Research ReportA study of Facebook behavior: What does it tell about your Neuroticism and Extraversion?
Introduction
Social Network Sites (SNS) facilitate interpersonal interaction and allow for the maintenance of ties that may have otherwise gone dormant (Ellison, Steinfield, & Lampe, 2007). The increased awareness of others’ activities could have profound implications for the way we keep connected with others and understand others (Berkovsky, Freyne, & Smith, 2012). People get an impression of a user based on the following SNS behaviors: (1) a “profile” created by this user, including basic demographics, personal interest and a list of friends this user chose to associate with. (2) Posted content, including videos, photos and “status updates” viewable to some or all of this user’s friends. Status updates (a.k.a. posts) are broadcast messages that are written for others’ consumption and usually are not tailored to a particular person. They are a major means to communicate with friends on SNS (Kramer, 2010). (3) Interactions with friends, such as “like” or “comment” on a post.
People are motivated to be seen as attractive, likeable, competent, and virtuous (Leary, 1996). Though SNS users strive to project a positive image of themselves (Barash, Duchenaut, Isaacs, & Bellotti, 2010), their portrayed images could still be quite telling of their underlying characteristics, such as personalities. Personality traits are consistent patterns of thoughts, feelings, or actions that distinguish people from one another (John, Robins, & Pervin, 2010). In a study conducted by Gosling, Gaddis, and Vazire (2007), 9 undergraduate research assistants rated personality traits of 133 subjects by examining their Facebook profiles. They found some consensus about profile-based personality assessment, with particularly strong consensus for Extraversion. Such personality assessment shows some accuracy, compared with traits reported by the studied subjects and their friends. This suggests that there exist some specific cues eliciting personality related behaviors and those cues are actually valid indicators of what someone is like.
The popularity of Social Network Sites (SNS) such as Facebook provides a great opportunity to examine personality inference using significant amounts of data. Indeed, recent SNS offer a wealth of behavioral indicators ranging from profiles to activity statistics that could reflect a user’s personality. This behavioral richness could be leveraged to infer the personality of the individual behind an SNS account. However, the problem of modeling personality from Facebook behaviors has achieved very little advance in terms of concrete data analysis. This is largely due to the difficulty in data collection. Researchers typically made their analysis based on the self-reported data from a small sample of students from a single university. There is no guarantee that the reported data is objective and reflects the real behaviors. The homogeneous, small-sized population likely leads to biased conclusion. The contradiction between some research results might be due to the above facts. For example, Ross et al. (2009) showed that Extroversion was found to belong to more Facebook groups and not necessarily be associated with more Facebook friends, while Hamburger and Vinitzky’s (2010) demonstrated that Extroversion had a positive effect on the number of friends, but no effect was found with regard to the use of Facebook groups.
In this paper, we use a data-driven approach to personality modeling and prediction for Facebook users. We developed a Facebook application to directly retrieve data from 1327 users. From a user’s profile, we extract a rich set of features and perform correlation analysis to discover which features are strongly correlated with the personality. The direct retrieval enables us to study objective, fine-grained signals embedded in users’ behaviors, such as writing styles or number of “likes”. Privacy concern often is a major hurdle to data collection and needs to be properly addressed. In our work, we have spent extra effort to guard personal information and designed a two-stage approach: (1) careful anonymization in the preprocessing stage to remove personally identifiable information (PII) such as name, address, email address and all numbers, and (2) the innovative design of an activity logger, which processes Facebook activities and retains only aggregated statistics. No raw content is logged in the feature set. Furthermore we assure that the aggregated statistics are sufficiently abstract such that the original content and/or meaning cannot be reconstructed from the feature set. These measures mitigate the privacy concern and are key to our data collection.
Some findings are expected and well in alignment with qualitative findings in social science studies. For instance, extroverts engage more actively in Facebook social activities. They share more photos, longer videos, and more status updates. Individuals high in Neuroticism are more likely to post accurate personal information. Our analysis also discovered some findings that we did not originally anticipate. For instance, we originally anticipated that neurotic users are more cautious and thus write less on Internet. but instead our analysis found that neurotic users tend to write longer posts, use more negative sentiment words and strongly subjective words in posts. Furthermore, we build a model to predict a Facebook user’s Neuroticism and Extraversion. Our predictor achieves a modest accuracy with correlation . Giving the noisy nature of the online data and the difficulty in personality analysis, the prediction accuracy is encouraging. In this paper, we also report some preliminary results on other personality traits, with the hope of inspiring future research.
Section snippets
Related work and theoretical background
Personality profiling. Personality traits are consistent patterns of thoughts, feelings, or actions that distinguish people from one another (John et al., 2010). Different theories make different predictions about how mean levels of personality traits change in adulthood (Srivastava, John, Gosling, & Potter, 2003), but it is generally agreed that the personality profile affects our activity (Hogan, Johnson, & Briggs, 1997). First, having a specific personality trait means reacting consistently
Research study design
Privacy-preserving data collection using Facebook app. Our data collection system was deployed as a web service, so users can easily participate from anywhere. Upon signing up for our study and giving informed consent, each participant first answered an online demographic and psychological survey. Participants rated themselves on 112 psychological items adopted from the International Personality Item Pool (Goldberg et al., 2006), using a Likert scale that ranged from 1 (Very Inaccurate) to 5
Results
From each participant’s Facebook account, we calculated 154 features, covering demographics, profile descriptions, produced content, and interactions with friends. It is important to mention that we are not claiming that we have extracted all possible variables for analysis in this paper, but rather, that we have extracted a meaningful and manageable subset of cues that cover a wide range of behaviors on Facebook. To answer RQ1, a description of derived variables, along with their means and
Conclusions
Our effort is inscribed in a broader scope of research seeking to establish correspondence between people’s real-world context and online social behaviors. Through careful design of direct retrieval of users’ Facebook activities, we were able to collect a large dataset from participants with diverse background. The purpose of this (re-)study was to use this large user sample to examine the nature of Facebook use, explore the connections between personality traits and behaviors of Facebook
Acknowledgments
The authors gratefully acknowledge support for this work from DARPA through the ADAMS (Anomaly Detection At Multiple Scales) program funded project GLAD-PC (Graph Learning for Anomaly Detection using Psychological Context). Any opinions, findings, and conclusions or recommendations in this material are those of the authors and do not necessarily reflect the views of the government funding agencies. We thank the anonymous reviewers for their constructive comments.
References (36)
- et al.
The international personality item pool and the future of public-domain personality measures
Journal of Research in Personality
(2006) - et al.
Who blogs? Personality predictors of blogging
Computers in Human Behavior
(2008) - et al.
An investigation of big five and narrow personality traits in relation to internet usage
Computers in Human Behavior
(2006) - et al.
Personality and motivations associated with Facebook use
Computers in Human Behavior
(2009) - et al.
A comparison study of user behavior on Facebook and Gmail
Computers in Human Behavior
(2013) - Barash, V., Duchenaut, N., Isaacs, E., & Bellotti, V. (2010). Faceplant: Impression (mis) management in Facebook status...
- Berkovsky, S., Freyne, J., & Smith, G. (2012). Personalized network updates: Increasing social interactions and...
- Burke, M., Kraut, R., & Marlow, C. (2011). Social capital on Facebook: Differentiating uses and users. In CHI-11 (pp....
- BusinessAndIndustryPortal. (2013). Who uses Facebook?...
Statistical power analysis for the behavioral sciences
(1988)
Personality and self-esteem as predictors of young people’s technology use
Cyberpsychology & Behavior
The benefits of Facebook “friends:” Social capital and college students’ use of online social network sites
Journal of Computer-Mediated Communication
A room with a cue: Personality judgments based on offices and bedrooms
Journal of Personality and Social Psychology
The relationship between extraversion and neuroticism and the different uses of the internet
Computers in Human Behavior
Social network use and personality
Computers in Human Behavior
Cited by (55)
Showing with whom I belong: The desire to belong publicly on social media
2023, Computers in Human BehaviorMaking friends on Facebook: Common group membership as a central decision rule
2022, Computers in Human BehaviorExtraversion, loneliness, and problematic game use: A longitudinal study
2021, Personality and Individual DifferencesCitation Excerpt :Their extroverted tendencies exhibited in the virtual world do not appear in games alone. Shen et al. (2015) analyzed 1327 Facebook users in the United States and found that extroverted users post more on social networking sites. They also received more comments or likes from their friends than do introverts (Shen et al., 2015).
The interplay between neuroticism, extraversion, and social media addiction in young adult Facebook users: Testing the mediating role of online activity using objective data
2020, Addictive BehaviorsCitation Excerpt :In particular, extraversion and neuroticism have shown consistent links with online activity (e.g., Amichai-Hamburger & Vinitzky, 2010; Kraut et al., 2002; Landers & Lounsbury, 2006). Concerning Facebook, when compared with introverted users, extraverts tend to show higher levels of posting activity, larger online networks, and higher Likes counts (Burrow & Rainone, 2017; Shen, Brdiczka, & Liu, 2015). Neurotics also tend to show increased posting activity, such as number of posted words, or comments, when compared with emotionally stable users, but lower number of received Likes (Amichai-Hamburger & Vinitzky, 2010; Burrow & Rainone, 2017; Shen et al., 2015).
Problematic Internet Use Among Adolescent Male and Female Psychiatric Inpatients: A Gender Perspective
2024, Child Psychiatry and Human Development