Correcting for measurement error in detecting unconscious cognition: comment on Draine and Greenwald (1998)

Abstract

A. G. Greenwald, M. R. Klinger, and E. S. Schuh (1995) have proposed a regression method for detecting unconscious cognition in experiments that obtain measures of indirect and direct effects of stimuli with suspected unconscious effects. Their indirect-on-direct-measure regression approach can produce misleading evidence for indirect effects in the absence of direct effects when the direct-effect measure has typical measurement error. This article describes an errors-in-variables variant of the regression method that corrects for error in the direct-effect measure. Applied to the uses of the regression method by S. C. Draine and A. G. Greenwald (1998) in this issue, the errors-in-variables method affirms substantial evidence for indirect effects in the absence of direct effects.