Reinforcement, reward and dopamine
Dopamine mainly signals the gap between expected and actual reward; using it to sell unpredictable behaviour in dating goes far beyond the research.
Evidence: PlausiblePlausible: One study, or indirect evidence from a nearby question. This is the weakest level among the claims below.
What the research shows
Research, mostly in monkeys and rodents, finds that dopamine neurons mainly signal reward prediction errors, the gap between what was expected and what arrived, and that dopamine matters more for wanting than for enjoying. A review of reinforcement-schedule research finds that intermittent reward does not reliably make behaviour harder to stop: results depend on the schedule and can reverse. One study of 75 women who had left abusive relationships found that intermittent abuse and power imbalance went with stronger attachment to the former partner, which describes harm. None of these findings was tested as a way to create attraction. What you can take from it: strong wanting is not the same as enjoyment or a good connection.12345
What dating marketing says it shows
A typical sales line, written for this page
A typical sales line promises: "Make your attention unpredictable and her dopamine will do the rest." Pages like this present dopamine as a pleasure switch and slot-machine rewards as a way to build attachment.
What doesn't follow
The reward research was done in animals and lab tasks and was never tested as a way to create attraction, and the one study linking intermittent treatment to attachment is about abuse: it describes harm, not a method.
Limits
Dopamine findings come mostly from animal recordings and rat experiments; the wanting-versus-liking view is influential but disputed. The schedule review is theoretical, by a proponent of one framework, and does not address relationships. The attachment study is one correlational, self-report study with no non-abusive comparison group.
Sources
Each numbered mark above opens the evidence behind that statement: the label, what it does not show and where it comes from.
Evidence: Supported with caveatsDopamine neurons mainly signal reward prediction errors, the gap between what was expected and what arrived, which drives learning about rewards; they are not simply a pleasure signal.
Why this label
Several studies agree, with stated limits such as student samples or lab settings. Schultz's review describes dopamine neurons firing more for better-than-predicted rewards, staying at baseline for fully predicted ones and dipping for worse-than-predicted ones. Berridge and Kringelbach add that dopamine may not generate pleasure itself.
What it does not show
- Recordings are mostly from monkeys and rodents; human evidence is indirect.
- Prediction error is one component of a more complex dopamine response and other accounts of dopamine function exist.
- Says nothing about attraction between people.
Who it applies to
A pattern across groups of people, not a statement about any one person.
Sources
- Schultz, W. (2016). Dopamine reward prediction error coding, Dialogues in Clinical Neuroscience.
- Berridge, K. C., Kringelbach, M. L. (2015). Pleasure Systems in the Brain, Neuron.
Evidence: Supported with caveatsThe claim that dopamine equals pleasure is contested: Berridge's research indicates that dopamine is more important for wanting (motivation to pursue) than for liking (enjoyment).
Why this label
Several studies agree, with stated limits such as student samples or lab settings. Rats with almost all accumbens dopamine depleted still showed normal liking reactions to sweet taste. Berridge and Kringelbach later summarise liking as generated by small limbic hot spots and wanting by a larger distributed system, adding that textbook dopamine-pleasure ideas may be wrong.
What it does not show
- Mostly rat evidence, including taste-reactivity measures; human wanting and liking are measured differently.
- The incentive-salience view is influential but disputed; rival accounts emphasise learning, effort or motivation more generally.
- Wanting and liking are technical terms and can differ from everyday use.
Who it applies to
A pattern across groups of people, not a statement about any one person.
Sources
- Berridge, K. C., Robinson, T. E. (1998). What is the role of dopamine in reward: hedonic impact, reward learning, or incentive salience?, Brain Research Reviews.
- Berridge, K. C., Kringelbach, M. L. (2015). Pleasure Systems in the Brain, Neuron.
Evidence: Supported with caveatsIntermittent reinforcement does not simply make behaviour harder to extinguish: in reinforcement-schedule research, resistance to extinction depends on the schedule and design, and effects can reverse.
Why this label
Several studies agree, with stated limits such as student samples or lab settings. Nevin's review notes that across multiple schedules, resistance to extinction rises with reinforcer rate, while the partial-reinforcement extinction effect opposes that, and single-schedule results run opposite to multiple-schedule ones.
What it does not show
- Theoretical review of laboratory reinforcement-schedule research; the abstract does not show human evidence.
- Review by a proponent of one theoretical framework (behavioral momentum).
- Does not address relationships or attraction.
Who it applies to
Depends on the setting and the people involved.
Source
- Nevin, J. A. (2012). Resistance to extinction and behavioral momentum, Behavioural Processes.
Evidence: PlausibleResearch on 'traumatic bonding' studied abusive relationships: in one study of 75 women who had left them, intermittent abuse and power imbalance predicted stronger attachment to the ex-partner. It describes harm, not a method for creating attraction.
Why this label
One study, or indirect evidence from a nearby question. Dutton and Painter tested the idea that intermittent maltreatment produces strong attachment. Relationship variables explained 55% of the variance in attachment six months after leaving. The reinforcement-schedule explanation is a theoretical interpretation of correlational data.
What it does not show
- Single correlational, self-report study of women recently separated from abusive partners; no non-abusive comparison group.
- Attachment after abuse is a trauma-related response, not evidence that unpredictability increases healthy attraction.
- Must never be framed as a technique; the app's guardrails exclude manufactured insecurity and coercion.
Who it applies to
Depends on the setting and the people involved.
Source
- Dutton, D. G., Painter, S. (1993). Emotional Attachments in Abusive Relationships: A Test of Traumatic Bonding Theory, Violence and Victims.
Evidence: PlausibleThe brain and memory effects that attraction marketing borrows were studied in other settings and were not tested as ways to create romantic attraction: dopamine in animal reward tasks, the Zeigarnik effect in recall of interrupted tasks, and novelty in couples who already know each other.
Why this label
One study, or indirect evidence from a nearby question. Demonstrates: dopamine neurons mainly signal reward prediction errors (see reinforcement-reward-dopamine-c1, c2 and c3), and uncertain reward cues drew more cue-directed wanting in rats; a 2025 meta-analysis found no memory advantage for interrupted tasks (see zeigarnik-rumination-c1); shared novel, self-expanding activities went with relationship benefits in established couples (see anticipation-novelty-c3 and c4). Marketing framing: leaving her 'hanging', anticipation and dopamine hacks create attraction (common framing, unattributed).
What it does not show
- Mixes animal, lab-task and couples evidence; no direct test on attraction to a specific person
- Animal and task findings do not transfer to dating by default
- Zeigarnik meta-analysis covers recall, not attraction
- Novelty studies cover established couples, mostly self-report
- No large registered replication of any of these applied to attraction found
Who it applies to
Depends on the setting and the people involved.
Sources
- Schultz, W. (2016). Dopamine reward prediction error coding, Dialogues in Clinical Neuroscience.
- Berridge, K. C., Kringelbach, M. L. (2015). Pleasure Systems in the Brain, Neuron.
- Anselme, P., Robinson, M. J. F., Berridge, K. C. (2013). Reward uncertainty enhances incentive salience attribution as sign-tracking, Behavioural Brain Research.
- Ghibellini, R., Meier, B. (2025). Interruption, recall and resumption: a meta-analysis of the Zeigarnik and Ovsiankina effects, Humanities and Social Sciences Communications.
- Tomlinson, J. M., Hughes, E. K., Lewandowski, G. W., Aron, A. (2018). Do shared self-expanding activities have to be physically arousing?, Journal of Social and Personal Relationships.