An empirical demonstration of the use of Monte Carlo techniques to overcome problems that often arise in cost analysis in developing estimating relationships. It is a common occurrence for the cost analyst to be faced with the situation in which the database is adequate enough to permit regression analysis to be used, but there is insufficient data available to test whether or not the standard assumptions are met. Thus, the analyst must proceed as though the classical conditions hold even though he may have serious reservations as to the validity of these assumptions. This paper examines the results of two distortions of the standard case and compares them with the standard case itself. Rather than contribute to the theory of mathematical statistics, the object of the study is to demonstrate, in a concrete manner, the effects of violating standard assumptions. 25 pp.