A PROGRAM FOR THREE-FACTOR ANALYSIS OF VARI- ANCE DESIGNS WITH REPEATED MEASURES ON TWO FACTORS FOR THE IBM 1130 COMPUTER

This program for three-factor analysis of variance with repeated measures on two factors is based on an unweighted means analysis, permitting the use of unequal sample sizes. Each factor may be fixed or random since the program will select an appropriate error term or will calculate a quasi-F ratio. The number of possible levels of each factor is extended over those in other programs. Rounding errors are minimized. A special three-factor analysis of variance design in which there are repeated observations on two factors has been used frequently in educational and psychological research. This design has been dis-cussed and illustrated in experimental design text books (Lindquist, 1953, pp. 292-297; Winer, 1971, pp. 539-559). Computer programs are available for this design. However, existing programs suffer from the following limitations: (1) they require equal sample sizes for each level of the nonrepeated factor; (2) their use is restricted to conditions where all factors are fixed or where the user can instruct the computer to select an appropriate error term; (3) on small com-puters like the IBM 1130 the limited core memory severely restricts the number of levels of each factor; and (4) rounding errors cumu-late to produce only approximate results