When a Movement Becomes a Party: Computational Assessment of New Forms of Political Organization in Social Media

When a Movement Becomes a Party
Computational Assessment of New Forms of
Political Organization in Social Media
Pablo Aragón*†, Yana Volkovich†, David Laniado†, Andreas Kaltenbrunner†
* Universitat Pompeu Fabra † Eurecat
The 10th International AAAI Conference on Web and Social Media (ICWSM-16)
Cologne, Germany, May 17-20

Motivation: Movement organizations
2
Networked social movement: Networks in multiple forms (multimodal, on/offline,
across platforms) without a central node, and with a decentralized structure.
(Castells, 2013)
Change from logic of collective action to a logic of connective action.
(Bennett et al, 2013)

Motivation: Movement networks
3
“Decentralized structure, based on
coalitions of smaller organizations”.
(González-Bailón et al, 2011)
“Decentralized organization without
stable leaders or representatives”.
(Aragón et al, 2013)
RT network of the 15M movement
May 15-22, 2011 (Aragón et al, 2015)

Motivation: Party organizations
4
Iron Law of Oligarchy: Political parties, like any complex organization, self-generate
an elite (“Who says organization, says oligarchy”).
(Michels, 1915)
Elite theory: Small minorities (elites) hold the most power in political processes.
(Pareto et al, 1935; Mosca, 1939; Mills 1999)

Motivation: Party networks
5
The Twitter party networks in the
2011 Spanish election presented:
 Isolated clusters for each party.
 Every party cluster was strongly
centralized around candidate
and/or party accounts.
(Aragón et al, 2013)
RT network of political parties in
the 2011 Spanish election (Aragón et al, 2013)

Motivation: The 2015 Spanish local elections
6
Grassroots parties emerged from
the 15M movement:
 Barcelona en Comú
 Ahora Madrid
 Zaragoza en Común
 Marea Atlántica
 Compostela Aberta
 Por Cádiz Sí se puede
 Guanyem Badalona en Comú

Research Question
7
Assuming that:
 Barcelona en Comú emerged from the 15M movement
 the 15M movement followed a decentralized structure
Has Barcelona en Comú…
 preserved a decentralized structure?
 adopted a conventional centralized organization?

Dataset
8
“Political parties share some interesting patterns of behavior, but also exhibit some
unique and interesting idiosyncrasies” (e.g. tagging practice of politicians)
(Lietz et al, 2015)
Sampling criteria
based on candidate
and party accounts:
 373 818 RTs
 RT network
- 6 492 nodes
- 16 775 edges

Computational assessment
9
Community detection
Identify the organization of nodes in clusters: political party networks.
Cluster characterization
Characterize the topology of the intra-network of each cluster.

First result with the Louvain Method (Blondel et al, 2008):
 Eight major clusters (seven parties)
 Every cluster contains some media accounts: media build weak ties
 Analysis of the ego-network of relevant media accounts:
 Public TV account retweeted by users from every cluster
 Private media mostly retweeted by users from like-minded parties
Community detection
10

Every cluster contains some media accounts
Each execution produces different results:
Some media do not always belong to the same cluster
Community detection
11

Community detection
12

Community detection
13

Community detection
14

We want the real intra-network structure of parties
Community detection
15

We want the real intra-network structure of parties
Louvain Method with Confidence Interval
 Run multiple executions (N=100)
 Validate the stability of major clusters
 Just consider nodes that appear in the same cluster more than times (1-ε=0.95)
Community detection
16

Community detection
17
Result with the Louvain Method with Confidence Interval
 Constant presence of eight major
clusters (seven parties) along the
100 executions

Community detection
18
100 executions
 Most media accounts
are now excluded from
major clusters

Community detection
19
100 executions
major clusters
 Two clusters for
Barcelona en Comú …

Community detection
20
100 executions
major clusters
 Two clusters for
Barcelona en Comú …

Movement-party structure
21
Party
Movement
Visualization of the two clusters of Barcelona en Comú

Cluster characterization
22
Inspired by the social dimensions of García et al. (2015):
 Hierarchical structure
In-degree centralization  Gini coefficient of the in-degree distribution
 Small world phenomenon (f.k.a. information efficiency)
Avg. path length + Clustering coefficient
 Coreness (f.k.a. social resilience)
Maximal k-core  Distribution of k-indices

Cluster characterization: Hierachical structure
23
Gini coefficient of the in-degree distribution

Cluster characterization: Small world
24

Cluster characterization: Coreness
25

Conclusions
26
For Barcelona en Comú, two paradigms co-exist:
 A centralized and low resilient party cluster
 A decentralized and resilient movement cluster
Polarized scenario likes previous studies of election campaigns on Twitter
Data preparation process accentuated the polarization effect
Media accounts build weak ties between clusters
Public media became more plural than private media

Open questions and future work
27
How did the dual network structure Barcelona en Comú was built over time?
 As the result of the confluence of minor parties and the 15M activists?
 As a party interface over a citizen decentralized system?
Is this dual paradigm observable in other grassroots parties?
Ahora Madrid? Zaragoza en Comú? … Syriza?
Is this dual paradigm observable in other OSNs?
Facebook? Youtube? Instagram?

References
28
 Castells, M. (2013). Networks of outrage and hope: Social movements in the Internet age.
J. Wiley & Son
 Bennett, W. L. and Segerberg, A. (2012). The logic of connective action: Digital media and the
personalization of contentious politics. Information, Communication & Society, 15(5):739–
768.
 González-Bailón, S., Borge-Holthoefer, J., Rivero, A., and Moreno, Y. (2011). The dynamics of
protest recruitment through an online network. Scientific reports, 1.
 Aragón P., Congosto M., & Laniado D. (2013). Evolución del sistema- red 15m a través de
topología de redes. In Toret, J., Calleja, A., Marín, O., Aragón, P., Aguilera, M., Barandarian, X.,
Lumbreras, A. & Monterde, A. (2015). Tecnopolítica y 15M. La potencia de las multitudes
conectadas, Barcelona: Editorial UOC. ISBN: 978-84-9064-458-4.
 Aragón, P., Kappler, K. E., Kaltenbrunner, A., Laniado, D., and Volkovich, Y. (2013).
Communication dynamics in twitter during political campaigns: The case of the 2011 spanish
national election. Policy & Internet, 5(2):183–206.

References
29
 Michels, R. (1915). Political parties: A sociological study of the oligarchical tendencies of
modern democracy. Hearst’s International Library Company.
 Pareto, V., Livingston, A., Bongiorno, A., Rogers, J. H., et al. (1935). Mind and society
 Mosca, G. (1939). The ruling class: elementi di scienza politica.
 Mills, C. W. (1999). The power elite. Oxford U. Press.
 Lietz, H., Wagner, C., Bleier, A., & Strohmaier, M. (2014). When Politicians Talk: Assessing
Online Conversational Practices of Political Parties on Twitter.
 Blondel, V. D., Guillaume, J.-L., Lambiotte, R., and Lefebvre, E. (2008). Fast unfolding of
communities in large networks. Journal of Statistical Mechanics: Theory and Experiment,
2008(10):P10008.
 Garcia, D., Abisheva, A., Schweighofer, S., Serdult, U., and Schweitzer, F. (2015). Ideological
and temporal components of network polarization in online political participatory media.
Policy & Internet, 7(1):46–79.

Thanks for your attention
Questions?

When a Movement Becomes a Party: Computational Assessment of New Forms of Political Organization in Social Media

Recomendados

Recomendados

Mais conteúdo relacionado

Mais procurados

Mais procurados (18)

Destaque

Destaque (17)

Semelhante a When a Movement Becomes a Party: Computational Assessment of New Forms of Political Organization in Social Media

Semelhante a When a Movement Becomes a Party: Computational Assessment of New Forms of Political Organization in Social Media (20)

Mais de Pablo Aragón

Mais de Pablo Aragón (16)

Último

Último (20)

When a Movement Becomes a Party: Computational Assessment of New Forms of Political Organization in Social Media