Main-01.png Read-01.png Recipes-01.png Tactics-01.png Cases-01.png Tools-01.png

Tactic: Issue Profiling



Detecting issues profiles of pro- and anti-camps in IP enforcement debates

Tomasso Venturini, Bernard Rieder, Vera Franz, Becky Hogge, Erik Borra, Noortje Marres


Question:
Which issue terms are specific to pro- and anti-camps in the SOPA/PIPA/ACTA/TPPA issue spaces?

Issue Advocacy Context

An issue on which a small group of people has suddenly come to prominence (SOPA/ACTA). Questions arise: what language of the issue drew the public in (issue uptake)?

1. The pro reform (anti SOPA)-camp has language difficulties: their issue terms are too complicated (stealing? people understand; Access to knowledge: what is that?). Which progressive terms have been successful and have appealed in public space?
2. Will we be able to detect changes in language on the anti-reform (pro SOPA) camp? This is important for future messages.
3. Monitoring the fate pro-reform and anti-reform issue language in real-time
If we can monitor in real-time the fate of pro-reform and anti-reform issue terms - then we have a good early warning system.

These questions require issue profiling, because we need to know pro and con issue language, their resonance (uptake), and their resonance over time. This would be to inform communication strategies of civil society organisations.

Issue Mapping Method
1. Collects tweets using a general issue tag (ex. #sopa)

2. Collecting seed URLs from experts

3. Enter seed URLs into Issue Crawler to draw the hyperlink network

4. Use a clustering algorithm to extract camps from the network and classify actors (represented by their URLs) as being pro & anti & undecided

5. Allocate tweets to camps based on their inclusion of a URL from a specific camp

6. Extracts words and related htags from tweets

7. Compute the fraction of each word (or hashtag) belonging to a specific camp.

8. Rank words by ‘polarizing power’ (the fact of being used almost exclusively by one camp)

9. Use the most polarizing words detect the leaning of other texts (ex. newspaper article)

Biased issue language:
stealing versus censorship

Data Collection

PRO URLs
www.eff.org
www.laquadrature.net
www.edri.org
www.openrightsgroup.org
www.tacd.org
panoptykon.org/
www.keionline.org
www.publicknowledge.org
www.michaelgeist.ca/
www.americancensorship.or g
www.fightforthefuture.org

CON URLs
http://www.riaa.com/
http://www.mpaa.org/
http://www.internationalpublishers.org
http://ec.europa.eu/trade /
http://www.ustr.gov/
http://www.creativeamerica.org/

sopa_linegraph_all_steal_censor.png
Topic revision: r3 - 22 May 2012 - 15:10:29 - NoortjeMarres