Loerch, Andrew G.Wallshein, Corinne C.2010-11-02NO_RESTRIC2010-11-022010-11-02https://hdl.handle.net/1920/6025Software Cost Estimation (SCE) uses labor hours or effort required to conceptualize, develop, integrate, test, field, or maintain program components. Department of Defense (DoD) SCE can use initial software data parameters to project effort hours for large, software-intensive programs for contractors reporting the top levels of process maturity, assuming these levels produce acceptable quality. Statistical analysis using ordinary least squares (OLS) proved initial parameters, such as estimated hours, initial peak staff, or estimated software size, could accurately predict actual effort hours. DoD cost estimating relationship (CER) equations differed by process maturity levels and differed for application area subsets by process maturity level. Grouping by application area subsets or adding Earned Value Management System‟s metrics (such as Schedule Performance Index and Cost Performance Index) did not consistently improve CER accuracy for the top two process maturity levels.en-USSoftware cost estimatingSoftware effort estimatingEarned value managementOrdinary least squaresDepartment of DefensePrediction accuracyDissertation