Investigating the effects of social media and recommended content on mental health

Social media use has been associated with adverse effects on mental health, though there are various gaps and conflicting findings in the literature. Previous research has found that positive effects on mental health can result from abstaining from social media for one week (Lambert et al., 2022). Conclusive causal links have not been drawn. 

Young people are particularly susceptible to socio-emotional problems (Knowles et al., 2022) and 75% of all mental health conditions start before the age of 24 (Kessler, 2005). Social media has become an integral element of young-adult life (Lee et al., 2022), with Ofcom (2021) suggesting that 88% of people aged 16-24 in the UK have a social media profile.  

Although some literature has attempted to infer causal links between social media use and adverse mental health outcomes through investigating social media abstinence (Tromholt, 2016; Lambert et al., 2022), these studies were of short duration, on small cohorts, and suffered from various potential confounds. More extensive studies would be necessary, with different types of interventions, sizes of cohorts, and durations. 

Social media algorithms are designed to keep users engaged by providing them with personalized content. It has been proposed that these algorithms can be a key driver of poor mental health by exposing people to negative stimuli (e.g., negative news and comments) (Kahout et al., 2023). Current evidence also suggests that social media’s effect on mental health may be influenced by who and how one uses it (Lee & Way, 2021; Verduyn et al., 2017). Habits of use may therefore provide an interesting behaviour for investigation. However, there is a lack of research looking at the extent to which changing one's social media habits can reduce the negative impact of social media on mental health and whether this is mediated by changes in the social media algorithm. For example, it is possible that asking people to create a new social media account or asking them to deliberately search/like/subscribe to certain types of content may influence the recommender system with positive impacts on mental health. 

We will begin with an extensive comparison of all known experimental results, and then propose further experiments based on those comparisons. One approach will be to integrate a theoretical foundation via a trans-diagnostic cognitive behavioural model of social media use proposed by Tibber & Silver (2020).  

This research aims to: a) expand on previous findings by investigating whether changing social media habits lead to improved mental and physical health via changes in the recommender system, b) utilise longer follow-up measures to assess the lasting impact of changing social media habits on mental health outcomes, and c) assess whether intensity and motivations for use influence social media’s effect on mental health and well-being.

Applicants should hold, or expect to receive, a master's degree or first or upper-second bachelor's degree in a relevant subject. Knowledge of Python and elementary statistics would be desirable, as well as previous experience in conducting cohort studies.

This project is associated with the UKRI Centre for Doctoral Training (CDT) in Accountable, Responsible and Transparent AI (ART-AI). The ART-AI CDT aims at producing interdisciplinary graduates who can act as leaders and innovators with the knowledge to make the right decisions on what is possible, what is desirable, and how AI can be ethically, safely and effectively deployed. We value people from different life experiences with a passion for research. The CDT's mission is to graduate diverse specialists with perspectives who can go out in the world and make a difference.

Formal applications should include a research proposal and be made via the University of Bath’s online application form. Enquiries about the application process should be sent to art-ai-applications@bath.ac.uk. Informal enquiries about the project should be directed to Dr Lambert and Prof Cristianini.

Start date: 2 October 2023.