Sampling with replacement means that each selection is from the whole population and so the same value (or person) can be chosen multiple times. It is the opposite of sampling without replacement.
Used in glossary entries: sampling without replacement