EU COHESION POLICY IN THE PUBLIC SPHERE: HOW DO THE MEDIA FRAME EU - PowerPoint PPT Presentation

EU COHESION POLICY IN THE PUBLIC SPHERE: HOW DO THE MEDIA FRAME EU COHESION POLICY? Results from the COHESIFY media analysis VasilikiTriga Cyprus University of Technology COHESIFY Final Conference, Brussels 26 April, 2018 The COHESIFY project (February 2016-April 2018) has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No. 693127 1

MEDIA DATASET: Overview Reactions 1 Cohesify Universe Volume Unique Users Web media articles (11 lang) 33,842 N/A N/A Framing Analysis Stratified sample (11 lang) 2,714 N/A N/A Computational Analysis (ENG/ES) Web media articles 4,092 33,183 N/A User Comments 33,183 N/A 7,945 Social Media (Facebook) 3,601 60,132 2,321 Social Media (Twitter) 19,653 37,886 13,298 Table: Dataset Overview Footnotes 1 Reactions: number of comments, retweets, likes, shares...etc. 2

1 FRAMING ANALYSIS 3

How is Cohesion Policy FRAMED in the news?

3 2 4 1 INCOMPETENCE ECONOMIC QUALITY OF LIFE CULTURE FRAMES OF LOCAL CONSEQUENCES GOVERNANCE Mismanagement Environment Job creation of funds Cultural heritage Social justice Development Bureaucracy SUBFRAMES Cultural development Public services Research & Innovation Fail to inform Infrastructure Financial Burden public/applicants Restore order

7 6 8 5 NATIONAL FRAMES COHESION POWER MISUSE OF FUNDS INTERESTS Civic participation/ External relations Political Leverage Collaboration Corruption SUBFRAMES Tackling brain drain Political Capital Social Awareness Fraud Sovereignty Empowerment Solidarity

Are the dominant frames POSITIVE or NEGATIVE ?

(All cases, n= 2714) 40% 34.2% 35% 27.3% 30% 25% 20% 13.4% 15% 9.0% 10% 4.1% 4.1% 3.8% 2.9% 5% 1.0% 0% Fund abuse Cohesion (Frame National Power (Frame 5) Incompetence of Culture (Frame 3) Quality of life Economic No frame (Frame 8) 7) Interests (Frame local/ national (Frame 2) consequences 6) authorities (Frame 1) (Frame 4) Cohesion policy mainly framed in The negative frames are less terms of economic gains (34%), frequent: ‘Incompetence’ (9%), and impact on citizens’ everyday ‘Power’ (5%) and ‘Fund Misuse’ (4%) lives (27%)

COMPARATIVE ANALYSIS: THE DOMINANT FRAMES 9

Quality of life (Frame 2) 49.2 60.0 46.9 42.2 43.2 50.0 31.1 40.0 27.9 24.3 21.5 21.7 21.8 20.2 30.0 17.9 18.3 12.5 20.0 10.0 0.0 The key positive and dominant frames are especially prevalent in Sl, CY, IE & UK, NL, IT Economic Consequences (Frame 1) 54.7 60 47.6 47.2 42.3 50 33.3 32.1 31.1 31.5 29.7 30.3 40 27.6 25.7 22.3 22.8 30 20 10 0

COMPARATIVE ANALYSIS: THE NEGATIVE FRAMES 11

‘Power’ frame by EU media is ‘Incompetence’ frame is twice 3 times higher than average - higher than average in Romania Emphasis on political bargaining Incompetence of local/ Power (Frame 5) national authorities (Frame 4) European/ … 12.9% Romania 21.7 Spain 7.2% Poland 12.4 Germany 6.7% Germany 11.5 Romania 5.5% Spain 10.6 Netherlands 5.4% European/ … 9.9 Slovenia 4.7% Total 9.3 Total 4.0% Italy 8.7 UK 3.6% Cyprus 7.8 Italy 3.2% Slovenia 7.8 Greece 2.5% Greece 6.3 Ireland 2.5% Netherlands 4.2 Poland 2.3% UK 4.0 Hungary 1.3% Ireland 2.5 Cyprus .6% Hungary 1.8 0% 5% 10% 15% 0 5 10 15 20

COMPARATIVE ANALYSIS: THE LEAST SALIENT FRAMES 13

National Interests (Frame 6) Cohesion (Frame 7) Fund abuse (Frame 8) Germany 20.2% Slovenia 3.1% Netherlands 10.7% Netherlands 6% Hungary 10.2% Germany 2.9% European/ … 5% Spain 1.3% Romania 8.9% UK 2.8% European/ … 8.9% Netherlands 1.2% Total 2.7% Spain 5.1% Total 1.0% Ireland 2.5% Total 3.4% Greece .8% Hungary 2.2% UK 2.4% UK .8% Slovenia 1.6% Italy 2.3% Poland .8% Romania .9% Germany 1.9% Italy .5% Greece .8% Poland 1% Hungary .4% Italy .5% Slovenia .5% Romania .4% Poland .3% Greece .4% European/ … 0% Cyprus 0% Cyprus 0% Cyprus 0% Spain 0% Ireland 0% Ireland 0% 0% 5% 10% 15% 20% 0% 5% 10% 15% 0% 1% 2% 3% 4%

TOWARS EU IDENTITY THROUGH POSITIVE NEWS? 15

Positive News Valence 100% Positive EU news promotes a 89% 90% sense belonging in a community 79% 78% 80% 75% 69% 70% 62% Majority of news is positive in most 60% 56% 55% 48% cases…. 50% 37% 40% 25% … less than 50% in some cases (EU 30% 16% 20% media, DE, IT, RO) – but large neutral 10% (factual) component, not negative 0%

TOWARS EU IDENTITY THROUGH EUROPEANISED NEWS? 17

Europeanisation of News 70% 58% 60% The EU dimension of news 50% contributes to a European public 40% sphere 30% 20% But only European media presents 19% 20% 12% 12% news from a European perspective 11% 8% 8% 10% 4% 3% 3% 3% 0%

FRAMING: Conclusions The media does frame Cohesion Policy 1 - shaping the way EU is understood, interpreted & evaluated 19

FRAMING: Conclusions 2 Cohesion Policy frames are rich & diverse 20

FRAMING: Conclusions 3 The framing of Cohesion Policy is positive overall (Economic consequences, Quality of life) 21

FRAMING: Conclusions 4 But Europeanised discourse is low (nationalised discourses) 22

FRAMING: Conclusions 5 Regional media frame positively (effects on daily lives) , national media focus more on criticism against the national government 23

2 COMPUTATIONAL TEXT ANALYSIS 24

COMPUTATIONAL TEXT ANALYSIS: Online news and social media Topic modelling Sentiment analysis Discovers topics from Opinion mining approach text documents (e.g. to determine polarity of news articles, tweets, text (positive, negative or posts, etc.) and can neutral) using a dictionary handle ‘big data’ of words 25

NEWS MEDIA: Topics & proportions Figure: Estimates of topic proportions based on structural topic model (n=4.092) Topics have been assigned short labels for facilitating interpretability. 26

THE TERRITORIAL DIMENSION: National vs regional news focus (a) Spain (b) UK Figure: Logit estimates of effect size of changing from one category to another. Note: Error bars that do not overlap with the zero line are statistically significant. Positive coefficients mean that the topic receives more emphasis at the “national” level, while negative coefficients imply that the regional level emphasises the topic more. 27

NEWS MEDIA: Sentiment analysis Figure: Comparison of sentiment per territorial level. The sentiment analysis was performed on the English language sources, which means that "Regional" and "National" refer to UK 28

USER COMMENTS: International Most commentary is neutral, though punctuated by NEGATIVE COMMENTARY Figure: Sentiment analysis of User Comments from international-focused media. 29

USER COMMENTS: UK Most of the sentiment associated with USER COMMENTS is negative, especially for the Daily Mail Figure: Sentiment analysis of User Comments from UK media. 30

FACEBOOK (ENG): Activity stats From 2012-2013, levels of FB activity are steadily INCREASING. Figure: Evolution of Facebook activity statistics over time 31

FACEBOOK: Sentiment analysis of Facebook posts Most sentiment is NEUTRAL or POSITIVE MORE POSITIVE posts in (a) English Spanish than English (b) Spanish 32

TWITTER (ES): Topic proportions & sentiment Figure: Sentiment analysis of tweets. 12,7K tweets in Spain (compared to 7,3K in UK) BUT NEED MORE TIME! 33

COMPUTATIONAL ANALYSIS: Conclusions (i) Large variation across 3 cases in Cohesion Policy topic emphasis and coverage Topics mirror thematic Objectives and broader EU political themes, e.g. Conditionality & EU affairs/Brexit& Irregularities Significant territorial differences in topic emphasis & sentiment analysis e.g. EU affairs focus at national level 34

COMPUTATIONAL ANALYSIS: Conclusions (ii) News comments contain morenegative sentiment (esp. UK) Facebook activity increasing over time, but most sentiment neutral - not surprising as mostly about objective information Twitter analysis reveals also neutral sentiment 35

THANKS FOR YOUR ATTENTION! ANY QUESTIONS ? vasiliki.triga@cut.ac.cy 36