In this article, the authors argue that turnover correlations do not need to be corrected. First, they maintain that correction formulas cannot correct for poor construct validity. Second, they discuss the original purposes of turnover correction formulas. Third, the authors describe the logical fallacies of correcting turnover correlations. Finally, they show why turnover correlations are not, as is widely believed, statistically limited to a maximum of .80.
The final, definitive version of this paper has been published in Organizational Research Methods, 1(1), 1998 by SAGE Publications Ltd., All rights reserved.
Williams, C.R., & Peters, L.H. (1998). Correcting turnover correlations: A critique. Organizational Research Methods, January, Vol. 1 No.1, pp, 88-103.