ArticlePDF Available

Abstract and Figures

Nowadays, millions of people around the world use social networking sites to express everyday thoughts and feelings. Many researchers have tried to make use of social media to study users' online behaviors and psychological states. However, previous studies show mixed results about whether self-generated contents on Facebook reflect users' subjective well-being (SWB). This study analyzed Facebook status updates to determine the extent to which users' emotional expression predicted their SWB-specifically their self-reported satisfaction with life. It was found that positive emotional expressions on Facebook did not correlate with life satisfaction, whereas negative emotional expressions within the past 9-10 months (but not beyond) were significantly related to life satisfaction. These findings suggest that both the type of emotional expressions and the time frame of status updates determine whether emotional expressions in Facebook status updates can effectively reflect users' SWB. The findings shed light on the characteristics of online social media and improve the understanding of how user-generated contents reflect users' psychological states.
Content may be subject to copyright.
Do Facebook Status Updates Reflect Subjective Well-Being?
Pan Liu, PhD,
1
William Tov, PhD,
2
Michal Kosinski, PhD,
3
David J. Stillwell, PhD,
4
and Lin Qiu, PhD
5
Abstract
Nowadays, millions of people around the world use social networking sites to express everyday thoughts and
feelings. Many researchers have tried to make use of social media to study users’ online behaviors and
psychological states. However, previous studies show mixed results about whether self-generated contents on
Facebook reflect users’ subjective well-being (SWB). This study analyzed Facebook status updates to determine
the extent to which users’ emotional expression predicted their SWB—specifically their self-reported satis-
faction with life. It was found that positive emotional expressions on Facebook did not correlate with life
satisfaction, whereas negative emotional expressions within the past 9–10 months (but not beyond) were
significantly related to life satisfaction. These findings suggest that both the type of emotional expressions and
the time frame of status updates determine whether emotional expressions in Facebook status updates can
effectively reflect users’ SWB. The findings shed light on the characteristics of online social media and improve
the understanding of how user-generated contents reflect users’ psychological states.
Introduction
Facebook is one of the most widely used online social
networking sites (SNS). Users frequently express and
share emotional experiences through their status updates.
1,2
With around 300,000 status updates published every minute,
3
Facebook provides a huge and natural record of users’ ev-
eryday emotional experiences.
4
Given that daily experiences
are an important predictor of subjective well-being (SWB),
5
it is likely that contents in Facebook status updates are re-
lated to users’ SWB. Judgments of SWB indicate how
people evaluate their quality of life, and SWB is an im-
portant predictor of many important aspects of life including
the probability of getting married and staying in marriage,
the likelihood of earning high income, and the tendency of
having a healthy lifestyle.
6–8
Understanding how Facebook
status updates reflect SWB can help the development of tools
to assess emotional states and quality of life automatically
on an unprecedented scale. Because status updates offer a
snapshot of users’ everyday lives from one period of time to
another, they provide real-time information about users’
mental states and offer the potential for studying changes in
SWB in broad segments of the population, without intrusive
survey methods.
Past research studying the relationship between Facebook
status updates and SWB has found mixed results. Kramer
9
proposed an index of Gross National Happiness (GNH) by
calculating the difference between the percentage of positive
and negative emotion words in millions of status updates
posted by users in a given country. GNH peaked on national
and cultural holidays such as Christmas and Thanksgiving,
and dipped on days of national tragedies such as the death of
Michael Jackson. It also followed a weekly cycle with peaks
on Friday and dips on Monday. This provided evidence of
face validity and suggested that the positive and negative
emotion word use in status updates was associated with
SWB. However, Wang et al.
10
challenged the validity of
Facebook GNH by comparing it with self-reported SWB
judgment of Facebook users. They showed that GNH did not
correlate with SWB scores aggregated by day and week, and
had a negative correlation with SWB scores aggregated by
month. Wang et al.’s study
10
suggested that Facebook GNH
does not accurately reflect SWB. However, their sample size
of 34 users on average every day is too small for a reliable
comparison.
Therefore, the present study aimed to examine the rela-
tionship between emotional expressions in Facebook status
updates and SWB further. Past research has shown that SWB
fluctuates with everyday positive and negative experi-
ences.
11,12
High levels of SWB are characterized by high
satisfaction with life, frequent experience of positive emo-
tions, and infrequent experience of negative emotions.
13
In
1
Behavioural Sciences Institute, Singapore Management University, Singapore, Singapore.
2
School of Social Sciences, Singapore Management University, Singapore, Singapore.
3
The Psychometrics Centre, Department of Psychology, University of Cambridge, Cambridge, United Kingdom.
4
Department of Computer Science, Stanford University, Stanford, California.
5
Division of Psychology, Nanyang Technological University, Singapore, Singapore.
CYBERPSYCHOLOGY,BEHAVIOR,AND SOCIAL NETWORKING
Volume 18, Number 7, 2015
ªMary Ann Liebert, Inc.
DOI: 10.1089/cyber.2015.0022
373
addition, Rutledge et al.
14
showed that emotional reactivity
to recent events predicted SWB based on evidence from a
computational model and functional fMRI. Suh et al.
15
asked
participants to report positive and negative events they ex-
perienced over the past 4 years, and found that life satis-
faction only correlated with events in the past 3 months. The
above findings suggest that only recent events are related to
SWB. Therefore, it was hypothesized that only recent
emotional expressions in Facebook status updates would be
associated with SWB.
Meanwhile, past research has shown that users often en-
gaged in impression management when they use Facebook.
16,17
Due to self-representational concerns, they selectively disclose
more positive than negative emotions to present a positive self-
image.
16,18
This desire to make a positive impression on others
may reduce individual differences in the expression of positive
emotion on Facebook, weakening the association between the
latter and self-reported SWB. In contrast, because there is rel-
atively less social pressure to express negative emotion (vs.
positive emotion), such expressions are more likely to reflect
how a person actually feels. Therefore, it was predicted that the
amount of negative (but not positive) emotional expressions
in Facebook status updates would be related to self-reported
SWB.
Many studies have used the Linguistic Inquiry and Word
Count (LIWC) text analysis software
19
to examine emotional
expressions in social media. LIWC counts words in pre-
defined categories that have been developed based on psy-
chological measurement scales and validated by independent
judges.
20
It has been widely used and proven reliable to
measure psychological attributes from writing samples, in-
cluding emotion, personality, thinking styles, and social re-
lationships.
21–23
A recent study shows that LIWC coding of
emotion in diary entries consistently correlated with self-
reported emotional experiences.
24
Chee et al.
25
used LIWC
to examine postings from illness groups in Yahoo! Groups
and revealed changes in sentiment after FDA approval of
certain drugs. Yu et al.
26
assessed overall sentiment in con-
gressional speeches to classify political party affiliation.
Golder and Macy
27
identified diurnal and seasonal mood
patterns in cultures across the globe from millions of tweets.
Qiu et al.
28
showed that extraverts expressed more positive
emotions in tweets than introverts. These studies suggest that
LIWC is a reliable tool for measuring emotional expressions,
and therefore it was used in the present study to analyze
Facebook status updates.
Method
Data were obtained from the myPersonality Facebook
application, which has been used by more than six million
users to voluntarily take a variety of psychological tests and
receive feedback.
29
All users provided consent to the anon-
ymous use of their Facebook data and test results for research
purposes upon installation of the myPersonality application.
They had the option to choose which Facebook data to dis-
close. Their data are only accessible to registered researchers
of the myPersonality project. This research protocol has re-
ceived IRB approval.
30
A total of 99,408 participants took the Satisfaction With
Life Scale (SWLS).
31
The scale has five items, including
‘‘I am satisfied with my life’’ and ‘‘If I could live my life
over, I would change almost nothing’’ with a 7-point Likert
response scale (1 =‘‘strongly disagree’ to 7 =‘strongly
agree’’). It is a well-established and widely used measure of
an individual’s own evaluation of life satisfaction and cog-
nitive judgment of SWB.
32,33
In the current study, SWLS
scores were highly reliable (Cronbach’s a=0.82; M=4.38,
SD =1.37), consistent with past studies.
34
Among all the participants, only 3,324 provided access to
their Facebook status updates. Their status updates posted in
the year before they completed the SWLS were downloaded.
Status updates were grouped into four 3 month periods.
Period 1 consisted of the most recent updates (i.e., those
posted within the 3 months prior to completing the SWLS).
Period 4 consisted of the oldest updates (i.e., those posted in
the 10–12 months prior to completing the SWLS). To keep
the sample size consistent across the four periods, only 1,124
participants were included who had at least one status update
in each of the four periods in the final analysis. In this
sample, the SWLS exhibited high reliability (Cronbach’s
a=0.83) and had an average score of 4.32 (SD =1.40),
comparable to that of the full sample. A total of 134,087
status updates were collected. The average number of words
per status update was 13.8 (SD =6.1). Among the 195 users
who reported their sex, 132 female were female. Among the
193 users who reported their age, the mean age was 26.2
years (interquartile range =6.5).
Results
First, the status updates of the sample (1,124 participants)
were analyzed using LIWC. The mean frequency of positive
and negative emotion words was 4.7% (SD =1.7%) and 2.5%
(SD =1.2%), respectively. To examine the representativeness
of the sample, the frequency of emotional expressions of a
larger sample was analyzed: 150,383 myPersonality users who
provided their status updates but did not provide their SWLS
ratings. The mean frequency of positive and negative emo-
tional words was 3.9% (SD=2.0%) and 1.8% (SD =1.1%),
respectively. In both samples, positive emotional words were
used about twice as often as negative emotional words.
Table 1 shows the descriptive characteristics and corre-
lations among the variables in this study. Results showed no
significant correlation between life satisfaction and positive
emotional expression at any of the four periods ( p>0.10).
In contrast, life satisfaction correlated negatively with neg-
ative emotional expression in the three most recent periods
(p<0.001), demonstrating that negative emotional experi-
ences within the past 9 months were related to SWB.
It is possible that the prediction of life satisfaction im-
proves monotonically as status updates are aggregated over
longer periods of time. For example, impression management
notwithstanding, people who are truly happy may express
positive emotion in their status updates more consistently
across time. Therefore, status updates were combined into in-
creasing periods of 1 month. Thus, the first period consisted of
only the most recent month of updates, whereas the 12th period
consisted of a full year of updates. LIWC emotion codings
were obtained for each cumulative period. If aggregation im-
proves the prediction of SWB, the correlation between life
satisfaction and emotional expression should increase as up-
dates are cumulated across more months. However, the results
did not completely support this prediction (see Fig. 1). The
374 LIU ET AL.
correlation between positive emotion and life satisfaction did
not improve as status updates were cumulated across time.
Even when updates were cumulated across the full year, the
correlation was not statistically significant ( p>0.05). Thus,
merely increasing the amount of status updates did not improve
the prediction of life satisfaction from positive emotional ex-
pression. The correlation between negative emotion and life
satisfaction increased gradually, but leveled out after aggre-
gating 10 months of updates. This is consistent with the cor-
relational analysis in Table 1 where the negative emotional
experiences from the 10th to 12th last month (i.e., NE4) did
not correlate with life satisfaction. Overall, the above results
confirmed the hypotheses that only negative emotional ex-
pressions in status updates are associated with SWB, and only
expressions within the most recent months (i.e., 9–10 months
prior) are related to SWB.
To evaluate the relationship between emotional expres-
sions and SWB further, a multiple mediation model was
tested to examine both the direct effects and indirect effects
of negative emotional expression at periods 1–4 on life
satisfaction. A direct effect refers to whether expression at
any given time period predicts life satisfaction, controlling
for expression at other periods. However, it is also possible
for expression at one period to have indirect effects on life
satisfaction through subsequent time periods. For example, a
car accident in period 4 might have trickle down effects on
subsequent periods (e.g., increased financial burden due to
repair costs and hospital fees), which ultimately affects life
satisfaction. SPSS was used with the PROCESS macro,
which employs a bias corrected bootstrapping method to
provide confidence intervals (CI) around the indirect ef-
fects.
35
To evaluate the direct and indirect effects of each
period, 5,000 bootstrap samples were specified and 95% CI
were constructed. A 95% CI that does not include 0 indicates
a significant (nonzero) indirect mediation effect.
36,37
A path diagram of the multiple mediation model is shown
in Figure 2. Note that these path estimates control for all
possible indirect effects among the different periods (e.g., the
path from period 4 to period 2 to life satisfaction). However,
to simplify the presentation and discussion, the focus is only
FIG. 1. Correlation between life
satisfaction and emotional expres-
sion in status updates cumulatively
combined across 12 months. Num-
bers represent the magnitude of
correlation between positive/nega-
tive emotion and life satisfaction;
*p<0.05; **p<0.01; ***p<0.001.
Table 1. Descriptive Characteristics and Correlations Among Life Satisfaction
and Positive and Negative Emotion in Each Period
LS PE1 PE2 PE3 PE4 NE1 NE2 NE3 NE4
LS —
PE1 0.043
PE2 0.037 0.296***
PE3 -0.003 0.216*** 0.191***
PE4 -0.017 0.149*** 0.130*** 0.264***
NE1 -0.145*** -0.114*** 0.036 -0.003 0.025 —
NE2 -0.105*** -0.027 -0.098** 0.079** 0.051 0.220***
NE3 -0.120*** -0.023 -0.014 -0.043 -0.040 0.194*** 0.223***
NE4 -0.037 -0.027 -0.042 -0.058 -0.087** 0.241*** 0.211*** 0.178***
M4.32 4.75 4.78 5.00 4.81 2.46 2.43 2.56 2.71
SD 1.40 2.58 3.24 3.92 4.60 1.82 1.95 2.58 3.60
**p<0.01; ***p<0.001.
LS, life satisfaction; PE1, positive emotion in period 1 (the 1st to 3rd last months); PE2, positive emotion in period 2 (the 4th to 6th last
months); PE3, positive emotion in period 3 (the 7th to 9th last months); PE4, positive emotion in period 4 (the 10th to 12th last months);
NE1 to NE4, negative emotion in periods 1–4.
FACEBOOK STATUS UPDATES REFLECT SUBJECTIVE WELL-BEING 375
on the direct effect of each period, as well as the indirect
effects through adjacent periods. The direct effect of each
period is indicated by the arrows running directly from each
period to life satisfaction. Significant direct effects were
observed for periods 1–3, suggesting that negative emotional
expressions at each period improves the prediction of life
satisfaction above and beyond each other. There were also a
number of indirect effects. For example, negative emotional
expressions in period 2 predicted expressions in period 1,
which in turn predicted lower life satisfaction. As shown in
Table 2, each period prior to period 1 exerted significant
indirect effects on life satisfaction. Although negative ex-
pressions in period 4 did not directly predict life satisfaction,
they indirectly predicted lower life satisfaction through their
effects on periods 3 through 1. This suggests that distant
negative experiences may be related to life satisfaction
through more recent negative experiences.
Discussion
Nowadays, millions of people use online media to express
thoughts and feelings. Understanding how user-generated
contents are related to psychological variables can improve
the understanding of online behavior and allow better use to
be made of online data. This study examined how emotional
expressions in Facebook status updates are related to SWB—
specifically, self-reported life satisfaction. The results show
that positive emotional experiences reported on Facebook
were not associated with life satisfaction. However, negative
emotional experiences in the last 9–10 months were nega-
tively related to life satisfaction. These findings have im-
portant theoretical and practical implications.
First, this study shows that Facebook status updates reveal
users’ SWB. This is consistent with past findings where us-
ers’ behaviors on Facebook reflect a wide range of traits and
attributes, including personality, ethnicity, gender, age, and
sexual orientation.
29,38–40
It suggests that Facebook data can
be a valid source to explore psychological processes and
phenomena. However, it is important to note that some
previously established relationships may not hold in social
media. For example, past research has shown that positive
emotion is related to SWB.
6
However, in the present study,
positive emotional expressions in Facebook status updates
were not associated with life satisfaction. This is probably
due to the use of impression management strategies to
present a positive social image in social media. Users have
been found to disclose more positive than negative emotions
on Facebook selectively.
16–18
In the present study, positive
emotional words were used about twice as often as negative
emotional words. This may reduce the predictive power of
0.1273***
0.1449*** 0.1418***
-0.0461*
-0.0469**
-0.0914***
0.0080 LS
NE4
NE3 NE1
NE2
FIG. 2. Path analysis for direct
and indirect prediction of life
satisfaction (LS) by negative
emotion in each period. NE1 to
NE4 =negative emotion in periods
1–4. Numbers beside each arrow
represent the path coefficients.
*p<0.05; **p<0.01; ***p<0.001.
Table 2. Prediction of Life Satisfaction by Negative Emotion in Each Period
Prediction of LS
Predictor Mediator(s) Direct Indirect Total
NE1 -0.0914*** -0.0914***
(-0.1382, -0.0445) (-0.1382, -0.0445)
NE2 NE1 -0.0461* -0.0130* -0.0590*
(-0.0897, -0.0025) (-0.0273, -0.0045) (-0.1024, -0.0157)
NE3 NE1 -0.0469** -0.0167* -0.0636*
NE2 (-0.0796, -0.0142) (-0.0305, -0.0080) (-0.0957, -0.0315)
NE4 NE1 0.0080
(-0.0156, 0.0315)
-0.0224*
(-0.0357, -0.0123)
-0.0144
(-0.0371, 0.0083)
NE2
NE3
*p<0.05; **p<0.01; ***p<0.001.
NE1, negative emotion in period 1 (the 1st to 3rd last months); NE2, negative emotion in period 2 (the 4th to 6th last months); NE3,
negative emotion in period 3 (the 7th to 9th last months); NE4, negative emotion in period 4 (the 10th to 12th last months). Numbers above
brackets represent the path coefficients. Numbers in brackets represent the 95% confidence intervals.
376 LIU ET AL.
positive emotions on SWB. The findings suggest that the re-
lationship between emotion and SWB can be context depen-
dent, and past observations made in offline settings may not
hold in online environments. This highlights the importance of
considering social contexts in research on well-being. In ad-
dition, the results provide new longitudinal evidence to support
the past finding that recent emotional experiences influence
SWB.
15
Although negative experiences that occurred more
than 9 months ago did not directly relate to life satisfaction,
they contributed indirectly by predicting subsequent negative
experiences. Thus, it suggests that distant events matter as
well—even if only in an indirect sense.
Second, SWB is an important measure of life quality and
has been found to affect health, income, and social rela-
tionships.
6
Understanding how contents in users’ Facebook
status updates reflect SWB provides researchers new op-
portunities for measuring SWB without self-report surveys.
Past research has attempted to measure SWB based on social
media but showed mixed results.
9,10
The present study pro-
vides empirical evidence supporting that Facebook status
updates can predict SWB, and found a similar effect size as
Kramer’s study
9
—although negative emotional expression
appears to be more diagnostic than positive emotional ex-
pression. It also suggests that it is the time frame, not merely
the amount of status updates, that determines the accuracy of
the prediction. Only recent status updates matter in predict-
ing current SWB. Including more distant status updates may
not improve the accuracy of prediction—particularly in the
case of positive emotional expression.
These findings provide important insights for optimizing
tools to predict well-being accurately from social media.
They open up the opportunity for health professionals to
monitor users’ psychological states naturally and provide
appropriate interventions if needed. Tools can be developed
to identify factors and events that influence SWB on a large
scale, and provide policy makers with concrete evidence so
that they can effectively formulate policies and create ac-
tivities to improve the well-being of citizens. This study also
illustrates an example of utilizing Big Data for psychological
research. Future research can incorporate more factors in
social media such as geographical information and network
structure to understand better the interaction between psy-
chological and environmental factors. It will also be impor-
tant to determine if there are ways to disambiguate positive
emotional expressions on status updates (e.g., by incorpo-
rating emoticons) to improve their correspondence with self-
reported well-being. The effects of positive emotion are
distinct from negative emotion. Positive emotions predict
longevity and have important implications for social rela-
tionships.
8
Thus, there is value in improving the ability to
detect positive emotional experiences accurately.
This study has several limitations. First, the results are
based on active Facebook users who posted at least one status
update in 3 months. Findings might be different for other
users who do not frequently post on Facebook. It is possible
that their status updates may not provide enough information
to reflect SWB. It is also possible that infrequent users en-
gage in less impression management, and therefore make
both of their positive and negative emotional expressions
predictive of their SWB. Future studies need to examine
further the patterns of different users groups and identify
possible variations.
Second, the study did not have detailed information about
the characteristics of the participants, including sex and per-
sonality. Previous studies have shown sex differences in
emotional expression, with women being more emotionally
expressive than men.
41
In this study, only 195 participants
reported their sex (132 female). No significant sex difference
was found in positive or negative emotional expression, ei-
ther for the total 12 months or any of the four 3 month
periods ( p>0.05). Past research has also shown close asso-
ciation between emotional experiences and personality traits,
especially extraversion and neuroticism.
42–44
Although these
are important lines of research, the main objective of the
present paper was to evaluate prediction and not establish
causality. It could very well be that both the expression of
negative emotions and low life satisfaction are reflections of
a common personality trait such as neuroticism. Irrespective
of whether this is the case, efforts to evaluate the quality of
life via social media
9
are predicated on the assumption that
emotional expressions on these platforms actually reflect
how users feel. The observed correlation between negative
emotional expression and self-reported life satisfaction is
critically important from this standpoint. Moreover, the an-
alyses suggest a critical window of 9–10 months, within
which negative expressions correspond with users’ well-
being. Finally, research has shown that different online SNS
have different user characteristics and usage patterns.
45,46
Therefore, it is important to examine if the findings can
generalize to other types of SNS and user groups.
Conclusion
The present study reveals the temporal relationship be-
tween emotional expressions in Facebook status updates and
SWB. It showed that users’ negative (but not positive)
emotional expressions in Facebook status updates from the
past 9–10 months were negatively related to their life satis-
faction. These results suggest that both the valence and the
time frame of emotional expressions determine whether
Facebook status updates can accurately reflect users’ sub-
jective well-being. The findings shed light on the character-
istics of online social media and improve the understanding
of how user-generated contents reflect users’ psychological
states.
Acknowledgment
This research was supported by Singapore Ministry of
Education Academic Research Fund Tier 1 Grant RGT37/13.
Author Disclosure Statement
No competing financial interests exist.
References
1. Ko
¨bler F, Riedl C, Vetter C, et al. (2010) Social connect-
edness on Facebook—an explorative study on status mes-
sage usage. In Proceedings of 16th Americas Conference
on Information Systems. Lima, Peru, Paper 247.
2. Carr CT, Schrok DB, Dauterman P. Speech acts within
Facebook status messages. Journal of Language & Social
Psychology 2012; 31:176–196.
3. Lichterman J. (2014) Facebook teams with Storyful to
highlight news content published on the social network.
FACEBOOK STATUS UPDATES REFLECT SUBJECTIVE WELL-BEING 377
www.niemanlab.org/2014/04/facebook-teams-with-storyful-
to-highlight-news-content-published-on-the-social-network/
(accessed June 21, 2014).
4. Kramer ADI, Chung CK. (2011) Dimensions of self-
expression in Facebook status updates. In Proceedings of
the Fifth International AAAI Conference on Weblogs and
Social Media. Barcelona, Spain, pp. 169–176.
5. Tov W. Daily experiences and well-being: do memories of
events matter? Cognition & Emotion 2012; 26:1371–1389.
6. Diener E. The remarkable changes in the science of sub-
jective well-being. Perspectives on Psychological Science
2013; 8:663–666.
7. Diener E, Chan MY. Happy people live longer: subjective
well-being contributes to health and longevity. Applied
Psychology: Health & Well-Being 2011; 3:1–43.
8. Lyubomirsky S, King L, Diener E. The benefits of frequent
positive affect: does happiness lead to success? Psycholo-
gical Bulletin 2005; 131:803–855.
9. Kramer ADI. (2010) An unobtrusive behavioral model of
‘‘gross national happiness.’’ In Proceedings of the SIGCHI
Conference on Human Factors in Computing Systems.Atlanta,
GA, pp. 287–290.
10. Wang N, Kosinski M, Stillwell DJ, et al. Can well-being be
measured using Facebook status updates? Validation of
Facebook’s Gross National Happiness Index. Social In-
dicators Research 2014; 115:483–491.
11. Headey B, Wearing A. Personality, life events, and sub-
jective well-being: toward a dynamic equilibrium model.
Journal of Personality & Social Psychology 1989; 57:731–
739.
12. Kahneman D, Krueger AB, Schkade DA, et al. A survey
method for characterizing daily life experience: the day
reconstruction method. Science 2004; 306:1776–1780.
13. Diener E. Subjective well-being. Psychological Bulletin
1984; 95:542–575.
14. Rutledge RB, Skandali N, Dayan P, et al. A computational
and neural model of momentary subjective well-being.
Proceedings of the National Academy of Sciences of the
United States of America 2014; 111:12252–12257.
15. Suh E, Diener E, Fujita F. Events and subjective well-
being: only recent events matter. Journal of Personality &
Social Psychology 1996; 70:1091–1102.
16. Lin H, Tov W, Qiu L. Emotional disclosure on social
networking sites: the role of network structure and psycho-
logical needs. Computers in Human Behavior 2014; 41:
342–350.
17. Bazarova NN, Taft JG, Choi YH, et al. Managing impres-
sions and relationships on Facebook: self-presentational and
relational concerns revealed through the analysis of lan-
guage style. Journal of Language & Social Psychology
2012; 32:121–141.
18. Qiu L, Lin H, Leung AK, et al. Putting their best foot
forward: emotional disclosure on Facebook. Cyberpsy-
chology, Behavior, & Social Networking 2012; 15:569–
572.
19. Pennebaker JW, Booth RJ, Francis ME. (2007) Linguistic
inquiry and word count (LIWC2007). Austin, TX. www
.liwc.net (accessed June 21, 2014).
20. Pennebaker JW, Chung CK, Ireland M, et al. (2007) The
development and psychometric properties of LIWC2007.
Austin, TX. www.liwc.net (accessed June 21, 2014).
21. Chung CK, Pennebaker JW. (2012) Linguistic inquiry and
word count (LIWC): Pronounced ‘‘Luke,’.and other
useful facts. In McCarthy P, Boonthum C, eds. Applied
natural language processing and content analysis: Identi-
fication, investigation, and resolution. Hershey, PA: IGI
Global, pp. 206–229.
22. Pennebaker JW, Mehl MR, Niederhoffer KG. Psychologi-
cal aspects of natural language use: our words, our selves.
Annual Review of Psychology 2003; 54:547–577.
23. Tausczik YR, Pennebaker JW. The psychological meaning
of words: LIWC and computerized text analysis methods.
Journal of Language & Social Psychology 2010; 29:24–54.
24. Tov W, Ng KL, Lin H, et al. Detecting well-being via
computerized content analysis of brief diary entries. Psy-
chological Assessment 2013; 25:1069–1078.
25. Chee B, Berlin R, Schatz B. (2009) Measuring population
health using personal health messages. In Proceedings of
the Annual American Medical Informatics Association
Symposium. San Francisco, CA, pp. 92–96.
26. Yu B, Kaufmann S, Diermeier D. Classifying party affili-
ation from political speech. Journal of Information Tech-
nology & Politics 2008; 5:33–48.
27. Golder SA, Macy MW. Diurnal and seasonal mood vary
with work, sleep, and daylength across diverse cultures.
Science 2011; 333:1878–1881.
28. Qiu L, Lin H, Ramsay J, et al. You are what you tweet:
personality expression and perception on Twitter. Journal
of Research in Personality 2012; 46:710–718.
29. Kosinski M, Stillwell DJ, Graepel T. Private traits and at-
tributes are predictable from digital records of human be-
havior. Proceedings of the National Academy of Sciences
of the United States of America 2013; 110:5802–5805.
30. Youyou W, Kosinski M, Stillwell D. Computer-based per-
sonality judgments are more accurate than those made by
humans. Proceedings of the National Academy of Sciences
of the United States of America 2015; 112:1036–1040.
31. Diener E, Emmons RA, Larsen RJ, et al. The Satisfaction
With Life Scale. Journal of Personality Assessment 1985;
49:71–75.
32. Palmer B, Donaldson C, Stough C. Emotional intelligence
and life satisfaction. Personality & Individual Differences
2002; 33:1091–1100.
33. Pavot W, Diener E. The Satisfaction With Life Scale and
the emerging construct of life satisfaction. The Journal of
Positive Psychology 2008; 3:137–152.
34. Pavot W, Diener E. Review of the Satisfaction with Life
Scale. Psychological Assessment 1993; 5:164–172.
35. Hayes AF. (2012) PROCESS: a versatile computational tool
for observed variable mediation, moderation, and conditional
process modeling [White paper]. www.personal.psu.edu/jxb14/
M554/articles/process2012.pdf (accessed June 21, 2014).
36. Preacher KJ, Hayes AF. SPSS and SAS procedures for
estimating indirect effects in simple mediation models.
Behavior Research Methods, Instruments, & Computers
2004; 36:717–731.
37. Preacher KJ, Rucker DD, Hayes AF. Addressing moderated
mediation hypotheses: theory, methods, and prescriptions.
Multivariate Behavioral Research 2007; 42:185–227.
38. Kern ML, Eichstaedt JC, Schwartz HA, et al. The online
social self: an open vocabulary approach to personality.
Assessment 2014; 21:158–169.
39. Kern ML, Eichstaedt JC, Schwartz HA, et al. From ‘‘Sooo
excited!!!’’ to ‘‘So proud’’: using language to study de-
velopment. Developmental Psychology 2014; 50:178–188.
40. Park G, Schwartz HA, Eichstaedt JC, et al. Automatic per-
sonality assessment through social media language. Journal
of Personality & Social Psychology 2015; 108:934–952.
378 LIU ET AL.
41. Kring AM, Gordon AH. Sex differences in emotion: ex-
pression, experience, and physiology. Journal of Person-
ality & Social Psychology 1998; 74:686–703.
42. Larsen RJ, Ketelaar T. Personality and susceptibility to
positive and negative emotional states. Journal of Person-
ality & Social Psychology 1991; 61:132–140.
43. Rapson Gomez, Andre Gomez, Cooper A. Neuroticism and
extraversion as predictors of negative and positive emotional
information processing: comparing Eysenck’s, Gray’s, and
Newman’s theories. European Journal of Personality 2002;
16:333–350.
44. Watson D, Clark LA. On traits and temperament: general
and specific factors of emotional experience and their re-
lation to the five-factor model. Journal of Personality 1992;
60:441–476.
45. Qiu L, Lin H, Leung AK-Y. Cultural differences and
switching of in-group sharing behavior between an American
(Facebook) and a Chinese (Renren) social networking site.
Journal of Cross-Cultural Psychology 2013; 44:106–121.
46. Wilson RE, Gosling SD, Graham LT. A review of Face-
book research in the social sciences. Perspectives on Psy-
chological Science 2012; 7:203–220.
Address correspondence to:
Prof. Lin Qiu
Division of Psychology
Nanyang Technological University
HSS-04-15
14 Nanyang Drive
Singapore 637332
Singapore
E-mail: linqiu@ntu.edu.sg
FACEBOOK STATUS UPDATES REFLECT SUBJECTIVE WELL-BEING 379
... While Facebook, Twitter, and Wikipedia have received considerable attention in various fields of psychology (e.g., Cress et al., 2016;Liu et al., 2015;Wang et al., 2016), other sources of big data have not yet been explored in a similar depth. Especially user-generated content in collaborative projects as in Wikipedia is only rarely examined but can be a valuable data source for a variety of research questions such as the development of social rules and norms in collaboration, exemplar theories, or contribution and correction processes in communities. ...
... This helps researchers to select exactly these elements that are relevant for further analyses. A major benefit of OpenStreetMap is that the data do not consist of unstructured text as in Wikipedia (Cress et al., 2016) or social-media networks (Liu et al., 2015;Wang et al., 2016), but rather contain numerical information for geographical properties as well as semantic tags in form of key-value pairs, which facilitates data preparation. By making OpenStreetMap data more accessible to psychology and the social sciences, OSM-Psychology offers researchers the possibility to test behavioral hypotheses using large-scale behavioral data obtained in a natural, ecologically valid environment rather than in artificial laboratory settings. ...
Preprint
Full-text available
Big data are not yet commonly used in psychological research as they are often difficultto access and process. One source of behavioral data containing both spatial andthematic information is OpenStreetMap, a collaborative online project aiming to developa comprehensive world map. Besides spatial and thematic information about buildings,streets, and other geographical features, the collected data also contains informationabout the contribution process itself. Even though such data can be potentially useful forstudying individual judgments and group processes within a natural context, behavioraldata generated in OpenStreetMap have not yet been easily accessible for scholars inpsychology and the social sciences. To overcome this obstacle, we developed a softwarepackage which makes OpenSteetMap data more accessible and allows researchers toextract data sets from the OpenStreetMap database as CSV or JSON files. Furthermore,we show how to select relevant map sections in which contributor activity is high and howto model and predict the behavior of contributors in OpenStreetMap. Moreover, wediscuss opportunities and possible limitations of using behavioral data fromOpenStreetMap as a data source.
... Thus, the presence of these demographic variables tends to be crucial for OSWB research to perform post-stratification and make results representative of the general population of the analyzed country. Lastly, there is still significant controversy about correspondence of digital traces to survey data [19,60,[71][72][73]. More specifically, it is often asked whether social media content really represents the state of affairs in the offline world. ...
Article
Full-text available
Policymakers and researchers worldwide are interested in measuring the subjective well-being (SWB) of populations. In recent years, new approaches to measuring SWB have begun to appear, using digital traces as the main source of information, and show potential to overcome the shortcomings of traditional survey-based methods. In this paper, we propose the formal model for calculation of observable subjective well-being (OSWB) indicator based on posts from a social network, which utilizes demographic information and post-stratification techniques to make the data sample representative by selected characteristics of the general population. We applied the model on the data from Odnoklassniki, one of the largest social networks in Russia, and obtained an OSWB indicator representative of the population of Russia by age and gender. For sentiment analysis, we fine-tuned several language models on RuSentiment and achieved state-of-the-art results. The calculated OSWB indicator demonstrated moderate to strong Pearson’s (r=0.733, p=0.007, n=12) correlation and strong Spearman’s (rs=0.825, p=0.001, n=12) correlation with a traditional survey-based Happiness Index reported by Russia Public Opinion Research Center, confirming the validity of the proposed approach. Additionally, we explored circadian (24 h) and circaseptan (7 day) patterns, and report several interesting findings for the population of Russia. Firstly, daily variations were clearly observed: the morning had the lowest level of happiness, and the late evening had the highest. Secondly, weekly patterns were clearly observed as well, with weekends being happier than weekdays. The lowest level of happiness occurs in the first three weekdays, and starting on Thursday, it rises and peaks during the weekend. Lastly, demographic groups showed different levels of happiness on a daily, weekly, and monthly basis, which confirms the importance of post-stratification by age group and gender in OSWB studies based on digital traces.
... Our analysis can only be used to understand the patterns of those who use Twitter or Weibo to communicate and lacks explanatory power for the least developed regions and elderly populations. Second, although social media expressed sentiment correlates with the affective aspects of subjective well-being, it cannot reliably measure the life satisfaction dimension of subjective well-being 39,40 . Due to the limitations in representativeness and measurement, social media sentiment analysis should serve as a complement rather than a substitute for self-reported measures of subjective well-being. ...
Article
Full-text available
The COVID-19 pandemic has created unprecedented burdens on people’s physical health and subjective well-being. While countries worldwide have developed platforms to track the evolution of COVID-19 infections and deaths, frequent global measurements of affective states to gauge the emotional impacts of pandemic and related policy interventions remain scarce. Using 654 million geotagged social media posts in over 100 countries, covering 74% of world population, coupled with state-of-the-art natural language processing techniques, we develop a global dataset of expressed sentiment indices to track national- and subnational-level affective states on a daily basis. We present two motivating applications using data from the first wave of COVID-19 (from 1 January to 31 May 2020). First, using regression discontinuity design, we provide consistent evidence that COVID-19 outbreaks caused steep declines in expressed sentiment globally, followed by asymmetric, slower recoveries. Second, applying synthetic control methods, we find moderate to no effects of lockdown policies on expressed sentiment, with large heterogeneity across countries. This study shows how social media data, when coupled with machine learning techniques, can provide real-time measurements of affective states. Using tweets in over 100 countries, Wang et al. examine evidence of global sentiment during the COVID-19 pandemic. They find that COVID-19 outbreaks caused a decline in sentiment worldwide, and the effects of lockdowns differed across countries.
... 4. Facebook status. Авторы [10][11][12] предлагают методику прогнозирования благополучия с использованием социальных сетей. На основе анализа статусов, открытых сообщений определяется семантическая соотнесенность ключевых слов в сообщениях, далее строится агрегированный индекс благополучия. ...
Article
This paper aims to develop a theory of statistical observation in terms of scientific and methodological approaches to processing big data and to determine the possibilities of integrating information resources of various types to measure complex latent categories (using the example of social comfort) and to apply this experience in practice through the use of the financial situation indicators in forecasting. The authors have built a social comfort model in which the choice of weights for its components is based on a modified principal component analysis . The assessment is based on Google Trends data and official statistics. Google Trends data analysis methods are based on the development of an integrated approach to the semantic search for information about the components of social comfort, which reduces the share of author’s subjectivity; methodology of primary processing, considering the principles of comparability, homogeneity, consistency, relevance, description of functions and models necessary for the selection and adjustment of search queries. The proposed algorithm for working with big data allowed to determine the components of social comfort (“Education and Training”, “Safety”, “Leisure and free time”), for which it is necessary to directly integrate big data in the system of primary statistical accounting with further data processing and obtaining composite indicators. The authors conclude that a stable significant correlation has been found for the “Financial Situation” component, which makes it possible to use it for further calculations and extrapolation of financial indicators. The scientific novelty lies in the development of principles and directions for the integration of two alternative data sources when assessing complex latent categories. The findings and the results of the integral assessment of social comfort can be used by state statistics authorities to form a new type of continuous statistical observation based on the use of big data, as well as by executive authorities at the federal, regional and municipal levels in terms of determining the priorities of socio-economic policy development.
... Social media and mobile technology provide sources of data that circumvent self-report (Luhmann, 2017). For example, automated linguistic analysis of Twitter and Facebook posts yields expected correlations between the valence of language use and dynamics of positive affect revealed in other research (e.g., day of week, regional variations; Bliss et al., 2012;Dodds et al., 2011;Frank et al., 2013;Mitchell et al., 2013) as well as self-reported hedonic wellbeing (satisfaction with life: Chen et al., 2017;Liu et al., 2015;Yang and Srinivasan, 2016;negative affect: Settanni and Marengo, 2015;cf. Wang et al., 2014). ...
Article
What does it mean to be “well” and how might such a state be cultivated? When we speak of wellbeing, it is of ourselves and fellow humans. When it comes to nonhuman animals, consideration turns to welfare. My aim herein is to suggest that theoretical approaches to human wellbeing might be beneficially applied to consideration of animal welfare, and in so doing, introduce new lines of inquiry and practice. I will review current approaches to human wellbeing, adopting a triarchic structure that delineates hedonic wellbeing, eudaimonic wellbeing, and social wellbeing. For each, I present a conceptual definition and a review of how researchers have endeavored to measure the construct. Drawing these three domains of research together, I highlight how these traditionally anthropocentric lines of inquiry might be extended to the question of animal welfare – namely by considering hedonic welfare, eudaimonic welfare, and social welfare as potentially distinguishable and complementary components of the broader construct of animal welfare.
... Guan et al. have identified Chinese microblog users with high suicide probability using internet-based profiles and linguistic features [20]. Liu et al. have proposed to detect suicide risk on social media using a Chinese suicide dictionary [21]. Wu et al. have analyzed Facebook status updates to determine the extent to which users' emotional expression predicted their SWB -specifically their self-reported satisfaction with life [22]. ...
Article
Today, with the development of internet technology, a new kind of social relations and interactions have been formed in the newly emerged social networks. Through social networks, the users can share different types of content, including personal information, text, image, video, music, poem, and other related information, which express their mental states, emotions, feelings, and thoughts. Thus, a new and essential aspect of human life is being formed in a virtual space in social networks, which must be explored from several viewpoints, such as mental disorders. Analyzing mental disorders according to the social network data can guide us to gain new approaches to improve the public health of the whole society. To this aim, developing mental health feature extraction (MHFE) methods in a social network is essential and is now becoming an active research area. Therefore, in this paper, a review of existing techniques and methods in MHFE is presented, and a comprehensive framework is provided to classify these approaches. Furthermore, to analyze and evaluate each approach in extraction methods, an appropriate set of functional criteria is proposed, which leads to a more accurate understanding and correct use of them.
Chapter
Mental health problems are widely recognized as a major public health challenge worldwide. This highlights the need for effective tools for detecting mental health disorders in the population. Social media data is a promising source of information where people publish rich personal information that can be mined to extract valuable psychological information. However, social media data poses its own set of challenges, such as the specific terms and expressions used on different platforms, interactions between different users through likes and shares, and the need to disambiguate between statements about oneself and about third parties. Traditionally, social media natural language processing (NLP) techniques have looked at text classifiers and user classification models separately, which presents a challenge for researchers wanting not only to combine text sentiment and user sentiment analysis but also to extract user’s narratives from the textual content.
Chapter
The research on big data in the social sciences and its impact has received much interest from practitioners and policy-makers. Data science can help find the answer to research questions in the social sciences because data is the lifeblood of the decision-making process; it is also the raw material for the accountability process. New data sources, new technologies, and new analytical approaches can make evidence-based decision-making more efficient and flexible. The analysis of this data plays a large role in discussing the challenges facing our societies today. This research provides an analysis of the six factors that influence happiness—GDP per capita, social support, life expectancy, freedom, corruption, and generosity. In this research, the World Happiness Report was studied in 2019, where its survey of the state of global happiness ranks 156 countries through their citizens’ happiness. Depending on six factors. This study focuses on analyzing factors that affect happiness and satisfaction with life.
Article
Previous studies have confirmed the powerful influence of social networks on individuals’ mental health and well-being. However, the relative contribution of different dimensions of network characteristics to psychosocial outcomes on mobile social media is unclear. The objectives of this current research are to investigate the effect of WeChat involvement on young people’s network characteristics (network diversity and network size) and to examine the influence of network characteristics on their online bridging capital, bonding capital and life satisfaction. Especially, the hypothesized research model is developed to unpack the mediating role of network characteristics in the association between WeChat usage and related psychosocial outcomes. Based on survey data collected from 566 young people, the study demonstrates that WeChat involvement positively predicts individuals’ network diversity and network size in their daily routine life. Interestingly, network diversity is also positively associated with degrees of online bridging capital, online bonding capital and satisfaction with life. Additionally, the mediating roles of network diversity and network size between WeChat involvement and psychosocial outcomes is revealed. The research findings provide fresh insight into understanding the implication of individuals’ mobile social media use and network network attibutes for well-being outcomes in contemporay mobile-mediated environment.
Article
Full-text available
Language use is a psychologically rich, stable individual difference with well-established correlations to personality. We describe a method for assessing personality using an open-vocabulary analysis of language from social media. We compiled the written language from 66,732 Facebook users and their questionnaire-based self-reported Big Five personality traits, and then we built a predictive model of personality based on their language. We used this model to predict the 5 personality factors in a separate sample of 4,824 Facebook users, examining (a) convergence with self-reports of personality at the domain- and facet-level; (b) discriminant validity between predictions of distinct traits; (c) agreement with informant reports of personality; (d) patterns of correlations with external criteria (e.g., number of friends, political attitudes, impulsiveness); and (e) test-retest reliability over 6-month intervals. Results indicated that language-based assessments can constitute valid personality measures: they agreed with self-reports and informant reports of personality, added incremental validity over informant reports, adequately discriminated between traits, exhibited patterns of correlations with external criteria similar to those found with self-reported personality, and were stable over 6-month intervals. Analysis of predictive language can provide rich portraits of the mental life associated with traits. This approach can complement and extend traditional methods, providing researchers with an additional measure that can quickly and cheaply assess large groups of participants with minimal burden. (PsycINFO Database Record (c) 2014 APA, all rights reserved).
Article
Full-text available
Significance A common question in the social science of well-being asks, “How happy do you feel on a scale of 0 to 10?” Responses are often related to life circumstances, including wealth. By asking people about their feelings as they go about their lives, ongoing happiness and life events have been linked, but the neural mechanisms underlying this relationship are unknown. To investigate it, we presented subjects with a decision-making task involving monetary gains and losses and repeatedly asked them to report their momentary happiness. We built a computational model in which happiness reports were construed as an emotional reactivity to recent rewards and expectations. Using functional MRI, we demonstrated that neural signals during task events account for changes in happiness.
Article
Full-text available
Objective: We present a new open language analysis approach that identifies and visually summarizes the dominant naturally occurring words and phrases that most distinguished each Big Five personality trait. Method: Using millions of posts from 69,792 Facebook users, we examined the correlation of personality traits with online word usage. Our analysis method consists of feature extraction, correlational analysis, and visualization. Results: The distinguishing words and phrases were face valid and provide insight into processes that underlie the Big Five traits. Conclusion: Open-ended data driven exploration of large datasets combined with established psychological theory and measures offers new tools to further understand the human psyche.
Article
Full-text available
We introduce a new method, differential language analysis (DLA), for studying human development in which computational linguistics are used to analyze the big data available through online social media in light of psychological theory. Our open vocabulary DLA approach finds words, phrases, and topics that distinguish groups of people based on 1 or more characteristics. Using a data set of over 70,000 Facebook users, we identify how word and topic use vary as a function of age and compile cohort specific words and phrases into visual summaries that are face valid and intuitively meaningful. We demonstrate how this methodology can be used to test developmental hypotheses, using the aging positivity effect (Carstensen & Mikels, 2005) as an example. While in this study we focused primarily on common trends across age-related cohorts, the same methodology can be used to explore heterogeneity within developmental stages or to explore other characteristics that differentiate groups of people. Our comprehensive list of words and topics is available on our web site for deeper exploration by the research community. (PsycINFO Database Record (c) 2013 APA, all rights reserved).
Article
The literature on subjective well-being (SWB), including happiness, life satisfaction, and positive affect, is reviewed in three areas: measurement, causal factors, and theory. Psychometric data on single-item and multi-item subjective well-being scales are presented, and the measures are compared. Measuring various components of subjective well-being is discussed. In terms of causal influences, research findings on the demographic correlates of SWB are evaluated, as well as the findings on other influences such as health, social contact, activity, and personality. A number of theoretical approaches to happiness are presented and discussed: telic theories, associationistic models, activity theories, judgment approaches, and top-down versus bottom-up conceptions.
Article
The science of subjective well-being (SWB) has grown dramatically in the last three decades, moving beyond the early cross-sectional surveys of the demographic correlates of SWB. Stronger methods are frequently used to study a broader set of psychological phenomena, such as the effects on SWB of adaptation, culture, personality, and genetics. One important new research finding is that SWB has beneficial effects on health and longevity, social relationships, and productivity. National accounts of SWB are being created to provide information to policy makers about the psychological well-being of citizens. The SWB accounts represent an opportunity for psychologists to demonstrate the positive effects their interventions can produce in societies. © The Author(s) 2013.
Article
This article reports the development and validation of a scale to measure global life satisfaction, the Satisfaction With Life Scale (SWLS). Among the various components of subjective well-being, the SWLS is narrowly focused to assess global life satisfaction and does not tap related constructs such as positive affect or loneliness. The SWLS is shown to have favorable psychometric properties, including high internal consistency and high temporal reliability. Scores on the SWLS correlate moderately to highly with other measures of subjective well-being, and correlate predictably with specific personality characteristics. It is noted that the SWLS is suited for use with different age groups, and other potential uses of the scale are discussed.
Article
Significance This study compares the accuracy of personality judgment—a ubiquitous and important social-cognitive activity—between computer models and humans. Using several criteria, we show that computers’ judgments of people’s personalities based on their digital footprints are more accurate and valid than judgments made by their close others or acquaintances (friends, family, spouse, colleagues, etc.). Our findings highlight that people’s personalities can be predicted automatically and without involving human social-cognitive skills.