On each trial you'll see a checkerboard with a cylinder casting a shadow. Two squares are marked A and B. Your only job: decide whether their actual surface color is the same or different — ignore the shadow and judge the paint itself.
All trials are logged below. Download the full row-level CSV, or copy the summary.
Summary shows proportion responded "different" (P“diff”) by context and signed δ. Positive δ = square B physically lighter (congruent with the illusion); δ=0 are catch trials where the squares are truly identical. The three contexts decompose the effect: isolated = patches on plain gray (context-free discrimination floor), board = checkerboard surround with no shadow (adds local-contrast induction), illusion = full shadow scene (adds the cast-shadow bias). So board − isolated estimates the surround contribution and illusion − board the shadow-specific contribution.