Uncertainty in Cost Performance as a Function of the Cusp Catastrophe in the NASA Program Performance Management System.

Cost performance by NASA programs can generate two stable states within the organization system used for program performance management. These states could be stable individually or show multistability with cusp-like behavior. To test this, we constructed a cusp catastrophe model in a multilevel modeling (HLM) procedure on the cost components of an eleven-month period of one NASA program. HLM was used for its ability to estimate random effects as replacements for unknown control parameters. The cusp catastrophe model was a good description of the data and there was some evidence that the overall budget size functioned as a control parameter. Results are presented in terms of three different forms of uncertainty: attractor strength, unistability or multistability, and control parameters. Recommendations and future directions are focused towards understanding the cause of uncertainty in complex management systems.