A multiple-baseline design across behaviors involves establishing multiple baselines for a single participant across different dependent variables (behaviors), and then introducing the intervention for each behavior at staggered times. By showing that an individual's behavior changes only when the treatment is applied to that specific behavior, researchers can establish causality. For example, if an intervention designed to improve productivity is introduced for an office worker's sales calls first, and later for their report writing, an increase in productivity for each task corresponding only to the staggered introduction time provides clear evidence that the treatment caused the improvement.