Plenary Panel Session:

Organizer/Session Chair: Robert G. Easterling, Sandia National Laboratories

Computer Models & Data Interface: Development, Validation, & Inference

Andrew J. Booker
Research & Technology Dept., Boeing Information & Support Services

Michael Cohen
Committee on National Statistic, National Academy of Sciencess

Dennis Cox
Dept. of Statistics, Rice Univ.

Alan Karr
National Institute of Statistical Sciences

Computer models play a growing role in industry and technology by providing an alternative to or complement to physical testing. For example, alternative designs of products or processes may be "virtually tested" via a computer model to help select or optimize a design. Data from physical testing can also play a role in developing or "validating" a computer model. Predictions from data-based or data-validated models and inferences drawn from an amalgam of computer runs and physical tests are all statistical in nature. The panel members and audience will discuss statistical design and analysis issues that arise in this situation.