Proceedings of the 2014 Industrial and Systems Engineering Research Conference Y. Guan and H. Liao, eds. Dynamic Guidelines to Statistically Control Peer Review Costs Constantine M. Koursaris, Ph.D. Department of Management Sciences College of Business Worldwide Embry-Riddle Aeronautical University Daytona Beach, FL Richard L. W. Welch, Ph.D. Chief Statistician & Senior Technical Fellow Northrop Grumman Corporation Melbourne, FL Abstract During the system/software development lifecycle process (SDLC), peer reviews are conducted throughout to ensure that true-positive defects are detected and flagged for re-work. Often, false-positive defects may be flagged for re- work as well, thus increasing operational costs. This paper examines two possible organizational baseline and operational scenarios sequence recommendations using statistical process control of peer review costs. The optimal organizational operational scenario will be selected based on the analysis of over four years of empirical data and 2517 data points collected during the time of the research. The conclusions will show that the log-transformed data scenario is the best, most optimal, operational scenario to implement within the organization. Although the research was sponsored by a Department of Defense prime contractor to the U.S. Air Force bound software systems, it can be applied in the civilian sector of aviation transportation software systems in the management of the supply chain. Identifying accurate defects and implementing this scientifically proven method of controlling costs within the organization is of tremendous value in both government and civilian aviation transportation systems. Keywords Industrial engineering, operations research, statistical process control 1. Introduction Identifying defects and controlling costs during peer reviews of a product being developed for the U.S. military ensures that quality is not compromised while the product is going through its development life cycle. Several process improvement guidelines have been established, but none specify what type of method to use in order to more accurately product detect defects and thus control costs within the organization developing the product and eliminating unnecessary re-work for the developers. The Capability Maturity Model Integration (CMMI) [1] has provided organizations with process improvement approaches. The application of high maturity statistical models and techniques within a CMMI Level 5 framework has been effectively employed and illustrated as a result of a continuing analysis and exploitation of peer review data. Statistically controlling peer reviews of requirements, software designs, code, and test artifacts, within all phases of the Software/System Development Life Cycle (SDLC) process has proven to have a direct relationship with controlling costs. Reducing and controlling variation directly affects the reduction of peer review cost and improves efficiency. Controlling the total cost of software development is one of the fundamental interests expressed by many organizations. Also, the necessity to improve internal processes within organizations, have given us structured methodologies to use and implement in order to become more efficient, increase productivity, and reduce costs. Using statistical process control (SPC) to our advantage, offers many benefits. Walter Shewhart developed control charts to indicate when special-cause variation exists in a process [2]. One notable benefit of using SPC is that it reduces programmatic risk. It gives superior insight into average performance, and variability of the controlled process. It provides higher confidence estimates. Use of SPC enhances predictability and stability in executing the job [3]. It enables proactive process improvement to meet management or customer performance targets. As a result, the common cause variation is removed from the process [4]. The focus of this paper is to recommend the best organizational baseline and operational scenario for statistically controlling peer review cost using dynamic guidelines. By using statistical process control methods, we can control costs, by first identifying defects. If the