Shaping the self: Identity and food choices in adolescence

Abstract

The decline in adolescent health in the UK represents an urgent and important public health issue with implications for both current and future generations of young people. This PhD used a range of methods to explore how identity formation and food choices interact during adolescence and the role of social media in that relationship. This thesis established identity formation as a core developmental task of adolescence and a key determinant of health behaviour change. A systematic review examined the use of identity theory within adolescent health behaviour change interventions. Two qualitative studies used story completion and narrative analysis to explore young people’s understanding of how identity and food choices are informed by and expressed through social media. Lastly, a novel method for large language model (LLM)-assisted narrative analysis was developed and tested. The systematic review found that identity-theory based intervention strategies, particularly those based on self-affirmation theory, were often combined with other behaviour change techniques such as implementation intentions and goal setting. Evidence for their efficacy in changing behaviour was limited, though identity-based components appeared to improve engagement with and receptivity to other intervention elements. The pilot story completion study demonstrated the feasibility of online story completion with young people and the efficacy of narrative analysis for analysing stories. The main story completion study revealed how young people viewed social media as a space for identity expression and a source of social pressure. Older participants portrayed characters with greater confidence and awareness of the potential harms of social media, whereas younger participants often portrayed characters who were more vulnerable to the harmful impacts of social media on mental and physical health. The final study found that two LLMs (Claude 3 Opus and GPT-o1) conducted high quality, credible narrative analysis, though Claude’s interpretations provided greater nuance and alignment with the interpretations of the human researcher. The study also established a set of practical guidelines for LLM-assisted analysis and outlined the key ethical and methodological issues associated with use of artificial intelligence in research. Overall, the findings of this PhD underscore the need for future research to develop and test identity theory-informed, co-produced health behaviour change interventions that empower young people to make healthier food choices, whilst supporting their autonomy and self expression through social media. Methodologically, this thesis demonstrates the potential for artificial intelligence to enhance the quality and efficiency of qualitative research with young people. It also highlights potential policy implications, which centre around the urgent need for both school- and national-level policies to integrate social media- and health-literacy into the education curriculum, supporting young people to use social media to confidently make informed, healthy choices. Collectively, the findings lay the foundation for future research and policy changes to support young people to develop positive, health-aligned identities and use social media in ways that support their mental and physical health.

More information

Identifiers

ORCID for Sarah Jenner: orcid.org/0000-0002-4644-5027

ORCID for Hanna Kovshoff: orcid.org/0000-0001-6041-0376

ORCID for Mary Barker: orcid.org/0000-0003-2976-0217

Catalogue record

Export record