Design, Testing and Estimation in Complex Experimentation. IV. A Computer Routine for Evaluating Incomplete Factorial Designs

摘要

A computer program for the evaluation of incomplete factorial experimental designs for a specified experimental situation is described. This program allows any number of alternative designs to be compared on the basis of variance of alias characteristics with a minimum of effort. Several different models, which differ in the interactions to be estimated, may be used. The designs may have up to 64 runs, and up to 64 parameters may be estimated; however, detailed instructions are included on how these and other limitations may be changed. The program takes a maximum of 1 minute of execution time per designmodel combination, but for a moderate number of parameters takes much less (3 seconds for 25 parameters). It is programmed in the Fortran IV system for the I.B.M. 7094 computer. (Author)