ABSTRACT
Suppose that we wish to develop a measure of cognitive ability. After receiving advice from an “expert” neighborhood psychologist, we will ask each participant to throw a baseball as far as he/ she can, and the distance, in feet, will be our measure of cognitive ability. A shrewd observer might remark that a given respondent might vary from one trial to the next in the distance that he/she throws the ball. Wanting to improve upon our measure, we decide to use the average of ten throws. Whereas there may be considerable intraindividual variability in the distance of a single throw, there will be little variability in the average distance over ten throws.
