2012 45th Hawaii International Conference on System Sciences
Self Discrepancy, Perceived Privacy Rights, and Contribution in Virtual
Communities
Ayoung Suh
Ewha Womans University
[email protected]
Kyung-shik Shin
Ewha Womans University
[email protected]
In a VC, anonymity and a lack of social cues enable
the individual to create an alternate identity in contrast
to his or her actual identity. Taking on a different
identity makes it possible to spread false rumors and
groundless slander. Conversely, people want to break
free from their daily personas, and the alternate identity
fostered in a virtual space allows them to freely do as
they please, without having to conform to social norms
or anxiety about social regulations [21]. We posit that
alternate identities may have positive or negative
repercussions in VCs; however, little research exists
that examines which consequences are caused by
taking on a different identity in a VC.
Considering the distinct properties of a VC, in which
people easily change their identities at little or no cost,
it is imperative to explore how taking on a different
identity is associated with people’s contributions to
their communities. To this end, we employ the concept
of self-discrepancy, from the social psychology theory
rooted in self-identity, and develop a theoretical model
to predict the quality of the contributions of individual
members. Such examination may reveal how best to
manage or control members’ virtual identities and
anonymous statuses in order to stimulate contributions.
While many critical research questions call for an
empirical examination with regard to one’s selfidentity and contribution in an online space [5, 16, 25],
a systematic theory relating discrepancy among actual
and virtual identities, perceived privacy rights, and
contribution in a VC is lacking. Furthermore, little
research has considered whether the mechanisms
underlying one’s self-identity and contributions vary
depending on the types of VCs. In this study, we raise
the following questions:
Abstract
Virtual communities enable one to pretend to be a
different person or to possess a different identity at
little or no cost. Despite the ubiquity of such
communities, there is limited theoretical and empirical
research on how taking on a different identity is
associated with one’s contributive behavior in those
communities. Drawing on the social psychology
literature, we adopt the concept of self-discrepancy
rooted in self-identity and derive an index for selfdiscrepancy by using the differences between actual
and virtual self-identities. Next, we link the selfdiscrepancy with perceived privacy rights and with the
quality and quantity of contribution. An analysis of 299
respondents showed that self-discrepancy significantly
influenced perceived privacy rights and indirectly
reduced quality and quantity of contribution in virtual
communities. Furthermore, sub-group analysis
revealed that the effects of self-discrepancy varied
depending on whether the virtual community was
utilitarian or hedonic.
1. Introduction
A virtual community (VC) can sustain itself for as
long as it offers shared resources and its members can
benefit from those resources. A community’s resources,
whether informational or emotional, are collective
goods generated by the members’ contributions. Thus,
contribution is a key to predicting a VC's sustainability.
Previous studies have focused on understanding the
motivations underlying member contributions,
including reputation, altruism, reciprocity, community
interest, and professional self-expression [15]. We
propose a novel perspective to explain the motivation
to contribute that highlights one’s perception of
privacy rights. People exercise these rights in an
anonymous environment where they can easily hide or
alter their real identities. Thus, we posit that perceived
privacy rights play a significant role in increasing
contribution and in guiding behavior by reducing the
fear of social evaluation.
978-0-7695-4525-7/12 $26.00 © 2012 IEEE
DOI 10.1109/HICSS.2012.520
(1) How does self-discrepancy between actual-self and
virtual-self influence perceived privacy rights?
(2) How do perceived privacy rights influence
contribution, especially in terms of quality and
quantity?
(3) Do the underlying relationships among selfdiscrepancy, perceived privacy rights, and contribution
vary across different types of VCs?
3520
not correspond with those associated with a different
self-state (e.g., the virtual self).
2. Theoretical background
In this study, we use the term “self-discrepancy” to
describe instances in a VC where a person has an
alternate identity and acts like a different person,
creating a virtual self distinct from his or her actual self.
Our point of departure is the difference between the
real-life identity (“actual identity”) and the alternate
identity represented in a virtual setting (“virtual
identity”). In the next section, we will discuss the
concepts of self-identity and self-discrepancy in order
to infer the consequences of assuming an alternate
identity.
2.3 Perceived privacy rights
Unlike offline community that are based on face-toface contact, people perceive few status differences,
less social pressure, and less fear of retribution in VCs
that are based on anonymous computer-mediated
communication [7, 24], and this makes it easier for
individual members to exercise their privacy rights in
VCs. A lack of privacy may negatively affect
information sharing processes, and consequently result
in a loss of VC members. Research suggests that three
types of privacy rights can be exercised under
anonymous conditions [21]. More specifically, people
perceive possession of privacy rights in a VC when
they experience (1) recovery from social injury, (2)
catharsis by expressing their emotions freely, and (3)
autonomy by trying out new behaviors, engaging in
creative activities, or breaking social norms and
shedding their inhibitions.
2.1 Self-identity
Self-identity is “the individual’s self-appraisal of a
variety of attributes along the dimensions of physical
and cognitive abilities, personal traits and motives, and
the multiplicity of social roles” [27]. Markus (1977)
argues that self-identity is the sum total of beliefs one
has about oneself, which is directly associated with the
term “self-concept” [16]. Self-concept is made up of
cognitive molecules called self-schemas: beliefs about
oneself that guide the processing of information
relevant to the self. Researchers explain the selfconcept using hierarchical self-schemas [9, 14] that
elucidate how we see ourselves. Theorists have
hypothesized that the self includes multiple
components, with the primary distinction being
between the personal self and the social self. The
former refers to the properties that constitute an
individual’s learning abilities, while the latter refers to
the properties that build an individual’s social
relationships.
2.4 Contribution
Contribution refers to the extent to which people
may be willing to exert themselves on behalf of a
community. Almost all VCs rely on voluntary
commitment, participation, and contributions; they
need visitors to return and members to interact with
others in order to maintain the community
infrastructure, generate new and updated information,
and provide social and emotional support to other
members. Regarding member contribution in a VC,
research suggests that quantity and quality should be
considered equally [22, 26]. As both the quality and
quantity of contribution are critical for the
sustainability of a VC, neither one should outweigh the
other.
2.2 Self-discrepancy
Self-discrepancy, however, arises from the mismatch
between the ways in which we see ourselves,
especially when we perceive that we have different
selves. Self-discrepancy theory [11] explains how selfdiscrepancy can be formed and how it influences an
individual’s psychological state. Higgins (1987) argues
that people have a psychological structure that helps
them understand the different types of negative
emotions that they experienced when they hold
conflicting beliefs—a discrepancy—about themselves
[12]. By applying the notion of self-discrepancy to this
study, which attempts to capture the discrepancy
between the “actual self” and the “virtual self,” we
assume that discrepancy occurs when the attributes
associated with one’s self-state (e.g., the actual self) do
3. Research model and hypotheses
We use the concept of self-discrepancy to capture
the degree of difference between actual identity and
virtual identity. As discussed above, we categorize
self-concept into (1) personal self and (2) social self.
The former represents the perception of a person’s
learning ability (e.g., intelligence, education, expertise),
while the latter reflects a person’s social relationships
with friends, family, and others (e.g., morality,
sociability, and accordance with social norms). Next,
we link the two types of self-discrepancy with three
dimensions of perceived privacy rights—autonomy,
recovery, and catharsis—which are considered
3521
have no job in the real world; conversely, some hide
their social status or intelligence and act like novices or
non-educated people. By taking on a different identity
in a VC, one can act freely without the fear of social
norms or regulations. Such behaviors allow people to
recover from social injury and express their emotions.
This is because people usually feel that they can
protect themselves from the hurtful remarks of others
and recover from negative social experiences by
adopting a different persona in a VC. Thus, we derive
the following hypotheses:
[Hypothesis 7]: Personal self-discrepancy is positively
associated with perceived autonomy.
[Hypothesis 8]: Personal self-discrepancy is positively
associated with perceived recovery.
[Hypothesis 9]: Personal self-discrepancy is positively
associated with perceived catharsis.
determinants of contribution quality and quantity in our
model.
2GTEGKXGF
2TKXCE[4KIJVU
5GNHFKUETGRCPE[
Personal
Selfdiscrepancy
+
Autonomy
Quality of
Contribution
+
+
Social
Selfdiscrepancy
+
+
+
+
+
%QPVTKDWVKQP
Recovery
+
+
+
Catharsis
Quantity of
Contribution
+
)LJXUH5HVHDUFK0RGHO
3.1 Antecedents of contribution
On the other hand, the social self reflects a person’s
morality, sociability, and sensitivity to social norms.
By changing these aspects in a VC, one is more likely
to be honest and to reveal one’s less polished sides due
to the lack of restraint by social norms or regulations;
this online environment fosters effective discussion [6].
People try out new behaviors and engage in creative
activities in their VCs or break social norms and shed
their inhibitions. Thus, the greater the extent to which
people perceive the discrepancy between their actual
and virtual identities, the greater the perception of
privacy rights in their VCs.
[Hypothesis 10]: Social self-discrepancy is positively
associated with perceived autonomy.
[Hypothesis 11]: Social self-discrepancy is positively
associated with perceived recovery.
[Hypothesis 12]: Social self-discrepancy is positively
associated with perceived catharsis.
Contribution is driven by individual-level
calculations of costs and benefits [6]. Writing and
posting take time and effort, but they provide benefits,
since contributors can conceal personal identities, and
confidentiality is maintained through anonymity. Such
benefits may cause diverse dysfunctional behaviors,
such as humiliation, antagonism, and selfishness, but in
the context of a VC, people are more likely to give
honest answers or disclose confidential information
when they can protect their privacy. Research has
suggested that people experience recovery, catharsis,
and autonomy as part of their privacy rights [21], and
that those privacy rights enable them to focus on the
merits of their contribution—as opposed to their status
or other social cues—and thus, foster more effective
discussion and knowledge contribution [6, 23].
[Hypothesis 1]: Perceived autonomy is positively
associated with the quality of contribution.
[Hypothesis 2]: Perceived autonomy is positively
associated with the quantity of contribution.
[Hypothesis 3]: Perceived catharsis is positively
associated with the quality of contribution.
[Hypothesis 4]: Perceived catharsis is positively
associated with the quantity of contribution.
[Hypothesis 5]: Perceived recovery is positively
associated with the quality of contribution.
[Hypothesis 6]: Perceived recovery is positively
associated with the quantity of contribution.
3.3 VC type
Researchers classify VCs into different categories
according to their underlying principles or focus. For
example, Hargel and Armstrong (1997) [10] indicate
that online communities meet four types of
participants’ needs: (1) interest, (2) relationship
building, (3) transaction, and (4) fantasy. More broadly,
needs are commonly classified into two types:
utilitarian and hedonic [3]. Utilitarian needs are related
to obtaining information, whereas hedonic needs are
related to fun, enjoyment, and pleasure [20, 26].
Accordingly, we classify VCs into the two categories:
utilitarian VC and hedonic VC. The online community
of book reviewers and consumers on Amazon serves as
an example of utilitarian VC. Community members
have common interests in particular books and are
interdependent because the members who consume
3.2 Self-discrepancy and perceived privacy
rights
The personal self represents the aspects of one’s
intelligence, education, and expertise. People alter their
identities in VCs, Some pretend to be professionals or
gurus, even though they are not fully educated or they
3522
rely on reviewers to post reviews that can inform their
purchase decisions, while the members who review
rely on other members for peer recognition in the form
of helpful votes.
Game communities in the Secondlife serve as an
example of hedonic VCs, where people seek more fun
and fantasy. For example, these VCs allow users to
meet, chat, play games and meet their recreational
needs. The two types of community users have
different needs; therefore, the underlying mechanisms
that operate for user contribution in the different types
of VCs should be different. Thus, we derive the
following hypothesis:
[Hypothesis 13]: The relationships among selfdiscrepancy, perceived privacy rights, and contribution
vary along with the types of VCs.
[21]. The function factors for perceived privacy rights
are categorized into three sub-dimensions: recovery,
catharsis, and autonomy. The responses to the
questions in these categories were assigned scores from
one to five. To measure the quality of contribution,
items were adapted from Chiu et al.’s quality of
knowledge scale [3]. To measure the quantity of
contribution, we asked the respondents how often they
made postings and replied to others’ postings. The
measurement items are shown in Table 1.
4.2
Data collection
A web survey was conducted to test the proposed
model proposed above. The sample population for this
study comprised of panel members of an Internet
survey company. E-mail messages were sent to the
people selected by a stratified sampling method. They
were solicited to visit a website for the survey. To filter
the proper respondents, we first used the filtering
question, “Are you now a member of a virtual
community?” If a respondent answered “no,” then the
survey did not proceed. We asked the respondents to
select one VC in which they were engaged as an active
member. Next, we asked questions to measure
perceived self-identity, perceived privacy rights, and
behaviors associated with contribution in their VCs. To
test our hypotheses, we used 299 questionnaires from
378 respondents who started to answer the survey
According to the results of the survey, 171 respondents
answered that their VCs are related to games and
socializing; thus, we categorized them as hedonic VC
users. The remaining 128 respondents are members of
VCs related to information sharing or business-related
VCs; thus, we categorized them as utilitarian VC users.
Table 2 provides an overview of the respondents’
characteristics. The results of independent sample t-test
indicate that there are no significant differences
attributed to gender, age, education, duration of
membership, and usage hours per day between the two
types of virtual communities (VCs).
4. Method
To test the proposed research model, we adopted the
cross-sectional survey method for data collection and
examined our hypotheses by applying the partial least
squares (PLS) method to the collected data. We chose
PLS from several other tools (e.g., AMOS, LISREL)
because it is more suitable when the objective is causal
predictive testing in situations of low theoretical
information, as opposed to testing an entire theory, and
because it is appropriate for the early stages of theory
development [1, 2]. Given that this study is an early
attempt to develop a theoretical model that predicts the
influence of multi-identity on one’s psychological state
and contribution behavior, PLS is appropriate. The
individual is the unit of analysis for this study.
4.1 Measurement
Based on the theoretical framework presented above,
new measures of self-discrepancy and perceived
privacy rights were developed. To measure selfdiscrepancy, we adapted and modified items from
Marsh et al.’s Tennessee Self-Concept Scale (TSCS)
and Marsh et al.’s Self-Description Questionnaire
(SDQ) [17, 18]. Based on these two methods, we
derived several scales to measure specific aspects
regarding self-concept. First, using seven items to
measure self-concept, we asked the respondents to
answer the question, “Who are you in the real world?”
Second, we asked, “Who are you in your VC?” To
capture the discrepancy between actual-self and
virtual-self, we derived a numerical index representing
self-discrepancy by calculating the difference between
actual-self and virtual-self.
For the measurement of perceived privacy rights, we
used the privacy function factors proposed by Pedersen
5. Results
5.1 Measurement model
Internal consistency was examined using the
composite scale reliability index developed by Fornell
and Larcker [8], which is a measure similar to
Cronbach’s alpha. All reliability measures were 0.8 or
higher, which is well above the recommended level of
0.7, indicating adequate internal consistency.
Discriminant validity was assessed by comparing the
correlation between the two constructs and the
3523
items DIS 3, 4, 5 are related to a person’s personal
learning ability (e.g., intelligence, education, expertise).
On the other hand, DIS 1,2,6 refer to a person’s social
aspects (e.g., morality, sociability, and social norms).
DIS 7 (social status) was not included in our final
analysis because of its low loading value. We also
eliminated one item from the catharsis scale (CAT3)
because of low loading value. Table 3 shows the final
items and loadings analyzed by factor analysis.
respective Average Variance Extracted (AVE). For
each construct, the square root of the average variance
extracted should exceed the construct’s correlation
with every other construct. This condition of
discriminant validity is upheld in our study, as shown
in Table 3.
As shown in Table 4, we identified two different
sub-dimensions
of
self-discrepancy.
This
categorization is consistent with the theory of selfconcept rooted in social psychology. The measurement
7DEOH0HDVXUHPHQW,WHPV
Constructs Reference(s)
Actual self-identity 1
adapted and modified items
from the Tennessee
Self-Concept Scale (TSCS)
and Self-Description
Questionnaire (SDQ)
(Marsh et al., 1983; 1985)
Items
In the real world, who are you?
1. I am satisfied with my moral behavior (Morality)
2. I make friends easily (Sociability)
3. I am smart and intelligent (Intelligence)
4. I have attained a high level of education (Education)
5. I am an expert in special skills and knowledge (Expertise)
6. I usually act in accordance with social norms (Social norms)
7. I have a high social position (Social Status)
1: Completely false; 2: Mostly false; 3: Partly false and partly true; 4: Mostly true 5: Completely true
Virtual self-identity
In your virtual community, who do you think you are?
1. I am satisfied with my moral behavior (Morality)
2. I make friends easily (Sociability)
3. I am smart and intelligent (Intelligence)
4. I have received a high level of education (Education)
5. I am an expert with special skills and knowledge (Expertise)
6. I tend to act in accordance with social norms (Social norms)
7. I have a high social position (Social Status)
1: Completely false; 2: Mostly false; 3: Partly false and partly true; 4: Mostly true 5: Completely true
Recovery
(Pedersen, 1997)
Catharsis
(Pedersen, 1997)
Autonomy
(Pedersen, 1997)
Quality of contribution
(Chiu et al., 2007)
1. I feel that I can protect myself from what others say in my community
2. I feel that I can recover from bad social experiences in my community
3. I feel that I can take refuge from the outside world in my community
4. I feel that I can build up high self-esteem in my community
1. I feel that I can express my emotions freely in my community
2. I express my anger when I am feeling angry or furious
3. I feel that I can determine what I want to be in my community (*)
1. I feel that I can attempt new behaviors in my community
2. I feel that I can engage in creative activities in my community
3. I feel that I can break some social norms
4. I feel that I can shed my inhibitions
1. The knowledge that I post in my community is reliable
2. The knowledge that I post in my community is relevant to the topics
3. I contribute to the development of my community
1: Never; 2: Rarely; 3: Occasionally; 4: Often; 5: Usually
Quantity of contribution
developed based on Wasko
and Faraj (2005)
1. On average, how many writings and commentaries do you post in your community per
month?
2. On average, how many replies do you post in your community?
1: Never; 2: Rarely; 3: Occasionally; 4: Often; 5: Very Often
VC Type
adapted from Lee et al.
(2007)
Considering the characteristics and purpose, which statement best describes your VC?
1. The VC is mainly related to games, socializing, or fun
2. The VC is mainly related to information seeking, knowledge sharing, or is business
related.
1
To measure the discrepancy between actual-self and virtual-self, we adapted and modified items from the Tennessee Self-Concept Scale
(TSCS) and Self-Description Questionnaire (SDQ). In the final analysis, we derived the numerical index representing self-discrepancy by
calculating the difference between actual-self and virtual-self.
All items are answered with a 5-point Likert Scale. The scale used for each item measures “Strongly Disagree” to “Strongly Agree” unless
otherwise indicated in the above table.
(*): deleted item in the final analysis 3524
7DEOH5HVXOWVRI7HVWLQJ'LVFULPLQDQW9DOLGLW\8VLQJ$9(
Total Sample
Item
Gender
Age
Education Level
Duration
of
Membership
Usage of VC
per day
Utilitarian VC
Hedonic VC
T-test between
Two Groups
Category
Frequency
Ratio
Frequency
Ratio
Frequency
Ratio
158
141
299
12
75
80
80
52
299
3
9
64
191
32
299
11
39
41
208
299
31
116
75
29
20
28
299
53%
47%
100%
4%
25%
27%
27%
17%
100%
1%
3%
21%
64%
11%
100%
4%
13%
14%
70%
100%
10%
39%
25%
10%
7%
9%
100%
54
74
128
6
42
37
24
19
128
2
3
32
81
10
128
6
17
20
85
128
14
54
27
12
7
14
128
61%
39%
100%
5%
33%
29%
19%
15%
100%
2%
2%
25%
63%
8%
100%
5%
13%
16%
66%
100%
11%
42%
21%
9%
5%
11%
100%
104
67
171
6
33
43
56
33
171
1
6
32
110
22
171
5
22
21
123
171
17
62
48
17
13
14
171
42%
58%
100%
4%
19%
25%
33%
19%
100%
1%
4%
19%
64%
13%
100%
3%
13%
12%
72%
100%
10%
36%
28%
10%
8%
8%
100%
1. Male
2. Female
Total
1. <20
2. 20-29
3. 30-39
4. 40-49
5. >50
Total
1. Primary
2. Secondary
3. College (2 year)
4. University (4 year)
5. Graduate School
Total
1. <1 month
2. 1-5 months
3. 6-12 months
4. >12 months
Total
1. <10 minutes
2. 10-30 minutes
3. 0.5-1 hour
4. 1-1.5 hours
5. 1.5-2 hours
6. >2 hours
Total
t
p
-.513
.608
-.299
.765
-.460
.646
-1.815
.071
.852
.395
7DEOH5HVXOWVRI7HVWLQJ'LVFULPLQDQW9DOLGLW\8VLQJ$9(
Constructs
Correlations
4
5
Mean
S.D.
C.R
AVE
Personal Self-discrepancy
Social Self-discrepancy
Total Recovery
Sample Autonomy
(n=299) Catharsis
Quality of Contribution
Quantity of Contribution
0.365
0.495
2.978
3.215
3.513
3.196
3.485
0.693
0.757
0.757
0.703
0.670
1.230
0.600
0.825
0.778
0.908
0.856
0.869
0.906
0.942
0.767
0.717
0.703
0.764
0.862
0.803
0.764
0.876
0.494
-0.297
-0.214
0.084
0.034
-0.182
0.847
-0.341
-0.25
0.066
-0.123
-0.137
0.839
0.561
-0.338
0.357
0.327
0.874
-0.466
0.372
0.345
0.928
-0.41
-0.221
0.896
0.293 0.874
Personal Self-discrepancy
Social Self-discrepancy
Utilitarian Recovery
VC
Autonomy
(n=128) Catharsis
Quality of Contribution
Quantity of Contribution
0.316
0.447
3.068
3.309
3.493
3.466
3.249
0.699
0.680
0.726
0.596
0.648
0.597
1.202
0.703
0.942
0.825
0.873
0.927
0.920
0.915
0.922
0.764
0.767
0.800
0.875
0.949
0.798
0.960
0.389
-0.369
-0.23
0.07
0.117
-0.098
0.874
-0.494
-0.27
0.063
-0.082
-0.181
0.876
0.584
-0.318
0.378
0.406
0.895
-0.501
0.42
0.256
0.935
-0.436
-0.205
0.928
0.378 0.893
Personal Self-discrepancy
Social Self-discrepancy
Hedonic Recovery
VC
Autonomy
(n=171) Catharsis
Quality of Contribution
Quantity of Contribution
0.481
0.610
2.764
3.099
3.563
3.530
3.068
0.668
0.778
0.791
0.708
0.722
0.608
1.294
0.942
0.890
0.903
0.852
0.890
0.873
0.905
0.764
0.823
0.839
0.810
0.815
0.800
0.784
0.874
0.528
-0.276
-0.219
0.132
-0.042
-0.233
0.907
-0.268
-0.234
0.076
-0.144
-0.11
0.916
0.544
-0.368
0.355
0.28
0.900
-0.435
0.359
0.379
0.903
-0.4
-0.239
0.895
0.257 0.885
1
2
3
6
7
* self-discrepancy = the value of actual-self minus the value of virtual-self, the minimum value of self-discrepancy is -2 and maximum value is
3.33
3525
7DEOH7KH5HVXOWRI)DFWRU$QDO\VLV
Recovery
REC1
REC2
REC3
REC4
AUT3
AUT4
AUT2
AUT1
QUL1
QUL2
QUL3
DIS2
DIS6
DIS1
DIS5
DIS3
DIS4
SQUN1
SQUN2
CAT1
CAT2
5.2
Autonomy
0.830
0.772
0.750
0.721
0.142
0.261
0.176
0.356
0.039
0.034
0.441
-0.187
-0.136
-0.032
-0.092
-0.084
-0.148
0.157
0.112
0.023
0.280
0.080
0.227
0.287
0.272
0.837
0.742
0.701
0.623
0.070
0.142
0.088
-0.084
-0.145
0.001
-0.071
-0.078
-0.029
0.162
0.125
-0.040
0.190
Quality of
Contribution
Social Self- Personal Self- Quantity of
discrepancy discrepancy Contribution
0.064
-0.065
0.090
0.260
0.213
-0.020
0.413
-0.108
0.863
0.861
0.779
-0.056
-0.029
0.017
0.058
0.061
-0.017
0.065
0.168
0.026
-0.024
-0.104
-0.144
-0.117
-0.065
-0.043
-0.116
0.048
-0.215
0.025
-0.016
-0.162
0.822
0.757
0.730
0.036
0.317
0.363
-0.060
-0.029
0.001
-0.081
-0.096
-0.224
-0.028
-0.095
-0.019
-0.104
-0.008
-0.128
-0.021
0.058
0.163
0.094
0.214
0.224
0.790
0.757
0.641
-0.032
-0.132
0.064
-0.023
0.072
-0.012
0.161
0.142
0.137
0.073
0.081
0.128
0.077
0.055
0.193
-0.086
0.107
-0.107
-0.131
-0.002
-0.025
0.875
0.872
0.042
0.093
Catharsis
0.006
0.144
0.096
0.079
-0.016
0.215
0.064
0.076
-0.076
0.024
-0.014
-0.003
-0.067
-0.013
-0.020
-0.015
-0.085
0.058
-0.022
0.897
0.787
Test of Structural Model
Figure 2 shows the results of the test of the
hypothesized structural model for the total sample
(n=299). The results of the PLS suggest that three
dimensions of Perceived Privacy Rights we suggested
in this study have significant effects on Quality and
Quantity of Contribution. Autonomy and Recovery
positively influence Quality and Quantity of
Contribution, supporting H1, H2, H, 3 and H4.
Contrary to our expectation, Catharsis negatively
influences Quality of Contribution and has no
significant effect on Quantity of Contribution, rejecting
H5 and H6. On the other hand, the results show that
both Personal Self-discrepancy and Social Selfdiscrepancy negatively influence Autonomy and
Recovery while they do not have any significant
influence on Catharsis.
In this study, the VC type was used as a global
moderator. To increase understanding of the
moderating role that the VC type might play, the base
model for subgroups of the sample (i.e., utilitarian vs.
hedonic VC) were analyzed. Figure 3 shows the results
of the test of the hypothesized structural model for
utilitarian VC users (n=128) and hedonic VC users
(n=171). In utilitarian VC, the results of analysis show
that Personal Self-discrepancy and Social Selfdiscrepancy negatively influence Recovery, which
positively increases Quantity of Contribution.
Catharsis negatively influences Quality of Contribution.
By contrast, in hedonic VCs, the results show that
Recovery increases Quality of Contribution while
Autonomy increases Quantity of Contribution.
Consistent with utilitarian VCs, Catharsis negatively
influences Quality of Contribution. On the other hand,
Personal Self-discrepancy negatively influences
Recovery, while Social Self-discrepancy negatively
influences Autonomy and Recovery. However, it is
notable that these relationships are weak (p<0.1).
)LJXUH6XPPDU\RI3/6$QDO\VLV7RWDO6DPSOH
3526
7VKNKVCTKCP8%
Personal
Selfdiscrepancy
Autonomy
R2 = .091
Recovery
R2 = .281
Social
Selfdiscrepancy
Quality of
Contribution
R2 = .268
Catharsis
R2 = .006
Quantity of
Contribution
R2 = 172
*GFQPKE8%
Personal
Selfdiscrepancy
Autonomy
R2 = .068
Recovery
R2 = .097
Social
Selfdiscrepancy
Quality of
Contribution
R2 = .223
* p<.1; **p<.05; ***p<.01
Catharsis
R2 = .017
Significant;
Quantity of
Contribution
R2 = 156
insignificant
)LJXUH6XPPDU\RI6XEJURXS$QDO\VLV
6. Discussion, implications and limitations
Relatively little attention has been paid to the role of
self-identity in a VC. The objective of this study was to
add to the understanding of self-identity and the
influences of identity discrepancy between actual-self
and virtual-self on one’s perceived privacy rights and
contributive behavior in a VC. Accordingly, we set out
to develop a theory-driven multi-dimensional measure
of self-discrepancy and perceived privacy rights that
included the full breath of the construct. The results of
the empirical study reveal that the degree of selfdiscrepancy between actual-self and virtual-self
negatively influences perceived privacy rights. The
major findings of this research are follows. First,
among the three dimensions of perceived privacy rights,
we found that autonomy and recovery enhanced not
only the quality of contribution but also the quantity of
contribution. The result implies that the more people
feel that they can recover from social injury and can try
out new behaviors, engage in creative activities, or
break social norms and shed their inhibitions in their
VCs, the more they will contribute themselves to their
VCs in terms of quality and quantity. However, it is
surprising that catharsis reduces the quality of
contribution. This result implies that the more people
express their emotions freely, the less relevant and
accurate information or knowledge they contribute to
their VCs.
Second, social self-discrepancy, which represents
the difference between actual-self and virtual-self in
terms of morality, sociability, and social norms, exerts
a significant negative influence on recovery and
autonomy in the total sample, while it does not have a
3527
significant influence on autonomy in utilitarian VCs.
Conventional wisdom tells us that people perceive a
higher level of privacy rights in VCs when engaging in
social behaviors without any constraints (e.g., social
norms, moral pressure, and regulations). Surprisingly,
the results of this study indicate that the participants
did not perceive a high level of privacy rights
reflecting catharsis, recovery, or autonomy, even
though they created different selves and engaged in
socially undesirable behaviors in a VC. The results
imply that people who pretend to be different from
their actual selves by engaging in socially undesirable
behaviors under their alternative identities are more
likely to suffer lower levels of psychological wellbeing
and thus experience lower levels of perceived privacy
rights (especially recovery) than others. Based on the
results of this analysis, we can explain the dysfunctions
of social self-discrepancy in VCs.
Third, personal self-discrepancy significantly
decreases the levels of perceived recovery and
autonomy in total sample. The negative relationship
between personal self-discrepancy and autonomy is
disappeared in hedonic VCs. This result implies that
those who pretend to be a different people in terms of
intelligence, education, or expertise also indirectly
debase the quality and quantity of their contribution by
decreasing perceived recovery.
These findings should be interpreted in light of the
study’s limitations. The first limitation is related to the
self-reported, perception-based measure that was
employed for capturing actual and virtual identities.
Future research may pursue an alternative route to the
conceptualization and operationalization of multiidentity. Second, it should be mentioned that in this
study, the quality of contribution is conceptualized as a
latent construct with reflective indicators. Although our
view of the quality of contribution appears reasonable
and consistent with past IS research, the results of our
study should be compared carefully with those based
on objective data [e.g., 20]. Finally, we did not include
the aspect of self-attributes such as confidence, belief,
and trust. We suggest that future research consider
those aspects to examine the effects of multi-identity
on a VC. Such aspects may be related to self-esteem,
which has been suggested as a critical aspect of the self.
An examination of the relationship between multiidentity and self-esteem would be a worthwhile
research endeavor.
This study holds several implications for the
academic world. First, the concept of self-discrepancy
studied in offline environments was applied to an
online setting. Furthermore, its impact on one’s
perceived privacy rights which represent one’s
psychological state in a VC, was empirically examined.
Although many researchers have raised issues related
to identity in virtual spaces [14], little research has
quantified the concept and investigated it using a large
sample.
Second, we extend the self-concept to the virtual self
by applying the conventional self-discrepancy theory
to the VC context. The results of this study empirically
show that self-discrepancy between actual and virtual
identities significantly reduces autonomy and recovery
as part of perceived privacy rights, which are critical
stimulators of the quality and quantity of contribution.
This understanding of the dysfunctions of selfdiscrepancy in a VC can potentially shed light on
collaboration among virtual teams or communities of
practice.
Third, this research contributes to the development
of VC research by corroborating the fact that the three
dimensions of perceived privacy rights—recovery,
catharsis, and autonomy—work in tandem to influence
the quality and quantity of contribution. Even though
several scholars have argued the importance of
individual members’ psychological states and
motivations to contribute, there is a lack of empirical
evidence to support this. Our findings on the important
effects of self-discrepancy on perceived privacy rights
and contribution in VCs open up rich and exciting
opportunities for theoretical extensions of the present
model and practical development of new VC features.
The fundamental motivation for an individual to create
different identities in online settings merits further
investigation.
From the pragmatic perspective, business
organizations are trying to generate value by
extensively investing their resources in developing the
infrastructure for VCs. For example, organizations rely
3528
on customer-based communities to increase loyalty and
to allow the customers to participate in product
development. In some cases, organizations are eager to
facilitate knowledge sharing among employees and
foster new ideas and innovations [4].
The results of this study offer several key
implications for practice. First, our findings offer
insight into the facilitators and inhibitors of
contribution, showing that each sub-dimension of
perceived privacy rights has a different effect on the
quantity and quality of contribution. In particular, it is
notable that catharsis has a negative influence on the
quality of contribution. In this regard, we suggest that
VC moderators should monitor whether members let
their feelings dissipate without any controls, while
facilitating members to perceive higher levels of
autonomy and recover.
Second, the results point to the self-discrepancy
between actual-self and virtual-self as a significant
factor that has implications for individuals’
contribution in VCs. The ability to assume a different
identity is one of the most common features of VCs,
and the important pragmatic guidelines regarding this
matter remain unclear. We suggest that community
design supporting effective identity verification and the
control of identity change will lead to the success of
VCs.
Finally, we point out the implications of our results
for VC managers. This research shows that the self
concept consists of two sub-aspects: personal and
social selves. Currently, many VCs have a
compensation mechanism for personal aspects, such as
expertise or skill. However, compensation for the
ratings in terms of social self has not been emphasized.
Thus, we propose that a device that can assess an
individual’s morality, sociability, social norms and
compensation mechanisms for a positive social self
must be developed. Through such a device, individuals
can form a positive virtual self and ultimately
contribute to improving the quality and quantity of
contribution.
7. Conclusion
Our knowledge of multi-identity in VCs is severely
limited compared with what we know about the
dynamics of VCs. This study presents a conceptual
framework that highlights the concept of selfdiscrepancy between actual and virtual identities,
assuming that they may influence individuals’
psychological states and contribution quality in VCs.
Our findings help resolve the conflicting views on the
effects of multi-identity in a VC. Furthermore, the
results suggest that VC managers should pay more
Social Psychology (Vol. 17, pp. 1-47), New York: Academic
Press, 1984.
[15] Ma, M., and Agarwal, R. “Through a Glass Darkely:
Information Technology Design, Identity Verification, and
Knowledge Contribution in Online Communities,”
Information Systems Research, Vol. 18, No. 1, 2007, pp. 4267.
[16] Markus, H. “Self-schemata and Processing Information
about the Self,” Journal of Personality and Psychology, Vol.
17, 1977, pp. 50-71.
[17] Marsh, H. W., Smith, I. D., and Barnes, J. “Multitraitmultimethod Analyses of the Self-Description Questionnaire:
Student-teacher Agreement on Multidimensional Ratings of
Student Self-concept,” American Education Research Journal,
Vol. 20, 1983, pp. 333-357.
[18] Marsh, H. W., Parker, J., and Barnes, J.
“Multidimensional
Adolescent
Self-concept:
Their
Relationship to Age, Sex, and Academic Measures,”
American Education Research Journal, Vol. 22, 1985, pp.
422-444.
[19] McMillan, D. W., and Chavis, D. M. “Sense of
Community: A definition and Theory,” Journal of
Community Psychology, Vol. 14, No. 1, 1986, pp. 6-23.
Oppenheim, A. N. Questionnaire Design and Attitude
Measurement, Heinemann, London, UK.
[20] O’Brien, H.L. “The Influence of Hedonic and Utilitarian
Motivations on User Engagement: The Case of Online
Shopping Experiences,” Interacting with Computers: Special
Issue on User Experience, Vol. 22, No. 4, 2010, pp. 344-352.
[21] Pedersen, D. M. “Psychological Functions of Privacy,”
Journal of Environmental Psychology, Vol. 17, 1997, pp.
147–156.
[22] Peddibhotla, N. B., and Subramani, M. R. “Contributing
to Public Document Repositories: A Critical Mass Theory
Perspective,” Organization Studies, Vol. 28, No. 3, 2009, pp.
327-346.
[23] Podsakoff, P. M., and Organ, D. W. “Self Reports in
Organizational Research: Problems and Prospects,” Journal
of Management, Vol. 12, No. 4, pp. 531-544.
[24] Postmes, T. and Lea, M. “Social Pressures and Group
Decision Making: Anonymity in Group Decision Support
Systems,” Ergonomics, Vol. 43, 2000, pp. 1252-1274.
[25] Rains, S. A. “The Impact of Anonymity on Perceptions
of Source Credibility and Influence in Computer-Mediated
Group Communication,” Communication Research, Vol. 34,
No. 1, 2007, pp. 100-125.
[26] Venkatesh, V. and Brown, S. (2001) A Longitudinal
Investigation of Personal Computers in Homes:
Adoption Determinants and Emerging Challenges. MIS
Quarterly, 25 (1), 71-102.
[27] Wasko, M. M., and Faraj, S. “Why Should I Share?
Examining Social Capital and Knowledge Contribution in
Electronic Networks of Practice,” MIS Quarterly, Vol. 29,
No. 1, 2005, pp. 35-57.
[28] Whitbourne, S. K., Connolly, A. The Developing Self in
Middle: Psychological and Social Development in Middle
Age. Academic Press, New York, 1999.
attention to the negative influences exerted by multiidentity on the quality of contribution, thereby
controlling the need to create alternative identities in
VCs. We hope that more research will be conducted on
this underexplored area of multi-identity and that our
theoretical framework will serve as a useful conceptual
tool for such endeavors.
8. References
[1] Barclay, D., Higgins, C., and Thompson, R. “The Partial
Least Squares (PLS) Approach to Causal Modeling, Personal
Computer Adoption and Use as an Illustration,” Technology
Studies, Vol. 2, No. 2, 1995, pp. 285-309.
[2] Chin, W. W. “Issues and Opinion on Structural Equation
Modeling,” MIS Quarterly, Vol. 22, No. 1, 1998, pp. VIIXVI.
[3] Chiu, C. M., Hsu, M. H., and Wang, T. G.
“Understanding Knowledge Sharing in Virtual Communities:
An Integration of Social Capital and Social Cognitive
Theories,” Decision Support Systems, Vol. 42, 2006, pp.
1872-1888.
[4] Constant, D. L., Sproull, L., and Kiesler, S. “The
Kindness of Strangers: The Usefulness of Electronic Weak
Ties for Technical Advice,” Organization Science, Vol. 7, No.
2, 1996, pp. 119-135.
[5] Dellarocas, C. “The Digitization of Word of Mouth:
Premise and Challenges of Online Feedback Mechanisms,”
Management Science, Vol. 49, No. 10, 2003, pp. 1407-1424.
[6] DeSanctis, G., and Gallupe, R. B. “A foundation for the
study of group decision support systems,” Management
Science, Vol. 33, 1987, pp. 589-609.
[7] Flanagin, A. J., Tiyaamoronwong, V., O’connor, J., and
Seibold, D. R. “Computer-mediated Group Work: The
Interaction of Member Sex and Anonymity,” Communication
Research, Vol. 29, 2002, pp. 66-93.
[8] Fornell, C., and Larcker, D.F. “Evaluating Structural
Equation Models with Unobservable Variables and
Measurement Error,” Journal of Marketing Research, Vol. 18,
No. 1, 1981, 39-50.
[9] Franzio, S. L. Social Psychology, Madison: Brown and
Bench Mark, 1996.
[10] Hagel J., and Armstrong, A.G. Net gain: expanding
market through virtual communities, Harvard Business
School Press, Boston, 1997.
[11] Higgins, E. T. “Self-discrepancy: A theory relating self
and affect,” Psychological Review, Vol. 94, 1987, pp. 319340.
[12] Higgins, E. T. Self-discrepancy Theory: What patterns
of Self-Beliefs Cause People to Suffer? In L. Berkowitz (Ed.),
Advances in Experimental Social Psychology, (Vol. 22, pp.
93-136). New York: Academic Press, 1987.
[13] Higgins, E. T. Self-Discrepancy: A Theory Relating Self
and Affect, In R. F. Baumeister (Ed.), The Self in Social
Psychology (pp. 150-181). Philadelphia, PA: Psychology
Press, 1999.
[14] Kihlstorm, J. F., and Cantor, N. Mental Representations
of the Self, In L. Berkowiz (Ed.), Advances in Experimental
3529