NotesFAQContact Us
Search Tips
ERIC Number: ED307296
Record Type: Non-Journal
Publication Date: 1989-Jan
Pages: 29
Abstractor: N/A
Reference Count: N/A
Some Examples of Invariance Procedures in Discriminant Analysis.
Jones, Gail
A brief historical background of discriminant analysis is given, with a description of the variety of roles that discriminant analysis can perform. Focus is on the classification role of discriminant analysis and how it can be performed by using Fisher's classification functions or the canonical discriminant functions. A small hypothetical data set consisting of two samples of 20 cases each and for whom the actual classifications (three groups) are known is used to illustrate these methods. The importance of measuring the effectiveness of the classification results by the use of invariance procedures is discussed. The first method of invariance is that of splitting the sample into two halves, developing the functions on one half and then "cross-validating" it on the other half of the sample. The second method of invariance is a comparison of the results obtained by Fisher's classification functions with those obtained by the canonical discriminant functions. The third method of invariance is performed by developing new discriminant function coefficients for the second sample and then rotating the two sets of coefficients from the two samples to "best fit." Contrary to statistical significance, which is often achieved only because a large sample is used and which is often misconstrued as an index of reproducibility, a successful invariance analysis enables the researcher to have confidence that the results will be stable and replicable across samples. Ten data tables are included. (Author/TJH)
Publication Type: Reports - Research; Speeches/Meeting Papers; Information Analyses
Education Level: N/A
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A