What this measures
A letter and a number appear together. If the pair sits in the top half, judge whether the letter is a vowel. If it sits in the bottom half, judge whether the number is even. The position changes without warning, so the rule you need changes with it.
This is a task-switching paradigm, and it isolates something the Stroop task does not: the cost of reconfiguring what you are doing, rather than suppressing a competing response.
Switch cost
The finding this design exists to demonstrate is that trials right after a switch are slower and more error-prone than trials that repeat the previous rule. The gap is called the switch cost, it is robust, and it is surprisingly large — commonly hundreds of milliseconds.
More interestingly, the cost does not disappear with warning or preparation time. Even when people know a switch is coming and have time to get ready, a residual cost remains. The current explanation is that the previous rule stays partly active and has to be overcome, not merely replaced.
This is the direct evidence against multitasking. What people call multitasking is rapid switching, and every switch is paying this tax.
Strategy
- Read position first, always. The most common error is applying the previous rule to the new position. Make "where is it?" a fixed first step.
- Do not pre-plan the answer. Deciding "the next one will probably be vowels" makes the switch cost worse, because you have to unwind the preparation.
- Keep the pace steady. Speeding up after a run of repeats is exactly when a switch catches you.
- Wrong answers subtract — the score is correct minus wrong over 45 seconds — so a rhythm you can sustain accurately beats a faster one you cannot.
- After an error, take one extra beat. Errors cluster: the disruption from one mistake causes the next.
Reading your result
The number to notice is not your total but where the errors fell. If your mistakes cluster on the trials right after the position changed, that is a clean switch cost and entirely normal. If they are spread evenly, the limiting factor is the rules themselves rather than the switching.
Frequently asked
Q. Why does this feel harder than Stroop?
A. Different demand. Stroop asks you to suppress one automatic response; this asks you to
hold two rule sets and swap between them. Most people find holding two rules harder.
Q. Can I train away the switch cost?
A. You can reduce it with practice on this specific pairing. A residual cost persists,
and the reduction does not generalise well to other switches.
Q. Does a large switch cost mean poor attention?
A. No. Everyone shows it. Only unusual magnitudes, measured under controlled conditions,
mean anything.
For inhibition rather than switching, try the Stroop task. For a rule that changes on you mid-task, number pressing.