a. 4.4%
b.

a
s
..

. .
at. 3:.
”WWI b. u .

.3”.

$1.? .
E. «up a
.3 .w

,. a; (.
wants:
A. gv‘fl‘a...v

.
v.1...

u.
.3551: n .

1 Ig.‘

 

 

.9

 

 

. M LIBRARY
, °‘ ,- ichigan State
*0“ University

 

 

 

This is to certify that the
dissertation entitled

A Socio-Cognitive Model of Information Disclosure in Human-
Computer Interaction

presented by

Doohwang Lee

has been accepted towards fulfillment
of the requirements for the

Ph.D. degree in Media and Information Studies

 

 

.47
1231”,;37 ,cfi-e.’

Major Professor’s Signature

Date

MSU is an affirmative-action, equal-opportunity employer

— _.-n-n-n-u--.---o-o—--—-------o-.-o-

 

 

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 DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K.IProj/AccaPres/CIRC/DateDue Indd

 

 

A SOCIO-COGNITIVE MODEL OF INFORMATION DISCLOSURE IN HUMAN
COMPUTER INTERACTION

By

Doohwang Lee

A DISSERTATION

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

DOCTOR OF PHILOSPHY
Media and Information Studies

2008

ABSTRACT

A SOCIO-COGNTIVE MODEL OF INFORMATION DISCLOSURE IN HUMAN-
COMPUTER INTERACTION

By
Doohwang Lee

This dissertation proposed and tested a social cognitive model of information
disclosure in human-computer interaction. Guided by Bandura’s (1991, 2001) Social
Cognitive Theory (SCT), this study conceptualized personalized social cues of Web site
interface as computer immediacy (LaRose & Whitten, 2000) and investigated social
cognitive processes of the effects of the personalized social cues of immediacy on the
likelihood of consumers revealing their personal and health information in the context of
a personal health record (PHR) Web site service. In the proposed model, personalized
social cues of immediacy was conceptualized as a social environmental influence which
reciprocally interacts with the other two cognitive determinants of human cognition
(outcome expectations and privacy self—efficacy) and human behavior (information
disclosure).

The results showed that, regardless of information type (embarrassing vs.
descriptive) that can be requested from PHR Web sites, consumers’ exposure to high
level of personalized social cues of immediacy significantly and positively motivated
their disclosure intentions even when privacy self-efficacy was controlled. However, the
significant direct effect of computer immediacy on the likelihood of information
disclosure was not evident in the final path model involving the potential mediators of
positive outcome expectations (social trust and customized service benefits) and negative

outcome expectations (embarrassment and information abuse). The null relationship

between computer immediacy and information disclosure intention suggests that the
effect of computer immediacy on disclosure intention is mediated by the social cognitive
variables of positive and negative outcome expectations.

Overall, the results of the present study supported the application of SCT to the
domain of human-computer interaction. Consistent with SCT, exposure to personalized
social cues of immediacy translated positively into one’s cognition toward the stimulus
PHR Web site by promoting expectations of positive consequences of social trust and
customized service benefits and reducing expectations of negative consequences of
information abuse from the Web site at the same time. This social cognitive model of
information disclosure extend the CASA model of human-computer interaction (Nass &
Moon, 2000; Reeves & Nass, 1996) which suggests that humans tend to respond to the
computer without making much social cognitive effort. Consequently, this dissertation
may untangle the relationship between personalized social cues of immediacy and human
motivation of information disclosure as a social cognitive process through which
consumers consider all personal, behavioral, and environmental influences associated
with their information disclosure behavior in dynamic human-computer interaction in the

Internet.

Copyright By
Doohwang Lee
2008

ACKNOWLEDGMENTS

The completion of this dissertation would not have been possible without many
people’s patience, support, and encouragement.

First and foremost, I would like to thank my mentor, Dr. Robert LaRose, whose
inspiration made my dream a reality. Without his unlimited support and guidance
throughout the entire process, I would not have been able to finish this work.

I also want to express my gratitude to the other members of my advisory
committee, Dr. Mark Levy, Dr. Hairong Li, and Dr. Wei Peng, for their kind advice on
and valuable critique to this project.

In addition, I am grateful for Kwonjoon Y00 and Kaitlan Chu who developed the
stimulus Web site for my dissertation experiment. My special thanks also go to Dr.
Federicco De Gregorio who professionally edited my manuscript.

I also extend many thanks to my colleagues at the Department of
Telecommunication and Film at the University of Alabama and my friends in both Korea
and the United States for their friendship and support during my study.

I also thank my parents, Hak-Nae Lee and J ung-Sook Shin, for their unlimited
love to and trust in their youngest child for over the years.

Finally, I dedicate this dissertation to my wife, Jungsun Ahn, who has been with
me throughout my doctoral program and will be always with me throughout my life.
Without her love, support and sacrifice, I would not have been able to complete this

dissertation. Further, I can’t thank her enough for giving birth to our precious baby, Lissa.

TABLE OF CONTENTS

LIST OF TABLES ........................................................................................................... viii
LIST OF FIGURES ........................................................................................................... ix
INTRODUCTION .............................................................................................................. 1
CHAPTER 1
COMPUTER IMMEDIACY AND INFORMATION DISCLOSURE IN HUMAN-
COMPUTER INTERACTION ........................................................................................... 8
Defining Computer Immediacy ................................................................................... 8
Immediacy in Human-Human Communication ................................................. 10
Computer Immediacy in Human-Computer Interaction ..................................... 13
Computer Immediacy in Other Media Theories ................................................. 15
Understanding Information Disclosure in Human-Computer Interaction ................. 18
Information Disclosure in Human-Computer Interaction .................................. 18
Social Cognitive Processes of Information Disclosure ...................................... 23
CHAPTER 2

INTEGRATED MODEL OF INFORMATION DISCLOSURE AND HYPOTHESES. 31
A Social Cognitive Perspective of Information Disclosure in Human-Computer

Interaction .................................................................................................................. 3 1
A Social Cognitive Model of the Effect of Computer Immediacy on Information
Disclosure .................................................................................................................. 37
CHAPTER 3
METHOD ......................................................................................................................... 50
Research Design ........................................................................................................ 50
Participants and Procedures ....................................................................................... 51
Stimulus Material ....................................................................................................... 54
Two Levels of Immediacy Manipulation ........................................................... 57
Types of Information Manipulation ................................................................... 58
Measurement .............................................................................................................. 59
Dependent Measure ............................................................................................ 59
Potential Mediating Measures ............................................................................ 59
Control Measure ................................................................................................. 62
Manipulation Check Measures ........................................................................... 62
Preparation for Data Analysis .................................................................................... 63
Data Analysis .............................................................................. I .............................. 64
CHAPTER 4
RESULTS ......................................................................................................................... 66
Manipulation Checks ................................................................................................. 66

vi

Computer Immediacy Manipulation ................................................................... 66

Information Type Manipulation ......................................................................... 66
Hypothesis Testing .................................................................................................... 68
CHAPTER 5
DISCUSSION ................................................................................................................... 83
Social Cognitive Process of the Effect of Computer Immediacy on Information
Disclosure Intention ................................................................................................... 84
Moderating Role of Information Type in the Social Cognitive Processes of the
Effect of Computer Immediacy on Self-Disclosure Intention ................................... 88
Implications ............................................................................................................... 92
Theoretical Implications ..................................................................................... 92
Practical Implications ......................................................................................... 93
Limitations and Future Research ............................................................................... 96
APPENDICES ................................................................................................................ 100
REFERENCES ............................................................................................................... 108

vii

LIST OF TABLES

Table 1. Paired Samples T-Tests for Information Type Manipulation ............................. 67

Table 2. Pearson Product-Moment Correlation Coefficients of the Variables Related to
Embarassing Information Disclosure ................................................................................ 70

Table 3. Pearson Product-Moment Correlation Coefficients of the Variables Related to
Descrptive Information Disclosure ................................................................................... 71

Table 4. Means and Standard Deviations for Information Disclosure Intentions by
Computer Immediacy and Information Type Conditions ................................................. 79

Table 5. Results of Hypothesis Testing ............................................................................ 82

viii

LIST OF FIGURES

Figure 1. Proposed Reserch Model ................................................................................... 49
Figure 2. A Screenshot of the Log-in Page of the E-Medfile Web Site ........................... 53
Figure 3. A Screenshot of the Greeting Page of the E-Medfile Web Site ........................ 55
Figure 4. A Screenshot of the Navigation Page of the E—Medfile Web Site .................... 56

Figure 5. A Screenshot of the Interactive Review Page of the E-Medfile Web Site ........ 58

Figure 6. Results of Standardized Path Analysis for the Proposed Model of
Embarassing Information Disclosure ................................................................................ 72

Figure 7. Results of Standardized Path Analysis for the Proposed Model of
Descriptive Information Disclosure .................................................................................. 73

INTRODUCTION

The rapid advance of information and communication technologies, particularly
the Internet, has increasingly required consumers to engage in some form of disclosure. A
growing number of organizations use their Web-based technologies to collect substantial
amounts of personal information about their customers (Cazier, Shao, & Louis, 2007;
Dinev & Hart, 2006a, 2006b; Dolnicar & Jordan, 2007; Hassanein & Head, 2007; Hui,
Teo & Lee, 2007; Metzger, 2004; Milne, Bahl, & Rohm, 2008; Moon, 2000; Ward,
Bridges, & Chitty, 2005). These Web sites not only execute a cost-effective way to
advertise their products and services, but also provide interactive capacities to facilitate
communication with consumers, tailor content to their needs, and collect information
about their customers (Ansari & Mela, 2003; Awad & Krishnan, 2006; Kalyanaraman &
Sundar, 2006; Metzger, 2004; Milne et al., 2008). In fact, many Web-based services
routinely ask existing members and would-be members to submit their personal
information in exchange for their customized products and services (Chellappa & Sin,
2005; Culnan & Bies, 2003; Milne et al., 2008), such as “free” personalized content
(Pastore 1999), product discounts and coupons (Phelps, Nowak & Ferrel, 2000),
monetary rewards (Hann, Hui, Lee, & Png, 2007; Hui et al., 2007; Milne & Culnan,
2004), loyalty program memberships (Earp & Baumer, 2003), interactions with
automated agents (Berendt, Gunther, & Spiekermann, 2005), or some other form of “fair
exchange” (Culnan & Bies, 2003).

In particular, health-related Web sites provide their visitors with a wide range of

health and medical information and ask them to respond to self-profiling questionnaires

about their sensitive personal medical records, health habits, and sexual activities as well
as personal demographics (Moon, 2000; Rohm & Milne, 2004). For example, many
branded direct-to-consumer drug (DTC) Web sites (e.g., viagra.com) collect visitors’
information about their personal demographics and symptoms using Web-based survey
tools and site registration processes (Macias & Lewis, 2003). Once they reveal their
personal information, visitors are often offered customized and tailored services, such as
an assessment of the likelihood that they have a particular illness or might contract it in
the future (Rohm & Milne, 2004). As such, the Internet is equipped with state-of-the-art
computer technologies to integrate, access, and distribute databases containing
consumers’ healthcare history, preferences, and behaviors. A wide range of health
organizations, including primary benefits managers, insurance companies, and health
maintenance organizations (HMOs), are expected to continue to expand the use of
Internet-based data collection to demand more personal and health information in the
name of authentication and customized offerings (Zahedi & Song, 2008).

One of the most effective data-collection technologies for Web-based services is
to utilize an intelligent computer agent], which can address an individual in a
personalized manner and respond to that individual in some contingent fashion (Deighton,
1996; Hassanein & Head, 2007; Moon, 2000). By introducing a computer agent that

enables one-to-one reciprocal interaction into their Web sites, organizations can build and

 

' Computer agents should be distinguished from avatars. While an agent is some type of mathematical or
computational algorithm designed to execute a specific command that is internally formulated, an avatar
can be perceived as a digitally embodied representation whose behaviors are controlled either by a human
user or executed by a computational agent within the computer (Bailenson & Blascovich, 2004). In this
research, computer agents are limited to the computational algorithm that enables personalized social cues
using the mode of text messages.

establish a valued relationship with consumers (Deighton, 1996; Dinev & Hart, 2006a,
2006b; Gefen & Straub, 2003; Hassanein & Head, 2007; Moon, 2000). Using such
personalized communication approaches, organizations not only attract their customers to
their Web sites (e.g., Gefen & Straub, 2003; Hassanein & Head, 2007; Kalyanaraman &
Sundar, 2006), but also gather customers’ information with much ease, ranging from
personal to medical_and psychiatric information (e.g., Heerwcgh, Vanhove, Matthijs, &
Loosveldt, 2005; Joinson & Reips, 2007; Moon, 2000; Weisband & Kiesler, 1986).
Consequently, personalized communication strategies demonstrate the importance of
unique customer databases in today’s age of the network society (cf. van Dijk, 1999).

In fact, much of the messages from Web sites highlight the social agency of the
immediate and responsive computer. For example, when logged in, Web sites are
programmed to address consumers’ names in their welcoming salutation, notify them of
the frequency of their visits, and respond to consumers’ previous requests automatically.
These personalized guidance and responses can provide highly interactive and immediate
social cues to lessen psychological distance between Web sites and consumers. These
practices mirror the traditional face-to-face data-driven direct marketing strategies that
have been effective in building relationship with customers. (e. g., Dahl, Manchanda, &
Argo, 2001; Doney & Cannon, 1997; Dwyer, Schurr, & Oh, 1987; Hassanein & Head,
2007; Moorman, Zaltman, Deshpande, 1994; Morgan & Hunt, 1994; Narayandas &
Rangan, 2004; Swan, Bowers, & Richardson, 1999). For example, when embedded with
highly interactive, immediate and personalized social cues, such systems can address
consumers by name, praise them, share personal anecdotes, and respond to their requests

even by possibly showing the smiling picture of the representative of the Web site.

Recent studies of Web-based survey administration show that such personalized
social cues can be used to increase response accuracy and rates as well (e. g., Heerwegh et
al., 2005; Joinson & Reips, 2007). These social cues can include the use of researcher
photographs, information about the researcher, the use of personal salutations and
questions, or URLs designed for specific individuals (Joinson & Reips, 2007). Such
social qualities of the Web service can offer the opportunity to provide real-time,
immediate, and responsive feedback to consumers’ needs and desires. In this process,
these personalized social cues of the Web site can evoke feelings of closeness and
warmness that promotes the establishment of relationships between the Web site and its
consumers. Consequently, through these relationship-building processes, consumers will
likely respond to the Web service and release their personal and health information to it
in as truthful a manner as possible.

However, it is still far less certain if the relationship-building effort using
personalized social cues always facilitate consumers’ willingness to disclose their
personal and health information. For example, Sproull, Subramani, Kiesler, Walker, and
Waters (1996) found that the effectiveness of a talking-face computer-administered
interview was moderated by gender difference. In their experiment, participants
expressed a more positive attitude toward a talking-face computer-administered interview
than a text—based interview, but female participants wrote more about themselves in the
no-talking text display condition while males wrote more in the talking computer-face
display condition. Furthermore, when combined, both female and male participants

responded more in the no-talking display condition than the talking display condition.

The findings may suggest that such social cues of the computer may unexpectedly
decrease consumers’ willingness to reveal their information under certain circumstances.
Such inverse or null relationships between personalized social cues of the

interface and consumers’ willingness to disclose may become more salient when they
interact with the kind of information that consumers are requested to reveal. For example,
Tourangeau, Couper, and Steiger (2003) found that personalized social cues in a
computer interface (e. g., photographs of the researcher and personalized statements and
feedback) reduced participants’ willingness to answer sensitive and embarrassing
questions. Similarly, in relation to online privacy concerns, Cranor, Reagle, and
Ackerman (1999) found that, when requested to be disclosed, certain types of personal
information were associated with different levels of comfort and sensitivity. Andreasen
(1970) also noted that personalization of mail surveys may reduce participants’ response
rates particularly when sensitive and embarrassing information is requested.

No research has empirically examined how personalized social cues of Web site
interfaces can instill psychological closeness and warmness to elicit consumers’
willingness to disclose their personal and health information to the Web site. Thus, this
study focuses on empirically investigating the effect of manipulating personalized social
cues of a Web site interface on consumers’ likelihood of disclosing information. Given
that many Web sites are utilize such personalized social cues through the immediate and
responsive computer agent, the current dissertation investigates whether and how
personalized social cues of a Web site interface motivate consumers to reveal their

personal and health information.

Specifically, this study first attempts to conceptualize the personalized social cues
of Web site interfaces as computer immediacy (LaRose & Whitten, 2000) within
Bandura’s (1991, 2001) social cognitive theoretical framework. Then, it investigates the
underlying mechanism of social cognitive processes of the effects of the personalized
social cues of immediacy on the likelihood of consumers revealing their information in
the context of a personal health record (PHR) Web service. Finally, it explores if the
information type (embarrassing vs. descriptive) that can be requested from Web sites
differently affects the social cognitive processes of the effects of the personalized social
cues on consumers’ willingness to disclose their personal and health information. By
proposing and testing a social cognitive model of information disclosure, this dissertation
seeks to explicate the social cognitive mechanisms of information disclosure in human-
computer interaction.

This dissertation is structured as follows. CHAPTER 1 introduces the concept of
computer immediacy and provides a theoretical foundation to propose a research model
for investigating the impact of manipulating personalized social cues of immediacy on
consumers’ willingness to reveal their personal and health information within Bandura’s
(1999, 2001) social cognitive theoretical framework. From a social cognitive perspective,
the concept of computer immediacy is also distinguished from other similar media
concepts and theories. It also reviews previous research on information disclosure in
human-computer interaction and Internet environments. Based on the review, CHAPTER
2 proposes an integrated model of information disclosure in human-computer interactions
and discusses the research hypotheses derived from it. CHAPTER 3 describes the method

to be used to test the research hypotheses. CHAPTER 4 presents the results of the

hypothesis testing. Finally, CHAPTER 5 provides discussions of the results presented in
CHAPTER 4, followed by theoretical and practical implications of the present study.

Research limitations and future study directions are also discussed.

CHAPTER 1

COMPUTER IMMEDIACY AND INFORMATION DISCLOSURE IN HUMAN-

COMPUTER INTERACTION

Defining Computer Immediacy

Computers are regarded as autonomous entities (e.g., Brock & Sulsky, 1994),
often described as “a well-trained English butler” who best knows consumers’ needs,
likes, and habits (N egroponte, 1995, p. 8). The entirely new form of social agent will
handle routine communications with consumers by answering their phones, recognizing
the callers’ voice and responding to everything they ask. Or in some cases, the computer
agent won't even ask.

The analogy of the computer as a human entity is not at all far-fetched. Indeed, it
is well established that computers in particular (Reeves & Nass, 1996) and interactive
technologies in general (LaRose & Bates, 1990) are capable of demonstrating personal
and social characteristics of reciprocal and immediate communications with humans.
Sherry Turkle’s Second Self ethnography (1984, 1995) further suggests that computers
function as extensions of ourselves. By equating the health of one’s computer with
personal health safety, Lee, LaRose, and Rifon (in press) also extended the issue of
computer virus threats to the human domain of computers’ self-protection. As the
computer is an extension of the self rather than an ordinary consumer electronics device,
the relevance of interpersonal social relationships between computer and human is more

than metaphorical, particularly in the age of the network society (cf. van Dijk, 1999). As

a result, human-computer interaction will strongly parallel the domain of human-human
relationships.

Indeed, computers are often viewed as cooperative partners (Clarke & Smyth,
1993; Johnstone, Berry, & Nguyen, 1994), coaches (Desmarais, Giroux, & Larochelle,
1993), secretaries (Bocionek, 1995), and teammates (Nass, Fogg, & Moon, 1996),
particularly when computers provide positive social cues to people, such as praising them,
addressing them by first name, giving prompt and frequent responses, etc. These findings
suggest that people may sense a positive social environment in human-computer
interactions, especially when positive, personalized social feedback is incorporated into
the computer interface design (Johnson, Gardner, & Wiles, 2004).

However, conversely people also can feel betrayed by and frustrated with
computers (Ceaparu, Lazar, Bessiere, Robinson, & Shneiderman, 2004; Lazar, Bessiere,
Ceaparu, Robinson, & Shneiderman, 2003). They even irately express anger to the
computer as well (Ferdig & Mishra, 2004), especially when computers exhibit negative
social cues, such as curt messages from the computer or its network, impersonal
notifications of error messages, slow network connections, abrupt refusals to perform
routine tasks, delayed feedback to inquiries, etc. As such, positive (e.g., feelings of
warmness, closeness, excitement, and satisfaction) and negative (e.g., feelings of
isolation, frustration, alienation, and anxiousness) responses to computers can come from
the computer users’ experiences with the personalized social cues that computers provide
(Rozell & Gardner, 2000). Subsequently, such experiences produce positive or negative
expectations about the computer itself, which will influence their intentions to use the

computer in the future. In this sense, only positive social experiences with personalized

social cues of the computer interface will likely promote positive social incentives to
enhance the feelings of psychological closeness and warmness. Eventually, these positive
social incentives will help build and develop a warm and friendly social relationship in

human-computer interaction.

Immediacy in Human-Human Communication

Within traditional social psychological and interpersonal communication
frameworks, such psychological social incentives to promote a close social relationship
between human and computer can be termed “immediacy” (Mehrabian, 1967;
Mehhrabian, 1971; Wiener & Mehhrabian, 1968), or more precisely “computer
immediacy” (LaRose & Whitten, 2000) in the context of human-computer interaction.
The concept of immediacy has been proposed to explain the effect of a given
psychological and physical distance on closeness between communicators. Originally
introduced by Mehrabian (1967), the term immediacy refers to a cluster of verbal and
nonverbal communication behaviors that enhance the “directness and intensity of
interaction between two entities” (p.325). That is, the more immediate the relational
behavioral cues are, the closer the communicators are psychologically as well as
physically. By expressing such immediate social behaviors, the participants increase the
level of sensory involvement in their relationship (Burgoon, Buller, & Woodall, 1989).

The theoretical and empirical research has examined immediacy social cues as
both nonverbal (Mehrabian, 1971) and verbal (Wiener & Mehrabian, 1968). Nonverbal
indicators of immediacy cues may be conveyed by the combinations of eye contact,

relaxed body position, body-lean behaviors, proximity, smiling and vocal expressiveness

10

(Burgoon, Buller, Hale & DeTurck, 1984), whereas verbal indicators may be carried by
spatiotemporally indicative demonstratives, denotative specificity, selective emphasis,
including use of humor, praise, asking questions that show interest and attentive listening,
and subtle variations of word choice using inclusive pronouns, such as “us” or “our”
(Wiener & Mehrabian, 1968). In particular, research on immediacy suggests that the
verbal channel can be more effective in relational closeness because it not only directly
conveys immediacy to the partners, but also may compensate for other nonverbal
channels’ immediacy reduction within the framework of equilibrium theory (Argyle &
Dean, 1965; Wiener & Mehrabian, 1968).

In most cases, these social cues of immediacy indicate behavioral approachability
that can enhance interpersonal closeness and warmth between communicators within the
verbal and nonverbal cues that convey one’s thoughts and feelings to another. For
example, Burgoon et al. (1984) found that high level of immediacy behaviors conveyed
greater intimacy, attraction, and social trust. Thus, immediacy behaviors should be
understood as social behavioral cues that not only increase physical and/or psychological
proximity in distance, but also contribute to perceptual involvement during interpersonal
interaction (Richmond & McCroskey, 2000; Richmond, Lane, & McCroskey, 2006).
Consequently, the closer two people are to each other, the more immediate their
interaction.

Immediacy has been intensively studied with respect to its effects on the
relationship between teacher and student in various instructional settings (e.g., Allen,
Witt, & Wheeless, 2006; Christophel, 1990; Christophel & Gorham, 1995; Frymier,

1994; Gorham, 1988; Kearney, Plax & Wendt-Wasco, 1985; Kelley & Gorham, 1988;

ll

Mottet & Beebe, 2006; Mottet, Parker-Raley, Beebe, & Cunningham, 2007; Richmond,
Gorham & McCroskey, 1987; Richmond, McCroskey, Plax, & Kearney, 1986;
Rodriguez, Plax, & Kearney, 1996; Zhang & Oetzel, 2006). In particular, research on
immediacy has consistently found that instructors’ immediacy behaviors are positively
associated with positive student affect toward instructors (Chesebro, 2003; Gorham,
1988; Myers, Martin, & Knapp, 2005; Richmond, 1990; Titsworth, 2001), instructor trust
(J aasma & Koper, 1999) and credibility (Myers, 2004), and positive evaluation ratings of
instructors (Abrarni, Leventhal, & Perry, 1982; Schrod & Witt; 2006; Teven & Hanson,
2004).

Previous research also found that the personalized approach of teacher immediacy
tends to involve students and teachers engaging in social exchanges and reciprocal self-
disclosure (e.g., Myers et al., 2005; Aylor & Oppliger, 2003; Fusani, 1994; Jaasma &
Koper, 1999; Myers, 2004; Myers, et al., 2005; Nadler & Nadler, 2000; Waldeck, 2006).
In other words, if students perceive their instructors as immediate and responsive,
students may be more motivated to engage in such communication with instructors
beyond the classroom (Aylor & Oppliger, 2003; J assma & Koper, 1999; Nadler & N adler,
2000). Because the teacher-student interaction may be defined as a form of interpersonal
communication (Dobransky & Frymier, 2004, Frymier & Houser, 2000; DeVito, 1986),
students may interpret the instructor’s immediacy as one way that the instructor
welcomes students’ participation, interaction, and involvement in communication. In
such interpersonal relationships, trust exists and communicators manifest specific social
behaviors that indicate certain levels of reliance or dependence on one another. As such,

immediacy behaviors not only enhance a trusting relationship between students and

12

instructors, but also motivate students to communicate with their instructors for relational
reasons. Thus, teacher immediacy is relevant to the current discussion in that it highlights
the effects of social cues of immediacy on trust building and disclosure behaviors

between communicators.

Computer Immediacy in Human-Computer Interaction

In the context of Web education, LaRose & Whitten (2000) introduced the term
“computer immediacy” and suggested that the concept of immediacy can be extended to
the domain of human-computer interaction. They argued that computers can be verbally
and nonverbally versatile in offering a cluster of personalized social cues and such social
cues can exert social environmental influences on building and enhancing social
relationships with users. Drawing on Bandura’s (1986) Social Cognitive Theory (SCT),
they proposed that the computer behaves like a social actor that is capable of expressing a
form of social interest that could evoke such feelings of closeness and warmness.

LaRose & Whitten (2000) also hypothesized that the personalized social cues of
immediacy in Web sites can promote social incentives to advance a valued social rapport
with users. For example, the social incentives of computer immediacy could include
expressions of social approval and support (e.g., “praised my action and comments,”
“provided feedback on my action and comments”), indication of social interest in the user
(e. g., “called on me by name”), invitation to enhance the status of the user’ social
affiliation thorough interaction with the computer (e.g., “invited me to visit a link only
those with a higher membership status can enter”), and disclosures of personal

information about the computer (e. g., “used personal anecdotes of system operation and

13

maintenance”). Richardson and Swan (2003) noted that such personalized social cues can
provide the needed social incentives that can encourage users to engage in social
interaction with Web sites and to build a psychologically safe online environment. This
social role is often shared with or transferred to consumers so that it can promote a high
degree of closeness and comfort that consumers desire to receive. Consequently, such
personalized social cues of computer immediacy can offer strong and positive
psychological social affect and build a warm and friendly Web environment that
continues to motivate users to interact with Web sites.

Further, LaRose and Whitten (2000) argued that the personalized social cues of
computer immediacy can reinforce consumers’ outcome expectations or judgments of the
likely consequences of human behavior (Bandura, 1986). That is, through interaction
with more immediate social environmental cues, users can build some forms of positive
social outcome expectations of social approval, support, and affiliation that should be
taken into account for their future interaction with the Web site. In other words, people
can assess the potential outcomes of their social behavior by directly experiencing
(enactive learning) and indirectly observing others’ interactions (observational learning)
with the Web site. These outcome expectations of social trust can play a pivotal role
when people do a self-reactive behavior while interacting with the social cues of the Web
site.

However, it is important to note that such outcome expectations do not always
produce positive social behavior. People can develop expectations of negative social
consequences online, such as privacy violations, as well. The expectations of negative

consequences can be enhanced depending upon their self-efficacy beliefs in their

14

capabilities of executing a course of action with the computer (Compeau & Higgins,
1995), the Internet (LaRose & Eastin, 2004), and specifically behaviors protecting against
privacy invasions (LaRose & Rifon, 2007) and virus attacks (Lee et al., in press). As such,
as autonomous beings, humans can interpret the meanings of social environmental cues

in various social situations and regulate their social behavior by predicting the expected
outcomes in advance. Clearly, this agentic perspective of computer immediacy highlights
Bandura’s (1991) self-regulatory mechanisms of human motivation and behavior in the
domain of human-computer interaction.

Thus, the present study defines the concept of computer immediacy as the
personalized social cues of Web site interfaces that enhance consumers’ feelings of
psychological closeness and warmness. Also, it theoretically places the concept within
Bandura’s (1991, 2001) SCT and seeks to explain how the social environmental influence
of the computer interface interacts with human cognition and behavior. It further
highlights self-regulatory mechanisms where human cognition plays a central role in

moderating one’s behavior when interacting with personalized social cues of the interface.

Computer Immediacy in Other Media Theories

The importance of personalized social cues of immediacy in relationship building
has been also highlighted in the context of computer-mediated communication (CMC). In
proposing the Social Information Processing (SIP) model, Walther (1992) argued that
text-based immediacy cues are enough to initiate interpersonal impression formation (of,
Goffman, 1959) and build relational growth as nonverbal social cues do in face-to-face

interaction. In other words, the lack of nonverbal cues is not necessarily detrimental to

15

relational development because people have inherent needs of affiliation and are likely to
develop relationships with communicators based on whatever social cues are available to
them in a given context. By reading existing text-based social cues more carefully,
communicators can foster a warm and friendly atmosphere for promoting relational
growth, given sufficient time (Walther & Burgoon, 1992), previous interaction history
(Walther, Slovacek & Tidwell, 2001), and anticipation of future interaction (Walther,
1994). Thus, SIP highlighted the effectiveness of text-based social cues on relationship
development.

From the social cognitive perspective, the dynamic role of immediacy in initiating
interpersonal impression formation and enhancing relational growth in CMC can be
extended to the domain of human-computer interaction. That is, social cues of immediacy
on a Web site interface evoke psychological closeness and building relational depth as
well as in human-computer interaction. For example, computer users are likely to expect
social incentives from their past experience with personalized social cues of the computer
and anticipate the feelings of psychological closeness and warmness from their previous
experiences. Consequently, such social incentives foster relational development with the
computer across time.

Computer immediacy is also implicit in the concept of social presence, which is
defined as “the degree of salience of the other person in a mediated communication and
the consequent salience of their interpersonal interactions” (Short, Williams & Christie,
1976, p. 65). Originally devised to describe characteristics of teleconferencing, social
presence indicates the degree to which a medium facilitates awareness of the other person

behind the medium during mediated communication. In fact, Short et a1. (1976)

16

suggested that immediacy in mediated communication may create and promote the
degree of the sense of “being with another” which, in turn, increases the level of intimacy
(cf. Argyle & Dean, 1965) between the communicators (Short et al., 1976). Similarly,
Danchak, Walther, and Swan (2001) showed that the level of intimacy was a function of
immediacy behaviors while holding the level of bandwidth of the medium constant.
These studies suggest that immediacy may affect intimacy directly or indirectly through
social presence.

From the social cognitive perspective, however, social presence may not be a
determining factor in developing social rapport between computer users and computers,
particularly in the context of human-computer interaction. This is because social presence
alone is not likely to provide social incentives to promote psychological closeness or
warmness that would lead to a valued relationship between the medium and its users. For
example, even if the computer interface creates a high degree of social presence with rich
modes of text, audio and video, the mere perception of ”someone is there” behind the
computer would not invoke psychological closeness and warmness to develop a
personalized rapport between the computer and its users. Rather it may only increase the
awareness of the salient other that is simply present or absent. For this reason, social
presence has been primarily examined in mediated communication that influences
individuals’ decisions to choose an appropriate medium when sending different types of
messages to increase efficiency and satisfaction (Rice, 1993; Short et al., 1976).

Moreover, the concept of social presence seems problematic in its
conceptualization and operationalization. Biocca, Harms, and Burgoon (2004) pointed

out that the definition of social presence is “vague, overly broad, or circular” because it

17

fails to specify the boundary dimensions of the construct, often leading to confounding
relationship between the correlates and effects of social presence. Besides, it is still
unclear if social presence should be conceptualized and measured as a judgment about
one’s fleeting psychological state within a medium or a stable property of the medium.
For example, while Short et al (1976) measured social presence as a property of the
medium (e.g., unsociable-sociable, insensitive-sensitive, cold-warm, and impersonal-
personal), Lee and Nass (2005) and Nowak (2000) measured social presence as one’s
fluctuating psychological state using previous indicators of intimacy, involvement and
immediacy (e.g., interesting, involving, intimate, etc). As such, social presence is not
only confounded with other related variables, but also inconclusive in terms of its
dimensions. This unclear conceptualization further deters successful establishment of
operationalization and measurement (Biocca et al., 2004).

In sum, computer immediacy provides social incentives to promote relational
depth in human-computer interaction. In this regard, the present study argues that
computer immediacy should be recognized as a more fundamental and legitimate

construct that may directly affect the relational depth between human and computer.

Understanding Information Disclosure in Human-Computer Interaction
Information Disclosure in Human-Computer Interaction
Humans’ psychological tendency to behave socially with a computer is well
documented in the literature on human—computer interaction. According to the
Computers-Are-Social-Actors (CASA) perspective, also known as the media equation,

individuals treat computers as human beings as if computer machines possess human-like

18

identities, feelings, intentions, or motivations (Nass & Moon, 2000; Reeves & Nass,
1996). That is, people automatically apply social heuristics and rules and exhibit
interpersonal social behaviors in their interactions with computers as if they were
interacting with real human beings. When primary social cues of human characteristics
(e.g., face, voice, gender, etc) are incorporated within the computer interface, these cues
mindlessly invoke schemata inherent in human-human interaction without
psychologically constructing a relevant human (Nass & Moon, 2000). From this point of
view, the human-computer relationship is fundamentally social and natural.

One possible explanation for the mindless socialness of human-computer
interaction is that humans are social animals who have an innate desire for a social
orientation (Nass & Moon, 2000; Reeves & Nass, 1996). This evolutionary explanation
bodes well when individuals interact with an intelligent computer with human—like
features and capabilities, such as language exchange, tum-taking ability, and immediate
interaction (Moon, 2000). Because such social cues invoke the most immediate and
responsive interpersonal communication actor of the typical human-human relationship,
individuals are likely to make the social attributions to their interaction with the computer
automatically (Reeves & Nass, 1996). As a result, individuals tend to display socially
reflexive responses to computers without consciously thinking of the computer as a
machine.

Another possibility is that people tend to avoid extensive information processing
and exercise convenient mental shortcuts for their mindless social orientation (Nass &
Moon, 2000). In this sense, people tend to be “cognitive misers” (Fiske & Taylor, 1984),

“lazy organisms” (McGuire, 1969), “economy minded souls” (Eagly & Chaiken, 1993),

19

and “lazy information processors” (Nass & Moon, 2000). By making immediate social
attribution to the computer machine for the actual communication actor behind the
computer, individuals engage in interpersonal behavior with the computer (cf. Langer,
1989) even when such social response is not rational in the given human—computer
interaction (Moon & Nass, 2000; Reeves & Nass, 1996). In this process, people respond
to computers in much the same way as they respond to other human beings in human-
human relationships. Consequently, individuals use the same social rules and conventions
of the typical human-human interaction when they assess and respond to the performance
of computers.

A substantial number of experiments have demonstrated strong evidence to
indicate the social formation of human-computer interaction in social domains, such as
politeness (Nass, Moon, & Carney, 1999; Reeves & Nass, 1996), flattery (Fogg & Nass,
1996), gender stereotyping (Nass, Moon, & Green, 1997; Lee, Nass, & Brave, 2000),
personality type (Moon & Nass, 1996), similarity attraction (Nass & Lee, 2001), and
intimate self-disclosure (Moon, 2000). Similarly, recent studies found that participants
exhibited greater liking of and trust in the computer agent when they interacted with a
computer interface exhibiting social and communicative characteristics of praise
(Bracken & Lombard, 2004) and caring (Brave, Nass, & Hutchinson, 2005; Lee, Nass,
Brave, Morishima, Nakajima, & Yamada, 2007). These findings suggest that
communication between humans and interactive media should be understood as a new
type of interpersonal interaction as well as mediated communication (Bracken &

Lombard, 2004).

20

Notably, based on the CASA model (Moon & Nass, 1996; Reeves & Nass, 1996),
Moon (2000) applied the principle of reciprocity in disclosure exchange (e.g., Chaikin &
Derlega, 1974; Cialdini, 1993). The study demonstrated that computers can elicit
consumers’ intimate information disclosures by revealing information about the
computers to the consumers in advance. That is, people who receive disclosures from the
computer are more likely to feel obligated to respond to the conversational partner of the
computer. Moon (2000) further highlighted the disclosure-attraction relationship that
people tend to be more attracted to those who disclose to them so that they reveal their
information to the conversational partner with an equal amount and degree of intimacy of
the disclosure. As a result, she concluded that computers’ self-disclosure enhanced social
attractiveness of the computer, which, in turn, promoted consumers’ disclosures.

However, Moon (2000) may have overlooked the role of individuals’ cognitive
processes in enacting the self-disclosure behavior. Rather, it may have confined this
domain of self-disclosure to “one notable exception” of the reciprocity effect as Moon
(2000, p. 324) acknowledged in her study. In other words, it may have demonstrated just
one particular social rule of the reciprocity effect that disclosure automatically begets
reciprocal disclosure and emphasized the reactive process of the attraction-disclosure
relationship in self-disclosure. Such process of disclosure may not always be generalized
in human-computer interaction because it does not address other possible social and
cognitive processes that mostly govern the extent to which it is appropriate to reveal
information in general social contexts (e.g., Chaikin & Derelga, 1974; Taylor & Altman,

1975).

21

Indeed, Moon (2000) noted that the reciprocity effect of human computer
interaction may be not be realistic because it does not address other important factors that
motivate such human behavior, such as anonymity or privacy (e. g., Rosenfeld, Booth-
Kewley, Edwards, & Thomas, 1996), trust (e.g., Wheeless, 1978), and the rewards and
risks offered in exchange for disclosure (e.g., Altman & Taylor, 1973; White, 2004),
which could influence the extent to which people are motivated to respond to computers.

In addition, she also pointed out that the study only addressed “self-disclosure
dynamics under specific circumstances” where the participants automatically follow the
norms of socially appropriate self-disclosure behavior due to the effect of reciprocity in a
controlled, computer-based interviewing situation. The participants may have fallen back
on the heuristics of their “lazy” mind (cf. McGuire, 1969; Nass & Moon, 2000) at the
point in time where their self-disclosure behavior was automatically evoked in the one-
shot laboratory experiment setting, free from distractions and from threats to privacy. In
other words, the participants in the experiment may have had low involvement in the
communication subjects and, accordingly, perceived minimal risks to privacy so that they
may have ignored any differences in the consequences between their revelations to a
human and to a computer.

Another possibility is that the participants may have paid more attention to the
experimental manipulation in the controlled environment of the laboratory (Tourangeau
at al., 2003). In an uncontrolled setting, such as the Internet, the participants could have
recognized the potential consequences of risks and benefits associated with their

disclosure behavior rather than the manipulation of the study itself.

22

Social Cognitive Processes of Information Disclosure

The self-disclosure research tradition has paid much attention to the social
exchange processes where personal information disclosure is a function of the perceived
potential risks as well as benefits associated with the behavior (Altman & Taylor, 1973;
Collins & Miller, 1994; Cozby, 1973; Jourard, 1971). Based on social exchange theory
(Emerson, 1976; Homans, 1958; Thibaut & Kelly, 1959), much of the research assumes
that people decide whether or not to carry out self-disclosure based on an assessment of
the interpersonal rewards and costs associated with the social act. That is, in a real world
setting, disclosure entails one’s social cognitive assessment of the potential positive and
negative consequences associated with the social act within the context of the social
relationship. In short, self-disclosure is a medium of social exchange that involves the
cognitive process of balancing social benefits and costs.

For example, disclosing personal information is related to the benefit of social
rapport building (Gibbs, Ellison, & Heino, 2006; Cozby, 1973; Wheeless, 1978;
Wheeless & Grotz, 1976). By making a disclosure to another person, mutual
understanding is increased (Laurenceau, Barrett, & Pietromonaco, 1998) and trust is
established (Rubin, 1975). However, when personal information is disclosed to another,
relational risk of increased vulnerability to privacy can follow as well (Cline, 1989). In
other words, when disclosing people normally weigh what they will socially gain or lose
associated with the result of the disclosure. Those social benefits often include the
potential consequences of the social rewards of validation, control, and close relationship
development (Derelga & Grzelak, 1979) and those social costs often include the

possibility of social risks such as being rejected, hurt, or embarrassed by the other

23

communication party (Baxter and Mongomery, 1996; Derlega & Chaikin, 1977). In most
cases, people are likely to avoid divulging their personal information particularly when
they expect high-risk of the loss of their control or being in a vulnerable position (Baxter
& Montgomery, 1996; Derlega & Chaikin, 1977).

The traditional interpersonal theory of self-disclosure has been applied to the
issue of consumers’ information disclosure, particularly in the context of e-commerce.
Much privacy research has investigated consumers’ disclosure behavior based on the
social exchange perspective (Emerson, 1976; Homans, 1961; Thibaut & Kelly, 1959) and
has suggested that information disclosure, like other interpersonal behaviors, is engaged
in and interpreted in terms of the social and economic costs and benefits of providing
information.

Similarly, Laufer and Wolfe (1977) proposed a model of privacy behavior and
suggested that the decision of whether or not to provide personal information largely
depends on a cognitive assessment of the benefits of convenient personalized/customized
services and the risks of information privacy concerns, such as transfer of their data to
third parties or use of their data for marketing solicitation. In this cognitive assessment
process, consumers consider the trade-off between the potential social and economic
benefits and costs of interactions.

Following Laufer and Wolfe (1977), Culnan and Armstrong (1999, p.106) also
argued that consumers’ willingness to provide personal information involves the process
of “privacy calculus” that assesses the possibility that “their personal information will
subsequently be used fairly and they will not suffer negative consequences”. The calculus

model is rooted in the notion of fair information practices that provide individuals with

24

the ability to control the risks of information privacy invasions by second and third
parties (Milne & Culnan, 2002). When consumers are informed by the seller’s
information practices that their information will not be misused, they are more willing to
provide personal information. Conversely, without the assurance of such fair information
practices, individuals are less likely to disclose personal information (Culnan &
Armstrong, 1999).

The privacy-related risks associated with disclosure may become more evident
within an e-commerce context. Indeed, the potential risk to privacy has been one of the
critical barriers to consumers’ willingness to disclose personal information online (Awad
& Krishnan, 2006; Dinev & Hart, 2006a, 2006b; Hann, et al., 2007; Hui et al., 2007;
LaRose & Rifon, 2007; Metzger, 2004). Surveys also show that consumers are very
concerned about threats to their online privacy because their information can be accessed
by, sold to, used for external third parties without permission (e.g., Cyber Security
Industry Alliance, 2006; Milne, Rohm, & Bahl, 2004).

In the online privacy literature, several studies have focused on the relationship
between privacy concerns and information disclosure (Awad & Krishnan, 2006; Dinev &
Hart, 2006a, 2006b; Hann, et al., 2007; Hui et al., 2007; LaRose & Rifon, 2007; Metzger,
2004). Dinev and Hart (2006a) refined the calculus model (Culnan & Armstrong, 1999)
and argued that consumers tend to maximize the difference of such costs and benefits of
providing information online. They found that perceived risk of privacy violations
decreased consumers’ willingness to provide personal information to transact on the

Internet.

25

In focusing on consumer and retailer relationships in particular, Culnan and Bies
(2003) also argued that individuals disclose personal information if they perceive that the
overall benefits of disclosure are at least balanced by, if not greater than, the assumed risk
of privacy invasions. They further argued that people are more likely to accept the loss of
privacy that accompanies any disclosure of personal information as long as an acceptable
level of risk accompanies the benefits” (Culnan & Bies, 2003, p. 327). Milne & Culnan
(2004) also noted that consumers’ information disclosure behavior are said to be weighed
in relation to monetary benefits and to privacy concerns (Milne & Culnan, 2004).
Similarly, McKnight, Choudhury, and Kacmar (2002) and Malhotra, Kim, and Agarwal
(2004) found that the risk of privacy violations was related to perceived uncertainty and
the disclosure of personal information.

Personal health record (PHR) services on the Internet also raise serious concerns
about privacy (Rothstein & Talbott, 2006; Sprague, 2006; Peel, 2007). Due to the vast
amount of personal information PHRs contains and the absence of federal laws
guaranteeing consumers’ right to privacy, PHRs are an attractive target for data mining
(Sprague, 2006; Peel, 2007). Numerous third parties such as health-related corporations,
government agencies, employers, insurers, and banks can access and use individuals’
PHRs (Peel, 2007). In addition, insecure Internet connections for remote access,
ingenious hackers, careless clerks, and technical limitations for individuals to limit the
scope of disclosures of their health information based on the type of information needed
by the third party further threaten privacy issues in the PHR domain (Rothstein & Talbott,

2006; Sprague, 2006).

26

It is important to note that privacy risks raised in the PHR domain may go beyond
the privacy risks in the e-commerce context. The consequence of privacy issues in the
PHR context can be adverse and multi-dimensional. The PHR is often a longitudinal
record that contains disparate reports and wide-ranging health conditions including
sensitive health information (e. g. mental health, genetic test results, sexually transmitted
diseases, HIV antibody status, sexual history, history of abortions and other reproductive
matters, domestic violence, and drug and alcohol abuse) (Rothstein & Talbott, 2006).
When such sensitive health information is accessed without individuals’ consent and
misused, the individuals may experience embarrassment, humiliation, shame, anxiety,
and depression (Rothstein & Talbott, 2006). Further, the individuals may lose
opportunities in their life: people with stigmatized or costly illnesses may become
incapable to obtain insurance or employment (Rothstein & T albott, 2006; Peel, 2007).

As such, disclosures of personal information online can be very consequential.
Unlike the “unreal” controlled setting of many human-computer interaction experiments
(e. g., Moon, 2000), people disclose their personal information within “real” natural
situations where they interact with Web sites on the Internet that have a far greater
capacity to spread information to both strangers and intimates who can harm people in
very serious ways. In this sense, the Internet possesses the capability of collecting, storing,
and distributing consumers’ personal information from individual computers that do not
necessarily evoke these potential threats to privacy.

It is important to note that consumers’ willingness to disclose their information
online involves their assessment of the positive outcomes of customization benefits and

social trust as well (Awad & Krshnan, 2006). For example, marketers not only offer their

27

customized and personalized products and services to motivate consumers’ information
disclosure (e.g., Chellappa & Sin, 2005; Culnan & Bies, 2003; Milne, Bahl, & Rohm, ,
2008) but also as attempts to build a trusting relationship with consumers that can
positively influence consumers’ information disclosure (e.g., Bart, Shankar, Sultan, &
Urban, 2005; Dinev & Hart, 2006a; Metzger, 2004; Earp & Baumer, 2003; Norberg,
Horne, & Home, 2007). In this sense, consumers’ disclosures should be positively
related to their assessment of the positive outcomes of the benefits of social trust
associated with their disclosure behavior.

In particular, consumers’ expectations of social trust toward Web sites may be an
important determinant for their information disclosure. Metzger (2004) noted that such
expectations of trust have a particularly important role in facilitating disclosure of
information in the Internet because social cues are reduced and any physical social
contact is replaced by computer-mediated communication. In short, the more consumers
expect a trusting relationship with Web sites, the more willing consumers may be to
disclose their information.

Previous research found that trust is positively related to consumer’s information
disclosure behavior (Bart et al., 2005, Dinev & Hart, 2006a; Earp & Baumer 2003; Gefen,
Karahanna, & Straub, 2003; Metzger, 2004; Milne & Boza, 1999; Norberg et al., 2007;
Schoenbachler & Gorden, 2002). For example, Norberg et al. (2007) found that when
consumers’ trust toward a Web site was salient, consumers were willing to provide
personal information. Schoenbachler and Gorden (2002) and Dinev & Hart (2006a) also
found trust increased consumers’ willingness to provide personal information. Similarly,

Earp and Baumer (2003) viewed the status of a company’s brand name in relation to trust

28

and found that consumers were likely to disclose personal information to well-known
companies. Gefen et al. (2003) also demonstrated the effect of a set of trust beliefs
(including a vendor’s honesty, caring for customers, and predictability) on an individual’s
intention to disclose information to complete an online transaction. These findings
suggest that dynamic trust building between consumers and Web sites is highly critical
for promoting consumers’ information disclosure because it affects consumers’
expectations that Web sites will provide the necessary conditions of social trust, such as
competence, benevolence, and integrity (e.g., Gefen et al., 2003) to rely upon and their
subsequent intention to provide their information. In this sense, consumers’ information
disclosures should be understood as a decision to carry social risks as well as rewards,
particularly in the Internet.

In summary, parallel to the basic premise of the CASA (Nass & Moon, 2000;
Reeves & Nass, 1996) perspective, the concept of computer immediacy assumes that
humans respond to computers in the same manner as they would toward other people,
especially when the computers manifest their personalities with a minimal set of social
cues (Nass & Moon, 2000; Reeves & Nass, 1996). In other words, humans socialize with
the computer, just as they would with another human, without much cognitive effort
when their social behavior is socially evoked by social cues of the computer interface.

However, at the same time, the concept of computer immediacy also assumes that
humans can be mindful of potential positive and negative consequences associated with
their social behavior, particularly when they interact with Web sites on the lntemet. In
such online environments, consumers tend to cognitively assess the potential positive

(e.g., social trust and customization benefits) and negative consequences (e.g., privacy

29

violations) of the disclosure behavior that could be very consequential to themselves in
the future. Thus, consumers' likelihood of information disclosure in human—computer
interaction on the Internet may have to be understood as a social behavior requiring social
cognitive efforts beyond the automatic social response to the computer itself.

Based on this reasoning, this dissertation extends the CASA model of human-
computer interaction and proposes a social cognitive model of information disclosure in
human-computer interaction. Specifically, this model seeks to address the question of
whether the computer-mediated personalized social cues of the Web site interface that
manifest immediacy can elicit consumers’ information disclosure behavior.

In this sense, the domain of information behavior in human-computer interaction on the
Internet may be beyond the initial social response to the computer itself that has been

suggested in the CASA model.

30

CHAPTER 2

INTEGRATED MODEL OF INFORMATION DISCLOSURE AND HYPOTHESES

A Social Cognitive Perspective of Information Disclosure in Human-Computer
Interaction

The present dissertation extends the CASA model of human-computer interaction
(Nass & Moon, 1996; Reeves & Nass; 1996) and proposes that information disclosure in
human-computer interaction should be understood as a social cognitive process that
underlies reciprocal interactions among human cognition, behavior, and social
environment influences.

This study applies Bandura’s (1991, 2001) Social Cognitive Theory (SCT) to the
domain of human-computer interaction and tests a model of information disclosure
behavior that describes human motivation for disclosure as socio-cognitive processes that
involve three major causal factors of social environmental influences, cognitive factors,
and social behavior. In the proposed model, it is expected that information disclosure in
human-computer interaction will be a function of personalized social cues of computer
immediacy as well as consumers’ cognitive evaluation of the benefits and risks associated
with information disclosure behavior. In the model, information disclosure will also be
directly influenced by consumers’ self-efficacy judgment of their capabilities to cope
with the potential risks of online privacy. Originally defined as the belief “in one‘s
capabilities to organize and execute the courses of action required to produce given

attainments” (Bandura, 1997, p. 3), self-efficacy has been recognized as an important

31

factor affecting human motivation to exert a course of behavior. The judgment of having
self-efficacy to cope with the risks of privacy will also directly influence the benefit and
risk perceptions associated with information disclosure.

SCT also highlights the moderating role of human cognition in human behavior.
That is, human agents intentionally make things happen through their actions by
exercising forethought, reflecting on their behavior, and applying self-reactive influences
(Bandura, 2001). Unlike the conventional stimulus-response model of motivation for
human behavior, this agentic perspective emphasizes the importance of cognitive
guidance as an influence of behavior. In this sense, the triadic reciprocal casual model is
mediated by humans’ self-regulatory capabilities to reflect on their behavior.

In fact, using the triadic reciprocal causation model, SCT claims that human
motivation for behavior is largely determined by mutual interactions of behavior,
cognition, and social environmental influences (Bandura, 1991). In the present context of
information disclosure in human-computer interaction, the three interdependent factors
are consumers’ information disclosure (behavior), their assessment of the likely positive
and negative consequences of the disclosure behavior and their perceived ability to
manage their personal and health information (cognition), and personalized social cues of
immediacy that the Web site interface provides (social environmental influences)

In SCT, two primary cognitive mechanisms that have been proposed to motivate
human behavior are outcome expectations and self-efficacy judgments (Bandura, 1991).
Outcome expectations are defined as the perceived likely consequences of performing a
behavior. In addition, outcome expectations provide both incentives and disincentives for

determining a course of behavior. In other words, people tend to judge the likely

32

outcomes of a behavior before engaging in the behavior because the potential
consequences can provide incentives or disincentives to guide future behavior. In
particular, positive outcome expectations provide incentives that promote people’s future
behavior whereas negative outcomes expectations provide disincentives that discourage
such future behavior. As such, outcome expectations focus on the potential consequences
of social behavior within SCT.

This incentive system of outcome expectations can be external (e.g., social and
economic risks and benefits) as well as internal (e.g., personal risks and benefits). For
example, consumers’ online information disclosure may lead to negative consequences of
third-party disclosures such as spam attacks, identify theft, or phishing, , which may
evoke social and economic consequences of consumers’ embarrassment and information
abuse due to privacy violations by others online. Indeed, the potential risk of privacy
violations has been found to deter consumers’ willingness to disclose personal
information online (Awad & Krishnan, 2006; Dinev & Hart, 2006a, 2006b; Hann, et al.,
2007; Hui et al., 2007; LaRose & Rifon, 2007; Metzger, 2004).

At the same time, however, consumers’ disclosures online may be followed by
positive economic consequences of customized service benefits for convenience and
consumers’ expectations of positive social consequences of trust between consumers and
Web sites (e. g., social affiliation and support), particularly when the Web sites recognize
consumers’ names and other personal and health information. For example, marketers not
only tailor their customized service benefit offerings to facilitate the likelihood of
consumer’s information disclosure (e.g., Chellappa & Sin, 2005; Culnan & Bies, 2003;

Milne, Bahl, & Rohm, 2008), but also to build social trust thereby motivating consumers’

33

information disclosure (e.g., Bart et al., 2005; Dinev & Hart, 2006a; Metzger, 2004; Earp
& Baumer, 2003; Norberg et al., 2007). As such, consumers’ disclosures should be
positively related to their assessment of the positive outcomes of customization and social
trust associated with their disclosure behavior. Previous research found that trust is
positively related to consumer’s information disclosure behavior (Bart et al., 2005, Dinev
& Hart, 2006a; Earp & Baumer 2003; Gefen, Karahanna, & Straub, 2003; Metzger, 2004;
Milne & Boza, 1999; Norberg et al., 2007; Schoenbachler & Gorden, 2002).

Such positive and negative outcome expectations may be shaped by our
experiences with our behavior and other social environmental influences (Bandura, 1991).
In other words, people can develop positive (e. g., social trust and customized service
benefits) or negative outcome expectations (e.g., embarrassment and information abuse
due to privacy violations) through their actual disclosure behavior (enactive learning in
social cognitive terms), through their observation of others’ disclosure behavior
(observational learning), or by persuasive messages about the incentives or disincentives
associated with their disclosure behavior (verbal persuasion). Consequently, such direct
and indirect influences contribute to the formation of the outcome expectations of
benefits and risks, which, in turn, affect information disclosure behavior in human—
computer interaction.

In fact, consumers’ willingness to reveal their information online involves their
assessment of the positive and negative consequences of their specific information
disclosure to Web sites that they interact with (Awad & Krshinan, 2006). It is recognized
as equally important to highlight the benefits as well as risks associated with information

disclosures (Culnan & Bies, 2003; Ward et al., 2005; LaRose & Rifon, 2007; Rifon,

34

LaRose, & Choi; 2005). For example, marketers not only provide customized service
benefits to enhance consumers’ willingness to disclosure their information, but also
present privacy seals, such as TRUSTe and BBBOnline, to lessen the perceived risks of
privacy violations associated with consumers’ disclosure. Applying this reasoning to the

present context of online disclosure behavior, the following hypotheses are proposed:

H1: Consumers’ willingness to reveal their information will be negatively
related to their expectations of the negative consequence of
embarrassment

H2: Consumers’ willingness to reveal their information will be negatively
related to their expectations of the negative consequence of
information abuse

H3: Consumers’ willingness to reveal their information will be positively
related to their expectations of the positive consequence of social trust

H4: Consumers’ willingness to reveal their information will be positively
related to their expectations of the positive consequence of
customized service benefits.

In SCT, self-efficacy is another key factor influencing how a person engages in
behavior. Along with outcome expectations, it is regarded as another self-evaluative form
of judgment about one’s capabilities to execute a task. Thus, it should reflect the degree
to which individuals are confident about executing a behavior with the skill they have
rather than the measure of the skill that they possess. In this sense, those who perceive
themselves to be highly efficacious with reference to a particular task will be more likely
to invest sufficient levels of effort to achieve successful outcomes than those with low

levels of self-efficacy.

35

In the present context, this cognitive construct should be privacy coping self-
efficacy, an assessment of one’s capability to cope with the potential risk of privacy
violations associated with consumers’ disclosure of personal and health information.
Previous studies demonstrated that privacy self-efficacy is a causal antecedent to the
motivation to reveal information. (LaRose & Rifon, 2007; Rifon, LaRose, & Choi, 2005;
Lee et al., in press). The greater consumers’ confidence in their privacy protection
behaviors, the more likely they will be willing to reveal their personal and health

information. Based on this reasoning, the following hypothesis is stated:

H5: Consumers’ willingness to reveal their personal and health information will
be positively related to their privacy self-efficacy.

It is also important to note that self-efficacy is also regarded as a direct predictor
of outcome expectations. Such relationship demonstrates the causal relationship among
the cognitive constructs within SCT. Previous research on the use of computer (Compeau
& Higgins, 1995), the Internet (LaRose & Eastin; 2004), video games (Lee & LaRose,
2007), and online protection behaviors (LaRose & Rifon, 2007; Rifon, et al., 2005; Lee et
al., in press) suggest that the stronger a person’s self-efficacy judgment, the more likely
he or she is to attempt to achieve the desired outcomes, whereas the weaker a person’s
self-efficacy, the less likely he or she is to try to achieve the desired outcomes.

In the current context, this means that privacy self-efficacy should be positively
related to the expectation of positive consequences associated with information disclosure
whereas privacy self-efficacy should be negatively related to the expectation of negative

consequences associated with information disclosure. In other words, consumers with

36

high privacy self-efficacy are more likely to believe that they will develop and maintain
positive social outcomes of trust towards a Web site they interact with and receive
customized service benefits. At the same time, those with high privacy self-efficacy are
more likely to be confident in protecting themselves from potential privacy violations so
that they are less likely to be embarrassed and their personal and health information is

less likely to be abused. Thus, the following hypotheses are formally proposed:

H6: Consumers’ privacy self-efficacy will be negatively related to
their expectations of the negative consequence of
embarrassment.

H7: Consumers’ privacy self-efficacy will be negatively related to their
expectations of the negative consequence of information abuse.

H8: Consumers’ privacy self-efficacy will be positively related to their
expectations of the positive consequence of social trust.

H9: Consumers’ privacy self-efficacy will be positively related to their

expectations of the positive consequence of customized service
benefits.

A Social Cognitive Model of the Effect of Computer Immediacy on Information
Disclosure
The present study specifically focuses on the social cognitive process of how
socio-environmental influences of computer immediacy cues affect the likelihood that
consumers will disclose their personal and health information. It is expected that
computer immediacy provides social incentives that promote social relationships with
consumers, in turn motivating disclosures. For example, when consumers experience

personalized social cues of immediacy that a Web site offers, they are likely to perceive

37

psychological closeness and warmness evoked by the interface. Then, exposures to the
personalized social cues of immediacy will likely lead consumers to develop a keen sense
of social affiliation with the Web site and feel that the Web site wants them as their
members.

From the social cognitive perspective, personalized social cues of computer
immediacy may signify social acceptance by offering social incentives that consumers
inherently desire to receive as they do in human-human interaction. Such social cognitive
influences can instill a sense of social trust that leads to meaningful interaction between
consumers and Web sites. Through the social cognitive process, consumers may develop
expectations of greater social trust towards the Web site, which will, in turn, likely induce
consumers’ disclosures. Consequently, such relationship-building process will motivate
consumers to reveal their information to the Web site. This assumption is consistent with
Bickmore & Cassell’s (2001) idea that a computer agent can promote positive social
expectations and thereby promote a trusting relationship between the computer agent and
its user. As a result, consumers’ expectations of social trust evoked by computer
immediacy will likely motivate consumers to disclose their personal and health
information.

It is important to note that such social outcome expectations may parallel the
traditional concept of trust between consumers and marketers “as the expectancypf
customer to rely upon database marketers to treat the consumer’s personal information
fairly” (Milne & Boza, 1999, p.8). Such trust perception is based on consumers’

expectations that marketers would not abuse their information without permission.

38

Particularly in Internet research, the concept of trust has been extended to the concept of
e-trust between consumers and Web sites (Gefen, et al., 2003).

However, the concept of trust has suffered from theoretical and operational
confusion (e.g., Hosmer, 1995, McAllister, 1995; Shapiro, 1987). Essentially, there is not
an agreed upon definition of trust (Mayer, Davis, & Schoorrnan, 1995; Perry & Mankin,
2004). For example, Gefen et al. (2003) reviewed 43 papers that offer different
conceptualizations of trust and identified four core conceptualizations of the construct.
The most popular definition of trust involves specific beliefs in the integrity, benevolence,
and ability of another party (Gefen & Silver, 1999). A second conceptualization involves
a general belief in consumers’ willingness to depend on another party based on optimistic
expectations of the party’s morally correct behavior (Hosmer, 1995), often leading to
trusting intentions (McKnight, Cummings, & Chervany, 1998). A third conceptualization
involves psychological affect in the form of feeling confident and secure in the other
party (Rempel, Holmes, & Zanna, 1985). A fourth conceptualization involves a
combination of these definitions.

Notably, all these involve vague and imprecise conceptualizations of trust that
may possibly confound the measures of two important social cognitive constructs -
social outcome expectations and self-efficacy. In social cognitive terms, trust should be
construed as specific social expectations about the positive consequences of social
affiliations with and support from a specific Web site that would eventually promote
consumers’ willingness to disclose their information to the Web site._From the social
cognitive perspective, matching outcome expectations to a specific target behavior is

crucial for validating a theoretical model with explanatory power. Thus, it is expected

39

that personalized social cues of computer immediacy will increase expectations of
positive social outcomes of trust and increase the likelihood of personal disclosure. Here,
it is proposed that there is a causal relationship between social trust, more properly
conceived as expected positive social consequences of social trust, and consumers’

information disclosures.

H10: Consumers who view a Web site with high computer immediacy
cues will report stronger expectation of social trust than consumers
who view a Web site with low computer immediacy cues

Once the relationship between social outcome expectations of social trust are
established through consumers’ interaction with the Web site, consumers may further
form another positive outcome expectation; namely, that they can rely on the computer
agent to quickly respond to and customize their requests. For example, many Web sites
provide customers with personalized services and products, but require consumers to
register in order to access them. When personalized social cues promote feelings of social
closeness and warmness and are explicitly presented by the Web site interface,
consumers will recognize the uniqueness of personalized social cues and expect that
those cues will likely create a sense of technological congruence between the consumers
and the Web site. In turn, this will likely induce positive outcome expectations of

customized service benefits. Based on this reasoning, the following hypothesis is stated.

H11: Consumers who view a Web site with high computer immediacy
cues will report stronger expectations of customization benefits than
consumers those who view a Web site with low computer immediacy
cues

40

Personalized social cues of computer immediacy can also affect consumers’
assessment of the risks of privacy violations (e.g., embarrassment and information abuse)
associated with information disclosure (Awad & Krishnan, 2006; Dinev & Hart, 2006a,
2006b; Hann, et al., 2007; Hui et al., 2007; LaRose & Rifon, 2007; Metzger, 2004). In
other words, as consumers experience immediacy cues on a Web site, they tend to feel
psychological closeness with and warmth from the site, and develop some degree of
social relational depth with the Web site. Such positive social outcome expectations may
not only reduce the social risk of embarrassment as a consequence of revealing their
private and embarrassing information as it would in human-human interaction, but also
mitigate the economic risk of privacy violations if their valued personal information is
abused. Although previous studies found that Web sites’ privacy statements generally
comply with FTC guidelines and assure consumers’ privacy (Milne & Culnan 2002;
Peslak, 2005), these statements do not provide adequate information about the potential
positive and negative ramifications of the risks associated with the sharing of personal
information (LaRose & Rifon, 2007).

However, computer immediacy may not always lessen consumers’ assessment of
the amount of risk associated with their online information disclosures. In fact, it might
intensify their expectations of negative consequences. In particular, it is more likely that
consumers’ expectations of embarrassment can be heightened when the topic of
information requested to be disclosed is sensitive and embarrassing. In fact, in traditional
communication situations it is often more difficult to relinquish private and intimate
information to close friends and family than to strangers (Rubin, 1975). This proposition

that individuals may not always relinquish intimate information to those with whom they

41

are in close relationships may be applied to the domain of human-computer interaction.
For example, Moon (1998) found that people were less likely to disclose personal
information to geographically adjacent computer servers than distant servers because they
felt more vulnerable to nearby servers.

Several previous studies of interpersonal communication indicated that
development and maintenance of close relationships is strongly related with topic
avoidance and privacy regulation (e.g., Afifi & Burgoon, 1998; Afifi & Guerreo, 2000;
Argyle, T rimboli, & Forgas, 1988; Baxter & Wilmot, 1985; Petronio, 1991, 2000;
Petronio, Ellemers, Giles, & Gallois, 1998; Rawlins, 1992; Rollof & Ifert, 1998, 2000;
Rosenfeld, 1979). Afifi and Guerrero (2000) defined topic avoidance as purposefully
concealing information on a particular topic from a partner and argued that this
interpersonal process reflects people’s attempt to protect their face as a function of their
appraising of the degree of self threat posed by communicating about the topic. For
example, when individuals have close relationship with others, some topics are often
considered “taboo” and commonly avoided for self or relational protection reasons. These.
avoidance topics include past relationships, sexual experiences, outside-relational
activities, and negative-life experiences (Afifi & Guerreo, 2000; Baxter & Wilmot, 1985;
Rawlins, 1983). Baxter and Wilmot (1985) also argued that people tend to conceal these
kinds of information from close others because they often desire to protect themselves
from the other’s humiliation, shame, and conflict. Similarly, Altman (1993) also argued
that one of the primary reasons for avoiding topics from the other was due to self—
protection against the fear of humiliation. These findings may imply that such reluctance

may be specific to certain types of information and may occur as a function of the

42

perceived depth of the relationship between the two parties (Argyle et al, 1988; Barrel] &
Jourard, 1976; Rosenfeld, 1979; White, 2004).

Recently, White (2004) argued that the type of information (sensitive
embarrassing information v. descriptive privacy-related information) is closely related to
the two types of potential disclosure-related risks (embarrassment and information abuse)
and found that the information type moderated the effects of consumers’ relationship
perceptions with marketers on the potential consequences of embarrassment and
information abuse. Specifically, when consumers were requested to disclose descriptive
information about themselves, the relational depth between consumers and marketers
reduced the negative outcome expectations of information abuse associated with the
disclosures. However, when consumers were asked to reveal embarrassing information
about themselves, the relational depth between consumers and marketers increased the
negative outcome expectations of embarrassment associated with the disclosures.

Similarly, Matthews, Derlega, and Morrow (2006) noted that two types of self-
disclosure are differently related to two types of information: descriptive disclosure of
private facts versus relational self-disclosure of personal feelings or opinions (e.g.,
Derlega, Metts, Petronio, & Margulis, 1993; Morton, 1978). As such, conceptually,
descriptive information is more concerned with individuals’ private facts and
embarrassing information is more concerned with individuals’ feelings or opinions.
When online privacy concerns are taken into account, Cranor, Reagle, and Ackerman
(1999) also found that consumers demonstrated different levels of comfort when asked to
reveal certain types of personal information associated with different levels of sensitivity.

Specifically, they found that consumers were more comfortable when providing some

43

personal information (e.g., their favorite television show, age, full name, email address,
or postal address) than when revealing others (e. g., health information, phone numbers,
income level, or credit card numbers). Milne (2000) also found that some personal
information (e.g., telephone numbers) was more private and sensitive than others (e.g.,
purchase behaviors). Earp and Baumer (2003) also indicated that consumers were more
willing to reveal less sensitive information (e. g., their gender and age) than more
sensitive information (e.g., their identification numbers). These findings may suggest that
the information consumers are requested to reveal online should be related to the level of
sensitivity or privacy concerns.

In social cognitive terms, information type might be an indicator to activate
consumers’ self-observation system in the self-regulatory mechanism of human behavior
(Bandura, 2001). In the process, consumers can be attentive to their own self-disclosure
behavior (self-observation), judge their self-disclosure behavior in relation to social
standards or norms (normative judgmental process), and adjust their behavior to the
social environment by regulating their heightened state of attentiveness to positive (e. g.,
feelings of warmness and satisfaction such as social trust and customized benefits) and
negative (e. g. feelings of anxiousness and frustration stemming from social
embarrassment and privacy violations ) responses to Web sites (self-reactive influences).
Consequently, corresponding degrees of expectations of potential consequences can be
matched to information disclosure of different types of information.

Guided by the logic of the studies discussed above, this dissertation predicts that
the type of information (embarrassing vs. descriptive) may moderate the effect of

perceived psychological closeness evoked by computer immediacy on negative outcome

expectations of embarrassment and information abuse in the domain of human computer
interaction. In other words, consumers may experience heightened risks of being
embarrassed when requested to reveal their sensitive and embarrassing information to a
Web site that approaches consumers with many personalized social cues of immediacy.
This is because embarrassing information may be more likely to be associated with the
relational risk of unwanted judgment and opinions of personal information by others.
Conversely, descriptive information may be more likely associated with economic risk of
unwanted misuse of personal information by other parties (e.g., Derlega, Metts, Petronio,
& Margulis, 1993, Mathews, Derlega, & Morrow, 2006; Morton, 1978). Thus, when
consumers are requested to reveal embarrassing information, consumers who visit a Web
site with high immediacy will perceive greater expectations of negative consequences of
embarrassment than those who visit a Web site with low level of immediacy. Also, when
consumers are requested to reveal their descriptive information, consumers who browse a
Web site with a high level of personalized social cues will demonstrate lower
expectations of the negative outcome of information abuse than those who browse a Web
site with lower levels of personalized social cues. Based on this, the following hypotheses

are formally stated.

H12: When requested to reveal embarrassing information, people in the
high computer immediacy condition will report greater expectations
of negative consequences of embarrassment than those in the low
computer immediacy condition

H13: When requested to reveal descriptive information, people in the high
computer immediacy condition will report lower expectations of
negative consequences of information abuse than those in the low
computer immediacy condition

45

Computer immediacy is also likely to affect the likelihood of consumers
disclosing their personal and health information. Drawing on Bandura’s (1991, 2001)
SC’T, LaRose & Whitten (2000) argued that a cluster of personalized social cues can play
the role of a social agent in promoting social relationships by evoking feelings of
closeness and warmth. The social environmental influences of personalized social cues of
immediacy can motivate consumers to reveal their personal and health information.

From this perspective, we expect that computer immediacy acts as a social incentive to
motivate consumers’ self-disclosures. Based on this reasoning, it is logical to predict that
computer immediacy will positively affect consumers’ intention to disclose their

information to Web sites.

H14: Consumers who view a Web site with high computer immediacy are
more likely to reveal their information than those who view a Web
site with low computer immediacy.

In the current context, the effect of immediacy on consumers’ motivation to
disclose information should be understood as a social cognitive process that involves the
reciprocal interactions of their information disclosure behavior, their cognitive evaluation
of the benefits and risks of that behavior, and the social environmental influences of
computer immediacy. From this social cognitive perspective, the present study also
attempts to explicate the underlying mechanism of disclosure in human-computer
interactions. Specifically, in accordance with Bandura’s (1991, 2001) SCT, this study
proposes that positive and negative outcome expectations mediate the direct effect of

computer immediacy on consumers’ disclosure intentions. By focusing on the mediating

46

role of one’s expectations about the likely consequences of disclosure behavior online,
this study predicts that the positive effect of computer immediacy on information
disclosure intentions will be mediated by the proposed social cognitive variables of

positive and negative outcome expectations.

H 15: The effect of personalized social cues of computer immediacy on
self-disclosure intention will be mediated by consumers’ positive and
negative outcome expectations.

In an attempt to test the direct and indirect relationships as hypothesized above,
this study proposed a structural model (See Figure 1) that integrates the social cognitive
variables of the underlying mechanisms of self-disclosure in human-computer interaction.
In the proposed model, self-disclosure intentions are directly and indirectly affected by
positive (social support and customized benefits) and negative (embarrassment and
information abuse) outcome expectations, privacy self-efficacy, and computer immediacy.

In summary, the original contribution of this dissertatiOn is that it uses the social
cognitive framework to conceptualize personalized social cues of Web site interfaces as
computer immediacy and explicates the underlying social cognitive mechanisms of the
effect of computer immediacy on consumers’ willingness to disclose their information in
the domain of human-computer interaction. By integrating existing theories of
information disclosure, this dissertation empirically tests the effect with respect to
consumers’ cognitive assessment of the benefits and risks associated with their
information disclosures as well as their judgment of self-efficacy to cope with the

potential risk of online privacy. In this sense, the social cognitive model of information

47

disclosure may extend the CASA model of human-computer interaction (N ass & Moon,
2000; Reeves & Nass, 1996) that suggests that humans tend to respond to the computer
interface without making much social cognitive effort.

This social cognitive model of information disclosure may be also differentiated
from the social exchange perspective of information disclosure in that it assumes
reciprocal interactions of the possible factors of social environmental influence
(personalized social cues of immediacy of Web site interface), human cognition (privacy
self-efficacy and outcome expectations) and human behavior (information disclosure).
Because the social exchange perspective of information disclosure primarily views
individuals’ social actions only as a function of reward, punishment, and the aggregate
outcomes associated with their disclosures, the social exchange perspective may not be
appropriate when applied to a dynamic social environment such as the lntemet. In such
an online environment, consumers not only assess potential positive and negative
consequences, but also actively process the potential impact of social environmental
influences of personalized social cues on their self-regulatory capacities to predict,
control, and manage the outcomes of their information disclosures. Consequently, this
dissertation seeks to untangle the relationship between personalized social cues of
immediacy and human motivation of information disclosure as a social cognitive process
through which consumers consider all personal, behavioral, and environmental influences

in the dynamic environment of the lntemet.

48

dag—~58 A<>OUZ<V 85538 he flax—mam mi? 332 m_ E: 6.82

 

 

20:88me mcoufiooaxm
2:330 2:850
5:32.230 59:. Eoom

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

a: a: 58533
v: n: monik—
/ 25.5
8:52: ll 3 i
958—85 ‘
8:255 2:
E ”, .23
V 52888—
”: E.— 55950
/ 9: NE
\ 2: \
acosfloomxm mcocfioomxm
050330 080050
82?. Stack—£5 waiver—35m
.252 5.383— 38:95

a «Swe—

49

CHAPTER 3

METHOD

The objective of this research is to propose a socio-cognitive model of
information disclosure in human-computer interaction and to investigate the underlying
social cognitive processes of the effect of computer immediacy on consumers’
willingness to disclose their personal and health information. The study design is further
detailed in this chapter via a discussion of the experimental method used. Specifically,
this section focuses on issues related to the research design, participants and their
recruitment, stimulus material development, and measures used in the study. Data were
collected in an online experimental setting that controlled for unwanted intervening
factors and observing the relations among the independent, mediating, and dependent

variables.

Research Design
An online experiment was conducted. The online experiment was chosen because
it was considered to be a natural setting where individuals are commonly requested to
disclose their personal and health information. In addition, participants are hardly
independent in a lab setting that requires them to disclose their personal and health
information with other participants present.
A 2 x 2 mixed factorial design was conducted. Since this research specifically

focused on the effects of the level of personalized social cues of immediacy (low or high)

50

and the type of information (embarrassing or descriptive) that can be requested from Web
sites, it considered these two variables as the treatment factors. In other words, the two
levels of personalized social cues of immediacy was the between-subjects factor and the
two types of information was the within-subjects factor.

In the mixed factorial design, participants were randomly assigned to either the
low immediacy or high immediacy treatment conditions. In each condition, they were
asked to complete the experiment and to fill out a follow—up online survey questionnaire
about their intentions to reveal their embarrassing and descriptive personal and health
information. Here, embarrassing information items were more concerned with the social
disclosure consequence of unwanted impressions to others and included sexually
transmitted disease history, mental illness history, and medical condition, such as obesity,
diabetes, and high blood pressure. On the other hand, descriptive information items were
more concerned with the social disclosure consequence of unwanted information
solicitation and abuse, and included social security number, health insurance policy
number, and doctors’ names and phone numbers. To avoid possible order effects, the

display order of the types of information requested was randomized.

Participants and Procedures
A total of 252 individuals were contacted and recruited by an established online
marketing company (www.zoomerang.com). Specifically, all participants were recruited
opt-ins during Web site registrations and were contacted through online announcements

and email contacts. Then, they were directed to visit a newly created personal health

51

record (PHR) Web site for the experiment (www.c-medfile.com), and were randomly
assigned to one of the two experimental conditions.

Two different versions of the e-medfile Web site were created to reflect the two
experimental conditions, namely low and high levels of personalized social cues of
computer immediacy. These Web sites were created to reflect the two levels of
immediacy manipulation.

With the help of a random assignment program placed in the log-in page, half of
the participants were randomly assigned to the low-immediacy manipulation condition
and the other half to the high-immediacy condition. In order to control for gender effects,
participants were also randomly assigned to each condition in terms of participants’
gender. Specifically, in order to recruit the same number of male and female participants
for each condition, female participants were first recruited until the number of the
participants reached a desirable and equal sample size for each condition. When 63
female participants visited each condition of the Web site and finished filling out the
online survey, the same number of male participants were recruited for the experiment
and asked to visit each condition and fill out the online survey. As a result, 63 male and
63 female participants were assigned to the high immediacy condition of the Web site
and the same number of male and female participants was assigned to the low immediacy
condition. Therefore, each condition contained equal numbers of males and females. In
total, 252 participants participated in the experiment.

Upon visiting the stimulus Web site, participants were first given brief
instructions instructing them to browse the Web site and evaluate it by filling out an

online survey. The instruction page explained that the study concerned the evaluation of a

52

newly launched personal health record Web site (www.c-medfile.com). The instruction
page also informed participants that the Web site would ask them to fill out an online
survey upon completion of browsing the Web site. When the participants were asked to
read a consent form, they indicated their voluntary participation by clicking a “next”
button at the bottom of the page. The moment when the participants clicked the button,
they were automatically and randomly assigned to the log-in page of one of the two
treatment conditions of computer immediacy (high vs. low). When participants were
directed to a registration page of the Web site, each was asked to provide his or her first

name to the Web site as a registration process for study participation (See Figure 2).

Figure 2

A Screenshot of the Log-in Page of the E-Medfile Web Site

@MEDFrIe

n— __—-... —_... -— ---———-v———n————-__L—‘-:‘- . 177:3—17;;ur¢=~: $.47 .4 .. Qaim.. '—_-_r 1..-... rq

s M . 1‘ -.' a "a r‘“ - ' ‘ - Iurl r w - .- 'i:

e-IIEDFile is a newly created Personal Health Record (PHR) website.

We want to know how viewers evaluate this website and its contents.
Please carefully review each page and give us valuable comments.

Please enter your first name and click the sign-in button for your Iogin.

 

First Name: 359" 1"

 

 

‘ ' ""é-HEDFilexom O 2007"

53

When they typed in their first names, they clicked a ‘Sign In’ button and went to
the first page of the Web site. After the participants browsed and experienced the Web
site, they were asked to fill out a post-test online survey questionnaire that collected their
perceptions of the Web site. Once participants finished the questionnaire, they were
directed to a debriefing page thanking them for their participation.

Most participants were white Caucasian (87.3%), followed by Black (4.8%),
Hispanic (4%), Asian (3.6%) and others (.3%). Participants’ ages ranged from 18 to 66
and the median age was 42 years (M = 41.2, SD =13.1). The sample of adult Internet
participants is considered appropriate for this experiment because it represents
individuals with some degree of Internet literacy.

To motivate participation in the study, the online marketing company has a point-
incentive system for participants. Participants can accumulate points and use them for
redemption for such incentives as a can-opener, a MP3 music download, a $10 discount
certificate on Greyhound, and others depending on how much they have accumulated in

redemption points.

Stimulus Material
A Web site (www.c—medfile.com) was created to contain personal health record
(PHR) services and served as the stimulus material for this dissertation. The e-medfile
Web site introduces what the online PHR service is and demonstrates how Internet users
can use the PHR service to store and access their personal health and medical record. The

site contained 12 pages: a log-in page to the site, one greeting page, two basic

54

introductory pages about the site and its PHR services, seven PHR topic pages, and one
interactive review session page.

In order to ensure that experiment participants went through the Web site, each
Web page was equipped with a hyperlink that required them to go to the next page or go
back to the previous page at the bottom of the page. For example, Figure 3 shows the

Web site’s greeting page.

Figure 3

A Screenshot of the Greeting Page of the E-Medfile Web Site

®MEDFIIe w.........

Kiln-1’ 7 . ' 'v-‘I hr!”

—. - —-- “I...“ r!— -m - -- - - u-u-II—sn... . “-1—;— -c- 5.“; . .- pn-q-I- . ----——

1.1553! We welcome your first visit to our e-IlEDFile website!

e-MEDFlle is a web based Personal Health Record (PHR) that stores your personal
health information for you. e-MEDFiIe is designed to promote the vision of personal
access and control over your health information.

e-MEDFiIe makes it easy to retrieve personal medical records, whenever you
- need your medical records quickly

- need your medical information away from home
- want your family and doctors to have access to your medical records 24/ 7

Lissa! Now our e—MEDFile will demonstrate you how this health technology works
for you.

 

 

(PREVIOUS NEXT >

e—HEDFIIéIcom "b 2007

55

Figure 4

A Screenshot of the Navigation Page of the E-Medfile Web Site

Erne'gency Card 8 Info
Personal 81. Insurance lrfo
Eme'genCy Contacts
Other Medical Conditions
Health Care Providers
Pam 84 Symptom Diary

My Shunt

Next, the seven PHR topic pages were presented in a two-column layout. The left

l.a —n.

G
F . .er' - I'm.’
A I | ‘

 

C" MEDFile
Emergency Information Card
http://wwwe-medfile . cum

 

 

 

 

 

 

 

Lon Moore, MD.
Michael Handler, ".0.

Emergency cm
Martha Smith

Carnations
Hydrocephalus

Subscriber's

Photo Here
Name: Jane Smith
Home Phone:
Birth Date: 5/28/1984
Health Plan: Blue Cross
Group Number: x3456
Member ID: 553 33 2233
Usemame:
Password:
“More Providers

SSS-555-5555
444-444-4444

333-333-3333

VP Shunt

 

 

 

Welcome Lissa

I Hat-lune:- -.- .
t . -h—‘a—kwln. ~

Lissa, this session allows you to interact with our Personal Health Record (PHR)
navigation system.

Please tick on each menu to learn more about what our e-HEDFile offers.

EMERGENCY CARD & INFO

If you are in an emergency situation, this
printable Emergency Information Card with
photo alerts emergency personnel to your
medical condition.

Imagine that you have a serious medical

condition but travel a lot. Now you don't

have to worry that an unfamiliar doctor in
an emergency might kill you.

column has a demo navigation system featuring the seven PHR service categories

including emergency card & info, personnel & insurance info, emergency contacts, any
other medical conditions, health care providers, pain & symptom diary, and my shunt.
Specifically, the left—hand navigation system allowed the participants to interact with the

e-medfile PHR services so that users could get to know the services, such as entering and

creating a printable emergency card with a photo, personal and health insurance

 

information, emergency contact information, medical condition history, healthcare
provider information, a pain and symptom diary, and shunt information (See Figure 3).
Appendices A-l through A-7 show the seven topic pages.

After the seven PHR topic pages, the Web site featured a brief interactive review
session for participants to check their general knowledge about the PHR services.
Content for the site was drawn and adapted from existing PHR Web sites, such as

www.followme.com (See Appendix B).

Two Levels of Immediacy Manipulation

Two versions of the Web site were created to reflect the two levels of computer
immediacy (low vs. high). In other words, all features and contents were identical in the
two versions of the e-medfile Web site, except for the immediacy manipulation. In the
high immediacy condition, participants were recognized at the top of the page upon
logging in with their first names. They were also personally greeted with their first name
on the right at the top of the page and in a welcoming salutation message on the main text
page (e.g., Welcome Lissa; Lissa! We welcome your first visit to our e-medfile Web
site!) (See Figure 3). The greetings and appropriate social cues of immediacy tailored for
each participant continued throughout all pages of the Web site. In the low immediacy
condition, participants were not recognized on any pages of the Web site.
Also, in the high immediacy condition, the interactive review session page demonstrated
appropriate social cues of immediacy. Following Bracken and Lombard (2004)’s
manipulation of a “parsing” computer in their study of its effects on children’s learning, ,

participants received praise (e. g., Excellent, Lissa! That is the right answer!) or

57

encouragement (e.g., Try again, Lissa. You can do it!) from the Web site about their
performance in the session (See Figure 5). However, in the low immediacy condition,
participants did not receive praise or encouragement, with the Web site remaining neutral

in addressing participants’ performance in the interactive review session (e.g., That is the

right answer or Try again).

Figure 5

A Screenshot of the Interactive Review Page of the E-Medfile Web Site

m 33..-...a—n... . ..__._._._.-. _ . -.- ....._._ __._

©8959

 

. .7
1" . m . l" -_..

Lissa, now we will ask you a couple of questions about our e-MEDFile service.

1.What does PHR stand for? mm lntemetfimlom &
6) Personal Health Representative l

6 Practical Health Reward
0 Personal Health Resource ' 11 Excellent Lissa! Thatisthe right answer!
Q Personal Health Record 1 ‘2

.—

 

 

 

 

  

2. Which of the following services
631 can access and control over your ea
(5. My family and doctors can have access to my medical records 24/7

61 can retrieve my medical records quickly, even away from home
C) All of the above

<PREVIOUS NEXT)
‘ ."t enema-nun” M -

 

 

 

Types of Information Manipulation

A Web-based questionnaire was also created and linked to the stimulus e-medfile

Web site. Six questionnaire items in the online survey represented two within-subject

58

experimental conditions, in which participants were asked to indicate their intention to
reveal both embarrassing and descriptive personal and health information. Based on a
pretest administered to 40 undergraduate students at a large Midwestern university, three
question items were selected for each information type. The three embarrassing
information items asked about focused on intention to reveal 1) mental illness history, 2)
sexually transmitted diseases history, and 3) medical conditions, such as obesity, diabetes,
and high pressure. The three descriptive information items asked about participants’
intention to reveal their 4) social security number, 5) health insurance policy number, and

6) doctors’ names and phone numbers.

Measurement

Dependent Measure

Information Disclosure Intention. The primary dependent measure of information
disclosure intention was utilized. To measure an individual’s intention to disclose their
personal and health information, participants were asked to rate the likelihood of
revealing each information item in the hypothetical profile to e-medfile.com. The
measurement item was adapted from LaRose and Rifon (2007)’s information disclosure
intention, using a 7-point likelihood scale with 1 representing very unlikely and 7

representing very likely.

Potential Mediating Measures

When discussing the effect of computer immediacy, four possible intervening

variables, namely negative outcome expectations of 1) embarrassment and 2) information

59

abuse, and positive outcome expectations of 3) social trust and 4) customization benefits,
were identified. Questions pertaining to each of these variables were included in order to
empirically verify whether these variables mediated the relationship between computer
immediacy and self-disclosure intention.

Negative outcome expectation of embarrassment. Participants rated the perceived
likelihood of the negative consequence of embarrassment occurring as a result of
providing information requested by the e-medfile Web site. Two items were adapted
from White (2004), consisting of a seven-point Likert type scale ranging from 1 for very
unlikely and 7 for very likely. The items included 1) revealing this information to the Web
site I just worked with will be embarrassing; and 2) revealing this information to the Web
site I just worked with will make others evaluate me negatively. The two items were
combined to form the scale for the negative outcome expectation of embarrassment. This
measure exhibited a high degree of internal consistency for each information type,
ranging from .88 to .97 (Cronbach’s alpha = .91 for mental illness history; .93 for
sexually transmitted diseases history; .88 for medical condition; .96 for social security
number; .95 for health insurance policy number; and .97 for doctor’s name and phone
number).

Negative outcome expectation of information abuse. Participants rated the
perceived likelihood of the negative consequence of information abuse occurring as a
result of providing information requested by the e-medfile Web site. The two items
adapted from White (2004) using a seven-point Likert type scale ranging from 1 for very
unlikely and 7 for very likely. The items included 1) revealing this information to the Web

site I just worked with will lead others to misuse my information without my permission

6O

and 2) revealing this information to the Web site I just worked with will result in an
increase in phone or mail or e-mail solicitations. The two items were combined to form
the scale for the negative outcome expectation of information abuse. This measure
demonstrated a high degree of internal consistency: 77 to .93 (Cronbach’s alpha = .87 for
mental illness history; .93 for sexually transmitted diseases history; 92 for medical
condition; .77 for social security number; .90 for health insurance policy number; and .92
for doctor’s name and phone number).

Positive Outcome Expectation of Customization Benefits. Using a seven-point
scale ranging from 1 for very unlikely and 7 for very likely, participants rated the
perceived likelihood of positive consequences occurring as a result of providing personal
information requested by the e-Medfile Web site. The two items included 1) revealing
this information to the Web site I just worked with will result in better customer service;
2) revealing this information to the Web site I just worked with will result in quick access
to personalized service. The two items were combined to form the scale for the positive
outcome expectations of customization benefits. This measure showed a high degree of
internal consistency for each information type that was requested to be hypothetically
disclosed from the e-medfile Web site, ranging from .77 to .94 (Cronbach’s alpha = .87
for mental illness history; .90 for sexually transmitted diseases history; .94 for medical
condition; .77 for social security number; 89 for health insurance policy number; and .93
for doctor’s name and phone number).

Positive Outcome Expectations of Social Trust. Using the same seven-point Likert
type scale ranging from 1 for very unlikely and 7 for very likely, participants rated the

perceived likelihood of positive social consequences associated with the use of the e-

61

Medfile Web site. Specifically, adapted from Gefen et al.’s (2003) e-trust items,
participants were asked to rate whether the Web site that they had just worked with
would 1) be trustworthy, 2) be willing to support them, 3) give advice in their best
judgment, 4) have good intentions toward them, and 5) be sincere. The five items were
combined to form the scale for the positive outcome expectations of social trust. This

measure demonstrated a high degree of internal consistency (Cronbach’s alpha = .87)

Control Measure

Online Privacy Self-Efrcacy. Using a seven-point scale ranging from 7 for
strongly agree to l for strongly disagree, privacy self— efficacy consisted of three items
that measured a respondent’s perceived ability to protect their own privacy on the
Internet. Adapted from LaRose and Rifon (2007), the three items included 1) I am able to
protect myself against the release of personal information; 2) It is easy to figure out
which sites you can trust on the Internet; 3) I am confident that I can protect my privacy
online. The three items were combined to form the privacy self-efficacy scale. This

measure demonstrated a high degree of internal consistency (Cronbach’s alpha = .91).

Manipulation Check Measures

Computer Immediacy Manipulation. Three items were created to check the
effectiveness of the immediacy level manipulation. Participants were asked to rate
whether the Web site that they had just worked with was 1) warm, 2) friendly, and 3)

intimate. These three questions were measured on the same 7-point scale with 1

62

representing strongly disagree and 7 representing strongly agree (Cronbach’s alpha
= .80).

Information type Manipulation. Two items were used to check the effectiveness
of the manipulation of the information type requested by the e-medifle Web site. Adapted
from White (2004), the type of consumer information should reflect either potentially
embarrassing kinds or privacy-related descriptive personal or health information that can
be requested from a PHR Web site such as e-medfile.com. Six items were selected on the
basis of a pretest administered to 40 undergraduate college students who received extra-
credit for their participation. Based on the pretest, three embarrassing items (e. g.,
sexually transmitted diseases, mental disease, and medical conditions such as obesity,
diabetes, and high blood pressure) and three privacy-related descriptive items (e. g., social
security number, insurance policy number, doctor’s names and phone numbers) were
selected for the main experiment. To examine the information type manipulation, this
study used the measures of negative outcome expectations of embarrassment and
information abuse and asked the participants to rate the perceived negative consequences
differently associated with their disclosure intention for their embarrassing and

descriptive information that can be requested by the PHR Web site.

Preparation for Data Analysis
An exploratory factor analysis was conducted upon the measure of information
disclosure intention regarding the six pieces of information (mental illness history,
sexually transmitted diseases history, medical conditions, social security number, health

insurance policy number, and doctor’s names and phone numbers). The factor analysis

63

yielded two factors accounting for 81% of the variance. The three items (mental illness
history, sexually transmitted diseases history, and medical conditions) loaded under the
first factor labeled “embarrassing information disclosure intention” and the other three
items (social security number, health insurance policy number, and doctor’s names and
phone numbers) loaded under the second factor labeled “descriptive information
disclosure intention.” The three items for each factor were combined to form the scales
for the two types of information disclosure intention. Each scale exhibited a high degree
of internal consistency (Cronbach’s alpha = .90 for embarrassing information disclosure
intention and .86 for descriptive information disclosure intention).

Since the scope of the research hypotheses may be attributable to differences
associated with the types of information that may be requested from the e-medfile Web
site, the proposed research model was tested using two path models of the two types of
information disclosure intention: one for embarrassing information disclosure intention
and another for descriptive information disclosure intention. The exploration of the two
path models allowed examination of the theorized moderating role of information type in

influencing consumers’ intention to reveal their personal and health information.

Data Analysis
First, AN OVA and paired samples t-tests were performed to check manipulations
of the immediacy level and information type, respectively.
Then, Pearson product-moment correlations were calculated to examine bivariate

relationships between the variables of interest using SPSS 16.

64

Next, with AMOS 8.0 and using maximum likelihood estimation, path analyses
were employed to test the hypotheses in the proposed research model. The path analysis
technique was chosen not only to examine the proposed direct and indirect relationships
of the hypotheses, but also to demonstrate whether or not the current data support casual
associations in the proposed model as theorized.

While testing the proposed research model, Analysis of Covariance (ANCOVA)
tests were also performed to examine the main effect of the immediacy level and
information type manipulations on information disclosure intention, with online privacy
self-efficacy as a covariate. By controlling for the effect of online privacy self-efficacy on
information disclosure intention, the ANCOVA test investigated the main effect of the
computer immediacy manipulation on information disclosure intention.

Finally, following Baron and Kenny’s (1986) suggestions for testing mediating
effects, a mediation analysis was employed to test whether the effect of computer
immediacy on consumers’ willingness to reveal their personal and health information was
mediated by the negative outcome expectations of embarrassment and information abuse
and positive outcome expectations of social trust and customized benefits. The next

chapter presents research findings using these analysis methods.

65

CHAPTER 4

RESULTS

Manipulation Checks

Computer Immediacy Manipulation

Computer immediacy was manipulated at two levels, high and low. To assess the
effectiveness of the manipulation, respondents were asked to rate the immediacy level of
the stimulus personal health record Web site (www.c-medfile.com) using the previously
validated three-item scale described in the previous section (Cronbach’s alpha = .80). As
expected, a one-way analysis of variance (AN OVA) revealed a statistically significant
effect for the manipulation (F (1, 250) = 24.797, p < .001). Specifically, participants’
perceived immediacy scores were significantly higher when the high computer
immediacy manipulation (M = 5.20, SD =.93, n: 126) was presented than when the low
immediacy manipulation (M = 4.61, SD =.96, n: 126) was presented on the e-medfile
Web site. These results were interpreted as successful manipulation of computer

immediacy.

Information Type Manipulation

As adapted from White (2004), the type of information was manipulated to reflect
either potentially embarrassing or descriptive personal or health information that can be
requested from the stimulus PHR Web site. Based on the pretest, three embarrassing

items (e. g., sexually transmitted diseases history, mental illness history, and medical

66

conditions, such as obesity, diabetes, and high pressure) and three descriptive items (e.g.,
social security number, insurance policy number, doctor’s names and phone numbers)
were selected for the main experiment. As shown in Table 1, a series of paired samples t-
tests compared the means of two related variables for one single group of participants and
revealed that they would expect that disclosing the three embarrassing information items
would be more likely to make them embarrassed whereas the three descriptive
information items would be more likely to lead to participants’ information being abused
if requested to disclose such information from the e-medfile Web site. These results were

interpreted as successful manipulation of information type.

Table l

Paired Samples T-Tests for Information Type Manipulation

 

 

Information Outcome
requested expectations Mean SD t-value P-value
Sexually Transmitted Embarrassment 4.37 1.53
Diseases History Information abuse 3.70 1.44 t (251) = 8.01 p < .001
Mental Illness History Embarrassment 4.24 1.39
Information abuse 3.74 1.50 t (251) = 6.02 p < .001
Medical Condition Embarrassment 3.76 1.49

Information abuse 3.56 1.42 t (251) = 2.34 p < .05
Social Security Number Embarrassment 2.66 1.46

Information abuse 4.57 1.47 t (251) = -17.29 p < .001

Health Insurance Policy Embarrassment 2.39 1.42
Number Information abuse 4.18 1.63 t (251) = -15.69 p < .001
Doctor’s Name and Embarrassment 2.36 1.43
Phone Number Information abuse 3.34 1.63 t (251) = -10.92 p < .001

 

67

Hypothesis Testing

Tables 2 and 3 show each matrix of Pearson product-moment correlation
coefficients for the variables related to information disclosure intentions of the two types
of information (embarrassing and descriptive information), respectively. Specifically,
Table 2 shows that embarrassing information disclosure intention was positively
associated with computer immediacy (r = .26, p < .001), online privacy self-efficacy (r
= .31, p < .001), social trust outcome expectations (r = .40, p < .001) as well as
customization benefit outcome expectations (r = .33, p < .001), and negatively related to
embarrassment outcome expectations (r = -.38, p < .001) and information abuse outcome
expectations (r = -.55, p < .001). On the other hand, Table 3 shows that descriptive
information disclosure intention was positively related to computer immediacy (r = .29, p
< .001), online privacy self-efficacy(r = .34, p < .001), social trust outcome expectations (r
= .37, p < .001) as well as customization benefit outcome expectations (r = .39, p < .001),
and negatively related to embarrassment outcome expectations (r = -.18, p < .001) and
information abuse outcome expectations (r = -.61 , p < .001).

Path analyses were conducted to examine the social cognitive processes of how
computer immediacy would influence the likelihood that consumers disclose their
personal and health information. Since the research hypotheses might be affected by the
types of information that can be requested from the e-medfile Web site, two path models
were tested: one for disclosure intentions of embarrassing information and another for
disclosure intentions of descriptive information.

The model fit indices indicated an adequate fit with the data for the path models

of disclosure intention for embarrassing information (AGFI = .913, CFI = .984, RMSEA

68

= .085 (90% C.I.:.028 — .146) and for descriptive information (AGFI = .924, CFI = .983,

RMSEA = .075 (90% C.I.:.009 - .137). Although the chi-square test statistics were
2
statistically significant for the model of embarrassing information disclosure (x = 11.225,

df = 4, p = .024) and for the model of descriptive information disclosure (x2 = 9.668, df=
4, p = .046), this is to be expected due to the relatively large sample size (Kline, 1998).
The two final models explained 35 % of the total variance in disclosure intention for
embarrassing information and 45% of the total variance in disclosure intention for
descriptive information. Figure 6 and Figure 7 demonstrate the results of path analyses
for the proposed model, with all paths reporting standardized coefficients. All the

research hypotheses, except for H14, were tested from the final models.

69

 

 

.05. 5. 05 8 EmoEcwfii ._o>o_ mo. 05 B 38.5%.? .802

8.. m3 8.. :8 in: ten. $8.- ......ch an- 88888. 8.850 884. 88:88. .5
..N. 8... 8.. to..- $8.- ...... .- 5.- in... 88888...» 2:850 858885.”. e
8.. .8... 8.. is... :8. ......8. ......mm 8.88%.”. 2:830 52.8 8828.80 .m
8. E... 8.. ......R. :8. to... 8.88%.". 2:850 .8; 88.0. ..V
N... o... 8.. ... .... . m. 88.0.3.8 .83... 2...:0 .m
8. 8. 8.. 3.8. 8.888. 8.880 .N
004 CY? DOA COB—.3:— OHSmOBmmn—utom :OMHMEHOME wEwmwh—anam .—
nm :82 b e m a n N _ 285$

 

9:50—85 :0555502— wfimmahanfim 3 e323— m03a2a> 2: he 35.05000 acts—9.500 acofic—Zéoauem :83?—

N 035,—.

70

.03. 5. 05 ... .§0c_=w_m** ._0>0_ mo. 05 .m EmoEcwmma. .082

 

 

8.. 85 8.. 3. ... 3.8.- 3.8.- 3.8.- 3.8.- 3. 6.- 5.8.5.8me 05850 9.8... 8.55.5.5. .5
8.. 5mm 8.. 8.- ...: .- 8.- 8.- 3... . .- m5.0.5.8....m 55850 5055885.”. .8
.N. R... 8.. 3.8. 3.8. 3.8. 3.8. 55.5586 8.850 .880 8.8.5950 .8
8. E... 8.. 3.8. 3.8. 3.8. m8.53.888. 05850 .5... .88 ...
N... o... 8.. .... 3.8. 88:50.8 8..-5.... 8.50 .m
8. 8. 8.. 3.8. 85.855. 8.2.500 .N
mm.— ©m.m 09— comm—:02: Dun—mo—omR—utom ComuaEuOmcu o>€mtomoa A
8 :82 h a m a m N _ 0.8.5.»

 

0.58.85 :060832— 0332809 3 603.0% 830......» 05 no 55.05000 5:20:00 3050230....05 58.30.—

m 030,—.

71

6.0.. I wmo. H...0 0800. mm... .1. <m.m.>.~. .03. n EU 5.0. u EU... .08. u a ... "b. .nmm... .1. ...
.O.VQ**.WO.VQ*

050m. .gom.0m.m 00.00.00. All:
00.00 .002...0w.m-000 00.00.00. A ..... ”0.02

 

 

      

      

0000080qu
0800.00
02.00.000.000

000008093
00.00.00
.000... .0.00m

      

 

 

 

 

 

 

 

 

 

:3”.
08.00.08
.8035
00:00
00000.0. 00000.85
00:00:80. ‘
w0.80t008m -

........... .26..
y. 85.855.

X 00.09000
0 30
:3 . \
gnu.- \ ~ \

 

 

 

 

 

00000.00gxm. 00000809...
0000.00 0800.00
0000... 00:00:80. 3.80.0000...

 

 

 

 

0.500.005 00:00:00.0. w0.0000000=_m ..0 .0002 000000.... 0... 00.. 0.0.0.005. 0.0a. 000.000.008m ..0 3.003.

0280

72

.5. I 08. H...0 Q08. 08. u $020 .000. n .00 .000. n E0< .03. u 0 .0 "00 .0000 u m
_O.VQ** .mo.VQ*

00.00 2000.030 00.00.00. All
00.00 .000¢.0w.0-000 00.00.00. A ..... ”0.02

 

 

0000300000 0000800000".
0800.00 00.00.00
00.00.80.000 .00; .0.00m

 

 

 

 

 

 

 

 

 

 

 

 

580000.00
000300
00:00
00000.0. 0.000.005
0000:0000.
300.0000
.004
00.00000:
00.00800

 

 

 

 

 

 

000..0.000xm. 000..0.000.~m
0800.00 00.50.00
0000< 000090.000. 3.0005005

 

 

 

 

 

 

0.500.005 003050000— 0>..0...0000 ..0 .0000). 00000000 0... 00.. 0.0%.00< ...00 000.000.008m ..0 0.300%

020$

73

H1 predicted that participants’ expectations of negative consequences of
embarrassment would be negatively related to their intention to disclose their personal
and health information to the e-medfile Web site. The hypothesized negative effect of
negative outcome expectations of embarrassment on disclosure intention was significant
at the .001 significance level when participants were requested to reveal embarrassing
information about themselves (B = -.23, p < .001). However, the effect was not significant
when participants were requested to reveal descriptive information about themselves ([3
= .07, p > .10). Thus, H1 was partially supported.

H2 predicted that participants’ expectations of negative consequence of
information abuse would be negatively associated with their intention to reveal their
information to the e-medfile Web site. As expected, negative outcome expectations of
information abuse were negatively associated with disclosure intention for both
embarrassing (B = -.23, p < .001) and descriptive information ([3 = -.52, p < .001). Thus,
H2 was supported.

H3 predicted that participants’ expectations of positive consequences of social
trust would be positively related to their intention to disclose their information to the e-
medfile Web site. As predicted, participants’ positive outcome expectations of social trust
with the e-medfile Web site promoted their disclosure intention when they were
requested to reveal embarrassing information ([3 = .16, p < .01). However, the positive
effect of positive social outcome expectations on disclosure intention was not significant
when participants were requested to reveal their descriptive information (B = .07, p> .10).

The results suggest a noticeable difference in the relationship between positive outcome

74

expectations of social trust and disclosure intention, depending on information type. Thus,
H3 was partially supported.

H4 predicted that participants’ expectations of positive consequence of
customization benefits would be positively associated with their intention to reveal their
information to the e-medfile Web site. Unexpectedly, depending on information type, the
results revealed a noticeable difference in the relationships between positive outcome
expectations of customization benefits and disclosure intention. Specifically, the
hypothesized positive effect of positive outcome expectations of customization benefits
on disclosure intention was not significant for the embarrassing information (B = .10, p
>.05), whereas the corresponding effect was significant for the descriptive information ([3
= .21, p < .001). Thus, H4 was partially supported.

H5 predicted that online privacy self-efficacy would promote participants’
intention to reveal their personal and health information. As expected, online privacy
self-efficacy was positively related to disclosure intention for both the embarrassing (B
= .12, p < .05) and the descriptive (B = .14, p < .05) personal and health information. The
results indicated that those with high privacy self-efficacy had greater intentions to reveal
their information to the e-medfile Web site than those with low privacy self-efficacy,
regardless of the information type. Thus, H5 was supported.

H6 predicted that participants’ high online privacy self-efficacy would lower their
expectations of negative consequences of embarrassment. However, depending on the
information type, there was an evident difference in the relationship between privacy
self-efficacy and negative outcome expectations of embarrassment. Specifically, the

hypothesized negative effect of privacy self-efficacy on negative outcome expectations of

75

embarrassment was significant only when participants were requested to reveal their
embarrassing personal and health information (B = -.16, p < .01), whereas the
corresponding effect was not significant when participants were requested to disclose
their descriptive information (B = -.O7, p > .10). Thus, H5 was partially supported.

H7 predicted that if participants have high online privacy self-efficacy, they will
have lower expectations of negative consequences of information abuse by the e-medfile
Web site. As hypothesized, the results showed that privacy self-efficacy had a negative
effect on the negative outcome expectations of information abuse for both the
embarrassing information (B = -.24 p < .001) and the descriptive information (B = -.20, p
< .001). Thus, H7 was supported.

H8 predicted that participants with high online privacy self-efficacy would report
greater positive outcome expectations of social trust than those with low privacy self-
efficacy. As expected, the positive effect of privacy self—efficacy on positive outcome
expectations of social trust was significant for both the embarrassing information and the
descriptive information (B = .34, p < .001). Thus, H8 was supported.

H9 predicted that if participants have high online privacy self-efficacy, they will
expect that they would receive customized benefits from the e-medfile Web site. As
predicted, privacy self-efficacy was positively associated with positive outcome
expectations for both the embarrassing information (B = .21, p < .001) and the descriptive
information (B = .26, p < .001). H9 was thus supported regardless the information type.

H10 predicted that participants who viewed the version of the e-medfile Web site
with high levels of personalized social cues of immediacy would report greater

expectations of social trust than those who viewed the version with low levels of

76

immediacy. As expected, the hypothesized positive effect of privacy self-efficacy on
expectations of social trust was significant for both the embarrassing information (B = .25,
p < .001) and the descriptive information (B = .25, p < .001). Thus, H10 was supported.

H11 predicted that participants who browsed the version of the e-medfile Web
site that had high levels of personalized social cues of immediacy would report greater
expectations that they will receive customized benefits from the Web site than those who
browsed the version with low levels of immediacy. As predicted, computer immediacy
was positively associated with expectations of customization benefits when participants
were requested to reveal their embarrassing information (B = .21, p < .001) and
descriptive information (B = .17, p < .001). H11 was thus supported regardless of the
information type.

H12 predicted that participants who viewed the version of the e-medfile Web site
featuring high levels of personalized social cues of immediacy would report greater
expectations of negative consequences of embarrassment than those who viewed the
version with low levels of immediacy. However, the hypothesized positive effect of
computer immediacy on the expectations of embarrassment was not significant for either
the embarrassing information (B = .04, p > .10) or the descriptive information (B = -.O4, p
> .10). Thus, H12 was not supported.

H13 predicted that participants who browsed the version of the e-medfile Web
site having high levels of personalized social cues of immediacy would report lower
expectations that their information will be abused than those who viewed the version
with low levels of immediacy. As predicted, computer immediacy was positively

associated with expectations of customization benefits when participants were requested

77

to reveal their embarrassing information (B = -.28, p < .001) and descriptive information
(B = -.24, p < .001). H13 was thus supported.

H 14 predicted that the participants who viewed the version of the e-medfile Web
site having high computer immediacy would be more willing to reveal their personal and
health information than those who viewed the version with low immediacy.

A mixed design 2 x 2 analysis of covariance (ANCOVA) test was performed to
examine the effects of the immediacy level and information type manipulations on the
two types of disclosure intention while controlling for the effect of privacy self-efficacy.
The results of the ANCOVA test showed significant main effects of the immediacy level
(F (1, 249) = 22.76, p < .001 and 772 = .08) and information type (F (l, 249) = 6.040, p
< .05 and If = .02) on information disclosure intention. In addition, no interaction effect
of the immediacy level and information type on disclosure intentions was found. The
results indicated that participants who browsed the Web site in the high immediacy
condition were more likely to disclose their embarrassing and descriptive information
than those who visited the Web site in the low immediacy condition. The results also
indicated that participants who browsed both high and low immediacy PHR Web sites
were more likely to disclose their embarrassing information than their descriptive
information. Another independent effect of the subjects’ online privacy self-efficacy on
the disclosure intention was also found (F (1, 249) = 34.66, p < .001 and 772 = .12) in the
analysis. Table 4 summarizes the appropriate means of the dependent measures across

experimental conditions.

78

Table 4

Means and Standard Deviations for Information Disclosure Intentions by
Computer Immediacy and Information Type Conditions

 

“Immediacy Level”

 

Low Immediacy High Immediacy

“Information Type”
Embarrassing Information

 

Sexually Transmitted Diseases 3.84 (1.95) 4.55 (1.67)
Mental Disease 3.67 (1.93) 4.63 (1.61)
Medical Conditions 4.46 (1.71) 5.28 (1.35)
Mean 3.99 (1.69) 4.82 (1.39)
Descriptive Informatifl
Social Security Number 2.59 (1.62) 3.51 (1.80)
Insurance Policy Number 3.53 (1.85) 4.40 (1.70)
Doctor’s Contact Numbers 4.08 (1.90) 5.06 (1.57)
Mean 3.40 (1.59) 4.32 (1.44)

 

Note: Means are computed based on the information disclosure intentions being measured on a 7
point scale with 1 representing very unlikely and 7 representing very likely. Standard deviations
are in parentheses.

H15 predicted that the direct effect of computer immediacy on self—disclosure
intentions would be mediated by the social cognitive mechanisms of self-disclosure
involving positive and negative outcome expectations. Following Baron and Kenny’s
(1986) suggestions for testing mediating effects, it was expected that the mediation
analysis would show the hypothesized direct effect of computer immediacy on
consumers’ willingness to reveal their personal and health information should become
non-significant after controlling for the potential mediators of negative outcome
expectations of embarrassment and information abuse and positive outcome expectations
of social trust and customized benefits.

First, the direct paths from computer immediacy to embarrassing and descriptive

information disclosure intentions were examined in the absence of the other mediators.

79

The path coefficients indicated that both direct paths to embarrassing information
disclosure intentions ([3 = .27, p < .001) and descriptive information disclosure intentions
(B = .29, p < .001) were significant.

However, when combined with the other mediators in the original path models,
the significant direct effects of computer immediacy on disclosure intentions were no
longer significant when participants were requested to reveal both embarrassing ([3 = .10,
p > .05) and descriptive personal and health information ([3 = .08, p > .05). Thus, the
relationship between computer immediacy and the dependent variable of information
disclosure significantly attenuated when the other mediators were incorporated in the
model. The results established support for mediation and H15 was supported.

In addition, the path analyses found that the personalized social cues of the Web
site affected participants’ willingness to reveal their embarrassing information only
through their expectations of the negative consequences of information abuse and of the
positive consequences of social trust, whereas the personalized social cues of the Web
site influenced participants’ intention to reveal their descriptive information only through
expectations of negative consequences of information abuse and of positive consequences
of customization benefits. Table 5 shows the summary of the hypothesis testing.

In summary, path analyses showed that privacy self-efficacy, an important
determinant affecting disclosure intention, not only directly affected the decision to
engage in some form of information disclosure but also indirectly affected it by
instigating the assessments of benefits (e.g., social trust and customized services) and
risks (e.g., embarrassment and information abuse) associated with information disclosure

behaviors. Path analyses also demonstrated that consumers’ exposure to personalized

80

social cues of computer immediacy enhanced their expectations of positive consequences
of social trust and customization benefits and reduced their expectations of negative
consequences of information abuse at the same time. Further, such outcome expectations
directly motivated the likelihood that consumers will reveal their personal and health
information. However, the influence of computer immediacy on the likelihood of
disclosure was not evident in the final path model, suggesting that the direct effect of
computer immediacy on disclosure intention should be mediated by the social cognitive
variables of positive and negative outcome expectations while controlling for the effects

of privacy self-efficacy.

81

Table 5

Results of Hypothesis Testing

 

 

Hypothesis Hypothesized Direction Results

H1 Negative effect of embarrassment on disclosure intention Partially supported

H2 Negative effect of information abuse on disclosure Supported
intention

H3 Positive effect of social trust on disclosure intention Partially supported

H4 Positive effect of customization benefits on disclosure Partially supported
intention

H5 Positive effect of privacy self-efficacy on disclosure Supported
intention

H6 Negative effect of privacy self-efficacy on Partially supported
embarrassment

H7 Negative effect of privacy self-efficacy on information Supported
abuse

H8 Positive effect of privacy self-efficacy on social trust Supported

H9 Positive effect of privacy self-efficacy on customization Supported
benefits

H10 Positive effect of computer immediacy on social trust Supported

H1 1 Positive effect of computer immediacy on customization Supported
benefits

H12 Positive effect of computer immediacy on Not supported
embarrassment

H13 Negative effect of computer immediacy on information Supported
abuse

H14 Main effect of computer immediacy on disclosure Supported
intention

H15 Indirect effect of computer immediacy on disclosure Supported

intention

82

CHAPTER 5

DISCUSSION

This dissertation proposed and tested a social cognitive model of information
disclosure in human-computer interaction. Guided by Bandura’s (1991, 2001) Social
Cognitive Theory (SCT), the present study attempted to develop a theoretical framework
for understanding and predicting the social cognitive mechanisms of information
disclosure in human-computer interaction within the context of personal health record
(PHR) Web site services.

Specifically, first, this dissertation sought to develop the concept of computer
immediacy (LaRose & Whitten, 2000) and establish its existence in the context of
information disclosure in human-computer interaction. Secondly, it investigated the
social cognitive processes of how the personalized social cues of immediacy on a Web
site interface would affect consumers’ willingness to disclose their personal and health
information through positive (social trust and customization benefits) and negative
(embarrassment and information abuse) outcome expectations associated with their
interaction with and their information disclosures to the Web site. Finally, this
dissertation explored how information type (embarrassing vs. descriptive information)
requested from the Web site would moderate the social cognitive processes of the effect
of computer immediacy on consumers’ likelihood of revealing their personal and health

information.

83

This dissertation started by conceptually defining computer immediacy as
personalized social cues of a Web site interface that enhance feelings of psychological
closeness and warmness between the Web site and its consumers. From the social
cognitive perspective, computer immediacy was also theoretically distinguished from
other similar media theories including social information processing (Walther, 1992),
social presence (Short, Williams, & Christie, 1976), and CASA (Computers-Are-Social-
Actors) model of human-computer interaction (Nass & Moon, 2000; Reeves & Nass,
1996), in that it is was regarded as a social environmental influence of the Web site
interface which reciprocally interacts with the other two cognitive determinants of human
cognition (outcome expectations and privacy self-efficacy) and human behavior
(disclosure). This proposition highlights humans’ self-regulatory mechanisms where
cognition plays a central role in moderating behavior when people interact with the

personalized social cues of computer immediacy.

Social Cognitive Process of the Effect of Computer Immediacy on Information
Disclosure Intention
Overall, the results of this study provide empirical evidence for the existence of

computer immediacy in human-computer interaction. First, the statistical significance on
the manipulation check of computer immediacy confirmed that Web users were able to
discriminate between the different levels of personalized social cues that the
manipulation of computer immediacy elicited. The results reported here provide some of
the first empirical evidence for the existence of computer immediacy in the domain of

human-computer interaction.

84

Further, the results of the present study confirmed that the application of
Bandura’s (1991, 2001) SCT to the domain of human-computer interaction can be a
promising theoretical framework for understanding how the social environmental
influence of personalized social cues of immediacy can influence consumers’ likelihood
of revealing their personal and health information to Web sites. That is, consistent with
SCT, exposure to personalized social cues of immediacy translated positively into one’s
cognition toward the stimulus e-medfile Web site by promoting expectations of the
positive consequences of social trust and customization benefits and reducing
expectations of the negative consequences of information abuse from the Web site at the
same time. Further, such outcome expectations directly motivated consumers’
willingness to reveal their information, which confirmed the outcome expectations of
human cognition as one of the main cognitive components affecting human-computer
interaction. Consequently, personalized social cues of immediacy triggered the social
cognitive self—regulatory mechanisms of information disclosure in human-computer
interaction where consumers can cognitively process the social environmental influence
of immediacy cues and evaluate the potential positive and negatively consequences
associated with information disclosure in the lntemet.

Consistent with prior research on online privacy protection behavior (LaRose &
Rifon, 2007; Rifon et al., 2005; Lee et al., in press), this study upheld a basic tenet of the
social cognitive model of online privacy protection behavior: that privacy self-efficacy is
another important determinant affecting disclosure intention directly and indirectly. That
is, consumers’ belief in their self-efficacy to protect their online privacy not only directly

influenced their decision to engage in some forms of disclosure online but also indirectly

85

affected them by instigating the assessments of benefits (social trust and customized
services) and risks (embarrassment and information abuse) associated with their
disclosures. By demonstrating the established causal links of the direct and indirect
relationships between privacy self—efficacy, outcome expectations, and information
disclosure intention, the current study highlighted the importance of privacy self-efficacy
as another important cognitive component in understanding self regulatory mechanisms
in online privacy protection behaviors.

The present study also showed that, regardless of whether the information
requested is embarrassing or descriptive in nature, exposure to a high level of
personalized social cues of immediacy significantly and positively motivated information
disclosure intention toward the stimulus e-medfile Web site, even after controlling for the
effects of privacy self-efficacy. However, the direct influence of computer immediacy on
the likelihood of information disclosure was not evident in the final path model which
involved the potential mediators of positive and negative outcome expectations. The null
relationship between computer immediacy and information disclosure intention in the
final model suggests that the effect of computer immediacy on disclosure intention is
mediated by the social cognitive variables of positive and negative outcome expectations.
As such, the mediating role of outcome expectations on the influences of consumers’
exposure to personalized social cues on their willingness to reveal personal and health
information was apparent.

The results extend the CASA model of human-computer interaction (Nass &
Moon, 2000; Reeves & Nass, 1996) suggesting that humans socialize with computers

without much cognitive processing effort so that social behavior is automatically evoked

86

by the computer itself. The present results of the study suggest that consumers can also be
mindful of the potential positive and negative consequences of providing their
information to specific Web sites on the lntemet. In other words, consumers are not
always mechanically responsive to a Web site’s social cues when enacting their
information disclosures. Rather, they can often actively process the social environmental
cues of the Web site interface and cognitively assess the consequences of their
information disclosures and make a self-reactive decision particularly when they have to
deal with the potential risks of privacy invasions and benefits of social trust and
customized service offerings in the lntemet. Thus, consumers' likelihood of disclosures in
human-computer interaction on the Internet may have to be understood as a human
behavior that requires social cognitive effort beyond the automatic social response to the
computer itself.

The social cognitive explanation of information disclosure in human-computer
interaction on the Internet should be differentiated from Moon’s (2000) CASA
perspective of the reciprocal process of disclosure exchange between a computer and its
user. She (2000) confines the domain of disclosure to one particular social rule - the
reciprocity effect that disclosure automatically generates reciprocal disclosure based on
the reactive process of the attraction-disclosure relationship. However, the social
cognitive model of information disclosure highlights the possible interactions of social
environmental, cognitive, and behavioral factors that may reciprocally determine
consumers’ information disclosures in a more dynamic human-computer interaction on
the lntemet. As such, Moon’s (2000) model of the reciprocity effect in disclosure

exchange may not be applicable to the domain of human computer interaction on the

87

Internet in situations where consumers actively assess other important determinants of
information disclosure, including privacy invasions (e.g., Dinev & Hart, 2006a; Hann, et
al., 2007; Rosenfeld et al., 1999), customized service benefits (e.g., Chellappa & Sin,
2005; Milne, et al., 2008), social trust (e.g., Metzger, 2004; Norberg, etal., 2007;
Wheeless, 1978), the rewards and risks offered in exchange for disclosures (e.g., Hui et
al., 2007; White, 2004), and privacy self-efficacy (e.g., LaRose & Rifon, 2007; Lee et al.,
in press) .

It is important to note that the current social cognitive model of human-computer
interaction agrees with the CASA perspective in terms regarding the computer as a social
actor. However, the social cognitive model does not conceptualize a person as one who

always engages in social behaviors as dictated by the computer’s social cues.

Moderating Role of Information Type in the Social Cognitive Processes of the Effect
of Computer Immediacy on Self-Disclosure Intention

In exploring the theoretical linkage of the social cognitive variables of computer
immediacy, online privacy self-efficacy, outcome expectations, and information
disclosure intentions in the path models, this study found that, depending on the
information type, consumers’ expectations of positive and negative outcome expectations
differently mediated the effect of computer immediacy on information disclosure
intention.

Specifically, the path analyses showed that, when consumers were requested to
disclose their descriptive information, privacy self-efficacy, positive outcome

expectations of customized service benefits, and negative outcome expectations of

88

information abuse were found to be significant predictors of the intention to disclose
descriptive information. However, positive outcome expectations of social trust and
negative outcome expectations of embarrassment had no significant effect on the
intention to disclose embarrassing information. These findings suggest that consumers
are not likely to reveal embarrassing information about themselves unless they feel
confident in their privacy protection behavior, they expect that they will receive social
support and trust from the Web site, and they believe that they will not be embarrassed
and their information will be not be abused after revealing the information. Also, the
findings suggest that consumes are not likely to reveal their descriptive information
unless they feel confident in their privacy protection behavior, they expect that they will
receive greater benefits of customization services from the Web site, and they believe that
their information will be not be abused after revealing the information.

Similarly, the results of the path analyses also confirmed that, depending on the
information type, exposure to the personalized social cues of immediacy differently
affected consumers’ intention to reveal their personal and health information through
different types of mediators.

Specifically, when consumers were requested to reveal their embarrassing
personal and health information, the effect of computer immediacy on disclosure
intention was evident by the increased level of expectations of social trust. However,
when requested to reveal their descriptive information, the effect was evident by the
increased level of expectations of customized service benefits instead. The findings may
suggest that consumers’ exposure to a high level of personalized social cues may activate

corresponding expectations of positive outcome expectations (social trust or

89

customization benefits) specifically linked to the two different types of information when
requested to disclose. For example, consumers are more likely to reveal their
embarrassing information when they expect that they can be socially supported from the
Web site than when they expect they can receive customized service benefits from the
Web site’s customization services. In contrast, consumers are more likely to reveal their
descriptive information when they expect that they can receive customized service
benefits from the Web site than when they expect that they can receive social trust from
the Web site. The results suggest that consumers’ expectations of social trust may be
more important in predicting their willingness to reveal their embarrassing information
whereas their expectations of customized service benefits may be more critical in
predicting their willingness to reveal their descriptive information. In this sense, the
different types of information may be significantly related to their corresponding positive
outcome expectations

In social cognitive terms, such discrepancies in the two path models may suggest
that the information type may have differently instigated the self-regulatory mechanisms
of human motivation for information disclosure (Bandura, 1991). Specifically, the two
different types of information may have motivated consumers to selectively attend to the
different types of risks and benefits associated with their own disclosure intentions (self-
observation), judge their disclosure intentions in relation to social norms (normative
judgmental process) related to the different types of information, and adjust their
disclosure intentions to social environmental cues of Web sites by regulating their

heightened psychological states depending on the information type, (self-reactive

90

influences). In this sense, the two different types of information may have activated
different social cognitive processes of information disclosure

However, when the negative outcome expectations of embarrassment and
information abuse were taken into consideration, only the expectations of information
abuse mediated the effect of computer immediacy on disclosure intention, regardless of
the information type. Specifically, as expected, when requested to reveal descriptive
information, personalized social cues of immediacy were significantly related to the
expectations of information abuse. However, when requested to reveal embarrassing
information, such immediacy cues were not significantly related to the expectations of
embarrassment. Instead, they were significantly associated with the expectations of
information abuse. The findings suggest that, regardless of the information type,
consumers’ expectations of information abuse were more important than their
expectations of embarrassment in predicting their willingness to reveal information.

One possible explanation of the null relationship between computer immediacy
and negative outcome expectations of embarrassment is that risks of information abuse
were so consequential to participants’ information disclosures that the risks may have
overridden the risk of embarrassment even when requested to reveal embarrassing
information. ,

Another possibility is that revealing embarrassing personal and health information
was not embarrassing enough to instigate the potential negative consequences of
embarrassment. The participants may have perceived that they could relinquish their
embarrassing information to the stimulus e-medfile Web site because the health-oriented

nature of the stimulus Web site may have reduced the participants’ perceived potential

91

risks of embarrassment in advance so that the participants may have not formulated the
negative consequences of social embarrassment as originally intended by the
personalized social cues of the Web site interface. This explanation may be consistent
with the results indicating that participants were more likely to disclose embarrassing
information (M = 4.40, SD =1.60) than their descriptive information (M = 3.86, SD

=1.58) to the e-medfile Web site (t (251) = 6.031, p < .001).

Implications
Theoretical Implications

Thus, the results of the study not only provide an overall picture of the social
cognitive processes involved in the effect of personalizes social cues of immediacy on
consumers’ willingness to reveal their persona and health information in the domain of
human-computer interaction, but also shed light on the social psychological antecedents
for disclosure behaviors in this area.

Specifically, the results suggest that personalized social cues of computer
immediacy is an important concept in understanding information disclosure in human-
computer interaction on the Internet, where consumers can actively process the social
environmental cues of immediacy and cognitively assess the potential to impact their
potential positive and negative consequences associated with their information
disclosures. This social cognitive perspective extend the CASA perspective (Nass &
Moon, 2000; Reeves & Nass, 1996) suggesting that humans respond to the computer
interface without exerting much social cognitive processing effort. In this sense,

computer immediacy should be recognized as a psychologically significant variable that

92

influences consumers’ willingness to reveal their information. Given that outcome
expectations mediate the relationship between computer immediacy and information
disclosure, the appeal of immediacy cues may go beyond simply attracting consumers’
initial attention and response to a Web site. The promise of the evolving relationship
between the Web site and its consumers motivates consumers’ interaction with the site
and helps develop and build social their relationships, especially over a period of time.

These considerations prompt a more dynamic conceptualization of personalized
social cues of computer immediacy, whereby the social cues not only provide
personalized information but also anticipate users’ interests and needs, and proactively
service such needs. In this sense, this construct of computer immediacy should be
regarded as an important aspect of users’ interaction with Web sites and may contribute
to a better understanding of how and why consumers are willing to reveal their

information.

Practical Implications

While the question is of obvious theoretical importance for understanding the
social psychological mechanisms of personalized social cues and the related decision
making processes of information disclosure, the findings of the study are also of practical
significance to those who are interested in promoting consumers’ information disclosures
on PHR Web sites as well as for those who are charged with online personal and health
data management systems.

Clearly, the results suggest that those who are in charge of consumers’ data

management on PHR Web sites should concentrate their efforts on utilizing personalized

93

social cues of the interface to promote consumers’ disclosure behavior regarding their
personal and health information. Based on the current results, technical solutions offered
by personalized social cues of computer immediacy may help encourage disclosures by
enhancing positive outcome expectations of social trust and customization benefits and
reducing negative outcome expectations of information abuse. In other words,
personalized social cues of computer immediacy may function to minimize the
perception of the risks of privacy invasions while enhancing the perceived benefits of
social trust and customized service benefits.

However, from the consumers’ perspective, there are also potential downsides of
privacy violations if Web sites abuse the personalized social cues of immediacy and
misuse consumers’ information. As Web sites capitalize on the use of personalized social
cues for collecting consumers’ information, consumers should be increasingly wary of
giving out their information online. Given that the issue of privacy may be even more
troubling, Web sites may have to be required to ensure consumers’ privacy by
participating in privacy seal programs such as TRUSTe and developing efficient privacy
warning statements that clearly describe their privacy policies (LaRose & Rifon, 2007).

In addition, the present findings suggest that PHR Web sites’ interventions aimed
at increasing consumers’ privacy self-efficacy beliefs are likely to yield more reliable
effects in eliciting their personal and health information. In other words, perceptions of
privacy protection efficacy could become a tool for encouraging consumers’ disclosure
behavior.

Three basic approaches are required to increase self-efficacy (Bandura, 1997):

verbal persuasion, anxiety reduction, and progressive mastery. In the current context,

94

verbal persuasion might be implemented through conventional information campaign
strategies targeted to PHR Web site users, making them aware that their privacy
protection is guaranteed by the Web site and instructing them in the ways that the Web
site implements protection measures.

Another approach of enhancing self-efficacy is to employ anxiety reduction
strategies that might focus on the relief of stress immediately prior to revealing personal
and health information. For example, to novice consumers, disclosing sensitive and
embarrassing personal and health information is stress-inducing. Perhaps their anxiety
may be reduced by tailored messages indicating how their information has been easily
secured by consumers’ minimal efforts in utilizing an interactive data management
system on the PHR Web site. Further, PHR Web sites can offer the possibility of tailoring
messages to consumers who tend to perceive high risk of information abuse associated
with their personal and health information disclosures. Based on the current results,
messages aimed at those who have previously experienced privacy violations can present
a potentially high payoff opportunity of receiving customization services from the Web
site, thus reducing such anxiety.

Finally, as the term implies, progressive mastery can build self-efficacy through
accomplishment of a series of successively difficult tasks related to personal health
record data management on PHR Web sites. For example, when novice consumers enter,
remove, and update their information by themselves on the Web site, they have to learn
all the complicated tasks of PHR data management. Thus, progressive mastery is a role
for help desk personnel who could be instructed to walk users through the data

management system rather than having users do it all on their own or doing it all for them.

95

An interactive curriculum session that provides contextual help for concerned and
frustrated PHR Web site users would represent such an approach. Such progressive
mastery could be maximized by the use of an avatar, a digitally embodied visual and
aural mode of representation (Bailenson & Blascovich, 2004), that can interactively guide

consumers through the PHW Web site in a highly personalized manner.

Limitations and Future Research

This dissertation created a personal health record (PHR) Web site and examined
how personalized social cues of a Web site interface would influence consumers’
willingness to reveal their personal and health information. However, the present research
is subject to several limitations that require further investigation.

First, this study recognizes that the utilization of Internet users as participants
might be a concern that limits the generalizability of the results. Although an established
online marketing company used their panels to recruit participants for the current study
through their online announcements and contacts, this sampling method may have posed
problems of imprecise and unreliable response rates. It is also very likely that the
recruited participants might have had higher privacy self-efficacy, higher disclosure
intention, and lower negative outcome expectations of embarrassment and information
abuse than ordinary Intemet users because the participants are accustomed to revealing
their information and opinions to a variety of online surveys and experiments whenever
requested. They may have already surrendered their privacy by agreeing to participate in
the panel for this study. However, when considering that PHR Web site services are not

familiar to other types of participants (e.g., college students), the sample of the current

96

research participants whose ages range from 18 to 66, with a median age of 42 years, can
be considered as an appropriate place to start. These participants can represent adult
Internet users who could be early adopters of PHR Web site services with some degree of
Internet literacy.

Second, the specific PHR Web site employed in this study may be not the most
ideal type for investigating the moderating roles of information type in the social
cognitive processes of the effect of personalized social cues on disclosure intention. Since
PHR Web sites are generally designed to promote consumers’ willingness to reveal
personal and health information that is inherently sensitive, judgmental, and embarrassing,
the use of the stimulus Web site might have confounded the potential interaction effects
of computer immediacy and information type on consumers’ disclosure intentions and
their expectations of potential consequences associated with their information disclosures.
Under such circumstances, the nature of the health-oriented stimulus PHR Web site
might have actually led the participants to a heightened state of involvement with the
benefits of providing their health information. By recognizing the benefits associated
with the PHR Web site, participants’ actual negative outcome expectations of
embarrassment associated their information disclosures may have been mitigated and
their subsequent disclosure intentions associated with their embarrassing information
may have increased. In this respect, future studies may attempt to employ other online
Web sites (e.g., portals, online stores) as a potential stimulus, which could control for
participants’ levels of involvement with the Web site which may influence outcome

expectations and disclosure intention associated with the information type in advance.

97

Another limitation of this study is that the personalized social cues of computer
immediacy were manipulated using only text. This constraint with respect to the textual
mode of personalized social cues may have mitigated the effects of personalized social
cues of immediacy on the negative outcome expectations of embarrassment and
information abuse. The focus of this study was to investigate how the initial difference
induced by the levels of immediacy is carried over to consumers’ disclosure intentions.
Thus, only the textual mode of computer immediacy on the stimulus Web site was
controlled for across the conditions. However, considering that visual and verbal modes
of personalized social cues (e.g., an avatar with voice) can play a critical role in inducing
computer immediacy in the current Web environment, a systematic empirical
investigation of such extensions to the other sensory modes of computer immediacy may
provide additional insights into the effect of computer immediacy on consumers’
willingness to reveal their personal and health information in future research.

Finally, this study used information disclosure intentions instead of actual
disclosure behavior as a dependent variable based on the assumption that behaviors are
strongly predicted by intentions (Ajzen & Fishbein, 1980). Since this study instructed the
participants to evaluate the newly launched PHR Web site, it was somewhat unrealistic to
ask them to reveal their personal and health information. However, it should also be
noted that intentions do not always predict behavior, especially in the domain of human-
computer interaction on the Internet where consumers are particularly concerned with
privacy violations. It is often reported that consumers are very concerned about for their
privacy, but they still disclose personal information even for small benefits (Norberg et

al., 2007). As such, consumers’ concern for privacy does not always determine their

98

behavior, a phenomenon often labeled the “privacy paradox” (Norberg etal., 2007).
Because the current study did not directly ask the participants to identify themselves, the
potential inconsistency between disclosure intentions and actual disclosure behavior
should be recognized as a lirrritation of this study. In this sense, future research should
employ more realistic scenarios where actual disclosure behaviors can be measured in the

domain of human-computer interaction.

99

Emergency Card & Info
Personal 8t Insurance Info
Emergency Contacts
Other Medical Conditions
Health Care Provrders
Pam. 8r Symptom Diary

My Shunt

APPENDIX A—l

EMERGENCY CARD & INFO

 

GMEDFiIe
Emergency Information Card
http://www.e-medfile.oom

 

 

 

 

 

 

 

Subscriber's
Photo Here
Name: Jane Smith
Home Phone:
Birth Date: 5/28/ 1984
Health Plan: Blue Cross
Group NumbEr: 30456
Member ID: 553 33 2233
Username:
Password:
Helm Providers
Lori Moore. MD. 555-555-5555
Michael Handler, MD. 444-444-4444
Enter-germ Contact:
Martha Smith 333-333-3333
Conditions
Hydrocephalus VP Shunt

 

 

 

100

If you are in an emergency situation, this

printable Emergency Information Card with
photo alerts emergency personnel to your

medical condition.

Imagine that you have a serious medical

condition but travel a lot. Now you don't

have to worry that an unfamiliar doctor in
an emergency might kill you.

Emergency Card 81 Info
Personal an Insurance Info
Emergency Contaas
other Medical Conditions
Health Care Provrders
Pain 5. Symptom Diary

My Shunt

 

 

PERSONAL 81 INSURANCE INFO

  

Fin! name:
un mm
Middle Initial!
ROI-honlhlpl
lino“:
Add“:
Mdmu
Ow:

Shh:

Zipi

Country:
Home phone:
Work phone:
Call phonon

 

APPENDIX A-2

“ V "V w% ‘5 .

Jon. 7 Sodomrsy
Smith Driver lkomor
[Mb P'ROI

Dough!“ unhdahi
nmvth’myhlnrocordlarg Gondon
4242 Donna Dov. shod typo:
Asp." Employer:
CO Employer oddmn
lmployor phenol

united Sun: of America Health plum

303 .61 6101 Group number:
Hamper Id!

203 090 0732

 
  
   

355 33 2213

Lo: Angola
3/ 2.! I 986
f

A!

Blue Cross
X3433
333 33 2233

You can store your personal and health Insurance lnfonnatlon that can be
transferred to your healthcare provider.

101

APPENDIX A-3

Emergency Card & info
Personal 8r Insurance Info
Emergency Contacts
Other Medical Conditions
Health Care Provuders
Pain In Symptom Diary

My Shunt

 

. an: , ' , _

7 Culley: United States of America
I _ Day Phone: 505 332 2122 , _ ,
Wane: $05 332 21227 7 7 _ _

_ ;.l

EM .1. ”in.“ r Jun 1 “my '2' “I

You can provide emergency contacts so that healthcare providers can
contact when you end up in the emergency room

 

102

Emergency Card Br Info
Personal 8‘ Insurance Info
Emergency Contads
other Medial Conditions
Health Care Providers
Pain & Symptom Diary

My Shunt

 

APPENDD( A-4

OTHER MEDICAL CONDITIONS

 

Data at moon/1935 7 _ , _ J
Gaunt: ,
Diamond by. Mark Stanley, M. D. _ g _ _ _’
Who Procadmas . , ._
Vanerqancy Card: Yea _ W
arm-ad by. Self ‘ , l
u). ._ .~ 1. a.-

You can enter your medical conditions and history including diagnoseso
your chronic illness

103

Emergency Card 3 Info
Personal 3. Insurance Info
Emergency Contact:
other Medical Conditions
Health Care Providers
Pain :1 Symptom Diary

My Shunt

HEALTH CARE PROVIDERS

  

 

APPENDIX A-5

   

   
 

llr-Jlthcare PIOVldi‘l s;

”as! “I " 'Jfi- -' V... r'

 

 

.6, w _. air -.
Wth‘ M. MD.
M ,
03.:- Mi _' ‘f’ _V 4
mm-PadauiCNaon gridrerglloapital __ W __,
Add"!!! 1056 E 19111 Avenue 8330 , _ 1

”may!
mm
11973021: j _, :7 _“ ' _
MIMI“ Statagof America .
«...... 303 as: 5101

Fwao; ss’i'eior _ _ ' _
.. so" _ __ N “i
_ _ W! . . - ...
inbred by:
1"! Se" ,_ ’9 f‘ .m ‘4- ‘ - ad

You can save your healthcare providers' contact Information.

Emergency Card 81 Info
Personal & Insurance Info
Emergency Contacts
Other Medical Conditions
Health Care Prayiders
Pain & Symptom Diary

My Shunt

 

 

APPENDIX A-6

PAIN & SYMPTOM DIARY

  

“' ..a "’ .3 3' if
i Event dau dd-mnr‘yyyy: 3/29[2003 »_
entdncwtion: Headache

- T!!! of:day

 

Agivity at gilget: ., _ -- ._ _ .... __~
i!a|1rat.i-ig:1-Md—~ _ 7 _ “A-” ’7 ,.,_.,_.
_ Pain Iocatgn: back othad _~ . ,__ _ __ __
Pandurationaomuoanw -.. _, 7_ l ll
l Wmdcafimfl’yiand __ W -.-L_, ._ __ _
.r “Entel'adbyll seir ,_ _ J
x Jr:- -* 4" w _’ Ff". Jude s-m

You can write your health diary by describing your symptoms in detail. Thls
allows both you and your healthcare provider to easily track and retrieve

 

your health conditions.

105

EmE'gemy Card 3 Icl:
Personal 3- Insurance Info
Eme'genf. Canto-Ts
Other Medical Conditions
Health Care Prowde’s
9am 8 Symptom Deny

My Shunt

APPENDIX A-7

MY SHUNT

 

Date origin-l Date of lay!

Shunt Type DIIgnOI" [hunt p revislon

‘v’entngu ogentgngl Arachnold Cyst 5.v'20,-"l939 9/20/1999

You can store scans and labs, and upload your medical documents such as
doctors' notes, test results, and CT images.

APPENDIX B

fol/owMe

   
 

 

 

’Have your private information
follow you wherever you go

Ollinercssto lie-Mme
Iedimlleoonlsi rounAnvwbere

ercessyourhealfl'iirrfo rmationwheneverltisneededby
you, yourdodororemergenw pummel.

0 Because physicians rfire. move, sell pradim and
records get “archived" and am to them an be
delayed.

. Manage important information withou having to store
and keep It in your home.

0 Perfed for patients or aregivers who manage a curred
or dimnic illness.

- A FollowMe 1m 00ij penniesperday.

 

 

6 20002007 Access Strategies, Inc. All rights reserved. CONTACY U! PRIVACV TERMS I CONDITIONS SWIM"

Powered by FollMe"

107

REFERENCES

Abrami, P. C., Leventhal, L., & Perry, R. P. (1982). Educational seduction. Review of
Educational Research, 52, 446-464.

Afifi,W. A.,& Burgoon, J. K. (1998).“We never talk about that”: A comparison of cross-
sex friendships and dating relationships on uncertainty and topic avoidance.
Personal Relationships, 5, 255—272.

Afifi, W. A., & Guerrero, L. K. (1998). Some things are better left unsaid II: Topic
avoidance in friendships. Communication Quarterly, 46, 231-249.

Ajzen, I., & Fishbein, M. (1980). Understanding Attitudes and Predicting Social
Behavior. Englewood Cliffs, NJ: Prentice-Hall, Inc.

Allen, M., Witt, P. L., & Wheeless, L. (2006). The role of teacher immediacy as a
motivational factor in student learning: Using meta-analysis to test a causal model.
Communication Education, 55, 21-31.

Altman, I., & Taylor, D. (1973). Social Penetration: The Development of Interpersonal
Relationships. New York: Holt, Rinehart, and Winston.

Andreasen, A. R. (1970). Personalizing mail questionnaire correspondence. Public
Opinion Quarterly, 34, 273-277.

Ansari, A., & Mela, C. F. (2003). E-customization. Journal of Marketing Research, 40,
131-145.

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

Argyle, M., T rimboli, C., & Forgas, J. P. (1988) Disclosure profiles in different
relationships. Journal of Social Psychology, 128, 117-124.

Awad, N., & Krishnan, M. (2006). The personalization privacy paradox: An empirical
evaluation of information transparency and the willingness to be profiled online
for personalization. MIS Quarterly, 30, 13-28.

Aylor, B., & Oppliger, P. (2003). Out-of-class communication and student perceptions of
instructor humor orientation and socio-communicative style. Communication
Education, 52, 122-134.

108

Bailenson, J. N., & Blascovich, J. (2004) Avatars. In W.S. Bainbridge’s (Eds),
Encyclopedia of Human-Computer Interaction (pp.64-68). Great Barrington, MA:
Berkshire Publishing Group.

Bandura, A. (1986). Social Foundations of Thought and action: A Social Cognitive
Theory. Englewood Cliffs, NJ: Prentice-Hall.

Bandura, A. (1991). Social cognitive theory of self-regulation. Organizational Behavior and
Human Decision Processes, 50, 248-287.

Bandura, A. (1997). Self-Efficacy: The Exercise of Control. New York: W. H. Freeman.

Bandura, A. (2001). Social cognitive theory: An agentic perspective. Annual Review of
Psychology, 52, 1-26.

Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in
social psychological research: Conceptual, strategic and statistical considerations.
Journal of Personality and Social Psychology, 51, 1173-1182.

Barrel], J. J ., & Jourard, S. M. (1976). Being honest with persons we like. Journal of
Individual Psychology, 32(2), 185-193.

Ban, 1. Y., Shankar, V., Sultan, F., Urban, G. L. (2005). Are the drivers and role of online
trust the same for all web sites and consumers? A large scale exploratory
empirical study. Journal of Marketing, 69, 133—152.

Baxter, L. A., & Mongomery, B. M. (1996). Relating: Dialgues and Dialectics. New
York: Guilford.

Baxter, L. A., & Wilmot, W. W. (1985). Taboo topics in close relationships. Journal of
Social and Personal Relationships, 2, 253-269.

Berendt, B., Gunther, 0., & Spiekermann, S. (2005). Privacy in E-commerce: Stated
preferences vs. actual behavior. Communications of The ACM, 48(4), 101-106.

Bickmore, T., & Cassell, J. (2001). Relational agents: A model and implementation of
building user trust. Proceedings of Human Factors Computing Systems (SIGCHI
01), ACM Press, New York, 396-403.

Biocca, F., Harms, C., & Burgoon, J. (2004). Towards a more robust theory and measure
of social presence: Review and suggested criteria. Presence: Teleoperators and

Virtual Environments, 12(5), 456-480.

Bocionek, S. R. (1995). Agent systems that negotiate and learn. International Journal of
Human-Computer Studies, 42, 265-288.

109

Bracken, C. C., & Lombard, M. (2004). Social presence and children: Praise, intrinsic
motivation and learning with computers. Journal of Communication, 54, 22-37.

Brave, S., N ass, C., & Hutchinson, K. (2005). Computers that care: Investigating the
effects of orientation of emotion exhibited by an embodied computer agent.
International Journal of Human-Computer Studies, 62(2), 161—178.

Brock, D. B., & Sulsky, L. M. (1994). Attitude toward computers: Construct validation
and relations to computer use. Journal of Organizational Behavior, I 5, 17-39.

Burgoon, J ., Buller, D., Hale, J ., & DeTurck, M. (1984). Relational messages associated
with nonverbal behaviors, Human Communication Research, 10(3), 351-378.

Burgoon, J ., Buller, D., &Woodall, W. (1989). Nonverbal communication, the unspoken
dialogue. New York: Harper and Row.

Cazier, J. A., Shao, B. B. M., & St. Louis, R. D. (2007). Sharing information and build
trust through value congruence. Information System Front, 9, 515-529.

Ceaparu, I., Lazar, J ., Bessiere, K., Robinson, J ., & Shneiderman, B. (2004). Determining
causes and severity of end-user frustration. International Journal of Human-
Computer Interaction, I 7(3), 333-356.

Chaikin A. L., & Derlega, V. J. (1974). Self-disclosure. Morristown, NJ: General
Learning Corporation.

Cialdini, R. B. (1993). Influence: Science and Practice (3rd ed.). New York, NY: Harper
Collins College Publishers.

Chellappa, R. K., & Sin, R. G. (2005). Personalization versus privacy: An empirical
examination of the online consumer’s dilemma. Information Technology and
Management, 6, 181-202.

Chesebro, J. L. (2003). Effects of teacher clarity and nonverbal immediacy on student

learning, receiver apprehension, and affect. Communication Education, 37, 323-
340.

Christophel, D. M. (1990). The relationships among teacher immediacy behaviors,
student motivation, and learning. Communication Education, 3 7, 323-340.

Christophel, D. M., & Gorham, J. (1995). A test-retest analysis of student motivation
teacher immediacy, and perceived source of motivation and demotivation in

college classes. Communication Education, 44, 292-306.

Clarke, A., & Smyth, M. (1993). A co-operative computer based on the principles of
human co-operation. International Journal of Man-Machine Studies, 38(1), 3-22.

110

Cline, R. J. (1989). The politics of intimacy: Costs and benefits determining disclosure
intimacy in male-female dyads. Journal of Social and Personal Relationships, 6,
5-20

Collins, N .L., & Miller, LC. (1994). Self-disclosure and liking: A meta-analytic review.
Psychological Bulletin, 116, 457-475.

Compeau, D. R., & Higgins, C. A. (1995). Computer self-efficacy: Development of a
measure and initial test. MIS Quarterly, 19(2), 189-211.

Cozby, P. C. (1973). Self-disclosure: A literature review. Psychological Bulletin, 79(2),
73-91.

Cranor, L. F., Reagle, J. & Ackerman, M. (1999). Beyond concern: Understanding net
users' attitudes about online privacy. Retrieved August 4, 2008, from
http://www.tprc.org/ABSTRACTS99/CRANORPAP.PDF

Culnan, M. J ., & Armstrong, P. K. (1999). Information privacy concerns, procedural
fairness, and impersonal trust: An empirical investigation. Organization Science,
10(1),104-115.

Culnan, M. J ., & Bies, R. J. (2003). Consumer privacy: Balancing economic and justice
considerations. The Journal of Social Issues, 59(2), 323-342.

Cyber Security Industry Alliance (2006). May 2006 DCS Survey Results. Retrieved
August 4, 2008, from

http://www.csialliance.org/publications/publications/surveys_and_polls/dci_surve
y_May2006/

Dahl, D. W., Manchanda, R. V., & Argo, J. J. (2001). Embarrassment in consumer purchase:
The roles of social presence and purchase familiarity. Journal of Consumer Research,
28(3), 473-481.

Danchak, M. M., Walther, J. B., & Swan, K. P. (2001, November). Presence in mediated
instruction: Bandwidth, behavior, and expectancy violations. Paper presented at
the annual meeting on Asynchronous Learning Networks, Orlando, FL.

Deighton, J. (1996). The future of interactive marketing. Harvard Business Review, 74, 4—
16.

Derlega V. J., & Chaikin, A. L. (1977). Privacy and self—disclosure in social relationships.
Journal of Social Issues, 33(3), 102-115.

Derlega, V. J ., & Grzelak, J. (1979). Appropriateness of self-disclosure. In G. J. Chelune

(Ed.), Self-Disclosure: Origins, Patterns, and Implications of Openness in
Interpersonal Relationships. San Francisco: Jossey-Bass.

111

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

Desmarais, M., Giroux, L., & Larochelle, S. (1993). An advice-giving interface based on
plan-recognition and user-knowledge assessment. International Journal of Man-
Machine Studies, 39, 901—924.

DeVito, J. A. (1986). The communication handbook: A dictionary. New York: Harper &
Row.

Dinev, T., & Hart, P. (2004). Internet privacy concerns and their antecedents -
measurement validity and a regression model. Behaviour & Information
Technology, 23(6), 413-423.

Dinev, T., & Hart, P. (2006a). An extended privacy calculus model for e-commerce
transactions. Information Systems Research, 17(1), 61-80.

Dinev, T., & Hart, P. (2006b). Privacy concerns and levels of information exchange: An
empirical investigation of intended e-services use. e-Service Journal, 4, 25-61.

Dobransky, N., & Frymier, A. B. (2004). Developing teacher-student relationship
through out of class communication. Communication Quarterly, 52, 211-223.

Dolnicar, S., & Jordaan, Y. (2007). A market-oriented approach to responsibly managing
information privacy concerns in direct marketing. Journal of Advertising, 36, 123-
149.

Doney, P. M., & Cannon, J. P. (1997). An examination of the nature of trust in buyer-
seller relationships. Journal of Marketing, 61 (2), 35-51.

Dwyer, F. R., Schurr, P. H., & Oh, S. (1987). Developing buyer-seller relationships.
Journal of Marketing, 51(2), 11-27.

Eagly, A. H., & Chaiken, S. (1993). The psychology of attitudes. Fort Worth, TX:
Harcourt Brace J ovanovich.

Earp, J, B., & Baumer, D. (2003). Innovative Web use to learn about consumer behavior
and online privacy. Communications of the ACM, 46(4), 81-83.

Emerson, R. (1976). Social exchange theory. Annual Review of Sociology, 2, 235-262.
Ferdig, R. E., & Mishra, P. (2004). Emotional responses to computers: Experiences in
unfairness, anger and spite. Journal of Educational Multimedia and Hypertext,

13(2), 143-161.

Fiske, S. T., & Taylor, S. E. (1984). Social Cognition, Boston: Addison-Wesley.

112

Fogg, B. J ., & Nass, C. (1996). Silicon sycophants: The effects of computers that flatter.
Paper presented at the annual meeting of the International Communication
Association, Chicago.

Frymier, A. (1994). A model of immediacy in the classroom. Communication Quarterly,
42, 133-144.

Frymier, A, & Houser, M. (2000). The teacher-student relationship as an interpersonal
relationship. Communication Education, 49, 207-219.

Fusani, D. S. (1994). Extra-class communication: Frequency, immediacy, self-disclosure,
and satisfaction in student-faculty interaction outside the classroom. Journal of
Applied Communication Research, 22, 232-255.

Gefen, D. (2002). Reflections on the dimensions of trust and trustworthiness among
online consumers. Database for Advances in Information Systems, 33(3), 38-53.

Gefen, D., & Silver, M. (1999). Lessons learned from the successful adoption of an ERP
system. Proceedings of the Proceedings of the 5th International Conference of the
Decision Sciences Institute, Athens, Greece, 1054-1057.

Gefen, D., & Straub, D. W. (2003). Managing user trust in B2C e-Services. e-Service
Journal, 2(2), 7—24.

Gefen, D., Karahanna, E., & Straub, D. W. (2003). Trust and TAM in online shopping:
An integrated model. MIS Quarterly, 27(1), 51-90.

Gibbs, J. L., Elison, N. B., & Heino, R. D. (2006). Self-presentation in online personals:
The role of anticipated future interaction, self-disclosure, and perceived success in
Internet dating. Communication Research, 33, 1-26.

Goffman, E. (1959). The Presentation of Self in Everyday Life, London: Penguin.

Gorham, J. (1988). The relationship between verbal teacher immediacy behavior and
student learning. Communication Education, 23, 40-53.

Hann, I. H., Hui, K. L., Lee, T. S. Y., & Png, I. P. L. (2007). Overcoming online
information privacy concerns: An information processing theory approach,
Journal of Management Information Systems, 42(2), 23-42

Hassanein, K., & Head, M. (2007). Manipulating social presence through the web

interface and its impact on attitude towards online shopping. International
Journal of Human-Computer Studies, 65(8), 689-708

113

Heerwcgh, D., Vanhove, T ., Matthijs, K., & Loosveldt, G. (2005). The effect of
personalization on response rates and data quality in web surveys. International
Journal of Social Research Methodology, 8, 85- 100.

Homans, G. C. (1958). Social behavior as exchange. American Journal of Sociology, 63
(6), 597-606.

Hosmer, L. T. (1995). Trust: The connecting link between organizational theory and
philosophical ethics. Academy of Management Review, 20, 379-403.

Hui, K. L., Teo, H. H., & Lee, S. Y. T. (2007). The value of privacy assurance: An
exploratory field experiment. MIS Quarterly, 31(1), 19—33.

J aasma, M. A., & Koper, R. J. (1999). The relationship between student-faculty out-of-
class communication to instructor immediacy and trust, and to student motivation.
Communication Education, 48, 41-47.

Johnson, D., Gardner, J ., & Wiles, J. (2004). Experience as a moderator of the media
equation: The impact of flattery and praise. International Journal of Human-
Computer Studies, 61, 237-258.

Johnstone, A., Berry, U., & Nguyen, T. (1994). There was a long pause: Influencing tum-
taking behavior in human-human and human-computer spoken dialogues.

International Journal of Human-Computer Studies, 41, 383-41 1.

J oinson, A., & Reips, U. (2007). Personalized salutation, power of sender and response
rates to Web-based surveys. Computers in Human Behavior, 23, 275-285.

J ourard, S. M. (1971). Self-Disclosure: An Experimental Analysis of the Transparent Self.
New York: Van Nostrand Reinhold.

Kalyanaraman, S., & Sundar, S. S. (2006). The psychological appeal of personalized
online content in Web portals: Does customization affect attitudes and behavior?

Journal of Communication, 56, 110-132.

Kearney, R, Plax, T. G., & Wendt-Wasco, N. J. (1985). Teacher immediacy for affective
learning indivergent college classes. Communication Quarterly, 33, 61-74.

Kelley, D. H., & Gorham, J. (1988). Effect of immediacy on recall of information.
Communication Education, 3 7, 198-207.

Kline, R. B. ( 1998). Principles and Practices of Structural Equation Modeling. New
York: Guilford Press.

Langer, E. J. (1989). Mindfulness. Boston: Addison-Wesley.

114

LaRose, R., & Bates, B. (1990). Virtual social distance and adoption of interactive

telecommunication technologies, Southemwestern Mass Communication Journal,
6, 34-43.

LaRose, R, & Eastin, M. S. (2004). A social cognitive theory of Internet uses and
gratifications: Toward a new model of media attendance. Journal of Broadcasting
& Electronic Media, 48, 399-420.

LaRose, R., & Rifon, N. (2007). Promoting i-safety: Effects of privacy warnings and
privacy seals on risk assessment and online privacy behavior. Journal of
Consumer Affairs, 41(1), 127—149.

LaRose, R., & Whitten, P. (2000). Rethinking instructional irmnediacy for web courses:
A social cognitive exploration. Communication Education, 49 (4), 320-338.

Laufer, R. S., & Wolfe, M. (1977). Privay as a concept and a social issues: A
multidimensional development theory. Journal of Social Issues, 33, 5- 10.

Laurenceau, J. P., Feldman Barrett, L., & Pietromonaco, P. R. (1998). Intimacy as an
interpersonal process: The importance of self-disclosure, partner disclosure, and
perceived partner responsiveness in interpersonal exchanges. Journal of
Personality and Social Psychology, 74, 1238-1251.

Lazar, J ., Bessiere, K., Ceaparu, 1., Robinson, J ., & Shneiderman, B. (2003). Help! I'm
lost: User frustration in web navigation. IT& Society, 1(3), 18-26.

Lee, D., & LaRose, R. (2007). A socio-cognitive model of video game usage. Journal of
Broadcasting & Electronic Media, 51(4), 632-650.

Lee, D., LaRose, R., & Rifon, N. (in press). Keeping our network safe: A model of online
protection behavior. Behaviour & Information Technology.

Lee, E., Nass, C., & Brave, S. (2000). Can computer-generated speech have gender? An
experimental test of gender stereotypes. In proceedings of the Computer-Human
Interaction ’00 conference.

Lee, K-M., & Nass, C. (2005). Social-psychological origins of feelings of presence:
Creating social presence with machine-generated voices. Media Psychology, 7(1),
3 1—45.

Lee, R., Nass, C., Brave, S., Morishima, Y., Nakajima, H., & Yamada, R. (2007). The
case for caring co-learners: The effects of a computer-mediated co-learner agent

on trust and learning. Journal of Communication, 57, 183-204.

Macias, W., & Lewis, L. (2003). A content analysis of direct-to-consumer (DTC)
prescription drug Web sites. The Journal of Advertising, 32(4), 43-56.

115

Malhotra, N. K., Kim, S. S., & Agarwal, J. (2004). Internet users’ information privacy
concerns (IUIPC): The construct, the scale, and a causal model information
systems research, Information Systems Research, 15, 336-355.

Mathews, A., Derlega, V. J ., & Morrow, J. A. (2006). What is highly personal
information and how is it related to self-disclosure decision making? The
perspective of college students. Communication Research Reports, 23(2), 85-92.

McAllister, D. J. (1995). Affect- and cognition-based trust as foundations for
interpersonal cooperation in organizations. Academy of Management Journal, 38,
24-59.

McGurire, W. J. (1969). The nature of attitudes and attitude change. In G. Lindsey, & E.
Aronson (Eds), Handbook of Social Psychology, Vol. 3 (pp.136-314). Boston:
Addison-Wesley.

McKnight, D. H., Choudhury, V., & Kacmar, C. (2002). The impact of initial consumer
trust on intentions to transact with a Web site: A trust building model. The
Journal of Strategic Information Systems. 11, 297-323.

McKnight, D. H., Cummings, L. L, & Chervany, N. L. (1998). Initial trust formation in
new organizational relationships. Academy of Management Review, 23(3), 472-
490.

Mehrabian, A. (1967). Orientation behaviors and nonverbal attitude communication.
Journal of Communication, 17, 324-332.

Mehrabian, A. (1971). Silent messages. Belmont, CA: Wadsworth.

Metzger, M. J. (2004) Privacy, trust and disclosure: Exploring barriers to electronic
commerce. Journal of Computer-Mediated Communication, 9 (4). Retrieved
August 4, 2008, from http://jcmc.indiana.edu/vol9/issue4/metzger.html

Mitchell, J. (1978 ). Social Exchange, Dramaturgy and Ethnomenthodology. Oxford:
Elsevier.

Millar, F., & Rogers, L. (1976). A relational approach to interpersonal communication. In
G. Miller (Ed.), Explorations in Interpersonal Communication (pp. 87-103).
Newbury Park, CA: Sage.

Milne, G. R. (2000). Privacy and ethical issues in database/interactive marketing and
public policy: A research framework and overview of the special issue. Joumal of

Public Policy & Marketing, 19(1), 1-6.

Milne, G. R., Bahl, S., & Rohm, A. J ,.(2008). Toward a framework for assessing covert
marketing practices. Journal of Public Policy & Marketing, 27 (1), 57- 62.

116

Milne, G. R., & Boza, M. (1999). Trust and concern in consumers' perceptions of

marketing information management practices. Journal of Interactive Marketing,
13, 5—24.

Milne, G. R., & Culnan, M. J. (2002). Using the content of online privacy notices to
inform public policy: A longitudinal analysis of the 1998-2001 US Web surveys.
Information Society, 18, 345-370.

Milne, G. R., & Culnan, M. J. (2004). Strategies for reducing online privacy risks: Why

consumers read (or don’t read) online privacy notices. Journal of Interactive
Marketing, 18(3), 15-29.

Milne, G. R., Rohm, A. J., & Bahl, S. (2004). Consumers' protection of online privacy
and identity. The Journal of Consumer Affairs, 38 (2), 217- 232.

Moon, Y. (2000). Intimate exchanges: Using computers to elicit self-disclosure from
consumers. Journal of Consumer Research, 26, 323—339.

Moon, Y., & Nass, C. (1996). "How real" are computer personalities? Psychological

responses to personality types in human-computer interaction. Communication
Research, 23(6), 651-674.

Moorman, C., Zaltman, G., & Deshpande, R. (1992). Relationships between providers
and users of market research: The dynamics of trust within and between
organizations. Journal of Marketing Research, 29, 314—28.

Morgan, R. M., & Hunt, S. D. (1994). The commitment-trust theory of relationship
marketing. Journal of Marketing, 58(3), 20-38.

Morton, T. (1978). Intimacy and reciprocity of exchange: A comparison of spouses and
strangers. Journal of Personal and Social Psychology, 36, 72-81

Mottet, T. P., & Beebe, S. A. (2006). The relationship between student responsive
behaviors, student socio-communicative style, and instructors’ subjective and

objective assessments of student work. Communication Education, 55(3), 295-
3 12.

Mottet, T. P., Martin, M. M., & Myers, S. A. (2004). Relationships among perceived
instructor verbal approach and avoidance relational strategies and students’

motives for communicating with their instructors. Communication Education, 53,
116-122.

Mottet, T. P., Parker-Raley, J ., Beebe, S. A., & Cunningham, C. (2007). Instructors who
resist “college lite”: The neutralizing effect of instructor immediacy on students’
course workload violations and perceptions of instructor credibility and affective
learning. Communication Education, 56(2), 145-167.

117

Myers, R. C., Davis, J. H., & Schoorman, D. (1995). An integrative model of
organizational trust. Academy of Management Review, 20(3), 709—734.

Myers, S. A. (2004). The relationship between perceived instructor credibility and
college student in-class and out-of-class communication. Communication Reports,
1 7, 129-137.

Myers, S. A., & Bryant, L. E. (2004). College students? Perceptions of how instructors
convey credibility. Qualitative Research Reports in Communication, 5, 22-27.

Myers, S. A., Martin, M. M., & Knapp, J. L. (2005). Perceived instructor in-class
communicative behaviors as a predictor of student participation in out of class
communication. Communication Quarterly, 53, 437-450.

Nadler, M. K., & Nadler, L. B. (2000). Out-of-class communication between faculty and
students: A faculty perspective. Communication Studies, 51, 176-188.

Narayandas, D., & Rangan, V. K. (2004). Building and sustaining buyer-seller
relationships in mature industrial markets. Journal of Marketing, 68(3), 63-77.

Nass,. C., Fogg, B. J., & Moon, Y. (1996). Can computers be teammaters? International
Journal of Human-Computer Studies, 45, 669-678.

Nass, C., & Lee, K. M. (2001). Does computer-synthesized speech manifest personality?
Experimental tests of recognition, similarity-attraction, and consistency-attraction.
Journal of Experimental Psychology: Applied, 7(3), 171- 181.

Nass, C., & Moon, Y. (2000). Machines and mindlessness: Social responses to computers.
Journal of Social Issues, 56(1), 81-103.

N ass, C., Moon, Y., & Carney, P. (1999). Are respondents polite to computers? Social
desirability and direct responses to computers. Journal of Applied Social
Psychology, 29(5), 1093-1 110.

Nass, C., Moon, Y., & Green, N. (1997). Are computers gender-neutral? Gender
stereotypic responses to computers. Journal of Applied Social Psychology, 27( 10),
864-876.

Negroponte, N. (1995). Being Digital. New York: Alfred P. Knopf.

Norberg, P. A., Home, D. R., & Horne, D. A. (2007). The privacy pardox: Personal

information disclosure intentions versus behaviors. Journal of Consumer Afiairs,
41(1), 100-126.

118

 

 

Nowak, K. (2000). The influence of anthropomorphism on mental models of agents and
avatars in social virtual enviorments. Unpublished Dissertation, Michigan State
University, East Lansing

Pastore, M. (1999). Consumers will provide information for personalization. Retrieved
February 21, 2002, from
http://cyberatlas.intemet.com/markets/advertising/article/0594 l_236 141,00.html

Peel, D. C. (2007). PHRs: Promise vs. problems. Retrieved August 4, 2008, from
http://www.aarpintemational.org/gra_sub/gra_sub_show.htm ?doc_id=54 1050

Perry, R. W., & Mankin, L. D. (2004). Understanding employee trust in management:
Conceptual clarification and correlates. Public Personnel Management, 33(3),
277-290.

Peslak, A. R. (2005). Privacy policies: A framework and survey of the Fortune 50.
Information Resources Management Journal, 18(1), 29-41

Petronio, S. (1991). Communication boundary management: A theoretical model of
managing disclosure of private information between marital couples.
Communication Theory, 1, 311-335.

Petronio, S. (2000). The boundaries of privacy: Praxis of everyday life. In S. Petronio
(Eds.), Balancing the Secrets of Private Disclosures (pp. 37-50). Mahwah, NJ:
Lawrence Erlbaum.

Petronio, S., Ellemers, N ., Giles, H., & Gallois, C. (1998). (Mis)communicating across

boundaries: Interpersonal and intergroup considerations. Communication
Research, 25, 571-595

Phelps, J, E., Nowak, G. J ., & Ferrell, E. (2000). Privacy concerns and consumer
willingness to provide personal information. Journal of Public Policy &
Marketing, 27(1), 27-41.

Rawlins, W. (1983). Openness as problematic in ongoing relationships: Two
conversational dilemmas. Communication Monographs, 50, 1-13.

Reeves, B. & Nass, C. (1996), The Media Equation: How People Treat Computers,
Television, and New Media Like Real People and Places. Cambridge, UK:
Cambridge University Press.

Rempel, J. K., Holmes, J. G., & Zanna, M. P. (1985). Trust in close relationships.
Journal of Personality and Social Psychology, 49(1), 95-112.

119

 

Rice, R. (1993). Media appropriateness: Using social presence theory to compare
traditional and new organizational media. Human Communication Research,
19(4), 451-484.

Richardson, J. C., & Swan, K. (2003). Examining social presence in online courses in
relation to students' perceived learning and satisfaction. Journal of Asynchronous
Learning Networks, 7(1), 68—88.

Richmond, V. P. (1990). Communication in the classroom: Power and motivation.
Communication Education, 39, 181-195.

Richmond, V. P., Gorham, J. S., 81. McCroskey, J. C. (1987). The relationship between
selected immediacy behaviors and cognitive learning. In M. A. McLaughlin (Ed.),
Communication Yearbook 10 (pp. 574-590). Newbury Park, CA: Sage.

Richmond, V. P., Lane, D. R., & McCroskey, J. C. (2006). Teacher imrniedacy and the
teacher-student relationship. In T. P. Mottet, V.P. Richmond, & J. C. McCroskey
(Eds.), Handbook of Instructional Communication: Rhetorical and Relational
perspectives (pp. 3-32). Boston: Allyn and Bacon

Richmond, V. P., & McCroskey, J. C. (2000). The impact of supervisor and subordinate

immediacy on relational and organizational outcomes. Communication
Monographs, 67, 85-95.

Richmond, V. P., McCroskey, J. C., Piax, T. G., & Kearney, P. (1986). Teacher
nonverbal immediacy training and student affect. World Communication, 15, 181-
194.

Rifon, N., LaRose, R., & Choi, S. M. (2005). Your privacy is sealed: Effects of Web
privacy seals on trust and false assurances. Journal of Consumer Afiairs, 39(2),
337-360.

Rodriguez, J. I., Plax, T. G., & Kearney, P. (1996). Clarifying the relationship between
teacher nonverbal immediacy and student cognitive learner: Affective learning as
the central causal mediator. Communication Education, 45, 293-305.

Rohm, A. J ., & Milne, J. R. (2004). Just what the doctor ordered: The role of information

sensitivity and trust in reducing medical information privacy concern. Journal of
Business Research, 31, 163-170.

Rollof, M. E., & Ifert, D. (1998). Antecedents and consequences of explicit agreements to
declare a taboo topic in dating relationships. Personal Relationships, 5, 191-205.

120

 

Rollof, M. E., & Ifert, D. (2000). Conflict management through avoidance: Withholding
complaints, suppressing arguments, and declaring topics taboo. In S. Petronio
(Eds), Balancing the Secrets of Private Disclosures (pp. 151-163). Mahwah, NJ:
Erlbaum.

Rosenfeld, L. B. (1979). Self-disclosure avoidance: Why am I afraid to tell you who I
am? Communication Monographs, 46, 63-74.

Rosenfeld, P., Booth-Kewley, S., Edwards, J. E., & Thomas, M. D. (1996). Responses on
computer surveys: Impression management, social desirability, and the big
brother syndrome. Computers in Human Behavior, 12, 263-274.

Rothstein, M. A., & T albott, M. K. (2006). Compelled disclosure of health information:
Protecting against the greatest potential threat to privacy. The journal of the
American Medical Association, 295, 2882-2885.

Rozell, E. J ., & Gardner HI, W. L. (2000). Cognitive, motivation, and affective processes
associated with computer-related performance: A path analysis. Computers in
Human Behavior, 16, 199-222.

Rubin, Z. (1975). Disclosing oneself to a stranger: Reciprocity and its limits. Journal of
Experimental Social Psychology, 11, 233-260.

Schoenbachler, D. D., & Gordon, G. L. (2002). Trust and customer willingness to
provide information in database-driven relationship marketing. Journal of
Interactive Marketing, 16, 2-16.

Schrodt, P., & Witt, P. L. (2006). Students’ attributions of instructor credibility as a
function of students’ expectations of instructional technology use and nonverbal
immediacy. Communication Education, 55, 1-20.

Shapiro, S. P. (1987). The social control of impersonal trust. American Journal of
Sociology, 93(3), 623-638.

Short, J ., Williams, E., & Christie, B. (1976). The Social Psychology of
Telecommunications. London: John Wiley & Sons.

Sprague, L. (2006). Personal health record: The people’s choice? NHPF Issue Brief, 820,
1-13.

Sproull, L., Subramani, M., Kiesler, S., Walker, J. H., & Waters, K. (1996). When the
interface is a face. Human-Computer Interaction, I 1, 97—124.

Swan, J. E., & Bowers, M. R., & Richardson, L. D. (1999). Customer trust in the

salesperson: An integrative review and meta-analysis of the empirical literature.
Journal of Business Research, 44(2), 93-107.

121

 

Taylor, D., & Altman, I. (1975). Self-disclosure as a function of reward-cost outcome.
Sociometry, 38(1), 18-31.

Teven, J. J ., & Hanson, T. L. (2004). The impact of teacher immediacy and perceived
teacher caring on teacher competence and trustworthiness. Communication
Quarterly, 52, 39-53.

Thibaut, J. W., & Kelley, H. (1959). The Social Psychology of Groups. New York: Wiley.

Titsworth, B. S. (2001). The effects of teacher immediacy, use of organizational lecture

cues, and students’ notetaking on cognitive learning. Communication Education,
50, 283-297.

Tourangeau, R., Couper, M. P., & Steiger, D. M. (2003). Humanizing self-administered
surveys: Experiments on social presence in Web and IVR surveys. Computers in
Human Behavior, 19, 1-24.

T urkle, S. (1984). The Second Self: Computers and Human Spirit. New York: Simon &
Schuster.

Turkle, S. (1995). Life on the Screen: Identity in the Age of the lntemet. New York:
Simon & Schuster.

van Dijk, J. (1999). The Network Society: Social Aspects of New Media. London: Sage.

Waldeck, J. H. (2006). What does “personalized education” mean for faculty and how
should it serve out students? Communication Education, 55, 345-352.

Walther, J. B. (1992). Interpersonal effects in computer-mediated interaction: A relational
perspective. Communication Research, 19, 52-90.

Walther, J. B. (1994). Anticipated ongoing interaction versus channel effects on
relational communication in computer-mediated interaction. Human
Communication Research, 20, 473-501.

Walther, J. B., & Burgoon, J. K. (1992). Relational communication in computer-mediated
interaction. Human Communication Research, 19, 50-88.

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, 105-34.

Ward, S., Bridges, K., & Chitty, B. (2005). Do incentive matter? An examination of on-

line privacy concerns and willingness to provide personal and financial
information. Journal of Marketing Communications, 11, 21-40.

122

 

 

Weisband, S., & Kiesler, S. (1986). Self-disclosure on computer forms: Meta-analysis
and implications. Proceedings of CHI96. Retrieved Oct 20, 2006, from
http://www.acm.org/sigchi/chi96/papers/Weisband/sw_txt.htm.

Wheeless, L. R. (1978). A follow-up study of the relationships among trust, disclusre,
and interpersonal solidity. Human Communication Research, 4, 143-157.

Wheeless, L. R., & Grotz, J. (1976). Conceptualization and measurement of reported self-
disclosure. Human Communication Research, 2, 338-346.

White, T. B. (2004). Consumer disclosure and disclosure avoidance: A motivational
framework. Journal of Consumer Psychology, 14, 41-51.

Wiener, M., & Mehrabian, A. (1968). Language within Language: Immediacy, a Channel
in Verbal Communication. New York: Appleton-Century-Crofts.

Zahedi, F., & Song, J. (2008). Dynamics of trust revision: Using health infomediaries.
Journal of Management Information Systems, 24(4), 225-248.

Zhang, Q., & Oetzel, J. G. (2006). A cross-cultural test of immediacy-learning models in
Chineses classrooms. Communication Education, 55, 313-330.

123