People who repeatedly post hostile content on social media tend to have criminal histories and childhood upbringings

When browsing social networking sites such as Twitter (formerly Twitter), you may notice users who repeatedly make excessively hostile comments. In a Danish study, the content of posts on Twitter and official records collected by the government were analyzed to find out what characteristics people who repeatedly make hostile comments on Twitter have.

The offline roots of online hostility: Adult and childhood administrative records correlate with individual-level hostility on Twitter | PNAS

https://www.pnas.org/doi/10.1073/pnas.2412277121

Hostile tweets linked to upbringing and legal troubles, study finds
https://www.psypost.org/hostile-tweets-linked-to-upbringing-and-legal-troubles-study-finds/

In the early stages of its development, social media held the promise of creating a world where people could feel closer to each other and more connected, but in recent years, it has become clear that the constant bickering, fighting, threats and insults on social media are hindering people from connecting with each other.

Previous studies have identified platform-side factors that incite conflict on social media, such as anonymity on the platform and company policies that emphasize engagement. However, the fact that most hostile posts are made by a small number of users, and that many users make few hostile posts, suggests that user characteristics are also related to aggression on social media.

So a research team from Denmark and Hungary combined extensive real-world records with social media behavior to investigate what makes users aggressive. First, the researchers randomly sampled adults living in Denmark and collected a variety of data from the Danish national registry, ranging from criminal records to childhood upbringing. They then used the adults' names to search for matching X-accounts, identifying those that had made at least one post during the study period.

In total, more than 1.3 million Danish-language posts from 4,931 users were collected, and a computer-based approach was used to measure the anger and hate contained in the posts, rating the hostility users displayed on X. A similar approach was also used to measure how frequently users posted about political issues.

The researchers then analyzed the levels of hostility displayed by participants online, along with factors such as their criminal history (drawn from national registries), academic achievement, parental divorce status, parental socioeconomic status, time spent in foster care or in care, as well as their age and gender.

The analysis revealed that people with more criminal records and those who spent more time in foster care tended to be more hostile on X. They also found that people who had better academic achievement in childhood and whose parents had higher socioeconomic status also posted more hostile content. Other findings included a tendency for men to be more hostile than women, and a slight tendency for younger people to prefer hostile content.

One might expect that people with higher family socio-economic status and academic excellence would be less hostile on online platforms, but our analysis also suggests that people with greater resources and skills are more likely to talk about politics on X, and that political discussion is associated with more confrontational and hostile posts.

This study is important because it directly links online platform behavior to individual-level characteristics over decades. One of the most interesting findings is that 'people who are more likely to engage in political debates tend to post more hostile posts because controversial topics elicit stronger language. This shows that social advantage characteristics, such as good academic performance in childhood and high family socioeconomic status, are associated with hostile online behavior.'

The study focused on Danish X users, so trends may differ in countries with different cultures and social structures. Also, the analysis of posts was done by a computer algorithm, so it is possible that more subtle forms of sarcasm, sarcasm, or indirect attacks were missed.