The mechanism
Before a response, the current task set must connect a stimulus to the correct decision and key. A cue can start that reconfiguration in advance, reducing but rarely removing the cost. Residual activation from the previous task can still compete with the new rule. Cue changes introduce an additional retrieval demand, which is why changing the symbol that names a task can look like a cognitive switch even when the rule itself repeats.
What to measure
Switch cost compares switch trials with repetition trials inside mixed blocks. Mixing cost compares repetition trials in mixed blocks with trials in a single-task block. Report both response time and error rate, because a participant may appear faster by accepting more errors. Preparation interval, cue repetition, response repetition, and rule complexity should be stored as trial-level variables. Analyze return switches separately when possible, because coming back to a recently abandoned task can carry additional interference.
Useful practice conditions
- Use clear, stable cues first; introduce alternative cues only when cue retrieval is an intended challenge.
- Randomize switches while keeping comparable numbers of switch and repetition trials.
- Give enough preparation time for learning, then vary it explicitly to distinguish advance preparation from reactive switching.
Limits of the conclusion
Smaller switch costs after practice show adaptation to the trained rules, cues, and timing. They do not automatically mean better real-world multitasking, where goals are less controlled and interruptions last longer. Switch costs also depend on the chosen baseline, so claims should name the exact comparison rather than use a general label such as mental flexibility.