ERIC Number: ED164612
Record Type: RIE
Publication Date: 1978-Mar
Reference Count: 0
Bias of the Canonical Redundancy Statistic: A Monte Carlo Study.
Dawson, Beth K.
A Monte Carlo study was conducted to discover the amount of bias to be expected in canonical correlations and in the canonical redundancy statistic. Six hundred random sample replications of 72 conditions were performed and the mean sample values compared with population values. Conditions were varied with respect to the number of predictor variables, the number of criterion variables, the size of the population correlations, and the size of the samples. The least biased estimates are those that use large sample sizes with moderate to high interbattery correlations. Both the Wherry and the Olkin-Pratt correction formulas were found to be satisfactory for correcting for bias in the total redundancy statistic. (Author/CTM)
Publication Type: Reports - Research
Education Level: N/A
Authoring Institution: N/A
Note: Paper presented at the Annual Meeting of the American Educational Research Association (62nd, Toronto, Ontario, Canada, March 27-31, 1978) ; Best copy available