Date of Award
Master of Science
Department of Mathematics and Statistics
Edward D. White III, PhD
Determining accurate cost and schedule is a crucial step to planning acquisition expenditures but history has shown that estimates are routinely low. Several researchers have attempted to forecast cost and schedule growth; we pick up this stream of research with a new approach. Our data collection and analysis focused on bringing in new data sources and added longitudinal variables to account for changes that took place over time. We assessed cost and schedule parameters for 37 major acquisition programs between Milestones II and III, resulting in 172 input variables and 5 regression models, 2 for schedule slippage and 3 for cost growth. All five models passed statistical scrutiny and exhibited an Adjusted r2 in excess of 0.80. The primary discriminator was the inclusion of strictly qualitative variables, taken from Selected Acquisition Report narratives and change justifications. We called these "soft" variables and coded them on a scale of 1 to 5 in the categories of funding problems, political problems, technical challenges, and contractor cost growth. Models with and without soft variables are presented to demonstrate their relative benefit. Finally, implications and implementation examples provide users a path to what-if analysis and decision-making.
DTIC Accession Number
Foreman, James D., "Predicting the Effects of Longitudinal Variables on Cost Schedule Performance" (2007). Theses and Dissertations. 2919.