ADHOC

Analyzing Dangerous Speech, Hate Speech and Offensive Speech in Romanian and Hungarian Public Facebook Contexts using Computational Sociology Approaches

ADHOC is funded by the Romanian Ministery of Research and Innovation, CNCS - UEFISCDI within PNCDI III
Project code: PN-III-P1-1.1-TE-2016-0892 | Contract number: 35/2018

Project Description

The ADHOC research project aims to use digital social research methodology to create an overview of the issue of hate speech on public Facebook contexts through a comparative analysis of Romanian and Hungarian Facebook Fan pages and associated comment threads. The proposal is anchored in computational sociology methodology to provide a large-scale data-driven analysis of contexts, emitters, contents and targets of dangerous speech, hate speech and offensive speech in social media with generalizable results and evidence that could drive policy in the matter. The analyzed samples include news outlets, satirical websites, social entertainment communities, political parties, governmental institutions and non-governmental organizations, activist groups. Data collection of a set estimated at 15-20 million posts and subsequent comments will be automated by using Facebook Open Graph API interrogation. The ADHOC mixed methods approach combines quantitative discourse analysis techniques, co-occurrence analysis with in-depth qualitative approaches to better understand multimodal content, tone and context of the prominent identified hate speech acts.

Objectives

The ADHOC research project aims to use digital social research methodology to create an overview of the issue of hate speech on public Facebook contexts through a comparative analysis of Romanian and Hungarian Facebook Fan pages. The specific objectives underpinning this goal are:

Context Identification

Identifying the most popular or most relevant Facebook Fan Pages in the Romanian and Hungarian linguistic contexts in different categories – entertainment, news media, politics (governmental, non-governmental, activist), brands, communities or sports.

Context Analysis

Analyzing the posts of each selected Facebook Fan Page in terms of type, time-frame, topic and engagement through comments, reactions and shares.

Term Classification

Identifying and classifying terms and phrases in Romanian and Hungarian that point towards antagonistic speech into dangerous speech, hate speech and offensive speech.

Target Classification

Identifying and classifying references to groups which are often targeted by dangerous speech, hate speech and offensive speech in messages or comments.

Co-occurrence Analysis

Analyzing the co-occurrence of references to target groups with different types of antagonistic speech in Facebook comments aiming to identify emitters, targets, content, context (including temporal).

Comparative Analysis

Comparing the findings for the Romanian and Hungarian linguistic contexts and putting them into perspective regarding previous findings in the region and in the respective countries.

Results

Conceptual framework

Normative and political concepts:
Free speech, Hate speech, Vulnerable group, Protected group, Hate crime, Bias against a member of community, Bias motivated crime, Prejudice against a member of community, Incitement against community, Right-wing extremism, Intolerance, Discrimination, Harassment, Solidarity and cohesion in society
Theoretical concepts
Speech act, Antagonistic speech, Offensive speech, Dangerous speech, Social group, Stereotype, Prejudice, Insult, Slur, Network agenda setting
Methodological concepts:
Niche dictionary, Semantic network, Similarity analysis, Co-occurrence network, Community detection, Discursive patterns, Text mining, Text classification 

State of the Art

Databases

Research Reports

Conference Participation and Workshops

Publications

Project Team

Grant Director
Associate Professor

Radu Meza

meza@fspac.ro

Experienced Researcher
Associate Professor

Andreea Mogoș

mogos@fspac.ro

Experienced Researcher
Associate Professor

Hanna Orsolya Vincze

vincze.orsolya@fspac.ro

Timeline

May 2018

Project Start

30 May - 1June 2018
Szeged, Hungary

Dissemination: CEECOM Conference

2 paper presentations
1. A Digital Social Science Approach to Studying Hate Speech on Facebook in Romania and Hungary
2. The self-representations of the Romanian politicians on Facebook.Roles, frames and social media engagement

May-September 2018

Sampling

May-June 2018 selection & collection of purposeful sample
Politicians, media & parties RO+HU posts & comments DB ~2M Records
July-September 2018 selection of popularity sample
300 Facebook pages in Romanian
300 Facebook pages in Hungarian

September-November 2018

Data Collection

Public figures: RO+HU posts & comments DB ~2 M Records
Media institutions: RO+HU posts & comments DB ~1.02 M Records
Political parties: RO+HU posts & comments DB ~ M Records
Governmental institutions: RO+HU posts & comments DB ~ M Records

October 19 2018
Cluj-Napoca

Dissemination: HumaNum Conference

1 presentation
Analysing dangerous, hateful and offensive speech in social media comments

August-November 2018

Dissemination

2 articles sent for publication:
Targets of Online Hate Speech in Context. A Comparative Digital Social Science Analysis of Comments on Public Facebook Pages in Romanian and Hungarian
Is the Eastern European Perspective that Different? Frame Variation in the Romanian News Coverage of the Refugee Crisis
1 article accepted for publication
A Triadic Formal Concept Analysis Approach to Analyzing Online Hate Speech in Facebook Comments 

November 16-17 2018
Cluj-Napoca

Dissemination: Transylvania International Conference in Public Administration

1 presentation
Discursive Patterns in Social Media Posts and Comments on the Facebook Pages of Institutions. A Comparative Analysis of Romania and Hungary

November-December 2018

Reporting

Phase 1 (Conceptual Framework and Methodology) Reporting

Contact

Departamentul de Jurnalism și Media digitală
Facultatea de Științe Politice, Administrative și ale Comunicării
Universitatea Babeș-Bolyai, Cluj-Napoca
Str. Traian Moșoiu, nr 71, cod poștal: 400132

Grant director:
Associate Professor Radu Meza, Ph.D.
meza@fspac.ro