The first part of this thesis was a meta-analysis examining the efficacy of virtual reality exposure The first part of this thesis was a meta-analysis examining the efficacy of virtual reality exposure therapy (VRET) for social anxiety. Effect sizes suggested that VRET was more effective than waiting list controls, and at least as effective as traditional therapeutic interventions. Some concerns were identified with regards to potential for bias. It was also noted that VRET involved the individual viewing a virtual reality (VR) scenario from a first-person perspective. As a recent study has investigated the use of third-person VR doppelgängers for public-speaking anxiety (PSA), this was identified as a key area of further research for the thesis to address. Some research has suggested that individuals who struggle to form vivid mental imagery (MI) are more likely to experience high levels of social anxiety, and may not benefit from visualisation techniques used in therapy. The next part of the thesis found that MI ability was a significant predictor of PSA. Five students with high PSA and low MI were then invited to engage in an enhanced version of the VR doppelgänger intervention, in order to examine the feasibility of incorporating this into the undergraduate psychology programme. The study suggested that the intervention could feasibly be rolled out as part of the course. Participants reported that they struggled with a standard visualisation script due to difficulties associated with MI, but the VR intervention compensated for this. They also discussed noticing significant improvements outside of the VR sessions, and that they felt safe and present in the simulation. A decline in PSA was recorded over time for all participants. The thesis concludes with a review of the theoretical implications of these findings and provides some suggestions for future research, before outlining the implications for clinical practice.