Once More with Feeling: Supportive Responses to Social Sharing on Facebook

Auteur-es

Moira Burke and Mike Develin

Nombre Auteurs

2

Titre

Once More with Feeling: Supportive Responses to Social Sharing on Facebook

Année de publication

2016

Référence (APA)

Burke, M., & Develin, M. (2016). Once More with Feeling : Supportive Responses to Social Sharing on Facebook. Proceedings of the 19th ACM Conference on Computer-Supported Cooperative Work & Social Computing, 1462‑1474. https://doi.org/10.1145/2818048.2835199

résumé

Life is more than cat pictures. There are tough days,heartbreak, and hugs. Under what contexts do people share these feelings online, and how do their friends respond? Using millions of de-identified Facebook status updates with poster-annotated feelings (e.g., feeling thankful or feeling worried), we examine the magnitude and circumstances in which people share positive or negative feelings and characterize the nature of the responses they receive. We find that people share greater proportions of both positive and negative emotions when their friend networks are smaller and denser. Consistent with social sharing theory, hearing about a friend’s troubles on Facebook causes friends to reply with more emotional and supportive comments. Friends’ comments are also more numerous and longer. Posts with positive feelings, on the other hand, receive more likes, and their comments have more positive language. Feelings that relate to the poster’s self worth, such as feeling defeated, feeling unloved, or feeling accomplished amplify these effects.

Mots-clés

Emotional support; social network sites; social sharing

URL

https://research.facebook.com/file/373780360921794/once-more-with-feeling-supportive-responses-to-social-sharing-on-facebook.pdf

doi

https://doi.org/10.1145/2818048.2835199

Accessibilité de l'article

Libre

Champ

Data Science, Human Computer Interaction, & UX

Type contenu (théorique Applicative méthodologique)

Théorique

Méthode

S1 : Two judges independently coded the top 200
feelings (which comprised 90% of feelings usage) on a 3-point valence scale (positive, ambiguous, or negative). Inter-
rater reliability was good (Cohen’s kappa = 0.77).

S2 : As in Study 1, the 14.2 million posts with feeling annotations were categorized according to the valence and self-relevance of the emotion.

Cas d'usage

Facebook

Objectifs de l'article

The objectives of the article are to explore the ways in which people share their emotions on Facebook and how their friends respond, and to examine the relationship between characteristics of the emotions shared and characteristics of the responses.

"The present research is the first very large-scale quantitative
study of social sharing of feelings in Facebook status
updates, examining (1) how network properties (size and
density) are associated with the valence of emotions shared, and (2) how characteristics of the emotions shared (valence and self-relevance) relate to how the audience responds (quantity of responses, emotional and supportive content in comments, and whether the responses come in network-visible or private channels)."

Question(s) de recherche/Hypothèses/conclusion

Research question : Do people share different kinds of emotions online depending on whom they think is listening?
Does social sharing of emotion online elicit emotion in viewers, and what form do their responses take?

The hypothesis is that people share more positive and negative emotions when their friend networks are smaller and denser, and that hearing about a friend's troubles on Facebook leads to more emotional and supportive comments.

The conclusion are that "the research demonstrates that when people have smaller, denser networks, they share more positive and negative emotions, and friends respond in greater volume to posts with emotion,
especially negative emotion. Responses to posts with negative emotion are more emotional themselves and contain more supportive language; more extreme expressions of negative self-worth (e.g., feeling unloved)
often elicit private responses from friends."

Cadre théorique/Auteur.es

The theoretical framework of the article is based on psychology (Pennebaker), on previous research on social sharing and emotional support, including studies by Rimé, Reis, and Shaver. The authors also draw on theories of social capital and social network analysis.

Concepts clés

Social sharing, Emotional valence, Social capital, Social network analysis

Données collectées (type source)

Large dataset of de-identified text-based status updates on Facebook's servers. One annotated and the other one without feeling annotations (as a baseline).

"We analyzed network properties and status updates by a random sample of people who used the feeling annotation tool. N=1,399,921 English speakers in the U.S.
(Mean age = 32.2, SD=10.9, 79% female) who posted at least one status update in June 2015 that included an emotion (using the annotation widget in Figure 1), had been Facebook users for at least six months, and had at least 20 friends, were randomly selected. Their status updates for one month (approximately 30 million) and friend networks were included in the analysis."

S2 : N=31.7 million de-identified text-based status updates, comprised of a random sample of 14.2 million posts with attached feeling annotations and a random sample of 17.5 million without feeling annotation.

Définition des émotions

Non

Ampleur expérimentation (volume de comptes)

14,5 millions posts annotated

17,5 millions posts not annotated


The sample represented 19.3 million unique posters aged 13-64 (M=29.9)

Technologies associées

Facebook and statistical analysis software.

Mention de l'éthique

Oui?

Finalité communicationnelle

The study found that people share more positive and negative emotions when their friend networks are smaller and denser, and that hearing about a friend's troubles on Facebook leads to more emotional and supportive comments. Posts with positive feelings receive more likes and positive language in comments.