A stopping rule is a predefined criterion used to determine when to stop a statistical experiment or data collection process. It ensures that decisions are not made based on arbitrary or post-hoc considerations, thereby maintaining the integrity of the statistical analysis.