Why removal works when willpower doesn’t.

The case for this phone rests on published research, and research deserves to be shown, not gestured at. This page lays out what the evidence says, what it does not say, and the popular numbers we refuse to use. Every study is named. Where we extrapolate, we say so.

It was built to be hard to stop.

In 1971, long before the smartphone, the economist Herbert Simon saw the trade coming: “a wealth of information creates a poverty of attention.” What he could not have predicted is that the people building the information would one day describe their own work in the past tense, on the record.

“The thought process that went into building these applications … was all about: ‘How do we consume as much of your time and conscious attention as possible?’”

“It’s a social-validation feedback loop … exploiting a vulnerability in human psychology.”

“The inventors, creators … understood this consciously. And we did it anyway.”

He was not confessing to a secret. A 2023 advisory from the US Surgeon General names push notifications, autoplay, infinite scroll, and algorithmic personalization as features designed to maximize engagement. Aza Raskin, who invented infinite scroll, has said publicly and repeatedly that he regrets it. And the mechanics can be measured one at a time: in a 2025 controlled experiment, researchers had streaming viewers turn off a single feature, autoplay on Netflix, and daily viewing fell by about 21 minutes.

Your brain isn’t broken. It’s predictable.

Neuroscience separates wanting from liking: the dopamine system drives seeking, while enjoyment lives elsewhere. That is the precise version of a familiar experience, scrolling long after the fun has stopped. The same system fires hardest not on reward but on the gap between what you expected and what arrived, which makes unpredictability the active ingredient. And the brain treats information itself as a reward: the circuitry responds to “what comes next” before anything has arrived.

Behavioral science adds the schedule. Rewards that arrive unpredictably produce the most persistent behavior known to the field, a principle established in the 1950s. Feeds run on that schedule. Whether a feed is fairly called a slot machine is an analogy, and we treat it as one; the schedule itself is not in dispute. Repetition then hands the behavior over to habit, where checking is cued by situation rather than decided, which is why it can feel involuntary.

None of this is a defect. It is standard human equipment, and the design described above is aimed straight at it.

Willpower was the wrong tool.

For years the story was that self-control is a muscle that tires out, so the fix for the feed was to try harder. Then twenty-three laboratories ran the experiment together, on more than two thousand people, and the effect came out close to nothing. What actually predicts self-control is quieter: the people who look disciplined mostly arrange their situations so the fight never starts, a finding the field now treats as its best answer.

Nor can you simply watch yourself out of it. Across 106 comparisons, self-reported phone use lines up with what the phone actually logs at a correlation of 0.38. People do not know their own usage, which is why tools that ask you to monitor yourself start from behind, and why changing the default beats keeping score.

What happens when the pull is removed.

In 2025, researchers ran the direct test. Four hundred sixty-seven people agreed to block the mobile internet on their own phones for two weeks, with calls and texts still working. Only 25.5% kept the block active for at least ten of the fourteen days; these were volunteers who had signed up to reduce their phone use, and the block came with an off switch. Even so, across the whole group, sustained attention improved (d=0.23), mental health improved (d=0.56), and day-to-day wellbeing improved (d=0.45), with the gains still measurable two weeks after the block lifted.

That result is not alone. Across 32 randomized trials with 5,544 participants, reducing or pausing social media improved depressive symptoms by a small but reliable amount (g=0.25). A four-week Facebook deactivation raised wellbeing modestly, and the freed time went to offline life, not to other apps. Honesty requires the next sentence too: when deactivation was repeated at far larger scale in 2024, the measured effect came out smaller. In this literature, bigger and better studies tend to find smaller effects, and a page like this one should say so.

Two more findings turn all of this into a product decision. First, the gains do not outlive the rules that produced them: in a 2025 randomized trial, three weeks of reduced use improved mood, and everything drifted back toward baseline within six weeks of the restriction ending. Second, the most reliable pattern in the behavioral literature is that defaults move behavior more than persuasion does. Flip organ-donation consent from a form you must find to a default you must decline, and agreement goes from 42% to 82% on identical people. Make retirement enrollment automatic and participation roughly doubles. Hard, externally set commitments outperform soft, self-imposed ones.

Put together: the benefits of removal are real and moderate; they last as long as the removal lasts; and removable rules mostly do not last. That is why the rule on the Kala Phone is not a setting. It is the default, and it does not expire.

What the evidence does not say.

• That phones cause depression. In the largest analysis to date, across 355,358 adolescents, technology use explained at most 0.4% of the variance in wellbeing, and whether the small association is causal is a live dispute among careful researchers. We build nothing on that claim.
• That this is an addiction in the medical sense. No smartphone or social-media addiction diagnosis exists in DSM-5 or ICD-11; the closest entries are scoped to gaming. We say compulsive or habitual use, because that is what the evidence supports.
• That the effects are dramatic. The honest range in randomized trials is real and moderate, and it shrinks in the biggest studies. A phone will not transform a life by itself. It changes the default the life runs on.
• That the Kala Phone itself has been studied. It hasn’t. What has been studied is what happens when the pull is taken away, and what happens when the rule is optional. We extrapolate from that component evidence, and this page is the extrapolation, shown in full.
• That blocking the phone can’t push the habit elsewhere. No study yet tests whether restricting a phone moves the scrolling to a laptop or tablet. It is the sharpest open question for a product like ours, and we would rather name it here than have you discover it.

Numbers we refuse to use.

This literature is full of confident numbers that do not survive a check, and several of them circulate in marketing for products like ours. If you ever catch one of these on a Kala Phone page, we broke our own rule.

• “You check your phone 96 / 150 / 352 times a day.” Every circulating count traces to vendor-commissioned self-report surveys; one vendor reported 96 in 2019 and 352 three years later, and self-reports barely track what phones actually log.
• “Attention spans have shrunk to eight seconds.” No study supports it, and the goldfish comparison is folklore. The widely quoted 47-second figure measures how often knowledge workers switch computer windows, which is not an attention span.
• “It takes 23 minutes and 15 seconds to refocus.” Traces to a 2006 magazine interview, not a paper.
• “Distraction costs the economy $650 billion.” One defunct firm’s 2005 survey, mutated by years of blog citation.
• “The mere presence of your phone drains your brainpower.” The 2017 finding failed its preregistered replication in 2022. We would love it to be true. It did not hold.
• “Scrolling gives you dopamine hits.” Dopamine tracks wanting and prediction error, not little jolts of pleasure. The people who built the feeds have used the phrase to describe what they intended. We do not use it to describe your brain.
• Any number for grayscale or minimal launchers. Grayscale rests on three small studies, and the most careful of them found most users disliked it; we ship it off by default and claim nothing for it. Minimal launchers have no controlled study at all; ours is design logic, honestly labeled, not a measured therapy.

Sources.

• Hagger et al. (2016), “A Multilab Preregistered Replication of the Ego-Depletion Effect,” Perspectives on Psychological Science 11(4): 546-573. PMC
• Duckworth, Gendler and Gross (2016), “Situational Strategies for Self-Control,” Perspectives on Psychological Science 11(1): 35-55. SAGE
• Fujita, Trope, Liberman and Levin-Sagi (2006), “Construal Levels and Self-Control,” Journal of Personality and Social Psychology 90(3): 351-367. PubMed
• Castelo, Kushlev, Ward, Esterman and Reiner (2025), “Blocking mobile internet on smartphones improves sustained attention, mental health, and subjective well-being,” PNAS Nexus 4(2): pgaf017. DOI
• Burnell et al. (2025), meta-analysis of 32 randomized trials of social-media reduction, N=5,544. ScienceDirect
• Allcott, Braghieri, Eichmeyer and Gentzkow (2020), “The Welfare Effects of Social Media,” American Economic Review 110(3): 629-676. AEA
• The 2024 large-scale deactivation experiment (Facebook n=19,857, Instagram n=15,585), PNAS. PNAS
• Three-week smartphone-reduction randomized trial with six-week follow-up, BMC Medicine (2025). BMC
• Johnson and Goldstein (2003), “Do Defaults Save Lives?,” Science 302: 1338-1339.
• Madrian and Shea (2001), “The Power of Suggestion: Inertia in 401(k) Participation and Savings Behavior,” Quarterly Journal of Economics 116(4): 1149-1187.
• Bryan, Karlan and Nelson (2010), “Commitment Devices,” Annual Review of Economics 2: 671-698. Annual Reviews
• Ariely and Wertenbroch (2002), “Procrastination, Deadlines, and Performance,” Psychological Science 13(3): 219-224.
• Berridge and Robinson (2016), “Liking, Wanting, and the Incentive-Sensitization Theory of Addiction,” American Psychologist 71(8): 670-679. PDF
• Schultz, Dayan and Montague (1997), “A Neural Substrate of Prediction and Reward,” Science 275: 1593-1599.
• Bromberg-Martin and Hikosaka (2009), “Midbrain Dopamine Neurons Signal Preference for Advance Information about Upcoming Rewards,” Neuron 63(1): 119-126. PMC
• Ferster and Skinner (1957), Schedules of Reinforcement.
• Wood, Quinn and Kashy (2002), “Habits in Everyday Life,” Journal of Personality and Social Psychology 83(6): 1281-1297; Yin and Knowlton (2006), Nature Reviews Neuroscience 7: 464-476.
• Parry et al. (2021), “A systematic review and meta-analysis of discrepancies between logged and self-reported digital media use,” Nature Human Behaviour 5: 1535-1547.
• Orben and Przybylski (2019), “The association between adolescent well-being and digital technology use,” Nature Human Behaviour 3: 173-182.
• Ruiz Pardo and Minda (2022), “Reexamining the ‘brain drain’ effect,” Acta Psychologica. ScienceDirect
• Sean Parker, on the record at an Axios event, November 9, 2017. Axios, CNBC
• US Surgeon General’s Advisory, “Social Media and Youth Mental Health” (2023). HHS (PDF)
• Autoplay controlled experiment, PACMHCI (2025). arXiv
• Reporting on Aza Raskin’s testimony, KOB (2026). KOB
• Herbert Simon (1971), “Designing Organizations for an Information-Rich World.”

Last checked July 29, 2026. If a citation here is wrong or has been superseded, tell us and we will fix it. The same standard we ask of the claims.