A "response window" technique is described and used to reliably demonstrate unconscious activation of meaning by subliminal (visually masked) words. Visually masked prime words were shown to influence judged meaning of following target words. This priming-effect marker was used to identify two additional markers of unconscious semantic activation: (i) the activation is very short-lived (the target word must occur within about 100 milliseconds of the subliminal prime); and (ii) unlike supraliminal prime-target pairs, a subliminal pair leaves no memory trace that can be observed in response to the next prime-target pair. Thus, unconscious semantic activation is shown to be a readily reproducible phenomenon but also very limited in the duration of its effect.

In 4 experiments, subjects classified visually presented target words as pleasant-unpleasant words or male-female first names. Prime words were similar (congruent) or dissimilar (incongruent) in meaning to targets. Brief duration of prime words (17, 33, or 50 ms), along with pre-and postmasking, prevented most subjects from perceiving their physical and semantic properties. By constraining response latencies to fall within a response window-a narrow time band that occurred earlier than subjects would ordinarily respond-these experiments consistently produced subliminal priming effects, indicated by greater error rates for incongruent than congruent priming trials. This conclusion was confirmed by analyzing magnitude of priming as a regression function of prime perceptibility using the method of A. G. Greenwald, M. R. Klinger, and E. S. Schuh (1995). The data of each experiment passed their significant-intercept criterion for demonstrating unconscious cognition.

Greenwald, Klinger and Schuh (1995) have proposed a regression approach for detecting unconscious cognition. An errors-in-variables approach is presented that corrects for measurement error in the predictor and takes into account that the latent predictor variable is assumed to be non-negative. The new approach requires the same input as the uncorrected regression analysis and provides consistent estimates of regression weights as well as valid statistical tests of their significance. In particular, the method yields a consistent estimate of the regression intercept that provides critical evidence for unconscious cognition. A simulation study illustrates these aspects of the new technique. Several data sets are then reanalysed by means of the new method.

(1998) demonstrated replicable unconscious semantic priming by combining a response window procedure, which increases priming effects by requiring rapid responding, and a regression analysis in which the regression intercept is a marker for unconscious cognition. The commentaries by B. A. Dosher (1998) and by P. M. Merikle and E. M. Reingold (1998) raise two questions about conclusions based on these methods: (a) Did Draine and Greenwald (1998) demonstrate an indirect effect (subliminal priming) in the absence of a direct effect (i.e., visibility of the subliminal priming words)? and (b) Did Draine and Greenwald (1998) demonstrate dissociation of conscious from unconscious cognition? The lint question has reassuring responses that are reviewed here. The second question is answered by pointing out that although Draine and Greenwald (1998) did not claim to have established such dissociation, they provided data that advance the plausibility of that conclusion.

(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.