https://www.dau.edu
Besides investigating Cost and Fee associated with a program (we do not investigate Price since Cost plus Fee constitutes Price), we also focus on the common Level 2 WBS of Data, Prime Mission Equipment (PME), Support Equipment (SE–sum of Peculiar and Common Support Equipment), Systems Engineering and Program Management (SEPM), System Test and Evaluation (ST&E), and Training. Due to lack of available data, we do not address fitting distributions to cost growths associated with Initial Spares (only 6 datapoints), Site Activation (only 5 datapoints), and Industrial Facilities (only 3 datapoints). This lack of data for site activation and initial spares is also noted by Markman et al. (2021) in their study. Prior to finalizing the database, preliminary histograms of the cost growth patterns reveal that extreme values could easily distort fitting a probability distribution. Figure 1 highlights an example of the histogram of the cost growth associated with the Cost WBS along with fitting a Rayleigh distribution (a specialized Weibull distribution) for illustration purposes. The far-right value near 25 greatly influences the population median of this fitted distribution, which calculated equals 3.5. The actual sample median of these cost growths is 1.5. We mitigate the effects of these extreme values by engaging the logic of Alpha Trimmed Filtering (ATF).
FIGURE 1. RAYLEIGH FIT TO WBS COST ELEMENT WITHOUT ATF
0.8
Histogram of Data Rayleigh Distribution
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
0
5
10
15
20 25
Note. ATF = Alpha Trimmed Filtering; WBS = Work Breakdown Structure.
37
Defense ARJ , Spring 2025, Vol. 32 No. 1
Made with FlippingBook - Online Brochure Maker