Attribution Theory: A Theoretical Framework for Understanding Information Systems Success

This paper discusses Attribution Theory and its relevance in understanding factors determining users’ attributions for information system-related outcomes, as well as the influence of these attributions and the nature of the system outcome on the level of users’ satisfaction with the system.

ORI GIN AL PA PER
Attribution Theory: A Theoretical Framework
for Understanding Information Systems Success
Ken C. Snead Jr.• Simha R.Magal• Linda F. Christensen•
Atieno A.Ndede-Amadi
Ó Springer Science+Business Media New York 2014
Abstract Information systems research often employs user satisfaction with, use of
perceived organizational benefits of, newly-developed systems as measures of info
system success.Further,this stream of research attempts to associate these measures
success with a myriad of hypothesized determinants involving organizational,personal,
task,and system characteristics,as well as characteristics of the implementation process
Initial research in thisarea wascriticized forthe dearth oftheoreticalunderpinning
employed. Subsequent to these criticisms, underlying theory from a variety of disc
now guides much of this research. Of particular interest to this research effort are t
of a well-established theory in the area ofsocialpsychology-attribution theory.While
attribution theory has been employed in some ofthe more recentworks investigating
factors related to information system success, none of these works simultaneously
the theory’s information and motivational antecedents along with the success/failu
of the system’s outcomes, users’ perceptions of the causes of the outcomes (attrib
and the reported level of user satisfaction with the system.In response,the current study
develops a modelfor the simultaneous empiricalexamination of these issues by incor-
porating them into a behavioral decision making methodology administered to Prof
MBA students.The study’s results support the relevance of attribution theory as a th
reticalframework forunderstanding those factorsdetermining users’attributionsfor
information system-related outcomes, as well as the influence of these attributions
nature of the system outcome on the level of users’ satisfaction with the system.
K. C. Snead Jr.(&)
College of Business,Bowling Green State University,Bowling Green,OH 43403-0262,USA
e-mail: [e-mail]
S. R. Magal
Seidman College of Business,Grand Valley State University,Grand Rapids,MI, USA
L. F. Christensen
School of Business,Christian Brothers University,Memphis, TN,USA
A. A. Ndede-Amadi
School of Business and Management Studies, Kenya Polytechnic University College, Nairobi, Kenya
123
Syst Pract Action Res
DOI 10.1007/s11213-014-9328-x

Keywords Attribution theory Causal attributions Information system success
Introduction
The incomparability/inconsistency of early information system success (ISS) studie
due theirlack of incorporating underlying psychological,motivational,and cognitive
factors that potentially mediate or moderate the relationship between measures re
and determinants of,system success.Further,they did notconsider/controleitherthe
nature of the outcomes resulting from using the system (such as the quality of the
decision or task completion) or the user perceptions of the causes of these outcom
an approach did not consider that users’/subjects’ responses may have been influe
the ultimate success or failure of the system. It seems logical that both the success
nature of the outcome of system use, and the user perceptions of factors causing t
of the outcome, would influence user evaluation of the system. For example, great
satisfaction islikely to be reported by a userwho experiencessuccess(successfully
completes a task or makes a good decision) upon using a supporting system.Further,the
extentto which the user feels personally responsible for bringing aboutthe outcome,as
opposed to the feeling that something external was primarily responsible for that o
will likely also impact his/her reported satisfaction with the system.
Attribution theory addresses these issues by modeling psychological and motiva
factors presumed to influence ISS,doing so by taking into accountthe impactof the
success/failure nature of system-related outcomes and user perceptions of the cau
these outcomes.This theory is employed here to provide a conceptualframework to
examine the impact of both the success/failure nature of system-related outcomes
information related to other system-related experiences, on user perceptions of the
of these outcomes.Further,both the causalperception and other system-related experi-
ences are examined to determine if either impact user system satisfaction.
The next section of this paper provides a discussion of the ISS literature pertinen
study.This is followed by a presentation ofattribution theory and accompanying
hypotheses, followed by a more detailed discussion of the theory’s elements incorp
into this study (antecedents and consequences).In addition,discussion of the literature
employing attribution theory in the ISS context and this study’s incremental contrib
will be presented.Sections presenting methodology,results,implications,and limitations
will then follow.
Information System Success
Early research in the ISS arena identified various organizational, personal, task, and
characteristics,as wellas characteristics of the implementation process,associated with
ISS (e.g., Swanson 1974; DeBrabander and Edstrom 1977; Ginzberg 1979; Swanson
Ives and Olson 1984; Franz and Robey 1986; Tait and Vessey 1988; Barki and Hart
1989; Wastell 1999). Organizational characteristics involve structure and top mana
supportissues,while personalcharacteristics relate to beliefs,attitudes,and experience
issues.Task characteristics refer to degree of structure,system characteristics dealwith
accuracy and reliability issues, while characteristics of the implementation process
userparticipation,training methods,etc.Of concern,was the observation thatresults
Syst Pract Action Res
123

across these ISS studies were often inconsistentand incomparable due to a myriad of
measuresbeing employed to measure‘‘success’’(e.g.,usersatisfaction,actualuse,
intentionsto use, information quality,individual/organizationalimpacts),and to the
atheoreticalresearch designs used.This concern motivated ISS researchers to examine
more reliable and valid ways to operationalize the notion ofsystem ‘‘success’’(e.g.,
DeLone and McClean 1992; Goodhue 1995; Goodhue et al. 2000; Abdinnour-Helm e
2005),as well as employ theoretical models to underpin the research designs used (
Davis et al. 1989—‘‘TechnologyAcceptanceModel’’; Sneadand Harrell 1994—
‘‘Expectancy Theory’’; Eccles and Wigfield 2002—‘‘Motivational Beliefs’’;DeLone and
McClean 2003—‘‘DeLone and McClean IS Success Model’’;Venkatesh etal. 2003—
‘‘Unified Theory of Acceptance and Use of Technology’’;Yi etal. 2006—‘‘Technology
Acceptance Model/Theory of Planned Behavior/Innovation Diffusion Theory’’;Khayun
et al.2012—‘‘The Delphi Technique’’).
Noteworthy to this study are the ongoing acknowledgments in the ISS literature
phenomena related to the human/computer interactions in the information technol
arena involves organizationalissues and complex roles that‘‘socialactors’’engage in
while adopting, adapting, and using information systems (Reeves and Nass 1996; M
et al. 2000; Bebebasat and Zmud 2003; Lamb and Kling 2003; Standing et al. 2006
and Benasat 2008; Sykes et al. 2009). Those acknowledging this social aspect of IT
that the tenets of attribution theory,a prominent theory within the domain of social psy-
chology,are relevant for the study of ISS.In fact,the social influence component of the
unified theory ofacceptance and use oftechnology (Venkahesh etal. 2003)considers
concerns of system users abouthow they willbe viewed by others as a resultof their
system use,particularly in the early stages when system use is mandatory.This notion is
quite similar to the ‘‘social actor’’ dynamics subsumed by the motivations anteced
attribution theory.Accordingly,a description ofthose tenets ofattribution theory per-
taining to this study (as developed by the seminal works formulating it) and corres
hypotheses are presented.
Attribution Theory & Corresponding Hypotheses
Attribution theory is the study of the process by which people associate causes to e
and outcomes thatthey experience (Jones and Davis 1965;Kelley and Michela 1980;
Swanson and Kelley 2001).A major goalof the attributionalprocess is to understand,
organize,and form meaningfulperspectives aboutoutcomes and to predictand control
them.This propensity to understand and controlevents and outcomes is more evident
in situations with unexpected or negative outcomes,when outcome dependency is high,
when involvement in the outcome is high,and when faced with an experience of lack of
control(Kelley 1967;Weary etal. 1989;Bogumil2001).It is noteworthy thatthese
situationsdescribe the circumstances/contextsurrounding system use (Standing etal.
2006).For purposes of this study,the ‘‘outcome’’ is the successful or unsuccessful com-
pletion of a task that was aided by a newly-developed information system.
Attribution theory suggests that certain factors, or ‘‘antecedents,’’ will influence
a person to infer the cause of an outcome in a particular way; these causal inferenc
referred to as‘‘attributions.’’Two of these antecedentspurported to influence one’s
attributions are ‘‘information’’ and ‘‘motivation.’’ The theory further suggests that
‘‘consequences’’ of causal attributions on an individual’s affective, or emotional rea
to the outcome.These elements of attribution theory are discussed in turn.
Syst Pract Action Res
123

Attributions
Attributions are the inferences of causation of a particular event or outcome.A common
dimension ofattributions foroutcomes involving success or failure is locus of control
(Jones and Davis 1965; Kelley 1967; Stajkovic and Sommer 2000), which is the deg
which a cause is thoughtto be related to factors within the person (internal) or to the
environment (external). Other dimensions of attributions, not germane to this stud
to the stability and controllability of the cause. Based upon the work of Heider (195
causal factors have been identified and determined to be relevant to the interpreta
achievement related outcomes. These are ability, effort, task difficulty, and luck. A
the characteristics of a person that describe his/her task related capabilities.Effort is the
personal characteristics related to the degree of persistence a person brings to bea
specific task.Task difficulty refers to the environmentalcharacteristics related to the
degree of challenge associated with task accomplishment,while luck has to do with the
influence of random (chance) environmental conditions. Relating these causal facto
locus of control dimension of attributions results in cataloging ability and effort as i
causes,and luck and task difficulty as environmental or external causes.
Antecedents
Antecedents are those factors impacting an individual’s inference ofthe cause ofan
outcome.The classes of antecedents relevant to this study are ‘‘information’’ and ‘‘m
vations.’’ These antecedents highlight important ways in which the nature of an ou
impacts an individual’s perception of the cause of that outcome.
Information
Kelley’s(1967)information ANOVA modelsuggeststhatcausalattributionswill be
associated with those factors perceived to vary systematically with outcomes. Furt
modelidentifies three characteristics of information thatare thoughtto influence causal
attributions-consistency,distinctiveness,and consensus.Each is illustrated using the fol-
lowing supposition:
A person Pð Þresponds to a stimulusSð Þat a point in time Tð Þ
A causal attribution or explanation of P’s (user’s) reaction to S (system-related outc
dependson the three information characteristics:consistency hasto do with this P’s
response to S at other T’s; distinctiveness refers to P’s response to other S’s; and c
is concerned with other P’s responses to S.In an ISS context,these information charac-
teristics represent specific aspects of user perceptions of past system experiences
dimensions: a user’s own experiences with similar systems at different times (cons
a user’s own experiences with differentsystems (distinctiveness),and other users’ past
experiences with similar systems (consensus). These aspects of information are pa
relevant for understanding a user’s causal attributions for system related outcomesince
past system experiences are commonly believed to influence user expectations rel
continued system use (e.g.,Kim and Malhotra 2005).The theory purports the nature of
these past experiences will influence a user’s attributions for a system-related succ
failure. For example, if a user’s past experiences with other systems have been suc
Syst Pract Action Res
123

whereas current efforts resultin failure,the user is likely to blame the system,and vice
versa.On the other hand,if a user knows thatothers have notbeen successfulin using
similar systems,he/she will likely take credit for any successes with the current system
The literature suggests that particular combinations of these information charact
systematically influence the internal/external orientation of the causal attribution (
and Michela 1980; Dixon 2001). Specifically, information revealing high consistenc
distinctiveness,and high consensus willevoke externalattributions.So, a userexperi-
encing a system-related outcome thatis similar to his/her own experiences with similar
systems (high consistency), different from his/her own experiences with dissimilar
(high distinctiveness),and similar to other users’ experiences with similar systems (hig
consensus) will likely attribute the cause of the outcome to the externally oriented
of task difficulty and luck. On the other hand, internal attributions are typically link
information profile consisting of high consistency, low distinctiveness, and low cons
(Kelley and Michela 1980).So, a userexperiencing a system-related outcome thatis
similar to his/her own experiences with similar systems (high consistency), similar
her own experiences with dissimilar systems (low distinctiveness), and dissimilar to
users’ experiences with similar systems (low consensus) will likely attribute causat
the internally oriented factors of ability and effort.
Based upon these theoretical tenets,we advance the following hypotheses:
H1a: The causal attributions of users for system-related outcomes will be external
the system-related outcome is:
similar to their experiences with similar systems—consistent;
different from their experiences with dissimilar systems—distinctive;
similar to other users’ experiences with similar systems—in consensus
H1b: The causal attributions of users for system-related outcomes will be internal
the system-related outcome is:
similar to their experiences with similar systems—consistent;
similar to their experiences with dissimilar systems—not distinctive;
different from other users’ experiences with similar systems—not in consensus
Motivation
Motivations associated with concerns for protection of self image will affect the pro
assigning causality. In a social context, attributions are said to be motivationally pr
since society requires an accountfor behavior thatis deviantfrom socialnorms (Zuck-
erman 1979).A classic,motivationally-driven pattern of attributions is the self-serving
bias, which predicts the predisposition of individuals to attribute failures to externa
and successes to internal causes (Miller 1976; Miller and Ross 1975; Zuckerman 19
This bias predicts thatusers willattribute system-related failure outcomes to external
causes (task difficulty and luck) and system-related success outcomes to internalcauses
(ability and effort).Based on these observations,we advance the following hypotheses:
H2a: The causalattributionsof usersfor system-related failureoutcomeswill be
external.
H2b: The causalattributionsof usersfor system-related successoutcomeswill be
internal.
Syst Pract Action Res
123

Consequences
Consequences of attributions have been observed in the areas of affective or emot
reactions and the expectancy ofexperiencing similaroutcomes.The locus ofcontrol
dimension of attribution is linked to the formation of affective or emotional reaction
the individualtoward the situation (Weineret al. 1972).Since locus ofcontrolis the
attribution dimension of interest in this study,only the affective reaction consequence is
considered.Although somestudieshaveconsidered job satisfaction asan affective
responsein attribution studies(Adler 1980),usersatisfaction with asystem isthe
affective response ofinterestin the currentstudy,given its prominentrole in the ISS
literature.
Weiner (1974) argues that attributions of success and failure to internal factors t
heighten ‘‘emotional reactivity’’ as compared to attributions to external factors.So,attri-
butions of success to the internalcauses of effortand ability are believed to resultin a
greater positive impact on the individual’s affective response than external attribu
the other hand,attributions of failure to internalcauses are expected to generate greater
negative affect than attributions of failure to external causes.This pattern has been con-
sistently observed in the attribution literature (Weiner 1974), and leads us to advan
following hypothesis:
H3: The causal attributions of users for system-related outcomes will be systemat
related to their satisfaction with the system.
A fourth hypothesis is advanced which addresses the intuitive impactof the success/
failure nature ofsystem outcomes experienced by a useron his/herevaluation ofthe
system:
H4: Successful system-related outcomes will be associated with greater user satis
than will failures.
While hypothesis H4 is not within the domain of attribution theory as such, it add
the issues related to the potential impact of system outcomes on system success m
initially expressed in the ISS literature by Tait and Vessey (1988).
Figure 1 summarizes the relationships suggested by hypotheses H1–H4 and high
the purported influence of both the success/failure nature of the system outcome,and the
information antecedent,on the attributions ofusers forthese system-related outcomes.
Furtherdepicted is the suggested influence ofboth the internal/externalnature ofthe
attribution, and the success/failure nature of the system outcome on user satisfact
information system.
Prior to discussing the study’s method and results,it is appropriate to presentthe
evidence for the direct relevance of attribution theory for ISS research, as demonst
the literature.
Attribution Theory and ISS Research
The recognition of the utility of employing the tenets of attribution theory in ISS res
is significant,as the theory has been found to have relevance in several facets of the
area.Specifically,studies use patterns of causal attributions to better understand the
between end user involvement and ISS (Magal and Snead 1993), to examine the in
of attributionalbiases forIS professional,end user,and managementattributions for
Syst Pract Action Res
123

system outcomes (Karsten 2002;Peterson etal. 2002;Snead and Ndede-Amadi2002;
Standing et al. 2006), and to identify antecedents contributing toward attribution e
et al. 2007).Further,Thatcheret al. (2008)employ tenetsof attribution theory to
decomposethe computerself-efficacy constructinto both an internaland external
dimension; they then empirically determine each has a separate influence on user
anxiety and perceived ease of use of IT.
Attribution theory has also been used to model the satisfaction of users/custome
IT-based system experiences.Fang etal. (2005) examine the impactof employee attri-
butions forcontrolsystem feedback on performance expectations and job satisfaction
while Anderson et al.(2009) find the nature of customer attributions for service failure
moderate the composition of overall customer satisfaction with the service provide
Hsieh (2012) incorporates attribution theory’s stability, locus, controllability dimens
causal attributions to help model the extent to which customers perceive they hav
treated unfairly (‘‘psychological contract violation’’) in an E-commerce electronic re
(e-Return) experience.
Further,the theory has been incorporated in understanding a more recentconstruct
related to ISS—user trust with IT related outcomes. Jarvenpaa et al. (2004) employ
tenets ofattribution theory which dealwith situationalfactors and theirinfluence on
people’ssocialperceptionsof othersand themselvesto betterunderstand how trust
influences user sentiments task performance in IT-enabled contexts.Wang and Benbasat
(2008) use the theory of social responses to computers (Reeves and Nass 1996) to
the nature of users’ attributions for their extent of trust with Recommendation Age
E-commerce setting,while Porter etal. (2013) use the consensus,consistency,distinc-
tiveness information dimensions to assess their influence on the level of trust cons
associate with firm-sponsored versus member-generated virtual communities.
Standing etal. (2006) offer an explanation for the emerging popularity of the use o
attribution theory in ISS research. They state—‘‘The IT context is a relevant one in
to study explanations for success and failure because: this area involves a continuo
of projects being undertaken; there are many contingencies in developing good IT
there are complex determinants for defining the success and failure of projects;the IT
environment(e.g.,funding of projects)is often unstable;many projectsdo ‘fail’;
responsibility is high as projects are often substantial and their success or failure h
impact on many users; there is a hierarch of responsibility for IT project failures; an
workers are, therefore, subject to continuous and often large motivational issues in
with the complexities of project failures.’’ (pp.1149–1150)
Information
(HHH or HLL)
H1 Attribution
(Internal or External)
H3 Affect
(Degree of User Satisfaction)
Outcome
(Success or Failure)
H4
H2
Fig. 1 Hypothesized relationships
Syst Pract Action Res
123

Noteworthy is thatonly one of these works consider the theory’s ‘‘information ante
cedent,’’as originally formulated by Kelley (1967),on userattributions ofIT related
outcomes.This is surprising given this antecedentrequires the sequentialupdating and
feedback mechanisms discussed by Bhattacherjee (2001) and Kim and Malhotra (2
in place in order for a user to assess the extentto which their currentIT experience is
consistentand/or distinctive with their other IT experiences,and whether or notit is in
consensus with the IT experiences of others. Further, none of these works simultan
consider the theory’s information and motivationalantecedents,consequences elements,
the success/failure nature of the system outcome,user perceptions of the causes of the
outcomes (attributions),and the reported levelof user satisfaction with the system.Par-
ticularly importantis examining the influence ofattributionalphenomena on the user
satisfaction construct,as this constructcontinues to occupy a key role in defining ISS;
further, it has been shown to be linked to system use (DeLone and McClean 1992,
Goodhue et al.2000; Kim and Malhotra 2005).
In response,the currentstudy develops a modelfor the simultaneous examination of
theseissuesby incorporating them into a behavioraldecision making methodology
administered to ProfessionalMBA students.This methodologicalapproach provides for
stronger levels of internal validity as compared to the methodologies employed in
works,and is discussed below.
Method
Data Collection
Data for this study was obtained using a behavioral decision making methodology,
typically presents subjects with various scenarios which require decisions to be ma
reported in an effort to evaluate various cognitive processes. This method is widely
a variety of contexts (Slovic et al. 1977) and has been specifically employed in attr
theory studies (Arrington etal. 1985;Harrison etal. 1988;Kelley and Michela 1980;
Weiner1974).A questionnaire (see Table 1)asked participants to assume they were
departmentalmanagers,in a major corporation,charged with the responsibility of pre-
paring the departmentalbudget.The questionnaire indicated thata newly-developed
computer-based information system, intended to support the department manager
budget preparation effort,was made available for his/her use.Participants were provided
with information about two experimentally controlled factors, each having two leve
type of system-related outcome: success or failure; and (2) type of information pro
consistency,high distinctiveness,high consistency (HHH) or high consistency,low dis-
tinctiveness,low consensus (HLL).
The success level for the outcome factor was operationalized by indicating to pa
pants that their initial use of the system was successful in that their budget was bo
and accurate. The failure level was operationalized by indicating to participants tha
initial use of the system was a failure in that their budget was neither timely nor ac
The information profile factor was operationalized by providing statements related
consistency,distinctiveness,and consensus dimensions ofKelley’s (1967)information
model.
The outcome and information factors were employed as between-subjects factor
four outcome/information profile experimental cells were formed by the study: (1)
HLL; (2) success/HHH; (3) failure/HLL; and (4) failure/HHH. Participants were random
Syst Pract Action Res
123

assigned to one ofthe fourcells.The top portion ofTable 1 presents the information
presented to participants for the success/HLL profile.
Participants provided causalattribution and user satisfaction measures related to the
hypothetical IT experience presented; they also provided a variety of demographic
sures.Attribution responses were obtained by having participants allocate a totalof 100
points among the four potential causes of system-related outcomes: (1) ability; (2)
(3) task difficulty; (4) luck (see the bottom portion of Table 1). This method of mea
subjectattributions is widely accepted (Adler 1980),(Harrison etal. 1988;Kaplan and
Reckers 1985).From this pointallocation,an attribution score variable (ATTRIB) was
formed by mathematically combining participants’ ability, effort, task difficulty, and
responses as follows:
Table 1 Instrument presenting the ‘‘Success’’ outcome with ‘‘High Consistency, Low Distinctiveness
Consensus’’ information (HLL)
PART 1: DESCRIPTION OF SETTING
Assume you are a departmental manager in a major corporation. One of your responsibilities is the
preparation of a budget for your department. A newly developed computer -based information system,
intended to support you in your budget preparation, was recently made available for you to use.
Your initial use of this system was VERY SUCCESSFUL. Your budget was much more timely
and it more accurately reflected your resource requirements.
You will be asked to make judgements about the extent to which certain factors contributed to your
successful experience using the system. The following three statements are designed to help with your
judgements:
YOUR PREVIOUS EXPERIENCES WITH SIMILAR
COMPUTER-BASED SYSTEMS HAVE BEEN................................................... VERY SUCCESSFUL
YOUR PREVIOUS EXPERIENCES WITH DIFFERENT
COMPUTER-BASED SYSTEMS HAVE BEEN................................................... VERY SUCCESSFUL
OTHER USERS' EXPERIENCES WITH SIMILAR
COMPUTER-BASED SYSTEMS HAVE BEEN....................................................................... FAILURES
PART 2: YOUR JUDGMENTS REGARDING THE SETTING
The following four general factors have been identified as possible causes for your successful experience
with the system. Please allocate a total of 100 points among these possible causes with the understanding
that the more points an item is assigned, the more important you perceive it to be the cause of your
successful experience with the system. Any one item can receive as many as 100 points, or as few as 0
points. Just be sure to allocate a totalof exactly 100 points among the possible causes.
POSSIBLE CAUSES POINTS
(1) You have above average ability. ______
(2) You exerted a great deal of effort. ______
(3) This task was very easy. ______
(4) Good luck, chance, or some other
irrational factor is important in
this case. ______
TOTAL POINTS 100
Syst Pract Action Res
123

ATTRIB ¼ Ability þ Effortð Þ Task Difficulty þ Luckð Þ:
Ability and effortpoints are combined since each represents an internalcause,while
task difficulty and luck points are also combined since each represents an external
Thus, the ATTRIB variable represents a net of internal and external causes. Becaus
of 100 points was allocated among the four causes, a perfect internal score would
while a perfect external score would be -100. This formulation is well established in
attribution research literature (Weiner 1974).
User satisfaction measures were obtained by having participants indicate their le
satisfaction with the computer-based system on a 11-point Likert scale (0–10).
Subjects
Subjects used in this study were students in the Professional MBA program of a Mid
US university, who are primarily employed full time in a management capacity. A t
usable responses were obtained during a regular class session. Demographic inform
obtained which included age, gender, employment information, and the level of ex
with computer-based information systems.There were 18 female and 63 male participants
and the average reported age was 31 years.Important to potential concerns related to this
study’s generalizability, it is noted that approximately 70 % of participants indicate
employed full-time in managementpositions,and 60 % indicated moderate to extensive
experience with computer-based information systems when performing their job.
Results
Hypotheses H1 and H2 were tested by subjecting the data collected to a generallinear
model procedure. The dependent variable of the model was the attribution score (A
for the participants.The modelcontained two explanatory factors,each attwo levels:
OUTCOME (success orfailure)and INFORMATION profile (HHH orHLL). Table 2
presents the results of the modelestimation.The overallmodelachieved a levelof sig-
nificance of0.0001 with an associated F value of10.81.Both the OUTCOME and
INFORMATION factors are related to the attribution variable in a significant way (Ty
III SS p values of 0.0002 and 0.0276,respectively).
Testing hypotheses H1a and H1b required focusing on the results of the INFORM
factor.ATTRIB cell means for each of its two levels are reported atthe top portion of
Table 3. The mean attribution for the HHH and HLL levels of INFORMATION are 2.59
and 34.14,respectively.Because the higherthe ATTRIB value the more internalthe
attribution, the pattern of cell means supports hypotheses H1a and H1b. Attribution
system-related outcome are more internal for users provided the HLL information p
and more external for users provided the HHH information profile. These research r
supportH1a and H1b suggestthatKelley’s(1967)information ANOVA modelmay
generalize to the ISS area of research.
Testing hypothesesH2a and H2b required focusing on the modelresultsfor the
OUTCOME factor.ATTRIB cell means for each of the two levels are reported atthe
bottom portion ofTable 3.The mean attribution forthe success and failure levels of
OUTCOME are positive 44.57 and negative 7.03, respectively. This pattern of cell m
offers supportfor hypotheses H2a and H2b.The self-serving bias of attribution theory
appears to generalize to the ISS area of research, suggesting that users experienci
Syst Pract Action Res
123

related failure outcomes provide more externalattributions than do users experiencing
successful outcomes.
Hypotheses H3 and H4 predict a systematic relationship between user satisfactio
attribution and system outcome,respectively.These hypotheses were examined by esti-
mating a generallinear modelwhere user satisfaction (SATISFACTION) serves as the
dependent variable and user attributions (ATTRIBUT) and system outcome (OUTCO
serve as factors. The OUTCOME factor was tested at two levels: success and failure
ATTRIBUT factor was also tested at two levels: internal and external. This dichotom
accomplished by assigning ‘‘internal’’to positive ATTRIB scoresand ‘‘external’’to
negative scores.The results are presented in Table 4.
As Table 4 shows, the overall model achieved a level of significance of 0.0001 w
associated F value of 46.20.Further,both the OUTCOME and ATTRIBUT factors are
related to the satisfaction variable in a significant way (Type III SS p values of 0.00
0.0220,respectively),supporting hypotheses H3 and H4.Thus,the hypothesized link
between the locus of controlattribution and user satisfaction (H3)is supported in this
study. To test the nature of this link, the mean user satisfaction is reported for each
Table 2 General linear model results (H1 and H2)
Dependent variable: ATTRIB
Source DF Sum of squares F value Pr [ F
Model 2 70217.6713629 10.81 0.0001
Error 78 253436.3533285
Corrected total 80 323654.0246914
Source DF Type I SS F value Pr [ F
Outcome 1 53836.7646181 16.57 0.0001
Information 1 16380.9067448 5.04 0.0276
Source DF Type III SS F value Pr [ F
Outcome 1 50221.7474086 15.46 0.0002
Information 1 16380.9067448 5.04 0.0276
Table 3 Mean attribution score by factor
Level of information N ATTRIB
Mean SD
HHH 37 2.5945946 68.2248975
HLL 44 34.1363636 56.2575435
Level of outcome N ATTRIB
Mean SD
Failure 39 -7.0256410 67.3152392
Success 42 44.5714286 48.7968126
Syst Pract Action Res
123

two ATTRIBUT levels atthe top portion ofTable 5.The mean usersatisfaction for
internal attributions is 5.27, compared to a mean level of 2.36 for external attributi
satisfaction ranged from a low of 0 to a high of 10).According to these results,the more
internal the attribution of users, the higher the satisfaction with the system, even a
trolling for outcome. With respect to outcome, the means reported in the bottom p
Table 5 confirm the intuitive expectations that successful system outcomes will be
with higher levels of user satisfaction (mean of 6.43) than with failures (mean of 2.
The theoretical framework of attribution theory developed and employed in this
considers the impact of both system related outcomes and their perceived causes
satisfaction.The model(Fig.1) predicts thatthe success/failure nature of the outcome
affects user satisfaction directly (H4), and indirectly through its motivational influen
userattributions through the self-serving bias (H2).Specifically,it was predicted that
attributions of system outcomes are directly influenced by a user’s past experience
system use comprising of consistency,distinctiveness,and consensus information (H1).
Attributions were also predicted to directly influence user satisfaction (H3). Accordi
the reported results, all hypothesized relationships (H1–H4) in this study were supp
Discussion and Implications
These findingshave importantimplicationsfor ISS research.First, the demonstrated
importance of the success/failure nature of outcome (in H2 and H4) supports the co
first posited by Tait and Vessey (1988) that the success/failure nature of system ou
needs to be monitored and controlled. Second, the evidence supporting the applica
Kelley’s (1967) information model(H1) implies thatusers willconsider both their past
system experiences and their perception of others’ past experiences when forming
attributions for system outcomes. Thus, the attribution theory model presents a co
framework that specifies the important attributes of a user’s past experiences likel
considered when evaluating system outcomes (consistency, distinctiveness, and co
and the likely impactof particularcombinations ofthese dimensions on theircausal
Table 4 General linear model results (H3 and H4)
Dependent variable: SATISFACTION
Source DF Sum of squares F value Pr [ F
Model 2 351.50015288 46.20 0.0001
Error 78 296.72206935
Corrected total 80 648.22222222
Source DF Type I SS F value Pr [ F
Outcome 1 330.70573871 86.93 0.0001
Attribut 1 20.79441417 5.47 0.0220
Source DF Type III SS F value Pr [ F
Outcome 1 216.02985669 56.79 0.0001
Attribut 1 20.79441417 5.47 0.0220
Syst Pract Action Res
123

attributions for current system outcomes experienced. What is interesting is that b
consistency constantin both information profiles (high in both—HHH and HLL),the
influence on attributions of the ‘‘lesser mentioned’’ dimensions of a user’s pastsystem
experiences are revealed.Specifically,information related to the user’s pastexperiences
with othersystems(distinctiveness),and information related to otherusers’system
experiences (consensus), were found to be important in the formation of attribution
implication is that many ISS research efforts may be focusing on an incomplete set
experiences which may further contribute to inconsistent results. As previously not
consistency, distinctiveness, and consensus information constructs subsume the se
updating and feedback mechanismsdiscussed in Bhattacherjee (2001)and Kim and
Malhotra (2005) and provide a more comprehensive explanation of elements of int
these updating and feedback mechanisms.Third,the support for hypothesis H3 indicates
that the nature of the causes a user perceives to be responsible for system outcom
influencehis/herreportedlevel of satisfaction.Therefore,thesecausalattributions
potentially act as an intervening mechanism between the host of factors thought to
system successidentified by previousISS research and measuresof system success.
Ignoring these causalattributions in ISS research willlikely introduce a confounding
source potentially contributing to inconsistent results.
The practical implications of this study suggest that managers need to be aware
attributional processes users evoke when reacting to the system outcomes they ex
The demonstrated applicability of the information antecedent in H1 suggests that u
rely on their perceptions of the quality of both their own previous system-related e
as well as these same perceptions for the quality experienced by other users, when
importantattributions for system-related outcomes.This would suggestthe importance of
holding ongoing IT outcome ‘‘debriefings’’ which involve all users of similar system
to facilitate the accuracy of user perceptions of the success/failure nature of the ou
user perceptions of the levels of consistency,distinctiveness,and consensus.Further,the
demonstrated applicability of the self-serving bias (H2) implies thatuser attributions and
corresponding impact on affect are potentially more a response to protect the user
than a logical response to the facts surrounding the situation. Consequently, mana
be aware of these tendencies to more effectively understand user reactions in orde
more accurately the merits of a system. For instance, system training sessions cou
techniques for influencing the attributional process thereby increasing the likelihoo
making attributions that are in the best interest of the organization. The results of
Table 5 Mean user satisfaction by factor
Level of Attribut N SATISFACTION
Mean SD
External 22 2.36363636 2.46007074
Internal 59 5.27118644 2.57862907
Level of outcome N SATISFACTION
Mean SD
Failure 39 2.38461538 1.78613645
Success 42 6.42857143 2.18802570
Syst Pract Action Res
123

suggestthatencouraging internalattributions may lead to more favorable user reactions
toward a system for either success or failure outcomes. This attribution ‘‘retraining
particularly desirable for newly implemented systems (Weary et al. 1989).
The strength of this study lies in its simultaneous consideration of the informatio
motivational antecedents of attribution theory in examining individual user attribut
system-related outcomes. Future research could examine another tenet of attribut
likely to have interesting and important implications for ISS research—the ‘‘actor-o
bias.’’ In an ISS context, this bias suggests that observers (such as a system design
user’s supervisor) are likely to attribute causal responsibility to the actor directly e
encing the outcome (such as a user), regardless of surrounding circumstances. Sig
differences between the observers’ and actors’ attributions are likely to cause cons
conflict within the organization (Tosi et al.1986).While this bias has been examined in
other works, (Karsten 2002; Standing et al. 2006), neither of these works employed
attribution theory’s antecedents; nor did they examine the consequences of the pa
attributions for system satisfaction.
Limitations
One limitation of this study relates to its scope.Only the locus of controldimension of
attribution (internalversus external) was considered.As previously noted,other dimen-
sions related to the stability and controllability of causes exist, with each dimension
a distinctconsequence.Anotherlimitation involvesthe study’smethodology.Some
externalvalidity was sacrificed given the hypotheticalnature ofthe decision making
exercise and the use of students as subjects, in order to gain the benefits of intern
The gains to internalvalidity were made possible by being able experimentally contro
simultaneously, key constructs related to the antecedents of attribution theory. Th
Professional MBA students as subjects and associated generalizability concerns is a
uated in this case, since most of the 81 participants were full time working profess
occupying managementroles,and having moderate to extensive experience with com-
puter-based technology.
References
Abdinnour-Helm S,Chaparro B,FarmerS (2005)Using the end-usercomputing satisfaction (EUCS)
instrument to measure satisfaction of web-based systems.Decis Sci 36(2):341–359
Adler S (1980) Self-esteem and causal attributions for job satisfaction and dissatisfaction.J Appl Psychol
65(3):327–332
Anderson S,BaggettL, Widener S (2009) The impactof service operations failures on customer satis-
faction: evidence on how failures and their source affect what matters to customers. Manuf Ser
Manag 11(1):52–69
Arrington CE,Bailey CD,Hopwood WS (1985) An attribution analysis of responsibility assessmentfor
audit performance.J Account Res 23(1):1–20
Barki H,Hartwick J (1989) Rethinking the concept of user involvement. MIS Q 13(1):53–63
Benbasat I,Zmud R (2003) The identity crisis within ghe IS discipline:defining and communicating the
discipline’s core properties.MIS Q 27(2):183–194
Bhattacherjee A (2001)Understanding information systemscontinuance:an expectation–confirmation
model.MIS Q 25(3):351–370
Bogumil DD (2001) Attribution and reciprocity in international relations: the attribution reciprocity m
N Am J Psychol 3(3):463
Syst Pract Action Res
123

Davis F,BagozziR, Warshaw P (1989) User acceptance of computer technology:a comparison of two
theoretical models. Manag Sci 35(8):982–1003
DeBrabanderB, Edstrom A (1977)Successfulinformation system developmentprojects.Manag Sci
24(2):191–199
DeLone W, McLean E (1992) Information system success: the quest for the dependent variable. Inf
3(1):60–95
DeLone W,McLean E (2003) The DeLone and McClean model of information system success: a 10-ye
update.J Manag Inf Syst 19(4):9–30
Dixon AL (2001) Successful and unsuccessful sales calls: measuring salesperson attributions and be
intentions.J Mark 65(3):64
Eccles JS,Wigfield A (2002) Motivational beliefs,values,and goals.Annu Rev Psychol 53(1):109
Fang E, Evans K, Landry T (2005) Control systems’ effect on attributional processes and sales outco
cybernetic information-processing perspective. Acad Mark Sci 33(4):553–574
Franz CR,Robey D (1986) Organizationalcontext,user involvement,and the usefulness of information
systems.Decis Sci 17:329–356
Ginzberg MJ (1979) A study of the implementation process. TIMS Stud Manag Sci 13:85–102
Goodhue D (1995) Understanding user evaluations of information systems.Manag Sci 41(12):1827–1844
Goodhue D, Klein B, Salvatore T (2000) User evaluation of IS as surrogates for objective performan
Manag 38(2):87–95
Harrison PD,WestSG, Reneau JH (1988) Initialattributions and information-seeking by superiors and
subordinates in production variance investigations.Account Rev 63(2):307–320
Heider F (1958) The psychology of interpersonal relations.Wiley, New York
Hsieh P (2012) Why e-Return services fail: a psychological contact violation approach. Cyberpsycho
Soc Netw 15(12):655–662
Ives B, Olson MH (1984) User involvementand MIS success:a review of research.Manag Sci
30(5):586–603
Jarvenpaa S,Shaw T, Staples D (2004) Toward contextualized theories of trust: the role of trust in glo
virtual teams.Inf Syst Res 15(3):250–267
Jones EE, Davis KE (1965) From acts to dispositions: the attribution process in person perception. A
Psychol 2:219–266
Kaplan SE, ReckersPM (1985) An examination ofauditorperformanceevaluation.AccountRev
60(3):477–487
Karsten R (2002) An analysis of IS professional and end user causal attributions for user-system ou
J End User Comput 14(4):51
Kelley HH (1967).Attribution theory in social psychology.In D. Levine (Ed.),Nebraska Symposium on
Motivation (University of Nebraska Press) pp.192–240
Kelley HH,Michela JL (1980) Attribution theory and research.Annu Rev Psychol 31:457–501
Khayun V,Ractham P,Firpo D (2012) Assessing e-Excise success with DeLone and McLean’s Model.
J Comput Inf Syst 52(3):31–40
Kim S, Malhotra N (2005) A longitudinal model of continued IS use: an integrated view of four mech
underlying postadoption phenomena.Manag Sci 51(5):741–753
Lamb R, Kling R (2003) Reconceptualizing users as social actors in information systems research. M
27(2):197–235
Magal S,Snead K (1993) The role of causal attributions in explaining the link between user participa
and information system success. Inf Res Manag J 1993 6(3):8–19
Marakas G,Johnson R,PalmerJ (2000)A theoreticalmodelof differentialsocialattributions toward
computing technology: when the metaphor becomes the model. Int J Hum Comput Stud 52(4):
Miller DT (1976) Ego involvement and attributions for success and failure. J Pers Soc Psychol 34:901
Miller DT,Ross M (1975) Self-serving biases in the attribution of causality: fact or fiction? Psychol Bu
82:313-225
Pan G,Pan S,Newman M (2007) Information systems project post-mortems: insights from an attribut
perspective. J Am Soc Inf Syst Technol 58(14):2255–2268
Peterson D,Kim C, Kim J, Tamura T (2002) The perceptions of information systems designers from the
United States,Japan,and Lorea on success and failure factors.Int J Manag 22:421–439
Porter C,DevarajS, Daewon S (2013) A testof two models of value creation in virtualcommunities.
J Manag Inf Syst 30(1):261–292
Reeves B, Nass C (1996) The media equation: how people treat computers, television, and mew me
real people and places. Cambridge University Press,New York
Slovic P,Fischhoff B,Lichtenstein S (1977) Behavioral decision theory.Annu Rev Psychol 28:1–39
Syst Pract Action Res
123

Snead K,HarrellA (1994) An application of expectancy theory to explain manager’s intention to use
decision support system.Decis Sci 4:499–513
Snead K, Ndede-Amadi A (2002) Attributional biases as a source of conflict between users and anal
an information system developmentcontext-hypothesesdevelopment.J Syst Pract Action Res
15(5):353–365
Stajkovic AD, Sommer SM (2000) Self-efficacy and causal attributions: direct and reciprocal links. J
Soc Psychol 30(4):707
Standing C,Guilfoyle A,Lin C, Love P (2006) The attribution of success and failure in IT projects.Ind
Manag Data Syst 106(8):1148–1159
SwansonEB (1974) Managementinformationsystems:appreciationand involvement.Manag Sci
21(2):178–188
Swanson EB (1982) Measuring user attitudes in MIS research: a review.Omega 10(2):157–165
Swanson SR, Kelley SW (2001) Attributions and outcomes of the service recovery process. J Mark T
Pract 9(4):50
Sykes T,Venkatesh V,Gosain S (2009) Modelof acceptance with peer support:a socialnetwork per-
spective to understand employees system use.MIS Q 33(2):371–393
Tait P, Vessey I (1988) The effect of user involvement on system success: a contingency approach.
12(1):91–108
Thatcher J, Zimmer J, Gundlach M, McKnight D (2008) Internal and external dimensions of compute
efficacy: an empirical examination.IEEE Trans Eng Manag 55(4):628–644
Tosi H,Rizzo J, Carroll S (1986) Managing organizational behavior. A Longman Company, New York
Venkatesh V, Morris M, Davis F (2003) User acceptance of information technology: toward a unified
MIS Q 27(3):425–478
Wang W, Benasat I (2008) Attributions of trust in decision support technologies: a study of recomm
agents for E-commerce. J Manag Inf Syst 24(4):249–273
WastellDG (1999)Learning dysfunctions in information systems development:overcoming the social
defenses with transitional objects.MIS Q 23(4):581
Weary G,Stanley MA, Harvey JH (1989) Attribution. Springer, New York
Weiner B (1974) Achievement motivation and attribution theory.General Learning Press,New Jersey
WeinerB, Heckhausen H,Cook R (1972)Causalascriptions and achievementbehavior:a conceptual
analysis of effort and reanalysis of locus of control.J Pers Soc Psychol 21(2):239–248
Yi M, Jackson J,Park J,Probst J (2006) Understanding information technology acceptance by individua
professionals: toward an integrated view.Inf Manag 43(3):350–363
Zuckerman M (1979) Attribution of success and failure revisited, or: the motivational bias is alive an
in attribution theory.J Pers 47:245–287
Syst Pract Action Res
123