Abstract When regression analysis is used to determine cost standards, two kinds of aberrant observations--outliers and Influential points--are often encountered. This paper examines the impact of these observations on cost estimation and on the cost variance-investigation decision. We review statistical techniques for detecting and classifying such deviant point. An application of the techniques to a set of cost data shows that, using two decision model, a failure o roped account for these properly accounts for these may lead a manager to suboptimal variance-Investigation decisions. The illustration is effective for demonstrating to students the importance of assuring that the sample cost data are appropriate for regression analysis.
Tomczyk et al. (Mon,) studied this question.