 4, 51-86.

Johnson, F. (1996). Friendships among women; Closeness in dialogue. In
J. T. Wood (Ed.), Gendered Relationships. Mountain View, CA: Mayfield. p. 79-
94.

Knapp. M. L. and Vangelisti, A. L. (1992). Interpersonal Communication
and Human Relationships, (2"d ed.). Boston, MA: Allyn & Bacon.

McCroskey, J. 0., Richmond, V. P., and Stewart, R. A. (1986). One on
one: The foundations of interpersonal Communication. Englewood Cliffs, NJ:
Prentice-Hall. p 241 -247.

Mehrabian, A. (1971). Silient Messages. Belmont, CA: Wadsworth.

Parks, M. R. and Floyd, K. (1998). Making friends in cyberspace. Journal
of Communication, 46(1), 0021 -9916/96.

Poster, M. (1990). The mode of information, Chicago, IL: University of
Chicago Press.

Reis, H. T. (1986). Gender effects in social participation: Intimacy,
loneliness and the conduct of social interaction. In R. Gilmour and S. W. Duck
(Eds.), The imerging field of personal realationships. Hillsdale, NJ: Lawrence
Erlbaum. p 91-105.

Rheingold, H. (1993). The virtual community. Reading, MA: Addison-
Wesley Publishing Company.

Rubin, R. B. (1977). The role of context in information seeking and
impression formation. Communication Monographs, 44(1), 81-90.

39

 

Rubin, R. B. (1979). The effect of context on Information Seeking across
the span of initial interactions. Communication Quarterly, 27(3) 13-20.

Sanders, J. A., Wiseman, R. J. and Matz, S. l. (1995). The influence of
gender on reported disclosure, interrogation, and nonverbal imediacy in same-
sex dyads: an empirical study of uncertainty reduction theory. Women’s Studies
in Communication, 85-108.

Slater, M., & Wilbur, S. (1997). A framework for immersive virtual
environments (FIVE): Speculations on the role of presence in virtual
environments. Presence, 6(6), 603-616.

Stoll, C. (1995). Silicon snake oil. New York, NY: Doubleday.

SUIer, J. (1996). Psychological Dynamics of Online Synchronous
Conversations in Text-Driven Chat Environments.
[http:l/www.rider.edu/users/suler/psycyber/texttalk.html1 990].

Turkel, S. (1995). Life on the screen: Identity in the age of the lntemet.
New York, NY: Simon and Schuster.

West, C., & Zimmerman, DH. (1991). Doing gender. In J. Lorber & S. A.
Farrell (Eds), The social construction of gender (pp. 13-37). Newbury Park, CA:
Sage Publications.

Wheeler, L., Reis, H. T., and Nezlek, J. (1983). Loneliness, socail
interaction, and sex roles. Journal of Personality and Social Psychology, 45, 943-
953.

Wood, J. T. (1997). Gendered lives. Belmont, CA: Wadsworth Publishing.

Wyer,, R. S. Jr., Swan, 8., and Gruenfeld, D. H. (1995). Impression
forrnatiopn in informal conversations. Social Cognition, 13(3), 243-272.

40

 

       

      

RRR ES

I.LIB I

488544

V

 

 

 

 

 

 

 

 

 

MICHIGAN STATE UN
llHIWIillllliilllllllllllilllll
31293020