ERIC Number: ED395956
Record Type: RIE
Publication Date: 1996-Jan-27
Reference Count: N/A
Evaluating Result Replicability: Better Alternatives to Significance Tests.
Gillaspy, James Arthur, Jr.
Three procedures for evaluating the replicability of descriptive discriminant analysis (DDA) results are discussed. The techniques include cross-validation, the jackknife, and the bootstrap. Discriminant analysis is a multivariate technique used when group membership or classification is the focus of the analysis. DDA is used to describe major differences among groups. Several procedures are available to researchers for evaluating the replicability of DDA results. The first is cross-validation, which involves randomly dividing the data into two roughly equal groups and deriving discriminant function coefficients for each. The discriminant function coefficients are switched and used to predict the group membership of the other sample. The bootstrap involves copying a data set over and over again into a megafile and then repetitively drawing different samples with different combinations of subjects. The jackknife differs in that different subsets or groups of subjects are repetitively dropped from the original data set. An example illustrates application of the jackknife. Appendix A presents the heuristic data set for the example, and Appendix B contains Statistical Package for the Social Sciences commands for the jackknife. (Contains 2 tables and 26 references.) (SLD)
Publication Type: Reports - Evaluative; Speeches/Meeting Papers
Education Level: N/A
Authoring Institution: N/A
Identifiers: Bootstrap Methods; Cross Validation; Descriptive Discriminant Analysis; Jackknifing Technique; Research Replication
Note: Paper presented at the Annual Meeting of the Southwest Educational Research Association (New Orleans, LA, January 27, 1996).