Presentation describing results of Pawel Kuczma, and Wlodzimierz Gogolek, Institute of Journalism, University of Warsaw presented on General Online Research Conference - GOR 11, March 14-16, 2011, Heinrich Heine University, Düsseldorf, Germany
Sociocosmos empowers you to go trendy on social media with a few clicks..pdf
Social Media Potential in Forecasting Presidential Election Results in Poland 2011
1. General Online Research Conference
GOR 11, March 14-16, 2011, Heinrich Heine University, Düsseldorf,
Germany
Pawel Kuczma, Institute of Journalism, University of Warsaw
Wlodzimierz Gogolek, Institute of Journalism, University of Warsaw
Social Media Potential in Forecasting Presidential Election Results in
Poland 2010
Contact: p.kuczma@id.uw.edu.pl
2. Social Media Potential in
Forecasting Presidential Election
Results in Poland 2010
Paweł Kuczma, Włodzimierz Gogołek
Institute of Journalism
University of Warsaw
Graphic source: http://jonnewman12.files.wordpress.com/2010/05/sm-crystal-ball.png
3. „Markets collect and diffuse information”
Hayek F., ’The use of knowledge in society”, American Economic Review, 1945;35 (4): 519–530/ Chen K.Y., Fine L.R., Huberman
B.A.,’Predicting the Future’, Information Systems Frontiers 5:1, 47–61, 2003
„What consumers are searching for online can also
predict their collective future behavior days or even
weeks in advance.”
Goel, S.; Hofman, J.M.; Lahaie, S.; Pennock, D.M.; Watts, D.J: Predicting consumer behavior with Web search,
http://research.yahoo.com/pub/3359 [2010]/
Inspirations
5. The purpose of this study:
to identify factors allowing prediction of the outcome of the presidential election
in Poland in 2010 basing on Social Media websites in pre-election period.
The research question was:
Is it possible to predict the action (in this case casting a vote for a candidate
in the presidential election) basing on content quantitative (number of
content related to the subject of the research) and qualitative (the
contexts in which they appear and their emotional values) analysis of
Social Media?
Study Details - Methodology
6. Study Details - Methodology
Type of content analysed:
Social Media websites (such as social networking sites, forums, blogs and microblogs)
News portals (websites with content written by professionals) for comparative analysis
Time of the study: Analysed content was published in pre-election period in 2010
(between April 10th
and July 5th
)
The following indicators of the content were examined :
Quantitative assessment:
- Amount of the content about candidates,
- Trends / dynamics of changes in the amount of the content,;
Qualitative assessment:
- Contexts analysis (finding topic content),
- Sentiment analysis (distinction between positive and negative content).
8. Qualitative potential of Social Media
2/3 of all references regarding presidential candidates in pre-election period
was generated in Social Media
Portale
Informacyjne
33%
Media
społecznościowe
67%
Social Media 67%
News Portals 33%
9. Share of voice by different types of Social Media
websites
poszczególnych rodzajów mediów
społecznościowych
Blip
0,8%
Komentarze
0,3%
Twitter
0,3%
Fora
21,8%
Blogi
74,1%
Facebook
2,7%
Most important sources were Blogs and Forums – 96% of content
Facebook was not very strong at that time (around 2 mil. users)
Forums
21,8%
Facebook
2,7%
Blip
0,8%
Comments
0,3%
Blogs
74,1%
12. More content regarding Bronislaw Komorowski
than Jaroslaw Kaczynski in news portals
Bogusław Ziętek
1%
Jarosław Kaczyński
34%
Bronislaw
Komorowski
36%
Kornel Morawiecki
1%
Waldemar Pawlak
7%
Andrzej Olechowski
4%
Janusz Korwin-
Mikke
3%
Grzegorz
Napieralski
9%
Andrzej Lepper
2%
Marek Jurek
3%
13. More content regarding Jaroslaw Kaczynski than
Bronislaw Komorowski in Social Media
Portale informacyjne 10.04-4.07.2010
Bogusław Ziętek
1%
Jarosław Kaczyński
34%
Bronislaw
Komorowski
36%
Kornel Morawiecki
1%
Waldemar Pawlak
7%
Andrzej Olechowski
4%
Janusz Korwin-
Mikke
3%
Grzegorz
Napieralski
9%
Andrzej Lepper
2%
Marek Jurek
3%
Social Media 10.04-4.07.2010
Grzegorz Napieralski
7%
Bronisław Komorowski
32%
Jarosław Kaczyński
41%
Andrzej Lepper
2%
Kornel Morawiecki
1% Boguslaw Zietek
1%
Marek Jurek
3%Waldemar Pawlak
5%
Janusz Korwin-Mikke
5%
Andrzej Olechowski
3%
Janusz Korwin-Mikke unexpectedly strong in Social Media. 4th place in the
amount of content and 4th place in election. This candidate was almost ignored
by mainstrem news portals.
14. There was much more content regarding two
candidates than all others in news portals …
0
500
1000
1500
2
010-04
-1
0
201
0-04
-1
7
201
0-04
-2
4
2
010-05
-0
1
201
0-05
-0
8
20
10-05
-1
5
20
10-05
-2
2
2
010-05
-2
9
20
10-06
-0
5
20
10-06
-1
2
2
010-06
-1
9
2
010-06
-2
6
2
010-07
-0
3
Andrzej Lepper Andrzej Olechowski Boguslaw Zietek Bronislaw Komorowski Grzegorz Napieralski
Janusz Korwin-Mikke Jaroslaw Kaczynski Kornel Morawiecki Marek Jurek Waldemar Pawlak
Election silence
# of pieces of content
15. …as well as in Social Media
Social Media All times
0
500
1000
1500
2000
2500
2010-04-10
2010-04-17
2010-04-24
2010-05-01
2010-05-08
2010-05-15
2010-05-22
2010-05-29
2010-06-05
2010-06-12
2010-06-19
2010-06-26
2010-07-03
Andrzej Lepper Andrzej Olechowski Boguslaw Zietek Bronislaw Komorowski Grzegorz Napieralski
Janusz Korwin-Mikke Jaroslaw Kaczynski Kornel Morawiecki Marek Jurek Waldemar Pawlak
Content regarding these two candidates prevailed: Jaroslaw Kaczynski and
Bronislaw Komorowski
# of pieces of content
16. The closer to the election day, the more content
appeared
The greatest intensity of content concerning presidential candidates
could be observed in June
1 243
446
1 544
658
2 942
1 587
2 627
1 473
0
500
1 000
1 500
2 000
2 500
3 000
Kwiecień Maj Czerwiec Lipiec
Media społecznościowe PortaleSocial Media News Portals
# of pieces of content
18. News Portals
Udział odniesień tematycznie 10.04-4.07.2010 - portale
informacyjne
Wybory 22,41%
Partie 21,59%
Polityka wewnętrzna
15,31%
Prezydent 15,27%
Katastrofa 4,29%
Gospodarka 4,98%
Media 8,79%
Polityka zagraniczna
3,53%
Powódź 2,13%
Rosja 1,71%
Smolensk Crash 4,29%
Foreign Policy
3,53%
Flood 2,12%
Russia 1,71%
Election 22,41%
Home Affairs
15,32%
Economy 4,98%
President 15,27%
Media
8,79%
Political Parties
21,59%
19. Social MediaLiczba odniesień (%) w mediach społecznościowych 10.04-
4.07.2010
Media
10,88%
Polityka
zagraniczna
5,39%
Wybory
18,30%
Prezydent
13,73%
Gospodarka
5,20%
Rosja
3,45%
Powódź
2,23%
Partie
19,08%
Katastrofa
6,11%
Polityka
wewnętrzna
15,63%
Smolensk Crash
6,11%
Foreign Policy
5,39%
Flood 2,23%
Russia
3,45%
Election
18,30%
Political Parties
19,08%
Home Affairs
15,63%
Economy
5,20%
President 13,73%
Media
10,88%
20. The intensity of the content in pre-election period
In pre-election period in Social Media the biggest number of pieces of
content concerned topics associated with political events at that time.
Smolensk plane crash and the flood in Poland did not eclipse pre-election
period. Foreign policy was hardly visible.
Liczba odniesień w mediach społecznościowych tematycznie [podział tygodniowy] 10.04-
4.07.2010
0
1 000
2 000
3 000
4 000
5 000
6 000
14 15 16 17 18 19 20 21 22 23 24 25 26 Tygodnie
Liczba
odniesień
Gospodarka Katastrofa Media
Partie Polityka wewnetrzna Polityka zagraniczna
Powodz Prezydent Rosja
Wybory
# of pieces of content
21. Social Media vs. News Portals: share of voice per
candidate
Udział treści dotyczących kandydatów w podziale na źródło [%]
68% 62% 56% 65% 61%
78% 71% 65% 70% 60%
32% 38% 44% 35% 39%
22% 29% 35% 30% 40%
0%
20%
40%
60%
80%
100%
A
ndrzejLepper
A
ndrzejO
lechow
ski
B
oguslaw
Zietek
B
ronislaw
K
om
orow
ski
G
rzegorz
N
apieralski
Janusz
K
orw
in-M
ikke
Jaroslaw
K
aczynski
K
ornelM
oraw
iecki
M
arek
Jurek
W
aldem
ar
P
aw
lak
Media społecznościowe Portale informacyjne
• Janusz Korwin-Mikke had the biggest share of content in Social Media sources
• Waldemar Pawlak (deputy prime minister) had very low share of content in Social Media
Social Media News Portals
22. Sentiment analysis in Social Media (in numbers)
Wydźwięk treści wiodących kandydatów w mediach społecznościowych 5.05-4.07
11 518
14 753
1 015835
2 245
1 730
0
2 000
4 000
6 000
8 000
10 000
12 000
14 000
16 000
18 000
20 000
Bronisław Komorowski Jarosław Kaczyński Kandydaci
Liczba
odniesień
Negative Neutral Positive
# of pieces of content
Candidates
Data gathered: 5.05-4.07.2010
23. Wydźwięk treści wiodących kandydatów w mediach społecznościowych
5.05-4.07
5,93% 5,63%
81,79% 81,90%
12,46%12,28%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Bronisław Komorowski Jarosław Kaczyński
Kandydaci
Liczba
odniesień (%)
negatywny neutralny pozytywny
Sentiment analysis in Social Media (%)
Candidates
Share of content
Negative Neutral Positive
Data gathered: 5.05-4.07.2010
24. 0
50
100
150
200
250
300
350
400
450
500
18 19 20 21 22 23 24 25 26 27 Tydzień
Liczba
odniesień
Jarosław Kaczyński Bronisław Komorowski
Positive content
Wydźwięk pozytywny w portalach informacyjnych
[podział tygodniowy]
0
200
400
600
800
18 19 20 21 22 23 24 25 26 27 Tydzień
Liczba
odniesień
Jarosław Kaczyński Bronisław Komorowski
Social Media
News Portals
# of pieces of content
# of pieces of content
Weeks
Weeks
Data gathered: 5.05-4.07.2010
25. 0
50
100
150
200
250
300
18 19 20 21 22 23 24 25 26 27 Tydzień
Liczba
odniesień
Jarosław Kaczyński Bronisław Komorowski
Wydźwięk negatywny w portalach informacyjnych
[podział tygodniowy]
0
50
100
150
200
250
18 19 20 21 22 23 24 25 26 27 Tygodnie
Liczba
odniesień
Jarosław Kaczyński Bronisław Komorowski
Negative content
Social Media
News Portals
Weeks
Weeks
# of pieces of content
# of pieces of content
Data gathered: 5.05-4.07.2010
26. Correlation between candidates and campaign topics
on Blogs
Source: Attentio.com, http://www.youtube.com/watch?gl=US&v=v0k0DWbddX8
27. Research Results
• Social Media is an extremely valuable source of information which
reflects public opinion - including those relating to social and
political phenomena, what is confirmed by this study.
• Although the hypothesis concerning the possibility of predicting
election results was not definitely proven - study helped to provide
the names of candidates who qualified for the second round of the
election.
• Moreover, it shows the importance of certain Web 2.0 forms in
terms of providing information and their competitiveness against
the traditional Internet resources.
28. Next Steps
This research is the first step towards creating a method supporting the
diagnosis of the condition and dynamics of changes of candidates/parties
/ideas preferences. Therefore it can be used to influence democratic
processes with the use of Social Media.