I divided each day into four six-hour blocks to calculate the average minutes late for comparison
purposes. This dependent variable gives an alternate method for controlling for effects other than
concentration. I thank one of the referees for suggesting this specification.