Fig. 1. Pathways from macro-level factors to loneliness.

a, Macro-level factors might influence loneliness through multiple indirect pathways, some leading to an increase (+) and some leading to a decrease (–) in factors that in turn either increase (+) or decrease (–) loneliness. The direction of the total effect of an indirect pathway is defined by the product of the two direct effects. For example, the indirect pathway involving offline interactions involves two negative effects (social media use reduces offline interactions and engaging in offline interactions reduces loneliness) that together result in a total positive effect (social media use increases loneliness). If positive (blue) and negative (red) pathways are approximately counter-balanced, the overall (net) effect of a macro-level factor is close to zero, even though specific causal effects might exist. b, Macro-level factors might have differential effects on different subgroups within a population. These differential effects might differ in strength and direction such that they lead to increased levels in loneliness in some subgroups (red) and to decreased levels of loneliness in other subgroups (blue). c, Most sociocultural changes occur gradually over time, but sudden changes are possible, for example in the context of historical events. Changes in loneliness might be similarly slow, and they might be delayed such that loneliness changes lag behind changes in macro-level factors. Whether these changes can be linked empirically depends on the time window examined. For example, the yellow time window would reveal no association between changes in the macro-level factor and changes in loneliness levels. By contrast, the purple time window would reveal a strong association between changes in the macro-level factor and changes in loneliness levels.