The Perversity of Missing Points in the 2(4) Design,
Abstract
The author would like better to understand the impact of missing data on estimability (and variance) in 2 to the n power factorials and 2 to the n-k power fractional factorials. Hoping to find generalizable results, the author examined the 2 to the 4th power design to determine what points could be deleted without losing estimability of main effects and 2-factor interactions (resolution V property). He was guided by a result of P. W. M. John which shows that if a fraction is missing from a 2 to the n power design, then estimable effects are those estimable from half replicates, and the least squares estimates are obtained by averaging the estimates from half replicates.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1983
- Accession Number
- ADP001588
Entities
People
- Carl T. Russell