Limiting the observed variation in a variable through selection, such as studying only a narrow stratum of senior executives. It can change measured correlations and limits the settings to which an observed relationship can generalize. Its effect is not universally a simple attenuation factor.
New to topics? Read the docs here!