Journal of Statistics: Advances in Theory and Applications
FACTORIAL AND 
FRACTIONAL FACTORIAL DESIGNSAuthor's: Maher Qumsiyeh and Kraig Kirchner
Pages: [131] - [147]
Received Date: June 13, 2011
Submitted by:
In 
factorial and 
fractional factorial designs, each effect
is dependent on every observation. Therefore, missing observations in
factorial designs can drastically alter these effects. To restore the
orthogonal structure to the design, estimation methods are needed.
After estimation, the half-normal plot for the effects needs to be
examined. If insignificant effects approximately point toward the
origin, then we have a proper estimation. Current, popular estimation
methods sacrifice effects in order to calculate missing observations.
In this paper, we have attempted new estimation methods without
explicitly sacrificing effects that seem to work well.
un-replicated factorial and fractional factorial experiments, missing values, effect estimation.
