NotesFAQContact Us
Search Tips
ERIC Number: ED312296
Record Type: Non-Journal
Publication Date: 1989-Nov
Pages: 19
Abstractor: N/A
Reference Count: N/A
Planned Comparisons as Better Alternatives to ANOVA Omnibus Tests.
Benton, Roberta L.
Analyses of data are presented to illustrate the advantages of using a priori or planned comparisons rather than omnibus analysis of variance (ANOVA) tests followed by post hoc or posteriori testing. The two types of planned comparisons considered are planned orthogonal non-trend coding contrasts and orthogonal polynomial or trend contrast coding. A hypothetical one-way design was devised to illustrate the superiority of planned comparisons. The data set used one dependent variable with five levels. Ten subjects were assigned evenly, two to each of five groups. The dependent variable consisted of hypothetical scores on a measure of performance. The examples demonstrate that planned comparisons using orthogonal coding may be preferable to classical ANOVA followed by post hoc testing. These procedures: (1) provide greater statistical power against Type II error; (2) help isolate specific sources of variance in the dependent variable; (3) help determine forms or patterns of relationships; and (4) make the researcher think about relationships when designing proposed studies. Researchers should consider using planned tests more often when factors have more than two levels. Five data tables and two appendices on non-trend comparisons and trend coding contrasts supplement the text. (SLD)
Publication Type: Reports - Evaluative; Speeches/Meeting Papers
Education Level: N/A
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A