Background: Empathy is fundamental to social cognition, driving prosocial behaviour and mental health. Self-reported empathy varies across cultures and there are differing reports of associations with demographic characteristics. We therefore aimed in a UK survey to characterise two main self-reported components of empathy, namely empathic concern (feeling compassion) and perspective taking (understanding others’ perspective). We hypothesised that empathy would be associated with age, gender, ethnicity, relationship status, employment, socio-economic status, education, and personality. / Methods: We asked participants in the COVID-19 Social Study - an internet-based survey of UK-dwelling adults aged ≥18 years - to complete the Interpersonal Reactivity Index subscales measuring empathic concern and perspective taking, and sociodemographic and personality questionnaires. We weighted the sample to be UK population representative and employed multivariable weighted linear regression models. Results: In 30,033 respondents, mean empathic concern score was 3.86 (95% confidence interval 3.85, 3.88) and perspective taking was 3.57 (3.56. 3.59), the correlation between these subscores was 0.45 (p < 0.001). In adjusted models, greater empathic concern was associated with female gender, non-white ethnicity, having more education, working in health, social-care, or childcare professions, and having higher neuroticism, extroversion, openness to experience and agreeableness traits. Perspective taking was associated with younger age, female gender, more education, employment in health or social-care, neuroticism, openness, and agreeableness. / Conclusions: Women and people working in caring professions have higher empathy levels. Perspective taking declines with age but empathic concern does not. Empathic compassion and understanding are distinct dimensions of empathy with differential associations with demographic factors.