https://www.dau.edu
The process for calculating and graphing %DCAC using EVM metrics is repeated using ES metrics, i.e., SPI CUM(t) and SCI (t) . Equations (8) through (11) show the ES variants of the equations presented previously. EAC SPI(t) = ACWP CUM + (BAC - BCWP CUM ) / SPI CUM (t) (8) EAC SCI(t) = ACWP CUM + [(BAC - BCWP CUM ) / (CPI CUM * SPI CUM (t))] (9) DCAC SPI(t) = (EAC SPI(t) - Final ACWP) / Final ACWP (10) DCAC SCI(t) = (EAC SCI(t) - Final ACWP) / Final ACWP (11) To assess the differences between utilizing ES versus EVM, the stability points for EAC SPI and EAC SCI are compared between the two techniques. The stability points are the percent complete values when the absolute value of %DCAC (|%DCAC|) reaches and maintains a value of less than or equal to 10% and less than or equal to 5% through project completion. These stability points are denoted as %EAC(10) and %EAC(5). The hypothesis tests include a Wilcoxon Rank Sum test to compare distribution locations. The Wilcoxon Rank Sum test assesses the equivalency between two distributions with respect to their responses. The null hypothesis is that the responses are comparable in value between the two groups. The alternative hypothesis is that the
responses of one group are larger than the other group. If comparing the responses between three or more groups, then the Kruskal-Wallis test is used. The null hypothesis in the latter test assumes the responses are comparable for all the groups. The alternative is that at least one group of responses is larger than another. The nonparametric Wilcoxon Rank Sum test (or Kruskal-Wallis test) is used in lieu of a parametric test due to the limited sample size of this study. The hypothesis tests were evaluated at a 0.05 significance level. Additionally, the Brown-Forsythe and Levene tests are used to compare variances.
63
Defense ARJ , Spring 2025, Vol. 32 No. 1
Made with FlippingBook - Online Brochure Maker