Why people watch, what sets off craving, and what works better than willpower

Why people say they watch

In March 2017, researchers asked 1,082 adults in Hungary how often they watch porn for each of eight reasons, on a scale from 1 for never to 7 for all the time¹. Sexual pleasure came out far ahead, with an average of 4.8¹. Watching to relieve stress averaged 2.5¹. Watching to push feelings away averaged 1.9, the lowest of the eight¹.

A much larger study, run online in 2021 and 2022, asked 75,117 adults in 42 countries and found the same order at both ends². Sexual pleasure came first and pushing feelings away came last². The usual reason for watching is the obvious one.

The more useful question is which reason goes with trouble. Bőthe and colleagues measured problematic use on an 18 item scale and put all eight reasons into one model side by side¹. Stress relief had the strongest link, a standardised beta of 0.43, which the authors call moderate¹. Pushing feelings away came next at 0.20, then boredom at 0.14 and fantasy at 0.12¹. Sexual pleasure, the most common reason, had the weakest link of the five at 0.09¹. Curiosity, self-exploration and lack of sexual satisfaction showed no significant link at all¹.

That turns the obvious reading around. The reason most people give tells you little about who struggles. A less common reason, reaching for porn to take the edge off a hard day, tells you much more¹. The survey caught one moment in time, so it cannot say that stress causes problem use, and it drew young volunteers recruited online¹. It does point to where a plan should aim. A goal like "watch less" says nothing about the evening when work has gone badly and the phone is on the bed.

A cue sets off the craving

Stress is one kind of cue. Addiction research has studied cues for decades by showing people reminders of their drug in a lab, then neutral images, and comparing the two. Carter and Tiffany pooled 41 of these studies in 1999, covering people addicted to alcohol, nicotine, cocaine or heroin³. Craving rose sharply after drug cues, with an effect size of 0.92³. The body answered too. Sweat gland activity rose with an effect of 0.40 and heart rate with 0.26, and skin temperature fell by 0.24³.

How far craving rose depended on the drug. It rose least for alcohol, at 0.53, and most for nicotine, cocaine and heroin, at 1.18 to 1.29³.

None of those studies used sexual cues, and there is a reason. High reactivity to sexual cues is not limited to people with pathological sexual behaviour, so a 2018 review of cue studies in behavioural addictions left porn studies out⁴. That review gives no pooled figure for porn cues⁴. We are reading across from drugs here.

Craving in a lab would matter less if it faded at the door. A 2022 review by Vafaie and Kober pooled 237 studies with 51,788 people to ask whether craving predicts what people do next⁵. Craving set off by a cue went with higher odds of later use or relapse, an odds ratio of 3.01⁵. That figure is per step up each study's own craving scale, and the scales differ, so treat the number as rough⁵. The direction is clear. The cue comes, the craving follows, and the craving goes with using again.

Holding on by willpower

The common answer to a craving is to resist it harder. For years psychology had a theory that made resisting sound like a fuel gauge. Self-control was a limited resource, and spending it on one task left less for the next. The classic experiment gave one group a tiring self-control task, like crossing out letters by a rule, then measured how both groups did on a second task⁶.

In 2010 Hagger and colleagues pooled 198 published tests and found an effect of 0.62⁶⁷. If that held, the evening after a hard day would be the worst time to rely on self-control, and there would be little to do about it.

It did not hold. In 2015 Carter and colleagues went looking for unpublished studies as well⁷. Of their 116 experiments, 48 had never been published, and once they corrected for missing studies the effect fell from 0.43 to 0.24 or less⁷. Then the field ran the experiment the hard way. In 2016, 23 labs ran one fixed version with 2,141 people and filed their plans before collecting any data⁸. The effect was 0.04⁸. In 2021, 36 labs tested 3,531 people with two versions, on plans filed before the analysis, and got 0.06 from the 2,461 they analysed⁹.

Near zero is not zero. In the 2021 study, people who had done the tiring task gave up on one task about 27 seconds sooner, an effect of 0.15⁹. These were lab tasks, not urges in daily life. But an effect of 0.62 that shrinks to 0.06 under fair testing should change how much weight anyone puts on willpower as a tank that empties⁶⁹. It should also change how people judge themselves after a slip. "I ran out of willpower" explains less than it seems to.

Deciding in advance

If resisting harder is not the answer, deciding ahead of time may be. An if-then plan names a situation and the response to it in advance. "If I am alone with my phone after 11, I will leave it in the kitchen." The plan is tied to the cue, the same cue that would otherwise set off the craving.

Gollwitzer and Sheeran pooled 94 tests in 2006 that compared people who made if-then plans with people who did not, 83 of them experiments¹⁰. The plans helped people reach their goals, with an effect of 0.65¹⁰. On the usual scale 0.2 is small, 0.5 medium and 0.8 large¹⁰. In 2016 Toli, Webb and Hardy looked only at people with mental health problems, 28 studies with 1,636 people¹¹. The effect there seemed larger, 0.99, after one very large result was left out¹¹.

That second finding is the one we would hold on to. You might guess that a simple planning method works for people whose lives are going fine and fails for people with mental health problems. In that review it seemed to work at least as well for them¹¹. And in the Hungarian survey, watching to relieve stress was the reason most tied to problem use¹.

Tracking came out lower, but it was still real. Harkin and colleagues pooled 138 studies with 18,398 people and found that prompting people to track their own progress toward a goal had an effect of 0.40, with a 95% interval of 0.32 to 0.48¹². Correcting for possibly missing studies cut it to 0.19, though the authors judge that bias modest¹². It helped more when people wrote their progress down or shared it¹².

Each of these bars comes from a different review, with different goals and different people, so compare them loosely. None of the studies measured pornography use¹⁰¹¹¹². What they share is a design. Most of their tests compared people given the method with people who were not, and the method came out ahead¹⁰¹¹¹².

All at once or bit by bit

A plan still needs a start date, and many people would rather taper off than stop. The cleanest test of that choice we know of comes from smoking. In a trial run from 2009 to 2011, nurses at 31 GP practices in England recruited 697 smokers and told them at random either to stop on a set day or to cut down over two weeks first¹³. The cut-down group aimed to halve their smoking in week one and halve it again in week two¹³. After quit day both groups got the same nicotine replacement and weekly nurse support¹³.

Stopping at once came out ahead at every check. Four weeks after quit day, 49% of the all-at-once group were still off cigarettes, against 39.2% of the cut-down group¹³. At 8 weeks it was 36.6% against 29.2%¹³. At 6 months it was 22% against 15.5%, or 78 of 355 people against 53 of 342¹³. A breath test confirmed every claim to have quit, slips in the first two weeks after quit day did not count, and anyone lost to follow-up counted as smoking¹³.

The authors put the gap down to fewer quit attempts¹³. 71% of the all-at-once group made a quit attempt of at least 24 hours, against 61% of those cutting down¹³. Once people had started, relapse rates were similar¹³. The two weeks of cutting down seem to have been where attempts stalled. Even smokers who said they would rather cut down did no worse when told to stop at once¹³.

This is smoking, not porn, and the trial has limits of its own. 94% of the smokers were white, nobody could be blinded to their group, and the groups also differed in nicotine replacement before quit day, not only in cutting down¹³.

How they found it

• The Hungarian survey asked 1,082 adults how often they watch porn for each of eight reasons, from 1 for never to 7 for all the time. A model then took all eight together and measured how each one tracked problematic use on an 18 item scale.
• Each cue study showed people with an addiction cues tied to their drug and neutral cues, then compared how much they craved and how their body responded. The 2022 review of 237 studies and 51,788 people asked whether that craving predicts later use or relapse.
• The classic willpower experiment gives one group a tiring self-control task, then measures how both groups do on a second task. In 2016, 23 labs ran one fixed version with 2,141 people. In 2021, 36 labs tested 3,531 people with two versions. The 2016 labs filed their plans before collecting data, and the 2021 team filed theirs before the analysis.
• An if-then plan names a situation and the response to it in advance. Each planning and tracking review pooled studies, mostly experiments, that compared people given the method with people who were not.
• Nurses at 31 GP practices in England recruited 697 smokers, who were randomly told to stop on a set day or to cut down over two weeks first. After quit day both groups got the same nicotine replacement and weekly nurse support.
• In the smoking trial, anyone lost to follow-up counted as smoking, and a breath test confirmed every claim to have quit. Slips in the first two weeks after quit day did not count.

What it cannot tell you

• Only the survey of reasons for watching measured pornography use. The cue studies covered alcohol, nicotine, cocaine and heroin, the planning reviews other goals, and the quit trial smoking.
• The survey of reasons was one survey at one point in time, of volunteers in Hungary, mostly young, recruited online. It shows which reasons go with problem use, not that stress causes it. We read its numbers in the authors' accepted manuscript.
• High reactivity to sexual cues is not limited to people with pathological sexual behaviour, so a 2018 review of behavioural addictions left porn studies out. It gives no pooled figure for porn cues.
• The 2022 odds ratio is per step up each study's own craving scale, and those scales differ.
• Near zero, not zero. In the 2021 willpower study, people who had done the tiring task gave up on one task about 27 seconds sooner, an effect of 0.15. These were lab tasks, not urges in daily life.
• Each bar on the willpower and planning charts comes from a different study, so compare them loosely. In the smoking trial 94% were white, nobody could be blinded to their group, and the groups also differed in nicotine replacement before quit day.

In P-Minus

When you log an urge in P-Minus, you tag how you were feeling, and once a pattern shows up the app offers to write a plan for it. Every plan starts from a cue you name, and you attach it before you need it. When you press the help button, it shows the plan whose cue matches the moment. P-Minus also asks for four standing changes to your phone and room, and for you to write, in your own words, what you are keeping out of reach from today.

Sources

1. Bőthe B, Tóth-Király I, Bella N, Potenza MN, Demetrovics Z, Orosz G. Why do people watch pornography? The motivational basis of pornography use. Psychol Addict Behav. 2021;35(2):172-86.
2. Koós M, Nagy L, Kraus SW, et al. Why do people watch pornography? Cross-cultural validation of the Pornography Use Motivations Scale (PUMS) and its short form (PUMS-8). J Sex Res. 2025;62(6):1049-65.
3. Carter BL, Tiffany ST. Meta-analysis of cue-reactivity in addiction research. Addiction. 1999;94(3):327-40.
4. Starcke K, Antons S, Trotzke P, Brand M. Cue-reactivity in behavioral addictions: a meta-analysis and methodological considerations. J Behav Addict. 2018;7(2):227-38.
5. Vafaie N, Kober H. Association of drug cues and craving with drug use and relapse: a systematic review and meta-analysis. JAMA Psychiatry. 2022;79(7):641-50.
6. Hagger MS, Wood C, Stiff C, Chatzisarantis NLD. Ego depletion and the strength model of self-control: a meta-analysis. Psychol Bull. 2010;136(4):495-525.
7. Carter EC, Kofler LM, Forster DE, McCullough ME. A series of meta-analytic tests of the depletion effect: self-control does not seem to rely on a limited resource. J Exp Psychol Gen. 2015;144(4):796-815.
8. Hagger MS, Chatzisarantis NLD, Alberts H, et al. A multilab preregistered replication of the ego-depletion effect. Perspect Psychol Sci. 2016;11(4):546-73.
9. Vohs KD, Schmeichel BJ, Lohmann S, et al. A multisite preregistered paradigmatic test of the ego-depletion effect. Psychol Sci. 2021;32(10):1566-81.
10. Gollwitzer PM, Sheeran P. Implementation intentions and goal achievement: a meta-analysis of effects and processes. Adv Exp Soc Psychol. 2006;38:69-119.
11. Toli A, Webb TL, Hardy GE. Does forming implementation intentions help people with mental health problems to achieve goals? A meta-analysis of experimental studies with clinical and analogue samples. Br J Clin Psychol. 2016;55(1):69-90.
12. Harkin B, Webb TL, Chang BPI, et al. Does monitoring goal progress promote goal attainment? A meta-analysis of the experimental evidence. Psychol Bull. 2016;142(2):198-229.
13. Lindson-Hawley N, Banting M, West R, Michie S, Shinkins B, Aveyard P. Gradual versus abrupt smoking cessation: a randomized, controlled noninferiority trial. Ann Intern Med. 2016;164(9):585-92.

Sources checked October 2026.