"Measurement error in subliminal perception experiments: Simulation analyses of two regression methods"—comment on Miller (2000).

Abstract

J. Miller (2000) presented simulation analyses of 2 regression methods for detecting unconscious cognition. He argued that both methods exhibit statistical bias and consequently do not provide valid statistical tests for nonzero regression intercepts that constitute critical evidence for unconscious cognition. In this reply, it is argued that the simulation analyses are not useful for evaluating the regression methods because they are based on unrealistic assumptions that are rarely met in actual empirical research.