The intertwined dynamics of social networks and mental health

Abstract

Most humans are born with a fundamental motivation to socialize with others. A lack of meaningful social ties, therefore, can be detrimental for an individual’s mental health. At the same time, an individual’s mental health can affect how and which social ties are created and maintained. Thus, the association between social networks (i.e., the collection of social ties) and mental health is bidirectional. In the worst case, this bidirectionality can turn into a vicious cycle, in which (the lack of) social ties and (poor) mental health affect each other in reinforcing feedback loops. This thesis presents four studies that aim to disentangle the bidirectional association between social networks and mental health by investigating social mechanisms on the level of friendship and social interaction ties. This disentanglement is not trivial, because, so far, theoretical and empirical endeavors have mostly been concerned with one direction of the vicious cycle only or have neglected the dynamic nature of these constructs. In order to overcome these challenges, this thesis introduces a new theoretical model and, throughout its four empirical studies, applies novel methodologies to acquire a better understanding of the bidirectional and dynamic nature of social networks and mental health. In the first study, we investigate how the mental-health-related mechanisms of perceived social integration, social influence, and social selection operate on the friendship networks of a graduate housing community. We show that individuals’ mental health strongly affects the formation of friendship ties (social selection), while the alternative directionality (perceived social integration and social influence) has no significant effect. Based on these findings, a reasonable next step is to understand in more detail how mental health affects the smallest building blocks of social life: social interactions. However, face-to-face social interactions are difficult to describe and measure, and many technological solutions for collecting data have not yet been validated. Thus, in a second study, we apply newly developed RFID badges that capture how individuals interact with one another, and we examine their validity by comparing their measures with video data and self-reports. We show that the RFID badges are moderately valid to measure face-to-face social interaction. We further discuss how the accuracy of this measurement can be improved with simple data processing strategies. In the third study, we apply these RFID badges to investigate how social interaction networks captured during a socializing weekend are affected by individuals’ depressive symptoms. This way, we examine social selection mechanisms on a fine-grained interaction level. We find that depressive symptoms are associated with (1) spending less time in social interactions, (2) spending time with others reporting similar levels of depression, and (3) spending time in pair-wise interactions as opposed to groups. These interaction-level results help us to better understand the social selection mechanisms operating between mental health and social networks. So far, these studies are within the bounds of specific social contexts. Thus, they do not necessarily reflect interaction patterns over more extended periods in daily life. In the fourth study, we thus investigate social interaction dynamics in daily life using experience-sampled data of individuals with residual depressive symptoms. In this study, we focus specifically on the absence of social interactions (i.e., solitude) and its bidirectional relationship with depressive symptoms. We introduce the concept of solitude inertia that captures individuals’ dynamic tendencies to remain in states of solitude. We show that solitude inertia and depressive symptoms are positively associated in cross-sectional analyses. When disentangling this association in longitudinal analyses, solitude inertia predicts depressive symptoms eight weeks later, while depressive symptoms do not predict solitude inertia within the same time frame. Altogether, the empirical studies of this thesis contribute to the literature by identifying and testing particular social mechanisms that clarify the bidirectional links between social networks and mental health. With the findings of these studies at hand, future scholarly efforts should particularly aim at testing social interventions to improve the structures of social networks and individual’s mental health.