With an ageing population, there are increasing numbers of experienced family carers (FCs) who could provide peer support to newer carers in a similar care situation. The aims of this paper are to: (i) use a cross-sectional study design to compare characteristics of volunteers and recipients of a peer support programme for FCs of people with dementia, in terms of demographic background, social networks and psychological well-being; and (ii) use a longitudinal study design to explore the overall impact of the programme on the volunteers in terms of psychological well-being. Data were collected from programmes run in Norfolk, Northamptonshire, Berkshire and four London boroughs between October 2009 and March 2013. The volunteer role entailed empathic listening and encouragement over a 10-month period. Both carer support volunteers (N = 87) and recipient FCs (N = 109) provided baseline demographic information. Data on social networks, personal growth, self-efficacy, service use and well-being (SF-12; EuroQol Visual Analogue Scale; Hospital Anxiety and Depression Scale; Control, Autonomy, Self-Realisation, Pleasure-19) were collected prior to the start of the intervention (N = 43) and at either 3- to 5 month or 10 month follow-up (N = 21). Volunteers were more likely than recipients of support to be female and to have cared for a parent/grandparent rather than spouse. Volunteers were also more psychologically well than support recipients in terms of personal growth, depression and perceived well-being. The longitudinal analysis identified small but significant declines in personal growth and autonomy and a positive correlation between the volunteers' duration of involvement and perceived well-being. These findings suggest that carers who volunteer for emotional support roles are resilient and are at little psychological risk from volunteering.