Predicting Collaboration Technology Use: Integrating Technology Adoption and

Collaboration Research
Author(s): SUSAN A. BROWN, ALAN R. DENNIS and VISWANATH VENKATESH
Source: Journal of Management Information Systems , Fall 2010, Vol. 27, No. 2 (Fall
2010), pp. 9-53
Published by: Taylor & Francis, Ltd.

Stable URL: https://www.jstor.org/stable/29780170

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

Taylor & Francis, Ltd. is collaborating with JSTOR to digitize, preserve and extend access to
Journal of Management Information Systems

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
Predicting Collaboration Technology Use:
Integrating Technology Adoption and
Collaboration Research
SUSAN A. BROWN, ALAN R. DENNIS, AND
VISWANATH VENKATESH

Susan A. Brown is an Associate Professor in MIS and a McCoy-Rogers Fellow in the

Eller College of Management at the University of Arizona. She received her Ph.D.
from the University of Minnesota and an MBA from Syracuse University. Her research
interests include technology implementation, individual adoption, computer-mediated
communication, technology-mediated learning, and related topics. Her research has
been published in MIS Quarterly, Information Systems Research, Organizational
Behavior and Human Decision Processes, IEEE Transactions on Engineering Man?
agement, Communications of the ACM, Journal of the AIS, and others. She has served
or is currently serving as an associate editor for MIS Quarterly, Information Systems
Research, Journal of the AIS, and Decision Sciences. She received MIS Quarterly's
Reviewer of the Year award for 2001, and in 2008 she received the Best Associate
Editor award from Information Systems Research.

Alan R. Dennis is a Professor of Information Systems and holds the John T. Chambers
Chair of Internet Systems in the Kelley School of Business at Indiana University.
Professor Dennis has written more than 100 research papers and has won numerous
awards for his theoretical and applied research. His research focuses on four main
themes: the use of computer technologies to support team creativity and decision
making, knowledge management, the use of the Internet to improve business and
education, and professional issues facing IS academics (e.g., business school rankings
and the difficulty of publishing and getting tenure in IS). He was a senior editor at
MIS Quarterly and is the founding publisher of MIS Quarterly Executive, a journal
focusing on applied research designed to improve practice.

Viswanath Venkatesh is a Professor and Billingsley Chair in Information Systems

at the Walton College of Business, University of Arkansas. He received his Ph.D.
from the University of Minnesota in 1997. His research focuses on understanding
the diffusion of technologies in organizations and society. His work has appeared
or is forthcoming in leading IS, organizational behavior, operations management,
marketing, and psychology journals. His articles have been cited about 12,000 times
per Google Scholar and about 4,300 times per Web of Science. Some of his papers
published in various journals (Decision Sciences 1996, Information Systems Research
2000, Management Science 2000, and MIS Quarterly 2003) are among the most cited
papers published in the respective journals. His MIS Quarterly (2003) paper has been

Authors are listed alphabetically. All authors contributed equally.

Journal of Management Information Systems /Fall 2010, Vol. 27, No. 2, pp. 9-53.

?2010M.E. Sharpe, Inc.

0742-1222 / 2010 $9.50 + 0.00.
DOI 10.2753/MIS0742-1222270201

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
 10 BROWN, DENNIS, AND VENKATESH

identified by ScienceWatch as the most influential paper in one of only four research
fronts in business and economics. He currently serves as a senior editor at Information
Systems Research and AIS Transactions on Human-Computer Interaction. He has
also served or serves on the editorial boards of Decision Sciences, Journal of the AIS,
Journal of Management Information Systems, Production and Operations Manage?
ment, Management Science, and MIS Quarterly. In October 2009 he launched an IS
research rankings Web site: http://vvenkatesh.com/ISRanking/.

Abstract: The paper presents a model integrating theories from collaboration research
(i.e., social presence theory, channel expansion theory, and the task closure model) with
a recent theory from technology adoption research (i.e., unified theory of acceptance
and use of technology, abbreviated to UTAUT) to explain the adoption and use of
collaboration technology. We theorize that collaboration technology characteristics,
individual and group characteristics, task characteristics, and situational characteristics
are predictors of performance expectancy, effort expectancy, social influence, and
facilitating conditions in UTAUT. We further theorize that the UTAUT constructs,
in concert with gender, age, and experience, predict intention to use a collaboration
technology, which in turn predicts use. We conducted two field studies in Finland
among (1) 349 short message service (SMS) users and (2) 447 employees who were
potential users of a new collaboration technology in an organization. Our model was
supported in both studies. The current work contributes to research by developing and
testing a technology-specific model of adoption in the collaboration context.

Key words and phrases: channel expansion theory, collaboration technologies, social
presence theory, task closure model, technology acceptance, technology adoption,
unified theory of acceptance and use of technology.

Technology adoption1 is one of the most mature streams in information systems

(IS) research (see [65, 76, 77]). The benefit of such maturity is the availability of
frameworks and models that can be applied to the study of interesting problems.
While practical contributions are certain to accrue from such investigations, a key
challenge for researchers is to ensure that studies yield meaningful scientific con?
tributions. There have been several models explaining technology adoption and use,
particularly since the late 1980s [76]. In addition to noting the maturity of this stream
of research, Venkatesh et al. identified several important directions for future research
and suggested that "one of the most important directions for future research is to tie
this mature stream [technology adoption] of research into other established streams
of work" [76, p. 470] (see also [70]).
In research on technology adoption, the technology acceptance model (TAM) [17]
is the most widely employed theoretical model [76]. TAM has been applied to a range
of technologies and has been very predictive of individual technology adoption and
use. The unified theory of acceptance and use of technology (UTAUT) [76] integrated
eight distinct models of technology adoption and use, including TAM. UTAUT extends
TAM by incorporating social influence and facilitating conditions. UTAUT is based in

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
 PREDICTING COLLABORATION TECHNOLOGY USE 11

the rich tradition of TAM and provides a foundation for future research in technology
adoption. UTAUT also incorporates four different moderators of key relationships.
Although UTAUT is more integrative, like TAM, it still suffers from the limitation of
being predictive but not particularly useful in providing explanations that can be used
to design interventions that foster adoption (e.g., [72, 73]).
There has been some research on general antecedents of perceived usefulness and
perceived ease of use that are technology independent (e.g., [69, 73]). But far less at?
tention has been paid to technology-specific antecedents that may provide significantly
stronger guidance for the successful design and implementation of specific types of
systems. Developing theory that is more focused and context specific?here, technol?
ogy specific?is considered an important frontier for advances in IS research [53,70].
Building on UTAUT to develop a model that will be more helpful will require a better
understanding of how the UTAUT factors play out with different technologies [7,76].
As a first step, it is important to extend UTAUT to a specific class of technologies
[70, 76]. A model focused on a specific class of technology will be more explanatory
compared to a general model that attempts to address many classes of technologies
[70]. Such a focused model will also provide designers and managers with levers to
augment adoption and use. One example is collaboration technology [20], a technology
designed to assist two or more people to work together at the same place and time or
at different places or different times [25, 26].
Technologies that facilitate collaboration via electronic means have become an
important component of day-to-day life (both in and out of the workplace). Thus, it
is not surprising that collaboration technologies have received considerable research
attention over the past decades [24,26,77]. Several studies have examined the adoption
of collaboration technologies, such as voice mail, e-mail, and group support systems
(e.g., [3, 4, 44, 56, 63]). These studies focused on organizational factors leading to
adoption (e.g., size, centralization) or on testing the boundary conditions of TAM
(e.g., could TAM be applied to collaboration technologies). Given that adoption of
collaboration technologies is not progressing as fast or as broadly as expected [20,
54], it seems a different approach is needed. It is possible that these two streams
could inform each other to develop a more complete understanding of collaboration
technology use, one in which we can begin to understand how collaboration factors
influence adoption and use.
A model that integrates knowledge from technology adoption and collaboration
technology research is lacking, a void that this paper seeks to address. In doing so,
we answer the call for research by Venkatesh et al. [76] to integrate the technology
adoption stream with another dominant research stream, which in turn will move us
toward a more cumulative and expansive nomological network (see [41, 70]). We
also build on the work of Wixom and Todd [80] by examining the important role of
technology characteristics leading to use. The current study will help us take a step
toward alleviating one of the criticisms of IS research discussed by Benbasat and Zmud,
especially in the context of technology adoption research: "we should neither focus our
research on variables outside the nomological net nor exclusively on intermediate-level
variables, such as ease of use, usefulness or behavioral intentions, without clarifying

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
 12 BROWN, DENNIS, AND VENKATESH

the IS nuances involved" [6, p. 193]. Specifically, our work accomplishes the goal of
"developing conceptualizations and theories of IT [information technology] artifacts;
and incorporating such conceptualizations and theories of IT artifacts" [53, p. 130]
by extending UTAUT to incorporate the specific artifact of collaboration technology
and its related characteristics. In addition to the scientific value, such a model will
provide greater value to practitioners who are attempting to foster successful use of
a specific technology.
Given this background, the primary objective of this paper is to develop and test a
model to understand collaboration technology adoption that integrates UTAUT with
key constructs from theories about collaboration technologies. We identify specific
antecedents to UTAUT constructs by drawing from social presence theory [64],
channel expansion theory [11] (a descendant of media richness theory [16]), and the
task closure model [66], as well as a broad range of prior collaboration technology
research. We test our model in two different studies conducted in Finland: the use of
short message service (SMS) among working professionals and the use of a collabora?
tion technology in an organization.

Background
Collaboration Technology
Collaboration technology is a package of hardware and software that can provide
one or more of the following: (1) support for communication among participants,
such as electronic communication to augment or replace verbal communication;
(2) information-processing support, such as mathematical modeling or voting tools;
and (3) support to help participants adopt and use the technology, such as agenda tools
or real-time training (e.g., [24, 26, 81]). A variety of terms have been used to refer
to collaboration technology over the years?such as group decision support systems,
group support systems, electronic meeting systems, groupware, computer-supported
cooperative work, and negotiation support systems?but these, as well as specific sys?
tems, such as e-mail, voice mail, and videoconferencing, are generally encompassed
under the larger umbrella term of collaboration technology.
Collaboration technology has been the subject of formal research at least since the
1970s, although its emergence as a key domain of research did not occur until the
1980s [18]. Many reviews of collaboration technology research have been published
over the years outlining the development of research and highlighting trends in the
empirical results [21,24,31,32]. While early collaboration technology research initia?
tives were centered on decision room environments [18], attention has more recently
turned to collaboration technologies that support virtual teams and distributed work
(e.g., e-mail, instant messaging, asynchronous discussion tools).
As collaboration technologies have evolved, our understanding of what contributes
to their use has not kept pace [62]. Past research has found that the use of collaboration
technology can produce strikingly different outcomes [21, 24, 31, 32]. Communica?
tion tools, such as e-mail, produce outcomes different from what is produced by more

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
 PREDICTING COLLABORATION TECHNOLOGY USE 13

complete systems that include information-processing support tools [21]. Likewise,

the nature of the task?such as decision making versus idea generation [21]?can
influence the value of a particular technology. Thus, the fit of the technology to the
task is important [24, 81]. Past research [24] suggests several key observations. First,
if the technology fits the needs of the task, then the use of collaboration technology
can improve decision quality and increase the number of ideas generated compared
to not using it [24]. But, if the technology is a poor fit, little is gained, at least initially
(see also [35]). Further, the aspects of the technology that the group chooses to use
and how they use them affect outcomes. Second, if groups new to a collaboration
technology receive no support in choosing what aspects of the technology to use and
guidance on how to use them, they take longer to complete tasks than groups working
without technology [24]. If these same groups receive support (or if they have prior
experience with the technology and task), they take less time and are more satisfied
[24]. Thus, use is a key factor affecting group performance.

Unified Theory of Acceptance and Use of Technology

Venkatesh et al. [76] proposed a unified model?namely, UTAUT?that incorporates
four key predictors of intention to use technology: performance expectancy, effort
expectancy, social influence, and facilitating conditions. Intention, in turn, predicts
technology use. Performance expectancy is the extent to which an individual perceives
that using a system will enhance his or her productivity, and thus lead to performance
gains?performance expectancy is conceptually and empirically identical to perceived
usefulness from TAM [76]. Effort expectancy is the extent to which using a system
is free from effort?effort expectancy is conceptually and empirically identical to
perceived ease of use from TAM [76]; note that high effort expectancy suggests high
ease of use and not high effort. Social influence is the extent to which an individual
perceives that important others think that he or she should use the target system [67,73].
Facilitating conditions is the perception regarding the availability of organizational and
technical resources to support use of the target system [76]. Further, UTAUT argues that
the various relationships are moderated by a combination of gender, age, experience,
and voluntariness [76]. Several studies have reported use of systems in organizations
as being either voluntary [69, 73, 76] or mandatory [9, 73, 76]. Given that our work
focuses on voluntary contexts, we exclude voluntariness from our model.
As noted earlier, the key question of interest in this paper is: Why do people choose
to use collaboration technology? In general, people adopt a technology because they
believe it will be useful in improving the effectiveness and efficiency of performing
some task [20, 76]. These effectiveness and efficiency motives correspond directly to
core underpinnings of performance expectancy and effort expectancy, respectively,
thus making UTAUT particularly suitable as the basis for the model development.
UTAUT also accounts for social influences and environmental factors not considered
in the original conceptualization of TAM. Further, UTAUT has explained over 70
percent variance in intention to use several different technologies [76], thus making
it a robust and comprehensive model.

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
 14 BROWN, DENNIS, AND VENKATESH

Model Development
UTAUT is a general model of technology Adoption and use. UTAUT and related
theories (e.g., TAM) have been successfully applied in a wide range of settings and
across diverse technologies. Yet there is nothing in the model that differentiates
across the characteristics of a use situation (i.e., a specific technology, its potential
users, and context of use). There is nothing in UTAUT by itself that directly helps
us in understanding what leads to the adoption of collaboration technology. UTAUT
argues that beliefs about performance and effort influence the decision to adopt and
use, but what influences these beliefs? In order to understand the factors that influ?
ence the performance and effort beliefs, we need to turn to theories that focus on the
situation of use. In this case, we need to begin with theories about collaboration and
link the key factors from these theories to the key factors of UTAUT to understand
how situational factors influence the ultimate decision to adopt and use collaboration
technology. In sum, we argue that UTAUT mediates the relationship between the
characteristics of a use situation and the ultimate adoption and use of a technology.
Therefore, our model of collaboration technology adoption and use begins with the
characteristics of situations in which the technology might be used. We focus on fac?
tors that have been important in past collaboration research?characteristics of the
collaboration technology, its potential users, their tasks, and the context. We argue
that these characteristics do not directly affect adoption and use, but rather influence
UTAUT factors (e.g., performance and effort expectancy), which in turn influences
adoption and use. Thus, UTAUT is the mediating mechanism through which the situ?
ational characteristics influence adoption and use.
Several factors have been suggested to influence the performance and satisfaction
of individuals and groups using collaboration technology [24, 31, 32]. These factors
largely fall into four major characteristics?technology, individual and group, task,
and situational (e.g., organizational context) [24, 25, 31, 32, 81]. Figure 1, adapted
from Dennis et al. [25], illustrates how these factors affect technology use and the
outcomes from group work. In developing a model to explain the adoption and use of
collaboration technology, we began with the four sets of factors argued to be important
in influencing the successful use of collaboration technology?technology, individual
and group, task, and situational (Figure 1). Research on collaboration technology has
examined how these factors affect performance (e.g., [37, 50]) but has not examined
how they influence adoption and use. We suggest that the mechanisms by which the
aspects of collaboration technology influence adoption and use are the cognitions
identified in UTAUT. Some evidence supporting this view is found in prior research?
for example, Fulk [34] found that perceptions of richness influenced perceptions of
usefulness, which ultimately affected use of e-mail.
We develop a model relating collaboration technology constructs to key constructs
in UTAUT using the framework of Dennis et al. [25]. Figure 2 presents our research
model. In the sections that follow, we define the key constructs and develop the theo?
retical arguments for the proposed relationships. We begin with a discussion of the
key UTAUT relationships and then move to the collaboration technology-specific
antecedents of the four key predictors in UTAUT.

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
 PREDICTING COLLABORATION TECHNOLOGY USE 15

Predicting Collaboration Technology Use:

Adapting UTAUT Hypotheses
Performance expectancy, the extent to which use is expected to improve work perfor?
mance, has been one of the most consistent predictors of behavioral intention across
technologies (see [76]), including communication technologies (e.g., [39]). The more
individuals expect that using a technology will improve their performance, the more
likely they are to use it [76]. Gender and age moderate this relationship [76]?men,
particularly younger men, have a greater focus on their tasks, productivity, and ef?
fectiveness [51, 75, 76]. Therefore, men, especially younger men, will place greater
importance on performance expectancy in evaluating IS in general [52,74]. This same
pattern can be expected in the effect of performance expectancy on intention to use
a collaboration technology, as such technologies have the potential to be minimally
disruptive to one's work in terms of time relative to alternatives, such as a face-to-face
meeting, and potentially help increase productivity. Thus, we hypothesize:2

Hypothesis la: The effect of performance expectancy on intention to use collabo?

ration technology will be moderated by gender and age such that it is strongest
for younger men.

Effort expectancy, the extent to which use is expected to be free of effort, has been
shown to be a predictor of intention [68,69; see 76 for a review]. Effort expectancy can
be particularly important in the context of personal technologies and non-workplace
settings [70] and has been identified as an important predictor in the context of com?
munication technologies [61]. Effort expectancy has both a direct effect on intention
and an indirect effect through performance expectancy [76]. As a technology is
perceived to take more effort to use, the less likely individuals are to intend to use it

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
16 BROWN, DENNIS, AND VENKATESH

Collaboration-related
constructs UTAUT

Figure 2. Research Model

[76]. Also, the more effort it takes to use a technology, the less useful the technology
is perceived to be [17, 69, 73].
As with performance expectancy, UTAUT hypothesized that the effect of effort
expectancy on intention to use will be moderated by gender and age; effort expended
using a technology is more salient to women than men [52, 74] and to older workers
than younger workers [76]. Therefore, the effect of effort expectancy on intention to
use a collaboration technology will be stronger for women and for older workers. We
expect such effects to possibly be intensified in the context of collaboration technology
use as women and older individuals particularly value communication as an important
aspect of their day-to-day functioning. Further, there is evidence to suggest that effort
expectancy has less of an effect on those with greater experience, because as experi?
ence increases, users have overcome the initial hurdles to use and effort expectancy
becomes less important [76]. Thus, we hypothesize:

Hypothesis lb: The effect of effort expectancy on intention to use collaboration

technology will be moderated by gender, age, and experience such that the effect
will be strongest for older women with little experience.

The role of social influence, the extent to which the individual perceives that im?
portant others believe he or she should use the system, has been somewhat unclear.
Initially, Davis et al. [17] did not include subjective norm, also called social influence
[76], in TAM, because it was the least understood aspect of the theory of reasoned
action (see [17]). However, other models have included various aspects of social influ

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
 PREDICTING COLLABORATION TECHNOLOGY USE 17

ence (see [68,76]). More recently, social influence has been incorporated in TAM, but
only in the presence of certain moderating variables, such as organizational mandate
[73] or gender [74]. Outside the workplace context, in predicting adoption of PCs in
homes, social influence has been found to be important [7, 70]. Despite the impor?
tance of social influence as a predictor of intention and behavior in certain situations,
it has been found to be of limited importance for those with significant technology
experience?that is, the views of others weigh heavily in adoption decisions before
one has acquired sufficient experience to feel confident about making an independent
decision [73, 74, 76].
We expect social influence will be important for collaboration technologies because
they are "social" technologies. Unlike the individual technologies studied in much prior
research, communication technologies cannot be used alone. If the normative pressure
is negative to the point of dissuading use, then there may be no potential communica?
tion partners. In such a case, intention to use a target collaboration technology could be
dampened. The converse is also true, where increased normative pressure, evidenced
via a "critical mass" of users [47], could lead to higher intention to use [44].
Gender, age, and experience moderate the relationship between social influence and
intention [76]. Women and older individuals not only place greater value on relation?
ships but they also value information from peers and friends more highly [76]. Like
effort expectancy, it can be expected that such an effect will play a role in the context
of collaboration technologies as such technologies, as noted earlier, are "social"
technologies and a critical mass of communication partners is quite important. Older
individuals, women in particular, will look to their peers and friends for reinforcing
messages to drive their own use. Increasing experience will dampen the relationship
between social influence and intention as experience allows the individual to rely on
his or her own judgment rather than that of others [76]. Thus, we hypothesize:

Hypothesis lc: The effect of social influence on behavioral intention to use col?
laboration technology will be moderated by gender, age, and experience such that
the effect will be strongest for older women with little experience.

The role of facilitating conditions, the extent to which the individual believes the
organization and technical infrastructure support use of the system, has also been
somewhat unclear. For example, Taylor and Todd [67] found support for perceived
behavioral control, which is conceptually similar to facilitating conditions [76], in
predicting both intention and use, whereas Thompson et al. [68] found facilitating
conditions to be a nonsignificant predictor of use and Venkatesh [69] found that effort
expectancy fully mediated the relationship between facilitating conditions and inten?
tion. Recently, Brown and Venkatesh [7] demonstrated the importance of facilitating
conditions in household adoption of PCs, even in the presence of effort expectancy.
While the results of prior work have been mixed, we expect that facilitating condi?
tions will be relevant for collaboration technology use. Due to the networked nature
of these technologies, the absence of technical resources to support their use will have
a strong negative effect on use. Likewise, if organizational support for collaborating
via technology is lacking, individuals will likely turn to the collaboration modes that

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
18 BROWN, DENNIS, AND VENKATESH

are supported within the organization. Venkatesh et al. [76] argue that, with experi?
ence, individuals are more able to seek out and find the assistance they need to use the
technology. Morris and Venkatesh [51] also show that as people age, the importance
to them of assistance to enable technology use increases. Thus, we hypothesize:

Hypothesis Id: The effect of facilitating conditions on collaboration technology

use will be moderated by age and experience, such that the effect is stronger for
older users, particularly those with little experience.

Antecedents of UTAUT Constructs in Prior Research

UTAUT is based on almost two decades of research on technology adoption and use.
Yet little is known about the key antecedents that influence the UTAUT constructs. As
noted earlier, some work has identified general, technology-independent antecedents
of performance expectancy and effort expectancy (e.g., [40,69,73]). Specifically, the
antecedents of performance expectancy were identified to be cognitions and social
influence [73], and the antecedents of perceived ease of use were identified to be in?
dividual's general technology beliefs and individual's perceptions of the system [69].
Other work has incorporated general psychological variables, such as trust tied to
the context of technology use (e.g., [43]). However, the constructs identified in prior
research were not tied to the technology or its conceptualization, and the research
did not consider the nuances associated with the specific type of technology or class
of technology being studied. To address this gap in the research, we develop a set of
antecedents of performance expectancy and effort expectancy that are drawn from
prior research on collaboration technologies. It should be noted that although social
influence and facilitating conditions are part of UTAUT, they represent external influ?
ences that relate to the social and organizational environment. Therefore, we do not
expect the collaboration technology-related constructs to influence either of those
constructs. Our model does, however, incorporate task and situational factors. As
noted earlier, in our model there are four sets of factors that we theorize to influence
the intention to use collaboration technology?technology characteristics, individual
and group characteristics, task characteristics, and situational characteristics. In the
sections below, we describe the characteristics and then identify specific constructs
within each set that we believe will have an effect on intention to use collaboration
technologies.

Technology Characteristics
Collaboration technologies have both innate physical and socially derived charac?
teristics [22, 34]. Many, or even most, of the characteristics that have been ascribed
to collaboration technologies are not innate physical characteristics, but are instead
socially derived characteristics, such as social presence and immediacy of feedback
[11,22, 34]. The perceptions of these socially derived collaboration technology char

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
 PREDICTING COLLABORATION TECHNOLOGY USE 19

acteristics can differ from person to person based on the person's skills, knowledge,
and personality and on the way he or she chooses to use the technology [34]. One
person may perceive that a specific collaboration technology tool has high social pres?
ence, while another person using the same tool may perceive that it has low social
presence. In fact, it is not uncommon for a given person's perceptions to change over
time, so a tool that is seen as having low social presence today may be seen as having
medium social presence next month?or even next week [11, 35]. Thus, in assessing
the characteristics of a collaboration technology, it is important to not focus on the
innate, supposedly "objective" physical characteristics of a specific technology, but
rather on the socially derived characteristics as perceived by individual users, which
typically differ from person to person [11, 22, 34].
A vast body of research has consistently shown that various characteristics of the
technology as experienced by users can potentially influence various outcomes [24,26,
69]. Such user perceptions of the technology characteristics comprise the first set of
factors that may influence adoption and use. In order to identify specific collaboration
technology characteristics that would be antecedents of the UTAUT constructs of per?
formance expectancy and effort expectancy, one of the most promising places to start is
with theories that attempt to explain why individuals choose to use one communication
medium over another?that is, media choice. In the current study, we examine three of
the more important theories that have shaped the choice of collaboration technologies
in general?social presence theory; media richness theory and its descendants, such
as channel expansion theory; and the task closure model. We leverage this research to
identify three specific technology characteristics?social presence, immediacy, and
concurrency. We theorize that these three characteristics of collaboration technology
will influence the intention to use it via the UTAUT cognitions of performance and
effort expectancies. These three characteristics are also expected to interact with the
task, which we discuss below.

Social Presence

Social presence theory argues that collaboration technologies differ in their ability
to convey the psychological impression of the physical presence of their users [64].
Collaboration technologies with high social presence convey a social and personal
environment for communication. Social presence is influenced by a technology's ability
to transmit nonword cues (e.g., voice inflection) and nonverbal cues (e.g., gestures,
facial expressions). Short et al. [64] argue that the greatest social presence is provided
by face-to-face communication, followed by technologies that provide both audio and
video communication, followed by those that provide only audio communication, and
least by those that provide only text communication. Social presence is an experiential
phenomenon in that it is possible for different users to perceive different levels of
social presence for a given technology (cf. [11]).
Use of collaboration technologies with low social presence can reduce effective?
ness, efficiency, and participant satisfaction because the collaboration technologies

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
20 BROWN, DENNIS, AND VENKATESH

can slow interaction and make communication more difficult [13, 33, 64]. Further,
with the collaboration technology acting as an interface between people, the greater
the social presence it exhibits, the more useful the technology is often seen [36].
Prior research has demonstrated the positive relationship between social presence and
usefulness (performance expectancy) [39]. Although there is no empirical evidence to
suggest that high social presence will affect effort expectancy, Kock [42] argues that
collaboration technologies are more difficult for individuals to use because they are far
removed from our natural face-to-face communication tendencies. Thus, collaboration
technologies that are higher in social presence will come closer to mimicking natural
communication and should be easier to use [42]. So, participants are likely to perceive
that collaboration technologies with low social presence (e.g., text-only technologies,
such as e-mail) have lower performance and effort expectancies than do technologies
with higher social presence (e.g., videoconferencing, telepresence) because of the
limitations that lower social presence spawns. Thus, we hypothesize:

Hypothesis 2a: Social presence will positively influence performance expectancy.

Hypothesis 2b: Social presence will positively influence effort expectancy.

Immediacy of Communication

Immediacy of communication refers to the extent to which a collaboration technology

enables the user to quickly communicate with others [22, 61, 66]. The task closure
model of media selection argues that people choose to use collaboration technologies
based on the ability to reach their communication partner and complete the task at
hand [66]. Although face-to-face meetings or telephone conversations may have greater
social presence, they require synchronous communication?that is, both parties must
be available at the same time [61]. Leaner technologies, such as voice mail and e-mail,
offer the ability to communicate asynchronously so that even if parties are not readily
available, communication may occur and may often prove a faster way to complete a
task rather than attempting to find a shared time to communicate [58].
As with social presence, immediacy is socially experienced. Immediacy depends on
capabilities inherent in the technology itself (it must be capable of immediacy) and also
on the way it is used. Although an e-mail may reach an individual's mailbox almost
instantaneously, the frequency with which he or she reads e-mail and the length of time
he or she chooses to take before responding are characteristics of use and not inherent
in the technology. Immediacy of communication is an important factor in the choice to
use a collaboration technology [66]. Technologies with higher immediacy capability
will be perceived to be more effective and efficient and, thus, will be perceived to have
greater performance and effort expectancies. Thus, we hypothesize:

Hypothesis 2c: Immediacy of communication will positively influence perfor?

mance expectancy.

Hypothesis 2d: Immediacy of communication will positively influence effort

expectancy.

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
 PREDICTING COLLABORATION TECHNOLOGY USE 21

Concurrency

Concurrency is the ability of a collaboration technology to enable an individual to

perform other tasks at the same time as using the technology. For example, one can
simultaneously engage in multiple separate "chat" sessions or chat while also us?
ing e-mail, talking on the telephone, or doing other work [59, 78]. Although truly
simultaneous work is probably impossible without some interference between the
tasks [55, 79], concurrent work in which the user focuses his or her attention on one
task for a few seconds and shifts to focus on another task for a few seconds and so on
is possible under some circumstances, thus leading to the performance of multiple
activities concurrently.
As with immediacy, concurrency is both a social and a technological capability?
that is, the technology must have the capability to support concurrent use and the user
must have the skills and desire to use it concurrently with other work. Further, the
social norms of the user's environment must permit concurrent use or the user must
be prepared to flaunt the norms. As with our arguments for social presence, the abil?
ity to work concurrently should be reflected in favorable assessments of performance
and effort expectancies given that enabling concurrency, when needed, should lead
to greater effectiveness, efficiency, and satisfaction, which should contribute to better
performance and effort expectancy cognitions. Thus, we hypothesize:

Hypothesis 2e: Concurrency will positively influence performance expectancy.

Hypothesis 2f: Concurrency will positively influence effort expectancy.

Individual and Group Characteristics

Individual and group characteristics are potentially important to the successful use
of collaboration technologies because different individuals and groups have different
needs [24, 25]. We focus on three specific factors that are likely to have the great?
est effect on intention to use collaboration technologies. They are two individual
characteristics?technology experience and self-efficacy?and one group-oriented
characteristic?familiarity with communication partners. Demographic factors, such
as age and gender, will also be important in understanding intention and use of a
collaboration technology [23]; as noted earlier, these are included as moderators in
UTAUT.

Individual Characteristics

Technology experience, the ability to use a specific type of technology, can play a role
in the selection and use of a technology, and in one's perceptions of the technology
[11, 16, 58]. When an individual first begins to use a new collaboration technology,
performance and satisfaction often decrease because its use requires new skills and
new patterns of interaction [19]. However, an individual will bring to bear his or her
experience from other related technologies; this mechanism is termed anchoring in

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
22 BROWN, DENNIS, AND VENKATESH

the psychology literature [69]. Over time, an individual's experience with the specific
technology will grow and it will gradually become easier to use, and performance will
also improve [19]. Thus, we hypothesize:

Hypothesis 3a: Collaboration technology experience will positively influence

performance expectancy.

Hypothesis 3b: Collaboration technology experience will positively influence

effort expectancy.

Computer self-efficacy, an individual's belief in his or her ability to use technol?

ogy to accomplish a task [14, 15], can also affect users' perceptions of performance
and effort expectancies [14, 69, 72]. Although computer self-efficacy is not specific
to collaboration technology, there is empirical evidence that shows that individuals
with greater computer self-efficacy perceive technologies to be easier to use [69,
72]. Thus, computer self-efficacy will positively influence effort expectancy [76].
Similarly, there is empirical evidence to suggest that computer self-efficacy has an
influence on performance expectancy. Compeau and Higgins [14] demonstrated that
computer self-efficacy had a positive effect on outcome expectations, a construct
similar to performance expectancy [76], because as an individual's perceptions of
his or her ability to use a technology increase, task performance also increases. Thus,
we hypothesize:

Hypothesis 3c: Computer self-efficacy will positively influence performance

expectancy.

Hypothesis 3d: Computer self-efficacy will positively influence effort expectancy.

Familiarity with Communication Partners

As individuals work together, they gradually develop an understanding of each other

and jointly develop a set of norms and expectations around the use of collaboration
technology [11, 19, 27]. Such shared norms reduce uncertainty and enable groups to
more quickly focus on the task at hand without needing to negotiate roles and expecta?
tions [49]. The development of this familiarity and shared norms enables groups to use
technologies more efficiently and effectively. Participants familiar with each other are
more likely to be able to use even lean technology to communicate rich messages than
those who lack familiarity with each other [11]. Thus, we expect that, as familiarity
with others increases, the performance and effort expectancies associated with using
collaboration technology increase. Thus, we hypothesize:

Hypothesis 3e: Familiarity with communication partners will positively influence

performance expectancy.

Hypothesis 3f: Familiarity with communication partners will positively influence

effort expectancy.

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC:34:56 UTC
All use subject to https://about.jstor.org/terms
 PREDICTING COLLABORATION TECHNOLOGY USE 23

Task Characteristics

Task has long been recognized as an important factor influencing performance [24,
25, 26, 31, 32, 81]. There are many ways in which we can examine and describe
tasks [25, 81]. Most research examining tasks has focused on specific tasks or spe?
cific task characteristics (e.g., equivocality, analyzability, complexity) depending
on the theoretical lens or collaboration technology under study. Following Dennis
et al. [24], we examine two types of tasks commonly performed with collaboration
technologies?idea generation/conferencing and decision making (see also [4]).
Idea-generation tasks are additive tasks in that the outputs of individual group mem?
bers are aggregated to form the group output; multiple, divergent results are desired
and the group need not come to consensus on one "correct" outcome. With decision
making, group members must work together to develop a shared understanding of
the issues and select among possible actions to choose one or more. While divergent
opinions may be useful as intermediate products, the ultimate outcome requires the
group to agree on a course of action. We posit that task plays an important role as a
moderator of the technology characteristics to performance expectancy relationship.
Generally, it is important for the technology to be appropriate for the task for which
it is used [24, 81]. Thus, when a task fits better with certain technology characteris?
tics, we expect that those technology characteristics will have a stronger influence
on performance expectancy.
Social presence is most important for task activities requiring high personal inter?
action [13, 33, 64]. Social presence is typically not important for activities that are
primarily information-processing activities requiring little interaction and feedback
[13, 33, 64]. Idea-generation tasks are primarily conveyance processes in which
group members provide information to others [22]; although group members need
to interact with each other, the group does not need to reach a shared consensus. In
contrast, decision-making tasks have a greater need for convergence processes in
which group members must understand each other and reach shared agreement [22].
Because decision-making tasks require group members to come to consensus and
engage in more interaction than idea-generation tasks, social presence will be more
important for decision-making tasks than for idea-generation tasks [24, 62]. Thus,
we hypothesize:

Hypothesis 4a: The effect of social presence on performance expectancy will be

moderated by tasks such that social presence will be more important for decision
making tasks.

As discussed earlier, immediacy will have a positive influence on performance

expectancy. As with social presence, we expect that this relationship will be stronger
for tasks that require interaction. Decision-making tasks, where all group members
must come to an agreement on a course of action(s), will have a greater need for im?
mediacy than will idea-generation tasks that are additive and do not require the group
to come to agreement, because members must converge on a shared understanding

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
24 BROWN, DENNIS, AND VENKATESH

and greater immediacy is more conducive to the development of shared understanding

[22, 24, 62]. Thus, we hypothesize:

Hypothesis 4b: The effect of immediacy of communication on performance ex?

pectancy will be moderated by tasks such that immediacy of communication will
be more important for decision-making tasks.

As with social presence and immediacy, concurrency is likely to be of greater value

for some tasks. Counter to our arguments for social presence and immediacy, we expect
that concurrency is less important for decision making, which requires the develop?
ment of group agreement, than for idea generation, which does not. Decision making,
and the development of group agreement, is best performed with technologies that
promote synchronicity, a shared pattern of coordinated synchronous behavior with a
common focus [22]. Concurrency inhibits the development of synchronicity because
it enables group members to work on different tasks simultaneously. Technologies
that inhibit concurrency are more likely to induce members to work together with a
common focus [22]. This suggests that technologies with higher concurrency will be
perceived to have fewer performance benefits for decision-making tasks. Thus, we
hypothesize:

Hypothesis 4c: The effect of concurrency on performance expectancy will be

moderated by tasks such that concurrency will be less important for decision
making tasks.

Situational Characteristics

Situational characteristics represent the context in which the collaboration technol?

ogy is implemented [25]. A variety of factors comprise the context [3, 4, 25, 56]. We
focus on co-worker factors and organizational environment factors. Co-worker factors
are the influence of peers and superiors. The influence of these important people in
the organizational context can directly affect the social influence, which ultimately
influences intention to use a collaboration technology [4, 67]. We expect peers and
superiors to be the key influences on the overall perception of social influence?when
peers and co-workers believe an individual should use the system, he or she will be
more likely to do so. Thus, we hypothesize:

Hypothesis 5a: The influence of peers will positively influence the perception of
social influence.

Hypothesis 5b: The influence of superiors will positively influence the perception
of social influence.

Other situational characteristics are experienced at the organizational level, such

as incentives, organizational culture, and the degree to which technology use is en?
couraged [4]. Consider, for example, an agile and innovative organization versus a
less innovative organization. The situational characteristics in these organizations

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
 PREDICTING COLLABORATION TECHNOLOGY USE 25

are likely to be markedly different when it comes to implementing and using tech?
nology. Likewise, organizations in which employees are rewarded for technology
use are likely to be different from those in which there are few incentives [3, 4, 56].
Two important aspects of the environment are resource- and technology-facilitating
conditions. Facilitating conditions, in the context of technology adoption, refers to
the extent to which various situational factors enable adoption and use of the system
[76]. Resource-facilitating conditions are the availability of money and infrastructure,
whereas technology-facilitating conditions relate to technical compatibility issues
[67]. As demonstrated by Taylor and Todd [67], these two components are expected
to contribute to perceptions of facilitating conditions in that as the resources and tech?
nology available to support system use increase, so will the perception of facilitating
conditions. Thus, we hypothesize:

Hypothesis 5c: Resource-facilitating conditions will positively influence the

perception of facilitating conditions.

Hypothesis 5d: Technology-facilitating conditions will positively influence the

perception of facilitating conditions.

System Use
System use was included in our model as the ultimate dependent variable for the sake
of completeness and also because use is typically measured objectively [76], thus
serving as a meaningful variable to assess criterion validity. The predictors of system
use have been well established in prior research, and the theoretical logic underlying
these hypotheses has also been extensively discussed in much prior research [76].
Specifically, UTAUT posits that there is a positive direct effect of behavioral intention
on use. Thus, we hypothesize:

Hypothesis 6: Behavioral intention will positively influence use.

Summary
Figure 2 presents our research model. In addition to contextualizing UTAUT to col?
laboration technologies, we present determinants of the four key UTAUT predictors?
performance expectancy, effort expectancy, social influence, and facilitating conditions.
Performance expectancy and effort expectancy are influenced by technology char?
acteristics (social presence, immediacy, and concurrency) and individual and group
characteristics (technology experience, computer self-efficacy, and familiarity with
communication partners). Task characteristics are expected to moderate the relation?
ship between technology characteristics and performance expectancy. The situational
variables attributed to co-workers (influence of peers and superiors) are expected to
influence social influence while the situational variables attributed to the environment

are expected to influence facilitating conditions.

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
 26 BROWN, DENNIS, AND VENKATESH

Method
We conducted two studies. The objective of the first study was to test our model in the
context of a general collaboration tool used to support day-to-day communication?
short message service (SMS). The first study did not incorporate a specific task and
the data were cross-sectional. The second study was conducted to complement the
first study, with the objective of testing the model in the context of an organizational
implementation of a collaboration technology that allowed for an examination of task
differences. Further, we collected use data six months after we collected data about
perceptions and intentions.

Measures
The survey instrument used previously validated measures where available. The items
used in Study 2, the organizational study, are shown in the Appendix, with similar
adapted items being used in Study 1. The constructs of intention to use, performance
expectancy, effort expectancy, social influence, and facilitating conditions were mea?
sured using scales adapted from Davis et al. [17] and Venkatesh et al. [76]. Our measure
of use was adapted and extended from prior research on technology adoption [17,
76] that used four items examining intensity, frequency, duration, and choice. These
items were reflective indicators of the latent variable of use. The choice item is new
and measured the percentage of time the individual chose to use SMS relative to the
overall need. We felt that different employees may need to collaborate to a different
extent, thus potentially constraining how much they would choose to use a collabora?
tion tool. By considering choice as part of use, we account for an important aspect of
an individual's collaboration in the context of work that is not typically considered in
measures of use, be it self-reported or actual [76]. By paying attention to all aspects
of the behavior?that is, the act of using and the choice to use?we are enhancing the
content validity of the measure relative to previous measures of use [38].
The scales for technology characteristics, individual and group characteristics,
and situational characteristics were adapted from prior research where possible.
The measures for social presence and familiarity with others (communication part?
ners) were adapted from Short et al. [64] and Carlson and Zmud [11], respectively.
Resource-facilitating conditions, technology-facilitating conditions, peer influence,
and superior influence were adapted from prior work [67,76]. Computer self-efficacy
was measured using a scale adapted from Compeau and Higgins [15]. The effect of
task was assessed only in the second study by asking participants to answer the survey
questions one time considering an idea-generation task and one time considering a
decision-making task. The order of the tasks was randomized and no significant order
effect was found. Gender and age were measured using single items. No scales were
available to measure experience, immediacy, and concurrency. We created the scales
for these three constructs using standard procedures of scale development [28]. We
created several candidate items, which we carefully examined for content validity.
These items were circulated for peer feedback and card sorts in order to arrive at the
final set of items, which possessed face validity and content validity.

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
 PREDICTING COLLABORATION TECHNOLOGY USE 27

Pretests and Pilot Study

Two pretests of the instrument were conducted to ensure that the measures were ap?
plicable in the current context. First, ten individuals (in two groups of five) affiliated
with the university were recruited to participate in this pretest. Each individual was
asked to complete the questionnaire and then provided the opportunity to comment on
any aspect of the questionnaire. The primary feedback from the first group was with
regard to the use of some "complex" English words/terms in the questions. Based on
this feedback, a few questions were slightly reworded. The updated questionnaire was
then validated with the second group of pretest participants and feedback solicited. No
significant suggestions were made and, thus, no further changes were made.
The revised survey was administered among 111 undergraduate students. The fo?
cus of the pilot study was to examine the reliability and validity of the scales in the
context of a collaboration technology?here, SMS. We were particularly interested
in establishing the reliability and validity of the new scales?immediacy and concur?
rency. The new scales were found to be reliable, with Cronbach's alpha exceeding
0.80. The other scales were also highly reliable, with similar Cronbach's alpha scores.
Next, a principal components analysis with varimax rotation was conducted among
the multi-item constructs from collaboration technology research?that is, social pres?
ence, media richness, immediacy, concurrency, and familiarity with others. A clean
factor structure was obtained, with loadings greater than 0.70 and cross-loadings less
than 0.35, thus supporting internal consistency and discriminant validity. A similar
analysis was conducted among the UTAUT predictors and a clean factor structure
was obtained there as well.
Given the total number of items from all multi-item constructs in the model, the
sample size in this pilot study was not sufficient to test internal consistency and dis?
criminant validity of all constructs in a single test using exploratory factor analysis.
However, this concern is somewhat alleviated for three reasons: (1) the new scales
were developed in the context of collaboration technology research and the likelihood
of overlap was more with the constructs in that domain, (2) the technology adoption
constructs and collaboration technology constructs come from very different bodies of
research where there has been minimal conceptual overlap thus far, and (3) the entire
model and scales will be validated in the actual data set using confirmatory factor
analysis in partial least squares (PLS).

Study 1
Setting and Target System

Our first study examined users of one emerging collaboration technology?SMS.

SMS is primarily a tool/service that allows two-way, near real-time communication
among people via mobile phones or computers [29].3 SMS enables users to send short
text messages that are displayed almost instantly on the target user's device. SMS is
similar to other collaboration technologies, such as instant messaging and Web confer?
encing, in that it is real time but can also be used to leave messages for absent users.

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
 28 BROWN, DENNIS, AND VENKATESH

SMS use is less intrusive than phone conversations because SMS message exchanges
can be silent and less distracting than e-mail [29]. One interesting social convention
around SMS use is that, in many cases, it is socially acceptable?or even expected?
to receive and send SMS messages while performing other activities, such as being
in meetings [30, 59]. Asia leads the way in the number of SMS messages exchanged,
with Europe coming in a close second. SMS use in North America is substantially
lower, but it is rapidly increasing now that technical issues of interoperability have
been resolved [60].

Participants

We surveyed voluntary users of SMS in Finland. Finland has a high penetration of

mobile phones and maturity of SMS use [2]. Data were collected during a one-week
period through an active solicitation process from individuals, primarily alumni, as?
sociated with a major university in Finland. The goal was to identify a broad range
of users from a wide variety of organizations to provide variance in terms of gender,
age, education, technology experience, and SMS experience. Of the 500 paper copies
distributed, 363 responses were received, and of these 349 were usable, resulting in
a return rate of almost 73 percent and a usable response rate of almost 70 percent.4
Of the 349 participants, 36 percent were women, and the average age of participants
was 34.3 (standard deviation [SD] = 9.01).

Data Collection Procedure

As noted earlier, we sought to identify participants such that they were representative
of the population of SMS users. We worked in collaboration with a leading university
in Finland and solicited participation from a list of individuals provided by the users.
Potential participants were provided with paper copies of the survey and asked to re?
turn the completed survey in one week to a specific individual who was coordinating
the administration of the survey. Due to privacy concerns, the university did not share
information about the participants who chose to complete the study (or who declined to
participate) or date of response. We were, therefore, unable to compare early and late
respondents or nonrespondents. As the responses were received within a week, which
is a fairly short amount of time, this issue is somewhat alleviated. Further, response
biases are somewhat alleviated given the high response rate. Overall, we deemed this
trade-off acceptable in order to collect real-world data.

Study 2
Setting and Target System

Our second study was conducted in a Fortune 500 technology company in Finland.
The company has a traditional, hierarchical structure and was organized as several
business units in two different geographic locations in Finland. The target system

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
 PREDICTING COLLABORATION TECHNOLOGY USE 29

was a collaboration technology that was developed in-house. The system design and
development process took about eight months and included employees at different
organizational levels as part of the design team. The objective was to provide an ad?
ditional option for employee collaboration beyond traditional options, such as tele?
phone, videoconferencing, and desktop messaging. As the employees were primarily
working in technology design, coding, testing, and related areas, collaboration was
an important aspect of their day-to-day work. They needed to collaborate with peers
and group members in the same location and at other organizational locations in
Finland, other parts of Europe, and the United States. Use of the system at the time
of the study was voluntary.
The system provided features to chat, conduct an audioconference, conduct a vid
eoconference, have a shared whiteboard, save meeting notes in multimedia format,
and use some of the functionalities of other organizational applications (application
exposure). The last feature was particularly important relative to an off-the-shelf
commercial tool (e.g., MSN messenger) as the organization had several unique
applications?some of which were developed in-house and some of which were pur?
chased from commercial vendors?that supported the work of the employees. Use
was not mandated by the organization. The initial eight-month period after the beta
testing was completed was designated as the trial period for the system, after which
the organization would make a decision regarding mandating system use.

Participants

The population of interest was knowledge workers. Our sampling frame was all
knowledge workers in a business unit in the firm where we were collecting data. As
the firm decided to follow a phased implementation plan, we were restricted to one
business unit for our data collection. There were 883 employees in that business unit
who were classified as knowledge workers. Of these, 830 agreed to participate and
participated in the initial survey, and 447 of them provided responses to the second
survey, which collected use data. This resulted in an effective final response rate just
under 51 percent relative to the entire sample and just under 54 percent relative to the
initial survey.5 Of the 830 participants, 227 were women (27.4 percent); 125 of these
women responded to the follow-up survey, resulting in about 28 percent of the final
sample being women. The average age of the participants was 33.8 (SD = 9.94) and
34.6 (SD = 10.41) in the initial and follow-up surveys, respectively. The key demo?
graphic characteristics were comparable across the two surveys. Thus, the threat of
nonresponse bias was diminished.

Data Collection Procedure

As noted earlier, the data were collected in conjunction with the rollout of a new col?
laboration tool in a business unit in a company in Finland. A one-day training class
was provided to employees, staggered over a period of three months, to accommodate
the entire business unit. The training was provided by the in-house IT group and each

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
30 BROWN, DENNIS, AND VENKATESH

training group comprised a primary instructor who conducted all the training sessions
and two technical assistants who were different across different training sessions.
These assistants only provided technical help when someone was stuck or when
someone had a procedural question. The training discussed the various features of
the collaboration tool, including features that enabled application and data exposure
across collaborators. The training included an opportunity for the participants to try
the system. Immediately after the training was completed, the employees responded
to a survey administered by the organization to gather feedback regarding the training
and the system. The perceptual data were collected in conjunction with the organiza?
tionally administered survey. We solicited participants' contact information in order
to follow up with the employees regarding their use of the system. Because use could
not be measured on the survey administered immediately after the training, a follow
up survey to measure use was conducted six months after the initial survey. As the
original training was staggered over three months, the follow-up was correspondingly
staggered. Data were collected via e-mail and phone calls to the various participants.
Up to six e-mails and six phone calls were attempted to contact the respondents over
a two-week period.
The employees of the company were generally quite proficient in English. However,
as both Finnish and Swedish are official languages of Finland, the firm suggested that
the respondents be offered those language options as well to fill out the questionnaire.
This was consistent with the organization's policy when it came to availability of vari?
ous policies and forms. Professional translators translated the instructions and questions
from English to Finnish and Swedish. Their translation procedure included a translation
back to English by a different translator, and no discrepancies were found.

Results
PLS Graph version 3, Build 1126, was used to analyze the data from both studies.
Testing interaction effects was possible in PLS. All constructs were modeled using
reflective indicators. Interaction terms were created using data at the indicator level
after the data were centered to minimize threats of multicollinearity [1]. We tested the
model shown in Figure 2?one difference is that task was examined only in Study 2.
In Study 2, repeated measures were treated as separate sample cases, which would be
problematic with ordinary least squares (OLS) regression, because this violates the
OLS assumption that the sample cases are independent [1]. However, with PLS "no
assumptions are made regarding the joint distribution of the indicators or the indepen?
dence of sample cases" [12, p. 332]. As a result, use of PLS is appropriate here.6
The measurement model results from both studies supported reliability and validity.
In both studies, the factor loadings and cross-loadings supported discriminant validity,
with loadings greater than 0.70 and cross-loadings lower than 0.30. Also, the internal
consistency reliability (ICR) of all the constructs was greater than 0.75, thus confirming
that the scales were reliable in both studies. Finally, in both studies, the average variance
extracted (AVE) for each construct modeled using reflective indicators was in excess
of 0.70 and the square root of the AVE for each construct exceeded all interconstruct

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
 PREDICTING COLLABORATION TECHNOLOGY USE 31

correlations. The descriptive statistics, ICRs, AVEs, and correlations are shown in
Tables la and lb for Studies 1 and 2, respectively. Most construct means were a little
over 4, with a standard deviation over 1. Most correlations were significant. UTAUT
constructs were more highly correlated with intention to use the system than they
were with collaboration constructs. Also, as expected, the collaboration constructs
were correlated with the various UTAUT constructs.
Due to the nature of the data collection, we tested for common method bias using
Harman's one-factor test [57]. If a substantial amount of common method variance
(CMV) exists, either a single factor will emerge from the factor analysis or a single,
general factor will account for the majority of the covariance in the independent and
dependent variables [57]. The single factor accounted for 24 percent of the variance
and did not account for the majority of the covariance, thus suggesting that common
method bias is not a concern in our data set.
To further alleviate concerns about common method bias, we employed the marker
variable technique [45, 46] and tested the hypotheses based on the corrected correla?
tions. Specifically, we chose the second-smallest positive correlation among the con?
structs as a conservative estimate of CMV to produce the CMV-adjusted correlation
matrix [45]. Following Malhotra et al. [46], we produced a CMV-adjusted correlation
matrix and then used it to estimate CMV-adjusted path coefficients and explained vari?
ance. The results show that after controlling for CMV effects, the explained variances
do indeed decrease, but the drop is not substantial and is just over 10 percent. The path
coefficients are consistent with those that were found without the CMV adjustment.
We conclude that concerns about common method bias are alleviated.
Prior to our model tests, consistent with the recommendation of Aiken and West [1],
we mean-centered the variables that were part of interaction terms. All variance in?
flation factors (VIFs) in our structural model tests were less than 5, thus alleviating
concerns about multicollinearity. Tables 2a and 2b show the results of our structural
model tests for both studies. The successful prediction of collaboration technology use
in these two studies provides criterion validity and is important given that many of the
constructs in this work were perceptual constructs. The results related to prediction
of behavioral intention are consistent with the UTAUT hypotheses that were adapted
to this context, thus supporting Hypotheses la, lb, and lc. Further, the results are
consistent with the key predictors of use in UTAUT, thus supporting Hypotheses 6
and Id. Specifically, behavioral intention had a positive, significant influence on use
(H6), and the effect of facilitating conditions on use was moderated by age and experi?
ence in both studies (Hid).
In Study 1, technology characteristics and effort expectancy7 predicted performance
expectancy, thus supporting Hypotheses 2a, 2c, and 2e. In Study 2, only immediacy
had a main effect on performance expectancy (regardless of task), thus supporting
Hypothesis 2c. All the technology characteristics were moderated by task and consis?
tent with the predictions of Hypotheses 4a and 4b, but were in the opposite direction
of Hypothesis 4c. Contrary to Hypothesis 4c, concurrency had a stronger effect on
performance expectancy for decision-making tasks. In both studies, effort expectancy
was predicted by all of the technology and individual and group characteristics,

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
32

ooooooooooo

oqcgi-T-i-i-i-cMOOi--^
????????????

COC\|t-OCV100-!-!-i-t-C\](MOO
0000000000000

ooooooooooooo

ooooooooo o o o o

00^t-(D^C0P)CV10)WU)NO^t0)i
COCOCMt-WCOCQCMt-Wt-t-i-i-t-CM
????O?????OO?O??

TfCDi-C0Oi-l0C0<J>LnCDCvJC005^-ir>C0
OO^J-CO^J-COCMCgT-CNCMCvlCVjT-t-OCgCM
?dddddddddddd?dd?

oooooooooooo o o o o

t-t-i-i-t-t-^t-Ot-O-i CD O i- 1- t

mw(DO)iflcooo)coi-oo) s m m n
0q05CD00l0CVJCT>O>0>l0C0i-<i-N-0>l0O
CO

oooooooooooo o o o o

Li. h

LU
CL
lu co 5:
>> ? to ? 13 co R?
CO
O CD^ o ^ cd r; ? CO? CD
o ?cdco ? cd co .E
q t_
cd ^ co .2
_ c
o
w
~ co X =?CO s?=
CL CD
cd eg
CD T3 CD
? _ q.
x 9t _? ^
c8?-1.co
1.?
CD 0 > C
8 o 55
2 O) >, CO ? CD 2
co g
-?
I* -
o .2 E cd ?
>. o t !5 ? CO
CO CO
I ? 3 o
u. CD O _c
O CD cd it: O "?
= x: too "o
03 O ? E E c O)
CD CD Q. W "h
CD 3 CD 2
O CQ Q_ 111 CO LL CO -E ? < q_ co cr f

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
 33

u
PL,
H

U
tu

?1
oo

o o o o

<

o o o o o
I I

Oh in CD O (O CD ^ CO
00 CO i- i- o o
? ? ? ? ? ? ?

00
U <? ? <? ? ? d

H o o o o o o
^ q
t3 1

c
o
o o o o o o o o o
U
a v
? ft,
II

>, ? I
5 o

LU
^-v LU 0T
2.1
O -t=
CL W CO LL ^ -E
o t? a. -a ?
O OQ
? cd
u o ?
~ " ??CO8
?
CO
o CD O) .?
Q.
CD ^ 0 ? ^ CD tO j=
X ^ ? ^
8 i II
?
+=: CO
CO o C c
? ?>,>,cd
w >cd
cd "5 c
c- ? i
00 2 c o> ^ >> CD
?2
suit
o o X
<D C
Z CO a = ?= ? CO E
C > 3 O
t: ? ??
?g 5 .
O CD CD 5= O
i_ (-j i_ j- CD CD
O m O c0 CD CO CD a op
O CQ0- LU CO O O LL o < CO LT

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
34

mcowooifimnincDON
cocvjcjcg-i-T-T-i-ooi-o

???????c^????
oooooooo o o o o

o o o o o o o
? ?o o o o o

l/)^CO^OCDO)rj-lflO)(Dc\l
COMOT-CjT-T-CVICMCMO't^T-i
???????????????

oooooooooo o o o o

oooooooooo p o o o o

(DOO^Or-WOOCOOJCOOinCOl^i-COM
CO^CT3^C0CNJCNi-(NC^C^C\|i-t-t-i-t-CN
??d????????o'op????

NCO^CMlfiQOm
ooin^cococvioj'r
????????????
<? ? ? o o o o

CMCOi-CMr-COOJO^W 1^ ^
t-OOCVJ^C00005x O) o <
T^l^T-^T^T^T^T^T^Ol-1 d -r^ z O T- T- T

NOOCOCOOCOOCOOlOOO O T- CO ^j- o
Ttin\foqo)i-oo(00)Ttr-T < CO <; O CM Tt in
CO

00COCOCOC0C000NNNCO00<<<00C000N
????????????^-^^-????

LH 8.1
Q.
o
111
LU CO
b o
^?1 ?
=1?
CO ^ <=! o
?30 =.2 CO
C Q_ ?
_ c .2 8 ~Li 8(D c(0 =I
la
Cl
CD "O CD
I 0 ?
r=cooc c X 2=
I s f8 'i^ 11
^ 0 q5
CO >, co ? ^
03 2
o .2
-D > fc CD
_ Q
??? OB
O 3 CO CD
-2 ?
1 3
"C $5 ^ CO 0 o c Cl
_C0 03 otoo
1'w 0O) m
II
_ _ w O
o o E c %0 0
"5 q 0 3= O 03 O c o co 3 0q. co
? CD CL HI CO LL CO JE ? LL O < F CL CO DC

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
 35

CM
00 CVJ
d d

co o
00 T- o I
odd

to ^
oo o
d d

o o o o I

o o o o

oooooooo

O 00 O 00
W ? CM r (M t
??????\

<? ? 0 ?

O O O O Q o o o o I

o2
LU ? .2
O AS
a. lu nT - ~ T5
lu CO Li
? to ?
?"5? ^ ! 0 x; J2 ;??
gu? ' o CO

II X 9. ?
CD CP
0
Q_
CO
? ? ? 8.
1 CO .

! _co ._

?2 ?? * 8 ? S
^ CO 1
SS >
co 2
o .2 t ? _
2
o_ bE
-Q >
o tl
_cg co too
"co
55o
3 2 1.1 * 11
- CO o
?(2?.2
o 0 0 *= o o 0 D
? CD d. LU COCO Q_ CO

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
36

o
CM

CO LO
CO
?

o co ^f- o cmNONCMCD't^CM
in o o i o o Ot-OOCMOOO
???O?????

CO o
CO o o
? d d d

LU Ql
LT X
>> O LU
? o cd
c5 LU co O
c c
c cd
o o
cd LU
c "5 o_ CD
0 o5 g ?D?
0 X <
?8 cd
o ?3 ? Q
c cd
LU X
"03
i_

g E <d ?
c ? cnDCujOCCOCLUCL
Q CDQ Q CD X
O <
's too
o "C .55
CD CD < UJ

11
cd o
X X X X
CM Q cd 5= O LU LU LU LU
O) CD LU LU
Q_ LU CO <FQ-(L ?_ LU LU LU

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
 37

* *
CO 00
T- CM
? d

to o
CO CO

* * * * *
i- CO O ID CM
CM CM CM T- CM
d d d d d

* *
* * *
CO 1- CD 00 ^
i? CO t? o o
d d d d d

CO^tCMt-CO^J-f-O)
OOOCOOOOCM
????????

o CM
d d

CO
c
o

c
o
q_ o
X X O)
lu lu ? o c
q_ x x
X lu lu q_
?4?' m?

lu O
Sm
< xm >> CO U
o
o

x x 5 . ? > w c
i
?- cl >^
X X lu q_ nr lu
U)
lu ?j
x x CD X Q CD -
<? CD<
6=58DI? t2 5.E=I c
o
o
dc lu dc x .2 E? ?c =E ?f
o o^
XI
* * c*O COo O?CD
o
q CD q lu O .?
CD < CD lu CO co c/3 c/3 u_

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
38

1^
?

o
C\J

CO
CO o

CO O) CM N C0NC0LOCAJ(DO)(DNW^
in OOO OOOOOOCMOOOO
? odd ???????????

CO
CO o o

lu QL
cl X
? LU
o CD
CO c fr CO O
c
o s?
o -
o
c

CD
X y CD t5 GL
? S g ?6? X
CD
|85C o Q

| SLjj O) LT LU GL" LU X
LU
E cd ? co CD O Q O CD CD <
< g CD < X

Si
o r J
too
CD 3= O
^O)sz xx
S LU LU
X X X X LT
LU LU LUQLU
0_ LU CO < F a. ?_ CL LU LU LU CD

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
 39

* *
CD CO
i- CO
? ?

CO i
00 CO

1- O) CM CO CM
CM CM CM i- CM
odd

* * *
* * * *

CD 00 CO CD N- CD CO CO
Ot-OO CM CM CO O O
dodo ? ? ? ? ?

N ^ O CO CM ^ N
O O CO O O O CM
0 0 0 0 0 0 0

o
d d

Q ?

Q co
CO CO
S Si
=3 ?
CO CD
Q_ CD > "g ?
X 2 > CD
2 ? ??D)|
?,
LU
a s CD
X LU X "55
^ CO
S
CO
2~ oiE
gCD ECO "
LU
X
q B X ? mL
3 ?i
CD ^ o =
CD ? CO >?
CO

? Si 8
CD
OC LU
Q CD Cl .?2 t -O CO ? CO ?c ? i O

$3
? "O Z3 CO 3 c O
CD < CO CD o ? o_ ? c
EES
i ? ii 6 ii CO CC F
UJ OC -C
CD Q 'o E ~
X X o
? E
<CDLUc7)COCOCOLLC7)i?F
CD

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
40 BROWN, DENNIS, AND VENKATESH

consistent with Hypotheses 2b, 2d, 2f, 3b, 3d, and 3f. Consistent with Hypotheses 5a
and 5b, peer influence and superior influence had a positive effect on social influence
in both studies. Also, consistent with Hypotheses 5c and 5d, facilitating conditions
were predicted by resource- and technology-facilitating conditions in both studies.
Although we focused on the overall model test and did not specifically theorize about
full or partial mediation, implicit in our model depiction is that the four predictors
in UTAUT?namely, performance expectancy, effort expectancy, social influence,
and facilitating conditions?will fully mediate the effect of the various collaboration
technology constructs on behavioral intention. Further, we expect behavioral inten?
tion to fully mediate the effects of performance expectancy, effort expectancy, and
social influence on technology use. In order to test for such full mediation, we used
the approach recommended by Baron and Kenny [5]. In addition to the results already
reported, we found that the various collaboration constructs had a similar effect on
intention as they did on the UTAUT predictors, and when the effects of the collabo?
ration technology constructs were included over and above the UTAUT predictors,
none of the collaboration constructs had an effect on behavioral intention. Likewise,
we found the UTAUT predictors had similar effects on technology use as they did on
behavioral intention. Further, when behavioral intention and facilitating conditions
(along with the moderators) were included as predictors of technology use, none of the
UTAUT predictors had an effect on technology use. Overall, these additional analyses
provide support for the mediation pattern shown in our model.

Discussion
The key objective of this paper was to develop and test a model to understand col?
laboration technology use. The model integrated UTAUT [76] with theories from
collaboration technology research [11, 16, 25, 64, 66]. The model was supported in
two studies examining two different collaboration technologies that used different
subject pools. The results from the two studies were similar. We found that UTAUT is
the conduit through which collaboration technology research constructs of technology,
individual/group, task, and situational characteristics influence behavioral intention
and use of that collaboration technology.

Contributions and Implications

This study makes several important contributions to IS research. Integrating two of the
most dominant streams of research in IS?that is, technology adoption and collabora?
tion technology?is a key contribution. Specifically, this work integrated UTAUT with
theories from collaboration research?social presence theory [64], channel expansion
theory [11] (a descendant of media richness theory [16]), and the task closure model
[66]. This study responds to a call for such work by recent articles that have provided
an in-depth analysis of the directions for future work in IS [6, 70]. An even more
general contribution, we hope, is that this paper serves as an example of integrative

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
 PREDICTING COLLABORATION TECHNOLOGY USE 41

research that ties together dominant streams and models of IS research and moves us
toward a cumulative tradition, a call that was issued at the first International Conference
on Information Systems (ICIS) in 1980 [41]. UTAUT and its generalizability have
not been in doubt, due to its foundation in TAM and related models of adoption and
use, and its robustness is furthered here by tying the UTAUT constructs to important
constructs from collaboration technology research. Moreover, we demonstrate that
UTAUT fully mediates the relationship between technology characteristics and use,
thus providing insights that could drive future research about the IT artifact and levers
influencing adoption and use of collaboration technologies.
This study complements previous models that use the general antecedents of per?
formance expectancy and effort expectancy (e.g., [40, 69, 73]) and provides evidence
that collaboration technology-specific factors play an important role in influencing
cognitions that drive technology use. First, the three collaboration technology char?
acteristics directly influenced performance expectancy and effort expectancy. More
interesting, perhaps, is that they also interacted with the task. In Study 1, higher social
presence, increased immediacy, and greater concurrency led to increased performance
expectancy and effort expectancy.8 In Study 2, which considered task interactions
for performance expectancy, we found that (1) higher social presence only increased
performance expectancy for decision-making tasks; (2) increased immediacy had
beneficial performance expectancy effects for both task types, but stronger effects for
decision-making tasks; and (3) counter to our hypotheses, greater concurrency led to
greater performance expectancy only for decision-making tasks.
We conclude that these three collaboration technology characteristics?social pres?
ence, immediacy, and concurrency?are important factors influencing the adoption
and use of collaboration technology. Social presence and immediacy have long been
linked to perceptions of performance and user satisfaction, particularly for decision
making tasks [13, 33, 64]. However, they have not been previously linked to the
decision to adopt or use a collaboration technology. This is an important contribution
of this work. More puzzling, perhaps, is the role played by concurrency?the ability
to perform multiple tasks. Concurrency is a newer construct that has received far less
attention. We have argued that decision-making tasks would benefit more than idea
generation from a shared focus of attention [22] and that concurrency would impede
the development of this shared focus. We found concurrency to have a direct effect
on effort expectancy but, counter to our hypotheses, to increase performance expec?
tancy for decision-making tasks and not idea-generation tasks. These results highlight
the important role of task and technology characteristics in research on technology
adoption and use.
Individual and group characteristics played an important role in influencing ef?
fort expectancy but, contrary to our hypotheses, not performance expectancy. Not
surprisingly, greater self-efficacy led to greater effort expectancy. However, it was
surprising that experience with collaboration technologies did not. It may be that the
use of collaboration technologies (e.g., audioconferencing, instant messaging, e-mail)
is relatively widespread and that once one gains even a little experience with them,
additional experience has little marginal benefit. Computer self-efficacy is a broader

This content downloaded from

159.224.180.78 on ffff:ffff on Thu, 01 Jan 1976 12:34:56 UTC
All use subject to https://about.jstor.org/terms
42 BROWN, DENNIS, AND VENKATESH

assessment of overall technology competency, especially as it incorporates the element

of confidence. UTAUT explicitly argues that computer self-efficacy will not influence
intention to use a technology [76]. Our research shows that computer self-efficacy
plays a role as an antecedent in influencing how different individuals perceive the ef?
fort expected of the same technology: individuals with greater computer self-efficacy
perceived both collaboration technologies to require less effort. Finally, group members
with greater familiarity with their communication partners perceived the collabora?
tion technology to require less effort. This is probably because they could rely on
well-established norms and deep understanding of each other so that communication
could be less explicit and express the same meaning with fewer words (cf. [11]). The
inclusion of individual and group characteristics in our model provides insights into
the antecedents of performance expectancy and effort expectancy in the context of
collaboration technologies. As an aside, we note that younger group members and
men were more likely to perceive collaboration technology to require less effort to
use, which is consistent with prior research with other technologies [51, 52].
The situational conditions surrounding collaboration technologies influenced inten?
tion and use. Co-workers, both peers and supervisors, worked through social influence
to affect intention to use. Both peer and supervisor opinion influenced intention?in
about equal proportion?so all potential communication partners exert an influence
on use. Environmental characteristics, such as facilitating conditions (moderated by
age, gender, and experience), also influenced intention to use. Technology-facilitating
conditions (compatibility with other technologies) had a greater effect than did
resource-facilitating conditions (time and money). The effect of technology compat?
ibility is a more immediate day-to-day consideration than the resources, which may
explain its greater effect in this context.
Overall, UTAUT proved effective in predicting intention to use. The three groups
of collaboration technology-specific antecedents?technology, task, and individual/
group characteristics?were significant antecedents influencing performance and ef?
fort expectancy. Performance expectancy, moderated by gender and age, and effort
expectancy, moderated by gender, age, and experience, had significant effects on the
intention to use. The consistency of findings across these two studies and technologies
contributes to the cumulative tradition and ongoing assessment of UTAUT.

Limitations

The strengths of this study are that it is a field study, conducted in two contexts,
with participants drawn from multiple organizations, using two different collabora?
tion technologies, with very different characteristics. However, our study has a few
limitations that should be noted. This study was conducted in Finland, a country at
the cutting edge in terms of technological sophistication. This raises a question re?
garding generalizability to other countries. This generalizability issue is even deeper
than just being an issue of external validity; it is possible that the sample studied here
represents mostly innovators and as the model is tested in countries that have less
technology-savvy populations, the pattern of findings or pertinent constructs may be

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
 PREDICTING COLLABORATION TECHNOLOGY USE 43

different. This calls for research to address the issue by an examination of the deeper
cross-cultural generalizability issues.
We examine two very different collaboration technologies?that is, SMS and a
proprietary collaboration tool?but they represent only a subset of the types of col?
laboration technologies available. Given the reasonably consistent results across the
two studies, we expect similar findings with other types of collaboration technologies.
However, the unique characteristics of blogs and wikis, for example, may suggest
refinements to the model. In addition, we characterized the collaboration tasks based
on their objectives?that is, idea generation and decision making. Other results may
become apparent if the focus turns to the nature of interaction with the tool (e.g., syn?
chronous versus asynchronous). This issue does raise an interesting question of just
how detailed research needs to be when theorizing about the IT artifact.
It is important to note that we analyzed user assessments of the technology character?
istics, thus relying on the socially derived characteristics rather than the innate physical
characteristics. By examining the socially derived characteristics, we were able to focus
on a single technology, yet still achieve variability in the assessments. It is conceivable
that an examination of the innate physical characteristics of the media would yield
different results. Thus, we encourage future research to examine the differential effects
of socially derived and innate physical characteristics of different media.
Finally, the model might have omitted constructs. Constraints imposed by the
research settings precluded us from having an excessively long survey instrument.
As a result, we had to necessarily scope our model to include key constructs and
keep the number of items per construct at a reasonable level. It is possible that other
collaboration constructs would offer alternative perspectives. However, this concern
is somewhat alleviated due to the strong ties to established theoretical perspectives
in our construct selection process. Thus, we call for future studies to consider other
constructs in the space of collaboration research to extend and augment the model
presented in this work.

Implications for Future Research

Several additional and important directions emerge from the findings in this work. Our
research begs the question about how use and its antecedents will unfold with even
greater experience. While we measured use and employed it as the ultimate dependent
variable in our model, the model and test here still represent only one point in time.
Future work with multiple waves of data is essential to deepen our understanding of
the antecedents of use as experience grows. This could predict trajectories of use based
on perceived characteristics of the collaboration technology or UTAUT constructs.
A number of opportunities exist for expanding our model. We selected three tech?
nology characteristics specific to collaboration technology that we believed were
important. Our results show them to be significant factors. However, there are many
other collaboration technology characteristics, such as rehearsability and synchronicity
[22], that may play a role. Future research needs to expand our model to investigate
other collaboration technology characteristics.

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
44 BROWN, DENNIS, AND VENKATESH

As we noted, task characteristics can be notoriously difficult, if not impossible, to

control in field studies, and therefore experimental research will be essential to further
investigation of the role of task. Moreover, it will be important to carefully consider
the nature of the task being studied. Incorporating aspects such as uncertainty or
interdependence could prove valuable in uncovering the nature of the role played by
task in influencing use.
Given the important moderating role of task, future research should clearly articulate
the focal task when trying to explain use. Prior work has demonstrated the importance
of the fit between task and technology, specifically in collaboration settings [24]. Prior
work has also shown that fit may be less important?or even unimportant?over time
[35]. Our results suggest that the nature of the task can alter the relationships between
the technology-specific characteristics and the UTAUT antecedents. So, although fit
may or may not be an important factor influencing ongoing performance (cf. [24, 35]),
fit does play a role in users' perceptions of performance expectancy and, ultimately,
their choice to use a collaboration technology. As additional technology-specific char?
acteristics are examined, future work should be mindful of the task and its potential
moderating effect on the choice to use a technology.
The range of situational characteristics that could be studied is expansive?some
examples include organizational culture, innovation culture, and voluntariness. While
some variance can indeed exist in perceptions about these situational characteristics
even within a single organization, we were limited by the practical constraints of
questionnaire length, and thus we chose to focus on certain constructs. The ideal ap?
proach to study situational characteristics will be to study collaboration technology
use in different organizations to gain variance in the situational characteristics?for
example, Venkatesh et al. [76] studied implementations in different organizations to
understand voluntary versus mandatory use situations.
We considered only one moderator of the relationships between collaboration con?
structs and UTAUT constructs?task. Considering additional moderators is an impor?
tant direction for future research. Given research on channel expansion theory [11], it
is possible that relationships among technology perceptions, technology experience,
and partner experience might have an important influence on use. This experience
evolves over time such that partners familiar with each other perceive lean media as
rich. This would argue for future research to examine these important interactions and
explore the effect of time and changing perceptions on our model.
Another next step will be to evaluate interventions related to the constructs in our
model and other commonly used interventions and their effect on the various constructs
studied here. For example, altering technology characteristics, such as social presence,
immediacy, and concurrency, or developing procedures to improve group member
familiarity, may prove valuable for enhancing collaboration technology use. It is also
important to consider that people often employ multiple collaboration technologies
to interact with the same communication partner(s). The use of one collaboration tool
may have important implications for the perceptions and use of another tool [48]. As
research moves forward, it will be important to consider the multimedia nature of
collaboration.

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
 PREDICTING COLLABORATION TECHNOLOGY USE 45

We conceptualized collaboration technology use quite broadly, and further research

is necessary to separate work use from leisure use and examine the antecedents of
each [71]. Recent research has shown that antecedents of these two types of use can
be quite different [70, 76] and the work versus leisure nature of communication is
likely to have an effect on collaboration technology use [8]. It will be important to
examine the generalizability of this model of general collaboration technology use to
different contexts while being sensitive to specific organizational context variables,
such as incentive systems and organizational culture, as recommended by Dennis et
al. [25].
Last but not least, given that collaboration is an inherently multiuser phenomenon,
we call for research to investigate collaboration technology use at the group level by
integrating both individual and group-level constructs in a multilevel model [10].
Group-level constructs are not simply aggregated individual values, but rather exist
at a separate level that the group as a whole co-creates (e.g., decision performance,
time, mindfulness). Future research needs to incorporate some aspect of "groupness"
that evolves from the interaction among members [63].

Implications for Practice

Much prior research on technology adoption has shown that performance expectancy
and effort expectancy are predictive of intention and use. While such predictive validity
is valuable, we argued that actionable guidance can better emerge by understanding
a set of technology-specific factors that influence these expectations. The support for
our collaboration technology model underscores the possibility that collaboration
technology use can be enhanced or eroded depending on the underlying technology,
individual/group, task, and situational characteristics. Thus, the results of this study
highlight the key levers that organizations can use to improve the adoption and use
of collaboration technologies.
One potential lever is training in the area of technology characteristics. Organiza?
tions will benefit by showing their employees how to exploit the characteristics of
various technologies. For example, across both studies, we found that immediacy is
an important antecedent to performance expectancy, regardless of task. Thus, it would
seem that educating employees about the speed of collaboration associated with the
use of various technologies could be valuable. There were noticeable differences in
the perceptions of immediacy within, as well as between, technologies, implying that
different respondents experience the same technology differently [11, 22]. Different
users are slower or faster in responding to messages, whether they are SMS mes?
sages or messages in a more full-featured collaboration system. Establishing group or
organizational norms for immediacy and educating users about them might influence
the adoption of collaboration technologies.
In addition, the significance of social presence suggests that providing employees
with examples of how to increase social presence would be valuable (e.g., via language
and emoticons). Further, when introducing collaboration technologies, organizations
will benefit by providing opportunities for employees to work with both the technology

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
46 BROWN, DENNIS, AND VENKATESH

and a set of familiar communication partners in order to positively influence percep?

tions of effort expectancy. Finally, the interaction of the technology characteristics with
the task in influencing performance expectancy highlights the need for organizations
to provide training regarding the fit between various collaboration technologies and
the tasks employees face.
Another lever is associated with designing collaboration tools. It is clear that per?
ceptions of the technology characteristics have a significant influence on both effort
expectancy and performance expectancy. As designers develop collaborative tools, they
should be aware of the importance of social presence, immediacy, and concurrency.
It should be possible to draw from the vast body of knowledge in human-computer
interaction to identify specific design practices that could target each of these three
factors. Thus, developing tools that enable a variety of interactions, potentially includ?
ing pictures and avatars, could be quite valuable.
Finally, organizations can also use the model presented here as a guide to evaluate
how effective their current training and design practices are in driving the key fac?
tors that influence collaboration technology use. Targeted upgrades could be made
to collaboration systems without overhauling the entire system so as to emphasize
key drivers. Similar focused modifications can be made to training programs that
emphasize the critical factors identified in this work. For instance, one such modifica?
tion to a traditional collaboration technology design may be to increase its ability to
be used concurrently with other tasks. SMS has greater concurrency than traditional
collaboration technology in part because it easily can be used on mobile devices
(e.g., one can use SMS on a mobile phone while walking to a meeting). Moving
traditional collaboration technologies to mobile devices would increase their con?
currency. Likewise, changing the organizational culture to encourage the use of col?
laboration technology in meetings to perform other tasks simultaneously [59] would
also increase concurrency.

Conclusions
We developed and validated a model of the use of collaborative technologies. The
model was developed by integrating UTAUT constructs with constructs drawn from
collaboration technology theories?specifically, social presence theory, media richness
theory and its descendants, and the task closure model. The constructs were selected
to be applicable to the general class of collaboration technologies. The model was
validated in two different settings, using two different collaboration tools. Specifically,
we found that UTAUT mediated the effects of various constructs from collaboration
research on intention to use a collaboration technology. The results from our study
have important implications for research on collaboration technologies and provide
practical guidance regarding collaboration technology use in general. This work in?
tegrates major streams of work into a single nomological network, and the proposed
integrated model provides guidance for the design of collaboration technologies to
foster adoption.

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
 PREDICTING COLLABORATION TECHNOLOGY USE 47

Notes
1. The terms "acceptance" and "adoption" are frequently used interchangeably in the
literature. In this paper, we primarily use the term "adoption," but we stay true to the original
sources when it comes to model names, such as the unified theory of acceptance and use of
technology (UTAUT).
2. We explicate the UTAUT hypotheses for the context of this study and highlight the
underlying logic. Given the vast amount of technology adoption research and our focus on the
integration of collaboration technology research and UTAUT, we refer the reader to Venkatesh
et al. [76] for a detailed discussion of the general logic underlying the UTAUT hypotheses.
3. SMS can also be used for providing automated message services, such as welcome mes?
sages, targeted advertising, and voice mail notifications, and can even support "commerce"
applications such as ticket purchases [29]. Our focus in this paper is on the use of SMS as a
collaboration technology, so we excluded non-collaboration-oriented uses of SMS.
4. As pointed out by an anonymous reviewer, this is a fairly high response rate for a survey
with no incentives. We attribute this to the respondents' loyalty to their university as well as
follow-up calls made by the university to encourage responses.
5. As in Study 1, this is a high response rate. In this case, strong top management support
for the collaboration technology and for our study are the reasons for the response rate.
6. In Study 2, we collected user reactions to the collaboration technology for two separate,
different task contexts, so we had two records per respondent in our data set. One concern
with including multiple responses from a single participant is that it is possible that there may
be correlated errors and consequent spurious relationships, although such correlation does not
impair the use of PLS [12]. Nonetheless, to ensure there was nothing unusual in our specific
data set that would cause problems, we conducted an additional empirical analysis to simulate
a between-subjects test of the model. We built 100 data sets that randomly included one of the
responses from each respondent and conducted a separate PLS analysis on each of these 100
data sets. For 98 of these data sets, the results were identical to what we have reported from our
full model test. Thus, the use of both responses from each respondent does not seem to cause
inflated significance for our data set.
7. Note that effort expectancy is coded such that a higher value means lower effort expec?
tancy or easier to use.
8. Remember that effort expectancy is reverse coded, so higher effort expectancy means
less effort is expected.

References
1. Aiken, L.S., and West, S.G. Multiple Regression: Testing and Interpreting Interactions.
London: Sage, 1996.
2. ATKearney. Mobinet 2005. London, October 2005 (available at www.atkearney.com/
images/global/pdf/Mobinet_2005_Detailed_Results .pdf).
3. Bajwa, D.S.; Lewis, L.F.; Pervan, G.; and Lai, V.S. The adoption and use of collaboration
information technologies: International comparisons. Journal of Information Technology, 20,
2 (June 2005), 130-140.
4. Bajwa, D.S.; Lewis, L.F.; Pervan, G; Lai, V.S.; Munkvoldm, B.E.; and Schwabe, G.
Factors in the global assimilation of collaborative information technologies: An exploratory
investigation in five regions. Journal of Management Information Systems, 25, 1 (Summer
2008), 131-165.
5. Baron, R.M., and Kenny, D.A. The moderator-mediator variable distinction in social psy?
chological research: Conceptual, strategic and statistical considerations. Journal of Personality
and Social Psychology, 51, 6 (1986), 1173-1182.
6. Benbasat, I., and Zmud, R.W. The identity crisis within the IS discipline: Defining and
communicating the discipline's core properties. MIS Quarterly, 27, 2 (June 2003), 183-194.
7. Brown, S.A., and Venkatesh, V. Model of adoption of technology in households: A
baseline model test and extension incorporating household life cycle. MIS Quarterly, 29, 3
(September 2005), 399-426.

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
48 BROWN, DENNIS, AND VENKATESH

8. Brown, S.A.; Fuller, R.; and Vician, C. Who's afraid of the virtual world? The role of
anxiety in computer-mediated communication use and satisfaction. Journal of the Association
for Information Systems, 5, 2 (February 2004), 81-109.
9. Brown, S.A.; Massey, A.R; Montoya-Weiss, M.; and Burkman, J. Do I really have to?
Understanding technology acceptance when use is mandatory. European Journal of Information
Systems, 11,4 (December 2002), 283-295.
10. Burton-Jones, A., and Gallivan, MJ. Toward a deeper understanding of system usage in
organizations: A multilevel perspective. MIS Quarterly, 31, 4 (December 2007), 657-679.
11. Carlson, J.R., and Zmud, R.W. Channel expansion theory and the experiential nature of
media richness perceptions. Academy of Management Journal, 42, 2 (April 1999), 153-170.
12. Chin, W.Y.W. The partial least squares approach for structural equation modeling. In
GA. Marcoulides (ed.), Modern Methods for Business Research. Mahwah, NJ: Lawrence
Erlbaum, 1998, pp. 295-336.
13. Christie, B. Human Factors of Information Technology in the Office. New York: John
Wiley & Sons, 1985.
14. Compeau, D.R., and Higgins, CA. Application of social cognitive theory to training for
computer skills. Information Systems Research, 6, 2 (June 1995), 118-143.
15. Compeau, D.R., and Higgins, C.A. Computer self-efficacy: Development of a measure
and initial test. MIS Quarterly, 19, 2 (June 1995), 189-211.
16. Daft, R.L., and Lengel, R.H. Organizational information requirements, media richness
and structural design. Management Science, 32, 5 (May 1986), 554-571.
17. Davis, F.D.; Bagozzi, R.R; and Warshaw, RR. User acceptance of computer technol?
ogy: A comparison of two theoretical models. Management Science, 35, 8 (August 1989),
982-1002.
18. Dennis, A.R., and Gallupe, R.B. A history of GSS empirical research: Lessons learned
and future directions. In L.M. Jessup and J.S. Valacich (eds.), Group Support Systems: New
Perspectives. New York: Macmillan, 1993, pp. 59-11.
19. Dennis, A.R., and Garfield, MJ. The adoption and use of GSS in project teams: Toward
more participative processes and outcomes. MIS Quarterly, 27,2 (June 2003), 289-323.
20. Dennis, A.R., and Reinicke, B. Beta versus VHS and the acceptance of electronic brain
storming technology. MIS Quarterly, 28, 1 (March 2004), 1-20.
21. Dennis, A.R., and Wixom, B.H. Investigating the moderators of group support system
use. Journal of Management Information Systems, 18, 3 (Winter 2001-2), 235-258.
22. Dennis, A.R.; Fuller, R.M.; and Valacich, J.S. Media, tasks, and communication processes:
A theory of media synchronicity. MIS Quarterly, 32, 3 (September 2008), 575-600.
23. Dennis, A.R.; Kinney, S.T.; and Hung, YC. Gender differences in the effects of media
richness. Small Group Research, 30, 4 (August 1999), 405-437.
24. Dennis, A.R.; Wixom, B.J.; and Vandenberg, R.J. Understanding fit and appropriation effects
in group support systems via meta-analysis. MIS Quarterly, 25, 2 (June 2001), 167-197.
25. Dennis, A.R.; George, J.F.; Jessup, L.M.; Nunamaker, J.F.; and Vogel, D.R. Information
technology to support electronic meetings. MIS Quarterly, 12, 4 (December 1988), 591-624.
26. DeSanctis, G, and Gallupe, R.B. A foundation for the study of group decision support
systems. Management Science, 33, 5 (May 1987), 589-609.
27. DeSanctis, G, and Poole, M.S. Capturing the complexity in advanced technology use:
Adaptive structuration theory. Organization Science, 5, 2 (May 1994), 121-147.
28. DeVellis, R.F. Scale Development: Theory and Applications. Newbury Park, CA: Sage,
2003.
29. Doyle, S. Using short message services as a marketing tool. Journal of Database Market?
ing, 8, 3 (April 2001), 273-277.
30. Ericsson, A.B. Loyal wireless enthusiasts. Ericsson, Stockholm, 2001.
31. Fjermestad, J., and Hiltz, S.R. An assessment of group support systems experimental
research: Methodology and results. Journal of Management Information Systems, 15, 3 (Winter
1998-99), 7-149.
32. Fjermestad, J., and Hiltz, S.R. Group support systems: A descriptive evaluation of case
and field studies. Journal of Management Information Systems, 17, 3 (Winter 2000-2001),
115-159.

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
 PREDICTING COLLABORATION TECHNOLOGY USE 49

33. Fowler, G., and Wackerbarth, M. Audio teleconferencing versus face-to-face conferenc?
ing: A synthesis of the literature. Western Journal of Speech Communication, 44, 3 (Summer
1980), 236-252.
34. Fulk, J. Social construction of communication technology. Academy of Management
Journal, 36, 5 (October 1993), 921-950.
35. Fuller, R.M., and Dennis, A.R. Does fit matter? The impact of task-technology fit and
appropriation on team performance in repeated tasks. Information Systems Research, 20, 1
(2009), 2-17.
36. Hassanein, K., and Head, M. Manipulating social presence through the Web interface
and its impact on attitude towards online shopping. International Journal of Human-Computer
Studies, 65, 8 (2007), 689-708.
37. Huang, W., and Wei, K.K. An empirical investigation of effects of GSS and task type on
social interactions from an influence perspective. Journal of Management Information Systems,
77, 2 (Fall 2000), 181-206.
38. Jarvis, C.B.; MacKenzie, S.B.; and Podsakoff, P.M. A critical review of construct indica?
tors and measurement model misspecification in marketing and consumer research. Journal of
Consumer Research, 30, 2 (September 2003), 199-218.
39. Karahanna, E., and Straub, D.W. The psychological origins of perceived usefulness and
ease of use. Information and Management, 35, 4 (April 1999), 237-250.
40. Karahanna, E.; Straub, D.W.; and Chervany, N.L. Information technology adoption across
time: A cross-sectional comparison of pre-adoption and post-adoption beliefs. MIS Quarterly,
23, 2 (June 1999), 183-213.
41. Keen, PG.W. MIS research: Reference discipline and cumulative tradition. In E. McLean
(ed.), Proceedings of the First Conference on Information Systems. New York: ACM Press,
1980, pp. 9-18.
42. Kock, N. The psychobiological model: Towards a new theory of computer-mediated com?
munication based on Darwinian evolution. Organization Science, 15, 3 (2004), 327-348.
43. Koufaris, M. Applying the technology acceptance model and flow theory to online con?
sumer behavior. Information Systems Research, 13, 2 (June 2002), 205-223.
44. Li, D.; Chau, P.Y.K.; and Lou, H. Understanding individual adoption of instant messag?
ing: An empirical investigation. Journal of the Association for Information Systems, 6, 4 (April
2005), 102-129.
45. Lindell, M.K., and Whitney, D.J. Accounting for common method variance in cross
sectional research designs. Journal of Applied Psychology, 86, 1 (2001), 114-121.
46. Malhotra, N.K.; Kim, S.S.; and Patil, A. Common method variance in IS research: A
comparison of alternative approaches and a reanalysis of past research. Management Science,
52, 12(2006), 1865-1883.
47. Markus, M. Toward a critical mass theory of interactive media: Universal access, inter?
dependence and diffusion. Communication Research, 14, 5 (October 1987), 491-511.
48. Massey, A.P, and Montoya-Weiss, M.M. Unraveling the temporal fabric of knowledge con?
version: A model of media selection and use. MIS Quarterly, 30, 1 (March 2006), 99-114.
49. McGrath, J.E. Time, interaction, and performance (TIP): A theory of groups. Small Group
Research, 22, 2 (1991), 147-174.
50. McGrath, J.E., and Hollingshead, A.B. Groups Interacting with Technology. Newbury
Park, CA: Sage. 1994.
51. Morris, M.G, and Venkatesh, V Age differences in technology adoption decisions: Im?
plications for a changing workforce. Personnel Psychology, 53, 2 (Summer 2000), 375-403.
52. Morris, M.G.; Venkatesh, V; and Ackerman, PL. Gender and age in technology adop?
tion and usage decisions: Toward the emergence of a unisex work force. IEEE Transactions on
Engineering Management, 52, 1 (2005), 69-84.
53. Orlikowski, W, and Iacono, C. Research commentary: Desperately seeking the "IT" in
IT research: A call to theorizing the IT artifact. Information Systems Research, 12, 2 (2001),
121-134.
54. Palen, L., and Grudin, J. Discretionary adoption of group support software: Lessons from
calendar applications. In B.E. Munkvold (ed.), Implementing Collaboration Technologies in
Industry. London: Springer Verlag, 2002, pp. 159-180.

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
50 BROWN, DENNIS, AND VENKATESH

55. Pashler, H. Dual-task interference in simple tasks: Data and theory. Psychological Bul?
letin, 116, 2 (1994), 220-244.
56. Pervan, G.; Bajwa, D.; and Lewis, L.F. A study of the adoption and utilization of seven
collaboration technologies in large organizations in Australia and New Zealand. Journal of
Global Information Technology Management, 8, 2 (2005), 5-26.
57. Podsakoff, P.M.; MacKenzie, S.B.; Lee, J.Y.; and Podsakoff, N.P. Common method biases
in behavioral research: A critical review of the literature and recommended remedies. Journal
of Applied Psychology, 88, 5 (2003), 879-903.
58. Reinsch, N.L., and Beswick, R. Voice mail versus conventional channels: A cost minimi?
zation analysis of individuals' preferences. Academy of Management Journal, 33, 4 (December
1990), 801-816.
59. Rennecker, J.; Dennis, A.R.; and Hansen, S. Reconstructing the stage: The use of instant
messaging to restructure meeting boundaries. In R.H. Sprague (ed.), Proceedings of the 39th
Annual Hawaii International Conference on System Sciences. Los Alamitos, CA: IEEE Com?
puter Society Press, 2006.
60. Rheingold, H. Smart Mobs: The Next Social Revolution. Jackson, TN: Basic Books,
2002.
61. Rice, R.E. Computer-mediated communications and organizational innovation. Journal
of Communication, 37, 4 (December 1987), 65-94.
62. Robert, L.P., and Dennis, A.R. Paradox of richness: A cognitive model of media choice.
IEEE Transactions on Professional Communication, 48, 1 (March 2005), 10-21.
63. Sarker, S.V.; Valacich, J.S.; and Sarker, S. Technology adoption by groups: A valence
perspective. Journal of the Association for Information Systems, 6, 2 (2005), 37-71.
64. Short, J.; Williams, E.; and Christie, B. The Social Psychology of Telecommunications.
London: John Wiley, 1976.
65. Straub, D.W., and Burton-Jones, A. Veni, vidi, vici: Breaking the TAM logjam. Journal
of the Association for Information Systems, 8, 4 (April 2007), 224-229.
66. Straub, D.W., and Karahanna, E. Knowledge worker communications and recipient
availability: Toward a task closure explanation of media choice. Organization Science, 9, 2
(March-April 1998), 160-175.
67. Taylor, S., and Todd, P.A. Understanding information technology usage: A test of compet?
ing models. Information Systems Research, 6, 4 (December 1995), 144-176.
68. Thompson, R.L.; Higgins, C.A.; and Howell, J.M. Personal computing: Toward a con?
ceptual model of utilization. MIS Quarterly, 15, 1 (March 1991), 124-143.
69. Venkatesh, V. Determinants of perceived ease of use: Integrating control, intrinsic motiva?
tion, and emotion into the technology acceptance model. Information Systems Research, 11,4
(December 2000), 342-365.
70. Venkatesh, V, and Bala, H. Technology acceptance model 3 and a research agenda on
interventions. Decision Sciences, 39, 2 (May 2008), 273-315.
71. Venkatesh, V, and Brown, S.A. A longitudinal investigation of personal computer adop?
tion in homes: Adoption determinants and emerging challenges. MIS Quarterly, 25, 1 (March
2001), 71-102.
72. Venkatesh, V, and Davis, ED. A model of the antecedents of perceived ease of use:
Development and test. Decision Sciences, 27, 3 (Summer 1996), 451-481.
73. Venkatesh, V, and Davis, ED. A theoretical extension of the technology acceptance model:
Four longitudinal field studies. Management Science, 46, 2 (February 2000), 186-204.
74. Venkatesh, V, and Morris, M.G Why don't men ever stop to ask for directions? Gender,
social influence, and their role in technology acceptance and usage behavior. MIS Quarterly,
24, 1 (March 2000), 115-139.
75. Venkatesh, V; Morris, M.G; and Ackerman, PL. A longitudinal field investigation of
gender differences in individual technology adoption decision making processes. Organizational
Behavior and Human Decision Processes, 83, 1 (September 2000), 33-60.
76. Venkatesh, V; Morris, M.G; Davis, G.B.; and Davis, F.D. User acceptance of information
technology: Toward a unified view. MIS Quarterly, 27, 3 (September 2003), 425-478.
77. Vessey, I.; Ramesh, V; and Glass, R.L. Research in information systems: An empirical
study of diversity in the discipline and its journals. Journal of Management Information Systems,
19, 2 (Fall 2002), 129-174.

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
 PREDICTING COLLABORATION TECHNOLOGY USE 51

78. Waskul, D., and Douglass, M. Cyberself: The emergence of self in online chat. Informa?
tion Society, 13, 4 (October-December 1997), 375-397.
79. Wickens, CD. Multiple resources and performance prediction. Theoretical Issues in
Ergonomics Science, 3, 2 (April 2002), 159-177.
80. Wixom, B.H., and Todd, P.A. A theoretical integration of user satisfaction and technology
acceptance. Information Systems Research, 16, 1 (March 2005), 85-102.
81. Zigurs, I., and Buckland, B.K. A theory of task/technology fit and group support systems
effectiveness. MIS Quarterly, 22, 3 (September 1998), 313-316.

Appendix: Constructs and Measures

Collaboration tool> is replaced with the actual system name in the company.
Study 2 items are shown; Study 1 items were similar.

Use
I rate my intensity of use of collaboration tool> to be: Very light. . . Very heavy
(seven-point scale)
How frequently do you use collaboration tool>: Never . .. Very frequently (seven
point scale)
On an average week, how much time (in hours) do you use collaboration tool>?
Of the opportunities you have to use collaboration tools, including a telephone,
what percentage of time do you choose collaboration tool>?

Intention to Use (seven-point Likert agreement scale)

I intend to use the collaboration tool> in the next 6 months.
I predict I would use the system in the next 6 months.
I plan to use the system in the next 6 months.

Performance Expectancy (seven-point Likert agreement

scale)
I believe collaboration tool> will be useful for communication.
Using collaboration tool> will enable me to accomplish work tasks more
quickly.
Using the collaboration tool will increase my productivity.

Effort Expectancy (seven-point Likert agreement scale)

Using collaboration tool> will not require a lot of mental effort.

I believe collaboration tool> will be easy to use.
Using collaboration tool> will be easy for me.

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
 52 BROWN, DENNIS, AND VENKATESH

Social Influence (seven-point Likert agreement scale)

People who influence my behavior think that I should use collaboration toolx
People who are important to me think that I should use collaboration toolx
The senior management of this business thinks I should use collaboration toolx

Facilitating Conditions (seven-point Likert agreement scale)

I have the resources necessary to use collaboration toolx

I have the knowledge necessary to use collaboration toolx
A specific person (or group) is available for assistance with difficulties with col?
laboration toolx

Social Presence (seven-point Likert agreement scale)

Using collaboration tool> to interact with others creates a warm environment for
communication.
Using collaboration tool> to interact with others creates a sociable environment
for communication.
Using collaboration tool> to interact with others creates a personal environment
for communication.

Immediacy (seven-point Likert agreement scale)

Collaboration tool> enables me to quickly reach communication partners.

When I communicate with someone using collaboration tool>, they usually re?
spond quickly.
When someone communicates with me using collaboration tool>, I try to respond
immediately.

Concurrency (seven-point Likert agreement scale)

I can easily use collaboration tool> while participating in other activities.

I can easily communicate using collaboration tool> while I am doing other things.
I can use collaboration tool> while performing another task.

Technology Experience (seven-point scale)

My experience with audioconferencing is: None at all. . . Very extensive
My experience with videoconferencing is: None at all. . . Very extensive
My experience with messaging tools (e.g., MSN messenger) is: None at all. . . Very
extensive
My experience with technologies similar to collaboration tool> is: None at all. . .
Very extensive

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms
 PREDICTING COLLABORATION TECHNOLOGY USE 53

Computer Self-efficacy (seven-point Likert agreement scale)

I could complete a task using a computer if there was no one around to tell me what
to do.
I could complete a task using a computer even if there was not a lot of time to
complete it.
I could complete a task using a computer if I had just the built-in help facility for
assistance.

Familiarity with Communication Partners (seven-point Likert

agreement scale)
I feel comfortable discussing personal or private issues with co-workers with whom
I collaborate.
I feel comfortable using informal communication (such as slang or abbreviations)
with co-workers with whom I collaborate.
Overall, I feel that I know my collaborators well.

Peer Influence (seven-point Likert agreement scale)

My friends think I should use collaboration tool>.

My peers think I should use collaboration toolx
My co-workers believe I should use collaboration toolx

Superior Influence (seven-point Likert agreement scale)

I believe the top management would like me to use collaboration toolx

My supervisor suggests that I use collaboration toolx
There is pressure from the organization to use collaboration toolx

Resource-Facilitating Conditions (seven-point Likert agreement

scale)
There isn't sufficient access to use collaboration technologyx
Using collaboration tool> is very resource intensive for me.
I am not able to use collaboration tool> when I need it.

Technology-Facilitating Conditions (seven-point Likert agreement

scale)
Collaboration tool> is not compatible with other tools and technologies that I use.
Collaboration tool> is not compatible with other software that I use.
I have trouble using collaboration tool> seamlessly with other applications.

This content downloaded from

159.224.180.78 on Fri, 01 Oct 2021 16:10:18 UTC
All use subject to https://about.jstor.org/terms