Collection
Search Tips
ERIC Number: ED445082
Record Type: Non-Journal
Publication Date: 2000-Jan
Pages: 16
Abstractor: N/A
Reference Count: N/A
ISBN: N/A
ISSN: N/A
Using Canonical Correlation To Explore Relationships between Sets of Variables: An Applied Example with Interpretive Suggestions.
Alexander, Erika D.
Canonical correlation analysis is a parsimonious way of breaking down the association between two sets of variables through the use of linear combinations. As a result of the analysis, many types of coefficients can be generated and interpreted. These coefficients are only considered stable and reliable if the number of subjects per variable is sufficiently large. The first of these coefficients, the canonical correlation, is the bivariate correlation between the composite scores for the two sets of variables. Two additional coefficients, the canonical function and structure coefficients, address the contribution a single variable makes to the explanatory power of the set of variables to which the variable belongs. The communality coefficient explains how useful the variable is in defining the canonical solution. The adequacy coefficient indicates how adequately the analysis represents the total variance in the unweighted set. The extent to which a variable contributes to explaining the composite of the variable set to which the variable of interest does not belong is the index coefficient. A final outcome from canonical correlation analysis is the redundancy coefficient, which indicates the average proportion of variance for variables in one set that is reproducible with the variables in the other set. While the coefficient is easy to calculate, it is not recommended for interpretation in most cases. (Contains 3 tables and 10 references.) (SLD)
Publication Type: Reports - Descriptive; Speeches/Meeting Papers
Education Level: N/A
Audience: N/A
Language: English