I've always wondered, have there been variations on these tests that control for camera exposure levels and lighting conditions, or try to separate color from luminosity from morphology? In reaction-delay-based tests, is the delay because of bias or because of something else about what's being presented (e.g. strange wording or visual layout) requiring additional mental processing?

Teasing out those differences could help e.g. layout information and design cameras and image pipelines to reduce the effects of bias.