Emotions and dialogue in a peer-production community: the case of Wikipedia

From Brede Wiki
Jump to: navigation, search
Conference paper (help)
Emotions and dialogue in a peer-production community: the case of Wikipedia
Authors: David Laniado, Carlos Castillo, Andreas Kaltenbrunner, Mayo Fuster Morell
Citation: WikiSym '12 - 8th International Symposium on Wikis and Open Collaboration, Linz, Austria  : 2012
Publisher: Define publisher
Meeting: WikiSym'12
Database(s): Google Scholar cites
DOI: Define doi.
Link(s): http://airwiki.ws.dei.polimi.it/images/b/b7/Emotions_and_dialogue_in_Wikipedia_2012.pdf
Web: DuckDuckGo Bing Google Yahoo!Google PDF
Article: Google Scholar PubMed
Restricted: DTU Digital Library
Format: BibTeX

Emotions and dialogue in a peer-production community: the case of Wikipedia is a study with text sentiment analysis on Wikipedia. They study the discussion on the "talk" pages and user talk pages.

The paper is briefly reviewed in the Wikimedia Research Newsletter [1] and labeled on Wikilit [2].

Slides are available from Slideshare [3].


[edit] Data

English Wikipedia where 871.485 was article talk pages (ATP) with a total of over 11 million comments.

Over 12.000 users had more than 100 comments on ATP.

[edit] Methods

  • Gender labeling
    • with Wikipedia API for around 2.000 of the around 12.000 prolific ATP
    • 1385 users labeled with Crowdflower.
  • Text sentiment analysis with ANEW: both valence, arousal and dominance.
  • There was a low correlation (r=0.10, p<10-8) with the logarithm of the edit count and the "dominance"
  • Admins were slightly more positive than non-admins. (Though no difference between female admins and non-admins)
  • Women had higher valence and dominance.
  • There is a difference in valence between different topic categories, e.g., computing and arts have high valence while politics, history and events has low. (Figure 4)
  • Females and admins would more often add links to Wikipedia policies.
  • Females write longer messages, especially female admins.
  • Answers tend to be more positive and with more dominance.
  • Homophily: (slide 29)
    • Gender: Females tend to interact with other females.
    • Users tend to interact more with users expressing similar emotions.
  • Emotions on the own personal page:
    • Women express more emotion: higher valence, arousal, dominance
    • Male admins more positive, dominance but less arousal.
    • Comments received on personal pages more positive

[edit] Results

For only 50% of the users the gender could be identified.

[edit] Related papers

  1. Increasing the willingness to collaborate online: an analysis of sentiment-driven interactions in peer content production

[edit] External links

  1. Wikimedia Research Newsletter, July 2012
Personal tools