Policy response to endemic HIV prevalence rates in Malawi has prioritised reducing HIV misconceptions through dissemination of factual information about HIV transmission. Yet, over three decades after the implementation of these strategies, we know little about how effective they have been in dispelling misconceptions that may be associated with new HIV infections. Using the 2004, 2010 and 2015/16 Malawi Demographic and Health Surveys and applying multivariate logistic regression, this study examined how endorsement of HIV misconceptions among males and females have changed over time. For women, we found at the bivariate level that the odds of endorsing misconception about HIV transmission were higher in 2004 (OR = 1.38, p < 0.001) but lower in 2015–16 (OR = 0.92, p < 0.001) compared to 2010. We also observed that the difference between 2015–16 and 2010 in misconception endorsement was largely supressed by socioeconomic characteristics (OR = 1.06, p < 0.01). It is also found at the bivariate level that, compared to men in 2010, men in 2004 (OR = 1.23, p < 0.05) were more likely to endorse misconception about HIV transmission although those in 2015–16 (OR = 0.92, p < 0.05) were less likely to do so. Once we accounted for behavioural variables, the significant impact for 2015–16 (OR = 0.98, p > 0.05), compared to 2010, was fully attenuated. Based on our findings, we recommend increasing efforts at reaching women and men with HIV information. It is also crucial to improve women's socioeconomic status to ensure a substantial reduction in their endorsement of HIV misconceptions in Malawi. [ABSTRACT FROM AUTHOR]