ABSTRACT
In the remainder of Section 3, we restrict the statistical analysis to two-level fractional designs. The computations are much simpler because of the design structure and the associated statistical models. The main steps of the analysis are illustrated with the 16-run design defined in Tables 9 and 10. The basic ideas and computations are directly applicable to all 2k-p designs. The statistical model for the 16-run design given by Table 9 can be written as
(2)
where y is the response variable; xi> x 2, ••• , x15 are the coded variables ( ± 1) defined by the columns of Table 9; l31, ••• , l315 are the coefficients and are called main or interaction halfeffects; E is the error component.
