The Multitasking Myth: What the Research Actually Shows

The Task-Switching Research

The foundational study: Rubinstein, Meyer, and Evans (2001), "Executive control of cognitive processes in task switching," Journal of Experimental Psychology: Human Perception and Performance, 27(4), 763–797.

Key finding: when participants switched between tasks, they consistently lost time. Every switch involved two measurable cost components:

The combined switch cost was small for simple, low-cognitive-demand tasks. For complex tasks (writing, analysis, programming, strategic thinking), the cost was significant. Each switch was not free; it had a measurable toll.

The Stanford Multitasker Study

Ophir, Nass, and Wagner (2009) at Stanford published "Cognitive control in media multitaskers" in Proceedings of the National Academy of Sciences, 106(37), 15583–15587. Their hypothesis: heavy multitaskers had developed superior attention management through practice.

The results were the opposite of the hypothesis. Heavy multitaskers performed worse than light multitaskers on every attention task measured:

Researcher Clifford Nass summarized the finding: "They're terrible at every aspect of multitasking. They get distracted constantly. Their memory is very disorganized."

The interpretation: heavy multitasking does not train superior attention management. It trains inferior attention management. Habitual multitaskers become worse at focusing because they have repeatedly practiced not focusing. The skill that develops is the opposite of the intended one.

Gloria Mark's Interruption Research

Gloria Mark (University of California, Irvine) conducted field research observing information workers in natural office environments, watching them work, timing task durations, and measuring behavior after interruptions across multiple studies from 2004 through 2023.

Her widely cited finding is that it takes a significant amount of time to fully return to a task after an interruption. A commonly cited figure of "23 minutes and 15 seconds" has circulated widely. However, it is worth noting that this specific number appears to have originated in a media interview rather than a published paper. Mark's published research (including "The Cost of Interrupted Work," CHI 2008, with Gudith and Klocke) documents substantial interruption costs without citing this precise figure. Reference her published research rather than this specific number when precision matters.

What her research consistently shows: after an interruption, workers typically shift to a second or third task before returning to the original. The sequence of interruption, handling it, difficulty re-engaging, finding another task, and eventually returning means interruptions compound in ways invisible to the person being interrupted.

Her more recent work (published in Attention Span, 2023) found that average self-directed attention on a single task before switching had decreased substantially compared to earlier measurements. A growing proportion of that switching is self-interruption: checking devices unprompted by any external notification. People have internalized the habit of interrupting themselves.

The Self-Interruption Problem

Most discussions of interruption focus on external sources such as colleagues stopping by, notifications, and phone calls. Mark's research reveals that self-interruption is now a significant fraction of total interruptions in knowledge work.

Self-interruption occurs when a person switches tasks or checks devices without any external prompt. Examples: checking email while writing a document, switching browser tabs during analysis, picking up the phone during a thinking session. These interruptions are invisible in organizational interventions because they have no external agent.

The implication: productivity improvements from reducing external interruptions are limited if self-interruption has become habitual. The behavioral pattern of checking devices and switching tasks has been sufficiently reinforced that it operates automatically. Structural solutions (turning off notifications, closing email) help; they are not complete solutions if the underlying checking behavior has been deeply conditioned.

The Email Study: Five Days Without

Mark, Voida, and Cardello (2012) conducted a study in which participants' email access was cut off for five days. Measured outcomes:

The Exception: When Dual-Task Performance Is Possible

Not all combinations of tasks are subject to the same switching costs. Research on dual-task performance identifies when simultaneous activity is possible:

Different cognitive modalities: Walking while talking uses separate neural systems and imposes minimal cost on either. Listening to non-lyrical music while doing physical work may not impair either task. These combinations work because they don't compete for the same cognitive resources.

Highly automated tasks: Driving a familiar route while listening to a podcast works because driving has become largely automatic. The same combination is dangerous in an unfamiliar city or in traffic, where driving requires conscious attention that competes directly with the podcast.

The rule: two tasks compete when they require overlapping cognitive resources, such as two verbal tasks, two visual tasks, or two tasks requiring active problem-solving. For cognitively demanding knowledge work, virtually all meaningful combinations compete. The feeling that you can write and monitor Slack simultaneously is a calibration error; the evidence is in the quality of what gets written.