NotesFAQContact Us
Search Tips
ERIC Number: ED314482
Record Type: Non-Journal
Publication Date: 1989-Oct
Pages: 22
Abstractor: N/A
Reference Count: N/A
Alternative Solutions for ANOVA Using 2-Way and 3-Way Completely Randomized Factorial Designs with Disproportional Cell Sizes.
Pugh, Richard C.; And Others
Three alternative solutions available for analysis of variance (ANOVA) using two-way and three-way completely randomized factorial designs with disproportional cell sizes across all two-way and three-way summary tables with no empty cells are discussed. These solutions are categorized as unique, sequential, and experimental. The procedures used, the Statistical Package for the Social Sciences-X (SPSSX) (release 3.1) ANOVA and multivariate analysis of variance (MANOVA) procedures and the Statistical Analysis System (SAS) (release 5.18) general linear models (GLM) procedure, report the same results for only the unique solution for each effect. Similarities as well as differences were found among the sequential and experimental solutions. Comparisons across the three software procedures were demonstrated by illustrating the adjustment procedures to effects through the use of multiple regression and orthogonal coding. It is evident that caution needs to be exercised in the interpretation of the results from the three software procedures and in relating the results to textbooks in the behavioral sciences. Eight data tables and three figures are provided. (TJH)
Publication Type: Reports - Evaluative; Speeches/Meeting Papers
Education Level: N/A
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A