Alternate between two rules applied to the same digit. Measures switch cost and mental agility.
Task switching measures cognitive flexibility - your ability to reconfigure mental rules quickly.
The same digit has two possible responses depending on the current rule (parity vs magnitude).
When the rule changes from the previous trial, responses slow - that slowdown is the switch cost.
Smaller switch costs are associated with better multitasking under high load.
Rogers & Monsell (1995) formalised the alternating runs paradigm used here.
Prefrontal cortex damage dramatically enlarges switch costs; training reduces them with transfer to unrelated tasks.
How much speed you lose the instant the rule changes.
Because both rules act on the same digit, the stimulus never tells you which to apply; only the cue does. The task isolates rule reconfiguration, your ability to load a new task-set and drop the previous one. A slow, error-prone switch trial reflects how much of the old rule lingers, not how well you know either rule on its own.
Unlike a Stroop or go/no-go task, which stress suppressing one fixed response, this measures shifting between two equally valid responses. The relevant correlate is everyday multitasking: moving between email and a spreadsheet, or alternating between adding and checking while cooking, where the holdover of a just-used rule briefly slows the next action.
Train the switch itself, not your speed on either rule.
Start with predictable, slow alternation so each rule is solid, then shorten the cue-to-digit gap and randomise when switches occur. Focus on the cue first, not the digit: read which rule applies before you look at the number. Level up only once your switch trials feel as automatic as your repeat trials, not merely fast.
The common mistake is reacting to the digit before you have locked in the rule, which produces fast but wrong switch responses, the most diagnostic error here. If switch errors climb while speed stays high, you are guessing the rule rather than reconfiguring it. Slow down momentarily at each cue rather than carrying the previous rule forward.
The number that matters is the gap, not the raw speed.
The key figure is your switch cost: the difference between switch-trial and repeat-trial performance. A small, stable gap with few switch errors signals efficient reconfiguration; a large gap means the old rule keeps intruding. Rather than chase an absolute benchmark, track your own switch cost over sessions and watch whether the gap, not just overall speed, shrinks.
Switch cost is highly state-dependent: fatigue, divided attention and time pressure all enlarge it, so a single session is a snapshot, not a trait. The score does not separate genuine flexibility from cautious slowing, nor does it tell you how you would cope switching among three or more rules, which taxes the system differently.