A resistant estimator is a type of statistical estimator that remains relatively unaffected by small changes or outliers in the dataset. It is particularly useful in robust statistics for providing reliable parameter estimates even when assumptions of underlying populations are violated.