AN EMPIRICALINVESTIGATIONINTOTHE

RELATIONSHIPBETWEENCOMPUTERSELF-EFFICACY,
ANXIETY,EXPERIENCE,SUPPORTAND USAGE
MARYHELENFAGAN
UnivenityofTexasatTyler,Texas
Tyler,Tens75799
STERNNEILL
UnivenityofWashington,Tacoma
Tacoma,Washington98402
BARBARAROSSWOOLDRIDGE
UnivenityofTampa
Tampa,Florida33606
ABSTRACT
Organizations make significantinvestments in information
technology. However, if individuaJs do not use infonnation
system applications as anticipated, successful implementation
can be hard to achieve. In order to investigatesome key factors
thought to affect an individual's use ofinformation technology.
this study draws 00 Bandura's Social Cognitive Theory (SCl"),
Triandis's Theory of Interpersonal Behavior (TIB). and the
computeranxiety literatUre todevelop itsconceptual model and
research hypotheses. An empirical invcstigation (0...,78) found
support for the majority ofthe hypothcses. As suggcsted by
scr. experience and suppon were positively related to
computer self-efficacy. and computer self-efficacy was
negatively related to anxiety and positively related to usage. As
suggested by TIE. experience was positively related to usage.
Furthermore.. computer anxiety was negatively related to
experience. By providing insight into these imponant
relationships. this research can help funher understanding of
theirroleintheacceptanceanduseof information h.. -chnology.
Keywords: computer self-cfficacy. computer anxiety.
computer experience, computer suppon, Social Cognitive
Theory.TheoryofInterpersonal Behavior
INTRODUCTION
Organizations make significant investments in information
technology(IT). However. if individualsdo not useinfonnation
system applications as anticipated. successful implementation
can be hard to achieve. Since the 1970s, infonnation systems
researchers have investigated a numberoffactors that influence
system usageand success(7, 22,60).andavarietyoftheoretical
and pragmatic explanations have been developed to help
understand and improve the successful deployment of IT.
Research on the factors that influenceinformationsystemsusa.ge
continuestobe thefocusof intensive,ongoingresearch(39.50).
In orderto investigatesomekeyfactorsthoughtto affectan
individual's use ofinformation technology. thjs study draws on
Bandura'sSocial Cognitive Theory (SCr), Triandis'sThcoryof
Interpersonal Behavior (11B), and the computer anxiety
lit.ernlure to develop its conceptual model and research
hypotheses. An empirical invcstiga'ion (n='i78) fouod
significant support for most ofthe hypothcsized relationships
between computer self.-efficacy. anxiety, experience, support
andusage. By combiningkeyconstructsfrom SCTand118into
one conceptual model. along wilh additional key relationships
addressingcomputeranxiety, lhis stud)' makesa contribution to
the growing literature in this area. This study provides insight
into key constructS and relalionships, and adds to the
understandingofthe role 'heyplayin 'heacceptance and uscof
informationtechnology.
This paper is organized as follows. First, background is
given on the twO theoretical perspectives. Social Cognitive
Theory and theTheory ofInlerpersonal Behavior, tha' underlie
the research. In addition, lhe literarure on computer anxiety is
briefly reviewed. Then, the study's conceptual model and
research hypothesesare presented. Nexl, derailsare providedon
the melhod and the analysis tha' support the study's empirical
investigations and its resullS. Finally. lhe outcomes and
implications of the study are discussed and conclusions are
dmwn.
THEORETICAL BACKGROUND
In orderto investigatesomekey factorsthoughtto affect an
individual's use ofinfonnation technology. this study draws on
two major theories from social psychology: Bandura's Social
Cognitive Theory and Triandis's llleory of Interpersonal
Behavior. In addition, the literamTe on computer an.xiety
provides support for the study's conceptual model and
hypotheses. This section pro ides a brief overview of the
relevant theoreticaJ and empirical literature that infonns this
study.
SocialCognitiveTheory
Social CognitiveTheory(SCf) is a "theoretical framework
for analyzing human motivation, thought and action" that
"embraces an interaetionaJ model of causation in which
environmental events, personal factors and behavior aU operate
as interactive determinants ofeach other" (IO. p. xi). A key
concept in SCT is perceived self-efficacy which refers to the
beliefan individual has in hislherabilityto successfuJlyperform
a certain behavior (8). Self-cfficaey is conceptualized by
Bandura as varying across tasks and situations. and has a
number ofdeterminanlS (27). Research thai has measured self
efficacy in regard. to specific tasks has proven to have more
prcdieativepower(12). Research has shown that an individual's
Winter20032004 JournalofComputerInformationSystems 95
self-efficaey has 8 direct influence on hislher choice oftaskand
persistence in achieving the task. Low self-efficacy beliefs, for
example. have been found to be negatively related to subsequent
task performance(8).
In developing an integrative framework for research on
computer selfefficaey, researchers found that Bandura and
others had identified over twenty-Lhree antecedent and
consequentfactorsthaiare theoreticallyrelaledtocomputerself-
efficacy (40). Figure I is a model ofa subset ofthe factors
related to self-efficaey, As the model illustrates, cnactive
mastery should be rdOled 10 self-efficacy since. as SCT posits,
these experiences "arc the most influential source of emeacy
information because they provide the most authentic evidence of
whetheronecan mUSIer whalever it takestn succeed" (8, p. 80).
Situational support should also be related 10 self-efficacy, since,
as SCT posits, "people who are socially persuaded that they
posses the capabilities to master difficult situations and are
provided with provisional aMs for effective action are likely to
mobilize grealer e[fort" (9, p. 198). Emotional arousal is
expected to have a negative effect on sclf-efficacy. resulting in
increased levels nfanxiety (II). Ifincreased anxiety leads to
subsequent increases in emotional arousal then a potentially
debilitating cycle ofanltiety can be CreaIed (40). Finally, high
levcls of anxicty can affect behavior. leading 10 lowered
performance(8).
FIGURE I
Self-eflie.cyModel lAd.pledfrom M.r....... Yi andJohnson(40)1
Enactive
Mastery

Situational Self-Eflicacy
Support
..
Emotional
V
Arousal
In the lale 1980sand early 1990s a numberofInformalion
Systems (IS) researchers investigated the relationship between
self-efficacy and computer-related behaviors and attitudes. The
rela'ionship between self-efficacy and software training
behavior was explored (28, 59) as well as the rdalionship
between compuler self-efficacy and the adoption of high
technology products (31). In addition, the concepl of self-
cfficacy plays a key role in the Technology Acceptance Model
(TAM), In his discussion of the theoretical foundations of
model's constructs, Davis indicated that TAM's perceived ease
of use concept is similar to Bandura's definition of self-efficacy
(21). Subsequenl research has investigated whether a person's
perceptions of case of use are anchored to their computer self-
efficacy(33.56). In 1995. SITwasused e.xpliciUy as the basis
for a research model that tested the relationship of computer
self-efficacy and seven other factors including support. anxiety,
and usage (20). Additional research has explnred the
determinants of computer selfefficacy (26) and tbe relationship
between general and specific computer self-efficacy (I). Two
lines of ongoing research may be seen in the IS arena now: I)
research building upon the TAM literature that views computer
selfeflicacy as an determinant of TAM's perceived ease of use
and 2) research based upon SCI'thai posits a cenlra1
role for computer seLf-efficacy as 8 direct dctenninant of
behavior.
Theory of Interpersonal Behavior
Triandis developed a Theory of Interpersonal Behavior
(TIB) !hat posits !hat habits, intentions and facilitating
conditions predict the probability that an act will be performed
(55). Also, in the 11B, social factors, and perceived
consequences of performing the behavior are postulated to
determine intention. TIS is a model of behavioral intention that
Behavior
AnXiety
\ l
provides a theorelical alternative to Ajzen's Theory of Planned
Behavior (TPB) (3). A model thai shows the key factors in the
11Bisprovided in Figure2[adaptedfrom (55)].
IS researchershaveused all orpart ofthcTIBasa basis for
studyingindividual ITacceptanceandusagebehavior(4. 13. 16.
17, 44, 51, 52). In particular, two factors have been utilized
extensively in IS reseaIch: I) habil (often operationalized as
prior computerexperience) and 2) facilita,ing conditions (often
operationalized as various fOnTIS of organizationaVcompul.Cr
support). According10 Triandis, habit strength is "measured by
the number oftimes the act has already been performed by the
person" (54. p. 9) and. thus, as performance of a behavior
increa.ses, its effect on later behavior is expected to increase.
And. although past behavior plays no theoretical role in TPB,
Ajzen found. after 8 review, that "researchers may want to
include a measure of prior behavior in our modelS 10 improve
predictability oflaLcr aclion" (2, p. 120). Similarly. Triandis's
conception of facilitating conditions has influenced IS research.
Facilitating conditions includes access to time. people, money.
or other resources needed to perfonn a behavior. and is imilar
to the perceived behavioml control construel in 11lP (3). Thus,
researchers have investigated the role that habit/computer
experience and facilitating conditions/organizational suppon
play in IT usage both within models based upon 11B and when
usingmodclsbasedonotherapproachessuch asTPB.
Computer Anxiety Literature
Bandura states that otself-efficacy theory suggcslS an
alternativeway oflookingaI human anxiel)l" (10. p. 439). Self-
efficacy theory. as described above, postulates relationships
between emotional arousal. self-efficacy. anxiet)', and behavior.
Many other factors have been aJso explored in relation to
anxiety within the larger theoretical BIld empiricallitcralun:.
Winter 2003-2004 Journal ofComputerInformation ystems 96
FIGURE2
TriandisModel(Adapted from Triandis)
Affect Habit
~
Social Factor.; ~
Intention Beha\1or
.. ..

Perceived Facilitating
V V
Consequences Conditions
Wilhin lile IS research area. anumber ofresearchers have
explored anxiety,oreven fear, thatsome people may experience
when they confront the possibility of using a compuu:r.
Computer anxiety is viewed as anegative emotional reaction or
e!Tect (53) and has been studied as part ofa larger research
stream often termed technophobia orcomputerphobia (46, 47).
Computer anxiety been shown to have8 significant relationship
to key IT constructs su has attitudes toward o p u t ~ usage
intention, usage behavior, and performance(14,18,23.30,57).
IT research has also explored computer anxiety's role as a
dctenninant of perceived case of usc, a key variable in the
Technoiogy Acceptance Model [56]. While computeranxicty is
recognized as an importa1l1 factoT, much remains to be
understood about its role from a lhcorclicaJ perspective. For
example. in his review Marakas concludes: "somewhat
counterintuitive. however, is the apparent lack of global
recognition by theCSE lcomputerself-efficacy) literatureofthe
imponance of thc anxiety relationship and by the
computerphobia literature of the potential value of C E
manipulation and enhancement in reducing anxiety," and he
argues that the complementary relationship between the
computerphobia and computer self-efficacy research calls for
furtherresearcb (40, p. 148). A recentaniclethat examined the
individual I.r8..ilS that are antecedent to computer anxiety ond
computer self-efficacy has begun to address this research gap
(SO).
CO CEPTUALMODELAND RESEARCH
HYPOTHESES
This study's conceptual model and research hypotheses
build upon the SCT. TIB and the computer anxiety Iiterarure.
Based upon the SCT literaruro, the autho..expect I)computer
experienceand organizational support to be positivelyrelated to
computer scLf-.cfficacy, 2) computer self-efficacy to be
negatively related to computer anxiety and positively related to
computcr usage, and 3) computer anxiety to be negatively
related to usage. Based upon the TIB literature, the authors
expect computer experience and organizational support to both
be directly and positively related to computer usage, finally,
based upon the computer anxiety literature. the authors expect
negative relationships between computer anxiety and two
determinants: experience and suppon. The study's conceprual
model is depicted in Figure3. (Note: each ofthe hypothesesare
designated by a label that is referenced later in the iext and
annotated with a (+) to indicate a positive hypothesi.zed
relationship and a(-) to indicate anegativeone.)The remainder
ofthis section provides an overview ofthe IS literature related
to each construct in the model. and lays out the studys
bypotheses.
ComputerSelf-Efficacy
SCT expects that an individual's self-efficacy has a direct
innuence on hisfller choice of task and their persistence in
achieving the task. In the computer acceptance literature. a
numberofstudieshave found that perceivedhigh computerself-
efficacy is related to the use ofa variety oftcchnologically
advanccdproducts(IS,20, 31).Thus, the following isproposed:
!:!!: Computer self-efficacy will be positively related to
computerusage.
Researchers have also bypothesized thaI computer self-
efficacy and computer anxiety arc inversely related. and studies
have found that individuals with lower levels ofanxiety will
have higher levels ofcomputer self-efficacy (36. 37, SO. 58).
Thus.thefollowing is proposed:
H2: Computer sclf-efficacy will be negatively related to
computeranxiety.
ComputerAO.J.iety
Anxiety isan unpleasantemotional reaction experienced by
individuals in threatening situations and the use ofacomputer
appean: to provide a fertile environment for such reactions (19,
41). Since anxiety will often cause people to avoid situations
that trigger these feelings, researche.. t)'pically expect an
inverse relationship between computer anxielY and computer
use. Many studies have found SlJppon for the expected
relationships (20, 33, 34, 37, 41, 58). Thus, the following is
proposed:
IlJ.: Computer anxiety will be negatively related to
computerusage.
Computer Experieou
10 SCf.coactivemastery is posited to predict sclf-efficacy.
In IS research. prior computer experience has been shown to be
ak.ey individual difference variable that predicts computer self
efficac), in a variety of IT applications (I. 20, 33, 40).
Researchers have found, for example. that prior Internet
experience was the strongest predictor ofInternet self-efficacy
(24).Thus,the followingisproposed:
H4: Computcr experience will be positively related to
computerself..-efficacy.
Winter20032004 Journal ofComputerInformation Systems 97
Vician and Davis report mat the most consistent finding of
studies of correlations with computer anxiety is that prior
computer experience has a negative relationship with computer
anxiety (57). In general, people with less experience are more
likely to beanxious when confronted with IT with which they
are unfamiliar(34,46).Thus,the following is proposed:
ill: Computer experience will be negatively related to
computer anxiety.
Furthermore. the TIB proposes that past behavior,
especially in the form of can be a primarydeterminantof
behavior (49). In the IS discipline, a number ofstudies have
found SUppoR for a direct relationship between prior experience
and computer usage (20. 35, 40, 48, 52). One study that
hypothesized no direcl effect of computer experience on
computer usage, based upon the expectation that other variables
would mediate the relationship. found. in fact a significant
relationship between computer experience and computer usage
(33).Thus,the followingisproposed:
H6: Computer experience will be positively related to
computer usage.
fiGURE3
Research Conceptu.al Model
H6(+)
I
Computer
H4 (+)
Computer
Experience
I H5 (-)
--.;
Self-effICacy
H2 (-)
HI (+)
r
Computer
I H7 (+)
Usage
OrganiZlltiona] Computer


Support
H8 (-)

Anxiety
I
H9 (+)
(+) indicatesa positivehypothesi2edrelatIonshipand(-)a negallveone
OrganiulionalSupport
SCT posits thai situational suppon is one of lhe factors that
affecl self-cfficacy. A number of IS researchers have found
support for the proposition that support. of various types.
increases lhe ability of end-users and thus resullS in increased
self-efficacy(20, 33).Thusthefollowing is proposed:
l!l: Organizational support will be positively related to
computer self-efficacy.
Many of the approaches used to reduce computer anxiety
involve makmgsure that situational support is provided so that
an individual perceives there is somewhere 10 lurn for help (18.
53). FOT an individual who is very anxious about intCTBClion
With a particularcomputing technology, further e'posureto the
technolog) alone may not result in reduced anxiety. As an
example, Vician and Davis expect that "developing an
appropriate learning environment for a computing intensive
course will be key to providing a beneficial situation for all
learners"(57,p. 47).Thus,thefollowingisproposed:
H8: Organizational support will be negatively related '0
computer anxiety.
A number of researchers have also recognized the role
organizational support can play in computer usage. and many of
Ihese studies build upon the lIBto conceptualize the role Ihal
facilitating conditions have in infonnation systems usage (55). A
recent study found 8 significant positive relationship between
facililating conditions and computer usage (17). However, !.he
results of otht:r studies have been mixed. One study found the
relationship to be non-signific8m (52) while another found a
negative relationship (S I). Building upon the work of Triandis
andothers,thefollowing isproposed:
H9: Organizational support will be positively related to
computer usage.
THESTUDY
Subjects
The sample consists of 978 business school students
attending a major Southeastern university. Data was collected
over a one-week period. and during this timeframe the survey
was distributed in all business classes. Students were not given
ex!Ttl credilto respond to this "",eyand therefore they had no
incentive to respond to more than one survey. This study's
questions were embedded in a larger survey on the overall
evaluation of me business school. The survey had a
response rate of 4st'1ct. While me use ofstudents in research is
not without controversy (29. ]2. 42), the autho'" believe that
business students are an appropriatc population for this research.
Hughes and Gibson (32) suggest that the use ofstudents needs
to be reviewed for each study for applicability and since the
behavior thaI was measured in this study was the usage ofthe
college computcr lab. business students are appropriate subjects.
Measures and Validation
A questionnaire was dcveloped for this sludy. Most of the
measures were adapted from existing scales that had
demonstrated validity and rcliabi lily in other studies: a)
computer self-cfficacy (42). b) computer anxiety (19) and c)
computer usage (21). The $ole measuring computer
Winler2003-2004 Journ.1 ofComputerInformation Systems 98
organizational support was developed specially for the study's To assess the unidimensionaJity, the covariance matrix ofa five
computer lab environment based upon measures used in a factor model - Computer Experience, Organizational Support.
number of prior studies, and its development has been Compuler Self-Efficacy. Computer Anxiety, and Computer
documentedextensivelyelsewhere(6). Usage- wasevaluated usingL1SREL VIII (38). Thefit statistics
Thequestionnairewasdistributed duringclass.and thus the and internal consistency were examined to assess model fit..
survey length was a key issue in the design. The survey needed discriminantvalidityandreliability.
to be able to be completed within 20 minutes, and therefore Fit statistics and internal consistency estimates for the five
manyofthescaleshad tohereduced in length.Thisreductinn in factor model arereported inTahle I. Due to thescnsitivityofi
lengthofthescalestookplaceduringan extensivepreteststage. to sample size, the root mean square error ofapproximation is
reported as an assessment ofoverall fit. The model is in the
RESULTS acceptable range(.05 to .08) at .052. Thegoodncss-<>f-fit (GFI)
and the adjusted-goodncss-of-fit (AGFI) are .94 and .92,
MeasurementModel Results respectively. Because of inconsistencies due to sample
characteristics. the Tucker-Lewis lndex (TLI) and the
As recommended byAnderson and Gerbing(5), " two-step comparative fit index (CFI) are reponed. For the five factor
approach was adopted. The measurement aspect ofthe model model. the indices are ncar the .90 range (.94 llnd .95.
wasestimated priortotestinglhcstructural asPCCllO preventany respectively)which isdeemed acceptable.
interaction between the two models due to mC8Suremenl error.
TABLE I
MeasurementModel Estimates
FitStatistics
x'
df RMSEA GFI AGYI TLI CFI
655.4 17 .05 .94 .92 .94 .95
2 9
InternllConsistencyMelsures
Como.n AYE
ComputerExperience .70 .46
OrganizationalSupport .82 .48
ComputerSelf-Efficacy .93 .74
ComputerAnxiety .85 .54
Comou,crUsaoe .75 .50
Note. df degreesoffreedom. RMSEA - root mean squareerrorofapprOXimation, GFI - goodness-or-fit mdex,
AGFI- adjusted-goodness-of-fit index;TLI- Tucker-Lewisindex;CFI= comparative-fit index;Compn = compositealpha;
AVE =averagevarianceextracted
Toensurethal the model ismeasuringdistinctconSlIUclS. it StructuralModel Results
was also assessed for discriminant validity. The most stringent
test was performed by checking ifthe square ofthe parameter Toassessthe structural model, threecriteria were used: (I)
estimate between two constructs (cf) is less than the average the fit indices, (2) the significance of the completely
AVE between any two constructs(25). In all cases, discriminant standardized path estimates, and (3) the amount of variance
validitywassupported. explained in each oftheendogenous constructs. Table2 reports
To assess internal consistency. items demonstrating high the correlations among the latent constructs in the StrUcturaJ
within/across factor correlated errors (standardiz.ed residual> aspectofthe model.
2.57) were examined as' candidates for removal. Before deletion The same indices used to evaluate the measurement model
from the study. each indicator was f i ~ examined for its (RMSEA. GFI, AGFI. TLI. and CFI) were estimated for the
conceptual contribution and ifdeemed negligible was removed structural portion and assessed using the same criteria The
from the srudy. In all, seven itemswereremoved from thestudy results inTable3 indicateadequatefit for thefive-factor model.
due to high standardized residuals. As further evidence of TheS1nIcturaJ equation modelingresultsindicatethereisno
internal consistency, the composite reliability estimates arc empirical relationship berween Organizational Support and
examined (see Table I). The reliability estimates ranged from ComputerAnxiety (B8)orherween Organizational Support and
.70 to .93, indicating acceptable reliability for the constructs. Computer Usage (H9). As indicated in Table 3, seven ofthe
Also, all items have significant t-value loadings for lheir nine paths BIe significant (p < .05 or bener). One hypothesis
respective construets (p < .01). Average variance extracted (B3) is statistically significant. but not in the hypothesized
estimates are also reported (Table I). AVE estimates of.45 or direction. Figure 4 shows the hypotheses that were supported
higher are an indication ofvalidity for 8 construct's measure with asolid!.inc. alongwith theirpathestimates.
(43). Eachconstructmeetsthesecrileria.
Winter2003-2004 JournalofComputer InformationSystems 99
AdditionaJly. the strueturdl equations account for 38% of DISCUSSION
me variance in Computer SelfEfficacy. 58% ofthe variance in
ComputerAnx.iety. and 4%oflhevariance in Computer Usage. This study developed a conceptual model based upon the
The faiJure ofthe model to account for greater variation in SCf, TTB and the computer anxiety literature. Six of the
computer usage indicates a need to examine additional factors hypotheses drdwn from this theoreticaJ and empirical literature
related to use. were supported and three hypotheses were not supported. The
resultsarc summarized in Table4.
TABLE2
CorrelationsAmongConstructs
Construct 1 2 3 4 5
(I) ComputerExperience 1.00
(2) Organizational Support -0.09 1.00
(3) ComputerSelf-Efficacy 0.44 0.04 1.00
(4) ComputerAnxiety -0.39 0.02 -0.53 1.00
(5) ComputerUsage 0.01 -0.03 0.06 0.09 1.00
TABLE3
Structural Model Estimates
FitStati!Jtics
x'
df RMSEA GFI AGFI TLI CFI
655.42 179 .052 .94 .92 .94 .95
Path Patb Estimates
ComputerExperience-.Computer YII .60'
ComputerExperience-+ ComputerAnxiety:121
-.4S'
ComplttcrExperience-.ComputerUsage:131
.13'
OrganiUltional Support...ComputerSelf-Efficacy:Y,1
.07'
.00n.s.
Organizational Suppon4'ComputerAnxiety:'(21
-.02n.s.
Organizational Support...ComputerUsage: Y"
-.39'
ComputerSelf-Efficacy...CompulerAnxiety: 6"
.23'
ComputerSelf-Efficacy...CompulerUsage: p"
_26'
ComnuterAnxiety...Comnuter Usa.e:
*p< .OS: n.s. - notSIgnificant
FIGURE4
HypothesizedRelationshipsAmongLatentConstructs
. - - -
1"
6
(.13')
Computer
H4 (.60")
Computer
Experience Self-efficacy

I H5 (,.48,) -
("omputer
112(,39')
,
)II
Usage
I H7 (D7') ,

, -
OrganizationaI Computer
,
;
113 (.26')
Support Anxiety

118 (DO)
,

. 119 (,02)
---
.J\.s expected from SeT,both priorcomputerexperienceand relationship with computer usage (Hl). As expected from 11"8,
organizational supporthad apositive relarionship with computer computer experience had a positive relationship with usage
self-efficacy (H4 and H7). and computer self-efficacy had 8 (H6). As expected from the computer anxiety literature.
negative relationship with computeranxiety (H2) and a positive computer experience had anegative relationship with computer
Winter2003-2004 Journalo(ComputerInformationSystelUs 100
anxiety (H5)_
The hypothesis drawn from SCT that p<>siled that computer
anxiety would have a negative relationship 'Hith usage was nOl
supp<>rted (H3)_ In fact. computer anxiety had a p<>sitive and
significant relationship with usage. The counterintuitive effect of
H) may be due 10 the fact that we measure amount oflime in the
lab. It may take anxious students more time to accomplish the
task than a less anxious student (even when taking into account
experience and computer self-efficacy). AIs<>, our study did nol
assess whether the participants usage behavior was voluntary or
not If participants were required to usc (e.g., to do
work for a class) then individuals would nOl be able to avoid
computer use when anxious. Instead. one might find that their
anxiety results in decreased perfonnanee. and even require more
time spent using the computer. Students might be
overcompensating to Qvt."fCome lheir fear. especially if they must
adopt the technology to succeed in their future career. Further
research would have to explore these possibilities.
TABLE 4
Results of Hypotheses Tests
HVDOth..i. Results
HI: Computer self-efficacy will be positively assoeiated with computer usa'e. Suooorted
H2: Comouter self-efficacv will be nel!ativelv associated with comouter anxietv. SUDoorted
In: Computer anxiety will be assoeiated with computer US8j!;e. Not Supp<>rtcd
1-14: Comoutcr exoerience will be oosilivelv associated with comDuter self-efficacv. SUDOOrtcd
H.5: Computer experience will be nC$t8tively associated with computer anxiety. Supported
116: ComDuter cxoerience will be oositivclv associwcd with COffiouter usa2.C. SUDOOrtcd
H7: Or1!ani11u.ional supoort will be positively assoeiated with compuler self-efficacy. SUPlXlrted
H8: Organizatiooal supp<>rt will be negatively assoeiated with compuler anxiety. NOl Supp<>rted
119: Organizational supp<>rt will be posilively associaled with computer usage. Not Supp<>rted
The hypothesis drawn from TIB that p<>sited that
organizational suppon would have a direct positive relationship
with computer usage was not supported (H9). There are several
p<>ssible explanalions for these results. II may be. as Taylor and
Todd suggest, that "the absence of facilitating resources
represents barriers to usage and may inhjbit the fannalian of
intention and usage: however the presence of faciliLaLing
resources may per so, encourage usage" (48, p. 153).
Another possible reason for the non-support of this hypothesis
might be the items that were used to measure organizational
support. The organizational suppon items were developed
specifically for this study (6), and il could be that items used in
other studies to measure individual's perceptions of facilitating
conditions (33) would have yielded differenl results. This
rationale might als<> explain why the hypothesis from the
computer anxiety literature that posited that organizational
support would have a negative relationship with computer
anxiety was not supported (H8). Funher research would have to
cxplore this p<>ssibility.
The results of this study provide supp<>rt for key hyp<>theses
drawn from Bandura's SeT. Practical implicalions of these
nndings suppon the training literature that encourages efforts to
build computer sclf-efficacy. When individuals have
experiences that build their mastery of IT applicatiOns and arc in
an environment with positive situational support. they tend to
have higber levels of computer self-efficacy. High computer
scII-efficaey. in tum.,. is associated with usage. The findings that
suppon Triandis's 118 also indicate that prior experience may be
a direct detenninant of usage. One possible implication from this
result may be to suggest that computer self--cfficacy is
panicularly important when a new IT application is being
adopted. However, in 8 situation when IT application usage has
become routine or habitual (e.g., checking one's e-mail) then
prior experience may become more important in usage behavior.
The study's other findings suggcst that organizational support
influences usage indirectly, through its relationship with self-
emcacy, but not directly. and not through any influence on
anxiety. Overall, the study's results suggest that organizations
and educators focus their efforts on building computer self
efficacy, and on modifying the detenninants of computer self-
efficacy in order to achieve higher levels of user acceptance of
computer technology.
CONCLUSIONS
Organizations are making signi ficant investments in IT.
However. if individuals do not use infonnation system
applications as anticipated, successful implementation can be
hard to achieve. Theoretjcal and empirical literature in the IS
field continues to investigate the factors that inHueoce the
adoption of IT. and this study contribulCS to that growing body
of literature. This study combines SCT, TIB and compuler
anxiety literature into one conceptual model and then tests the
related bypotheses using a field survey approach. The results
provide partial supp<>rt for hypotheses derived from CT. TIB
and the computer anxiety studies that served as the research
foundation for the study. The theoretically based research model
and the supp<>rt obtained for the majority of the hypotheses
represent a contribution to this area ofstudy.
The resulls lend supp<>rt to the research literalure that
suggests that computer self-efficacy plays a key role in user
acceptance of technolOgy. For example. in training
environments. one could potentially reduce computer anxiety
and increase computer usage by interventions designed to
improve computer self-efficac)'. In the future, longitudinal
research could be designed lO lest causal hypotheses regarding
computer sclf--cfficaey and the other key fBCtOrs involved in
computer usage. By providing insight into the important
relationships between computer sclf-efficacy. anxiety.
experience. supp<>rt and usage. this research can help further
work that explores their role in the acceptance and use of
infonnation technology.
REFERENCES
1. Agarwal. R., V. Sambamurthy, and R.M. Stair. "Research
Report: The Evolving Relationship Between GeneraJ and
Specific Compuler Self-Efficacy An Empirical
Winter 2003-2004 Journal of Computer InformatioD Systems 101
Assessmenl,.. Information Systems Research, II:4, 2000,
pp.418-430.
2. Ajzen, I. "Residual Effects of Past on Later Behavior:
Habituation and Reasoned Action Perspeclives...
PersonalityandSocial Psychology Review,6:2.2002, pp.
107-122_
3. Ajzen. I. "The Theory of Planned Behavior,"
Organizational Be.bavior and Human Decision
Proc.....,50, 1991,pp. 179-211.
4. AlKhaldi, M.A. and R.S. D. Wallace. "The InOuence of
Altitudes on Personal Computer Utilization among
Knowledge Wori<ers: The Case nf Saudi Anlbia.'
Information& Management,36, 1999,pp. 185-204.
5. Anderson, J.C. and D.W. Geroing. " tructural Equation
Modeling in Practice: A Review and Recommended Two-
Approach," Psycbologiul Bulktin, 103:3. 1988. pp.
411-423.
6. Author{s)(Date). Pleasenote: thedetailsofthispublication
have been withheld 10 ensure the aUlhors are anonymous
duringthereviewprocess.
7. Bailey, J.E. and S.W. Pearson. "DevelopmentofaTool for
Measuring and Analyzing Computer User Satisfaction,"
ManagementSeience,29:5. 1983, pp. 530-545.
8. Bandura. A_ Self-Efficacy: TbeElerciseofControL New
York: W.H. Freeman. 1997.
9. Bandura. A. " elf-Efficacy: Toward a Unifying Theory of
BehaYioral Change," Psycbological Review, 84:2, 1977,
pp.191-215.
10. Bandura. A Social Fouodation'ofThought and Action:
A Social Cognitive Theory. Englewood Cliffs, NJ:
PrenticeHal!, 1986.
11. Bandura. A. N.E. Adams. and 1. Beyer. "Cognitive
Processes Mediating Behavioral Change," Journal of
Personality and Social P.ychology, 35:3, 1977, pp. 125-
139.
12. Bandura. A. and D. DelVone. "Differential Engagement of
Self-Reactive lnfluences in Cognitive Motivation,"
Organizational Behavior and Human Decision
Proc......38:1,1986.pp. 92-113.
13. Bergeron, F.. S. Raymond. and M.F. Gara. "Determinants
of EIS Use: Testing a Bebavioral Mode!," Oecision
SupportSystems. 14. 1996.pp. 131-146.
14. Brosnan, MJ."IlleImpact ofComputerAnxiety and Self-
Efficacy Upon Performance," Journal of Computer
Assisted Learning, 14, 1998.pp. 223-234.
15. Burldtardt, M.E. and DJ. BI1lSS. "Changing Panems or
PanemsofChange:TheEffectsofaChange in Technology
on Social NetWork Structure and Power,"
SeienceQuarterly,35. 1990,pp. 104-127.
16. Chang, M.K. and W. Cheung. "Determinants of lhe
(ntention 10 Use InlemetIWWW at Work: AConfirmalory
ludy." I.form.tio.& M..agemeat,39.2001.pp. 1-14.
17. Chcung, W.. M.K. Chang, and V.S. Lai. "Prediction of
Internetand World Wide Web Usageal Work: ATestofan
ExtendedTriandisModel."DecisionSupportSy.tem>,30.
2000.pp. 83-100.
18. Coffin. RJ. and P.O. Macintyre. "Motivational InOuences
on Computerrelated Affective Slates,n Computers in
Human Behavior, 15. 1999,pp. 549-569.
19. Cohen, B.A. and G.W. Waugh. "A.se.sing Computer
Anxiety," P.ychological Reports,65. 1989.pp. 735-738.
20. Compeau. D.R. and C.A. Higgins. "Computer Self
Efficacy: DevelopmentoraMeasureand lnit.ial Test." MIS
Quarterly, 19, 1995,pp. 189-21I.
21. Davis, F. "Perceived Usefulness, Perceived Ease ofUse,
and User Acceptance of Information Technology," MIS
Quarterly, 13:3, 1989.pp.319340.
22. Dickson, G.W., J.A. Senn, and N.L. ChelVany. "Research
in Management Information Syslems: The MinnesoLa
Experiments,"Manageme.tSeience,23:9.pp. 913-923.
23. Durnell. A. and Z. Haag. "Computer Self..,fficacy.
Computer Anxiety, Aniludes Toward the Internet and
Reponed Experience with the Internet, by Gender, in an
East European Sample,"Computers in Human Behavior,
18,2002,pp. 521-535.
24. Eastin, M.S. and R. LaRose. "InternetSelf-Efficacyand the
Psychology ofthe Digilal Divide," Jour.alofComputer
Mediated Communications. 6:1. 2000. Online:
hnp:J/www.ascusc.orWcmdvol6lissuel/eastin.htmL
25. Farnell, C. and D.F. Latcker. "Evaluating Structural
Equation Models with Unobservable Variables and
Measurement Error." Journal of Marketing Ranrcb.
18:1,1981,pp. 39-50.
26. Gardner, W.L. and EJ. Rozell. "Computer Efficacy:
DclcnninanlS, Consequences. and Malleability," JournaJ
ofHigb Tecbnology Mageme.tResearcb. II: I, 2000.
pp. 109-136.
27. Gist. M.E. and R.T. Mitchell. "Self.Efficacy:ATheoretical
AnaJysis ofits and Malleability." Ac.ade.my
ofManagement Review, 17:2, 1992,pp. 183-21L
28. Gist. M.E., C.E. Schwoerer, and B. Rosen. "Effects of
Alternative Training Melbods on Self-efficacy and
Perfonnance in Computer Software Training.," Journal of
Applied P.ychology,74:6, 1989,pp.884-891.
29. Gordon. M.E., L.A. Slade. and N. Sehmitt, "'Ille 'Science
of the Sophomore' Revisited: From Conjecture to
Empiricisl1l.," Atldemy of Manage-ment Rniew, It: II.
1986,pp. 191-207.
30. Harrison. A.W. and R.K. Ranier. "An Examination ofthe
Factor Structures and Concurrent Validities for the
Computer Altilude Scale. the Computet Anxiety Rating
Scaleand theComputerSelf-Efficacy calc," Educational
and P'ychological Mea.urement, 52:3, 1992, pp. 735
746.
31. Hill. T. N.D. Smith. and M.F_ Mann. "Role Efficacy
Expecta1ions in PrediC1ing the Decision 10 Use Advanced
Technologies: The Case of Computers," Journal of
Applied Psycbology. 72:2. 1987.pp.307-313.
32. Hughes.C.T.and M.L. Gibson. "Studentsas urrogatesfor
Managers in a Decision-making Environment: An
Experimental Study," Journal of bnagement
InformatioaSy.tem" 8:2. 1991.pp. 153-166.
33. Igbaria, M. and J_ livara. "The Effects ofSelf-Efficacyon
ComputerUsage,"Omega.23:6.1995,pp.587-605.
34. Igbaria, M. and J. Parasuraman. "A Path Analytic Studyof
Individual Characteristics,ComputerAnxie!) and Anitudes
TowardComputers."Journ.1ofManagement, 15:3. 1989.
pp.373-388.
35. Igbaria. M., 1. Parasununan. and J. Baroudi. "A
Motivational Model ofMicrocomputer Usage,"Journalof
Manageme.t Information Sy,tem" 13:1, 1996, pp. 127-
143.
36. Igbaria, M., F.N. Payri. and S.L. Huff. "Microcomputer
Applications: An Empirical Look al Usage," Information
& Management, 16:4. 1989. pp. 187-196.
37. Johnson, R.D. and G.M. Marakas. "Research Report: The
Role of Behavioral Modeling in Computer Skills
Acquisition - Toward Refinement of the Model,"
InformationSystem.! Research. 11:4.2()(X). pp. 402-417.
38. Joreskog. K. and D. Sarhom. LlSRE!. 8: User',
Reference Guide. Chicago: Scientific Software
International. 1996.
Winter2003--2004 JournalofComputerInformationSystems 102
39. Koons. K.. '. and L.C. Liu."AFactorAnalysisofAttrihutes
Affecting Computing and Infonnation Technology Usage
for Decision Making in Small Business." Intunational
Journal of Computu Applie:ations in Technology,
12:213/4/5,1999.pp81-89.
40. Marak... G.M. M.V. Vi, and R.D. Johnson. "The
Multilcvel and Multifaceted Character ofComputer Self
Effieaey: Toward Clarification of rhe ConslIUct and an
Integrative Framc\'vork for Research.'" Information
ystemsResearch.9:2.1998.pp. 126-163.
41. Marcouldies, G.A. "Measuring Computer Anxiety: The
ComputerAnxietyScale,<>1 Educationaland Psychological
Measurement,49,1989, pp.733-739.
42. Murphy. CA.D. Coover. and S.V. Owen. "Development
and Validation of the Computer SelfEmcacy Scale."
Education and Psychological 49. 1989.
pp.893-899.
43. Nelemeyer. R.G. W.O. Bearden, and S. Sharma. Scaling
Procedure:s: Issues and Applications. Thousand oaks.
CA:Sage.2003.
44. Pare. G. and J. Elam. "Discretionary Use of Personal
Computers by KnowJedge Workers: Testing of 8 Social
Psychological Theoretical Model." Behavior and
I"formationTecbnology, 14:4, 1995,pp. 215-228.
45. Rosen, L.D. and P. Maguire. "Myths and Realities of
ComputerPhobia: A Anxiety R.esearch, 3,
1990.pp. 175-191.
46. Rosen. L.D.. D.C. Sears. and M.M. Wcil. "Treating
Technophobia: A LongilUdinal Evaluation of the
Computerphobia Reduction Progr'dm," Computers in
Human Bebavior,9,1993,pp. 27-50.
47. Sears. D.O. "College Sophomores in the Laboratory:
Influences ofaNarrow Data Base on Social Psychology's
View of Human Nature." Journal of Personality and
SocialP,yehology.51:3.1986,pp. 515-530.
48. Taylor. S. and PA Todd. "Unde=ding Information
Technology Usage: A Test of Competing Models."
InformationSystemsResearch.6:2.1995.pp. 144-176.
49. Taylor, . and PATodd. "Assessing IT Usage: The Role
ofPriorExperience; MISQuarterly, 19:4. 1995. pp. 561-
670.
50. Thatcher. J.B. and P.l. Perrewe. "An Empirical
Examination of Individual TrailS as Antecedents to
Computer Anxiety and Computer Self-Efficacy." MIS
Quarterly.26:4.2002.pp. 381-396.
51. Thompson,R.l.C.A. Higgins.andJ.M. Ho"ell."Influence
ofPersonal Experience on Personal Computer Utilization:
Tcsting a Conceptual Model" Journal of Management
Sy,tems, 11:1. 1994,pp. 167-187.
52. Thompson. R.L.. CAHiggins, andJ.M. Howell. "Personal
Computing: Toward a Conceptual Model of Utilization,"
MISQuarterly,15:1.199l.pp.I25-143.
53. Torkzadeh. G. and I.E. Angulo. "The Conccpl and
Correlates of Computer Anxiety," Behavior and
InformationTechnology, II:2, 1992. pp. 99-108.
54. Triandis. ItC. Inttrpeor5onal Behavior. Monterrey. CA:
Brooks/Cole. 1977.
55. Triandis. H.C. "Values. Altitudes and Interpersonal
Behavior'" Nebraska ymposium on Motivation. In Page.
M.M. (Ed.). !klier,Attitudes and Values. lincoln. NE:
UniversityofNebraska Press. 1980.pp. 195-259.
56. Venkatesh, V. "Determinants of Perceived Ease of Use:
IntegratingControl, Intrinsic Motivation and Emotion into
theTechnology Aceeplance Model." Information System,
Researcb, II:4,2000.pp. 342-365.
57. Vician. C. and L.R. Brown. ""Investigating Computer
Anxietyand Communication ApprehensionasPerfonnance
Antecedents in a Computinglnttnsivc Learning
Journal of Computer Information
Sy,tems,42:2.2002-2003.pp. 51-57.
58. Webster. J.. J.B. Heian. and J.E. Michelman. "Computer
Training'and ComputerAnxiety in the Educational Process:
An ExperimenlaJ Analysis.." Proceedings ofthe: Eleventh
International Conference: on Information ystems,
1990.pp. 171-182.
59. Webster, J. and J.1. Manocchio. "Microcomputer
Playfulness: Development of8 Measure with Workplace
Implications."MISQuarterly, 16:2. 1992,pp. 201-226.
60. Zmud. R.W. "Individual Differences and MIS Success. A
Review of the Empirical Literature." Management
Seienc..25:10.pp.966-979.
SCALEITEMS
Compute:rUsage(itemsstandardized beforerunningCFA)
usageI How ollendo you work in the EBA Micro-Lab?
usage2 Currentlyon averagehowmany hoursaweek doyou use the lab?
usage3 Ofyourcomputerwork for classassignments" whatpercentage isdone in the Micro-Lab?
ComputerSelf-Efliacy
cflieaI Ifeel confident callingupadalafile to viewon themonitorscreen
effica2 I feel confidentusingthecomputerto write:an essayoralener
eflieal I feel confidententeringand savingdarn(numbersorwords) inlO a file
effica4 I feel confidentmovingthecursoraround me monitorscreen
effienS Ifeel confidentmakingselectionsfrom an onscrecn menu
effica6 I fccl confidentescaping/exitingfrom aprogramorsoftware
effica7 Ifeci confidentworkingon apersonaJ computer(microcompul'cr)
effienS Ifeel confidentusingaprintertomakea"hardcopy- ofmy work
Removed
Rt.moved
Remove:d
Winter2003-2004 JournalofComputerInformationSystems 103
ComputtrADJitt}
anxiety I I am confiden, in my ability to use computers IREV)
anxiety2 I tr} to avoid using a compmer whenever possible
anxiety3 Iworry about maJcing mistakeson a computCf
anxiety4 I enjoy working with computers [REV]
anxietyS I feel o\'erwhelmed whenever I am working on a computer
anxiety6 Jfeel anxious whenever Iamusing acomputer
anxiety7 I feci tense whenever working on a computer
anx.iery8 I feel comfortable with computers [REV]
Computer Experieneo (items standardized before running CFA)
"".perill
Indicate your overall computer literacy
z x ~ r t l How many years ago did you first begin using computers?
zexperil3 How knowledgeable are you about computer and software?
Orpniutionll Support
Second-order factor comprised of five dimensions: assisUlncc, access to equipment.
atmosphere. and reliability(print dimension removed)
access I Availability of equipment when needed
access2 Waiting time for 8 computer
access) Townumberof computers
access4 Balance of Pentiums to 386/486 computers
assist I AssiSl8llcc in equipment usc
assist2 Attitudetoward students
assisL3 Suppon in assisting wiLh problems
assist4 Knowledge abou' software
assistS Knowledge about lab opera,ion
assist6 Response time to problems
assist7 Student orientation
alma! Ability laconcentrate on work
atm02 Quiet workingenvironment
atmo3 onfidemialityofwark
hours I Number afhours open
hours2 Evening hours/closing time
hoursJ Open when needed
hours4 Hours of opel'81ion on week-ends
print! Printer quality
print2 3lisfy your printing needs
printJ Printers' output quality
print4 Printer speed
relial FulVcomplele software functionality
relia2 Working order of computers
relial Maintenanceof equipment
rclia4 Reliability ofequipment
reliaS Rei iability ofsoftware
relia6 Dependability ofhardware
Removed
Removed
Removed
boutS of operation, quality of printed output,
Removed
"
Winter 2003-2004 Journal of Computer Information Systems 104
Copyright of Journal of Computer Information Systems is the property of International
Association for Computer Information Systems and its content may not be copied or
emailed to multiple sites or posted to a listserv without the copyright holder's express
written permission. However, users may print, download, or email articles for individual
use.