Personality and Individual Differences 105 (2017) 175–178

Contents lists available at ScienceDirect

Personality and Individual Differences

journal homepage: www.elsevier.com/locate/paid

Personality differences in online and offline self-disclosure preference

among adolescents: A person-oriented approach
Wu Chen a,b,1, Xiao-Chun Xie c,1, Fan Ping d, Ming-Zhong Wang e,⁎
a
School of Marxism, Wuhan University, Wuhan 430072, China
b
Institute of Developmental and Educational Psychology, Wuhan University, Wuhan 430072, China
c
Department of Psychology, Renmin University of China, Beijing 100872, China
d
School of Marxism, Central China Normal University, Wuhan 430079, China
e
School of Educational Science, Qufu Normal University, Qufu 273165, China

a r t i c l e i n f o a b s t r a c t

Article history: This study analyzed the relationship between adolescents' personality types and their preference for online
Received 30 June 2016 or offline self-disclosure. A sample of 1644 Chinese adolescents completed offline/online self-disclosure
Received in revised form 25 September 2016 questionnaire, and personality scale (FFI-60). Latent profile analyses were conducted to get the personality
Accepted 26 September 2016
types. Results revealed that adolescents' personality could be classified into four types: Overcontrollers, resilients,
Available online xxxx
undercontrollers, and the average. Adolescents with a preference for offline self-disclosure showed a high prob-
Keywords:
ability of being resilients, and adolescents with a preference for online self-disclosure showed a high probability
Self-disclosure preference to be undercontrollers and overcontrollers. The implications and limitations are discussed.
Personality type © 2016 Published by Elsevier Ltd.
Adolescents
Person-oriented analysis

1. Introduction
Personality traits emphasizing single personality dimensions such as
extraversion or agreeableness, are variable oriented, whereas personality
types emphasizing the combination of different personality dimensions
such as high extraversion and low agreeableness, are person-oriented.
Base on Block's (1971) theory of ego resiliency and ego control,
Robins, John, Caspi, Moffitt, and Stouthamer-Loeber (1996) classified
the Big Five traits into three personality types such as resilients,
overcontrollers and undercontrollers. The resilients tend to show low
neuroticism and average or high levels of extraversion, openness to ex-
perience, agreeableness, and conscientiousness. The overcontrollers are
characterized by high neuroticism, agreeableness, and conscientious-
ness, and high extraversion. The undercontrollers generally display
low agreeableness and conscientiousness and high extraversion. Prior
research demonstrated that resilients were generally well adjusted,
whereas overcontrollers were vulnerable to internalizing symptoms
and undercontrollers displayed the most externalizing symptoms
(Van den Akker, Deković, Asscher, Shiner, & Prinzie, 2013). These
three types of personality have been established in many studies (e.g.
⁎ Corresponding author.
E-mail address: 943963988@qq.com (M.-Z. Wang).
1
These two authors contributed equally to the study. They should be regarded as joint
first authors.
Leikas & Salmela-Aro, 2014), and also have been validated consistently
across the cultures and ages (e.g. Specht, Luhmann, & Geiser, 2014).
Self-disclosure is associated with many aspects of adolescents' psy-
chosocial development (Xie, Sun, & Zhou, 2013). Offline self-disclosure
is defined as the sharing of thoughts, feelings and experiences with
close friends in face-to-face settings (Valkenburg, Sumter, & Peter,
2011). In recent years, the proliferation of computers and smart devices
has been accompanied by an increase in online self-disclosure among
adolescents all over the world. As the extension of the definition of tra-
ditional offline self-disclosure, online self-disclosure can be defined as a
variety of cyber behaviors (e.g. instant communication, post in micro-
blog) used to convey information and to maintain communication or
to satisfy social needs in cyber space (Xie et al., 2013).
Disclosure preferences are important for understanding both online
and offline social behavior. Adolescents tend to be absorbed in excessive
online communication given the three most robust adolescent behav-
ioral changes: Increased risk taking, increased sensation seeking, and a
move away from parent toward greater peer affiliation (Giedd, 2012).
Thus, updated theories and research are required to understand how
adolescents' high engagement with the virtual world (e.g. too much on-
line disclosure but lack of offline disclosure) may influence their devel-
opment (George & Odgers, 2015). However, as Koo, Woo, Yang, and
Kwon (2015) noted, many studies on social interaction in online space
have not considered the context of social interactions in offline space.
According to Internet paradox (Kraut et al., 1998), excessive use of
online communication was associated with declines in individuals'

http://dx.doi.org/10.1016/j.paid.2016.09.048
0191-8869/© 2016 Published by Elsevier Ltd.
176 W. Chen et al. / Personality and Individual Differences 105 (2017) 175–178
offline communication with family members or friends. Just as Turkle
(2011) stated, too much online interaction would interfere with indi-
viduals' (especially young people's) ability to achieve closeness in
offline relationships. Koo et al. (2015) found that online social behaviors
decreased well-being under conditions of low engagement in offline
social behaviors. In contrast, a higher level of online social behavior
predicted greater well-being among individuals with high engagement
in offline social behaviors. Adolescents who spend too much time on
online disclosure are characterized as spending time “alone together”
and missing out on important offline disclosure and socialization
experiences.
Van den Akker et al. (2013) suggested that overcontrollers were vul-
nerable to internalizing symptoms, and undercontrollers exhibited the
most externalizing symptoms. Overcontrollers and undercontrollers
were maladjusted. Cyber space provides an alternative environment
for self-disclosure, especially for maladjusted individuals (e.g. low self-
esteem, social withdrawal and anxiety) (Joinson, McKenna, Postmes, &
Reips, 2007). The properties of online space, such as the absence of
physical presence, permission for delayed feedback, and greater infor-
mation manipulation, can help maladjusted adolescents perceive online
space as a viable, safe, and efficient context to meet friends, thus en-
abling them to engage more in online than face-to-face interactions
(Joinson et al., 2007). Therefore, we hypothesize that adolescents who
prefer online self-disclosure are more likely to be undercontrollers or
overcontrollers, whereas adolescents who prefer offline self-disclosure
are more likely to be resilients.
2. Method
2.1. Participants and procedure
The present research recruited participants from 36 classes dis-
persed over 12 middle schools in southwest China. Participants were
1644 adolescents (754 were girls) whose age ranged from 10 to
16 years old (Mage = 13.21 ± 1.23 years). The investigation was con-
ducted in classrooms after informed consent was obtained from school
authorities and the participants. The study was approved by the Ethical
Committee for Scientific Research.
2.2. Measures
2.2.1. The NEO Five-Factor Inventory
Adolescents' personality was assessed by the Chinese version of the
NEO Five-Factor Inventory (NEO-FFI; Nie, Lin, Zheng, Ding, & Peng,
2008), which was adapted from the original NEO Five-Factor Inventory
(Costa & McCrae, 1992). As the most classic personality questionnaire in
China, NEO-FFI-60 has been widely used in measuring Chinese adoles-
cents' personality (e.g. Nie et al., 2008; Nie, Yang, & Zeng, 2011). NEO-
FFI includes 60 items, with 12 items for each of the five dimensions.
Participants rated the items on a five-point Likert scale (1 = completely
not true to 5 = completely true). In the present study, the Cronbach's α
coefficients for extraversion, agreeableness, conscientiousness, neuroti-
cism, and openness were 0.72, 0.64, 0.79, 0.78, and 0.60 respectively.
2.2.2. Online self-disclosure
The 10-item Self-disclosure Index questionnaire (Miller, Berg, &
Archer, 1983) was adapted to assess adolescents' online self-disclosure
to close friends (Li et al., 2006). Respondents were asked “How much do
you usually tell your close friends about your personal habits, worst
fears, secrets, etc. when you are on the Internet (online)?” Items were
rated on a five-point Likert scale (1 = never tell to others to 5 = tell
others in detail), with higher summary scores on the total scale indicat-
ing greater levels of self-disclosure. We found satisfactory reliability and
validity for the online self-disclosure to close friends (Cronbach's α =
0.91, χ2/df = 8.45, CFI = 0.97, RMSEA = 0.067).
2.2.3. Offline self-disclosure
The present study adjusted the items used to measure adolescents'
online self-disclosure (Li et al., 2006) by changing the reference to the
Internet to a reference to in-person communication (see Valkenburg
et al., 2011). We asked adolescents “How much do you usually tell
your close friends about your (personal habits, worst fears, secrets,
etc.) in face-to-face meetings (offline)?” Items were rated on a five-
point Likert scale (1 = never tell to others to 5 = tell others in detail).
We found good levels of reliability and validity for the offline self-disclo-
sure to friends (Cronbach's α = 0.93, χ2/df = 8.03, CFI = 0.98,
RMSEA = 0.065).
3. Results
3.1. Descriptive results
The first step in the analyses was to determine adolescents' self-
disclosure preference. Consistent with Valkenburg et al. (2011), we
subtracted the offline self-disclosure scores from the online self-disclo-
sure scores, creating a variable that ranged from −4.00 to 4.00. We then
calculated the standard error (SE) of this variable and multiplied this SE
(0.025) by 1.96 to create a 95% confidence interval (CI; − 0.049 to
0.049) for the middle score of the new categorical variable. The newly
formed variable had three levels: (a) the lowest score to − 0.049
(adolescents who preferred offline self-disclosure), (b) − 0.049 to
0.049 (adolescents with no self-disclosure preference), and (c) 0.049
to the highest score (adolescents who preferred online self-disclosure).
The number for adolescents who preferred offline self-disclosure, online
self-disclosure, and had no preference were 1137 (69.20%), 74 (4.50%),
433 (26.30%), respectively.
3.2. Results of person-oriented analyses
Before conducting the person-oriented analyses we obtained classi-
fications based on personality traits. According to Van den Akker et al.
(2013), the scores of each personality trait were first standardized. La-
tent profile analysis was then executed to analyze the emerging person-
ality types. We next compared the 2-class, 3-class, 4-class, and 5-class
models, and the 4-class model was chosen to be the most appropriate
model of adolescents' personality types. Table 1 showed the results of
the model comparison, and Fig. 1 showed the personality type profile.
According to the characteristics of each personality type and tenets
by Van den Akker et al. (2013), we named these four personality
types as follows: Overcontrollers (6.20%) had high scores for neuroti-
cism, low scores for extraversion, and average scores for the remaining
three traits (conscientiousness, agreeableness, and openness to experi-
ence); resilients (24.80%) had low scores for neuroticism and average
scores for the remaining four traits; the Average group (64.10%) had av-
erage scores for all the five personality traits; and undercontrollers
(4.90%) had low scores for conscientiousness and agreeableness, and
average scores for the remaining three traits (Table 2).
Multiple logistic regression was conducted to analyze the relation-
ship between self-disclosure preference and personality types. Because
the largest group of participants had average scores on all the five traits,
the average personality type was then taken as the reference group. The
Table 1
Criteria for latent profile models of personality types.
Model BIC SSBIC Entropy LMR-LRT (p-value) BLRT (p-value)
2-class 22,629.03 22,578.20 0.68 b0.001 b0.001
3-class 22,561.73 22,491.84 0.71 0.04 b0.001
4-class 22,516.57 22,427.62 0.70 0.03 b0.001
5-class 22,497.45 22,389.45 0.70 0.06 b0.001
Note. N = 1644.
 W. Chen et al. / Personality and Individual Differences 105 (2017) 175–178 177

The same pattern of results was obtained in analyses using resilients,

undercontrollers and overcontrollers as the reference groups. In sum-
mary, no matter which type was used as the reference, there was a sim-
ilar relationship between self-disclosure preference and personality
types, in that adolescents with online self-disclosure preference were
more likely to be overcontrollers or undercontrollers than adolescents
with offline self-disclosure preference and no disclosure preference.

4. Discussion

The present study analyzed the relationship between self-disclosure

preference and personality using person-oriented analyses. We found
that there were four types of personality among Chinese adolescents,
which were resilients, overcontrollers, undercontrollers, and the aver-
age. Most adolescents were in the average group and the fewest were
in the undercontroller group. Although this study did not find evidence
of exactly the same three types of personality as Robins et al. (1996).
However, the average, as an additional personality type, was also iden-
tified in some adolescent studies (e.g. De Clercq, Rettew, Althoff, & De
Bolle, 2012). In addition, Chang et al. (2011) declared that conformity
and compliance are especially encouraged by Chinese teachers. Thus,
Fig. 1. The latent profile of adolescents' personality types. N = 1644. Overcontrollers (n =
Chinese adolescents may be more likely to observe the social norm,
103): high scores for neuroticism, low scores for extraversion, and average scores for the
remaining three traits (conscientiousness, agreeableness, and openness to experience);
and most Chinese adolescents would fall in the average group. Results
Resilients (n = 407): low scores for neuroticism and average scores for the remaining of multiple logistic regression further revealed that compared to adoles-
four traits; Undercontrollers (n = 80): low scores for conscientiousness and cents with no self-disclosure preference, adolescents showing a prefer-
agreeableness and average scores for the remaining three traits; the Average group ence for offline self-disclosure were more likely to be resilients. By
(n = 1054): average scores for all five personality traits.
contrast, adolescents showing a preference for online self-disclosure
were more likely to be undercontrollers as well as overcontrollers.
Valkenburg et al. (2011) examined gender differences in disclosure
six predictor variables were the remaining three personality types, preference. To the best of our knowledge, they were the first to investi-
gender, self-disclosure preference, and the interaction between gender gate disclosure preference by comparing the prevalence of online and
and self-disclosure preference. The forward selection method was offline self-disclosure. As an extension of their study, the current study
used. Self-disclosure preference and gender were both included in the investigated individual differences in self-disclosure preference by
first step, and their interaction was included in the second step. Because using a person-oriented approach. Personality traits are organized and
the interaction was not significant, it was not included in the final integrated within individuals. The person-oriented approach provides
model. The results revealed that gender was not significant as a predic- an efficient classification system and a taxonomy for organizing people
tor of personality type classification (ps N 0.05). Self-disclosure prefer- (as opposed to traits) at a high level of abstraction.
ence was significant, such that adolescents who preferred offline self- It is important to make a distinction between adolescents with high
disclosure were more likely to be resilients, compared to adolescents level of online self-disclosure, but with different levels of offline self-dis-
with no self-disclosure preference (B = 0.58, SE = 0.15, OR = 1.79, closure. According to Caplan's (2005) proposition, some individuals
p b 0.001). Additionally, adolescents who preferred online self-disclo- were using online space to compensate for their lack of offline social in-
sure were more likely to be undercontrollers, compared to adolescents teractions, and they were more likely to experience negative outcomes
with no self-disclosure preference (B = 1.06, SE = 0.48, OR = 2.88, from such use. On the contrary, strong offline self-disclosers using on-
p = 0.027). Lastly, adolescents who preferred online self-disclosure line space as an extension of offline relationships were more likely to
were more likely to be overcontrollers, compared to adolescents experience positive outcomes. Thus, the resilient adolescents seem to
who preferred offline self-disclosure (B = 0.92, SE = 0.39, OR = 2.52, use online self-disclosure to supplement offline self-disclosure, whereas
p = 0.018). for undercontrollers and overcontrollers, online self-disclosure occurs
at the expense of offline self-disclosure. So, educators and parents
should pay more attention to adolescents who prefer online-disclosure
but lack offline disclosure, given that they were more likely to be
undercontrollers or overcontrollers.
Table 2 The present study should be interpreted in light of certain limita-
Means and standard deviations of each personality dimension for each personality type. tions. First, our study only focused on adolescent disclosure to close
n
friends. In other words, we did not take other disclosure targets (e.g.,
N E O A C (%) parents) into consideration. Future research is needed to examine
whether the current findings could generalize to other disclosure tar-
Overcontrollers 1.26 −1.22 −0.38 −0.83 −0.88 103 (6.20%)
(0.77) (0.84) (0.90) (0.85) (0.78) gets. Secondly, because all measures were self-reported, it was possible
Resilients −0.92 0.71 0.58 0.69 1.00 407 (24.80%) that method variance played a role in the findings observed. Future
(0.77) (0.84) (0.90) (0.85) (0.79) studies should simultaneously employ multiple informants and multi-
Average 0.12 −0.19 −0.28 −0.04 −0.30 1054 (64.10%)
ple methods to collect data to reduce the potential effects of shared
(0.77) (0.84) (0.90) (0.85) (0.79)
Undercontrollers 0.83 0.56 0.77 −1.15 0.02 80 method variance, and to improve internal validity. Third, it should be
(0.77) (0.84) (0.90) (0.85) (0.79) (4.90%) noted that the Cronbach's α coefficients for agreeableness and openness
Note. N = 1644. The five personality dimensions are neuroticism (N), extraversion (E),
were not high, though Cronbach's α has been criticized as
openness (O), agreeableness (A), and conscientiousness (C). Standard deviations are in underestimating true reliability (Peterson & Kim, 2013), future studies
parentheses. should avoid such issues.
178 W. Chen et al. / Personality and Individual Differences 105 (2017) 175–178

Acknowledgments Leikas, S., & Salmela-Aro, K. (2014). Personality types during transition to young adult-
hood: How are they related to life situation and well-being? Journal of Adolescence,
5, 753–762.
This work was supported by Education Youth Project of National Li, D., Yu, M., Wang, C., Xie, X., Zhou, L., & Zhu, X. (2006). Self-disclosure and self-conceal-
Social Science Foundation of P. R. China [Project name: The impact of ment in adolescence and the relationship between them and subjective well-being.
Psychol. Dev. Educ., 22(4), 83–90.
mobile social network on adolescent self-control; project number: Miller, L. C., Berg, J. H., & Archer, R. L. (1983). Openers: Individuals who elicit intimate self-
CBA150157]. No competing financial interests existed. disclosure. Journal of Personality and Social Psychology, 44, 1234–1244.
Nie, Y., Lin, C., Zheng, X., Ding, L., & Peng, Y. (2008). The relationship between adolescent
social adaptive behavior and the “big five” personality. Psychological Science, 31(4),
References 774–779.
Nie, Y., Yang, A., & Zeng, M. (2011). The relationship among adolescents' meta-cognition,
Block, J. (1971). Lives through time. Berkeley: Bancroft.
big five personality and academic adaptive behavior. Psychol. Dev. Educ., 27(2),
Caplan, S. E. (2005). A social skill account of problematic internet use. Journal of
151–157.
Communication, 55, 721–736.
Peterson, R. A., & Kim, Y. (2013). On the relationship between coefficient alpha and com-
Chang, L., Mak, M. C. K., Li, T., Wu, B. P., Chen, B. B., & Lu, H. J. (2011). Cultural adaptations
posite reliability. Journal of Applied Psychology, 98(1), 194–198.
to environmental variability: an evolutionary account of east–west differences.
Robins, R. W., John, O. P., Caspi, A., Moffitt, T. E., & Stouthamer-Loeber, M. (1996). Resil-
Educational Psychology Review, 23, 99–129.
ient, overcontrolled, and undercontrolled boys: Three replicable personality types.
Costa, P. T., Jr., & McCrae, R. R. (1992). The NEO-PI-R: Professional manual. Odessa, FL: Psy-
Journal of Personality and Social Psychology, 70, 157–171.
chological Assessment Resources.
Specht, J., Luhmann, M., & Geiser, C. (2014). On the consistency of personality types across
De Clercq, B., Rettew, D., Althoff, R. R., & De Bolle, M. (2012). Childhood personality types:
adulthood: Latent profile analyses in two large-scale panel studies. Journal of
Vulnerability and adaptation over time. Journal of Child Psychology and Psychiatry,
Personality and Social Psychology, 3, 540–556.
53(6), 716–722.
Turkle, S. (2011). Alone together: Why we expect more from technology and less from each
George, M. J., & Odgers, C. L. (2015). Seven fears and the science of how mobile technol-
other. New York, NY: Basic Books.
ogies may be influencing adolescents in the digital age. Perspectives on Psychological
Valkenburg, P. M., Sumter, S., & Peter, J. (2011). Gender differences in online and offline
Science, 10(6), 832–851.
self-disclosure in pre-adolescence and adolescence. British Journal of Developmental
Giedd, J. N. (2012). The digital revolution and adolescent brain evolution. Journal of
Psychology, 29, 253–269.
Adolescent Health, 51, 101–105.
Van den Akker, A. L., Deković, M., Asscher, J. J., Shiner, R. L., & Prinzie, P. (2013). Personal-
Joinson, A. N., McKenna, K. Y. A., Postmes, T., & Reips, U. (2007). The Oxford handbook of
ity types in childhood: Relations to latent trajectory classes of problem behavior and
internet psychology. Oxford: Oxford University Press.
over reactive parenting across the transition into adolescence. Journal of Personality
Koo, H. J., Woo, S., Yang, E., & Kwon, J. H. (2015). The double meaning of online social
and Social Psychology, 4, 750–764.
space: Three-way interactions among social anxiety, online social behavior, and
Xie, X. C., Sun, X. J., & Zhou, Z. K. (2013). The type, function and influencing factors of on-
offline social behavior. Cyberpsychology, Behavior and Social Networking, 9, 514–520.
line self-disclosure. Advances in Psychological Science, 2, 272–281.
Kraut, R., Patterson, M., Lundmark, V., Kiesler, S., Mukopadhyay, T., & Scherlis, W. (1998).
Internet paradox. A social technology that reduces social involvement and psycholog-
ical well-being? American Psychologist, 53(9), 1017–1031.