The Notification Tax, Measured

Every notification is a small interruption, and small interruptions are the easiest costs to dismiss. This page is the receipt: what researchers actually measured when they interrupted people — the attention that stays stuck on the old task, the error rates that climb after a two-second buzz, and the arousal that never quite returns to baseline. If you want to know whether silencing your phone is worth it, these are the numbers the parent topic gestures at.

The Daily Stream, Counted

Before asking what an interruption costs, start with how many exist. The most detailed measurement tracked 94 Android users for five days, logging every tap, swipe, and session (dscout, 2016): 2,617 touches per day on average, across about 76 sessions, and 5,427 for the heaviest users. Notifications arrive more slowly but steadily: in an in-situ study logging real phones for weeks, the median was 63.5 per day (Pielot et al., 2014). These are exposure numbers, not harm numbers — they define the size of the stream. They matter because each event is a candidate micro-interruption, and the studies below measure what a single one does.

Attention Residue: The Half-Switch Problem

The core mechanism has a name: attention residue. Sophie Leroy had participants switch between work tasks and measured how much attention stayed behind on the task just left (Leroy, 2009). Two findings stand out. First, residue is the default — part of attention keeps chewing on the old task after you have "moved on." Second, it is worst when the abandoned task is unfinished and the deadline is loose enough that the brain feels licensed to keep worrying. A notification is a textbook residue generator: it arrives mid-task, pulls you toward an open loop (who is it? what do they want?), and leaves the original task unfinished. The cost is not the glance — it is the partial attention that stays allocated afterward.

Even the briefest interruptions do measurable damage. Altmann and colleagues interrupted people mid-task for just 2.8 seconds — and found the post-interruption error rate roughly doubled relative to uninterrupted work (Altmann et al., 2014). Being derailed was enough to reset the cognitive state needed to resume. And the phone does not even need you to answer it. Stothart and colleagues compared people who received a notification during a demanding attention task with those who did not: the notification group's error rate climbed sharply — roughly three times the controls' in the headline comparison — even though almost nobody actually picked up the phone (Stothart et al., 2015). The buzz alone is the event.

The Interruption Studies, One by One

The table collects the studies this page leans on. Read the last column as confidence, not drama: these are honest but mostly small lab findings, and their exact percentages will not transfer to your Tuesday afternoon.

Two honest glosses. Lab tasks exaggerate some things and hide others: a controlled test makes the effect crisp, while real workdays add recovery time that softens it. And the workplace study behind the famous "23 minutes" figure (Mark et al., 2008) is a small observation, not a law of nature — an order-of-magnitude marker, as the parent topic treats it.

The Stress-Physiology Question

So interruptions cost attention — but is the notification stream a stress exposure? The honest answer is that the direct physiology is thinner than the attention data. What has been measured: self-reported stress falls when email checks are batched (Kushlev & Dunn, 2015); interrupted workers report more frustration and work faster to compensate (Mark et al., 2008); and when a phone rings and cannot be answered, heart rate and reported anxiety climb (Clayton et al., 2015). What has not been measured at scale is a cortisol time-series pinned to notification frequency. The cortisol topic owns the curve science; this page's job is the more modest claim the data actually support: notifications produce small, frequent, measurable arousals and attentional costs, and their cumulative effect is that the stress system never fully settles — the failure mode the Downshift topic says undermines recovery. Call it ambient vigilance, not acute alarm.

Why Small Taxes Compound

• 🔁 Repetition is the multiplier. A three-times error rate for ten seconds of your day is trivia; the same event embedded in 60-plus daily notifications becomes the texture of the day.
• 🧠 Residue stacks. Each switch leaves a fragment of attention behind, so late afternoon thinking carries the morning's unfinished loops along with it (Leroy, 2009).
• 🤳 The checks self-generate. People reach for the phone without any notification — the habit becomes its own interruption stream, which is why the fix has to be architectural (see Detox That Sticks).
• 🌙 The evening is the expensive end. The same small tax lands harder after dark, when it collides with the cortisol descent and sleep onset — the reason the Notifications Curfew exists.
• 📱 Presence alone bills you. Even a silent, face-down phone on the desk measurably reduces working memory — the brain keeps a reserve of attention pointed at it (Ward et al., 2017).

Measuring Your Own Exposure

• 📝 Run a one-day tally. Every time you pick up the phone, make a mark on paper. Most people's guess is off by half — the tally is the correction.
• 📊 Read the report your phone already writes. The weekly screen-time summary lists pickups, notifications received, and hours — compare it with your mental estimate.
• ⏳ Try the checkless hour. Put the phone in another room for one working hour. Note how many times the urge to check appears — each urge is a residue event you had been paying for invisibly.
• 🌇 Notice the 7pm contrast. When the stream is silenced for the evening (the curfew page has the schedule), compare how your body feels at 9pm versus a stream-on evening. Subjective, but it is your dose-response.
• 🛠️ Then spend the settings ten minutes. The tax is paid by default; removing it is a settings menu plus a charger location (the friction page and the bedroom rule).

Questions, Answered Briefly

• 🗓️ Is this just productivity talk? No — the physiology links are weaker than the attention data, but real; and the evening half of the stream has the clearest health relevance, through sleep (the parent topic documents the crossover).
• 🚫 Do I need to quit my phone? No. The measured harm comes from interruptions, not from the device — removing the optional stream while keeping calls is the design the whole series uses.
• ⏱️ How fast would I notice a change? The batching studies measured lower self-reported stress within about a week; the habit-level relief takes longer and belongs to the final page of this series.

Related Topics