Aims: Ensuring unbiased measurements that represent the perspectives of diverse people is foundational to person-centred practice, organization and governance. As part of our research on “Equitable People-Centred Health Measurement” (https://www.healthyqol.com), we examined emotional well-being and the extent to which: a) responses to emotional well-being items are heterogeneous, and b) social determinants of health (SDOH) and health-related variables explain measurement bias. Methods: Data were obtained via an online survey of 10,076 adults in Canada. The questionnaire included: a) the “Emotional Well-Being” item bank (43 items) of the CAT-5D-QOL; b) The “Screening for Poverty and Related social determinants to improve Knowledge of and links to resources” tool to collect information about SDOH, including demographics (e.g., immigration, gender identity, racial background), social needs (e.g., finances, housing, social isolation, transportation) and disability; and c) health-related variables (health conditions, healthcare utilization, medications). Latent variable mixture models were used to examine heterogeneity in measurement model parameters across latent classes. Measurement bias was estimated as the differences between standardized emotional well-being scores from a 1-class model (no heterogeneity) and a k-class model (accommodating heterogeneity). Multivariable regression was used to explain measurement bias. Results: The sample was heterogeneous, with optimal results obtained for a 4-class model (class proportions = .09, .08, .41, and .41; entropy = .88). For 41% of the sample, measurement bias was positive (ranging from 0.01 to 0.54 for the 10th and 90th percentiles), with SDOH and health-related variables explaining 17% and 10% of the variance. For 59% of the sample, measurement bias was negative (ranging from −0.01 and −0.31 for the 10th and 90th percentiles), with SDOH and health-related variables explaining 5% and 4% of the variance. Conclusions: Ignoring SDOH and various health-related differences could result in biased measurements of emotional well-being, leading to some people’s perspectives of their emotional well-being being misrepresented.