Disentangling perceptual and response biases: The act of detecting a stimulus contaminates measures of conscious experience

A central challenge in consciousness research is determining whether observers truly experience a stimulus. However, present/absent detection judgments are often biased by contextual factors, making it difficult to isolate conscious perception from non-perceptual influences. Traditional psychophysical methods cannot disentangle these components. To address this, we conducted in-person experiments (N=505) in which participants detected and reproduced dim stimuli under three contextual manipulations: attentional cues, asymmetrical base rates, and payoff schemes. Using a reproduction task and a Hurdle-Gaussian model, we were able to separate perceptual from non-perceptual effects. We found that statistical priors (base rate) and reward structures (payoff) induced non-perceptual shifts in reproduction behavior, whereas attentional cues selectively led to shifts that reflect changes in conscious experience. Critically, we show that the mere presence of a detection context contaminates reports of conscious experience with non-perceptual effects. This highlights the need for caution in using and interpreting results relying on detection judgments, despite their central role in consciousness research. We propose that reproduction is better suited to assay conscious experience than detection.