The outcome changes what ‘prediction’ means
Two real studies with different settings. Neither is a personal relapse calculator or proof that craving and action correspond one to one.
| Evidence | Outcome examined | Limit of the inference |
|---|---|---|
| 2023 repeated daily-life assessments | A smoking lapse during a quit attempt | Within-person association; different scales and time gaps |
| 2012/2013 laboratory cue study | Smoking during a 15-minute opportunity | People not planning to quit; not sustained relapse |
What daily-life research actually links together
A 2023 review examined repeated assessments during quit attempts: how changes in psychological states and context within a person related to smoking lapses. It included 61 studies, of which 19 contributed to meta-analyses. Craving and environmental or social cues showed positive associations with lapses; confidence was reduced by weaknesses in study quality. This is not the same as simply ranking people from ‘low craving’ to ‘high craving’.
The review concerned discrete smoking events, not one uniform definition of sustained relapse. Some assessments were simultaneous, others separated in time; scales and reporting windows differed. Its pooled odds ratio cannot be converted into your percentage chance of smoking, or read as a universal increase for one point on a personal rating. It also did not establish a causal chain in which craving alone produces the event.
A cue effect and a craving–behavior association can differ
In a laboratory study published online in 2012 and in a 2013 issue, the analysis included 207 daily smokers who did not intend to quit within three months; the experiment used six cue-session conditions. After viewing images, they could smoke up to two cigarettes over 15 minutes. Smoking images increased craving compared with neutral images, but cue type did not show significant main differences in whether, how soon or how much participants smoked.
Yet higher craving levels, and increases from before to after viewing the cues, were associated with several later smoking measures. The findings are compatible: comparing assigned images asks a different question from comparing measured urges. Some associations were curved rather than a straight ‘more urge, proportionally more smoking’ line. This short laboratory opportunity was not a long-term quit attempt, and its results do not supply a personal threshold.
Two moments do not make a risk calculator
Consider a fictional person who feels a strong urge at home but does not smoke, then smokes during a familiar break despite a milder urge. Neither event disproves the research association. They show why knowing one feeling is not knowing the whole situation or the eventual behavior. The opposite mistake is to treat the stronger urge as proof that the first moment must end in smoking.
A report saying ‘craving predicted smoking’ needs an outcome and a time window: smoking in the next laboratory period, one lapse later that day, or sustained smoking over follow-up are not interchangeable. A statistically supported association alone also says little about a model's accuracy for new people or settings. There is no validated personal cut-off or risk calculation offered here.
Taking an urge seriously without treating it as destiny
Not being able to forecast your next action does not make a troubling urge unimportant. Stronger or weaker feelings are not a verdict on success, failure or character, and a mild urge is not a certificate of safety. Do not deliberately expose yourself to smoking cues or smoke to test a prediction.
If urges are difficult to manage or you want support with quitting, discuss your circumstances with a qualified professional rather than using a population estimate to choose treatment. In England, NHS Better Health links to local stop-smoking services; confirm access and available support. Elsewhere use qualified local services. This page needs no craving scores, smoking dates or health details.
What to keep in mind
Sources
The central claims on this page were checked against the sources below.
- Perski and colleagues · Addiction: Within-person associations between psychological and contextual factors and lapse incidence: EMA systematic review and meta-analysis (2023)
Sources checked: 2026-10-06
- Shiffman and colleagues · Journal of Abnormal Psychology: Smoker Reactivity to Cues: Effects on Craving and on Smoking Behavior (2012 online; 2013 issue)
Sources checked: 2026-10-06
- NHS: Ready to quit smoking: local support
Sources checked: 2026-10-06
General explanation of craving–behavior evidence, not an individual relapse forecast, a risk score, a diagnosis or treatment selection. Do not test the association by smoking.