Week 6: The blank that isn't

Doomscrolling, and the course's central suspicion. The thing that feels like it empties your head is the thing that keeps the loop running

29 March 2027

Open the slides

Where. Outdoors, venue posted by the Friday before. Opens with the ten-minute detox.

Start with the experience, because everyone in the room has had it. You are spiralling about something. You pick up your phone and scroll. News, feeds, whatever is in front of you. After a while your head feels blank. Not calm exactly. Emptied. And it feels like an improvement on the spiralling.

The suspicion this lecture is built on is that the blank is not an improvement, and that reaching for it is how a loop stays open for three days instead of closing in twenty minutes.

Week 5 argued that thinking finishes itself in quiet, undemanding slots. This lecture is about the thing that now occupies those slots, and it is the last piece of the course’s argument.

It is a real thing, recently#

Doomscrolling has only been a measurable construct for a few years. Sharma, Lee and Johnson (2022) built the first proper scale for it, persistently attending to negative news and crisis content in a feed, and found it tracks with neuroticism and negativity bias. Satici and colleagues (2023) developed and validated another across three studies, and found psychological distress mediated the relationship between doomscrolling and wellbeing.

So there is a real construct with real instruments behind it. Also a young field, and a scoping review of it notes there is still no single agreed measure, with plenty of studies using one-off scales. Hold the numbers loosely.

The break that isn’t a break#

This is the strongest evidence in the lecture and it is worth knowing precisely.

Kang and Kurtzberg (2019) gave 414 people a set of anagrams with a break in the middle. The break’s activity and length were held constant. The only thing that changed was what people did it on.

Read that last row again. Taking a break on your phone left people performing about as well as people who got no break at all. Whatever a phone break is, it is not rest. And because the researchers held the activity fixed, this is not a story about how long the break was or what people did during it.

Kang and Kurtzberg measured problem-solving, not rumination. They did not study doomscrolling, spirals, or feelings. What their study establishes is narrower and stranger than the headline. A break can feel like a break and measure like no break at all.

Why the content skews grim#

Robertson and colleagues (2023, Nature Human Behaviour) ran a registered report across roughly 105,000 headline variants and 5.7 million clicks. Each additional negative word in a headline of average length raised click-through by about 2.3%. Positive words lowered it.

Two things that paper does not say, both of which you will hear people claim it does. It does not show that the algorithm feeds you doom. It is about what humans click, which is a different and less comfortable finding. And the authors are explicit that preferring negative headlines does not mean readers felt worse for reading them.

The connection this course is actually making#

Here is the argument, and the parts of it that are mine rather than anyone’s data are marked.

The week 5 finding was specific. Incubation happened during an undemanding task, and did not happen during a demanding one, during rest, or during no break. A scroll is not undemanding. It is a stream of novel, emotionally loaded, attention-capturing material, much closer to the demanding condition than the dishes.

Which would mean scrolling does not fail as a break by being a waste of time. It fails by being the wrong kind of occupied. It takes the one slot where the loop could have finished and fills it with the one thing that guarantees it won’t.

The blank itself is what a fully occupied attention feels like from the inside. Nothing has been processed. The loop has been paused, which feels like relief and is why we keep doing it, and it is waiting exactly where you left it. Leroy’s attention residue, from week 5, predicts precisely that. You left an unresolved thing to go and do something else, and unresolved is the condition where residue is worst.

That paragraph is a hypothesis this course finds persuasive, not a finding. Nobody has connected Kang and Kurtzberg to Baird, or run the study that would test it. The “mind goes blank” experience in particular has, as far as we can find, no research behind it whatsoever. It is a thing people report and nobody has measured. If you want a genuinely open question to work on, that is the best one in this course.

What this lecture will not tell you#

The honest part, and it cuts against the lecture’s own thesis.

• The effects are small. Orben and Przybylski (2019) ran a specification curve across data from 355,358 adolescents and found digital technology use explained at most about 0.4% of variance in wellbeing. Anyone telling you phones are the defining catastrophe of your generation is not describing this literature.
• Self-reported screen time is unreliable, and unreliable in the direction that inflates these findings.
• The causation is genuinely unresolved, and this is the big one. Anxious, ruminating people plausibly scroll more, the spiral causing the scrolling rather than the reverse. Within-person longitudinal studies are mixed. Some find no link at all, others find browsing predicting worse mood hours later. So “doomscrolling has made overthinking worse” is a claim this course cannot make, and does not make. What it can say is that the two travel together and that the felt relief is not doing what it seems to.
• The passive-scrolling story is contested. The most cited experiment here, Verduyn and colleagues (2015), found passive use predicted declines in wellbeing where active use did not, but the mechanism they measured was envy, not numbing. And Valkenburg and colleagues (2022) argue the whole active/passive distinction doesn’t hold up. There is a published adversarial exchange on this that ends unresolved, which is what an honest open question looks like.

Discussion#

• The blank is unstudied. Describe yours precisely enough that someone could design a study on it. What would they measure?
• If anxious people scroll more and scrolling makes it worse, you get a loop that needs no external cause. How would you tell that apart from the simple story in either direction?
• Every technique from How you stopped that worked involved ending something. A timer, sending the message, going to sleep. Scrolling ends nothing. Is that the whole explanation?
• Given a 0.4% variance figure, is any of this worth a lecture? Make the case against this lecture.