Posted by Using Dummy Variables in CER Development
From the Journal of Cost Analysis and Parametrics: Volume 10 | Issue 1 | October 2021
JCAPv10i1-UsingDummyVariables-Hu
Abstract:
Dummy variables (also referred to as indicator variables) are commonly used in regression analysis to stratify data into distinct categories. The dummy variable t-test validates the assumption that distinct categories in the data set share the same sensitivity (rate of change) for the ordinary independent variable; the only difference is in the response level. However, many analysts ignore this test when specifying dummy variables in their cost estimating relationships (CER). Consequently, the fit statistics can be misleading. This paper explains when to use dummy variables and how to use them properly when deriving CERs. Specific guidelines are proposed to help analysts determine if dummy variables are appropriate for their data set and common errors analysts experience when applying dummy variables to real examples are explored. The paper also explains how to use the Chow test and dummy variable t-test to validate the CER and discusses using dummy variables in splines (to derive the fitted equation as well as the intersection).
Authors:
Ph.D. in Mathematics, emphasis in Statistics. Shu-Ping Hu retired as the Chief Statistician of Tecolote Research, Inc. She has over 30 years cost analysis experience and has published many technical papers, including the PING Factor (adjust log-linear CER result to reflect the mean), adjusted R-square for MUPE and ZMPE CERs, and Generalized Degrees of Freedom for the constrained minimization process. Shu-Ping is a recipient of the Frank Freiman Lifetime Achievement Award.
Bachelor in Mechanical Engineering, Master’s of Science in Naval Architecture. Alfred Smith retired as a General Manager with Tecolote Research, Inc. He managed the development, distribution, and training of tools such as ACE, CO$TAT, RI$K, and JACS. He was the contractor lead for handbooks, including Joint Agency Cost Schedule Risk and Uncertainty (JA CSRUH), JA CER Development, and DoD Cost Estimating Guide. Alfred is a recipient of the Frank Freiman Lifetime Achievement Award.