|
Abstract:
|
When planning an experimental investigation, we often encounter practical or economic constraints that result in a restriction on randomization accommodated with a split-plot design structure. In this talk, case studies from aerospace research are used to illustrate split-plot design considerations and construction strategies. Candidate second-order, split-plot designs are evaluated based on their estimation and prediction efficiency. In addition, the clarity of structure, ease of execution, and simplicity of analysis of the competing designs are considered. Throughout this talk, subjective compromises among design evaluation criteria are emphasized. Developing effective experimental strategies in a restricted randomization context depends on our ability to balance multiple design evaluation criteria, thereby guiding the design selection process.
|
