Moral networks: An analysis of punitive attitudes using big data

Ezquerra Silva, Pablo 2025. Moral networks: An analysis of punitive attitudes using big data. PhD Thesis, Cardiff University. Item availability restricted.

Preview	PDF (Thesis) - Accepted Post-Print Version Download (13MB) | Preview
	PDF (Cardiff University Electronic Publication Form) Restricted to Repository staff only Download (131kB)

Abstract

Punishment is a universal practice across all human societies, playing a central role in maintaining social order and shaping cultural norms. Previous research has linked citizens' punitive attitudes to the legitimacy of the political system, the implementation of ineffective criminal policy, and the prevalence of vigilante justice. Research on punitive attitudes has relied on data from public opinion surveys. While this approach offers the advantage of statistical representativeness, it typically provides a static snapshot, failing to capture the dynamic, real-time shifts in public opinion. Surveys also focus on individuals in isolation, making it difficult to account for the communication networks, exchanges, and discussions that shape these attitudes. Additionally, they are often susceptible to social desirability bias, where respondents may tailor their answers to align with what they believe is expected by the interviewer or their community. Using more than 100,000 crime-related tweets from the media and political system in Uruguay and the United Kingdom, alongside around 1.5 million citizen responses to both types of communications in 2022, this thesis explores the potential of social media data to develop and analyse alternative measurements of public attitudes toward punishment, advancing both methodological and theoretical innovation within Computational Criminology. It employs Natural Language Processing techniques, such as Topic Modelling, to analyse the main topics related to crime news and political communications on social media. Additionally, it uses Transformers Models to infer punitive attitudes from user-generated textual data and applies generalised linear models to compare the key determinants and consequences of these attitudes between the UK and Uruguay. The research argues that the expression of punitive attitudes occupies a relevant place in the discussion of crime and punishment on social media. It further observes that both the characteristics of news stories and user profiles influence the shaping of these attitudes, highlighting convergences and divergences between this new form of measurement and conventional strategies. Finally, the study emphasises that these communications are highly dynamic, fluctuating over time in response to specific crimes that act as triggers, and often covary with the crime-related perspectives expressed by the political system.

Item Type:	Thesis (PhD)
Date Type:	Completion
Status:	Unpublished
Schools:	Schools > Social Sciences (Includes Criminology and Education)
Subjects:	H Social Sciences > H Social Sciences (General)
Date of First Compliant Deposit:	10 October 2025
Last Modified:	10 Oct 2025 15:59
URI:	https://orca.cardiff.ac.uk/id/eprint/181597

Actions (repository staff only)

Edit Item