際際滷

際際滷Share a Scribd company logo
TWITTER, POLITICS AND GENDER: WHY
ARE ITALIAN WOMEN POLITICIANS LESS
INTERACTIVE THAN MEN
Stefania Spina

Universit per Stranieri, Perugia, Italia
Twitter and Italian politicians
≒ first tweet of a Member of Parliament: march 2007


≒ December 2012: 70% of Members of Parliament use
 Twitter




                   #Lutwit - Lancaster University,10 - 12 April 2013
Conversationality
≒ Twitter is based on interaction
≒ short, public conversations that share a double audience:
    ≒ Followers
    ≒ pecific users (@mention)



≒ source of change for political discourse (Spina 2012)




                     #Lutwit - Lancaster University,10 - 12 April 2013
Three perspectives
≒ Gender
   ≒ linguistic resources people deploy to present themselves as certain
      kinds of women or men. (Eckert-McConnel-Ginet 2003:5)
≒ Political discourse
   ≒ Chilton (2004):
     ≒ representation in order to be accepted in the political arena (Wodak 2003);
     ≒ interaction

≒ Computer-mediated discourse (Herring 2008)
   ≒ paradigm of interaction: share horizontally dynamic flows of
      conversations, new forms of interpersonal relationships


≒ Gendered attitudes in computer-mediated political discourse

                             #Lutwit - Lancaster University,10 - 12 April 2013
Research questions

≒ Does gender affect the way political actors participate to
 this flow of conversations, with the aim of interacting with
 others and representing themselves as reliable?

≒ Are there different paths women and men take in the
 context of social media interactions to gain a positive
 representation of themselves and of their political role?




                     #Lutwit - Lancaster University,10 - 12 April 2013
Methodology
≒ corpus-based


≒ 24 politicians (12 male and 12 female)


≒ Balanced:
   ≒ 1 leader
   ≒ 6 Members of Parliament;
   ≒ 5 regional and local administrations




                         #Lutwit - Lancaster University,10 - 12 April 2013
TwitteR (R package)
≒ last 3000 tweets
    ≒ sspina <- userTimeline(@sspina, n=3000)


  1: Reply to other tweets in which the author has been mentioned




  2: new conversation through the mention of another user




                        #Lutwit - Lancaster University,10 - 12 April 2013
Corpus composition
≒ 21191 tweets
    ≒ men: 12128
    ≒ women: 9063


≒ 345.000 tokens


≒ Xml annotation (date, author, sex and type of tweet)
≒ Pos-tagging




                     #Lutwit - Lancaster University,10 - 12 April 2013
Analysis of dialogic attitude

1. distribution of tweets that are a response to other tweets
    where the politicians have been mentioned;
2. distribution of tweets that start a new conversation with
    specific users, selected through @mention;
3. gender and the role of addressees;
4. distribution of selected conversational features.




                    #Lutwit - Lancaster University,10 - 12 April 2013
Results: use of @mention




                           per 100 tweets
Results: responses to tweets and new
tweets + @mention




     per 100 tweets         P values < 0.0001
Results: involved linguistic features




 per 100 tweets                P values < 0.0001
Results: informational linguistic features




per 100 tweets                        P values < 0.001
Discussion/1
≒ different strategies taken by women and men within the
 stream of conversations

≒ Conversation: any exchange of messages between two
 or more participants, where the messages that follow bear
 at least minimal relevance to those that preceded or are
 otherwise intended as responses (Herring 2010)




                   #Lutwit - Lancaster University,10 - 12 April 2013
Twitter, politics and gender
Mention
≒ deictic marker of addressivity (Herring, Honeycutt 2009)
    ≒ associated with conversational activity
    ≒ assures coherence to the exchanges, supporting users in tracking
       conversations
    ≒ tweets disrupted by other intervening messages: strategy for
       relating one tweet to another




                        #Lutwit - Lancaster University,10 - 12 April 2013
#Lutwit - Lancaster University,10 - 12 April 2013
tweets that contain a mention
≒ more focused on an
   addressee
≒ more likely to provide
   information for others
≒ more likely to exhort others
   to do something
≒ their content is more
   interactive
tweets without mentions
≒ more self-focused
    ≒ (Herring, Honeycutt 2009)




                         #Lutwit - Lancaster University,10 - 12 April 2013
Discussion/2
≒ gender identity is constructed in interaction (Wodak 2003)
≒ Twitter:
    ≒ A place where we perform our online identities in order to connect
       with others (Zappavigna 2012)
    ≒ each tweet conveys a stance, which reflects the author, the topic,
       and the audience (Bamman et al. 2012)
≒ context of tweets exchanged by politicians:
    ≒ who are the addressees of their messages?
    ≒ who do they answer to?




                        #Lutwit - Lancaster University,10 - 12 April 2013
Male and female interlocutors




per 100 replies                   per 100 new conversations
Roles of addressees
Discussion/3
≒ Involed/informational linguistic features
≒ gendered attitudes in the way politicians manage social
   relations and participate to the flow of conversations on
   Twitter
≒ Data replicates previous findings on online and offline
   gender patterns




                     #Lutwit - Lancaster University,10 - 12 April 2013
Women use more involved features
 ≒ expressive lengthening and puntctuation
     ≒ @November_67 grazieeeeee! // Buongiorno a tutti !!!!


 ≒ Emoticons: more types and different functions

 men                                            women
 :)   :-) ;-) ;) :D :-( :-D                     :) :-) ;-) :( ;) :* :D :P -_- :'D o_o ^___^ :-|   -__-

      ≒ Intensifiers:
          ≒ grazie ;))))
      ≒ Search for conviviality
          ≒ A Trapani in partenza x Strasburgo ! Buon inizio settimana ;))
         ≒ [In Trapani leaving to Strasblourg! Have a good beginning of the week ;)) ]
      ≒ Evaluation
          ≒ I mercati ci amano :-) #benvenutasinistra #iovotoSEL http
         ≒ [Markets love us :-) #benvenutasinistra #iovotoSEL http]
      ≒ 

                                    #Lutwit - Lancaster University,10 - 12 April 2013
Discourse markers
≒ members of a functional class of verbal (and non verbal)
   devices which provide contextual coordinates for ongoing
   talk (Schriffin 1987).
≒ Difference in frequency between men and women: non
   significant. Why?

≒ ma, mica, beh, a quanto pare, ecco.*, scusa[te], appunto,
 in effetti, effettivamente, senti, mah, come no, boh, vabb竪
≒ [well, at all, as far as it seems, here, here you are, sorry, thats it, indeed, in
 fact, as a matter of fact, listen, who knows, sure, dunno, ok]




                             #Lutwit - Lancaster University,10 - 12 April 2013
Comparison between:
≒ Discourse markers used in interactive tweets (with
 mention)
    ≒ @akappa beh, per嘆 ammetterai anche tu che per un profano qualsiasi
      quellarticolo filava abbastanza bene :) cmq mi informer嘆 ;)
    ≒ [@akappa Well, you will admit as well that for someone that is not informed that article made
      sense :) however I will ask about it]



≒ Discourse markers used in tweets without mentions
    ≒ Vabb竪 non ci siamo capiti
    ≒ [Ok/thats fine, you don't seem to have understood.]
Use in dialogic vs. non-dialogic tweets




Women: diffuse conversational style rather than concrete instantiations of
conversations
Conclusions/1
≒ Twitter and politicians: new approach to audience, based
   on horizontal and pervasive forms of conversation;
≒ in a gender perspective: window on how women and men
   construct their gender and political identity within
   conversational flows (Wodak 2003).
≒ different attitudes towards Twitter interactions:
  ≒ men: seeking for interaction and for dialogue and replying to more
     questions
  ≒ women: less oriented to conversation, adopt a more self-focused
     approach




                       #Lutwit - Lancaster University,10 - 12 April 2013
Conclusions/2
≒ Women:
   ≒ Frequency of involved features
   ≒ spread across dialogic and non-dialogic tweets
≒ search for informal and convivial styles, and attitude of
   personal involvement.
≒ conversationalization and informalization of public
   discourse (Fairclough 1994): different strategy compared
   to the instantiation of real conversations.
≒ Twitter double audience:
  ≒ women address more to the general audience of followers
    modelling their public discourse upon the conversational practices
    of ordinary life (Faiclough 1994:253)

                        #Lutwit - Lancaster University,10 - 12 April 2013
Conclusions /3
≒ Indirectness
≒ Conformity to traditional styles
≒ Lack of competitiveness and authority


≒  but mostly
≒ Lack of willingness to interact with others


≒ Further research needed:
    ≒ More data (more politicians)
    ≒ Deeper analysis of different discursive practices:
     ≒ When interacting with male/female users


                         #Lutwit - Lancaster University,10 - 12 April 2013
Thank you for your attention!

          @sspina
References
≒ Bamman, D., Eisenstein, J., Schnoebelen, T. (2012). Gender in Twitter: Styles, stances, and
     social networks. eprint arXiv:1210.4567.
≒   Eckert, P., McConnel-Ginet, S. (2003). Language and Gender. Cambridge: Cambridge
     University Press.
≒   Fairclough, N. (1994). Conversationalization of public discourse and the authority of the
     consumer. In R. Keat & N. Whitely & N. Abercrombie (Eds.), The authority of the consumer.
     London: Routledge.
≒   Herring, S. C. (2008). Computer-Mediated Discourse. In D. Schiffrin, D. Tannen, & H. E.
     Hamilton (Eds.), The Handbook of Discourse Analysis (pp. 612-634). Malden, MA: Blackwell.
≒   Honeycutt, C., Herring, S. C. (2009). Beyond Microblogging: Conversation and Collaboration
     via Twitter. Proceedings of the Forty-Second Hawaii International Conference on System
     Sciences (HICSS-42). Los Alamitos, CA: IEEE Press.
≒   Schriffin, D. (1987). Discourse Markers. Cambridge: Cambridge University Press.
≒   Spina, S. (2012). Openpolitica. Il discorso dei politici italiani nellera di Twitter. Milano:
     FrancoAngeli.
≒   Wodak, R. (2003). Multiple Identities: The Roles of Female Parliamentarians in the EU
     Parliament. In Holmes-Meyerhoff. The Handbook of Language and Gender, 671-698.
     Malden: Blackwell.
≒   Zappavigna, M. (2012). The Discourse of Twitter and Social Media. How We Use Language
     to Create Affiliation on the Web. London: Continuum.
Second person pronouns
Evolution of replies
Age and replies
Exception: links

Men            women
36,4%          65,1%

move the interaction to other non-
dialogic spaces outside the social
network (blog, newspapers articles)
CMC abbreviations
nn = non (not); prox = prossimo (next); msg =
messaggio (message)

More Related Content

Twitter, politics and gender

  • 1. TWITTER, POLITICS AND GENDER: WHY ARE ITALIAN WOMEN POLITICIANS LESS INTERACTIVE THAN MEN Stefania Spina Universit per Stranieri, Perugia, Italia
  • 2. Twitter and Italian politicians ≒ first tweet of a Member of Parliament: march 2007 ≒ December 2012: 70% of Members of Parliament use Twitter #Lutwit - Lancaster University,10 - 12 April 2013
  • 3. Conversationality ≒ Twitter is based on interaction ≒ short, public conversations that share a double audience: ≒ Followers ≒ pecific users (@mention) ≒ source of change for political discourse (Spina 2012) #Lutwit - Lancaster University,10 - 12 April 2013
  • 4. Three perspectives ≒ Gender ≒ linguistic resources people deploy to present themselves as certain kinds of women or men. (Eckert-McConnel-Ginet 2003:5) ≒ Political discourse ≒ Chilton (2004): ≒ representation in order to be accepted in the political arena (Wodak 2003); ≒ interaction ≒ Computer-mediated discourse (Herring 2008) ≒ paradigm of interaction: share horizontally dynamic flows of conversations, new forms of interpersonal relationships ≒ Gendered attitudes in computer-mediated political discourse #Lutwit - Lancaster University,10 - 12 April 2013
  • 5. Research questions ≒ Does gender affect the way political actors participate to this flow of conversations, with the aim of interacting with others and representing themselves as reliable? ≒ Are there different paths women and men take in the context of social media interactions to gain a positive representation of themselves and of their political role? #Lutwit - Lancaster University,10 - 12 April 2013
  • 6. Methodology ≒ corpus-based ≒ 24 politicians (12 male and 12 female) ≒ Balanced: ≒ 1 leader ≒ 6 Members of Parliament; ≒ 5 regional and local administrations #Lutwit - Lancaster University,10 - 12 April 2013
  • 7. TwitteR (R package) ≒ last 3000 tweets ≒ sspina <- userTimeline(@sspina, n=3000) 1: Reply to other tweets in which the author has been mentioned 2: new conversation through the mention of another user #Lutwit - Lancaster University,10 - 12 April 2013
  • 8. Corpus composition ≒ 21191 tweets ≒ men: 12128 ≒ women: 9063 ≒ 345.000 tokens ≒ Xml annotation (date, author, sex and type of tweet) ≒ Pos-tagging #Lutwit - Lancaster University,10 - 12 April 2013
  • 9. Analysis of dialogic attitude 1. distribution of tweets that are a response to other tweets where the politicians have been mentioned; 2. distribution of tweets that start a new conversation with specific users, selected through @mention; 3. gender and the role of addressees; 4. distribution of selected conversational features. #Lutwit - Lancaster University,10 - 12 April 2013
  • 10. Results: use of @mention per 100 tweets
  • 11. Results: responses to tweets and new tweets + @mention per 100 tweets P values < 0.0001
  • 12. Results: involved linguistic features per 100 tweets P values < 0.0001
  • 13. Results: informational linguistic features per 100 tweets P values < 0.001
  • 14. Discussion/1 ≒ different strategies taken by women and men within the stream of conversations ≒ Conversation: any exchange of messages between two or more participants, where the messages that follow bear at least minimal relevance to those that preceded or are otherwise intended as responses (Herring 2010) #Lutwit - Lancaster University,10 - 12 April 2013
  • 16. Mention ≒ deictic marker of addressivity (Herring, Honeycutt 2009) ≒ associated with conversational activity ≒ assures coherence to the exchanges, supporting users in tracking conversations ≒ tweets disrupted by other intervening messages: strategy for relating one tweet to another #Lutwit - Lancaster University,10 - 12 April 2013
  • 17. #Lutwit - Lancaster University,10 - 12 April 2013
  • 18. tweets that contain a mention ≒ more focused on an addressee ≒ more likely to provide information for others ≒ more likely to exhort others to do something ≒ their content is more interactive tweets without mentions ≒ more self-focused ≒ (Herring, Honeycutt 2009) #Lutwit - Lancaster University,10 - 12 April 2013
  • 19. Discussion/2 ≒ gender identity is constructed in interaction (Wodak 2003) ≒ Twitter: ≒ A place where we perform our online identities in order to connect with others (Zappavigna 2012) ≒ each tweet conveys a stance, which reflects the author, the topic, and the audience (Bamman et al. 2012) ≒ context of tweets exchanged by politicians: ≒ who are the addressees of their messages? ≒ who do they answer to? #Lutwit - Lancaster University,10 - 12 April 2013
  • 20. Male and female interlocutors per 100 replies per 100 new conversations
  • 22. Discussion/3 ≒ Involed/informational linguistic features ≒ gendered attitudes in the way politicians manage social relations and participate to the flow of conversations on Twitter ≒ Data replicates previous findings on online and offline gender patterns #Lutwit - Lancaster University,10 - 12 April 2013
  • 23. Women use more involved features ≒ expressive lengthening and puntctuation ≒ @November_67 grazieeeeee! // Buongiorno a tutti !!!! ≒ Emoticons: more types and different functions men women :) :-) ;-) ;) :D :-( :-D :) :-) ;-) :( ;) :* :D :P -_- :'D o_o ^___^ :-| -__- ≒ Intensifiers: ≒ grazie ;)))) ≒ Search for conviviality ≒ A Trapani in partenza x Strasburgo ! Buon inizio settimana ;)) ≒ [In Trapani leaving to Strasblourg! Have a good beginning of the week ;)) ] ≒ Evaluation ≒ I mercati ci amano :-) #benvenutasinistra #iovotoSEL http ≒ [Markets love us :-) #benvenutasinistra #iovotoSEL http] ≒ #Lutwit - Lancaster University,10 - 12 April 2013
  • 24. Discourse markers ≒ members of a functional class of verbal (and non verbal) devices which provide contextual coordinates for ongoing talk (Schriffin 1987). ≒ Difference in frequency between men and women: non significant. Why? ≒ ma, mica, beh, a quanto pare, ecco.*, scusa[te], appunto, in effetti, effettivamente, senti, mah, come no, boh, vabb竪 ≒ [well, at all, as far as it seems, here, here you are, sorry, thats it, indeed, in fact, as a matter of fact, listen, who knows, sure, dunno, ok] #Lutwit - Lancaster University,10 - 12 April 2013
  • 25. Comparison between: ≒ Discourse markers used in interactive tweets (with mention) ≒ @akappa beh, per嘆 ammetterai anche tu che per un profano qualsiasi quellarticolo filava abbastanza bene :) cmq mi informer嘆 ;) ≒ [@akappa Well, you will admit as well that for someone that is not informed that article made sense :) however I will ask about it] ≒ Discourse markers used in tweets without mentions ≒ Vabb竪 non ci siamo capiti ≒ [Ok/thats fine, you don't seem to have understood.]
  • 26. Use in dialogic vs. non-dialogic tweets Women: diffuse conversational style rather than concrete instantiations of conversations
  • 27. Conclusions/1 ≒ Twitter and politicians: new approach to audience, based on horizontal and pervasive forms of conversation; ≒ in a gender perspective: window on how women and men construct their gender and political identity within conversational flows (Wodak 2003). ≒ different attitudes towards Twitter interactions: ≒ men: seeking for interaction and for dialogue and replying to more questions ≒ women: less oriented to conversation, adopt a more self-focused approach #Lutwit - Lancaster University,10 - 12 April 2013
  • 28. Conclusions/2 ≒ Women: ≒ Frequency of involved features ≒ spread across dialogic and non-dialogic tweets ≒ search for informal and convivial styles, and attitude of personal involvement. ≒ conversationalization and informalization of public discourse (Fairclough 1994): different strategy compared to the instantiation of real conversations. ≒ Twitter double audience: ≒ women address more to the general audience of followers modelling their public discourse upon the conversational practices of ordinary life (Faiclough 1994:253) #Lutwit - Lancaster University,10 - 12 April 2013
  • 29. Conclusions /3 ≒ Indirectness ≒ Conformity to traditional styles ≒ Lack of competitiveness and authority ≒ but mostly ≒ Lack of willingness to interact with others ≒ Further research needed: ≒ More data (more politicians) ≒ Deeper analysis of different discursive practices: ≒ When interacting with male/female users #Lutwit - Lancaster University,10 - 12 April 2013
  • 30. Thank you for your attention! @sspina
  • 31. References ≒ Bamman, D., Eisenstein, J., Schnoebelen, T. (2012). Gender in Twitter: Styles, stances, and social networks. eprint arXiv:1210.4567. ≒ Eckert, P., McConnel-Ginet, S. (2003). Language and Gender. Cambridge: Cambridge University Press. ≒ Fairclough, N. (1994). Conversationalization of public discourse and the authority of the consumer. In R. Keat & N. Whitely & N. Abercrombie (Eds.), The authority of the consumer. London: Routledge. ≒ Herring, S. C. (2008). Computer-Mediated Discourse. In D. Schiffrin, D. Tannen, & H. E. Hamilton (Eds.), The Handbook of Discourse Analysis (pp. 612-634). Malden, MA: Blackwell. ≒ Honeycutt, C., Herring, S. C. (2009). Beyond Microblogging: Conversation and Collaboration via Twitter. Proceedings of the Forty-Second Hawaii International Conference on System Sciences (HICSS-42). Los Alamitos, CA: IEEE Press. ≒ Schriffin, D. (1987). Discourse Markers. Cambridge: Cambridge University Press. ≒ Spina, S. (2012). Openpolitica. Il discorso dei politici italiani nellera di Twitter. Milano: FrancoAngeli. ≒ Wodak, R. (2003). Multiple Identities: The Roles of Female Parliamentarians in the EU Parliament. In Holmes-Meyerhoff. The Handbook of Language and Gender, 671-698. Malden: Blackwell. ≒ Zappavigna, M. (2012). The Discourse of Twitter and Social Media. How We Use Language to Create Affiliation on the Web. London: Continuum.
  • 35. Exception: links Men women 36,4% 65,1% move the interaction to other non- dialogic spaces outside the social network (blog, newspapers articles)
  • 36. CMC abbreviations nn = non (not); prox = prossimo (next); msg = messaggio (message)