Date of Award


Document Type


Degree Name

Master of Science


Department of Mathematics and Statistics

First Advisor

Edward D. White, PhD


This study explores a two-step procedure for assessing defense acquisition program cost growth using historical data. Specifically, we seek to predict whether a program will experience cost growth and, if applicable, how much costs will increase. We compile programmatic data from the Selected Acquisition Reports (SARs) between 1990 and 2000 for programs from all defense departments. We focus our analysis on cost growth in research and development dollars for the Engineering Manufacturing Development phase of acquisition. We further limit our study to only one of the seven SAR categories of cost growth engineering cost growth. We explore the use of logistic regression in cost analysis to predict whether cost growth will occur. Using this methodology, we produce a statistically significant model that accurately predicts approximately 70 percent of our validation data. For those programs that have cost growth, we use a multiple regression model (an adjusted R2 of 0.4645), with a natural log transformation, to predict the expected amount of cost growth. We discover the two-step logistic and multiple regression approach produces desirable results. Finally, we find schedule variables to have the most predictive ability from the 78 candidate independent variables analyzed.

AFIT Designator


DTIC Accession Number


Included in

Risk Analysis Commons