McElroy et al./Dispositional Factors in Internet Use
RESEARCH NOTE
DISPOSITIONAL FACTORS IN INTERNET USE:
PERSONALITY VERSUS COGNITIVE STYLE1
By:
James C. McElroy
Department of Management
Iowa State University
Ames, IA 50011-1350
U.S.A.
[email protected]
Anthony R. Hendrickson
Information Systems and Technology
Creighton University
Omaha, NE 68178
U.S.A.
[email protected]
Anthony M. Townsend
Management Information Systems
Iowa State University
Ames, IA 50011-1350
U.S.A.
[email protected]
Samuel M. DeMarie
Department of Management
Iowa State University
Ames, IA 50011-1350
U.S.A.
[email protected]
1
Ron Weber was the accepting senior editor for this paper. Dennis Galletta
was the associate editor. Henry C. Lucas and Rex Kelly Ranier, Jr., served
as reviewers. The third reviewer chose to remain anonymous.
Abstract
This study directly tests the effect of personality and cognitive
style on three measures of Internet use. The results support
the use of personality—but not cognitive style—as an antecedent variable. After controlling for computer anxiety, selfefficacy, and gender, including the “Big Five” personality
factors in the analysis significantly adds to the predictive
capabilities of the dependent variables. Including cognitive
style does not. The results are discussed in terms of the role
of personality and cognitive style in models of technology
adoption and use.
Keywords: Human factors, individual differences, individual characteristics, personality, cognitive style, end-user computing
Introduction
For decades, information system researchers have recognized
how important users’ personal factors are for predicting
technology adoption and use (e.g., Lucas 1981). However,
researchers have historically focused on a relatively mutable
subset of personal factors: individual attitudes (attitude
toward computers) and personal perceptions (such as perceived ease of use and perceived usefulness). It was not until
the 1990s that a less mutable subset of personal factors—
dispositional factors—was incorporated into predictive
models. These dispositional factors include locus of control
(Flaherty et al. 1998), self-efficacy (Agarwal et al. 2000;
Compeau et al. 1999), general personality factors (e.g., Amiel
and Sargent 2004), and cognitive style (e.g., Taylor 2004).
Our research compares the roles of two general types of
MIS Quarterly Vol. 31 No. 4, pp. 809-820/December 2007
809
McElroy et al./Dispositional Factors in Internet Use
dispositional factors—personality and cognitive style—in the
acceptance of the Internet.
Lucas’ (1973) early work on IS implementation indicated that
personal factors, decision style, and users’ attitudes toward a
system affect its adoption. He argued that personal factors
can impede implementation simply because some people find
computers incomprehensible. Zmud (1979) also showed that
there was considerable interest among researchers in the
effect of dispositional factors, specifically personality, on
MIS success. However, the role of personality was extracted
primarily from research in non-MIS contexts and was limited
to its connection to cognitive style as an antecedent to IS
success. Until recently, dispositional personal factors, such
as personality and cognitive style, were largely ignored.
A notable exception has been research on the role of negative
attitudes toward computers—what Cambre and Cook (1985)
label computer anxiety. Computer anxiety, later research
showed, is a significant predictor of computer achievement
(Marcoulides 1988) and computer use (Howard and Mendelow 1991). These studies extend Lucas’ position: Some
people find computers not only incomprehensible but
frightening as well.
The lack of targeted research on the role of dispositional
factors in IS adoption and use may be a by-product of the
Huber-Robey debate over whether cognitive style is a suitable
basis for management-information and decision-support
system design. Huber (1983) argued that cognitive style is
not necessarily applicable to MIS design, in part because of
the low amount of variance explained in outcomes. He
concluded that cognitive style should be considered only in
contexts where it can explain significant variance (such as
career choice). Conversely, Robey (1983) argued that
steering research away from cognitive style threatens to
ignore the human side of the human/computer interface.
Later research supports Robey’s prediction: The predominant
personal-factor variable used in subsequent models of IT
adoption has been perceptions rather than the more constant
factors of personality or cognitive style. For example, Davis’
(1989) original technology acceptance model (TAM), its
revised form (Davis et al. 1989), and TAM2 (Venkatesh and
Davis 2000) all rely heavily on perceptions of the technology’s usefulness and ease of use as primary determinants
of its adoption and acceptance. Over time, a number of
variables have been added to TAM, including some personal
factors such as computer anxiety and self-efficacy (see Lee et
al. 2003). Nonetheless, most of these added variables involve
perceptions (e.g., subjective norms, image, job relevance, and
output quality). Similarly, Seddon’s (1997) interpretation of
810
MIS Quarterly Vol. 31 No. 4/December 2007
DeLone and McLean’s (1992) model of IS success uses
perceptions of usefulness, satisfaction, and system quality to
predict IS use. In 2003, in an effort to resolve competing
models of IS use and adoption, Venkatesh, Morris, Davis, and
Davis proposed a unified theory of acceptance and use of
technology (UTAUT). Nonetheless, UTAUT also overlooks
personal, dispositional factors, in favor of perceptions of
performance, effort required, and others (the only nonperceptual variables were gender, age, and experience). Even
though the role of perceptions continues to dominate models
of technology acceptance and use, individual attitudes such as
user satisfaction and attitudes toward computers are also
acknowledged as important variables (Venkatesh et al. 2003).
Only recently, however, have dispositional personal factors
reentered the picture. In the 1990s, nearly 20 years after
Lucas’ work, interest turned toward single, relatively stable
personality traits. Self-efficacy, for example, is essential to
both the theory of planned behavior (Mathieson 1991; Taylor
and Todd 1995) and social cognitive theories of user reaction
to IT (Compeau and Higgins 1995a, 1995b; Compeau et al.
1999). Self-efficacy, in both general (Bandura 1977) and
computer-specific (Compeau and Higgins 1995b) forms, has
also been examined as a personal factor affecting new IT use
as well as a dependent variable to be honed through training
(Agarwal et al. 2000).
Recent research has examined general dispositional factors as
determinants of technology adoption and use. Mirroring the
Huber-Robey debate, this research generally defines disposition in terms of either general personality traits or cognitive
style. Consequently, this paper will compare the ability of
these two competing conceptual frameworks to explain
Internet adoption and use.
The Role of Personality in IS Use
Personality is a stable set of characteristics and tendencies
that determine peoples’ commonalities and differences in
thoughts, feelings, and actions (Maddi 1989). Many individual traits have been identified, but this study focuses on the
so-called “Big Five” personality factors: agreeableness, conscientiousness, extraversion, neuroticism, and openness to
experience. These factors theoretically capture the essence of
one’s personality (Digman 1990).
We selected these personality factors for three reasons. First,
using the Big Five allows us to examine whether personality,
in general, affects Internet use. We are interested in defining
personality’s direct effect on IS use, not in casting personality
McElroy et al./Dispositional Factors in Internet Use
as a precursor to cognitive behaviors and attitude formation
(Zmud 1979). Second, using general personality factors
allows us to revisit the cognitive style debate. Pitting the Big
Five personality dimensions directly against cognitive style
instruments such as the Meyers-Briggs Type Indicator
(Gardner and Martinko 1996) will inform this debate. Finally,
comparing the effects of a general personality measure to a
general cognitive style ensures consistency in the levels of
abstraction.
videos) and more time online engaged in academic pursuits
(Landers and Lounsbury 2006).
Each Big Five personality factor represents a collection of
unique personality traits. Agreeableness represents the tendency to be sympathetic, good-natured, cooperative, and
forgiving. Highly agreeable people help others and expect
help in return. Conscientiousness represents the tendency to
be self-disciplined, strong-willed, deliberate, and reliable.
Conscientious people actively plan, organize, and carry out
tasks. Extraversion represents sociability, cheerfulness, and
optimism. Extraverts seek out new opportunities and excitement. Neuroticism represents a lack of psychological adjustment and emotional stability. Highly neurotic people tend to
be fearful, sad, embarrassed, distrustful, and have difficulty
managing stress. Finally, openness to experience represents
one’s curiosity and willingness to explore new ideas. Open
individuals tend to devise novel ideas, hold unconventional
values, and willingly question authority (Costa and McCrae
1992). Taken together, the Big Five capture the essence of
one’s personality.
The distrust inherent in people with neurotic personalities has
tended to limit the amount of time they spend online
exchanging information and playing games (Swickert et al.
2002). Nonetheless, neurotic people do spend extensive time
on the Internet seeking to gain a sense of belonging (Amiel
and Sargent 2004).
Organizational behavior researchers have long been interested
in personality because of its established link to behaviors and
cognitions. For example, nearly a decade’s worth of research
has linked various dimensions of the Big Five to job performance (Barrick and Mount 1991), job choice (Spector et al.
1995), implicit leadership theories (Keller 1999), absenteeism
(Judge et al. 1997), occupational stress, and other workrelated attitudes and values (see Tokar et al. 1998).
Only recently has research begun to link personality traits to
IS adoption and use. For example, agreeable people tend not
to use the Internet for email as much as others. Although perfectly willing to help others, agreeable people may not necessarily feel compelled to establish that connection (Swickert et
al. 2002). Nonetheless, the forgiving nature of highly agreeable people may make them more persistent in the face of
frustrating Internet investigations (e.g., sites that are difficult
to navigate), keeping them online longer (Landers and
Lounsbury 2006).
Conscientious people are less likely to use the Internet for
what they see as unproductive activities. They tend to spend
less time online in leisure pursuits (e.g., watching YouTube
Extraverts prefer face-to-face interaction and typically spend
less time on the Internet (Landers and Lounsbury 2006), especially for online social activities such as chat rooms (Hamburger and Ben-Artizi 2000). Nonetheless, they will use the
Internet as a tool to acquire things to share with others, such
as information and music (Amiel and Sargent 2004).
Finally, some research shows that open people are attracted to
online activity to sate their curiosity and seek out new forms
of adventure (Tuten and Bosnjak 2001).
These studies suggest that the Big Five personality dimensions have at least some utility as determinants of information
system use. That is, people dominated by different personality characteristics will use the Internet to varying degrees
and for different purposes.
The Role of Cognitive Style in IS Use
As some researchers began investigating the influence of
personality on IS use, others developed a renewed interest in
the influence of cognitive style. Cognitive style research is
based on psychiatrist Carl Jung’s (1921) premise that the
mental functions related to information gathering and
decision-making are central to one’s personality. Consequently, people are “typed” according to how they perceive
and form judgments. The Myers-Briggs Type Indicator, or
MBTI, is the primary instrument used to capture Jung’s concepts (Wheeler et al. 2004). Despite criticism of the dichotomous nature of its typing (Gardner and Martinko 1996), the
MBTI has been validated extensively and shown to be reliable
(Harvey 1996).
In a representative MBTI study, Ludford and Terveen (2003)
showed that a small sample of 20 individuals used various
task-oriented technologies differently depending on their
MBTI type. For example, perceivers were more likely to save
task-related e-mail once a project was complete, while judgers
were more likely to delete them. Moreover, thinkers were
more likely to use editorial reviews in evaluating CDs on
MIS Quarterly Vol. 31 No. 4/December 2007
811
McElroy et al./Dispositional Factors in Internet Use
Amazon.com, while feelers were more likely to rely on their
own listening experience. In a larger study of 257 software
development employees, Taylor (2004) found that cognitive
style affected use of the company’s knowledge management
infrastructure. Specifically, analytical people were more
likely than intuitive people to use the company’s data mining,
knowledge repository, and Lotus Notes features. As with the
research on personality, these cognitive style studies offer
additional support for a dispositional view of personal factors
as a determinant of information system adoption, and suggest
that the way people process information and make judgments
affects their Internet use.
Method
Participants were 92 MBA students and 61 senior undergraduate students from a variety of majors. We collected the
data using a lengthy survey (participating students received a
few extra credit points). The length of the survey may have
affected sample size because some questions were left blank.
Nonetheless, of 153 subjects, 132 answered every question.
Measures
Control Variables
Propositions
Much research links either personality or cognitive style to IS
use and adoption. However, we could find no studies that
examine the relative contribution of both factors. Therefore,
the current study fills that gap by comparing the relative
effects of these two types of dispositional factors. Rather than
focusing on professional users of specialized IS technologies,
we focus on use of the Internet, which is the most widely used
IS technology. We do not examine the specific role of each
personality factor or cognitive-style type. Instead, we seek to
determine if personality and cognitive style are, in fact, antecedents of IS adoption and use and, if so, to evaluate their
relative worth as predictors of Internet use. Consequently,
our research addresses the following propositions:
Proposition 1. Personality explains variation in Internet use
beyond the traditional antecedent measures of computer
attitudes, self-efficacy, and gender.
Proposition 2. Cognitive style explains variation in Internet
use beyond the traditional antecedent measures of computer attitudes, self-efficacy, and gender.
The purpose of this study was to determine how well personality and cognitive style predict IT (i.e., Internet) adoption and
use. Therefore, we used three variables common to earlier
adoption-and-use models as control variables: computer
anxiety, self-efficacy, and gender.
Computer anxiety was measured by a 20-item, attitudetoward-computers scale as defined by Dambrot et al. (1981).
Significantly, this scale captures general anxiety about the
implications involved in using computers, rather than the
specific anxiety associated with actually using a computer.
We avoided using the more-specific Computer Anxiety
Rating Scale (Heinssen et al. 1987), however, because it
includes items that might have overlapped with our selfefficacy scale (such as “I am confident that I can learn
computer skills”). The Dambrot scale employs a five-point
Likert response format. Sample items include “Computers
intimidate and threaten me” and “Even though computers are
valuable I fear them.” An average score was computed for
each respondent such that a high score is indicative of high
computer anxiety. Coefficient alpha for the sample was .72.
The debate between Huber (1983) and Robey (1983)
regarding the role of cognitive style in MIS design leads us to
predict that personality and cognitive style will affect information technology adoption and use differently. Huber called
for abandoning cognitive style as a determinant of IS design
due to its lack of explanatory power. If Huber is correct,
personality should have greater power to predict Internet use.
Therefore,
Self-efficacy refers to a person’s belief in his or her capacity
to perform, which influences the person’s desire to try
(Bandura 1986; Torkzadeh et al. 2003). Self-efficacy can be
measured both in terms of general beliefs, such as one’s sense
of overall effectiveness, and by context-specific beliefs, such
as computer self-efficacy. We elected to use a general measure of self-efficacy because we wanted to remain parallel
with our general measures of personality, cognitive style, and
computer anxiety. We used a 12-item form of Sherer et al.’s
(1982) self-efficacy scale, which has statements such as “If I
can’t do a job the first time, I keep trying until I can.”
Coefficient alpha for the scale was .80.
Proposition 3. Personality will better predict Internet use
than cognitive style.
For our final control variable, we used a one-item scale
measuring gender (1 = male, 2 = female). The interpretation
812
MIS Quarterly Vol. 31 No. 4/December 2007
McElroy et al./Dispositional Factors in Internet Use
of personality scores—like some MBTI preferences—varies
by gender (Costa and McCrae 1992; Gardner and Martinko
1996).
Personality
We measured personality with Costa and McCrae’s (1992)
Revised NEO Personality Inventory. This widely used
instrument is a 240-item questionnaire that describes the
respondent’s personality according to the Big Five factors. Its
validity and reliability are well documented (Costa and
McCrae 1992). Coefficient alphas for the scales used in this
study ranged from .89 for agreeableness to .93 for neuroticism.
Cognitive Style
A number of instruments measure cognitive style, such as the
Kirton Adaptors Innovators (KAI) scale (Kirton 1976) and the
Cognitive Style Index (CSI) (Allinson and Hayes 1996). The
most common, however, is the Myers-Briggs Type Indicator
(Wheeler et al. 2004). We opted to use the MBTI for this
study for several reasons. First, the MBTI focuses on how
one makes judgments and arrives at conclusions, rather than
on how creativity is expressed (the focus of the KAI scale).
Second, even though instruments such as the CSI are shorter,
they are based on and are highly correlated with the MBTI
(Allinson and Hayes 1996). Third, the MBTI is extremely
popular in industry, and it is administered to more than three
million people a year (more than 40 percent from major
corporations) (Gardner and Martinko 1996). Researchers
have argued that studies based on such widely used instruments have more relevance to organizations than studies
based on seldom-applied research measures (Garfield et al.
2001), such as the KIA and CSI.
We used the MBTI Form M, a 93-item instrument (Myers and
Myers, 1998). (Gardner and Martinko assess its reliability
and validity.) This version of the MBTI uses a forced-choice
format, meaning respondents select which of two statements
for each item most applies to them. Because the dimensions
are opposites, we calculated a difference score for each
respondent on four dimensions: thinking/feeling, extraversion/introversion, sensing/intuition, and judgment/
perceptions. A high score indicates thinking, extraversion,
sensing, and judgment. Given the dichotomous nature of the
scoring scheme, KR-20 estimates of reliability were used.
Values for these scales ranged from .73 for the extraversion/
introversion scale to .92 for the judgment/perceptions scale.
Dependent Variables
We designed three instruments to measure Internet use. The
first had seven items that used a five-point Likert response
scale. Four items asked respondents to indicate how often
they surfed the Internet, visited chatrooms, visited bulletin
boards, and looked up specific information. The remaining
three items asked respondents how comfortable they were
looking up information, visiting chat rooms, and surfing the
Internet. Responses were averaged to yield a single score
indicating each respondent’s level of overall Internet usage.
Coefficient alpha for this sample was .71.
The second instrument had two items and a five-point Likert
response scale that measured respondents’ willingness to buy
products online. Respondents were asked to respond to the
statements “I would be willing to buy products on the
Internet” and “I buy products on the Internet.” Coefficient
alpha for this e-buy scale was .84.
Finally, a two-item instrument with a five-point Likert
response scale measured respondents’ willingness to sell
products on-line. Respondents indicated their degree of
agreement with the statements “I would be willing to sell
products on the Internet” and “I have sold items on the
Internet.” As with the other scales, responses were averaged
to yield a single e-sell score. Coefficient alpha for this
instrument was .70.
Results
Because data were collected from both graduate and undergraduate students, t-tests were conducted to determine
whether the overall data could be pooled and treated as a
single sample. Because no significant differences existed
between graduate and undergraduate student means for all
dependent variables, the samples were combined.
Table 1 shows the means, standard deviations, and correlations for the variables. To determine how much, comparatively, personality and cognitive style help explain variance
in Internet use, we conducted two hierarchical regressions. In
step one, we entered computer anxiety, self-efficacy, and
gender as control variables because they are well established
in the IS adoption and use literature. In step two, we entered
the Big Five personality factors and cognitive style in each
analysis to test whether they add significantly to the model.
Tables 2 and 3 report the results.
MIS Quarterly Vol. 31 No. 4/December 2007
813
119.68
85.77
115.16
1.99
2.28
-.16
5.45
3.54
4.11
2.52
NEOPIRE
(Extrav)
NEOPIRN
(Neuro)
NEOPIRO
(Open)
MBTI T-F
MBTI E-I
MBTI S-N
MBTI J-P
Internet Use
E-buy
E-sell
Std. Dev.
1.22
.96
.58
12.83
13.68
12.46
13.62
20.40
23.67
21.16
18.47
18.65
.58
.44
.36
CA
-.25**
-.25**
-.29***
.15
.21*
.04
-.13
-.26**
.15
-.24**
-.11
-.07
.02
-.23**
.72
Self Eff
.04
.20*
.13
.10
-.06
.13
.19*
.24**
-.52***
.36***
.59***
.19*
-.03
.80
Gender
-.01
.21*
-.06
.16
.04
-.04
-.25**
.14
.17*
.06
.13
.09
na
.03
.12
-.08
.09
.08
.02
-.23**
.02
-.32***
.06
.27***
.89
Agree
.04
.07
.03
.40***
.26**
.00
.29***
.04
-.38***
.34***
-.90
Numbers in bold along the diagonal are coefficient alpha values for the scales.
Sample size of correlations varies due to missing data.
**p < .01
***p < .001
*p < .05
126.84
NEOPIRC
(Consc)
1.55
Gender
116.30
4.00
Self-Efficacy
(Self Eff)
NEOPIRA
(Agree)
2.32
Mean
Computer
Anxiety (CA)
Variable
Consc
Table 1. Descriptive Statistics and Correlations (N = 153)
Extrav
.11
.04
.25**
-.08
-.20*
.67***
-.12
.44***
-.27***
.91
Neuro
.21*
.08
.04
.14
.21*
-.14
-.26**
-.13
.93
Open
.03
.17*
.30***
-.31***
-.61**
.22**
-.14
.90
T-F
.08
.04
.05
.21*
.15
-.19*
.85
E-I
.08
-.10
.21*
-.10
-.11
.73
S-N
.13
.04
-.17*
.61***
.91
J-P
.05
.16
-.10
.92
Use
.30***
.30***
.71
E-buy
.38***
.84
.70
E-sell
McElroy et al./Dispositional Factors in Internet Use
Table 2. Regression Analyses of the Big Five Personality Factors on Internet Use, E-Buy, and E-Sell,
Controlling for Computer Anxiety, Self-Efficacy, and Gender (N = 137)
Internet Use
Step 1
Computer Anxiety
Self-Efficacy
Gender
-.27**
.06
-.05
Step 2
Agreeableness
Conscientiousness
Extraversion
Neuroticism
Openness
*p < .05
**p < .01
E-Sell
Step 2
Step 1
Step 2
-.22*
.07
-.12
-.20*
.17*
.20*
-.20*
.33*
.15
3.61***
3.09**
.10
.14
Step 1
-.26**
-.10
.00
6.03***
6.03***
.12
.10
3.57***
1.96
.06
.14
Step 2
-.25**
-.09
-.10
.11
.17
.15
.36***
-.00
.12
-.08
-.08
.25***
.08
-.07
.02
.14
.19
.21*
4.13**
4.13**
.09
.07
F
F Change
Change in R²
Adjusted R²
E-Buy
3.03*
3.03*
.07
.04
3.23**
3.20**
.11
.12
***p < .001
Table 2 shows that after controlling for computer anxiety,
self-efficacy, and gender, personality significantly helps
explain Internet use variance in general, and online selling in
particular. Openness is a significant predictor of general
Internet use; open individuals are more likely to use the
Internet. Neuroticism also adds to the explanation for Internet
use variance, even though it is just below the conventional
standard of statistical significance (p = .059). Neuroticism is
a strong predictor of e-selling; neurotic individuals are more
likely to sell products online. The addition of the Big Five
personality factors in step 2 resulted in significant F changes
for both Internet use and E-selling activities, which provides
support for Proposition 1.
Table 3 shows that after controlling for computer anxiety,
self-efficacy, and gender, cognitive style indicators failed to
significantly predict any of the dependent variables. There
was a significant beta for the extraversion/introversion
dimension as a predictor of general Internet use, but the
MBTI as a whole failed to add significantly to the model.
Thus, Proposition 2 is not supported. Taken together, the
results shown in Tables 2 and 3 support Proposition 3: The
Big Five personality factors better predict Internet use than
cognitive style.
To further support Proposition 3, we conducted a usefulness
analysis, which reports the incremental change in variance for
all possible orderings of applicable variables entered into
hierarchical regression analysis (Darlington, 1968). In our
primary analysis, the Big Five or the MBTI was entered into
the regression equation in step 2. In our usefulness analysis,
this step is followed by the insertion of the other possible
explanatory variables.
Table 4 reports on the incremental change in variance
explained when the Big Five is entered both before and after
the MBTI (and vice versa). Table 4 shows that the Big Five
adds significantly to the model, whether they are entered
before or after the MBTI. In contrast, the MBTI fails to add
significant explanatory power regardless of when it is entered.
The lone exception to this pattern is e-buying behavior.
Recall that the Big Five did not add significantly more
explanatory power beyond that explained by computer
anxiety, self-efficacy, and gender (Table 2). This is illustrated
here by the lack of significance for the Big Five in step 2 of
our primary analysis. The MBTI, however, does add significantly to the model in our usefulness analysis, but only after
the Big Five are taken into account. In other words, the
MBTI is only relevant to e-buying behavior in the presence of
the Big Five. The analysis in Table 4 supports Proposition 3:
personality is a more important predictor of Internet use than
cognitive style. These results support Huber’s (1983) contention that cognitive style is of little value as a basis for MIS
and DSS design.
MIS Quarterly Vol. 31 No. 4/December 2007
815
McElroy et al./Dispositional Factors in Internet Use
Table 3. Regression Analyses of the MBTI on Internet Use, E-Buy, and E-Sell, Controlling for Computer
Anxiety, Self-Efficacy, and Gender (N = 137)
Internet Use
Step 1
Step 1
Computer Anxiety
Self-Efficacy
Gender
Step 2
-.29***
-.01
-.00
-.29***
.02
-.04
Step 2
MBTI T-F
MBTI E-I
MBTI S-N
MBTI J-P
Step 1
-.20*
.16
.22**
3.05**
1.98
.05
.14
4.35**
4.35**
.09
.09
**p < .01
E-Sell
Step 2
-.21
.16
.18*
Step 1
Step 2
-.31***
-.12
.02
-.27**
-.09
-.01
.08
.14
.23*
-.04
-.01
-.12
-.03
.17
.07
.21*
-.05
-.08
F
F Change
Change in R²
Adjusted R²
*p < .05
E-Buy
6.32***
6.32***
.13
.11
3.56**
1.43
.04
.12
2.63*
2.11
.06
.08
3.22*
3.22*
.07
.05
***p < .001
Table 4. Usefulness Analysis: Comparison of the R² Incremental Change for Each Step (N = 132)
Incremental Change Explained in†
Step: Predictor Variables
Internet Use
E-Buy
E-Sell
Control Variables
.09**
.12***
.07*
Personality Entered in First
Step 2: Big Five Factors
Step 3: MBTI
.10**
.04
.06
.07*
.11**
.06
Cognitive Style Entered in First
Step 2: MBTI
Step 3: Big Five Factors
.06
.08*
.04
.09*
.06
.11**
†
The values represent the additional change in R² achieved by entering the variables specified at each step.
*p < .05
**p < .01
***p < .001
Discussion
The results of this study show that general personality factors
predict aspects of Internet use while cognitive style does not.
Two general issues merit attention. First, why would personality affect Internet use and Internet selling (but not Internet
buying) behavior? We believe that these differential results
are due to both measurement and familiarity issues. Most
items measuring Internet use in this study dealt with how fre-
816
MIS Quarterly Vol. 31 No. 4/December 2007
quently people use the Internet. Consequently, even though
computers are virtually unavoidable today, personality may
still explain extreme behaviors in terms of using them. Ebuying and e-selling, in contrast, were measured in an
absolute sense; people do it or they do not. Buying products
on the Internet has become commonplace, particularly for the
age group in this study—indeed, online retail spending is
expected to reach $144 billion and 50 percent of total retail
sales by the year 2010 (Anfuso 2006). Selling products over
McElroy et al./Dispositional Factors in Internet Use
the Internet, however, is a relatively new phenomenon.
Future research might examine the role of personality in
online social networks, such as Facebook and MySpace, and
in other Internet innovations. If verified, our results suggest
personality may predict adoption of new technologies as well
as extreme uses of established ones.
The second issue that merits further attention is the extent to
which personality can predict Internet use. Table 2 demonstrates two important points. First, according to the change in
R2 values, personality explains an additional 10 percent of
Internet use variance beyond the 9 percent explained by the
control variables. With respect to e-selling, personality
explains 11 percent beyond the 7 percent explained by the
control variables. Second, personality increases the variance
explained by the controls by a factor of 2 for Internet use and
a factor of 3 for e-selling behavior, according to the adjusted
R2 values. This is an increase from .07 to .14 and from .04 to
.12, respectively. Some might argue that adjusted R2 values
of .12 and .14 are not particularly high, but explaining 12 to
14 percent of the variance in highly complex human behavior
is meaningful. Moreover, the purpose of this paper was to
comparatively test personality and cognitive style as predictors of Internet use. These results clearly show that personality is a superior predictor. Cognitive style fails to explain
significant amounts of variance beyond what was explained
already by the control variables of computer anxiety, selfefficacy, and gender.
The results suggest several more specific observations about
the role of personality. Neuroticism is a strong predictor of eselling behavior and just missed the standard statistical cut-off
for Internet use significance. Seeking information, socializing, and selling goods online may enable neurotic people to
escape the stress of face-to-face interaction (this is consistent
with the findings of Amiel and Sargent 2004). Openness
plays a smaller, although still significant, role in predicting
Internet use. Curious and open-minded people are more
likely to use the Internet to seek information, and visit
chatrooms and bulletin boards (this is consistent with the
finding of Tuten and Bosnjak 2001). One might have expected openness to also predict e-buying and e-selling
behaviors, but the financial, effort, and time obligation of
such behavior may overcome their curiosity. Visiting chat
rooms and bulletin boards, in contrast, carries no financial
obligation. Although care must be taken in drawing causal
inferences, personality clearly temporally precedes Internet
use and e-commerce behaviors.
Our findings suggest that future models of IS implementation
and adoption deserve a dispositional component, but one that
is based on personality rather than cognitive style.
This study addressed only the direct effects of personality and
cognitive style on Internet use. The role of personality, if
replicated, may prove even more salient if one also examines
its indirect effects. Future models of IS adoption and use may
be improved by incorporating personality along with existing
attitudinal and situational determinants. In addition to the
independent effects shown here, personality may explain IS
adoption and use, because it is an antecedent of attitudes such
as user satisfaction. Thus, the indirect effects of personality
may explain additional adoption and use variance.
While cognitive style does not seem promising for predicting
Internet adoption and use, other personal factors, such as
personal values, might be useful. In addition, other measures
of cognitive style might yield different results. Future
research should examine whether personality is best captured
using broad dimensions, as was the case in this study, or
specific personality traits, such as self-esteem and tolerance
for ambiguity. Previous research suggests that specific traits
may be particularly good predictors in cases where a
theoretical rationale posits an expected relationship between
a given trait and a given criterion (George 1992).
Future research should also examine more specific types of IS
adoption and use. Personal, dispositional factors may predict
some types of IS adoption and use better—or at least differently—than others. The adoption of new IS technologies
and the pervasiveness of existing technologies may be
affected by individual differences such as personality. While
millions of people currently shop online, some are more likely
to do so than others. Is this a function of their personality?
Moreover, the type of products bought and sold online may be
personality-dependent. For example, buying books, music, or
travel tickets online may involve different antecedents than
buying specialized products or selling items on eBay.
Limitations
The results of this study must be viewed in light of its limitations. First, participant behavior was self-reported during a
single session. Consequently, common-method bias is a potential weakness (future researchers could test for this bias by
using a larger sample). Ideally, in addition, future researchers
will measure Internet use behaviorally. Second, it is true that
student samples can sometimes be problematic, but they are
appropriate in this study because college students are real
Internet users and provide a representative sample of this
population. If this sample has any bias, it is due to students’
above-average familiarity with the Internet (this may explain
the lack of significance with respect to buying online). Internet purchasing is so common to the sample generation that
age may be a more salient individual variable than personality.
MIS Quarterly Vol. 31 No. 4/December 2007
817
McElroy et al./Dispositional Factors in Internet Use
Additional potential weaknesses of this study involve the way
self-efficacy and computer anxiety, two of the control variables, were operationalized. We used a general measure of
self-efficacy. The results might have been different had we
used the more-specific computer self-efficacy as a control.
However, we were able to perform a post hoc analysis to
address this potential weakness because our data set included
Torkzadeh, Koufteros, and Pflughoeft’s (2003) measure of
advanced computer skills self-efficacy. Substituting their
computer self-efficacy measure (coefficient alpha = .90) for
our general measure of self-efficacy revealed identical results.
Personality significantly predicted computer use and e-sell
(but not e-buy) behavior. Cognitive style did not significantly
add to the model for any of the three dependent variables.
The only major difference when we substituted computer selfefficacy was that neuroticism (rather than openness) was the
main contributing factor in personality’s prediction of Internet
use.
In a second post hoc analysis, we attempted to assess whether
a more specific measure of computer anxiety would affect the
results. The computer anxiety measure we used technically
measured general attitudes toward computers, although, as
noted, these measures were subsumed under the label of
“computer anxiety” (Cambre and Cook 1985). Three items
from our measure are common to other, more-specific measures of computer anxiety (Heinssen et al. 1987). Consequently, we used these items to form a three-item computer
anxiety instrument that had a coefficient alpha of .71. Substituting it for the more complete computer anxiety measure
produced results nearly identical to our original findings. The
only major difference was that computer anxiety, as measured
by the shorter scale, was not a significant predictor of Internet
use in its own right.
While this second, post hoc analysis addresses the validity of
our computer anxiety measure, it does not address the issue of
specificity. Using a measure focused on the anxiety associated with using a computer, such as Marcoulides, Rosen,
and Sears’ (1985) computer anxiety rating scale, as opposed
to one measuring anxiety about the implications of using computers, may produce differential results. However, because
we were investigating general aspects of personality and
cognitive style, we opted for general measures of self-efficacy
and computer anxiety to be consistent across concepts.
Implications
A literature review on the role of personality in organizational
life concluded that personality is essential to understanding
some classes of organizational phenomena (George 1992).
818
MIS Quarterly Vol. 31 No. 4/December 2007
Our study provides evidence for including Internet use in this
set of phenomena. Current e-buying and e-selling vendors
have focused on developing applications that attempt to customize each consumer’s shopping experience. They seek to
provide users with information that is tailored to their specific
needs (Callaghan 2000). This trend is evidenced by the proliferation of “portal” software designed to provide users with
preprogrammed, personalized information. However, the
trend is clearly still in what Amazon CEO Jeff Bezos calls the
early, “Kitty Hawk stage” of personalization (Du Bois 2000).
Firms that can understand their customers’ personality and
buying behavior will have a competitive advantage in the
marketplace. A richer understanding of the effect of personality on IS use will allow Internet firms to better target their
products and services to specific markets.
More research is needed to investigate promising individual
differences. The first step is to reintroduce dispositional
personal factors into models of technology use and adoption.
TAM2 (Venkatesh and Davis 2000) and UTAUT(Venkatesh
et al. 2003) are two viable candidates. This study suggests
that personality, not cognitive style, provides a promising
avenue for such research.
Acknowledgments
The authors would like to thank the senior editor, Ron Weber,
associate editor, Dennis Galletta, and the reviewers for their assistance in the development of this manuscript. We also acknowledge
the editorial assistance of Francis and Stephanie Storrs.
References
Agarwal, R., Sambamurthy, V., and Stair, R. M. “Research Report:
The Evolving Relationship Between General and Specific
Computer Self-Efficacy: An Empirical Assessment,” Information Systems Research (11:4), 2000, pp. 418-430.
Allinson, C. W., and Hayes, J. “The Cognitive Style Index: A
Measure of Intuition-Analysis for Organizational Research,”
Journal of Management Studies (33:1), 1996, pp. 119-135.
Amiel, T., and Sargent, S. L. “Individual Differences in Internet
Usage Motives,” Computers in Human Behavior (20:6), 2004,
pp. 711-726.
Anfuso, D. “Report: Online Retail Spending to Grow,” iMedia
Connection, February 7, 2006 (available online at http://www.
imediaconnection.com/news/8191.asp)
Bandura, A. “Self-Efficacy: Toward a Unifying Theory of
Behavioral Change,” Psychological Review (84:2), 1977, pp.
191-215.
Bandura, A. Social Foundations of Thought and Action, PrenticeHall, Englewood Cliffs,NJ, 1986.
McElroy et al./Dispositional Factors in Internet Use
Barrick, M. R., and Mount, M. K. “The Big Five Personality
Dimensions and Job Performance: A Meta-Analysis,” Personnel
Psychology (44:1), 1991, pp. 1-26.
Callaghan, D. “Personalization is Key to Web Success,” eWeek
(17:27), July 3, 2000, p. 23.
Cambre, M. A., and Cook, D. L. “Computer Anxiety: Definition,
Measurement, and Correlates,” Journal of Educational Computing Research (1:1), 1985, pp. 37-54.
Compeau, D. R., and Higgins, C. A. “Application of Social Cognitive Theory to Training for Computer Skills,” Information
Systems Research (6:2), 1995a, pp. 118-143.
Compeau, D. R., and Higgins, C. A. “Computer Self-Efficacy:
Development of a Measure and Initial Test,” MIS Quarterly
(19:2), 1995b, pp. 189-211.
Compeau, D. R., Higgins, C. A., and Huff, S. “Social Cognitive
Theory and Individual Reactions to Computing Technology: A
Longitudinal Study,” MIS Quarterly (23:2), 1999, pp. 145-158.
Costa, P. T., and McCrae, R. R. Revised NEO Personality Inventory
(NEO-PI-R) and NEO Five-Factor Inventory (NEO-FFI) Professional Manual, Psychological Assessment Resources, Odessa,
FL, 1992.
Dambrot, F. H., Watkins-Malek, M. A., Siling, S. M., Marshall, R.
S., and Garver, J. “Correlates of Sex Differences in Attitudes
Toward and Involvement with Computers,” Journal of Vocational Behavior (27:1), 1985, pp. 71-86.
Darlington, R. B. “Multiple Regression in Psychological Research
and Practice,” Psychological Bulletin (69:3), 1968, pp. 161-182.
Davis, F. D. “Perceived Usefulness, Perceived Ease of Use, and
End User Acceptance of Information Technology,” MIS
Quarterly (13:3), 1989, pp. 319-340.
Davis, F. D., Bagozzi, R. P., and Warhsaw, P. R. “User Acceptance
of Computer Technology: A Comparison of Two Theoretical
Models,” Management Science (35:8), 1989, pp. 982-1003.
DeLone, W. H., and McLean, E. R. “Information System Success:
The Quest for the Dependent Variable,” Information Systems
Research (3:1), 1992, pp. 60-95.
Digman, J. M. “Personality Structure: Emergence of the FiveFactor Model,” Annual Review of Psychology (41:1), 1990, pp.
417-440.
Du Bois, G. “Portals Getting Personal: Personalization Technology
Goes Corporate,” eWeek (17:27), July 3, 2000, pp. 1, 22.
Flaherty, L. M., Pearce, K. J., and Rubin, R. B. “Internet and Faceto-Face Communication: Not Functional Alternatives,” Communication Quarterly (46:3), 1998, pp. 250-266.
Gardner, W. L., and Martinko, M. J. “Using the Myers-Briggs Type
Indicator to Study Managers: A Literature Review and Research
Agenda,” Journal of Management (22:1), 1996, pp. 45-83.
Garfield, M. J., Taylor, N. J., Dennis, A. R., and Satzinger, J. W.
“Research Report: Modifying Paradigms—Individual Differences, Creativity Techniques, and Exposure to Ideas in Group
Idea Generation,” Information Systems Research, (12:3), 2001,
pp. 322-333.
George, J. “The Role of Personality in Organizational Life: Issues
and Evidence,” Journal of Management (18:2), 1992, pp.
185-213.
Hamburger, Y. A., and Ben-Artizi, E. “The Relationship Between
Extraversion and Neuroticism and the Different Uses of the Internet,” Computers in Human Behavior (16:4), 2000, pp. 441-449.
Harvey, R. J. “Reliability and Validity,” In MBTI Applications: A
Decade of Research on the Myers-Briggs Type Indicator, A. L.
Hammer (ed), Consulting Psychologists Press, Palo Alto, CA,
1996, pp. 5-29.
Heinssen, R. K., Glass, C. R., and Knight, L. A. “Assessing
Computer Anxiety: Development and Validation of the Computer Anxiety Rating Scale,” Computers in Human Behavior
(3:1), 1987, pp. 49-59.
Howard, G. S., and Mendelow, A. L. “Discretionary Use of Computers: An Empirically Derived Explanatory Model,” Decision
Sciences (22:2), 1991, pp. 241-265.
Huber, G. P. “Cognitive Style as a Basis for MIS and DSS Designs:
Much Ado About Nothing?” Management Science (29:5), 1983,
pp. 567-579.
Judge, T. A., Martocchio, J. J., and Thoresen, C. J. “Five-Factor
Model of Personality and Employee Absence,” Journal of Applied Psychology (82:5), 1997, pp. 745-755.
Jung, C. G. Psychological Types, Princeton University Press,
Princeton, NJ, 1921.
Keller, T. “Images of the familiar: Individual Differences and
Implicit Leadership Theories,” The Leadership Quarterly (10:4),
1999, pp. 589-607.
Kirton, M. J. “Adaptors and Innovators: A Description and Measure,” Journal of Applied Psychology (61:5), 1976, pp. 622-629.
Landers, R. N. and Lounsbury, J. W. “An Investigation of Big Five
and Narrow Personality Traits in Relation to Internet Usage,”
Computers in Human Behavior (22:2), 2006, pp. 283-293.
Lee, Y., Kozar, K. A., and Larsen, K. R. T. “The Technology
Acceptance Model: Past, Present, and Future,” Communications
of the Association for Information Systems (12:50), 2003, pp.
752-780.
Lucas, H. C., Jr. Computer Based Information Systems in Organizations, Science Research Associates, Inc., Chicago, IL, 1973.
Lucas, H. C., Jr. Implementation, the Key to Successful Information
Systems, Columbia University Press, New York, 1981.
Ludford, P., and Terveen, L. “Does an Individual’s Myers-Briggs
Type Indicator Preference Influence Task-Oriented Technology
Use?,” in Human-Computer Interaction INTERACT ’03, G. W.
M. Rauterberg, M. Menozzi, and J. Wesson (eds.), IOS Press,
Amsterdam, 2003, pp. 623-630.
Maddi, S. R. Personality Theories: A Comparative Analysis (5th
ed.), Dorsey, Homewood, IL, 1989.
Marcoulides, G. A. “The Relationship Between Computer Anxiety
and Computer Achievement,” Journal of Educational Computing
Research (4:2), 1988, pp. 151-158.
Marcoulides, G. A., Rosen, L. D., and Sears, D. C. The Computer
Anxiety Scale, California State University, Dominguez Hills, CA,
1985.
Mathieson, K. “Predicting User Intentions: Comparing the Technology Acceptance Model with the Theory of Planned Behavior,”
Information Systems Research (2:3), 1991, pp. 173-191.
Myers, P. B., and Myers, K. D. Myers-Briggs Type Indicator
Form M, Consulting Psychologists Press, Inc., Palo Alto, CA,
1998.
MIS Quarterly Vol. 31 No. 4/December 2007
819
McElroy et al./Dispositional Factors in Internet Use
Robey, D. “Cognitive Style and DSS Design: A Comment on
Huber’s Paper,” Management Science (29:5), 1983, pp. 580-582.
Seddon, P. B. “A Respecification and Extension of the DeLone and
McLean Model of IS Success,” Information Systems Research
(8:3), 1997, pp. 240-253.
Sherer, M., Maddux, J. E., Mercandenk, B., Prentice-Dunn, S.,
Jacobs, B., and Rogers, R. W. “The Self-Efficacy Scale: Construction and Validation,” Psychological Reports (51:2), 1982,
pp. 663-671.
Spector, P. E., Jex, S. M., and Chen, P. Y. “Relations of Incumbent
Affect-Related Personality Traits with Incumbent and Objective
Measures of Characteristics of Jobs,” Journal of Organizational
Behavior (16:1), 1995, pp. 59-65.
Swickert, R. J., Hittner, J. B., Harris, J. L., and Herring, J. A.
“Relationship Between Internet Use, Personality and Social
Support,” Computers in Human Behavior (18:4), 2002, pp.
437-451.
Taylor, S., and Todd, P. A. “Understanding Information Technology Usage: A Test of Competing Models,” Information Systems
Research (6:2), 1995, pp. 144-176.
Taylor, W. A. “Computer-Mediated Knowledge Sharing and Individual User Differences: An Exploratory Study,” European
Journal of Information Systems (13:1), 2004, pp. 52-64.
Tokar, D. M., Fischer, A. R., and Subich, L. M. “Personality and
Vocational Behavior: A Selective Review of the Literature,
1993-1997,” Journal of Vocational Behavior (53:2), 1998, pp.
115-153.
Torkzadeh, G., Koufteros, X., and Pflughoeft, K. “Confirmatory
Analysis of Computer Self-Efficacy,” Structural Equation
Modeling (10:2), 2003, pp. 263-275.
Tuten, T., and Bosnjak, M. “Understanding Differences in Web
Usage: The Role of Need for Cognition and the Five Factor
Model of Personality,” Social Behavior and Personality (29:4),
2001, pp. 391-398.
Venkatesh, V. and Davis, F. D. “A Theoretical Extension of the
Technology Acceptance Model: Four Longitudinal Field
Studies,” Management Science (46:2), 2000, pp. 186-204.
Venkatesh, V., Morris, M. G., Davis, G. B., and Davis, F. D. “User
Acceptance of Information Technology: Toward a Unified
View,” MIS Quarterly (27:3), 2003, pp. 425-478.
Wheeler, P. R., Hunton, J. E., and Bryant, S. M. “Accounting
Information Systems Research Opportunities Using Personality
Type Theory and the Myers-Briggs Type Indicator,” Journal of
Information Systems (18:1), 2004, pp. 1-19.
Zmud, R. W. “Individual Differences and MIS Success: A Review
of the Empirical Literature,” Management Science, (25:10),
1979, pp. 966-97.
820
MIS Quarterly Vol. 31 No. 4/December 2007
About the Authors
James C. McElroy is a University Professor of Management at
Iowa State University. He received his Ph.D. from Oklahoma State
University in 1979. His current research focuses on the role of
individual differences, particularly job satisfaction and work commitment, on human behavior in organizations. His research has
been published in many journals, including Academy of Management Journal, Academy of Management Review, Journal of Applied
Psychology, Journal of Vocational Behavior, Journal of Marketing,
Journal of Organizational Behavior, Computers in Human
Behavior, and Journal of Educational Computing Research.
Anthony R. Hendrickson is Dean of the College of Business
Administration at Creighton University. He received his Ph.D. in
Computer Information Systems and Quantitative Analysis from the
University of Arkansas in 1994. His current research is in virtual
organizations, information system usage, information systems
psychometrics, and object-oriented technologies. He has published
in a variety of journals, including MIS Quarterly, Academy of
Management Executive, Information Systems Research, DATA
BASE, and Journal of International Business Studies, and has
presented his work at numerous conferences.
Anthony M. Townsend is an associate professor of Management
Information Systems at Iowa State University. He received his
Ph.D. in Organizational Behavior and Industrial Relations from
Virginia Polytechnic Institute and State University in 1993. His
current research examines the role of technology in organizations
and smaller workgroups. He has published in journals including
Information Systems Research, Journal of International Business
Studies, Industrial Relations, Academy of Management Executive,
and Communications of the ACM.
Samuel M. DeMarie is an associate professor of Management at
Iowa State University. He received his Ph.D., specializing in
strategic management, from Arizona State University in 1995. His
current research is in virtual organizations, the strategic implications
of emerging technologies, and competitive dynamics. He has
published in a variety of journals including Academy of Management Review, Academy of Management Executive, Organizational
Dynamics, and Communications of the ACM, and has presented his
work at numerous conferences.