Stroop test with your interference in milliseconds

The 1935 experiments did not contain the comparison everyone now calls the Stroop effect. Stroop compared naming the ink of conflicting color words against naming the color of solid squares, and reading conflicting color words against reading the same words in black — and the striking half of the result was the asymmetry. Reading was barely touched by the ink; naming was slowed enormously by the word. That is the finding: one of the two operations is automatic enough to run through interference and the other is not. The congruent condition, where the word names its own ink, arrived later, and with it the difference this page leads with. Anybody quoting a figure for “the Stroop effect” is quoting one of several different subtractions, which is a large part of why the published numbers span the range they do: 50-200 ms, per MacLeod (1991), Half a century of research on the Stroop effect: an integrative review, Psychological Bulletin.

Three things move that figure around, and a page that does not name them is not describing its own task. The first is the response channel — saying the color out loud produces far more interference than pressing a key, because speech and the printed word compete for the same output, and a keyboard version therefore sits low in the range on purpose. The second is whether the conditions are blocked or mixed: a printed card of nothing but incongruent items lets you settle into a strategy, while a shuffled run denies you that, and the two designs are not measuring quite the same thing. The third is the control condition, and it is the one browser versions usually drop. Without a neutral cell there is no way to say how much of a difference was a mismatched word hurting and how much was a matching word helping, so a two-condition test reports a quantity it cannot decompose and calls it interference. This run keeps the third cell.

The reason no rank or percentile appears beside your milliseconds is specific and worth stating: the effect is robust as a phenomenon and unreliable as an individual measure. The same person tested twice produces two interference scores that agree far less well than the size of the group effect implies, which is exactly the property that makes a task good for demonstrating a mechanism and bad for sorting people — Hedge, Powell & Sumner (2018), The reliability paradox: Why robust cognitive tasks do not produce reliable individual differences, Behavior Research Methods sets this out directly. So the page hands you a difference and a confidence interval on it, and leaves the ranking alone. If you want to see the same control mechanism measured through a different conflict, the neighbors are worth a run each: the flanker task puts the distraction beside the target rather than inside it, and the Simon task uses a distractor with no relevance to the answer at all — only to the hand that gives it. A broader sweep across several of these paradigms at once is what the executive function assessment assembles, and the selective attention test asks the filtering question without a speeded response at all.

A reaction time here is the interval between the frame that painted the stimulus and the timestamp the browser attached to your key, both read from the same monotonic clock. What neither can see is the display pipeline behind it, so on a 60 Hz screen roughly 16 ms of every figure below is the machine rather than you. That is the timing floor: two numbers closer together than that are the same number, and this page reports no precision it cannot support.

This is a measurement exercise, not a clinical assessment. It reports what you did on this page against a stated reference and nothing more — it cannot establish an attention disorder, a reading difficulty or a frontal-lobe problem. Only a qualified professional, working with more than a browser, can make that judgment.