ABSTRACT
This chapter examines the assumption of independence between activity durations and introduces correlation. It generates 5000 projects, represented by the 5000 iterations of a Monte Carlo simulation, using uncorrelated random variables. In we introduce two confounding risks that will be assigned one to Design and one to Test. The chapter assumes that each pair of activities is highly correlated at 90 percent. It begins the analysis that simulates the schedule without correlation. In traditional schedule risk analysis, the values of correlation coefficients are usually determined in interviews or risk workshops where project participants talk about risks and their impacts on the project schedule. Some of the retail software that provides Monte Carlo simulation capabilities uses the Spearman Rank Order approach to correlation, an approach that is theoretically inferior to the Pearson Product Moment approach.
