Nico Seifert

• When COVID-19 was declared a global pandemic by the World Health Organization (WHO), concerns quickly arose that restrictions on social contact and physical distancing could lead to a mental health crisis due to the potentially higher prevalence of social isolation and loneliness in the general population. These concerns were based on the belief that social isolation and loneliness pose significant risks to the mental health of people who experience them, commonly referred to as the causation hypothesis. Extensive research has linked both conditions to outcomes such as depression, anxiety, and suicide. However, evidence of causation has been limited because few studies have attempted to rule out alternative explanations for the observed associations. It is equally plausible that people experience social isolation or loneliness because of the presence of mental health problems (direct selection hypothesis) or because of stable characteristics that make them vulnerable to these problems (indirect selection hypothesis). Against this background, the purpose of this thesis is to improve our understanding of why social isolation and loneliness are associated with mental health problems. As an initial contribution, theoretical arguments for causation and selection have been reviewed and systematised into a broader set of mechanisms. Then, a brief review of empirical research to assess the extent to which previous studies distinguish between the three explanations has been presented. The review revealed that empirical evidence of the negative effects of social isolation and loneliness on mental health is not as clear-cut as often assumed. Finally, four empirical studies were conducted using advanced techniques of panel data analysis to rule out selection as an alternative explanation and to provide new evidence for the causation hypothesis. In addition, the empirical studies examined whether the conclusions regarding causation remained robust across multiple measurement approaches and broader sociodemographic groups (i.e., men and women, older and younger people). Overall, the results support the causation hypothesis and show that social isolation and loneliness are associated with worse mental health, even when selection is considered as a competing explanation. Convincing evidence of causation emerged across all sociodemographic groups, and measurement approaches did not affect the substantive conclusions. These findings have several practical implications. First, concerns about the COVID-19 pandemic triggering a mental health crisis by increasing the prevalence of social isolation and loneliness in the general population were well founded. Second, the critical importance of addressing social isolation and loneliness is highlighted by strong support for the causation hypothesis. Because of the continued high prevalence of social isolation and loneliness worldwide after the pandemic and their impacts on mental health, both conditions cause great human suffering and impose significant costs on health care systems.