Testing Dependent Correlations with Nonoverlapping Variables: A Monte Carlo Simulation

Autor/inn/en	Silver, N. Clayton; Hittner, James B.; May, Kim
Titel	Testing Dependent Correlations with Nonoverlapping Variables: A Monte Carlo Simulation
Quelle	In: Journal of Experimental Education, 73 (2004) 1, S.53-69 (17 Seiten)Infoseite zur Zeitschrift PDF als Volltext Verfügbarkeit
Sprache	englisch
Dokumenttyp	gedruckt; online; Zeitschriftenaufsatz
ISSN	0022-0973
Schlagwörter	Monte Carlo Methods; Simulation; Effect Size; Sample Size; Correlation; Predictor Variables; Statistical Analysis; Population Distribution; Tests; Statistics + Suchen Sie Ihr Suchwort? Monte-Carlo-Methode; Simulation program; Simulationsprogramm; Korrelation; Prädiktor; Statistische Analyse; Demographical distribution; Bevölkerungsverteilung; Examination; Prüfung; Examen; Statistik
Abstract	The authors conducted a Monte Carlo simulation of 4 test statistics or comparing dependent correlations with no variables in common. Empirical Type 1 error rates and power estimates were determined for K. Pearson and L. N. G. Filon's (1898) z, O. J. Dunn and V. A. Clark's (1969) z, J. H. Steiger's (1980) original modification of Dunn and Clark's z, and Steiger's modification of Dunn and Clark's z using a backtransformed average z procedure for sample sizes of 10, 20, 50, and 100 under 3 different population distributions. For the Type 1 error rate analyses, the authors evaluated 3 different magnitudes of the predictor-criterion correlations (rho[subscript 12] = rho[subscript 34] = 0.10, 0.30, and 0.70). Likewise, for the power analyses, 3 different magnitudes of discrepancy or effect sizes between correlations with no variables in common (rho[subscript 12] and rho[subscript 14] were examined (values of 0.10, 0.40, and 0.60). All of the analyses were conducted at 3 different levels of predictor intercorrelation. Results indicated that the choice as to which test statistic is optimal, in terms of power and Type 1 error rate, depends not only on sample size and population distribution but also on (a) the predictor intercorrelations and (b) the effect size (for power) or the magnitude of the predictor-criterion correlations (for Type 1 error rate). For the conditions examined in the present study, Pearson and Filon's z had inflated Type 1 error rates when the predictor-criterion correlations were low to moderate. Dunn and Clark's z and the 2 Steiger procedures had similar, but conservative Type 1 error rates when the predictor-criterion correlations were low. Moreover, the power estimates were similar among the 3 procedures. (Author).
Anmerkungen	Heldref Publications. 1319 Eighteenth Street NW, Washington, DC 20036-1802. Tel: 800-365-9753; Tel: 202-296-6267; Fax: 202-293-6130; e-mail: subscribe@heldref.org; Web site: http://www.heldref.org
Erfasst von	ERIC (Education Resources Information Center), Washington, DC
Update	2017/4/10