Involves intentionally making equivalent or balancing the levels of a specific variable, between different experimental groups to control for potential confounding effects and ensure that observed differences are attributable to the manipulated independent variable.