Published April 1, 2024 | Version Version 1. Watermarked
Project deliverable Open

D1.4 NETWORK ANALYSIS REPORT

  • 1. ROR icon University of West Attica
  • 2. ROR icon Trilateral Research & Consulting
  • 3. ROR icon University of Central Lancashire

Description

This deliverable presents the results of the research conducted in the context of T1.4 and examines the factors that influence public trust in science by studying data from online social networks. The main objectives of the analysis are:

  1. Find potential ways to identify the credibility of information circulated over online social networks.
  2. Identify social media users’ characteristics and recognise their typology.
  3. Distinguish communities (group of users) of the same beliefs.
  4. Identify social media users’ topics of high interest.
  5.  Identify the profile of social media accounts that tend to promote unreliable sources in terms of impact.
  6. Detect potential factors that can influence the judgement of users by addressing their sentiments and stances.

Initially, a literature review was conducted of how public trust in science was researched through social media, investigating methodologies and findings from previous corresponding research. Then, an analysis was performed involving the selection of a dataset on the scientific topic of COVID-19 vaccination. This analysis approach considered various key aspects: i) evaluating the reliability of information within social networks, ii) exploring online social networks and detecting potential communities, iii) analysing the text of posts as well as metrics such as number of likes and reshares, iv) developing artificial intelligence models to draw conclusions about users’ sentiment and assess the subjectivity of their opinions. Additionally, the COVID-19 analysis methodology included the examination of survey data concerning public trust in science, specifically in relation to social networks.


The study yielded important results that highlight the need to take action to strengthen trust in science through social media platforms as they constitute the second preferred source of information about science and technology issues after TV. The literature review has revealed that a highly significant factor impacting trust in science is political ideology, as it seems to reinforce and perpetuate scepticism toward scientific information. A second notable finding from the literature review is the observed variation in trust in science across different social platforms. An important quantitative finding of the analysis studying the trustworthiness of external sources in social media is the ratio of 3:1 of trustworthy posts in relation to untrustworthy posts. The third noteworthy finding of analysing the reactions to the messages is that there was greater user interaction with fake news at the beginning of the very critical vaccination period in relation to trustworthy news, while the methodology that focused on the text analysis captured specific themes of public concern by identifying specific conspiracy theories of high acceptance and the sources that were circulating them. The fifth important finding answering the question of who is involved in trustworthiness in social media is that the most untrustworthy messages of the analysis are related to people with high visibility, while in the case of credible messages, a lesser-known person seemed to have great acceptance. In the case of the methodology of social network analysis, users of untrustworthy messages tend to have a more closed community, creating echo chambers than those of credible news. Lastly, the methodology that focused on the survey analysis showed that users of online social platforms are more inclined to trust fake news than those who use traditional media for scientific information.

Files

D1.4 Network Analysis Report- final watermarked.pdf

Files (2.8 MB)

Additional details

Related works

Documents
Dataset: 10.5281/zenodo.10902347 (DOI)