Increasing dispersal duration should result in increasing dispersal distance, facilitating higher gene flow among populations. As such, it has long been predicted that genetic structure (e.g. FST) among populations of marine species should be strongly correlated with pelagic larval duration (PLD). However, previous studies have repeatedly shown a surprisingly poor correspondence. This result has been frequently interpreted as evidence for larval behaviours or physical oceanographic processes that result in larvae failing to reach their dispersal potential, or error inherent in estimating PLD and FST. This study employed a computer modelling approach to explore the impacts of various uncertainties on the correlation between measures of genetic differentiation such as FST and PLD. Results indicate that variation resulting from PLD estimation error had minor impacts on the correlation between genetic structure and PLD. However, variation in effective population size between species, errors in FST estimation and non-equilibrium FST values all had major impacts, resulting in dramatically weaker correlations between PLD and FST. These results suggest that poor correlations between PLD and FST may result from variation and uncertainty in the terms associated with the calculation of FST values. As such, PLD may be a much stronger determinant of realized larval dispersal than suggested by the weak-to-moderate correlations between PLD and FST reported in empirical studies.