This study used causal inference to estimate the longitudinal effects of contagion in cohabitants and family members on university students’ mental health and academic self-efficacy during the COVID-19 pandemic.

A prospective longitudinal study including a baseline online measurement in May 2020, and online follow-ups after 5 months and 10 months. Participants were recruited through open-access online advertising.

Public universities and university colleges in Sweden.

The analytical sample included 2796 students.

Contagion in cohabitants and in family members was assessed at baseline and at the 5-month follow-up. Mental health and academic self-efficacy were assessed at the 5-month and 10-month follow-ups.

Mild symptoms reported in cohabitants at baseline resulted in negative mental health effects at follow-up 5 months later, and mild baseline symptoms in family members resulted in negative effects on academic self-efficacy at follow-ups both 5 and 10 months later.

Notwithstanding the lack of precision in estimated effects, the findings emphasise the importance of social relationships and the challenges of providing students with sufficient support in times of crisis.