Task Switching Costs: Why Changing Tasks Takes Effort
You are comparing two versions of a report when a message arrives. Answering it takes only a moment, but returning to the report is slower: which paragraph were you checking, and why did the number in the margin matter? The interruption has ended, yet the original task is not immediately restored.
Task switching costs describe measurable differences between performing a task after switching from another task and performing it after repeating the same task. In experiments, those differences often appear in response time or accuracy. In everyday work, a switch can also involve finding materials, recovering an intention, and reconstructing a partially completed argument.
The research offers a useful reason to organize work thoughtfully. It does not establish one universal percentage of productivity lost, or prove that every combination of activities is impossible.
What researchers mean by a switch cost
Imagine a simple experiment using digits. On some trials you classify a digit as odd or even; on others you classify it as greater or less than five. The screen tells you which rule applies. Researchers compare trials where the rule repeats with trials where the rule changes.
The number itself may be easy to recognize. The difficulty concerns applying the appropriate rule, selecting a response, and avoiding interference from the previous task. This makes task switching different from merely doing something difficult.
Rubinstein, Meyer, and Evans examined switching between task rules in a 2001 experimental paper. Switching costs varied with rule complexity and advance cues. That result supports the importance of task conditions; it does not supply a fixed penalty for every email, meeting, or conversation.
A laboratory comparison isolates particular operations. Real work usually combines those operations with additional demands. Calling both situations “multitasking” can hide that difference.
Switching, doing two things together, and background activity
There are at least three situations worth separating. First, you can alternate between tasks: write a sentence, answer a message, then resume writing. Second, two tasks can overlap in time, such as following spoken instructions while inspecting a diagram. Third, an activity can continue largely in the background while another receives deliberate attention.
These situations need not produce identical interference. Familiarity matters, and tasks may compete for different resources. Walking along a clear, familiar path while talking is not equivalent to navigating an unfamiliar intersection while solving a demanding problem.
Even an activity that normally feels automatic can require more attention when conditions change. A conversation may become harder to follow when the walking surface becomes uneven. The useful question is therefore not whether someone “can multitask,” but which processes overlap, under what conditions, with what consequences.
Our guide to selective attention explains how goals help prioritize information. Switching adds another question: how efficiently can the current priority and its associated rules be replaced?
Why preparation helps but does not guarantee an effortless transition
If you know that the next task requires a different rule, you can prepare. In ordinary work, a visible agenda or a short transition note serves a related practical purpose: it makes the next demand less ambiguous.
However, preparation and performance are different stages. Knowing that you are about to edit a spreadsheet does not mean that every relevant formula, assumption, and unfinished check is already available. Some information must still be recovered once the task begins.
This distinction explains why a planned transition may feel smoother without being free. It also explains why advice to “just focus harder” often misses a design problem. If essential context is scattered across several windows, the person must repeatedly rebuild it.
A worked example: reviewing a document
Consider a reviewer checking whether a report's conclusions follow from its data. The reviewer needs the current claim, the relevant table, the meaning of its columns, and the reason a particular comparison matters.
A message asking for tomorrow's meeting time creates a new, simpler task. After answering, the reviewer returns to the same paragraph, but the paragraph alone does not contain the entire previous train of reasoning. Rereading is part of restoring the task.
Now change the setup. Before leaving, the reviewer writes: “Check whether Table 3 excludes incomplete responses; paragraph assumes all participants.” Returning no longer requires reconstructing the entire question. The note supplies a specific next action.
This is an illustrative workflow, not a trial demonstrating a particular amount of time saved. Its value can be assessed by whether the reviewer resumes accurately and finishes with fewer omissions.
Reduce unnecessary switches without becoming rigid
Start by identifying a task that depends on keeping several relationships active. Protect a manageable interval for that task, and keep the materials it requires close together. The interval should fit the work and your responsibilities; no single timer duration follows from switching research.
Separate interruptions that require immediate action from those that can wait. A collaborator reporting a serious problem may deserve an immediate response. Routine notifications can often be checked at a convenient boundary. The aim is appropriate prioritization, not absolute isolation.
When you must switch, leave a useful stopping point. Record the next action, unresolved question, or location of the relevant evidence. “Continue report” is less informative than “Compare the two estimates before revising the conclusion.”
If different tasks use conflicting rules, consider grouping similar operations. For example, checking references together may involve fewer changes of criterion than repeatedly alternating between reference checking, stylistic editing, and numerical verification. Still review the completed document as a whole afterward.
Measure output rather than the feeling of busyness
A busy session can feel productive because many actions occurred. That feeling is not the same as completing the intended work accurately. Conversely, a quiet session may produce valuable progress that is less immediately visible.
Choose a simple outcome before changing your workflow: a completed section, correctly checked records, or a finished draft with known issues listed. Compare similar tasks across several sessions, while recording major differences such as deadlines or unusually difficult material.
Avoid treating one good afternoon as proof of a universal method. Work varies, and a home comparison cannot control every influence. It can still reveal practical problems worth fixing, such as spending substantial time repeatedly locating the same information.
When switching is useful
Remaining with one task is not always the best choice. A missing answer from another person may block progress. A new priority may be genuinely urgent. A fresh activity can also be appropriate after reaching a natural stopping point.
The goal is not to eliminate transitions. It is to make the reason for a transition clear and reduce avoidable reconstruction. A deliberate switch made after recording the current state differs from following every notification without noticing that the original task remains unfinished.
Likewise, switching tasks is different from mind-wandering. You can deliberately switch between two external tasks, or remain on one screen while your thoughts drift elsewhere. Both may interrupt progress, but they call for different observations.
Questions people often ask
Does multitasking damage the brain?
A switch cost is a performance finding, not evidence that ordinary task switching causes brain damage. Claims about lasting structural harm require different studies and cannot be inferred from slower responses in a switching experiment.
Can practice remove every cost?
Practice can change how a task is performed, but improvement on a particular combination does not establish unlimited capacity. To evaluate a claim, compare accuracy as well as speed, and test the actual combination of activities involved.
Should I batch every activity?
No. Batching is a practical option when tasks share requirements and delays are acceptable. It can be unhelpful when a response is time-sensitive or when waiting creates additional work for someone else. Choose a workflow around the actual demands.
What should I try first?
Choose one demanding activity, reduce one avoidable interruption, and write a precise resumption note whenever you leave it. Evaluate whether this improves completion and accuracy. That is a more informative starting point than trying to prove you are either good or bad at multitasking.
Sources
- Rubinstein, J. S., Meyer, D. E., & Evans, J. E. (2001). Executive control of cognitive processes in task switching. https://pubmed.ncbi.nlm.nih.gov/11518143/