The Art of Soft-Censorship in Singapore

Snippets from my methodology chapter from my PhD thesis at the University of Melbourne. Primary data was gathered via a self-written homebrew Python crawler using Beautiful Soup to scrape moderated comments from social media and news sites. The crawler is run every 20 minutes and each comments section is compared with previous iterations in order to determine the types of comments that were moderated during that 20 minute delay. It also demonstrates my ability to create a complex and self-funded survey using Qualtrics, and using the results from both to create a causal Probit Regression model to determine socio-demographic and political characteristics which encourage self-censorship. Stats were analysed on SPSS and Stata. Download

Understanding the relation between suicide, surface‐air temperature and gross state product in Australia

This study (University of Melbourne) aims to establish if there is a relationship between suicide (outcome variable), gross state product (GSP) (explanatory variable), and temperature (explanatory variable) using multiple linear regression analysis across seven Australian cities. It uses the ACORN-SAT meteorological dataset. All modelling and analysis was done on Python. Download

Longitudinal analysis of state indices in predicting news trust from 2017 to 2022

Snippets from a study done with University of Canberra’s News Media and Research Centre which took several country-level indices (35 countries total) ranging from economic, demographic, political and even behavioural characteristics in order to create a model to predict news trust using Digital News Report (DNR) data. Analysis was done using Stata. Download

The relationship between news trust, mistrust, and audience disengagement

This study addresses this puzzling phenomenon by applying the concept of ‘mistrust’ and differentiating it from low or lack of trust. Mistrust is defined as exercising scepticism towards news and deferring trust. Confirmatory factor analysis of an online survey (n = 4401) conducted in Australia, Singapore and UK between March and April 2023 reveals that trust and mistrust are empirically separate constructs and that they co-exist. Quantitative analysis was done on SPSS. https://doi.org/10.1177/14648849241299775

Experiment on news articles’ heuristics on article trust in Australia, Singapore and the UK

Snippets of results from a WIP paper where respondents were exposed to different articles on a Website or Facebook platform and later asked on their trust of certain articles based on different heuristics. Each heuristic was then grouped into Preconceptions, Content-based, Non-content-based and Source-based. Regressions were done on SPSS. Download

Understanding the moderating effect of child abuse and poor mental health on the use of adolescent family violence: Findings from an Australian study

Snippets of results and explanation of results from a submitted paper (Monash University) looking at socio-demographic factors which lead to the perpetration of abuse. Key finding included an explanation that perpetration of abuse was wrongly attributed to having a disability in several papers, when there is a clear mediation effect between disability and perpetration of child abuse with having suffered/witnessed abuse. Respondents with disability were more likely to have suffered or witnessed abuse, thereby increasing the likelihood of perpetration. Download