Social relationships are important predictors of a range of individual outcomes, such as wellbeing and health. These social relationships are conceptualised in different ways, such as (inter-personal) forms of social support, identification with groups, or social capital. What is the overlap among these concepts and in what ways do they differ? The present work aims to clarify this with empirical evidence from two panel studies (N = 3934; N = 2912). The studies include central measures of social relationships (group identification, group membership, social support and social capital). Empirical differences and overlap were studied by evaluating the factor structure of the data with both confirmatory factor analyses and bi-factor analyses. Results showed that the different concepts had a large amount of empirical overlap (together accounting for over 60% of common variance). Surprisingly, results also revealed that subcomponents were identifiable based on who they target and not based on their conceptualisation. For example, items about identification with neighbourhood factored together with support items from the neighbourhood, and not with other identification items. Accordingly, we conclude that in addition to a general factor, it is possible to meaningfully distinguish components of social relations based on which group is targeted by the items (e.g. neighbourhood or family and friends). For future research on the relationship between social relations and health, the present measures are unlikely to be sufficiently precise to disentangle whether health effects are caused by identification, support or capital. Differences between targets appear to be more important than differences between these concepts for understanding the relationship between social relations and health and wellbeing.