2**(3-1) FRACTIONAL FACTORIAL DESIGN NUMBER OF LEVELS FOR EACH FACTOR = 2 NUMBER OF FACTORS = 3 NUMBER OF OBSERVATIONS = 4 (A SATURATED DESIGN) RESOLUTION = 3 (CAUTION! MAIN EFFECTS ARE CONFOUNDED WITH 2-TERM INTERACTIONS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . FACTOR DEFINITION CONFOUNDING STRUCTURE 1 1 1 + 23 2 2 2 + 13 3 12 3 + 12 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 12 + 3 13 13 + 2 23 23 + 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . DEFINING RELATION = I = 123 REFERENCE--BOX, HUNTER & HUNTER, STAT. FOR EXP., PAGE 410, 387 NOTE--THIS DESIGN IS EQUIVALENT TO A TAGUCHI L4 DESIGN REFERENCE--TAGUCHI, SYS. OF EXP. DES., VOL. 2, PAGE XXX. NOTE--IF POSSIBLE, THIS (AS WITH ALL EXPERIMENT DESIGNS) SHOULD BE RUN IN RANDOM ORDER (SEE DATAPLOT'S RANDOM PERMUTATION FILES). NOTE--TO READ THIS FILE INTO DATAPLOT-- SKIP 50 READ 2TO3M1.DAT X1 TO X3 DATE--DECEMBER 1988 NOTE--IN THE DESIGN BELOW, "-1" REPRESENTS THE "LOW" SETTING OF A FACTOR "+1" REPRESENTS THE "HIGH" SETTING OF A FACTOR NOTE--ALL FACTOR EFFECT ESTIMATES WILL BE OF THE FORM AVERAGE OF THE "HIGH" - AVERAGE OF THE "LOW" X1 X2 X3 ---------- -1 -1 +1 +1 -1 -1 -1 +1 -1 +1 +1 +1