Three variables, one effect : what happens to the Stroop effect when the internal validity of the "Stroop protocol" tool is re-established?

If everyone knows what attention is, all psychologists know what the Stroop effect is: it's the interference (conflicting overlap) caused by the automatic reading of an incongruent word on the more controlled process of color naming (when BLUE is written in YELLOW, the automatic reading of the word "BLUE" would interfere with the naming of the color "YELLOW"). With no fewer than 26,200 citations of the original article to date, the Stroop effect is undoubtedly one of the best-known, most widely used and most studied effects since its discovery 89 years ago. Adopting a resolutely offbeat approach to this monument of scientific literature in psychology, we questioned the internal validity of the "Stroop protocol" tool for studying overlap between processes. Analysis of the tool revealed an original violation of the protocol's historical dimension, manipulating not one (Nature of the dimension supporting the color; the factor consensually considered to explain the Stroop effect) but three factors (in addition: familiarity and set size). The corrected version of the tool involved the design of two sets of neutral items (i.e., 4 familiar neutral items and 12 unfamiliar neutral items). Analyses show that (1) color-naming times for colored words (whether congruent or incongruent) are significantly longer than for colored strings, (2) color-naming times for familiar items are significantly shorter than for unfamiliar items. In conclusion, we assert that (1) the variations in color-naming times for colored words is not due to the overlaping between processes, and (2) the Stroop effect is just a particular familiarity effect. Implications of these results are discussed.