Why Task Switching Costs More Than It Appears

Switching between tasks feels efficient because each individual switch is quick. Laboratory measurement shows a consistent cost, and its two components explain why the cost accumulates.

The cost is measured against repetition

Experiments present a sequence of simple tasks, sometimes repeating and sometimes alternating, and compare response times between the two conditions.

Responses following a switch are reliably slower and less accurate than responses following a repetition, using identical stimuli.

The difference isolates the cost of switching itself, since everything else about the trials is matched. It appears consistently across task types.

The magnitude is modest per switch, typically a fraction of a second. It becomes substantial only because switches are frequent, and the cost is paid on each one.

Preparation removes only part of it

Giving advance notice of an upcoming switch reduces the cost, indicating that some of it involves reconfiguring in advance of the stimulus.

Extending the preparation interval further produces diminishing returns. A residual cost remains no matter how much warning is given.

The residual portion appears to require the stimulus itself before it can be resolved, which is why preparation cannot eliminate it entirely.

The previous task persists

Settings established for one task do not switch off cleanly. They remain partially active and interfere with the new task's requirements.

Interference is strongest when the two tasks use the same stimuli with different rules, since the same input activates two competing responses.

The persistence decays over time, which is why returning to a task after a longer gap involves less interference from what was done between.

Interruption resembles switching

An interruption imposes a switch that was not prepared for, combining the full reconfiguration cost with disruption of whatever was held in mind.

Recovering the interrupted context requires reconstructing it, and that reconstruction takes considerably longer than the interruption itself.

This is why brief interruptions consume disproportionate time, and why the measured recovery period after an interruption exceeds its duration substantially.

Practice reduces but does not eliminate the cost

Extensive practice at switching between two specific tasks reduces the measured cost, largely by making each task more automatic.

The reduction is specific to the practised pair. Switching between different tasks shows the original cost, which limits how far the gain generalises.

Grouping similar work together therefore has a mechanical basis. Fewer switches means fewer reconfigurations and less interference carried between them.