In a pandemic such as Covid-19, effective communication plays a key role both in crisis management and in the efforts to ensure public health and safety. Consequently, understanding public trust in different information sources is one of the most important aspects in designing effective communication strategies. This article explores public trust in different information sources in the context of Covid-19 in Bangladesh, and asks: Firstly, which information sources do people mostly trust in a pandemic? Secondly, does trust in different information sources vary across peoples’ gender, age, education level, stream of education, employment status, socio-economic background, religion and political affiliation, and if so, how does trust in information sources vary across these socio-demographic characteristics? Using primary data from an urban sample, we employ binary probit models to obtain the determinants of individuals’ most trusted information source during the Covid-19 pandemic. Our paper finds that there are differences in the trust that individuals place in different information sources. These differences vary in magnitude and significance across different socio-demographic characteristics, and our evidence insists that we take into account these differences when steering policy. In particular, we find that most of the urban population of Bangladesh mostly trust the government and administration as a source of information during a pandemic such as Covid-19. On the other hand, media celebrities and politicians are among the least trusted information sources. In addition, our results reveal that the government and administration face, as an information source, a trust deficit among females, Muslim people and Bangla medium and English medium educated people as compared to people with no formal education. Our findings provide valuable information to policymakers on how to best disseminate critical information to reach different populations, e.g., females, non-Muslims, not formally educated individuals. Based on our findings, we recommend targeted community engagement, tailored according to the trust in information sources of different groups of individuals, and we advocate for the increased involvement of government authorities, administration and academicians in disseminating Covid-19 related information.