A Bicycle built for one. Spatial dimensions and prediction of loneliness in the Netherlands

Abstract

Loneliness, encompassing emotional, social, and general dimensions, is increasingly recognized as a critical public health issue linked to mental and physical health risks, including depression and premature mortality. Despite assumptions that urban environments offer greater social opportunities, a growing body of research suggests that urban living may paradoxically exacerbate feelings of isolation. This thesis investigates how environmental, socio-economic, and behavioral factors have influenced the trajectory of loneliness across Dutch neighborhoods between 2016 and 2022. Drawing on high-resolution neighborhood-level data from CBS and RIVM, this study employs both linear regression and spatial lag models to examine the predictors of loneliness over time and account for spatial dependencies.

Findings reveal that socio-demographic and behavioral factors, such as household size, volunteerism, smoking, and the proportion of non-Western migrants, are consistently strong predictors of loneliness across all three dimensions. Infrastructural accessibility, including proximity to train stations and retail services, plays a significant but context-dependent role. Environmental variables such as green space show mixed effects, with increases sometimes associated with higher loneliness, particularly in rural or less socially cohesive contexts. The results further indicate that emotional loneliness is increasingly shaped by individual or micro-social factors, while social loneliness remains tightly linked to broader community and environmental conditions.

Temporal comparison highlights how predictors of loneliness have evolved, with certain variables (e.g., theater access, caregiving burden, and urban density) shifting in direction or strength over time. Spatial lag models confirm that loneliness is not randomly distributed but instead clusters geographically, especially in relation to social and general loneliness. These patterns suggest that neighborhood-level interventions, targeting social infrastructure, behavioral health, and equitable urban planning, are essential for reducing loneliness in the population. By combining spatial analysis with longitudinal insights, this study contributes to a more nuanced and actionable understanding of loneliness as a spatially and socially embedded phenomenon.