Using mandatory stay-at-home orders in Australia as a natural experiment and data from a long-running panel study, this paper investigates the causal link between physical isolation and loneliness. We exploit variations in the number of lockdown days in 2020 the respondent had experienced up until the interview date to estimate the causal link and find, based on difference-in-differences analyses with three-way fixed-effects estimations, that the number of days in lockdown does not significantly affect loneliness. Further, we use triple differences to examine heterogeneous effects. For income, age, personality, living arrangements, and remoteness, we find insignificant effects; for extroverts and young people, we find weak significance. We investigate exclusion restrictions through channels such as social contacts, internet access, job industry, and household characteristics on loneliness. Whereas many believe that 'being alone' and 'being lonely' are similar concepts, our study provides the first empirical causal evidence of no links between the two. Our findings also refine understanding of social isolation and demonstrate that it likely encompasses factors other than physical isolation.