Cruel Optimism and Your Attention

Why We Engage · 18 October 2026 · Reviewed by Faisal Zaidi

The book that calls individual focus advice a cruelty also admits, in its own introduction, that nobody ever measured the decline it describes. We went to see what the tested record says, and it lands between the two sides rather than on either.

Answer to the question

What is cruel optimism in attention advice?

It is the charge that offering an upbeat individual fix for a structurally caused problem makes people blame themselves when it fails. Johann Hari presses it in Stolen Focus, published in 2022. The term comes from Lauren Berlant's 2006 article, reaching him through Ronald Purser. Hari concedes in his own introduction that nobody has measured the attention decline his case describes.

Key takeaways

Johann Hari concedes in the introduction to Stolen Focus, published in 2022, that nobody ran the study that would measure a decline in sustained attention, and four searches found no population-level measure four years on.
A 2024 meta-analysis led by Andree Hartanto pooled 166 effect sizes from 53 samples across 33 studies and 4,368 participants and found no detectable effect of a phone's mere presence on cognition.
Victoria Goodyear's 2025 study of 30 English secondary schools and 1,227 pupils aged twelve to fifteen found no evidence of a difference in mental well-being between restrictive and permissive phone policies.
Yu Luo's 2023 meta-analysis of 184 field experiments covering 2,245,373 people found that interventions changing effort outperform interventions changing intrinsic motivation.

The charge is cruel optimism, and it is a serious one. Offer somebody a cheerful individual fix for a problem whose causes are structural, the argument runs, and when the fix fails they will blame themselves rather than the structure. Johann Hari makes that charge in Stolen Focus, published in 2022, against an advice industry that tells distracted people to install a blocker and build better habits. It is the sharpest thing in the book, and the book is better than its reputation, because in its introduction the author has already conceded the thing that would settle the whole dispute. Nobody, he writes, ran the study.

The concession in the introduction

That concession is the most useful sentence in the book, and almost nobody quotes it. The case for a decline in sustained attention is assembled from separate factors, Hari says, because the study that would measure the decline directly was never run. Credit the admission. Then notice that the conclusion travels onward as though it had been measured.

The prescription is where the strain shows. A book that concedes the trend has never been measured goes on to propose banning a business model, legislating a four-day week and rebuilding childhood, and nowhere in the body text is the strength of the prescription conditioned on the weakness of the measurement.

The endnotes are more honest than the body, which is an unusual thing to be able to say about a book. Counter-evidence is repeatedly parked there rather than in the argument, including a qualification on diet failure and a counter-finding on stimulant effects in rats. Hari states plainly that the printed notes are partial, with more on the book's website, and a long stretch of the body carries no notes at all.

Several figures in it cannot be used. The ten-IQ-point claim rests on a study his own note describes as unpublished and privately commissioned. A twenty per cent texting figure has no researcher and no sample attached. A four-day-week result comes from one uncontrolled company trial and a company press claim.

Use the book as a map rather than as a source.

The twelve-cause structure is the part worth keeping, and it is genuinely useful: it names forces that act at once, which is how most real environments work. The charge of cruel optimism sits inside that structure, and it is the claim this piece is about.

A population measure nobody has

What would settle it? A measure of sustained attention taken in the same way, in a large slice of a population, repeated across decades. The conscription-based intelligence series that researchers use to study the rise and fall of test scores is roughly the right shape of thing, though it measures intelligence tests rather than attention, and we have verified only that it exists.

We went looking for the attention equivalent, four different ways, and found nothing of the kind.

So Hari's concession was correct when he made it, and it is still correct now. That is the central fact of this whole argument and it is a negative one.

The nearest thing in the literature is often mistaken for the real thing. Philipp Lorenz-Spreen led a 2019 paper in Nature Communications reporting that across longitudinal datasets covering multiple decades, cultural items receive collective attention in steeper and shorter bursts, driven by more content being produced and consumed. Hari credits the paper's last author throughout; the first author is Lorenz-Spreen.

The unit there is the cultural item, not the person.

It measures how long a topic holds the public's attention, which is a fact about a media environment rather than about anybody's capacity to concentrate.

It measures how long a topic holds the public's attention, which is a fact about a media environment rather than about anybody's capacity to concentrate. Gloria Mark's work gets borrowed the same way. Her 2005 observational study followed twenty-four information workers in detail and found their work heavily fragmented, with 57 per cent of their working spheres interrupted and more than two intervening activities before an interrupted task was resumed. That is a property of a working day. It is not a measurement of an attention span, and twenty-four people is a small base for either.

Her 2008 work should have travelled and did not, and it is a laboratory experiment rather than field observation. Forty-eight participants, most of them German students, worked through a simulated email task. They completed the interrupted versions in less time, with no difference in quality the study could detect, apparently by working faster, and partly by writing shorter emails. The cost showed up as stress, frustration and time pressure instead.

The chain of custody on a phrase

The phrase cruel optimism is itself borrowed twice over. Lauren Berlant, a scholar of literary and cultural theory who died in 2021, published an article called Cruel Optimism in the journal differences in 2006. Ronald Purser carried the idea into the critique of the mindfulness industry, and Stolen Focus takes it from there. We verified Berlant's published record and did not read the article, so nothing here paraphrases her argument.

The other key term in the book has no published source we could find. Attentional pathogenic culture is a phrase the researcher Joel Nigg gave Hari in conversation, and four searches turned up no paper using it. It reaches readers as an interviewee talking, which is a different grade of claim from a published construct, and it should be attributed that way.

Why fuss over custody? Because a phrase that has passed through three hands tends to arrive with its original qualifications stripped off, and in this case the original was a work of theory rather than a finding.

Eyal's argument in his own words

The target of the cruel optimism charge is the individual-remedies approach, and Nir Eyal's Indistractable, published in 2019, is its best-known statement. His own framing is not the one the charge assumes.

Distraction, in his account, is an escape from internal discomfort, and the device is the proximate cause rather than the root one. Blaming the smartphone, he writes, is like blaming the pedometer for the stairs. Time management is pain management. Boredom and the fading of satisfaction are presented as the evolutionary default rather than as a personal failing, which is an argument against self-blame rather than for it.

He also declines the model the charge assumes he is using. He uses Daniel Wegner's white-bear work to dismiss thought suppression, builds on acceptance-based methods instead, and argues against the idea that willpower is a fuel tank that empties. Precommitment is explicitly his last resort rather than his first move, and effort pacts, which raise the cost of the distracting option, come after three other steps.

His workplace chapter is a structural argument outright. Workplace distraction is described there as a symptom of a dysfunctional culture, with Leslie Perlow's consulting work showing that a cycle of constant responsiveness is fuelled by culture rather than by the messaging tool, and psychological safety sitting underneath the other factors. Technology, he writes, is the accelerant.

An author who says the workplace problem is cultural is not a pure individualist.

Two further facts of record belong here together. Eyal discloses investments in several products he recommends, each in a footnote, and states that his recommendations are not advertiser-influenced. And Hari published Eyal's objection inside Stolen Focus alongside his own reply, and posted the full interview audio on the book's website.

Our own criticism of Eyal's work is about an instrument rather than a person. The manipulation matrix in his earlier book Hooked asks a maker whether they would use the product themselves and whether it improves lives, and both axes are answered by the maker about the maker, so the instrument cannot catch anyone who believes their own pitch. Most of Hooked is an argued framework assembled from case studies, and it should be cited as a taxonomy rather than as evidence.

Which sets up the fairest sentence available. Neither the twelve causes nor the four-step hook model has ever been evaluated as a model, and we searched for tests of both.

The individual remedies under test

Here the ledger starts filling, and it fills against almost everybody, including the people the cruel optimism charge was aimed at.

Andree Hartanto led a 2024 meta-analysis on whether the mere presence of a phone drains cognition. It pooled 166 effect sizes from 53 samples across 33 studies, covering 4,368 participants, and found nothing the analysis could detect, with no moderation by demographics, by how attached people were to their phones, or by how the presence was manipulated. The authors go further than most would: there is little reason at present, they write, to think that complete isolation from smartphones in a work environment would improve performance.

The best-known single finding behind that idea did not survive a direct test. Ana Cecilia Ruiz Pardo and John Paul Minda ran a pre-registered replication in 2022, varying both where the phone sat and whether it was switched on, and found no difference between locations on either of two cognitive tasks. A contrary single study was published in 2023, and we have verified its record without reading it; the meta-analysis is the better evidence.

Mindfulness has the one trial that measured sustained attention head on. Donal MacCoon's team randomised 63 adults to eight weeks of mindfulness-based stress reduction or to a structurally matched active control, with assessors blind to group. Attention did not differ between the groups over time.

That null had the power to find a difference, which is what makes it worth printing.

The authors are careful about its edge, noting that another aspect of attention, vigilance, was not addressed. Stepping away from platforms fares no better on the outcomes it has been measured against. Laura Lemahieu's preregistered 2025 meta-analysis pooled ten studies covering 4,674 people and found no effect of social media abstinence on positive affect, negative affect or life satisfaction, and no relationship with how long the abstinence lasted. Note the outcome, which is well-being rather than attention.

Theda Radtke's systematic review of digital detox, meaning planned timeouts from devices, covered 21 studies and 3,625 participants and found effects varying across them, including some negative consequences for well-being. It is a narrative synthesis with no pooled figure, so it cannot be quoted as a verdict either way.

The structural remedies under test

If the cruel optimism argument is right, this is the column that should fill differently. The structural side has fewer tests, and the biggest one went the same way. Victoria Goodyear led the SMART Schools study, published in 2025, comparing 30 English secondary schools, twenty with restrictive phone policies and ten permissive, across 1,227 pupils aged twelve to fifteen. On the primary outcome, a standard mental well-being scale, there was no evidence of a difference between the two groups.

Pupils at restrictive schools did use phones and social media less during the school day. Across whole weekdays and weekends, the difference disappeared. The study is cross-sectional and observational rather than a trial, and its authors say so, concluding that the policies in their current form are not supported by the evidence and require further development.

The one large structural result that did move an outcome came bundled with a cost. Hunt Allcott's team recruited 2,743 Facebook users, then randomised the 61 per cent whose asking price for four weeks off the service sat below about 102 dollars, paying that group to deactivate for the four weeks before the 2018 United States midterm election. The randomised group was therefore a price-filtered, paid subsample, not the recruited sample.

Neither side of this dispute has a comfortable home for that trade-off.

What was tested	Scale	What it showed
A phone merely present	4,368 people, 166 effect sizes	Nothing the meta-analysis could detect
Mindfulness against an active control	63 adults, eight weeks	No difference in sustained attention
Social media abstinence	4,674 people, ten studies	No effect on affect or life satisfaction
School phone policies	1,227 pupils, 30 schools	No difference in mental well-being
Four weeks off Facebook	2,743 recruited, the 61 per cent under the price threshold randomised and paid	Well-being up, news knowledge down

Friction, the one surviving finding

One result survives both columns, and it belongs to neither camp.

Yu Luo led a 2023 meta-analysis organising behavioural interventions by the cognitive process they target, with attention as one of six alongside perception, memory, effort and two kinds of motivation. Across 184 field experiments with real behavioural outcomes, covering 2,245,373 people, interventions that changed effort outperformed interventions that changed intrinsic motivation. The authors flag a possible publication bias.

Stéphanie Mertens's 2022 meta-analysis found the same ordering from a different angle. Across more than 200 studies, more than 440 effect sizes and more than two million observations, choice-architecture interventions produced a small to moderate change overall, and interventions targeting the structure of the choices beat those describing the alternatives, which in turn beat those reinforcing people's behavioural intentions. The authors report a moderate publication bias towards positive results.

Most individual attention advice falls in that last and weakest category.

Then the dissent, which is severe and gets full weight. Maximilian Maier and colleagues published a PNAS Letter in 2022 arguing that the publication bias was the real headline, and that a model-averaging Bayesian reanalysis leaves no evidence of an overall nudging effect, with evidence against the information and assistance categories. It is a Letter rather than a research article. Their own caveat travels with it: almost every category shows heterogeneity, which implies some interventions work even where the average does not.

Note which category they leave undecided rather than refuted. Structure.

Three independent analyses point at friction and away from motivation, and the category that survives a hostile reanalysis is structure.

Three independent analyses, then, point at friction and away from motivation. Angela Duckworth and two co-authors named the mechanism in a 2016 review, which is a theory paper with no sample: situational self-control, meaning choosing or changing your circumstances so that an impulse is weakened before it grows. Their own example is the student who goes to the library without a phone. Their closing observation is the best line in this literature.

> We may underappreciate situational self-control for the same reason it is so effective, namely that by manipulating our circumstances to advantage, we minimise the in-the-moment struggle.

Limits, and our position between them

What this piece cannot tell you is as important as what it can. We cannot say attention has declined, and we cannot say it has not, because the measure does not exist. We cannot print a sample size for the pre-registered replication, because its abstract does not give one. The school study is observational. A ritual study cited in Indistractable was retracted in 2024 and is not used. The twenty-three minute version of the interruption-recovery figure appears nowhere in the published record of either paper it is credited to, so it does not appear here, and nor does the researcher's name attached to it in the book.

Take both sides at their strongest. Hari's is that an environment built to interrupt people is a political fact, that telling individuals to try harder inside it produces self-blame, and that the forces act together, so no single-cause study will ever capture them. Eyal's is that the discomfort people are escaping is real and internal, that the device is downstream of it, that willpower is the wrong model, and that workplace distraction is cultural rather than technological.

Both are describing environments. Neither is really arguing for trying harder.

Our position is that the dichotomy does not hold, and the evidence is what dissolves it. Changing your own situation is an individual act with structural mechanics, and that is the category that keeps doing best in the ledger above. Eyal's effort pacts are friction. Duckworth's situational strategies are friction. Luo's effort interventions are friction too, and the pooled figure of 184 experiments and more than two million people covers all six of her categories, so the effort subset is a fraction of it.

Her effort category and the other team's decision-structure category are different taxonomies pointing the same way, which is the honest claim rather than one construct measured twice. What fails, on both sides of the argument, is anything that works by making people want the outcome more, which is where most attention advice and a good deal of attention policy currently sits.

So the charge of cruel optimism lands, but not where it was aimed. It lands on motivation talk, wherever it comes from, including the structural kind that asks a school or a parliament to want a better outcome without changing a single step. Our articles on both sides of the screen-time argument, on why willpower is not a fuel tank, and on how temptation bundling needs a lock all run into the same wall, as does what cognitive load costs at checkout from the other direction.

Which gives you something to apply, and it is a short test:

Does the change remove a step between you and what you want?
Does it add a step between you and what you do not?
Or does it only ask you to want the right thing more?

The first two have field evidence behind them from pooled literatures covering more than two million people, with the friction subsets a fraction of that total. The hostile reanalysis is no endorsement either: it reports no evidence of an overall nudging effect, evidence against the information and assistance categories, and a heterogeneity caveat that applies right across the categories rather than singling structure out.

The third has a literature that keeps producing nulls and a rhetoric that keeps producing guilt. So the test worth applying to any piece of attention advice, including ours, is simple to state: does it remove a step, add a step, or does it only ask you to want the outcome more than you wanted it yesterday?

Common questions

What is cruel optimism in attention advice?

Does putting your phone away help you concentrate?

Not on the evidence available. A 2024 meta-analysis led by Andree Hartanto pooled 166 effect sizes across 33 studies and 4,368 participants and found no effect of a phone's mere presence on cognition. A pre-registered replication in 2022 also found no difference between phone locations. The authors say isolation from phones at work is unlikely to improve performance.

Do school phone bans improve wellbeing?

The largest English study found no difference. Victoria Goodyear's SMART Schools paper, published in 2025, compared 30 secondary schools and 1,227 pupils aged twelve to fifteen, and found no evidence of a difference in mental well-being between restrictive and permissive policies, and no difference in phone use across whole weekdays. It is observational, and its authors say the policies need further development.

What actually works against distraction?

Friction, on the best available evidence. Yu Luo's 2023 meta-analysis of 184 field experiments found that interventions changing effort beat interventions changing intrinsic motivation. Its pooled total covers all six of her categories, so the effort subset is a fraction of it. Stéphanie Mertens's analysis of over 440 effect sizes found the same ordering, with structural changes beating attempts to reinforce intentions. Remove or add a step rather than trying to want it more.

Sources

Berlant, L. (2006). 'Cruel Optimism.' differences, 17(3), 20-36. doi:10.1215/10407391-2006-009. [Published record verified at OpenAlex and the Duke University Press landing text; sole author; published 1 December 2006 in differences: A Journal of Feminist Cultural Studies, Brown University. ABSTRACT NOT AVAILABLE and the article was NOT read]: this fixes the date of the term and nothing more. Berlant's field was literary and cultural theory, not history. She died in 2021. Do not paraphrase her argument.

Lorenz-Spreen, P., Monsted, B. M., Hovel, P., & Lehmann, S. (2019). 'Accelerating dynamics of collective attention.' Nature Communications, 10(1), 1759. doi:10.1038/s41467-019-09311-w. [Abstract read in full; OPEN ACCESS, PDF at nature.com; author order confirmed, Lorenz-Spreen first, Max Planck Institute for Human Development; Lehmann last, University of Copenhagen]: 'In longitudinal datasets across various domains, covering multiple decades, we find increasing gradients and shortened periods in the trajectories of how cultural items receive collective attention', driven by 'increasing production and consumption of content, resulting in a more rapid exhaustion of limited attention resources'. THE UNIT IS THE CULTURAL ITEM, NOT THE PERSON. Hari credits Lehmann throughout; the first author is Lorenz-Spreen. NOT VERIFIED: the specific datasets, time spans and gradients.

Mark, G., Gonzalez, V. M., & Harris, J. (2005). 'No task left behind? Examining the nature of fragmented work.' Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (CHI '05), 321-330. doi:10.1145/1054972.1055017. [Abstract read in full; author order confirmed at OpenAlex, Gloria J. Mark first, UC Irvine]: detailed observation of 24 information workers. 'We found work to be highly fragmented: people average little time in working spheres before switching and 57% of their working spheres are interrupted... Though most interrupted work is resumed on the same day, more than two intervening activities occur before it is.' THE TWENTY-THREE-MINUTE FIGURE IS NOT IN THIS ABSTRACT and was not located at the published record. N = 24, which is small.

Mark, G., Gudith, D., & Klocke, U. (2008). 'The cost of interrupted work: more speed and stress.' Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (CHI '08), 107-110. doi:10.1145/1357054.1357072. [Abstract read in full; author order confirmed, Mark first, UC Irvine; Gudith and Klocke at Humboldt-Universitaet zu Berlin]: 'context does not make a difference but surprisingly, people completed interrupted tasks in less time with no difference in quality. Our data suggests that people compensate for interruptions by working faster, but this comes at a price: experiencing more stress, higher frustration, time pressure and effort.' NOT VERIFIED: sample size.

Hartanto, A., Lua, V. Y. Q., Kasturiratna, K. T. A. S., Koh, P. S., Tng, G. Y. Q., Kaur, M., Quek, F. Y. X., Chia, J. L., & Majeed, N. M. (2024). 'The effect of mere presence of smartphone on cognitive functions: A four-level meta-analysis.' Technology, Mind, and Behavior, 5(1), 1-14. doi:10.1037/tmb0000123. [Abstract read in full; OPEN ACCESS, PDF at tmb.apaopen.org; author order confirmed, Hartanto first, Singapore Management University]: 166 effect sizes, 53 samples, 33 studies, 4,368 participants. 'the mere presence of smartphone had no significant effect on cognitive outcomes (d = -0.02, SE = 0.02, 95% CI [-0.06, 0.01], p = .246)', not moderated by demographics, trait smartphone dependency, or manipulation or assessment method. Authors' own scope: 'there is little reason at present to think that complete isolation from smartphones in a work environment would improve productivity and performance.'

Ruiz Pardo, A. C., & Minda, J. P. (2022). 'Reexamining the "brain drain" effect: A replication of Ward et al. (2017).' Acta Psychologica, 230, 103717. doi:10.1016/j.actpsy.2022.103717. [Abstract read in full; OPEN ACCESS at ScienceDirect; author order confirmed, Ruiz Pardo first, Western University]: a pre-registered direct replication (OSF osf.io/5fq4r) assigning both smartphone location and power for six conditions, with an automated operation span task and a cue-dependent go/no-go task. 'the brain drain effect did not replicate: there was no difference between smartphone location conditions on performance on either the o-span task or the go/no-go task.' NOT VERIFIED: sample size.

Radtke, T., Apel, T., Schenkel, K., Keller, J., & von Lindern, E. (2022). 'Digital detox: An effective solution in the smartphone era? A systematic literature review.' Mobile Media & Communication, 10(2), 190-215. doi:10.1177/20501579211028647. [Abstract read in full; OPEN ACCESS, PDF at journals.sagepub.com; author order confirmed, Radtke first, University of Wuppertal]: PRISMA systematic review of seven databases, k = 21 studies, total N = 3,625, of which 12 were randomised controlled trials. 'the effects from digital detox interventions varied across studies on health and well-being, social relationships, self-control, or performance... some studies found positive intervention effects, whereas others found no effect or even negative consequences for well-being.' A narrative synthesis with NO pooled effect size.

Lemahieu, L., Vander Zwalmen, Y., Mennes, M., Koster, E. H. W., Vanden Abeele, M., & Poels, K. (2025). 'The effects of social media abstinence on affective well-being and life satisfaction: a systematic review and meta-analysis.' Scientific Reports, 15(1), 7581. doi:10.1038/s41598-025-90984-3. [Abstract read in full; OPEN ACCESS; author order confirmed, Lemahieu first, University of Antwerp]: preregistered; ten studies, N = 4,674, 38 effect sizes. 'The analyses revealed no significant effects of social media abstinence interventions on positive affect, negative affect, or life satisfaction. Relationships between social media abstinence duration and the three outcomes were also non-significant.' Authors' own caveat: 'important methodological differences between studies should be considered when interpreting these results.' NOTE THE OUTCOME: well-being, not attention.

MacCoon, D. G., MacLean, K. A., Davidson, R. J., Saron, C. D., & Lutz, A. (2014). 'No Sustained Attention Differences in a Longitudinal Randomized Trial Comparing Mindfulness Based Stress Reduction versus Active Control.' PLoS ONE, 9(6), e97551. doi:10.1371/journal.pone.0097551. [Abstract read in full; OPEN ACCESS; author order confirmed, MacCoon first, University of Wisconsin-Madison; ClinicalTrials.gov NCT01301105]: eight weeks of mindfulness training measured by a visual continuous performance task, MBSR against the Health Enhancement Program, 'a structurally equivalent, active control condition', randomised, researchers blind to assignment. 63 community participants randomised. 'Our hypotheses were not confirmed... Our study had sufficient power to demonstrate that intervention groups do not differ in their improvement over time in sustained attention performance.' Their limit: 'it is unclear whether mindfulness might positively affect another aspect of attention, vigilance.'

Allcott, H., Braghieri, L., Eichmeyer, S., & Gentzkow, M. (2020). 'The Welfare Effects of Social Media.' American Economic Review, 110(3), 629-676. doi:10.1257/aer.20190658. [Abstract read in full; OPEN ACCESS PDF at aeaweb.org; author order confirmed, Allcott first; Gentzkow last]: a randomised experiment deactivating Facebook for the four weeks before the 2018 US midterm election. Deactivation '(i) reduced online activity, while increasing offline activities such as watching TV alone and socializing with family and friends; (ii) reduced both factual news knowledge and political polarization; (iii) increased subjective well-being; and post-experiment Facebook use. Deactivation reduced post-experiment valuations of Facebook.' NOT VERIFIED: sample size and all effect sizes.

Goodyear, V. A., Randhawa, A., Adab, P., Al-Janabi, H., Fenton, S. A. M., Jones, K., Michail, M., Morrison, B., Patterson, P., Quinlan, J. I., Sitch, A. J., Twardochleb, R., Wade, M., & Pallan, M. (2025). 'School phone policies and their association with mental wellbeing, phone use, and social media use (SMART Schools): a cross-sectional observational study.' The Lancet Regional Health - Europe, 51, 101211. doi:10.1016/j.lanepe.2025.101211. [Abstract read in full; OPEN ACCESS; author order confirmed, Goodyear first, University of Birmingham; ISRCTN77948572; NIHR Public Health Research Programme]: 30 English secondary schools, 20 restrictive and 10 permissive, 1,227 participants aged 12 to 15, primary outcome the Warwick-Edinburgh Mental Well-Being Scale. 'no evidence of a difference between groups (adjusted mean difference -0.48, 95% CI -2.05 to 1.06, p = 0.62)'. Restrictive schools had lower phone and social media time DURING SCHOOL TIME, 'but there was no evidence for differences when comparing usage time on weekdays or weekends'. Authors: 'The findings do not provide evidence to support the use of school policies that prohibit phone use during the school day in their current form.' CROSS-SECTIONAL AND OBSERVATIONAL, not a trial.

Luo, Y., Li, A., Soman, D., & Zhao, J. (2023). 'A meta-analytic cognitive framework of nudge and sludge.' Royal Society Open Science, 10(11), 230053. doi:10.1098/rsos.230053. [Abstract read in full; OPEN ACCESS; author order confirmed at OpenAlex, Luo first, University of British Columbia]: interventions organised along six cognitive processes, attention, perception, memory, effort, intrinsic motivation and extrinsic motivation; meta-analysis of field experiments with real behavioural measures, 184 papers, k = 184 observations, N = 2,245,373, 2008 to 2021. 'interventions changing effort are more effective than interventions changing intrinsic motivation, and nudge and sludge interventions had similar effect sizes.' Authors' caution: 'these results need to be interpreted with caution due to a potential publication bias.'

Mertens, S., Herberz, M., Hahnel, U. J. J., & Brosch, T. (2022). 'The effectiveness of nudging: A meta-analysis of choice architecture interventions across behavioral domains.' PNAS, 119(1), e2107346118. doi:10.1073/pnas.2107346118. [Abstract read in full at the PNAS site; author order confirmed at Crossref and OpenAlex, Mertens first, University of Geneva]: more than 200 studies, over 440 effect sizes, n = 2,148,439. 'choice architecture interventions overall promote behavior change with a small to medium effect size of Cohen's d = 0.43 (95% CI [0.38, 0.48])'. And: 'interventions that target the organization and structure of choice alternatives (decision structure) consistently outperform interventions that focus on the description of alternatives (decision information) or the reinforcement of behavioral intentions (decision assistance).' Authors' own limitation: 'a moderate publication bias toward positive results in the literature'.

Maier, M., Bartos, F., Stanley, T. D., Shanks, D. R., Harris, A. J. L., & Wagenmakers, E.-J. (2022). 'No evidence for nudging after adjusting for publication bias.' PNAS, 119(31), e2200300119. doi:10.1073/pnas.2200300119. [FULL TEXT READ at pnas.org, free access, CC BY; author order confirmed at Crossref, Maier first, University College London; Wagenmakers last; PubMed 35858389. IT IS A LETTER, NOT A RESEARCH ARTICLE]: 'We propose their finding of moderate publication bias (p. 1) is the real headline; when this publication bias is appropriately corrected for, no evidence for the effectiveness of nudges remains.' Mertens et al.'s own sensitivity analysis 'indicates that the true effect size could be as low as d = 0.08 (if publication bias is severe)'. Using model-averaging Bayesian meta-analysis: 'an absence of evidence for an overall effect and evidence against an effect in the information and assistance intervention categories, whereas the evidence is undecided for structure interventions.' Their caveat: 'all intervention categories and domains apart from finance show evidence for heterogeneity, which implies that some nudges might be effective, even when there is evidence against the mean effect.' The PNAS record notes the original article has a reply, which was NOT read.

Duckworth, A. L., Gendler, T. S., & Gross, J. J. (2016). 'Situational Strategies for Self-Control.' Perspectives on Psychological Science, 11(1), 35-55. doi:10.1177/1745691615623247. [Abstract read in full; OPEN ACCESS via PubMed Central PMC4736542; author order confirmed, Duckworth first, University of Pennsylvania; Gendler at Yale; Gross at Stanford]: a review and theory paper with NO sample and no test. 'enacting self-control is not always difficult, particularly when it takes the form of proactively choosing or changing situations in ways that weaken undesirable impulses or potentiate desirable ones', with 'the student who goes to the library without a cell phone' among their own examples. And: 'we may underappreciate situational self-control for the same reason it is so effective, namely, that by manipulating our circumstances to advantage, we are often able to minimize the in-the-moment experience of intrapsychic struggle typically associated with exercising self-control.'

RECORD-ONLY, published record verified but content NOT read: Skowronek, J., et al. (2023). 'The mere presence of a smartphone reduces basal attentional performance.' Scientific Reports, 13(1), 9363. doi:10.1038/s41598-023-36256-4. Included for balance against the meta-analytic null; the meta-analysis is the better evidence.

RECORD-ONLY, published record verified but content NOT read: Bratsberg, B., & Rogeberg, O. (2018). 'Flynn effect and its reversal are both environmentally caused.' PNAS, 115(26), 6674-6678. doi:10.1073/pnas.1718793115. The nearest genuine population-level longitudinal cognitive series, and it measures intelligence test scores in conscription data, not attention. No figure from it is printed.

SEED BOOKS, read in full with reliability audits in the vault: Hari, J. (2022). Stolen Focus: Why You Can't Pay Attention. London: Bloomsbury. [Full read-through, pages 1 to 333 including endnotes, logged at Books/Stolen Focus/Stolen Focus.md]. Eyal, N., with Li, J. (2019). Indistractable: How to Control Your Attention and Choose Your Life. London: Bloomsbury. [Full read-through, pages 1 to 263, logged at Books/Indistractable/Indistractable.md]. Eyal, N., with Hoover, R. (2014/2019). Hooked: How to Build Habit-Forming Products. London: Penguin Business. [Full read-through, pages 1 to 246, logged at Books/Hooked/Hooked.md; the hub's standing rule is that the Hook Model is an argued framework and cannot fail replication because it was never an empirical claim].

NOT FOUND, and we looked: any population-level longitudinal measure of sustained attention (four OpenAlex search formulations); any published source for Joel Nigg's phrase 'attentional pathogenic culture' (three formulations); any peer-reviewed evaluation or empirical test of the Indistractable model or the Hook Model (two formulations); the twenty-three-minute interruption-recovery figure at the published record; a pooled digital-detox effect size on attention specifically.