User Acceptance of Information Technology: Toward a Unified View

Author(s): Viswanath Venkatesh, Michael G. Morris, Gordon B. Davis and Fred D. Davis
Reviewed work(s):
Source: MIS Quarterly, Vol. 27, No. 3 (Sep., 2003), pp. 425-478
Published by: Management Information Systems Research Center, University of Minnesota
Stable URL: http://www.jstor.org/stable/30036540 .
Accessed: 09/11/2012 04:51

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
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 support@jstor.org.

Management Information Systems Research Center, University of Minnesota is collaborating with JSTOR to
digitize, preserve and extend access to MIS Quarterly.

http://www.jstor.org
 Venkateshet al./UserAcceptanceof IT

MIS

Quarterly
RESEARCH ARTICLE

USER ACCEPTANCEOF INFORMATION

TECHNOLOGY:TOWARDA UNIFIEDVIEW1

By: Viswanath Venkatesh Abstract

Robert H. Smith School of Business
University of Maryland Informationtechnology (IT)acceptance research
Van Munching Hall has yielded many competing models, each with
College Park, MD 20742 differentsets of acceptance determinants.Inthis
U.S.A. paper, we (1) review user acceptance literature
vvenkate@rhsmith.umd.edu and discuss eight prominentmodels, (2) empiri-
cally compare the eight models and their exten-
Michael G. Morris sions, (3) formulatea unifiedmodel thatintegrates
Mclntire School of Commerce elements across the eightmodels, and (4) empiri-
University of Virginia cally validatethe unifiedmodel. Theeight models
Monroe Hall reviewed are the theory of reasoned action, the
Charlottesville, VA 22903-2493 technology acceptance model, the motivational
U.S.A. model, the theory of planned behavior, a model
mmorris@virginia.edu combiningthe technologyacceptance model and
the theoryof planned behavior, the model of PC
Gordon B. Davis utilization,the innovationdiffusiontheory,and the
Carlson School of Management social cognitive theory. Using data from four
University of Minnesota organizationsover a six-monthperiod withthree
321 19'" Avenue South points of measurement, the eight models ex-
Minneapolis, MN 55455 plainedbetween 17 percent and 53 percent of the
U.S.A. variance in user intentions to use information
gdavis@csom.umn.edu technology. Next, a unified model, called the
UnifiedTheoryof Acceptance and Use of Tech-
Fred D. Davis nology (UTAUT),was formulated,withfour core
Sam M.Walton College of Business determinantsof intentionand usage, and up to
University of Arkansas fourmoderatorsof key relationships.UTAUTwas
Fayetteville, AR 72701-1201 then tested using the originaldata and found to
U.S.A. outperformthe eight individualmodels (adjusted
fdavis@walton.uark.edu R2 of 69 percent). UTAUTwas then confirmed
withdata fromtwo new organizationswithsimilar
results (adjustedR2of 70 percent). UTAUTthus
'CynthiaBeath was the accepting senior editorfor this
paper. provides a useful tool for managers needing to

Vol.27 No.3, pp. 425-478/September

MISQuarterly 2003 425
 et al./UserAcceptanceof IT
Venkatesh
assess the likelihoodof success for new techno-
logy introductionsand helps them understandthe
driversof acceptance in orderto proactivelyde-
sign interventions(includingtraining,marketing,
etc.) targetedat populationsof users thatmay be
less inclinedto adopt and use new systems. The
paper also makes several recommendationsfor
future research including developing a deeper
understandingof the dynamicinfluences studied
here, refiningmeasurementof the core constructs
used in UTAUT,and understandingthe organiza-
tionaloutcomes associated withnew technology
use.
Keywords: Theory of planned behavior, inno-
vation characteristics, technology acceptance
model, social cognitive theory, unified model,
integratedmodel
Introduction
The presence of computerand informationtech-
nologies in today's organizationshas expanded
dramatically.Some estimates indicatethat,since
the 1980s, about 50 percent of all new capital
investmentin organizationshas been in informa-
tion technology (Westlandand Clark2000). Yet,
fortechnologies to improveproductivity,
they must
be accepted and used by employees in organi-
zations. Explaininguser acceptance of new tech-
nology is often described as one of the most
matureresearch areas in the contemporaryinfor-
mation systems (IS) literature(e.g., Hu et al.
1999). Research in this area has resulted in
several theoreticalmodels, withroots in informa-
tion systems, psychology, and sociology, that
routinelyexplainover40 percentof the variancein
individualintentionto use technology (e.g., Davis
et al. 1989; Taylorand Todd 1995b; Venkatesh
and Davis 2000). Researchers are confronted
with a choice among a multitudeof models and
find that they must "pickand choose" constructs
across the models, or choose a "favoredmodel"
and largely ignore the contributions from
alternativemodels. Thus, there is a need for a
reviewand synthesis in orderto progress toward
a unifiedview of user acceptance.
426 Vol.27 No.3/September
MISQuarterly 2003
The currentworkhas the followingobjectives:
(1) To review the extant user acceptance
models: The primarypurpose of this review
is to assess the currentstate of knowledge
with respect to understanding individual
acceptance of new informationtechnologies.
This reviewidentifieseight prominentmodels
and discusses their similarities and dif-
ferences. Some authorshave previouslyob-
served some of the similarities across
models.2 However,our review is the firstto
assess similaritiesand differences across all
eight models, a necessary first step toward
the ultimategoal of the paper: the develop-
ment of a unifiedtheory of individualaccep-
tance of technology. The reviewis presented
in the followingsection.
(2) To empiricallycompare the eight models:
We conduct a within-subjects,longitudinal
validation and comparison of the eight
models using data from four organizations.
This providesa baseline assessment of the
relative explanatorypower of the individual
models against which the unifiedmodel can
be compared. The empiricalmodel compari-
son is presented in the thirdsection.
(3) To formulatethe Unified Theory of Accep-
tance and Use of Technology (UTAUT):
Based upon conceptual and empiricalsimi-
laritiesacross models, we formulatea unified
model. The formulationof UTAUTis pre-
sented in the fourthsection.
(4) To empiricallyvalidateUTAUT:An empirical
test of UTAUTon the originaldata provides
preliminarysupport for our contention that
UTAUToutperformseach of the eight original
models. UTAUTis then cross-validatedusing
data fromtwo new organizations.The empiri-
cal validationof UTAUTis presented in the
fifthsection.
2Forexample,MooreandBenbasat(1991)adaptedthe
perceivedusefulnessandease of use itemsfromDavis
et al.'s(1989)TAMto measurerelativeadvantageand
complexity,respectively,in theirinnovationdiffusion
model.

-----
f--------'--------'---------------------
Individualreactions to Intentions to use
using information information
technology technology
Fiur 1. 0
s0Cncptndrly Sg UserAc
Review of ExtantUser
Acceptance Models
Descriptionof Models
and Constructs
IS research has long studied how and why indivi-
duals adoptnew informationtechnologies. Within
this broadarea of inquiry,there have been several
streams of research. One stream of research
focuses on individualacceptance of technologyby
using intentionor usage as a dependent variable
(e.g., Compeau and Higgins 1995b; Davis et al.
1989). Other streams have focused on
implementationsuccess at the organizationallevel
(Leonard-Barton and Deschamps 1988) and task-
technology fit (Goodhue 1995; Goodhue and
Thompson 1995), among others. While each of
these streams makes important and unique
contributionsto the literatureon user acceptance
of informationtechnology,the theoreticalmodels
to be includedin the present review,comparison,
and synthesis employ intentionand/or usage as
the key dependent variable. The goal here is to
understandusage as the dependentvariable.The
role of intentionas a predictorof behavior(e.g.,
usage) is criticaland has been well-establishedin
IS and the reference disciplines(see Ajzen 1991;
Sheppard et al. 1988; Taylorand Todd 1995b).
Figure1 presents the basic conceptualframework
underlyingthe class of models explainingindivi-
dual acceptance of informationtechnology that
forms the basis of this research. Our review re-
sulted in the identificationof eight key competing
theoreticalmodels. Table 1 describes the eight
Venkateshet al./UserAcceptance of IT
Actual use of
information
technology
cepta nceModel
models and defines theirtheorized determinants
of intention and/or usage. The models hypo-
thesize between two and seven determinantsof
acceptance, for a total of 32 constructs across
the eight models. Table 2 identifies four key
moderatingvariables (experience, voluntariness,
gender, and age) that have been found to be
significantin conjunctionwiththese models.
PriorModel Tests and
ModelComparisons
There have been manytests of the eight models
but there have only been four studies reporting
empirically-basedcomparisonsof two or more of
the eight models publishedin the majorinforma-
tion systems journals. Table 3 provides a brief
overview of each of the model comparison
studies. Despite the apparentmaturityof the re-
search stream, a comprehensive comparison of
the key competing models has not been con-
ducted in a single study. Below, we identifyfive
limitationsof these priormodeltests and compari-
sons, and howwe address these limitationsin our
work.
Technology studied: The technologies that
have been studied in many of the model
development and comparison studies have
been relatively simple, individual-oriented
informationtechnologies as opposed to more
complex and sophisticated organizational
technologies thatare the focus of managerial
concern and of this study.
MIS QuarterlyVol.27 No. 3/September2003 427
Ven

the p. of
302). 1112).
a a itself,
activity activity
p.
p. 1989,
perform an achieving an process
(evaluative
not 1975, usingjob using only. promotions" 1992,
in activity the
(Davis or al.
her that
that the
or TAM2 perform pay, perform thanet
feelings Ajzen
should in to to
his effort" from
or and of
believes believes instrumental want other
(Davis
320). want
negative p. free be will se"
should Included will distinct
or enhance be to
he (Fishbeinperson
a
person
1989,
a areperformance,
1112).
per
users users
would would thatjobp.
think that reinforcement
that
positive
216). which(Davis
which activity
TRA/TPB. perceived
p. him question"to to is 1992,
the
to in system system from it al.
improved
1975, outcomeset apparent
degree degree perception as perception
individual's no
"an Ajzen important
Definitions particular
behavior"the performance"
particular
"the 320). Adapted The"because
such
valued performing
"for
(DavisThe

of

Norm Ease
Usefulness Norm
Toward Motivation Motivation
Constructs

BehaviorSubjective
Core Attitude Perceived
Perceived Use Subjective Extrinsic Intrinsic

pre-
onTAM in- of a for
most al. the to sup-
in
largely
et techno- of explain
by to
thebehavior.of has motiva-
'1*ividualf.Acceptane1~
behaviorswasTRA usage Vallerand adoption
of TAM (Venkatesh
better motivational
fundamental
information
1999).
of Davis and to predictor
applied
Ind111 onehuman designed explanation
thethe
f
is of an of applied
range was settings Speier
o[ acceptance
explained (TAM) order widely as
extended users. psychology contexts.
TRA review). employed in examinedWithin technology
(TRA) wide a and conceptualization
additional in
ls review(1992)and
theories had acceptance been
and
a for Model TAM2an have
theory specific al. new
hereE final as base.
variance
that
individual has
mandatory et
Action 1988 theconstructof
contexts, (MM)research for
to
predict the it excellent
psychology, IS normTAM of studies DavisVenkatesh
to al.
influential
behaviors. an
et TRA thatstudies to technology
TRA, case motivation understand
technologies
Model
S'[
oelsan and
social Acceptance attitude body to also
theoretical
Reasonedused with
other of
the2000). adapted
Several
1. found theparsimoniously.
Unlike in thisdomain,
(see
of from applied of tailored subjectiveset generalandpresents
been of theory
: Sheppard
and is information Davis use
job.
has significant
Tabe~1.TheoryDrawn It (see(1989)
fundamental logyconsistent
context
TAMdicttheexcludes
Technology intention
intention
cludinganddiverse A ported
behavior.
Motivational theory systems
tenets
(1997) and
tional

428
 Venka

IS
theof constraints
149).
p.
context
external
theand
performing
of In 1995b,

188). Todd
internal
p. of
difficulty
or and
1991,
TRA. TRA. ease TRA/TPB.
(Taylor TRA/TPB. TAM.
TRA/TPB.

from from
"perceptionsfrom from
(Ajzen from from
perceived
behavior"
Definitions
Adapted Adapted behavior"
"the research,
on Adapted Adapted
Adapted Adapted

Norm Behavioral Norm Usefulness

Behavioral
(ontiued
Toward Toward
Constructs

Core Behavior
Attitude Subjective Perceived
Control Behavior
Attitude Perceived
Subjective
Perceived
Control

TPBof tech-
control
of Taylorpre- atti-
control
perceived
intention
predict of contrast
of of In technology hybrid
with
to settings. a
review different
1991;
Decomposed
ndvidalAcceptane of terms
a TPB
behavioral theIn TPB. behavioral TPB
understanding within of
manyis to provide
construct used theof "decomposes"
variety to
(TPB) the determinantto Mathieson
presented
perceived model
(DTPB). structure
wide usage perceived(C-TAM-TPB)
DTPB TAM
a identical predictors
1995a).
1997; is
TPB, (1991)in applied
adding andal. related andbelief TPBthefrom
Behavior
by In additional
successfullyet A BehaviorTAM,
an to Todd
DTPBnorm, and
that
TRA be Ajzenbehavior and
to similar TAM combines
Planned
MdelSad-Teoresof control. acceptance
andsuccessfully Planned underlying
1995b).
(Harrison contexts.
usefulness
of studies of but
intention,
the
subjective model(Taylor
1
extendedbehavior.been Todd its
theorized TPB
Tabe
TheoryTPB intention
is andseveral
behavioral nologies
hasindividual dicting
andTheory intoadoption
to tude, This
Combined model
perceived

429
VAnke

to et
of
[a her to
joy, act" agree
or of culture, is
degree128). individual returning (Thompson
using difficult fee"provision
his p. (Thompson individual's facilitating
of "the displeasure, the
that particular no of
a "the observers
1991, "feelings situations"
subjective
that
future" is example,type
relatively
al. are that whencontext,
utilization"
(1971), disgust,
with
believes as et the use social For IS one
129). in group's anbe
performance factors
p. In system
the toward 127).
individual 126).
agreements facilitated
p.
specific is
p. environment may
individual pay-off depression,
1991, perceived
Shoemaker a
(Thompson an social in item.
an is or reference theaccomplish.
PCs
al. and affectby
1991, the 1991,
in theof influence
to online
enhance
et use"
have
of
al. Triandis,others,
al. can129).
which interpersonal easy
to can andthat 129). et et users p.
return
Rogers pleasure,
innovationp. Triandis, with factors that
from act to for
on an on or associated anpurchased 1991,
extent(Thompson specific
made
1991, hate al.
"the job"Based
Definitions
technology]
which al. Based
understand
"Outcomes elation, Derived
internalization
or(Thompson andhas(Thompson
make
items
Objectivecharged
support
condition
et

Use
Conditions
(honinuIIed)
Constructs Towards Factors

Long-term
Core Job-fit Complexity Affect
Consequences Social Facilitating
is
Acceptanc.Le
al. con- a al.
et IS
human of et thanroots,
these
of perspective
for particularly
use of different
SI'[*Individual it rather examination
the
utilization.
and theory's
theory model anof
effect
of PC
Thompson Thompson
the
makes thesuch
competing behavior
a TPB.
(1977) with
predict
Triandis'
model
acceptance Also,
(MPCU) and to
Theor1i:es usage comparison
examine
the
presents of keeping fair
and
Triandis'
TRArefined
model in will a
technologies.
by and predict intention.
individual
from the
model
Utilization nature to on ensure
Modls research
PC this thepredict
used however,to
information
largely
1. of adapted to of sought
proposed
and
current
that
Model
TabJIle behavior,
Derived
to (1991)However,
texts range
suited (1991)
intention; models.
thedeterminants
important

430
 Venka

to as
p.
being being system being
(Moore
as 1991, as the as past
Benbasat
195). system"
Benbasat
p. perceived perceived
is andand is Benbasat
using
social and innovation,
Benbasat
perceived 1991,
perceived perceived the and
is is is needs,
(Moore
and one's others
Moore
innovation
in 195). communicability"
innovation
see using (Moore
an p. from
Benbasat values, of and the
(Moore of of will"
innovationinnovation status can adopters" 203).
innovation
and useor 1991,
an an one an resultsp. usefree
(adapted existing of
the or
which which
precursor" which
(Moore image which thepotential
which 1991,
which
of observability
to its to to Benbasat
to to of to
use" one's with
195). their 195).
than to and
organization p. p.
voluntary,
degree degreedegree degree degree tangibility
degree
Benbasat
the
Definitions
better
"the 195).
"the "The (Moore
enhance
difficult Thein "the 1991,
consistent
experiences
"the
including
and"the 1991,
being

Use
of

Advantage Demon-
Constructs of

Core Relative Ease Image Visibility Compatibility

Results Voluntariness

in
1999;1991).
be Moore
been alsoal.
et
hasinnovations, (1991) could predictive
presented (see
of
IndvidualAccpace(otiud
Within that the
of
1995) for
acceptance.
Benbasat
organizational
variety 1982). Karahanna
(IDT) a to innovations
(Rogers and constructs
support
Klein of 1998;
characteristics
studytools of
IDT technology
Theoryto andMooreset found1997,
adThois a
1960s innovation
(1996)
individual
odel*s agricultural
sociology, systems,characteristics
refined Prasad
2001).
Diffusion
in the (Tornatzky al.
. from theandstudy theseandet
of
since to Benbasat
Tabe
used
Innovation
Grounded information
Rogers
innovation
ranging used
adapted Plouffe
andvalidity
Agarwal

431
Venkat

and
1995b).
job-
behavior.Specifically,
esteem (e.g., it
(e.g.,
thewith task.
1995b). Higgins or
of when
deal
and job
individual
behavior. behavior
Higgins technology
thethe a
andof reactions
with useparticular
expectations(Compeau
consequences to a particular
a
deal
for emotional
(Compeau ability
or
liking
consequences
performance accomplish
one's
of to
accomplishment use). anxious
expectations
outcomesof
performance-related
personal individual's
Definitions Thepersonal
TheSpecifically,
related senseJudgment
computer) Evoking
Ancomputer p

Expec-
Cotined
Constructs

Core
Outcome
tations- Expectations-
Personal
Performance
Outcome Affect Anxiety
is al. to
it
theet is andthe
to em-
butallow Compeau and
Compeaualsothus
behavior
SCT of predictive
use
andCompeau
Compeau theory predicting
the models.
information
dependent
human 1986). of a of intention
also(1995a) model
of the
ofIndvidalAccetace
of extended as
use
computer of
(see research. spirit
examine
and Higgins
Bandura underlying
performance andoriginal
the
(SCT) theories usagewillcontext
the The
studied
(see and current withwethe
applied study used in comparison
utilization
to the and
Theory
powerful
theory of modelacceptance fair
was keeping
general. a
model
(1995b) it
Compeau model
to (1995b)
in
most goal in
MoelsandTheris computer the acceptance,
theallow
of the(1995b) but
1 thecognitivewhile
Cognitive SCT, of of to
of Higgins Higgins
extended
Tabe
Social
One social 1999);
ployed
andcontext Higgins'
nature
outside andvariable
individual
be technology usage
validity

432
 Vanke

AgeN/A N/A N/A

of of
in- evi- moreMorriswaswomen a
salient
useful-perceived
original
not was andeffect forstages
(i.e.,
the more women norm interaction).
was
in Empirical while
usefor The
demonstrated early
salient
was of
perceived
men the
N/A
Gender Gender
cludeddence
TAM. for ease
thatness 2000).
salient more
(Venkatesh
subjective N/A
in experience
three-way

the a
not notin and
was norm cases
notand norm less
original
in
in (i.e.,
was was Within
suggestedusebe
the only
settings
important
to only interaction).
experience
in included system
Hartwick TAM.
subjective
Although system
(1994)more
subjective salientthen the
limited
included
Voluntariness
TRA. Barki
tested,
Voluntariness thatwas perceived
when voluntary.
explicitly
original
TAM2,
Voluntariness mandatory
was even
of withthree-way N/A

of
in in
and
of Davis attitude pro- with
became that
lsI3 role by salience
included that have
experience.
(1989)
included
thecontrast,norm al.
theexamined
in In increasing et showing
analysis found others,
nonsignificant
xistinigMode
explicitly with increasing
explicitly
found.
However,change (1999)subjectiveDavis
not empirically not evidence
inE al. with among
No was becomes
wasTRA. et while was
was important TAM. experience,
cross-sectional (1996),use
a (1989). empirical
of
original moreimportant
original
al.
Experience
Experience
theexperience
using Karahanna
et determinants
was lessExperience
theSzajna
experience, vided
ease
increased N/A

RolefMoei[aISJtors
2. of (and

TableModelReasoned
Theory
Action Model
Technology
TAM2)
Acceptance Motivational
Model

433
Venkat

a
atti- was
salient older salient
(i.e.,
that for
Venkateshworkers
control more interaction).
more Subjective
women
andfound perceivedsalientwas
was
younger older
AgeMorris
(2000)
tude behavioral
forwhile more norm N/A
to three-way
workers.

was more
in
(2000)men.
norm exper-
al. for beha- were of
et attitude womenthree-way
that for (i.e.,
control
salient
perceivedstages
subjective

Gender Both
more
Venkatesh
found andvioral ience
salient
early N/A
interactions).

not to
more (Hart-
As
original was system 1994).
was
tested,be
discussion
the to when
norm
in TRA,
not Barki
the perceived
voluntary
DTPB.
in
or was and
less
included
TPB
Voluntariness regarding
noted
Voluntariness although usebe wickN/A
important
subjective
suggested

(onined in that thein

studies to control
been this users). became
norm (1999) toward
has 2000). behavioral
included increasing
similar
al.
intodesign
norm
It relationship is et experience
follow-on and increasing
behavioral
demonstratedwith
via the This attitude
explicitly with
DTPB. has norm
xstigMoes
subjective
inexperienced 1995a).
i notor TPBVenkatesh incorporated subjective
increasing
that
important
into Karahanna
TRA. and perceived
salient
TPB and moderates
was of of was between-subjects
usefulness, Todd
with
while
evidence such
lessexperience. a
subjective and
more and
Morris of in
all salient
odeatos original context
Experience
Experience
theincorporated
(e.g.,
Empirical
between
intention,
becomes
experience levels the Experience
suggestionmodel
(experienced
Perceived
behavior,
were less(Taylor
experience
of

2.Rol
of

Tabl
ModelPlanned
TheoryBehavior Combined
TAM-TPB

434
 Venkat

AgeN/A N/A N/A

N/A
Gender N/A N/A

not have
was to on
moderator,
a
shown
effect
as
d
was
direct
N/A
Voluntariness buta intention.
Voluntariness
tested N/A

(oninS the trial-

other on
com- more a found visi-
the
factors, study usage showed use, relative
predictors
that all On conse-
importantto adoptionof and
andandof
Moes
only
significant.
stn
social
were results ease
found conducted
experience.
of were
usage,
use, long-term Thesignificant
characteristics
comparison
increasingly
(1999) predictors
experience)
experience) the for
inEx (1994) experience.
conditions
about levels image
al. toward al. the demonstrability,
advantage,
et less et in and
became innovation
(no/low behavior.
(greater contrast,
affect with of adoption,
results
In
Sodeatos concern relative
facilitating increasing for
usage
of
Experience
plexity, hand,
andsalient
Thompson withKarahanna
quences between-subjects
impactbehavior
adoption
differences bility.
ability,
vs.thatwere N/A
advantage

PC
2.Rol of

Tabl
ModelUtilization
Model Innovation
Diffusion
Theory Social
Cognitive
Theory

435
Tal

TRA and
by and51%, was62% was was45%
intentionand intention intention
57%, intention
in 26%, in TAM was in TAM in TAMwas
by TPB by was60% by IDT
and47%
explained
TPB was
use32%
variance was variance
and variance and
variance
Findings
Theandwas respectively.
TAM 70%
Theexplained DTPB
Theexplained
52%, Theexplained
33%

or thein
at
points
Longitudinal
Cross-
Sectional
Analysis
Cross-
sectional
analysis sectional
Cross-
twotime Cross-
sectional Cross-
sectional

Points mea-
center
students
of student.
weeks all the
14 three-month per
surveyed-
a multiple
Measurement
Number
of Two;
apart One Forperiod, i.e.,sures
visiting
were One

as the
of to the per-
to tech-
a to task with months
famil-
new with partici-
had a students
already 10
technology use
Newness
Studied
Technology
Participants
the
were iarity
Some pant
each
technology Many
nology
choose
form were center
familiar Survey
administered
after
of

students students students

Participants
107 262 786 176merchants
to to
word
Study a usethe mar- card.
center a an pay-
of of to in
of of
intention
compari- intention
compari-
spread-
and intention
compari- behavioral
computing compari- smart
system
of use of a of a of use trial
odel

(Incl.
Context
Technology)
model
sonandprocessor
Between-subjects
Within-subjects model
sonusesheet
calculator model
sonuseresource
Within-subjects model
sonintentionketelectronic
andcontext
Within-subjects using
ment

TAM TPB IDT

Models
Compared
Theories/
TRA, TAM, TAM,
TPB/DTPB TAM,

al.
al.
and(1995b) et
et

Model
Comparison
Studies
Tale3.ReieoPio Davis
(1989) (1991)
Mathieson Taylor
Todd Plouffe
(2001)

436

* Participants: While there have been some
tests of each modelinorganizationalsettings,
the participantsin three of the four model
comparison studies have been students-
only Plouffe et al. (2001) conducted their
research in a nonacademic setting. This
research is conducted using data collected
fromemployees in organizations.
Timingof measurement: Ingeneral, most of
the tests of the eight models were conducted
well after the participants'acceptance or
rejection decision rather than during the
active adoption decision-making process.
Because behavior has become routinized,
individualreactionsreportedin those studies
are retrospective(see Fiske and Taylor1991;
Venkateshet al. 2000). Withthe exceptionof
Davis et al. (1989), the model comparisons
examined technologies that were already
familiarto the individualsat the time of mea-
surement. Inthis paper,we examine techno-
logies fromthe timeof theirinitialintroduction
to stages of greaterexperience.
* Nature of measurement: Even studies that
have examined experience have typically
employed cross-sectional and/or between-
subjects comparisons(e.g., Daviset al. 1989;
Karahannaet al. 1999; Szajna 1996; Taylor
and Todd 1995a; Thompson et al. 1994).
This limitationapplies to model comparison
studies also. Our work tracks participants
throughvarious stages of experience with a
new technologyand compares all models on
all participants.
* Voluntaryvs. mandatorycontexts: Most of
the model tests and all four model com-
parisons were conducted in voluntaryusage
contexts.3 Therefore,one must use caution
when generalizing those results to the
3NotableexceptionsareTRA(Hartwick andBarki1994)
and TAM2(Venkateshand Davis 2000) as well as
studiesthathaveincorporated as a direct
voluntariness
effect(on intention) in orderto accountfor perceived
nonvoluntary adoption(e.g.,Agarwal andPrasad1997;
Karahanna et al. 1999;MooreandBenbasat1991).
Venkatesh
et al./UserAcceptanceof IT
mandatorysettings that are possibly of more
interest to practicing managers. This re-
search examines both voluntaryand man-
datoryimplementationcontexts.
EmpiricalComparisonof the
Eight Models
Settings and Participants
Longitudinalfield studies were conducted at four
organizationsamong individualsbeing introduced
to a new technology in the workplace. To help
ensure our results would be robust across
contexts, we sampled for heterogeneity across
technologies, organizations,industries,business
functions, and nature of use (voluntary vs.
mandatory).Inaddition,we capturedperceptions
as the users' experience with the technology
increased. At each firm,we were able to time our
data collection in conjunction with a training
program associated with the new technology
introduction. This approach is consistent with
priortrainingand individualacceptance research
where individualreactions to a new technology
were studied (e.g., Davis et al. 1989; Olfmanand
Mandviwalla1994; Venkatesh and Davis 2000).
A pretested questionnairecontainingitems mea-
suring constructs from all eight models was
administered at three different points in time:
post-training(T1),one monthafterimplementation
(T2),and three monthsafterimplementation(T3).
Actualusage behaviorwas measuredoverthe six-
monthpost-trainingperiod. Table 4 summarizes
key characteristicsof the organizationalsettings.
Figure2 presents the longitudinaldata collection
schedule.
Measurement
A questionnairewas created withitems validated
in priorresearch adapted to the technologies and
organizationsstudied. TRAscales were adapted
from Davis et al. (1989); TAM scales were
adapted from Davis (1989), Davis et al. (1989),
Vol.27 No.3/September
MISQuarterly 2003 437
Venkateshet al./UserAcceptance of IT

x 0 X 0 X 0 X 0
Training User System User System User System Usage
Reactions Use Reactions/ Use Reactions/ Use Measurement
Usage Usage
Measurement Measurement
1 week 1 month 3 months 6 months

Figure 2L g d Sche

Tal 4 ecrpio fStde

Functional Sample
Study Industry Area Size System Description
Voluntary Use
Onlinemeeting managerthat could be
Product used to conductWeb-enabled video or
la Entertainment 54
Development audio conferences in lieu of face-to-face
or traditionalphone conferences
Database applicationthat could be used
Telecomm to access industrystandardsfor particular
Services productsin lieu of other resources (e.g.,
technical manuals, Web sites)
Mandatory Use
Business Portfolioanalyzerthat analysts were
2a Banking Account 58 requiredto use in evaluatingexisting and
Management potential accounts
Public Proprietaryaccountingsystems on a PC
2b Administration Accounting 38 platformthat accountantswere required
to use for organizational bookkeeping

and Venkateshand Davis(2000); MMscales were perceived voluntariness was measured as a

adapted from Davis et al. (1992); TPB/DTPB
scales were adapted from Taylor and Todd
(1995a, 1995b);MPCUscales were adaptedfrom
Thompsonet al. (1991); IDTscales were adapted
fromMooreand Benbasat (1991);and SCTscales
were adaptedfromCompeauand Higgins(1995a,
1995b) and Compeau et al. (1999). Behavioral
intentionto use the system was measured using
a three-item scale adapted from Davis et al.
(1989) and extensively used in much of the
previousindividualacceptance research. Seven-
point scales were used for all of the aforemen-
tioned constructs'measurement,with1 being the
negative end of the scale and 7 being the positive
end of the scale. In additionto these measures,
manipulationcheck per the scale of Mooreand
Benbasat (1991), where 1 was nonvoluntaryand
7 was completely voluntary. The tense of the
verbs in the variousscales reflectedthe timingof
measurement: futuretense was employedat T1,
present tense was employed at T2 and T3 (see
Karahannaet al. 1999). The scales used to mea-
sure the key constructs are discussed in a later
section where we performa detailed comparison
(Tables 9 through 13). A focus group of five
business professionals evaluated the question-
naire, following which minor wording changes
were made. Actual usage behavior was mea-
sured as durationof use via system logs. Due to
the sensitivity of usage measures to network

438 MISQuarterlyVol.27 No. 3/September2003


availability,inallorganizationsstudied,the system
automaticallylogged off inactive users after a
period of 5 to 15 minutes, eliminatingmost idle
time fromthe usage logs.
Results
The perceptionsof voluntarinesswere veryhighin
studies la and lb (la: M = 6.50, SD = 0.22; lb:
M = 6.51, SD = 0.20) and very low in studies 2a
and 2b (la: M = 1.50, SD = 0.19; 1b: M = 1.49,
SD = 0.18). Giventhis bi-modaldistributioninthe
data (voluntaryvs. mandatory),we created two
data sets: (1) studies la and Ib, and (2) studies
2a and 2b. This is consistent withVenkateshand
Davis (2000).
Partial least squares (PLS Graph, Version
2.91.03.04) was used to examine the reliability
and validityof the measures. Specifically, 48
separate validitytests (twostudies, eight models,
three time periods each) were run to examine
convergent and discriminantvalidity. In testing
the various models, only the direct effects on
intention were modeled as the goal was to
examine the predictionof intention rather than
interrelationshipsamong determinantsof inten-
tion; further,the explained variance (R2) is not
affected by indirectpaths. The loading pattern
was found to be acceptable with most loadings
being .70 or higher. All internal consistency
reliabilitieswere greaterthan .70. The patternsof
results found in the currentworkare highlycon-
sistent withthe results of previousresearch.
PLS was used to test all eight models at the three
points of measurement in each of the two data
sets. In all cases, we employed a bootstrapping
method (500 times) that used randomlyselected
subsamples to test the PLS model.4Tables 5 and
6 present the model validationresults at each of
the points of measurement. The tables report the
varianceexplainedand the beta coefficients. Key
4Theinterestedreaderis referredto a moredetailed
andhowitcomparestoother
ofbootstrapping
exposition
suchas jackknifing
techniquesof resampling (see Chin
1998;EfronandGong1983).
Venkateshet al./UserAcceptance of IT
findingsemerged fromthese analyses. First,all
eight models explained individualacceptance,
withvariance in intentionexplained rangingfrom
17 percent to 42 percent. Also, a key difference
across studies stemmed from the voluntaryvs.
mandatorysettings-in mandatorysettings (study
2), constructs related to social influence were
significantwhereas inthe voluntarysettings (study
1), they were not significant. Finally,the deter-
minants of intentionvaried over time, with some
determinantsgoing from significantto nonsigni-
ficantwithincreasingexperience.
Followingthe test of the baseline/originalspecifi-
cations of the eight models (Tables 5 and 6), we
examined the moderatinginfluences suggested
(eitherexplicitlyor implicitly)inthe literature-i.e.,
experience, voluntariness, gender, and age
(Table2). Inorderto test these moderatinginflu-
ences, stay true to the model extensions
(Table 2), and conduct a complete test of the
existing models and their extensions, the data
were pooled across studies and time periods.
Voluntariness was a dummy variable used to
separate the situational contexts (study 1 vs.
study2); this approachis consistentwithprevious
research (Venkatesh and Davis 2000). Gender
was coded as a 0/1 dummyvariable consistent
with previous research (Venkatesh and Morris
2000) and age was coded as a continuous vari-
able, consistent with priorresearch (Morrisand
Venkatesh 2000). Experience was operationa-
lizedvia a dummyvariablethattookordinalvalues
of 0, 1, or 2 to capture increasing levels of user
experience with the system (T1, T2, and T3).
Using an ordinal dummy variable, rather than
categorical variables, is consistent with recent
research (e.g., Venkatesh and Davis 2000).
Pooling the data across the three pointsof mea-
surement resulted in a sample of 645 (215 x 3).
The results of the pooled analysis are shown in
Table 7.
Because pooling across time periods allows the
explicitmodelingof the moderatingrole of exper-
ience, there is an increase in the variance ex-
plainedin the case of TAM2(Table 7) compared
to a main effects-only model reported earlier
(Tables 5 and 6). One of the limitationsof pooling
is that there are repeated measures from the
MIS QuarterlyVol.27 No. 3/September2003 439
Venkateshet al./UserAcceptance of IT

I Table 5.0Stu -y1 .Peicig 0Intention in Voluntary Settings

Time I (N = 119) Time 2 (N = 119) Time 3 (N = 119)

Models Independent variables R2 Beta R2 Beta R2 Beta
TRA Attitude toward using tech. .30 .55*** .26 .51*** .19 .43***
Subjectivenorm .06 .07 .08
TAM/ Perceived usefulness .38 .55*** .36 .60*** .37 .61**
TAM2 Perceived ease of use .22** .03 .05
Subjectivenorm .02 .06 .06
MM Extrinsic motivation .37 .50*** .36 .47*** .37 .49***
Intrinsic motivation .22** .22** .24***
TPB/ Attitude toward using tech. .37 .52*** .25 .50*** .21 .44***
DTPB Subjectivenorm .05 .04 .05
Perceived behavioralcontrol .24*** .03 .02
C-TAM- Perceived usefulness .39 .56*** .36 .60*** .39 .63***
TPB Attitudetowardusing tech. .04 .03 .05
Subjectivenorm .06 .04 .03
Perceived behavioralcontrol .25*** .02 .03
MPCU Job-fit .37 .54*** .36 .60*** .38 .62***
Complexity(reversed) .23*** .04 .04
Long-termconsequences .06 .04 .07
Affecttowarduse .05 .05 .04
Social factors .04 .07 .06
Facilitatingconditions .05 .06 .04
IDT Relativeadvantage .38 .54*** .37 .61*** .39 .63***
Ease of use .26** .02 .07
Result demonstrability .03 .04 .06
Trialability .04 .09 .08
Visibility .06 .03 .06
Image .06 .05 .07
Compatibility .05 .02 .04
Voluntariness .03 .04 .03
SCT Outcomeexpectations .37 .47*** .36 .60*** .36 .60***
Self-efficacy .20*** .03 .01
Affect .05 .03 .04
Anxiety -.17*" .04 .06

Notes: 1. *p < .05; **p < .01; ***p< .001.

2. When the data were analyzed separately for studies 2a and 2b, the patternof results was
very similar.

440 MISQuarterlyVol.27 No. 3/September2003

 Venkateshet al./UserAcceptance of IT

Tableg 6. Studys2: Predicting Intention inMandatorySettins

Time 1 (N = 96) Time 2 (N = 96) Time 3 (N = 96)

Models Independent variables R2 Beta R2 Beta R2 Beta
TRA Attitude toward using tech. .26 .27*** .26 .28*** .17 .40***
Subjectivenorm .20** .21** .05
TAM/ Perceived usefulness .39 .42*** .41 .50*** .36 .60***
TAM2 Perceived ease of use .21* .23** .03
Subjectivenorm .20* .03 .04
MM Extrinsicmotivation .38 .47*** .40 .49*** .35 .44***
Intrinsicmotivation .21** .24** .19**
TPB/ Attitudetowardusing tech. .34 .22* .28 .36*** .18 .43***
DTPB Subjectivenorm .25*** .26** .05
Perceived behavioralcontrol .19* .03 .08
C-TAM- Perceived usefulness .36 .42*** .35 .51*** .35 .60***
TPB Attitudetowardusing tech. .07 .08 .04
Subjectivenorm .20* .23** .03
Perceived behavioralcontrol .19* .11 .09
MPCU Job-fit .37 .42*** .40 .50*** .37 .61***
Complexity(reversed) .20* .02 .04
Long-termconsequences .07 .07 .07
Affecttowarduse .01 .05 .04
Social factors .18* .23** .02
Facilitatingconditions .05 .07 .07
IDT Relativeadvantage .38 .47*** .42 .52*** .37 .61**
Ease of use .20* .04 .04
Result demonstrability .03 .07 .04
Trialability .05 .04 .04
Visibility .04 .04 .01
Image .18* .27** .05
Compatibility .06 .02 .04
Voluntariness .02 .06 .03
SCT Outcomeexpectations .38 .46*** .39 .44*** .36 .60***
Self-efficacy .19** .21*** .03
Affect .06 .04 .05
Anxiety -. 18* -. 16* .02

Notes: 1. *p < .05; **p < .01; ***p< .001.

2. When the data were analyzed separately for studies 2a and 2b, the patternof results was
very similar.

MIS QuarterlyVol.27 No. 3/September2003 441

Venkateshet al./UserAcceptance of IT

7 e tn I0n - 0 - *O m a
0dlC I..
0s0n0ig D
0
PoldAcrssSude ( 65

Model Version Independent Variables R2 Beta Explanation

TRA 1 Attitude(A) .36 .41*** Directeffect
Subjective norm (SN) .11
Experience (EXP) .09
Voluntariness (VOL) .04
A x EXP .03
SN x EXP -.17* Effectdecreases withincreasing
experience
SN x VOL .17* Effectpresent only in mandatory
settings
TAM 2a Perceived usefulness (U) .53 .48*** Directeffect
TAM2 Perceived ease of use (EOU) .11
Subjective norm (SN) .09
Experience (EXP) .06
Voluntariness (VOL) .10
EOU x EXP -.20** Effectdecreases with increasing
experience
SN x EXP -.15
SN x VOL -.16* Cannotbe interpreteddue to
presence of higher-order term
EXP x VOL .07
SN x EXP x VOL -.18** Effectexists only in mandatory
settings but decreases with
increasing experience
2b Perceived usefulness (U) .52 .14* Cannot be interpreteddue to
TAM presence of interaction term
incl. Percd. ease of use (EOU) .08
gender Subjective norm (SN) .02
Gender (GDR) .11
Experience (EXP) .07
U x GDR .31*** Effectis greaterfor men
EOU x GDR -.20** Effect is greater for women
SN x GDR .11
SN x EXP .02
EXP x GDR .09
SN x GDR x EXP .17** Effectis greaterfor women but
decreases with increasing
experience

442 MISQuarterlyVol.27 No. 3/September2003

 Venkateshet al./UserAcceptance of IT

Pooled.Across Studies (N = 64 5
(CoS n
Sined

Model Version Independent Variables R2 Beta Explanation

MM 3 Extrinsicmotivation .38 .50*** Directeffect
Intrinsic motivation .20*** Direct effect
TPB/ 4a Attitude(A) .36 .40*** Directeffect
DTPB TPB Subjectivenorm(SN) .09
incl. vol Percd. behrl.control .13
(PBC)
Experience(EXP) .10
Voluntariness(VOL) .05
SN x EXP -.17** Effectdecreases withincreasing
experience
SN x VOL .17** Effectpresent only in mandatory
settings
4b Attitude(A) .46 .17*** Cannotbe interpreteddue to
TPB presence of interaction term
incl. Subjectivenorm(SN) .02
gender Percd. behrl.control(PBC) .10
Gender(GDR) .01
Experience(EXP) .02
A x GDR .22*** Effectis greaterfor men
SN x EXP -.12
SN x GDR .10
PBC x GDR .07
PBC x EXP .04
GDR x EXP .15* Termincludedto test higher-
orderinteractionsbelow
SN x GDR x EXP -.18** Both SN and PBC effects are
higherfor women, butthe effects
PBC x GDR x EXP -.16* decrease withincreasing
experience
4c Attitude(A) .47 .17*** Cannotbe interpreteddue to
TPB presence of interaction term
incl. Subjective norm (SN) .02
age Percd. behrl. control (PBC) .10
Age (AGE) .01
Experience (EXP) .02
A x AGE -.26*** Effectis greaterfor younger
workers
SN x EXP -.03
SN x AGE .11
PBC x AGE .21** Effectis greaterfor older
workers

MIS QuarterlyVol.27 No. 3/September2003 443

Venkateshet al./UserAcceptance of IT

Tabe. paisn ncudn Mdeatrs Dt

a 1 &
Peicignenin-oe. l Co.m
Model Version Independent Variables R2 Beta Explanation
AGE x EXP .15* Termincludedto test higher-
orderinteractionbelow
SN x AGE x EXP -.18** Effectis greaterfor older
workers,but the effect
decreases withincreasing
experience
C- 5 Perceived usefulness (U) .39 .40*** Directeffect
TAM- Attitude (A) .09
TP3B Subjective norm (SN) .08
Perceived beholdercontrol .16* Cannotbe interpreteddue to
(PBC) presence of interaction term
Experience (EXP) .11
U x EXP .01
A x EXP .08
SN x EXP -.17* Effectdecreases withincreasing
experience
PBC x EXP -.19** Effectdecreases withincreasing
experience
MPCU 6 Job-fit(JF) .47 .40*** Directeffect
Complexity (CO) (reversed) .07
Long-termconsequences (LTC) .02
Affect toward use (ATU) .05
Social factors (SF) .10
Facilitating condns. (FC) .07
Experience (EXP) .08
CO x EXP(CO-reversed) -.17* Effectdecreases withincreasing
experience
LTC x EXP .02
ATU x EXP .01
SF x EXP -.20** Effectdecreases withincreasing
experience
FC x EXP .05
IDT 7 Relativeadvantage (RA) .40 .49*** Directeffect
Ease of use (EU) .05
Result demonstrability (RD) .02
Trialability(T) .04
Visibility (V) .03
Image (1) .01
Compatibility (COMPAT) .06
Voluntariness of use (VOL) .11

444 MISQuarterlyVol.27 No. 3/September2003


Tabl-7.Preictng ntetio
q'.s1PoldAcrs ~iI~~T3i~~
i,o,hr
Suie p1j
I IFI I{Oz.evr~
N ri rr
.
45)(otiud
Model Version Independent Variables
Experience (EXP)
EU x EXP
RD x EXP
V x EXP
I x EXP
Voluntariness of use (VOL)
SCT 8 Outcome expectations
Self-efficacy
Affect
Anxiety
Notes: 1. *p < .05; **p< .01; ***p< .001.
2. The significanceof the interactionterms was also verifiedusing Chow'stest.
same individuals,resultinginmeasurementerrors
that are potentially correlated across time.
However,cross-sectional analysis using Chow's
(1960) test of beta differences(p < .05) fromeach
time period (not shown here) confirmed the
patternof results shown in Table 7. Those beta
differenceswitha significanceof p < .05 or better
(when using Chow's test) are discussed in the
"Explanation" columnin Table 7. The interaction
terms were modeled as suggested by Chinet al.
(1996) by creatinginteractionterms that were at
the level of the indicators. For example, if latent
variableA is measured by fourindicators(Al, A2,
A3, and A4) and latentvariableB is measured by
three indicators(B1, B2, and B3), the interaction
termA x B is specified by 12 indicators,each one
a productterm-i.e., Al x B1, Al x B2, Al x B3,
A2 x Bl, etc.
Withthe exceptionof MMand SCT, the predictive
validityof the models increased afterincludingthe
moderatingvariables. Forinstance, the variance
explained by TAM2increased to 53 percent and
TAMincluding gender increased to 52 percent
when compared to approximately35 percent in
Venkateshet al./UserAcceptance of IT
-Moel C mpa 6so0Inludng Sodeatos: Dat
R2 Beta Explanation
.03
-.16* Effectdecreases withincreasing
experience
.03
-.14* Effectdecreases withincreasing
experience
-.14* Effectdecreases withincreasing
experience
.05
.36 .44*** Directeffect
.18* Directeffect
.01
-.15* Direct effect
cross-sectionaltests of TAM(withoutmoderators).
The explained variance of TRA, TPB/DTPB,
MPCU,and IDTalso improved. Foreach model,
we have only included moderators previously
tested in the literature.For example, in the case
of TAM and its variations, the extensive prior
empiricalworkhas suggested a largernumberof
moderatorswhen compared to moderatorssug-
gested for other models. This in turn may have
unintentionallybiased the results and contributed
to the high variance explained in TAM-related
models when compared to the other models.
Regardless, it is clear that the extensions to the
various models identified in previous research
mostly enhance the predictive validity of the
variousmodels beyondthe originalspecifications.
In looking at technology use as the dependent
variable,in additionto intentionas a key predictor,
TPB and DTPB employ perceived behavioral
controlas an additionalpredictor.MPCUemploys
facilitatingconditions, a constructsimilarto per-
ceived behavioral control, to predict behavior.
Thus, intentionand perceived behavioralcontrol
were used to predictbehaviorin the subsequent
MISQuarterlyVol. 27 No. 3/September2003 445
Venkateshet al./UserAcceptance of IT

Table PredictingUa hai=

8.-
Use12 Use23 Use3
Independent Variables R2 Beta R2 Beta R2 Beta
Studies Behavioralintentionto use (BI) .37 .61*** .36 .60*** .39 .58***
la and lb
(voluntary) Perceived behavioralcontrol(PBC) .04 .06 .17*
(N=119)
Studies Behavioralintentionto use (BI) .35 .58*** .37 .61*** .39 .56***
2a and 2b
(mandatory) Perceived behavioralcontrol(PBC) .07 .07 .20*
(N = 96)

Notes: 1. BI,PBC measured at T1 were used to predictusage betweentime periods 1 and 2 (denoted
Use12);BI, PBC measured at T2 were used to predictusage between time periods 2 and 3
(Use23);BI, PBC measured at T3 were used to predictusage between time periods 3 and 4
(Use3).
2. *p < .05; **p < .01; ***p< .001.

time period: intentionfromT1 was used to predict
usage behaviormeasured betweenT1and T2 and
so on (see Table 8). Since intentionwas used to
predictactual behavior,concerns associated with
the employmentof subjectivemeasures of usage
do not apply here (see Straub et al. 1995). In
additionto intentionbeing a predictorof use, per-
ceived behavioral control became a significant
directdeterminantof use overand above intention
withincreasingexperience (at T3) indicatingthat
continued use could be directly hindered or
fostered by resources and opportunities.A nearly
identical patternof results was found when the
data were analyzed using facilitatingconditions
(from MPCU) in place of perceived behavioral
control(the specific results are not shown here).
TRA and TPB/DTPB, perceived usefulness in
TAM/TAM2 and C-TAM-TPB, extrinsicmotivation
in MM,job-fitin MPCU,relativeadvantage in IDT,
and outcome expectations in SCT. Second,
several other constructswere initiallysignificant,
but then became nonsignificantover time, in-
cludingperceivedbehavioralcontrolinTPB/DTPB
and C-TAM-TPB, perceived ease of use in TAM/
TAM2,complexityin MPCU,ease of use in IDT,
and self-efficacyand anxietyin SCT. Finally,the
voluntary vs. mandatory context did have an
influenceon the significanceof constructsrelated
to social influence:subjective norm (TPB/DTPB,
C-TAM-TPBand TAM2),social factors (MPCU),
and image (IDT) were only significant in
mandatoryimplementations.

Having reviewed and empiricallycompared the

eight competing models, we now formulate a
unifiedtheory of acceptance and use of techno-
logy (UTAUT).Towardthisend, we examine com-
monalitiesacross models as a firststep. Tables
5, 6, 7, and 8 presented cross-sectional tests of
the baseline models and theirextensions. Several
consistent findings emerged. First, for every
model, there was at least one constructthat was
significantin all time periods and that construct
also had the strongest influence-e.g., attitudein
Formulationof the Unified
Theoryof Acceptance and
Use of Technology
(UTAUT)
Seven constructsappearedto be significantdirect
determinantsof intentionor usage in one or more
of the individualmodels (Tables 5 and 6). Of
these, we theorizethat fourconstructswillplay a

446 MISQuarterlyVol.27 No. 3/September2003


Performance
Expectancy
Effort
Expectancy
ExpectancyBehavioral
Social
Influence
Facilitating
Conditions
Gender
Age
oe
- -ue .Rseac
significant role as direct determinants of user
acceptance and usage behavior: performance
expectancy, effort expectancy, social influence,
and facilitatingconditions. As will be explained
below, attitude toward using technology, self-
efficacy,and anxietyare theorizednotto be direct
determinantsof intention.The labels used forthe
constructsdescribe the essence of the construct
and are meantto be independentof any particular
theoreticalperspective. In the remainderof this
section, we define each of the determinants,
specify the role of key moderators(gender, age,
voluntariness,and experience), and providethe
theoretical justification for the hypotheses.
Figure3 presents the research model.
PerformanceExpectancy
Performanceexpectancy is definedas the degree
to whichan individualbelieves that using the sys-
tem will help him or her to attain gains in job
Venkateshet al./UserAcceptanceof IT
Use
Intention , Behavior
Experience
Voluntariness
ou s
of Use
performance. The five constructs from the dif-
ferent models that pertain to performance
expectancyare perceivedusefulness (TAM/TAM2
and C-TAM-TPB), extrinsicmotivation(MM),job-fit
(MPCU), relative advantage (IDT),and outcome
expectations (SCT). Even as these constructs
evolved in the literature,some authors acknowl-
edged theirsimilarities:usefulness and extrinsic
motivation(Davis et al. 1989, 1992), usefulness
andjob-fit(Thompsonet al. 1991), usefulness and
relativeadvantage (Davis et al. 1989; Mooreand
Benbasat 1991; Plouffe et al. 2001), usefulness
and outcomeexpectations(Compeauand Higgins
1995b;Daviset al. 1989), andjob-fitand outcome
expectations (Compeauand Higgins1995b).
The performance expectancy construct within
each individualmodel (Table 9) is the strongest
predictorof intentionand remainssignificantat all
pointsof measurementinbothvoluntaryand man-
datory settings (Tables 5, 6, and 7), consistent
with previous model tests (Agarwaland Prasad
2003
Vol.27 No.3/September
MISQuarterly 447
Venkateshet al./UserAcceptance of IT
3.
Table 9 .PrfraneExecacy *oo Cntrcs
Construct Definition
Perceived The degree to which a
Usefulness person believes that using
(Davis 1989; Davis et a particularsystem would
al. 1989) enhance his or her job
performance.
ExtrinsicMotivation The perceptionthat users
(Davis et al. 1992) willwant to performan
activitybecause it is per-
ceived to be instrumental
in achievingvalued out-
comes that are distinct
fromthe activityitself, such
as improvedjob perfor-
mance, pay, or promotions
Job-fit Howthe capabilitiesof a
(Thompsonet al. system enhance an indi-
1991) vidual'sjob performance.
448 MISQuarterlyVol.27 No. 3/September2003
, eintosandScae
Items
1. Using the system in myjob would
enable me to accomplishtasks more
quickly.
2. Using the system would improvemy job
performance.
3. Using the system in my job would
increase my productivity.
4. Using the system would enhance my
effectiveness on the job.
5. Using the system would make it easier
to do myjob.
6. I wouldfindthe system useful in myjob.
Extrinsicmotivationis operationalizedusing
the same items as perceived usefulness
fromTAM(items 1 through6 above).
1. Use of the system willhave no effect on
the performanceof my job (reverse
scored).
2. Use of the system can decrease the
time needed for my importantjob
responsibilities.
3. Use of the system can significantly
increase the qualityof outputon myjob.
4. Use of the system can increase the
effectiveness of performingjob tasks.
5. Use can increase the quantityof output
for the same amountof effort.
6. Consideringall tasks, the general extent
to which use of the system could assist
on the job. (differentscale used for this
item).
 Venkateshet al./UserAcceptance of IT

3.
Table 6.ProraceEpetny:RotC nsrcs , eintosandScae

(Continued)

Construct Definition Items

RelativeAdvantage The degree to which using 1. Using the system enables me to
(Mooreand Benbasat an innovationis perceived accomplishtasks more quickly.
1991) as being betterthan using 2. Using the system improvesthe qualityof
its precursor. the workI do.
3. Using the system makes it easier to do
myjob.
4. Using the system enhances my
effectiveness on the job.
5. Using the system increases my
productivity.
Outcome Outcome expectations IfI use the system...
Expectations relate to the consequences 1. I willincrease my effectiveness on the
(Compeauand of the behavior. Based on job.
Higgins 1995b; empiricalevidence, they 2. I willspend less time on routinejob
Compeau et al. 1999) were separated into per- tasks.
formanceexpectations 3. I willincrease the qualityof outputof my
(job-related)and personal job.
expectations (individual 4. I willincrease the quantityof outputfor
goals). For pragmatic the same amountof effort.
reasons, fourof the highest 5. Mycoworkerswillperceive me as
loadingitems fromthe competent.
performanceexpectations 6. I willincrease my chances of obtaininga
and three of the highest promotion.
loadingitems fromthe 7. I willincrease my chances of getting a
personal expectations raise.
were chosen fromCom-
peau and Higgins(1995b)
and Compeau et al. (1999)
for inclusionin the current
research. However,our
factoranalysis showed the
two dimensions to load on
a single factor.

1998; Compeau and Higgins 1995b; Davis et al.
1992; Taylorand Todd 1995a; Thompson et al.
1991;Venkateshand Davis2000). However,from
a theoretical point of view, there is reason to
expect thatthe relationshipbetween performance
expectancy and intentionwill be moderated by
genderand age. Research on gender differences
indicatesthat men tend to be highlytask-oriented
(Mintonand Schneider 1980) and, therefore,per-
formance expectancies, which focus on task
accomplishment,are likelyto be especially salient
to men. Gender schema theory suggests that
such differences stem from gender roles and
socialization processes reinforced from birth
ratherthan biologicalgender per se (Bem 1981;
Bem and Allen 1974; Kirchmeyer1997; Lubinski
et al. 1983; Lynottand McCandless2000; Moto-
widlo1982). Recent empiricalstudies outside the

MISQuarterlyVol.27 No. 3/September2003 449

 et al./UserAcceptanceof IT
Venkatesh
ITcontext (e.g., Kirchmeyer2002; Twenge 1997)
have shown that gender roles have a strong
psychological basis and are relativelyenduring,
yet open to change over time (see also Ashmore
1990; Eichingeret al. 1991; Feldmanand Aschen-
brenner1983; Helson and Moane 1987).
Similar to gender, age is theorized to play a
moderating role. Research on job-related atti-
tudes (e.g., Halland Mansfield1975; Porter1963)
suggests that younger workers may place more
importanceon extrinsicrewards. Genderand age
differences have been shown to exist in techno-
logyadoptioncontextsalso (MorrisandVenkatesh
2000; Venkatesh and Morris2000). Inlookingat
gender and age effects, it is interestingto note
that Levy (1988) suggests that studies of gender
differencescan be misleadingwithoutreferenceto
age. Forexample, given traditionalsocietal gen-
der roles, the importanceof job-relatedfactors
may change significantly (e.g., become sup-
planted by family-orientedresponsibilities) for
workingwomen between the time that they enter
the labor force and the time they reach child-
rearingyears (e.g., Barnettand Marshall1991).
Thus,we expect thatthe influenceof performance
expectancywillbe moderatedby bothgender and
age.
Hl: The influence of performance ex-
pectancy on behavioralintentionwill
be moderated by gender and age,
such that the effect willbe stronger
formen and particularlyforyounger
men.
EffortExpectancy
Effortexpectancy is definedas the degree of ease
associated withthe use of the system. Threecon-
structs from the existing models capture the
concept of effortexpectancy: perceived ease of
use (TAM/TAM2),
complexity(MPCU),and ease
of use (IDT).As can be seen in Table 10, there is
substantialsimilarityamong the constructdefini-
tions and measurement scales. The similarities
among these constructshave been noted in prior
research (Davis et al. 1989; Mooreand Benbasat
1991; Plouffeet al. 2001; Thompsonet al. 1991).
450 MISQuarterly
Vol.27 No.3/September
2003
The effortexpectancyconstructwithineach model
(Table 10) is significant in both voluntaryand
mandatoryusage contexts;however,each one is
significantonly duringthe firsttime period (post-
training, T1), becoming nonsignificant over
periods of extended and sustained usage (see
Tables 5, 6, and 7), consistent with previous
research (e.g., Agarwaland Prasad 1997, 1998;
Davis et al. 1989; Thompson et al. 1991, 1994).
Effort-orientedconstructs are expected to be be
more salient inthe earlystages of a new behavior,
when process issues represent hurdles to be
overcome, and later become overshadowed by
instrumentalityconcerns (Davis et al. 1989;
Szajna 1996; Venkatesh 1999).
Venkateshand Morris(2000), drawinguponother
research (e.g., Bem and Allen 1974; Bozionelos
1996), suggest that effort expectancy is more
salient for womenthan formen. As noted earlier,
the gender differences predicted here could be
drivenby cognitionsrelatedto gender roles (e.g.,
Lynottand McCandless 2000; Motowidlo1982;
Wong et al. 1985). Increased age has been
shown to be associated with difficultyin pro-
cessing complexstimuliandallocatingattentionto
informationon the job (Plude and Hoyer 1985),
both of which may be necessary when using
software systems. Priorresearch supports the
notionthatconstructsrelatedto effortexpectancy
willbe strongerdeterminantsof individuals'inten-
tion for women (Venkatesh and Morris 2000;
Venkatesh et al. 2000) and for older workers
(Morrisand Venkatesh 2000). Drawingfromthe
arguments made in the context of performance
expectancy, we expect gender, age, and exper-
ience to workin concert (see Levy 1988). Thus,
we propose that effort expectancy will be most
salientforwomen,particularly those who are older
and with relatively little experience with the
system.
H2: The influence of effort expectancy
on behavioral intention will be
moderated by gender, age, and
experience, such thatthe effect will
be strongerfor women, particularly
younger women, and particularlyat
early stages of experience.
 Venkateshet al./UserAcceptance of IT

Tabe 1.IEfot Epecanc: SootCostrcs efntosandScae

Construct Definition Items

Perceived Ease of Use The degree to whicha 1. Learningto operate the system would be
(Davis 1989; Davis et person believes that easy for me.
al. 1989) using a system would be 2. I wouldfind it easy to get the system to
free of effort. do what I want it to do.
3. Myinteractionwiththe system would be
clear and understandable.
4. I wouldfindthe system to be flexibleto
interactwith.
5. Itwould be easy for me to become
skillfulat using the system.
6. Iwouldfindthe system easy to use.
Complexity The degree to whicha 1. Using the system takes too much time
(Thompsonet al. 1991) system is perceived as frommy normalduties.
relativelydifficultto 2. Workingwiththe system is so
understandand use. complicated,it is difficultto understand
what is going on.
3. Using the system involvestoo much time
doing mechanicaloperations(e.g., data
input).
4. Ittakes too long to learn how to use the
system to make it worththe effort.
Ease of Use The degree to which 1. Myinteractionwiththe system is clear
(Mooreand Benbasat using an innovationis and understandable.
1991) perceived as being 2. I believe that it is easy to get the system
difficultto use. to do what I want it to do.
3. Overall,I believe that the system is easy
to use.
4. Learningto operate the system is easy
for me.

Social Influence the explicitor implicitnotionthat the individual's

Social influenceis definedas the degree to which
an individual perceives that important others
believe he or she should use the new system.
Social influenceas a directdeterminantof behav-
ioralintentionis representedas subjectivenormin
TRA,TAM2,TPB/DTPBand C-TAM-TPB,social
factors in MPCU,and image in IDT.Thompsonet
al. (1991) used the term social norms in defining
their construct,and acknowledge its similarityto
subjective norm within TRA. While they have
differentlabels, each of these constructscontains
behavior is influenced by the way in which they
believe others willview them as a resultof having
used the technology. Table 11 presents the three
constructs related to social influence:subjective
norm(TRA,TAM2,TPB/IDTPB, and C-TAM-TPB),
social factors (MPCU),and image (IDT).
The currentmodel comparison(Tables 5, 6, and
7) foundthatthe social influenceconstructslisted
above behave similarly. None of the social influ-
ence constructs are significantin voluntarycon-
texts; however, each becomes significant when

MIS QuarterlyVol.27 No. 3/September2003 451

Venkateshet al./UserAcceptance of IT

T 0l 1 o Sl0I 0fec R ot ns Defini

0ucts, n d c

Construct Definition Items

Subjective Norm
(Ajzen 1991; Davis et al.
1989; Fishbeinand Azjen
1975; Mathieson1991;
Taylorand Todd 1995a,
1995b)
Social Factors
(Thompsonet al. 1991)
Image
(Mooreand Benbasat 1991)
The person's perception
that most people who are
importantto himthinkhe
should or should not
performthe behaviorin
question.
The individual'sinter-
nalizationof the reference
group'ssubjectiveculture,
and specific interpersonal
agreements that the indivi-
dual has made withothers,
in specific social situations.
The degree to which use of
an innovationis perceived
to enhance one's image or
status in one's social
system.
1. People who influencemy
behaviorthinkthat I should use
the system.
2. People who are importantto me
thinkthat I should use the
system.
1. I use the system because of the
proportionof coworkerswho use
the system.
2. The senior management of this
business has been helpfulin the
use of the system.
3. Mysupervisoris very supportive
of the use of the system for my
job.
4. Ingeneral, the organizationhas
supported the use of the system.
1. People in my organizationwho
use the system have more
prestige than those who do not.
2. People in my organizationwho
use the system have a high
profile.
3. Havingthe system is a status
symbol in my organization.

use is mandated. Venkatesh and Davis (2000)
suggested that such effects could be attributed
to compliance in mandatory contexts that
causes social influences to have a directeffect
on intention; in contrast, social influence in
voluntarycontexts operates by influencingper-
ceptions about the technology-the mech-
anisms at play here are internalizationand
identification. In mandatory settings, social
influence appears to be importantonly in the
early stages of individualexperience with the
technology, with its role eroding over time and
eventually becoming nonsignificantwith sus-
tained usage (T3), a patternconsistent withthe
observations of Venkatesh and Davis (2000).
The role of social influence in technology
acceptance decisions is complexand subjectto
a wide range of contingent influences. Social
influence has an impacton individualbehavior
throughthree mechanisms: compliance, inter-
nalization,and identification(see Venkateshand
Davis 2000; Warshaw 1980). While the latter
two relateto alteringan individual'sbelief struc-
ture and/orcausing an individualto respond to
potential social status gains, the compliance
mechanismcauses an individualto simplyalter
his or her intentionin response to the social
pressure-i.e., the individualintends to comply
with the social influence. Priorresearch sug-
gests that individualsare more likelyto comply

452 MISQuarterlyVol.27 No. 3/September2003


with others' expectations when those referent
others have the abilityto reward the desired
behavior or punish nonbehavior(e.g., French
and Raven 1959; Warshaw1980). This view of
compliance is consistent with results in the
technologyacceptance literatureindicatingthat
relianceon others'opinionsis significantonly in
mandatorysettings (Hartwickand Barki1994),
particularlyin the early stages of experience,
when an individual'sopinions are relativelyill-
informed(Agarwaland Prasad 1997; Hartwick
and Barki1994; Karahannaet al. 1999; Taylor
and Todd 1995a; Thompsonet al. 1994; Venka-
tesh and Davis 2000). This normativepressure
will attenuate over time as increasing exper-
ience providesa more instrumental(ratherthan
social) basis for individualintentionto use the
system.
Theory suggests that women tend to be more
sensitive to others' opinions and thereforefind
social influenceto be moresalientwhen forming
an intentionto use new technology(Miller1976;
Venkateshet al. 2000), withthe effect declining
with experience (Venkatesh and Morris2000).
As in the case of performanceand effortexpec-
tancies, gender effects may be driven by psy-
chologicalphenomenaembodiedwithinsocially-
constructed gender roles (e.g., Lubinskiet al.
1983). Rhodes' (1983) meta-analyticreviewof
age effects concluded that affiliationneeds in-
crease withage, suggesting that older workers
are more likelyto place increased salience on
social influences, withthe effect decliningwith
experience (Morris and Venkatesh 2000).
Therefore,we expect a complex interactionwith
these moderatingvariablessimultaneouslyinflu-
encing the social influence-intentionrelation-
ship.
H3: The influence of social influence
on behavioral intention will be
moderated by gender, age,
voluntariness, and experience,
such that the effect will be
stronger for women, particularly
older women, particularlyin man-
datorysettings in the earlystages
of experience.
Venkateshet al./UserAcceptanceof IT
FacilitatingConditions
Facilitatingconditionsare definedas the degree
to which an individualbelieves that an organi-
zational and technical infrastructureexists to
supportuse of the system. This definitioncap-
tures concepts embodiedby threedifferentcon-
structs: perceived behavioral control (TPBI
DTPB, C-TAM-TPB),facilitating conditions
(MPCU),and compatibility(IDT). Each of these
constructsis operationalizedto includeaspects
of the technologicaland/ororganizationalenvi-
ronmentthatare designed to removebarriersto
use (see Table 12). Taylorand Todd (1995b)
acknowledged the theoretical overlap by
modelingfacilitatingconditionsas a core com-
ponent of perceived behavioral control in
TPB/DTPB. The compatibilityconstruct from
IDTincorporatesitems that tap the fit between
the individual'swork style and the use of the
system in the organization.
The empiricalevidence presented in Tables 5,
6, 7, and 8 suggests that the relationships
between each of the constructs (perceived
behavioral control, facilitatingconditions, and
compatibility)and intentionare similar. Specifi-
cally, one can see that perceived behavioral
controlis significantin bothvoluntaryand man-
datory settings immediatelyfollowingtraining
(T1), but that the construct's influence on
intention disappears by T2. It has been
demonstratedthat issues relatedto the support
infrastructure-a core concept withinthe facili-
tatingconditionsconstruct-are largelycaptured
withinthe effortexpectancyconstructwhichtaps
the ease with which that tool can be applied
(e.g., Venkatesh 2000). Venkatesh (2000)
found supportfor full mediationof the effect of
facilitating conditions on intention by effort
expectancy. Obviously,if effortexpectancy is
not present in the model (as is the case with
TPB/DTPB),then one would expect facilitating
conditions to become predictive of intention.
Ourempiricalresults are consistent withthese
arguments. For example, in TPB/DTPB,the
construct is significant in predictingintention;
however, in other cases (MPCUand IDT),it is
nonsignificantin predictingintention. Inshort,
MISQuarterly
Vol.27 No.3/September
2003 453
 et al./UserAcceptanceof IT
Venkatesh

I Se
12.
TaS Fai it Sn C tos S Sto r s De io ad l

Construct Definition Items

Perceived Behavioral
Control
(Ajzen 1991; Taylorand
Todd 1995a, 1995b)
FacilitatingConditions
(Thompsonet al. 1991)
Compatibility
(Mooreand Benbasat
1991)
Reflects perceptionsof
internaland external
constraintson behavior
and encompasses self-
efficacy, resource facili-
tatingconditions,and
technologyfacilitating
conditions.
Objectivefactors in the
environmentthat
observers agree make
an act easy to do,
includingthe provision
of computersupport.
The degree to whichan
innovationis perceived
as being consistent with
existing values, needs,
and experiences of
potential adopters.
1. I have controlover using the system.
2. I have the resources necessary to use
the system.
3. I have the knowledgenecessary to
use the system.
4. Giventhe resources, opportunities
and knowledgeit takes to use the
system, it wouldbe easy for me to use
the system.
5. The system is not compatiblewith
other systems I use.
1. Guidancewas availableto me in the
selection of the system.
2. Specialized instructionconcerningthe
system was availableto me.
3. A specific person (or group)is
availablefor assistance withsystem
difficulties.
1. Using the system is compatiblewith
all aspects of my work.
2. I thinkthat using the system fits well
withthe way I liketo work.
3. Using the system fits into my work
style.

when both performanceexpectancy constructs intention). In fact, the effect is expected to

and effort expectancy constructs are present,
facilitatingconditions becomes nonsignificantin
predictingintention.
H4a: Facilitatingconditions will not
have a significantinfluence on
behavioralintention.5
The empiricalresults also indicatethatfacilitating
conditions do have a direct influence on usage
beyond that explained by behavioral intentions
alone (see Table 8). ConsistentwithTPB/DTPB,
facilitatingconditionsare also modeledas a direct
antecedent of usage (i.e., not fully mediated by
increase withexperience as users of technology
find multiple avenues for help and support
throughoutthe organization,thereby removing
impedimentsto sustained usage (Bergeronet al.
1990). Organizationalpsychologists have noted
that older workers attach more importance to
receivinghelpand assistance on the job (e.g., Hall
and Mansfield1975). This is furtherunderscored
in the context of complex IT use given the
increasingcognitiveand physicallimitationsasso-
ciated withage. These argumentsare in linewith
empiricalevidence from Morrisand Venkatesh
(2000). Thus,when moderatedby experience and
age, facilitatingconditionswill have a significant
influenceon usage behavior.

H4b: Theinfluenceof facilitatingcon-

5Totest the nonsignificant we performa
relationship, ditionson usage willbe mode-
poweranalysisinthe resultssection.

454 Vol.27 No.3/September

MISQuarterly 2003

rated by age and experience,
such that the effect will be
strongerforolder workers,par-
ticularlywithincreasing exper-
ience.
Constructs TheorizedNot to Be
DirectDeterminantsof Intention
Althoughself-efficacyand anxietyappearedto be
significantdirectdeterminantsof intentionin SCT
(see Tables 5 and 6), UTAUTdoes not include
them as directdeterminants. Previous research
(Venkatesh 2000) has shown self-efficacy and
anxietyto be conceptuallyand empiricallydistinct
from effort expectancy (perceived ease of use).
Self-efficacyand anxiety have been modeled as
indirectdeterminantsof intentionfullymediatedby
perceivedease of use (Venkatesh2000). Consis-
tent with this, we found that self-efficacy and
anxiety appear to be significantdeterminantsof
intentionin SCT-i.e., withoutcontrollingfor the
effect of effortexpectancy. We thereforeexpect
self-efficacyand anxietyto behave similarly,that
is, to be distinct from effort expectancy and to
have no direct effect on intention above and
beyond effortexpectancy.
H5a: Computerself-efficacy willnot
have a significantinfluence on
behavioralintention.6
H5b: Computeanxiety willnot have
a significantinfluenceon beha-
vioralintention.7
Attitudetowardusing technology is defined as an
individual'soverall affective reaction to using a
system. Fourconstructsfromthe existingmodels
align closely with this definition:attitudetoward
behavior(TRA,TPB/DTPB,C-TAM-TPB), intrinsic
6Totest the nonsignificant we performa
relationship,
poweranalysisinthe resultssection.
7Totest the nonsignificant we performa
relationship,
poweranalysisinthe resultssection.
et al./UserAcceptanceof IT
Venkatesh
motivation8(MM),affecttowarduse (MPCU),and
affect (SCT). Table 13 presents the definitions
and associated scale items for each construct.
Each constructhas a componentassociated with
generalizedfeeling/affectassociated witha given
behavior (in this case, using technology). In
examiningthese fourconstructs,it is evidentthat
they all tap into an individual'sliking,enjoyment,
joy, and pleasureassociated withtechnologyuse.
Empirically,the attitude constructs present an
interestingcase (see Tables 5, 6, and 7). Insome
cases (e.g., TRA, TPB/DTPB, and MM), the
attitude construct is significantacross all three
time periodsand is also the strongest predictorof
behavioralintention. However,in othercases (C-
TAM-TPB,MPCU,and SCT), the constructwas
not significant. Upon closer examination, the
attitudinalconstructs are significant only when
specific cognitions-in this case, constructs
relatedto performanceand effortexpectancies-
are not includedin the model. There is empirical
evidence to suggest thataffectivereactions(e.g.,
intrinsicmotivation)may operate through effort
expectancy (see Venkatesh2000). Therefore,we
consider any observed relationship between
attitudeand intentionto be spuriousand resulting
from the omission of the other key predictors
(specifically, performance and effort expec-
tancies). This spurious relationshiplikelystems
fromthe effect of performanceand effortexpec-
tancies on attitude(see Davis et al. 1989). The
non-significance of attitude in the presence of
such other constructs has been reportedin pre-
vious model tests (e.g., Taylorand Todd 1995a;
Thompson et al. 1991), despite the fact that this
findingis counterto what is theorizedin TRAand
TPB/DTPB.Giventhatwe expect strongrelation-
ships inUTAUTbetween performanceexpectancy
and intention,and between effortexpectancy and
intention,we believe that,consistentwiththe logic
developed here, attitudetowardusing technology
8Some perspectivesdifferon the role of intrinsic
Forexample,Venkatesh(2000)modelsitas
motivation.
a determinantof perceived ease of use (effort
expectancy).However,inthe motivational model,it is
shownas a directeffecton intentionand is shownas
suchhere.
Vol.27 No.3/September
MISQuarterly 2003 455
Venkateshet al./UserAcceptance of IT

Tale1. Attitud Toward sin eholoy:Rot5ontrctefniiosSn

Construct Definition Items

AttitudeToward
Behavior
(Davis et al. 1989;
Fishbeinand Ajzen
1975; Taylorand Todd
1995a, 1995b)
IntrinsicMotivation
(Davis et al. 1992)
AffectTowardUse
(Thompsonet al. 1991)
Affect
(Compeauand Higgins
1995b; Compeau et al.
1999)
An individual'spositiveor
negative feelings about
performingthe target
behavior.
The perceptionthat users
willwant to performan
activityfor no apparent
reinforcementother than
the process of performing
the activityper se.
Feelings of joy, elation, or
pleasure;or depression,
disgust, displeasure, or
hate associated by an
individualwitha particular
act.
An individual'slikingof
the behavior.
1. Using the system is a bad/good idea.
2. Using the system is a foolish/wise idea.
3. I dislike/likethe idea of using the
system.
4. Using the system is unpleasant/
pleasant.
1. I find using the system to be enjoyable
2. The actual process of using the system
is pleasant.
3. I have fun using the system.
1. The system makes workmore
interesting.
2. Workingwiththe system is fun.
3. The system is okay for some jobs, but
not the kindof job Iwant. (R)
1. I likeworkingwiththe system.
2. I look forwardto those aspects of my
job that requireme to use the system.
3. Using the system is frustratingfor me.
(R)
4. Once I startworkingon the system, I
find it hardto stop.
5. I get bored quicklywhen using the
system. (R)

will not have a direct or interactiveinfluence on Behavioral Intention

intention.
Consistentwiththe underlyingtheoryforall of the
H5c: Attitude toward using techno- intention models1odiscussed in this paper, we
logy willnot have a significant expect that behavioral intention will have a
influence on behavioral significantpositiveinfluenceon technologyusage.
intention.9
H6: Behavioral intention will have a
significant positive influence on
usage.

9 To test the absenceof a relationship,

we perform
a 10Forexample,see Sheppardet al. (1988) for an
poweranalysisinthe resultssection. extendedreviewof the intention-behavior
relationship.

456 MISQuarterly
Vol.27 No.3/September
2003

Empirical Validation of UTAUT
Preliminary Test of UTAUT
Using the post-trainingdata (T1) pooled across
studies (N = 215), a measurement model of the
seven direct determinantsof intention(using all
items that relatedto each of the constructs)was
estimated. All constructs, with the exception of
use, were modeled using reflectiveindicators.All
internal consistency reliabilities (ICRs) were
greaterthan .70. The square roots of the shared
variance between the constructs and their mea-
sures were higher than the correlationsacross
constructs, supporting convergent and discri-
minantvalidity-see Table 14(a). The reverse-
coded affect items of Compeau and Higgins
(1995b) had loadings lower than .60 and were
dropped and the model was reestimated. With
the exception of eight loadings, all others were
greater than .70, the level that is generally con-
sideredacceptable (Fornelland Larcker1981;see
also Compeau and Higgins 1995a, 1995b; Com-
peau et al. 1999)-see Table 15. Inter-itemcorre-
lation matrices (details not shown here due to
space constraints)confirmedthat intra-construct
item correlationswere very high while inter-con-
struct item correlations were low. Results of
similaranalyses fromsubsequent timeperiods(T2
and T3)also indicatedan identicalpatternand are
shown in Tables 14(b) and 14(c).
Althoughthe structuralmodel was tested on all
the items, the sample size poses a limitationhere
because of the number of latent variables and
associated items. Therefore,we reanalyzedthe
data using only fourof the highest loading items
from the measurement model for each of the
determinants;intentiononly employs three items
and use is measured via a single indicator.
UTAUTwas estimated using data pooled across
studies at each time period(N = 215). As withthe
model comparison tests, the bootstrapping
method was used here to test the PLS models.
An examinationof the measurement model from
the analysis using the reduced set of items was
similarto that reported in Tables 14 and 15 in
terms of reliability,convergent validity,discrimi-
nant validity, means, standard deviations, and
Venkateshet al./UserAcceptance of IT
correlations. The results fromthe measurement
model estimations (with reduced items) are not
shown here in the interestof space.
An examinationof these highest loading items
suggested that they adequately represented the
conceptualunderpinningsof the constructs-this
preliminarycontent validitynotwithstanding,we
will returnto this issue later in our discussion of
the limitationsof this work. Selection based on
item loadings or corrected item-totalcorrelations
are often recommended in the psychometric
literature(e.g., Nunnallyand Bernstein 1994).
This approach favors building a homogenous
instrument with high internal consistency, but
could sacrifice content validity by narrowing
domaincoverage." The itemsselected forfurther
analysis are indicatedvia an asterisk in Table 15,
and the actual items are shown in Table 16.
Tables 17(a) and 17(b) show the detailed model
test results at each time periodfor intentionand
usage, respectively, including all lower-level
interactionterms. Tables 17(a) and 17(b) also
show the model with direct effects only so the
reader can comparethat to a modelthat includes
the moderating influences. The variance ex-
plained at various points in time by a direct
effects-only model and the full model including
interactionterms are shown in Tables 17(a) and
17(b) for intentionand usage behavior, respec-
tively.12We pooled the data across the different
11BagozziandEdwards(1998)discusspromising new
alternatives forcopingwiththeinherent
to thisapproach
tension betweensample size requirements and the
of
numberof items,suchas fullor partialaggregation
items.
12SincePLSdoes notproduceadjustedR2,we usedan
alternative
procedure to estimatean adjustedR2. PLS
estimatesa measurement modelthatinturnis usedto
generatelatentvariableobservationsbased on the
loadings;these latentvariableobservationsareusedto
estimatethe structuralmodelusingOLS.Therefore, in
orderto estimatethe adjustedR2,we used the latent
observations
variable generated fromPLSandanalyzed
the datausinghierarchical regressionsin SPSS. We
reportthe R2and adjustedR2fromthe hierarchical
regressions. This allowsfor a directcomparisonof
varianceexplainedfromPLSwithvarianceexplained
(both R2 and adjusted R2) from traditionalOLS
regressionsand allows the readerto evaluatethe
varianceexplained-parsimony trade-off.
MISQuarterlyVol.27 No. 3/September2003 457
Venkateshet al./UserAcceptance of IT

Tale14SMasreen Mde*Etiatonfo te reimnay es o UAU

(a) T1 Results (N = 215)

ICR Mean S Dev PE EE ATUT SI FC SE ANX BI
PE .92 5.12 1.13 .94
EE .91 4.56 1.40 .31*** .91
ATUT .84 4.82 1.16 .29*** .21** .86
SI .88 4.40 1.04 .30*** -.16* .21** .88
FC .87 4.17 1.02 .18* .31*** .17* .21** .89
SE .89 5.01 1.08 .14 .33*** .16* .18** .33*** .87
ANX .83 3.11 1.14 -.10 -.38*** -.40*** -.20** -.18** -.36*** .84
BI .92 4.07 1.44 .38*** .34*** .25*** .35*** .19** .16* -.23** .84
(b) T2 Results (N = 215)
ICR Mean S Dev PE EE ATUT SI FC SE ANX BI
PE .91 4.71 1.11 .92
EE .90 5.72 0.77 .30*** .90
ATUT .77 5.01 1.42 .25** .20** .86
SI .94 3.88 1.08 .27*** -.19* .21** .88
FC .83 3.79 1.17 .19* .31*** .18* .20** .86
SE .89 5.07 1.14 .24** .35*** .19** .21** .33*** .75
ANX .79 3.07 1.45 -.07 -.32*** -.35*** -.21** -.17* -.35*** .82
BI .90 4.19 0.98 .41*** .27*** .23** .21** .16* .16* -.17* .87
(c) T3 Results (N = 215)
ICR Mean S Dev PE EE ATUT SI FC SE ANX BI
PE .91 4.88 1.17 .94
EE .94 5.88 0.62 .34*** .91
ATUT .81 5.17 1.08 .21** .24** .79
SI .92 3.86 1.60 .27*** -.15 .20** .93
FC .85 3.50 1.12 .19* .28*** .18* .22** .84
SE .90 5.19 1.07 .14* .30*** .22** .20** .33*** .77
ANX .82 2.99 1.03 -.11 -.30*** -.30*** -.20** -.24** -.32*** .82
BI .90 4.24 1.07 .44*** .24** .20** .16* .16* .16* -.14 .89

Notes: 1. ICR:Internalconsistency reliability.

2. Diagonalelements are the square rootof the shared variance between the constructsand
theirmeasures; off-diagonalelements are correlationsbetween constructs.
3. PE: Performance expectancy; EE: Effort expectancy; ATUT: Attitude toward using
technology;SI:Social influence;FC:Facilitatingconditions;SE: Self-efficacy;ANX:Anxiety;
BI:Behavioralintentionto use the system.

458 MISQuarterlyVol.27 No. 3/September2003

 Venkateshet al./UserAcceptance of IT

Tal 0Ie L 0s 0
( =2 a.m e

Items T1 T2 T3 Items TI T2 T3
U1 .82 .81 .80 *SN1 .82 .85 .90
U2 .84 .80 .81 C *SN2 .83 .85 .84
U3 .81 .84 .84 SF1 .71 .69 .76
U4 .80 .80 .84 . *SF2 .84 .80 .90
U5 .81 .78 .84 c SF3 .72 .74 .77
*U6 .88 .88 .90 *SF4 .80 .82 .84
*RA1 .87 .90 .90 11 .69 .72 .72
RA2 .73 .70 .79 0 12 .65 .75 .70
M RA3 .70 .69 .83 13 .71 .72 .69
RA4 .71 .74 .74 PBC1 .72 .66 .62
x *RA5 .86 .88 .94 *PBC2 .84 .81 .80
"' JF1 .70 .71 .69 *PBC3 .81 .81 .82
JF2 .67 .73 .64 V PBC4 .71 .69 .70
M - JF3 .74 .70 .79 o *PBC5 .80 .82 .80
E: JF4 .73 .79 .710 FC1 .74 .73 .69
JF5 .77 .71 .73 FC2 .78 .77 .64
JF6
JF6 81
.81 .78 81
.81
,
.78 *FC3 .80 .80 .82
OE1 .72 .80 .75 Cl .72 .72 .70
OE2 .71 .68 .77 - C2 .71 .74 .74
OE3 .75 .70 .70 .wMr
.
C3 7 .70
OE4 .70 .72 .67
OE5 *SE1 .80 .83 .84
OE5 .72
.72 .72
.72 .70
.70
0E SE2 .78 .80 .80
OE6 .69 .79 .74
OE7 .86 .87 .90
E" SE3 .72 .79 .74
EOUI .90 .89 .83 4.*S8W.
w
Li ? SE4
SE5 .77
.80 .74
.84 .69
.84
EOU2 .90 .89 .88
*SE6 .81 .82 .82
*EOU3 .94 .96 .91
*SE7 .87 .85 .86
EOU4 .81 .84 .88 . .
SSE8 .70 .69 .72
c *EOU5 .91 .90 .90
* .92 .92 .93 *ANX .78 .74 .69
O
4) EU6
E1 4.84
.84
4 .80
. 8_<8*ANX3
"zX
*ANX2 .71
.72
.70
.69
.72
.73
SW EU2 .83 .88 .85 *ANX4 .77
EU3 .74 .72
ti .89 .84 .80
o *EU4 .91 .91 .92 *BI1 .88 .84 .88
w C01 .83 .82 .81 o *B02 .82 .86 .88
CO02 .83 .78 .80 *BI3 .84 88 7
87
CO03 .81 .84 .80
CO04 .75 .73 .78
*Al .80 .83 .85
A2 .67 .64 .65
r- A3 .64 .64 .71
A4 .72 .71 .64
S IM1 .70 .78 .72
IM2 .72 .72 .78
>,
O O IM3 .73 .75 .81
- _
_*AF1 .79 .77 .84
*AF2 .84 .83 .84
AF3 .71 .70 .69
.D
S *Affect1 .82 .85 .82
Affect2 .67 .70 .70
Affect3 .62 .68 .64

Notes: 1. The loadingsat T1, T2, and T3 respectivelyare fromseparate measurementmodeltests and relateto
Tables 14(a), 14(b), and 14(c) respectively.
2. Extrinsicmotivation(EM)was measured using the same scale as perceivedusefulness (U).
3. Items denoted withan asteriskare those thatwere selected for inclusionin the test of UTAUT.

MISQuarterlyVol.27 No. 3/September2003 459

Venkateshet al./UserAcceptance of IT

Tale16 te s se n stmaig U AU

Performance expectancy
U6: I wouldfindthe system useful in my job.
RA1: Using the system enables me to accomplishtasks more quickly.
RA5: Using the system increases my productivity.
OE7: IfI use the system, I willincrease my chances of gettinga raise.

Effort expectancy
EOU3: Myinteractionwiththe system would be clear and understandable.
EOU5: Itwould be easy for me to become skillfulat using the system.
EOU6: I wouldfindthe system easy to use.
EU4: Learningto operate the system is easy for me.

Attitude toward using technology

Al: Using the system is a bad/good idea.
AF1: The system makes workmore interesting.
AF2: Workingwiththe system is fun.
Affect1: I likeworkingwiththe system.
Social influence
SN1: People who influencemy behaviorthinkthat I should use the system.
SN2: People who are importantto me thinkthat I should use the system.
SF2: The senior managementof this business has been helpfulin the use of the system.
SF4: Ingeneral, the organizationhas supportedthe use of system.
the,
Facilitating conditions
PBC2: I have the resources necessary to use the system.
PBC3: I have the knowledgenecessary to use the system.
PBC5: The system is not compatiblewithother systems I use.
FC3: A specific person (or group)is availablefor assistance withsystem difficulties.

Self-efficacy
I could complete a job or task using the system...
SE1: Ifthere was no one aroundto tell me what to do as I go.
SE4: If I could call someone for help if I got stuck.
SE6: If I had a lot of time to complete the job for whichthe softwarewas provided.
SE7: If I had just the built-inhelp facilityfor assistance.

Anxiety
ANX1: I feel apprehensiveabout using the system.
ANX2: Itscares me to thinkthat I could lose a lot of informationusing the system by hitting
the wrongkey.
ANX3: I hesitate to use the system forfear of makingmistakes I cannot correct.
ANX4: The system is somewhat intimidatingto me.

Behavioral intention to use the system

BI1: I intendto use the system in the next <n> months.
B12: I predictI would use the system in the next <n> months.
B13: I planto use the system in the next <n> months.

460 MISQuarterlyVol.27 No. 3/September2003


pointsof measurement by convertingtime period
(experience) into a moderator. The columntitled
"Pooled Analysis" reports the results of this
analysis. Cautionis necessary when conducting
such analyses and the reader is referred to
AppendixA for a discussion of the potentialcon-
straintsof pooling. Also reportedin AppendixA
are the statisticaltests thatwe conducted priorto
poolingthe data for the analysis in Table 17.
As expected, the effect of performanceexpec-
tancy was in the form of a three-way inter-
action-the effect was moderatedby gender and
age such that it was more salient to younger
workers, particularlymen, thus supporting HI.
Note that a direct effect for performanceexpec-
tancy on intentionwas observed; however,these
main effects are not interpretabledue to the
presence of interactionterms (e.g., Aiken and
West 1991). The effect of effortexpectancy was
via a three-wayinteraction-the effectwas moder-
ated by gender and age (more salient to women
and more so to older women). Based on Chow's
test of beta differences(p < .05), effortexpectancy
was more significantwithlimitedexposure to the
technology (effect decreasing with experience),
thus supportingH2. The effect of social influence
was via a four-wayinteraction-with its role being
more importantin the context of mandatoryuse,
moreso amongwomen, and even moreso among
olderwomen. The Chow'stest of beta differences
(p < .05) indicatedthat social influencewas even
more significantin the early stages of individual
experience with the technology, thus supporting
H3. Facilitatingconditionswas nonsignificantas
a determinantof intention,thus supportingH4a.13
As expected self-efficacy, anxiety, and attitude
did not have any direct effect on intention,thus
supportingH5a, H5b,and H5c. Three of the four
13Sincethesetwohypotheseswereaboutnonsignificant
relationships,the supportiveresultsshouldbe inter-
pretedwithcaution,bearinginmindthe poweranalysis
reported inthetext. Inthiscase, theconcernaboutthe
lackof a relationship is somewhatalleviatedas some
previousresearchin this area has foundattitudeand
conditions
facilitating to notbe predictors in
of intention
the presence of performance expectancyand effort
expectancyconstructs(see Taylorand Todd1995b;
Venkatesh2000).
Venkateshet al./UserAcceptance of IT
nonsignificant determinants (i.e., self-efficacy,
anxiety, and attitude) were dropped from the
model and the modelwas reestimated;facilitating
conditions was not dropped from the model
because of its role in predictinguse.14 The re-
estimated model results are shown in Table 17.
In predictingusage behavior (Table 17b), both
behavioralintention(H6)andfacilitating conditions
were significant, with the latter's effect being
moderated by age (the effect being more impor-
tant to olderworkers);and, based on Chow'stest
of the beta differences (p < .05), the effect was
stronger with increasing experience with the
technology(H4b). Since H4a, H5a, H5b,and H5c
are hypothesized such that the predictoris not
expected to have an effect on the dependent
variable, a power analysis was conducted to
examine the potentialfor type IIerror. The likeli-
hood of detecting medium effects was over 95
percentforan alpha level of .05 and the likelihood
of detecting small effects was under50 pecent.
Cross-Validation of UTAUT
Data were gathered from two additionalorgani-
zations to further validate UTAUT and add
externalvalidityto the preliminary test. The major
details regarding the two participatingorgani-
zations are providedin Table 18. The data were
collected on the same timelineas studies 1 and 2
(see Figure 2). The data analysis procedures
were the same as the previous studies. The
resultswere consistent withstudies 1 and 2. The
items used inthe preliminary test of UTAUT(listed
in Table 16) were used to estimate the mea-
surement and structuralmodels in the new data.
This helped ensure the test of the same model in
both the preliminarytest and this validation,thus
limitingany variationdue to the changingof items.
UTAUTwas then tested separatelyfor each time
14Sincethesetwohypotheseswereaboutnonsignificant
the supportiveresultsshouldbe inter-
relationships,
pretedwithcautionafterconsideration of the power
analysesreportedinthetext. Theresultsarepresented
hereformodelcompletenessandto allowthe readerto
link the current research with existing theoretical
perspectives in this domain.
MISQuarterlyVol.27 No. 3/September2003 461
Venkateshet al./UserAcceptance of IT

Tal 17. Pei o A

(a) Dependent Variable: Intention

T1 T2 T3 Pooled
(N = 215) (N = 215) (N = 215) (N = 645)
DONLY D+I DONLY D+I DONLY D+l DONLY D+I
R2(PLS) .40 .51 .41 .52 .42 .50 .31 .76
R2(hierarchicalregrn.) .39 .51 .41 .51 .42 .50 .31 .77
AdjustedR2(hierarchicalregrn.) .35 .46 .38 .46 .36 .45 .27 .69

Performanceexpectancy(PE) .46*** .17* .57*** .15* .59*** .16* .53*** .18*

Effortexpectancy(EE) .20* -.12 .08 .02 .09 .11 .10 .04
Social influence(SI) .13 .10 .10 .07 .07 .04 .11 .01
Facilitatingconditions(FC) .03 .04 .02 .01 .01 .01 .09 .04
Gender(GDR) .04 .02 .04 .01 .02 .01 .03 .01
Age (AGE) .08 .02 .09 .08 .01 -.08 .06 .00
Voluntariness(VOL) .01 .04 .03 .02 .04 -.04 .02 .00
Experience(EXP) .04 .00
PEx GDR .07 .06 .02
O.17*
PE x AGE d 13 .04 j
06k 10E .01
x
GDRxAGE .07 .02 .02 06
PE x GDR x AGE .52*** 55*** .57*** .55***
EE x GDR 17 08 .09 .02jY
EE xAGE 104
.08 .02 -~ .04
a .02
EEyxEXstP
GDRx AGE(includedearlier) Earlier Earlier Earlier Earlier
GDRx EXP .02
AGE x EXP .01
EE x GDRx AGE 22** 20*** .18*.01
E x GDR x EXP -.10
EExAGE
xEXP-.02
GDRx AGE x EXP -.06
EE x GDRx AGE x EXP -.27***
SI x GDR .11
.00 .02 .02
Slx AGE c060u 01 .02
Sl xVOL
O
.0V .01 ~ .06
f e
Slx EXP
GDR x AGE (included earlier) Earlier ..04 arlier Earlier
Eaaier
GDR x VOL 01 .04 02 .01
GDR x EXP (included earlier) Earlier
AGExVOL .0 .02 06 .02
AGE x EXP (included earlier) A Earlier
VOL x EXP .02
SI x GDR x AGE -.10 .02 .04 .04
SI x GDR x VOL .03 .02.01
-.01
SI x GDR x EXP .01
SI x AGE x VOL -.17* .02 .06 .06
SI x AGE %M.01
x EXP
SI x VOL x EXP .00

462 MISQuarterlyVol.27 No. 3/September2003

 Venkateshet al./User Acceptance of IT

- - A
T 1 l yTso (Cntnud

T1 T2 T3 Pooled
(N = 215) (N = 215) (N = 215) (N = 645)
DONLY D+1 DONLY D+l DONLY D+I DONLY D + I
GDR x AGE x VOL .02 02 01 .00
GDRx AGE x EXP(included Earlier
earlier)
GDR x VOLx EXP .00
AGE x VOLx EXP .01
SI x GDR x AGE x VOL .25** 23** 20* .04
GDRx AGE x VOLx EXP .02
SI x GDRx AGE x VOLx EXP .28**
(b) Dependent Variable: Usage Behavior
R2(PLS) .37 .43 .36 .43 .39 .44 .38 .53
R2(hierarchicalregrn.) .37 .43 .36 .43 .39 .43 .38 .52
AdjustedR2(hierarchicalregrn.) .36 .41 .35 .40 .38 .41 .37 .47

Behavioralintention(BI) .61*** 57*** .60*** .58*** .58*** .59*** .59*** .52***

conditions(FC)
Facilitating .05 .07 .06 .07 .18* .07 .10 .11
Age (AGE) .02 .02 .01 .02 .04 .13 .04 .08
Experience(EXP) .06
FC x AGE .22* 24* .27** .02
FC x EXP .00
AGE x EXP .01
FC x AGE x EXP .23**

Notes: 1. D ONLY:Directeffects only;D + I:Directeffects and interactionterms.

2. "Includedearlier"indicates that the term has been listed earlierin the table, but is includedagain for
completeness as it relates to higher-orderinteractionterms being computed. Grayedout cells are not
applicableforthe specific column.

Tabl 1 8. Dec ip oftuie

Functional Sample
Study Industry Area Size System Description

Voluntary Use

3 Financial Research 80 This software application was one of the

Services resources available to analysts to conduct
research on financial investment opportunities
and IPOs

Mandatory Use

4 Retail Customer 53 Application that customer service representa-

Electronics Service tives were required to use to document and
manage service contracts

MISQuarterlyVol.27 No. 3/September2003 463

Venkateshet al./UserAcceptance of IT

Tabe 9.Meaurmet.Mde Etimtin or heCrss-Vaidaion fS AU

(a) T1 Results (N = 133)

ICR Mean S Dev PE EE ATUT SI FC SE ANX BI
PE .90 4.87 1.20 .88
EE .90 3.17 1.09 .30*** .93
ATUT .80 2.67 0.87 .28*** .16* .80
SI .91 4.01 1.07 .34*** -.21* .20* .89
FC .85 3.12 1.11 .18* .30*** .15* .14 .84
SE .85 4.12 1.08 .19** .37*** .17* .22** .41*** .82
ANX .82 2.87 0.89 -.17* -.41** -.41*** -.22** -.25** -.35*** .80
BI .89 4.02 1.19 .43*** .31*** .23*** .31*** .22* .24** -.21** .85
(b) T2 Results (N=133)
ICR Mean S Dev PE EE ATUT SI FC SE ANX BI
PE .90 4.79 1.22 .91
EE .92 4.12 1.17 .34*** .89
ATUT .84 3.14 0.79 .32*** .18** .78
SI .92 4.11 1.08 .30*** -.20* .21** .91
FC .88 3.77 1.08 .19* .31*** .13 .11 .86
SE .87 4.13 1.01 .21** .38*** .25** .20** .35*** .80
ANX .83 3.00 0.82 -.17* -.42*** -.32*** -.24** -.26* -.36*** .81
BI .88 4.18 0.99 .40*** .24** .21** .24** .20* .21** -.22** .88
(c) T3 Results (N = 133)
ICR Mean S Dev PE EE ATUT SI FC SE ANX BI
PE .94 5.01 1.17 .92
EE .92 4.89 0.88 .34*** .90
ATUT .83 3.52 1.10 .30*** .19** .80
SI .92 4.02 1.01 .29*** -.21" .21** .84
FC .88 3.89 1.00 .18* .31*** .14 .15 .81
SE .87 4.17 1.06 .18** .32*** .21* .20* .35*** .84
ANX .85 3.02 0.87 -.15* -.31*** -.28*** -.22* -.25* -.38*** .77
BI .90 4.07 1.02 .41*** .20** .21* .19* .20* .20* -.21** .84

Notes: 1. ICR:Internalconsistency reliability.

2. Diagonalelements are the square rootof the shared variancebetween the constructsand
their measures; off-diagonalelements are correlationsbetween constructs.
3. PE: Performance expectancy; EE: Effort expectancy; ATUT: Attitude toward using
technology;SI:Social influence;FC:Facilitatingconditions;SE: Self-efficacy;ANX:Anxiety;
BI:Behavioralintentionto use the system.

464 MISQuarterlyVol.27 No. 3/September2003

 Venkateshet al./UserAcceptance of IT

Tabe2. I
Load rmS mhtme

Items T1 T2 T3 Items T1 T2 T3
U6 .91 .92 .91 PBC2 .84 .88 .85
Performance RA1 Facilitating PBC3
.90 .89 .88 .88 .89 .88
Expectancy Conditions
(PE) RA5 .94 .89 .90 (FC) PBC5 .86 .89 .84
OE7 .89 .90 .91 FC3 .87 .78 .81
EOU3 .91 .90 .94 SE1 .90 .84 .88
Effort EOU5 .92 .91 .90 Self-Efficacy SE4 .88 .82 .81
Expectancy
Expectancy)
EOU6 .93 .90 .89 (SE) SE6 .80 .85 .79
EU4 .87 .87 .90 SE7 .81 .77 .75
Attitude Al .84 .80 .86 ANX1 .80 .84 .80
Toward AF1 .82 .83 .77 ANX2 .84 .84 .82
Anxiety
Using
.80 .80 .76 (ANX) ANX3 .83 .80 .83
TechnologyAF2
Technology
(ATUT) Affectl .87 .84 .76 ANX4 .84 .77 .83
SN1 .94 .90 .90 BI1 .92 .90 .91
Social SN2 .90 .93 .88 Intention B12 .90 .90 .91
Influence
(S) SF2 .89 .92 .94 (BI) B13 .90 .92 .92
SF4 .92 .81 .79

Note: The loadings at T1, T2, and T3 respectively are from separate measurement model tests and
relate to Tables 18(a), 18(b), and 18(c) respectively.

-abl- 2-

on.oh g
(a) Dependent Variable: Intention
T1 T2 T3 Pooled
(N = 133) (N = 133) (N = 133) (N = 399)
DONLY D+I DONLY D+I DONLY D+I DONLY D+I
R2(PLS) .42 .52 .41 .52 .42 .51 .36 .77
R2(hierarchicalregrn.) .41 .52 .41 .52 .42 .51 .36 .77
AdjustedR2(hierarchicalregrn.) .37 .48 .36 .47 .36 .46 .30 .70

Performanceexpectancy(PE) .45*** .15 .59*** .16* .59*** .15* .53*** .14

Effortexpectancy(EE) .22** .02 .06 .06 .04 .01 .10 .02
Social influence(SI) .02 .04 .04 .04 .02 .01 .02 .02
Facilitatingconditions(FC) .07 .01 .00 .08 .01 .00 .07 .01
Gender(GDR) .02 .06 .01 .04 .07 .02 .03 .03
Age (AGE) .01 .00 .07 .02 .02 .01 .07 .01
Voluntariness(VOL) .00 .00 .01 .06 .02 .04 .00 .01
Experience(EXP) .08 .00
PE x GDR .14 17* 18 .01
PE x AGE .06 .01 .02 -.04
GDR x AGE .02 .04 .04 -.02
EE x GDR -.06 02 .04 .09
EE x AGE -.02 .01 .02 -.04
EE x EXP .01

MISQuarterlyVol.27 No. 3/September2003 465

Venkateshet al./UserAcceptance of IT

Tal 21
Es-ala][tI~ eI'1 i. 'n[e ofJLIi FAUI E(C]ont inue[d)
T1 T2 T3 Pooled
(N = 133) (N = 133) (N = 133) (N = 399)
DONLY D + I DONLY D+I DONLY D+I DONLY D+I
GDR x AGE(includedearlier) Earlier Earlier Earlier Earlier
GDR x EXP .02
AGE x EXP .06
EE x GDRx AGE .21** .18" .16" .04
EE x GDRx EXP .00
EE x AGE x EXP .00
GDRx AGE x EXP .01
EE x GDRx AGE x EXP -.25***
SI xGDR -.06.00 .02 -.07
SI x AGE .02 .01 .04 .02
SI xVOL .01 .04 .00 .00
SI x EXP .02IN .
GDR x AGE(includedearlier) Earlier Earlier gk Earlier Earlier
GDRx VOL .04 .02 -.03 .02
GDR x EXP(includedearlier) Earlier
AGE x VOL .00 .01 .00 .07
AGE x EXP(includedearlier) Earlier
VOLx EXP .02
SI x GDRx AGE -.03 .01 -.07 .00
Sl x GDRx VOL .04 -.03 .04 .00
SI x GDRx EXP .00
SI x AGE x VOL .06 .02 .00 .01
SI x AGE x EXP M.07
SI x VOLx EXP .02
GDRx AGE x VOL .01 .06 .01 .04
GDRx AGE x EXP(included Earlier
earlier)
GDR x VOLx EXP ... ....
.. .. .01
AGE x VOL
.VO..L
x..x
EXP
.E
X..P .......................... ......... .0
...........0
SI x GDRx AGE x VOL .27*** .21** .16* .01
GDR x AGE x VOLx EXP .00
SI x GDR x AGE x VOLx EXPa -.299*
(b) Dependent Variable: Usage Behavior
R2(PLS) .37 .44 .36 .41 .36 .44 .38 .52
R2(hierarchicalregrn.) .37 .43 .36 .41 .36 .44 .37 .52
R2
Adjusted (hierarchicalregrn.) .36 .41 .35 .38 .35 .41 .36 .48

Behavioralintention(BI) .60*** .56*** .59*** .50*** .59*** .56*** .59*** .51**

Facilitatingconditions(FC) .04 .11 .01 .01 .02 .06 .14* .08
Age (AGE) .06 .02 .06 -.03 -.03 -.01 .06 .02
Experience(EXP) .10
FC x AGE .17" .21** .24** .01
FC x EXP .06
AGExEXP -.07
FC x AGE x EXP .2 *

Notes: 1. D ONLY:Directeffects only;D + I: Directeffects and interactionterms.

2. "Includedearlier"indicatesthat the term has been listed earlierin the table, but is includedagain for
completeness as it relates to higher-orderinteractionterms being computed. Grayedout cells are not
applicableforthe specific column.

466 MISQuarterlyVol.27 No. 3/September2003


period (N = 133 at each time period). The mea-
surementmodels are shown in Tables 19 and 20.
The patternof results in this validation(Tables
21(a) and 21(b)) mirrorswhat was found in the
preliminarytest (Table 17). The last column of
Tables 21(a) and 21(b) reportsobservationsfrom
the pooledanalysisas before. AppendixB reports
the statisticaltests we conducted priorto pooling
the data for the cross-validationtest, consistent
with the approach taken in the preliminarytest.
Insofaras the no-relationshiphypotheses were
concerned, the power analysis revealed a high
likelihood(over 95 percent) of detecting medium
effects. The variance explained was quite com-
parable to that found in the preliminarytest of
UTAUT.
Discussion
The present research set out to integrate the
fragmented theory and research on individual
acceptance of informationtechnology into a uni-
fied theoreticalmodel that captures the essential
elements of eight previouslyestablished models.
First, we identified and discussed the eight
specific models of the determinantsof intention
and usage of informationtechnology. Second,
these models were empiricallycompared using
within-subjects,longitudinaldatafromfourorgani-
zations. Third,conceptual and empiricalsimi-
larities across the eight models were used to
formulatethe UnifiedTheoryof Acceptance and
Use of Technology(UTAUT).Fourth,the UTAUT
was empiricallytested using the originaldata from
the four organizationsand then cross-validated
using new data from an additionaltwo organi-
zations. These tests provided strong empirical
support for UTAUT,which posits three direct
determinants of intention to use (performance
expectancy, effort expectancy, and social influ-
ence) and two direct determinants of usage
behavior (intention and facilitatingconditions).
Significantmoderatinginfluences of experience,
voluntariness,gender, and age were confirmedas
integralfeatures of UTAUT. UTAUTwas able to
account for 70 percent of the variance (adjusted
R2)inusage intention-a substantialimprovement
Venkateshet al./UserAcceptance of IT
over any of the originaleight models and their
extensions. Further,UTAUTwas successful in
integratingkeyelements fromamong the initialset
of 32 main effects and four moderators as
determinantsof intentionand behaviorcollectively
posited byeightalternatemodels intoa modelthat
incorporated four main effects and four
moderators.
Thus, UTAUTis a definitivemodel that synthe-
sizes what is knownand providesa foundationto
guide future research in this area. By encom-
passing the combined explanatorypower of the
individualmodels and key moderatinginfluences,
UTAUTadvances cumulative theory while re-
taining a parsimoniousstructure. Figure 3 pre-
sents the model proposed and supported.
Table 22 presents a summaryof the findings. It
should be noted that performanceexpectancy
appears to be a determinantof intentionin most
situations: the strengthof the relationshipvaries
with gender and age such that it is more signi-
ficantfor men and youngerworkers. The effect of
effortexpectancy on intentionis also moderated
by gender and age such that it is more significant
for women and older workers,and those effects
decrease with experience. The effect of social
influence on intention is contingent on all four
moderatorsincludedhere such thatwe founditto
be nonsignificantwhen the data were analyzed
withoutthe inclusionof moderators. Finally,the
effect of facilitatingconditionson usage was only
significantwhen examinedin conjunctionwiththe
moderatingeffects of age and experience-i.e.,
they only matterfor older workersin laterstages
of experience.
Priorto discussing the implicationsof this work,it
is necessary to recognize some of its limitations.
One limitationconcerns the scales used to mea-
sure the core constructs. Forpracticalanalytical
reasons, we operationalized each of the core
constructsin UTAUTby using the highest-loading
items from each of the respective scales. This
approach is consistent with recommendationsin
the psychometric literature(e.g., Nunnallyand
Bernstein1994). Such pruningof the instrument
was the only way to have the degrees of freedom
necessary to model the various interactionterms
MISQuarterlyVol 27 No. 3/September2003 467
Venkateshet al./UserAcceptance of IT

-abl22.Summayo

Hypothesis Dependent Independent

Number Variables Variables Moderators Explanation
H1 Behavioral Performance Gender, Age Effectstrongerfor men and
intention expectancy youngerworkers
H2 Behavioral Effort Gender,Age, Effectstrongerfor women,
intention expectancy Experience olderworkers,and those with
limited experience
H3 Behavioral Social Gender,Age, Effectstrongerfor women,
intention influence Voluntariness, olderworkers,underconditions
Experience of mandatoryuse, and with
limited experience
H4a Behavioral Facilitating None Nonsignificantdue to the effect
intention conditions being capturedby effort
expectancy
H4b Usage Facilitating Age, Effectstrongerfor older
conditions Experience workerswithincreasing
experience
H5a Behavioral Computer None Nonsignificantdue to the effect
intention self-efficacy being capturedby effort
expectancy
H5b Behavioral Computer None Nonsignificantdue to the effect
intention anxiety being capturedby effort
expectancy
H5c Behavioral Attitude None Nonsignificantto the effect
intention towardusing being capturedby process
tech. expectancy and effort
expectancy
H6 Usage Behavioral None Directeffect
intention

at the item level as recommended by Chin et al.
(1996). However,one danger of this approachis
that facets of each constructcan be eliminated,
thus threateningcontentvalidity.Specifically,we
found that choosing the highest-loading items
resulted in items from some of the models not
being represented in some of the core constructs
(e.g., items from MPCUwere not represented in
performanceexpectancy). Therefore,the mea-
sures for UTAUTshould be viewed as preliminary
and future research should be targeted at more
fullydevelopingand validatingappropriatescales
for each of the constructs with an emphasis on
content validity,and then revalidatingthe model
specified herein(or extendingit accordingly)with
the new measures. Ourresearch employedstan-
dard measures of intention,but future research
should examine alternativemeasures of intention
and behavior in revalidatingor extending the
research presented here to other contexts.
Froma theoreticalperspective, UTAUTprovides
a refinedviewof howthe determinantsof intention
and behaviorevolve over time. It is importantto

468 MISQuarterlyVol.27 No. 3/September2003


emphasize that most of the key relationshipsin
the model are moderated. Forexample, age has
received very little attention in the technology
acceptance research literature,yet our results
indicatethat it moderates all of the key relation-
ships in the model. Gender, which has received
some recent attention,is also a key moderating
influence;however,consistent withfindingsin the
sociology and social psychology literature(e.g.,
Levy1988), itappears to workin concertwithage,
a heretoforeunexaminedinteraction. For exam-
ple, prior research has suggested that effort
expectancy is more salient for women (e.g.,
Venkatesh and Morris2000). While this may be
true, our findingssuggest this is particularlytrue
forthe oldergenerationof workersand those with
relativelylittleexperience with a system. While
existing studies have contributedto our under-
standing of gender and age influences indepen-
dently,the present research illuminatesthe inter-
play of these two key demographicvariablesand
adds richness to our currentunderstandingof the
phenomenon. We interpretour findingsto sug-
gest that as the younger cohort of employees in
the workforcemature,gender differences in how
each perceives information technology may
disappear. This is a hopeful sign and suggests
thatoft-mentionedgenderdifferencesinthe use of
informationtechnologymay be transitory,at least
as they relateto a youngergenerationof workers
raised and educated in the DigitalAge.
The complex natureof the interactionsobserved,
particularlyfor gender and age, raises several
interestingissues to investigateinfutureresearch,
especially given the interest in today's societal
and workplaceenvironmentsto create equitable
settings for women and men of all ages. Future
research should focus on identifyingthe potential
"magicnumber"for age where effects begin to
appear (say for effort expectancy) or disappear
(say for performanceexpectancy). Whilegender
and age are the variables that reveal an inter-
esting patternof results, futureresearch should
identifythe underlyinginfluentialmechanisms--
potential candidates here include computer
literacyand social or culturalbackground,among
others. Finally,althoughgender moderatesthree
key relationships,it is imperativeto understand
Venkateshet al./UserAcceptance of IT
the importanceof gender roles and the possibility
that "psychologicalgender"is the root cause for
the effects observed. Empiricalevidence has
demonstratedthat gender roles can have a pro-
foundimpacton individualattitudesand behaviors
both withinand outside the workplace(e.g., Baril
et al. 1989; Feldman and Aschenbrenner 1983;
Jagacinski1987;Keys1985; Roberts1997; Sachs
et al. 1992; Wong et al. 1985). Specifically,gen-
der effects observed here could be a manifesta-
tion of effects caused by masculinity,femininity,
and androgyny ratherthan just "biologicalsex"
(e.g., Lubinskiet al. 1983). Futureworkmightbe
directedat moreclosely examiningthe importance
of gender roles and exploringthe socio-psycho-
logical basis for gender as a means for better
understandingits moderatingrole.
As is evidentfromthe literature,the role of social
influence constructs has been controversial.
Some have arguedfortheirinclusionin models of
adoption and use (e.g., Taylorand Todd 1995b;
Thompson et al. 1991), while others have not
includedthem (e.g., Davis et al. 1989). Previous
work has found social influenceto be significant
only in mandatory settings (see Hartwickand
Barki1994; Venkatesh and Davis 2000). Other
work has found social influence to be more
significant among women in early stages of
experience (e.g., Venkatesh and Morris2000).
Still other research has found social influenceto
be more significantamong older workers (e.g.,
Morrisand Venkatesh 2000). This research is
among the firstto examinethese moderatinginflu-
ences in concert. Ourresults suggest that social
influences do matter; however, they are more
likely to be salient to older workers, particularly
women, and even then during early stages of
experience/adoption.This patternmirrorsthat for
effortexpectancywiththe addedcaveat thatsocial
influences are more likely to be importantin
mandatory usage settings. The contingencies
identifiedhere providesome insightsintothe way
in which social influences change over time and
may help explain some of the equivocal results
reportedin the literature.By helpingto clarifythe
contingentnatureof social influences,this paper
sheds light on when social influence is likelyto
play an importantrole in driving behavior and
when it is less likelyto do so.
MISQuarterlyVol.27 No. 3/September2003 469
Venkateshet al./UserAcceptance of IT
UTAUTunderscores this pointand highlightsthe
importanceof contextualanalysis in developing
strategies for technology implementationwithin
organizations. Whileeach of the existing models
in the domain is quite successful in predicting
technology usage behavior, it is only when one
considers the complex range of potentialmoder-
ating influences that a more complete pictureof
the dynamicnatureof individualperceptionsabout
technology begins to emerge. Despite the ability
of the existing models to predict intentionand
usage, currenttheoreticalperspectives on indivi-
dual acceptance are notably weak in providing
prescriptiveguidance to designers. Forexample,
applyingany of the models presented here might
informa designer that some set of individuals
mightfind a new system difficultto use. Future
research should focus on integratingUTAUTwith
researchthathas identifiedcausal antecedents of
the constructs used within the model (e.g.,
Karahannaand Straub 1999; Venkatesh 2000;
Venkatesh and Davis 2000) in orderto providea
greaterunderstandingof howthe cognitivepheno-
mena that were the focus of this research are
formed. Examplesof previouslyexamineddeter-
minants of the core predictors include system
characteristics(Daviset al. 1989) and self-efficacy
(Venkatesh 2000). Additionaldeterminantsthat
have not been explicitlytied into this stream but
merit consideration in future work include task-
technologyfit(Goodhueand Thompson1994) and
individualabilityconstructssuch as "g"-general
cognitiveability/intelligence(Colquittet al. 2000).
While the variance explained by UTAUTis quite
high for behavioralresearch, furtherworkshould
attempt to identifyand test additionalboundary
conditionsof the model in an attempt to provide
an even richerunderstandingof technologyadop-
tionand usage behavior. Thismighttake the form
of additionaltheoreticallymotivated moderating
influences,differenttechnologies (e.g., collabora-
tive systems, e-commerce applications),different
user groups(e.g., individualsindifferentfunctional
areas), and other organizationalcontexts (e.g.,
publicor governmentinstitutions). Results from
such studies will have the importantbenefit of
enhancing the overall generalizabilityof UTAUT
and/orextendingthe existingworkto account for
additionalvariancein behavior.Specifically,given
470 MISQuarterlyVol 27 No. 3/September2003
the extensive moderating influences examined
here, a researchstudythatexamines the general-
izabilityof these findings with significant repre-
sentation in each cell (total numberof cells: 24;
two levels of voluntariness,three levels of experi-
ence [no, limited,more],two levels of gender, and
at least two levels of age [youngvs. old])wouldbe
valuable. Such a study would allow a pairwise,
inter-cellcomparison using the rigorousChow's
test and provide a clear understanding of the
nature of effects for each construct in each cell.
Related to the predictivevalidityof this class of
models in general and UTAUTin particularis the
role of intentionas a key criterionin user accep-
tance research-future research should investi-
gate otherpotentialconstructssuch as behavioral
expectation (Warshawand Davis 1985) or habit
(Venkatesh et al. 2000) in the nomologicalnet-
work. Employingbehavioralexpectationwillhelp
account for anticipated changes in intention
(Warshawand Davis 1985) and thus shed light
even inthe earlystages of the behaviorabout the
actual likelihoodof behavioralperformancesince
intention only captures internal motivations to
performthe behavior. Recent evidence suggests
that sustained usage behavior may not be the
result of deliberated cognitions and are simply
routinizedor automaticresponses to stimuli(see
Venkatesh et al. 2000).
One of the most importantdirectionsfor future
research is to tie this maturestream of research
into other established streams of work. For
example, littleto no research has addressed the
link between user acceptance and individualor
organizationalusage outcomes. Thus, while it is
often assumed that usage will result in positive
outcomes, this remainsto be tested. The unified
model presented here mightinformfurtherinquiry
intothe short-and long-termeffects of information
technology implementationon job-related out-
comes such as productivity,job satisfaction,
organizational commitment, and other perfor-
mance-oriented constructs. Future research
should study the degree to which systems per-
ceived as successful from an IT adoption per-
spective (i.e., those thatare likedand highlyused
by users) are considered a success from an
organizationalperspective.

Conclusion
Followingfromthe UnifiedTheoryof Acceptance
and Use of Technology presented here, future
research should focus on identifyingconstructs
that can add to the predictionof intentionand
behaviorover and above what is already known
and understood. Given that UTAUTexplains as
much as 70 percentof the variance in intention,it
is possible that we may be approaching the
practicallimitsof our abilityto explain individual
acceptance and usage decisions inorganizations.
Inthe study of informationtechnology implemen-
tations in organizations, there has been a
proliferationof competingexplanatorymodels of
individualacceptance of informationtechnology.
The presentworkadvances individualacceptance
research by unifyingthe theoreticalperspectives
common in the literatureand incorporatingfour
moderators to account for dynamic influences
includingorganizationalcontext,user experience,
and demographiccharacteristics.
Acknowledgements
The authors thank Cynthia Beath (the senior
editor),the associate editor,and the three anony-
mous reviewers. We also thankHeatherAdams,
Wynne Chin,DeborahCompeau, Shawn Curley,
RuthKanfer,PhillipAckerman,TracyAnn Sykes,
PeterTodd,ShreevardhanLele, and PaulZantek.
Last, but not least, we thankJan DeGross for her
tireless efforts in typesetting this rather long
paper.
References
Agarwal,R., and Prasad, J. "AConceptualand
OperationalDefinitionof Personal Innovative-
ness inthe Domainof Information Technology,"
InformationSystems Research (9:2), 1998, pp.
204-215.
Agarwal,R., and Prasad, J. "TheRole of Innova-
tion Characteristicsand Perceived Voluntari-
ness in the Acceptance of InformationTechno-
logies," Decision Sciences (28:3), 1997, pp.
557-582.
Venkatesh
et al./UserAcceptanceof IT
Aiken, L. S., and West, S. G. MultipleRegres-
sion: Testing and InterpretingInteractions,
Sage, London,1991.
Ajzen, I. "The Theory of Planned Behavior,"
OrganizationalBehaviorand HumanDecision
Processes (50:2), 1991, pp. 179-211.
Ashmore,R. D. "Sex,Gender,and the Individual,"
in Handbookof Personality,L. A. Pervin(ed.),
GuilfordPress, New York,1980, pp. 486-526.
Bagozzi, R. P., and Edwards,J. R. "AGeneral
Approach to Construct Validationin Organi-
zationalResearch: Applicationto the Measure-
ment of Work Values," OrganizationalRe-
search Methods(1:1), 1998, pp. 45-87.
Bandura,A. Social Foundationsof Thoughtand
Action: A Social Cognitive Theory, Prentice
Hall,EnglewoodCliffs,NJ, 1986.
Baril, G. L., Elbert, N., Mahar-Potter,S., and
Reavy, G. C. "AreAndrogynous Managers
Really MoreEffective?,"Groupand Organiza-
tion Studies (14:2), 1989, pp. 234-249.
Barnett,R. C., and Marshall,N. L. "TheRelation-
ship betweenWomen'sWorkand FamilyRoles
and TheirSubjectiveWell-Beingand Psycho-
logical Distress,"in Women, Workand Health:
Stress and Opportunities,M.Frankenhaeuser,
V. Lundberg, and M. A. Chesney (eds.),
Plenum, New York,1981, pp. 111-136.
Bem, S. L. "The BSRI and Gender Schema
Theory: A Reply to Spence and Helmreich,"
Psychological Review (88:4), 1981, pp. 369-
371.
Bem, D. J., and Allen,A. "OnPredictingSome of
the People Some of the Time:The Search for
Cross-SituationalConsistencies in Behavior,"
Psychological Review (81:6), 1974, pp. 506-
520.
Bergeron,F., Rivard,S., and De Serre, L. "Inves-
tigating the Support Role of the Information
Center,"MIS Quarterly(14:3), 1990, pp. 247-
259.
Bozionelos, N. "Psychologyof ComputerUse:
Prevalence of Computer Anxiety in British
Managers and Professionals," Psychological
Reports (78:3), 1996, pp. 995-1002.
Chin, W. W. "The Partial Least Squares Ap-
proach for StructuralEquation Modeling,"in
Modern Methods for Business Research,
George A. Marcoulides (ed.), Lawrence
2003
Vol.27 No.3/September
MISQuarterly 471
 et al./UserAcceptanceof IT
Venkatesh
ErlbaumAssociates, New York,1998, pp. 295-
336.
Chin,W. W., Marcolin,B. L., and Newsted, P. R.
"A Partial Least Squares Latent Variable
ModelingApproachfor MeasuringInteraction
Effects: Results froma MonteCarloSimulation
Studyand Voice MailEmotion/Adoption Study,"
in Proceedings of the InternationalConference
on InformationSystems, J. I.DeGross,A. Srini-
vasan, and S. Jarvenpaa (eds.), Cleveland,
OH, 1996, pp. 21-41.
Chow, G. C. "Testsof EqualityBetween Sets of
Coefficients in Two Linear Regressions,"
Econometrica(28:3), 1960, pp. 591-605.
Colquitt,J. A., LePine, J. A., and Noe, R. A.
"Towardan IntegrativeTheoryof TrainingMoti-
vation: A Meta-AnalyticPath Analysis of 20
Yearsof TrainingResearch,"JournalofApplied
Psychology (85:5), 2000, pp. 678-707.
Compeau, D. R., and Higgins,C. A. "Application
of Social CognitiveTheoryto TrainingforCom-
puter Skills,"InformationSystems Research
(6:2), 1995a, pp. 118-143.
Compeau, D. R., and Higgins,C. A. "Computer
Self-Efficacy: Developmentof a Measureand
InitialTest," MIS Quarterly(19:2), 1995b, pp.
189-211.
Compeau, D. R., Higgins, C. A., and Huff, S.
"SocialCognitiveTheoryand IndividualReac-
tionsto ComputingTechnology:A Longitudinal
Study,"MIS Quarterly(23:2), 1999, pp. 145-
158.
Davis, F. D. "PerceivedUsefulness, Perceived
Ease of Use, and User Acceptance of Informa-
tion Technology,"MIS Quarterly(13:3), 1989,
pp. 319-339.
Davis, F. D., Bagozzi, R. P., and Warshaw,P. R.
"Extrinsicand IntrinsicMotivationto Use Com-
puters in the Workplace,"Journal of Applied
Social Psychology (22:14), 1992, pp. 1111-
1132.
Davis, F. D., Bagozzi, R. P., and Warshaw,P. R.
"UserAcceptance of ComputerTechnology: A
Comparisonof TwoTheoreticalModels,"Man-
agement Science (35:8), 1989, pp. 982-1002.
Efron,B., and Gong, G. "ALeisurelyLookat the
Bootstrap,the Jackknife,and Cross-Validation,"
TheAmericanStatistician(37:1), 1983, pp. 36-
48.
472 Vol.27 No.3/September
MISQuarterly 2003
Eichinger, J., Heifetz, L.J., and Ingraham, C.
"SituationalShifts in Sex Role Orientation:
Correlates of Work Satisfaction and Burnout
Among Women in Special Education,"Sex
Roles (25:7/8), 1991, pp. 425-430.
Feldman, S. S., and Aschenbrenner,B. "Impact
of Parenthoodon VariousAspects of Mascu-
linityand Femininity: A Short-TermLongitu-
dinalStudy,"DevelopmentalPsychology(19:2),
1983, pp. 278-289.
Fiske, S. T., and Taylor,S. E. Social Cognition,
McGraw-Hill, New York,1991.
Fishbein,M.,and Ajzen, I. Belief, Attitude,Inten-
tion and Behavior: An Introductionto Theory
and Research, Addison-Wesley,Reading, MA,
1975.
Fornell,C., and Larcker,D. F. "EvaluatingStruc-
tural Equation Models with Unobservable
Variables and Measurement Error: Algebra
and Statistics,"Journalof MarketingResearch
(18:3), 1981, pp. 382-388.
French, J. R. P., and Raven, B. "TheBases of
Social Power,"in Studies in Social Power, D.
Cardwright(ed.), Institutefor Social Research,
AnnArbor,MI,1959, pp. 150-167.
Goodhue, D. L. "Understanding User Evaluations
of Information Systems,"ManagementScience
(41:12), 1995, pp. 1827-1844.
Goodhue, D. L., and Thompson, R. L. "Task-
Technology Fit and IndividualPerformance,"
MISQuarterly(19:2), 1995, pp. 213-236.
Hall,D., and Mansfield,R. "Relationshipsof Age
and Senioritywith Career Variables of Engi-
neers and Scientists," Journal of Applied
Psychology (60:2), 1995, pp. 201-210.
Harrison, D. A., Mykytyn,P. P., and Riemen-
schneider, C. K. "ExecutiveDecisions About
Adoptionof InformationTechnology in Small
Business: Theoryand EmpiricalTests," Infor-
mation Systems Research (8:2), 1997, pp.
171-195.
Hartwick,J., and Barki,H. "Explaining the Role of
User Participationin InformationSystem Use,"
ManagementScience (40:4), 1994, pp.40-465.
Helson, R., and Moane, G. "PersonalityChange
in Women fromCollege to Midlife,"Journalof
Personality and Social Psychology (53:1),
1987, pp. 176-186.

Hu, P. J., Chau, P. Y. K., Sheng, O. R. L., and
Tam, K.Y. "Examining the TechnologyAccep-
tance Model Using Physician Acceptance of
TelemedicineTechnology,"JournalofManage-
ment Information Systems(16:2), 1999, pp. 91-
112.
Jagacinski, C. M. "Androgynyin a Male-Domi-
nated Field: The Relationshipof Sex-Typed
Traitsto Performanceand Satisfactionin Engi-
neering, Sex Roles (17:X), 1987, pp. 529-547.
Karahanna,E., and Straub, D. W. "ThePsycho-
logical Origins of Perceived Usefulness and
Ease of Use," Informationand Management
(35:4), 1999, pp. 237-250.
Karahanna,E., Straub,D. W., and Chervany,N.
L. "Information
Technology AdoptionAcross
Time: A Cross-SectionalComparisonof Pre-
Adoption and Post-Adoption Beliefs," MIS
Quarterly(23:2), 1999, pp. 183-213.
Keys, D. E. "Gender,Sex Role, and Career
Decision Makingof CertifiedManagementAc-
countants,"Sex Roles (13:1/2), 1985, pp. 33-
46.
Kirchmeyer,C. "Changeand Stabilityin Mana-
ger's Gender Roles," Journal of Applied
Psychology (87:5), 2002, pp. 929-939.
Kirchmeyer,C. "GenderRoles in a Traditionally
Female Occupation: A Study of Emergency,
Operating, Intensive Care, and Psychiatric
Nurses, Journalof VocationalBehavior(50:1),
1997, pp. 78-95.
Leonard-Barton,D., and Deschamps, I. "Mana-
gerial Influencein the Implementationof New
Technology," Management Science (34:10),
1988, pp. 1252-1265.
Levy,J. A. "Intersectionsof Gender and Aging,"
The Sociological Quarterly(29:4), 1988, pp.
479-486.
Lubinski,D., Tellegen, A. and Butcher, J. N.
"Masculinity, Femininity, and Androgyny
Viewed and Assessed as DistinctConcepts,"
Journalof Personalityand Social Psychology
(44:2), 1983, pp. 428-439.
Lynott,P. P., and McCandless,N. J. "TheImpact
of Age vs. LifeExperienceson the GenderRole
Attitudesof Women in DifferentCohorts,"Jour-
nalof WomenandAging(12:2),2000, pp. 5-21.
Mathieson,K. "PredictingUser Intentions:Com-
paringthe TechnologyAcceptance Modelwith
Venkateshet al./UserAcceptance of IT
the Theoryof Planned Behavior,"Information
Systems Research (2:3), 1991, pp. 173-191.
Miller, J. B. Toward a New Psychology of
Women,Beacon Press, Boston, 1976.
Minton,H. L., and Schneider, F. W. Differential
Psychology, Waveland Press, Prospect
Heights, IL,1980.
Moore,G. C., and Benbasat, I. "Developmentof
an Instrumentto Measure the Perceptions of
Adoptingan InformationTechnology Innova-
tion," InformationSystems Research (2:3),
1991, pp. 192-222.
Moore,G. C., and Benbasat, I. "IntegratingDiffu-
sion of Innovationsand Theory of Reasoned
Action Models to Predict Utilizationof Infor-
mationTechnologyby End-Users,"in Diffusion
and Adoption of InformationTechnology, K.
Kautzand J. Pries-Hege (eds.), Chapmanand
Hall,London,1996, pp. 132-146.
Morris, M. G., and Venkatesh, V. "Age Dif-
ferences in Technology Adoption Decisions:
Implicationsfor a ChangingWorkforce,"Per-
sonnel Psychology (53:2), 2000, pp. 375-403.
Motowidlo,S. J. "Sex Role Orientationand
Behavior in a Work Setting,"Journal of Per-
sonality and Social Psychology (42:5), 1982,
pp. 935-945.
Nunnally,J. C., and Bernstein,I.H. Psychometric
Theory(3rded.), McGraw-Hill, New York,1994.
Olfman, L., and Mandviwalla,M. "Conceptual
Versus Procedural Software Training for
GraphicalUser Interfaces:A LongitudinalField
Experiment,"MIS Quarterly(18:4), 1994, pp.
405-426.
Plouffe, C. R., Hulland,J. S., and Vandenbosch,
M. "ResearchReport: Richness Versus Parsi-
mony in ModelingTechnologyAdoptionDeci-
sions-Understanding MerchantAdoptionof a
Smart Card-Based Payment System," Infor-
mation Systems Research (12:2), 2001, pp.
208-222.
Plude, D., and Hoyer,W. "Attentionand Perfor-
mance: Identifyingand LocalizingAge Defi-
cits," in Aging and Human Performance, N.
Charness (ed.), JohnWiley&Sons, NewYork,
1985, pp. 47-99.
Porter, L. "JobAttitudesin Management: Per-
ceived Importanceof Needs as a Functionof
Job Level," Journal of Applied Psychology
(47:2), 1963, pp. 141-148.
MIS QuarterlyVol.27 No. 3/September2003 473
Venkateshet al./UserAcceptance of IT
Rhodes, S. R. "Age-RelatedDifferencesin Work
Attitudesand Behavior:A Reviewand Concep-
tual Analysis," Psychological Bulletin (93:2),
1983, pp. 328-367.
Roberts, R. W. "Plasteror Plasticity:Are Adult
WorkExperiencesAssociated withPersonality
Change in Women?," Journal of Personality
(65:2), 1997, pp. 205-232.
Rogers, E. Diffusionof Innovations,Free Press,
New York,1995.
Rogers, E. M., and Shoemaker, F. F. Com-
munication of Innovations:A Cross-Cultural
Approach,Free Press, New York,1971.
Sachs, R., Chrisler, J. C., and Devlin, A. S.
"Biographicand Personal Characteristicsof
Women in Management,"Journalof Vocational
Behavior(41:1), 1992, pp. 89-100.
Sheppard,B. H., Hartwick,J., and Warshaw,P. R.
"The Theory of Reasoned Action: A Meta-
Analysis of Past Research with Recommen-
dationsforModifications and FutureResearch,"
of Consumer Research
Journal (15:3), 1988,
pp. 325-343.
Straub, D., Limayem, M., and Karahanna, E.
"MeasuringSystem Usage: Implicationsfor IS
TheoryTesting,"ManagementScience (41:8),
1995, pp. 1328-1342.
Szajna, B. "EmpiricalEvaluationof the Revised
TechnologyAcceptance Model,"Management
Science (42:1), 1996, pp. 85-92.
Taylor,S., and Todd, P. A. "Assessing ITUsage:
The Role of PriorExperience,"MIS Quarterly
(19:2), 1995a, pp. 561-570.
Taylor,S., and Todd, P. A. "UnderstandingInfor-
mation Technology Usage: A Test of Com-
petingModels,"Information Systems Research
(6:4), 1995b, pp. 144-176.
Thompson,R. L., Higgins,C. A., and Howell,J. M.
"Influenceof Experienceon PersonalComputer
Utilization: Testing a Conceptual Model,"
Journal of ManagementInformationSystems
(11:1), 1994, pp. 167-187.
Thompson,R. L., Higgins,C. A., and Howell,J. M.
"Personal Computing:Toward a Conceptual
Model of Utilization,"MIS Quarterly(15:1),
1991, pp. 124-143.
Tornatzky,L. G., and Klein, K. J. "Innovation
Characteristicsand Innovation Adoption-Imple-
mentation:A Meta-Analysisof Findings,"IEEE
474 MISQuarterly
Vol.27 No.3/September
2003
Transactions on Engineering Management
(29:1), 1982, pp. 28-45.
Triandis,H. C. InterpersonalBehavior, Brooke/
Cole, Monterey,CA, 1977.
Twenge, J. M. "Changes in Masculine and
FeminineTraitsOverTime: A Meta-Analysis,"
Sex Roles (35:5/6), 1997, pp. 305-325.
Vallerand,R. J. "Towarda HierarchicalModelof
inAdvances
Intrinsicand ExtrinsicMotivation,"
in Experimental Social Psychology (29), M.
Zanna(ed.), AcademicPress, New York,1997,
pp. 271-360.
Venkatesh,V. "CreatingFavorableUser Percep-
tions: Exploringthe Role of IntrinsicMoti-
vation,"MIS Quarterly(23:2), 1999, pp. 239-
260.
Venkatesh, V. "Determinantsof Perceived Ease
of Use: Integrating Perceived Behavioral
Control,ComputerAnxietyand Enjoymentinto
the Technology Acceptance Model,"Informa-
tion Systems Research (11:4), 2000, pp. 342-
365.
Venkatesh, V., and Davis, F. D. "ATheoretical
Extension of the Technology Acceptance
Model: FourLongitudinalFieldStudies,"Man-
agement Science (45:2), 2000, pp. 186-204.
Venkatesh,V., and Morris,M.G. "WhyDon'tMen
Ever Stop to Ask For Directions? Gender,
Social Influence,and TheirRole in Technology
Acceptance and Usage Behavior," MIS
Quarterly(24:1), 2000, pp. 115-139.
Venkatesh,V., Morris,M.G., and Ackerman,P. L.
"A LongitudinalField Investigationof Gender
Differencesin IndividualTechnologyAdoption
Decision MakingProcesses," Organizational
Behavior and Human Decision Processes
(83:1), 2000, pp. 33-60.
Venkatesh, V., and Speier, C. "ComputerTech-
nology Trainingin the Workplace: A Longi-
tudinalInvestigationof the Effectof the Mood,"
OrganizationalBehaviorand HumanDecision
Processes (79:1), 1999, pp. 1-28.
Warshaw, P. R. "A New Model for Predicting
BehavioralIntentions:An Alternativeto Fish-
bein,"Journal of MarketingResearch (17:2),
1980, pp. 153-172.
Warshaw,P. R., and Davis, F. D. "Disentangling
Behavioral Intentionand Behavioral Expec-
tation,"Journalof ExperimentalSocial Psych-
ology (21:3), 1985, pp. 213-228.

West, S. G., and Hepworth,J. T., "Statistical
Issues in the Study of TemporalData: Daily
Experiences," Journal of Personality (59:2),
1991, pp. 609-662.
Westland, J. C., and Clark,T. H. K. GlobalElec-
tronicCommerce: Theoryand Case Studies,
MITPress, Cambridge,MA,2000.
Wong, P. T. P., Kettlewell,G., and Sproule, C. F.
"Onthe Importanceof Being Masculine: Sex
Role, Attribution, and Women's Career
Achievement,"Sex Roles (12:7/8), 1985, pp.
757-769.
About the Authors
Viswanath Venkatesh is Tyser Fellow,associate
professor, and directorof MBAConsultingat the
RobertH.SmithSchool of Business at the Univer-
sity of Maryland,and has been on the facultythere
since 1997. He earned his Ph.D. at the University
of Minnesota'sCarlson School of Management.
His research interests are in implementationand
use of technologies in organizationsand homes.
His research has been publishedin many leading
journals including MIS Quarterly, Information
Systems Research, ManagementScience, Com-
munications of the ACM, Decision Sciences,
OrganizationalBehavior and Human Decision
Processes, PersonnelPsychology. Heis currently
an associate editor at MIS Quarterlyand Infor-
mation Systems Research. He received MIS
Quarterly's"Reviewerof the Year"awardin 1999.
Michael G. Morris is an assistant professor of
Commercewithinthe Information Technologyarea
at the MclntireSchool of Commerce,Universityof
Virginia. He received his Ph.D. in Management
InformationSystems from Indiana Universityin
1996. His research interestscan broadlybe clas-
sified as socio-cognitive aspects of human
response to informationtechnology,includinguser
acceptance of informationtechnology, usability
engineering,and decision-making. His research
has been published in MIS Quarterly,Organiza-
tionalBehaviorand HumanDecision Processes,
and Personnel Psychology, among others.
Venkatesh
et al./UserAcceptanceof IT
Gordon B. Davis, Honeywell Professor of
ManagementInformationSystems in the Carlson
School of Managementat the Universityof Minne-
sota, is one of the founders of the academic
disciplineof informationsystems. He has lectured
in 25 countriesand has written20 books and over
200 articles,monographs,and bookchapters. He
participatedin and helped form the majoraca-
demic associations related to the field of man-
agement informationsystems. He has been
honored as an ACM Fellow and an AIS Fellow,
and is a recipient of the AIS LEO award for
lifetime achievement in the field of information
systems. His research interests include concep-
tual foundationsof informationsystems, informa-
tion system design and implementation,and
managementof the IS function. He has a Ph.D.
fromStanfordUniversityand honorarydoctorates
fromthe Universityof Lyon,France,the University
of Zurich,Switzerlandand the StockholmSchool
of Economics, Sweden.
Fred D. Davis is DavidD. Glass Endowed Chair
Professorin InformationSystems and Chairof the
InformationSystems Departmentat the Sam M.
Walton College of Business at the Universityof
Arkansas. Dr. Davis earned his Ph.D. at MIT's
Sloan School of Management,and has served on
the business school faculties at University of
Michigan,Universityof Minnesota,and University
of Maryland.He has taughta wide range of infor-
mation technology (IT) courses at the under-
graduate, MBA,Ph.D., and executive levels. Dr.
Davis has served as associate editor for the
scholarly journals Management Science, MIS
Quarterly,and InformationSystems Research. He
has publishedextensively on the subject of user
acceptance of ITand IT-assisteddecision making.
His research has appeared in such journals as
MIS Quarterly,InformationSystems Research,
Management Science, Journal of Experimental
Social Psychology, Decision Sciences, Organiza-
tionalBehaviorand HumanDecision Processes,
Communicationsof the AIS, and Journalof MIS.
Currentresearch interests includeITtrainingand
skillacquisition,managementof emergingIT,and
the dynamicsof organizationalchange.
MISQuarterly
Vol.27 No.3/September
2003 475
 et al./UserAcceptanceof IT
Venkatesh

Appendix A

Cautions Related to Pooling Data and Associated Statistical

Tests for PreliminaryTest of UTAUT(Studies 1 and 2)

The most criticalconcern when poolingrepeated measures fromthe same subjects is the possibilityof
correlatederrors. West and Hepworthdescribe this problemas follows:

A perusalof the empiricalliteratureoverthe past 30 years revealssome of the statistical

approaches to repeated measures over time that have been used by personality
researchers. The first and perhaps most classic approach is to aggregate the
observations collected over time on each subject to increase the reliabilityof
measurement (1991, p. 610).

However,in manycases, these observationswilloften violatethe assumptionof independence, requiring

alternateanalyses (see West and Hepworthfor an extended discussion of alternateapproaches).

When errorterms can be shown to be independent,West and Hepworthnote that "traditionalstatistical

analyses such as ANOVAor MR[multipleregression]can be performeddirectlywithoutany adjustment
of the data."(pp. 612-613). The best way to determinewhetherit is appropriateto pool data is to conduct
a test for correlatederrors. In orderfor it to be acceptable to pool the data, the errorterms should be
uncorrelated.A second approachuses bootstrappingto select subsamples to conducta between-subjects
examinationof the within-subjectsdata. Inorderfor it to be acceptable to pool the data, this second test
should yield identicalresultsto the test of the modelon the complete data set. Belowwe reportthe results
fromthe specific findingsfromthe correlatederrorstest and the patternof findingsfromthe second test.

CorrelatedErrorsTest
We computed the errorterms associated withthe predictionof intentionat T1, T2, and T3 in studies 1
(voluntarysettings) and 2 (mandatorysettings). Further,we also calculatedthe errorterms when pooled
across both settings-i.e., for cross-sectional tests of the unifiedmodel at T1, T2, and T3 respectively.
These computationswere conductedbothforthe preliminary test andthe cross-validation(reportedbelow).
The errortermcorrelationsacross the intentionpredictionsat variouspointsintimeare shown below. Note
that all error correlationsare nonsignificantand, therefore, not significantlydifferentfrom zero in all
situations.

Between-Subjects Test of Within-SubjectsData

Whilethe results above are compelling,as a second check, we pooled the data across differentlevels of
experience (T1, T2, and T3) and used PLS to conducta between-subjectstest using the within-subjects
data. We used a DOS version of PLS to conductthe analyses (the reason for not using PLS-Graphas in
the primaryanalyses in the paper was because it did not allow the specification of filteringrules for
selecting observationsin bootstraps). Specificallyforthe between-subjectstest, we applieda filteringrule
that selected any given respondent'sobservationat only one of the three time periods. This approachen-

476 MIS QuarterlyVol.27 No. 3/September2003

 Venkateshet al./UserAcceptance of IT

PeriodsSS
Tabe l.CoreatinsBewen r oir Terms of~ ~E
Intention Construct
S*
at VariousTimel

Study I (Voluntary) T1 T2 T3
T1
T2 .04
T3 .11 .09
Study 2 (Mandatory) T1 T2 T3
T1
T2 .07
T3 .08 .13
Study 1 and 2 (Pooled) T1 T2 T3
T1
T2 .06
T3 .09 .10

sured thatthe data includedto examine the interactiontermswithexperience didnot includeany potential
forsystematiccorrelatederrors. Using50 such randomsubsamples, the modelwas tested and the results
derivedsupportedthe patternobserved when the entiredata set was pooled across experience.

Taken together,the analyses reportedabove supportthe poolingof data (see Table 17) across levels of
experience and eliminatethe potentialstatisticalconcerns noted by West and Hepworthin the analysis of
temporaldata.

MISQuarterlyVol.27 No. 3/September2003 477

Venkateshet al./UserAcceptance of IT

Appendix B

Statistical Tests Priorto Pooling of Datafor

Cross-Validationof UTAUT(Studies 3 and 4)

As withthe test of the preliminarymodel, priorto poolingthe data forthe cross-validationstudies (studies
3 and 4), we conducted statisticaltests to examine the independence of observations (as detailed in
AppendixA). The table below presents the errorterm correlationmatrices for intentionfor studies 3
(voluntary)and 4 (mandatory)as well as pooled across both settings at T1, T2, and T3 respectively.

As inthe preliminary test of UTAUT,the errorcorrelationmatricesabove suggest thatthere is no constraint

to pooling in the cross-validationstudy of the model. As before, a between-subjects test of the within-
subjects data was tested using PLS (as described inAppendixA), andthe resultsof thattest corroborated
the independence of observationsin this sample. In lightof both sets of results,we proceeded withthe
pooled analysis as reportedin the body of the paper (see Table 21).

ToBoIoV

Perod

Study 3 (Voluntary) T1 T2 T3
T1
T2 .01
T3 .07 .11
Study 4 (Mandatory) T1 T2 T3
T1
T2 .04
T3 .02 .08
Study 3 and 4 (Pooled) TI T2 T3
T1
T2 .03
T3 .05 .10

478 MISQuarterlyVol.27 No. 3/September2003