Why Paying Attention May Be Better Than Trying to Impress

In 2017, three relationship researchers gave machine-learning models the sort of information a dating service might dream about: more than 100 measures of each person's traits and romantic preferences. Then the participants went speed dating.

The models predicted a limited portion of the differences in people's general tendencies to desire others and to be desired. What they could not predict was which particular person would uniquely click with which other particular person.

That failure contains the most honest answer to the question of how to attract someone.

There is no reliable switch you can flip inside another human being. Attraction is partly influenced by qualities that many people notice, but a substantial part belongs to the encounter itself: two specific people talking, reacting and discovering how it feels to be together.

Research does not establish a single best attraction strategy. It does, however, suggest one evidence-backed way to improve an interaction: stop treating the conversation as a performance and become genuinely responsive to the person in front of you.

Attraction has more than one moving part

Researchers studying speed dating often divide romantic desire into three components. An actor effect describes whether someone tends to like many or few people. A partner effect describes whether someone tends to be liked by many or few people. A relationship effect is the intriguing one: the extra attraction one particular person feels for another that cannot be explained by either person's general tendencies.

In the machine-learning study, the models predicted between 4% and 18% of actor-effect variance and between 7% and 27% of partner-effect variance. Despite having more than 100 pre-date measures to work with, they did not successfully predict relationship-specific desire before the dates occurred.

This does not prove that chemistry is mystical or random. It shows that information about two people considered separately is not equivalent to information about what happens when they meet.

That distinction gets lost in a great deal of dating advice. Becoming more broadly appealing may increase the opportunities someone receives. It cannot ensure that one particular person will feel attraction.

Yes, appearance matters

Any evidence-based account has to acknowledge the obvious. In a 2009 speed-dating study of 108 college students, Shanhong Luo and Guangjian Zhang compared similarity, reciprocity, attachment-related characteristics and physical appearance. A partner's physical attractiveness was the strongest measured predictor of initial attraction for both men and women.

Other live-dating studies complicate the familiar claim that appearance matters vastly more to men. Men and women often report different priorities when describing an ideal partner. Yet Paul Eastwick and Eli Finkel found that those stated differences did not predict what generated desire for the actual people participants met during speed dating.

Looks influence first impressions. That result does not produce a universal beauty formula, however, and it says little by itself about who will form a satisfying relationship. It also leaves much of attraction unexplained, especially the pair-specific component that the prediction models did not see coming.

Profiles are not complete forecasts

The prediction findings suggest a reason not to treat profiles and checklists as complete forecasts: pair-specific attraction becomes visible only after people interact.

In two experiments, Harry Reis and his colleagues varied how long previously unacquainted people interacted. More interaction produced more interpersonal attraction. The change was connected to growing comfort, satisfaction, perceived knowledge and, importantly, the feeling that the other person was responsive.

These experiments involved same-sex strangers and measured interpersonal rather than romantic attraction. They did not directly compare profiles with in-person dating, and they do not prove that repeated contact will turn an uninterested acquaintance into a romantic partner. Repeated unwanted contact is harassment, not a psychological technique.

What the research does show is that liking can develop through positive interaction rather than arriving fully formed at first sight.

That fits another striking result. Across seven samples totaling 1,897 people, roughly two thirds of respondents who were in relationships said the relationship had started as a friendship.

The number carries substantial caveats. The reports were retrospective, the samples came from Canada and the United States, and people may define friendship differently. It certainly does not mean that pretending to be someone's friend is a clever route around rejection. Friendship is not a waiting room in which romantic interest is eventually owed a turn.

The more defensible inference is that recurring social environments can give compatibility time to become visible. Clubs, classes, volunteer groups, sports, communities and networks of friends allow people to encounter qualities that a profile cannot display.

Curiosity can improve the conversation

Once two people are talking, asking questions is one behavior with unusually direct evidence behind it.

In live getting-acquainted experiments, Karen Huang and her colleagues instructed some participants to ask more questions. Their partners liked them more and perceived them as more responsive. Follow-up questions were especially useful because they depended on hearing and responding to the previous answer.

The researchers also analyzed face-to-face speed dates. People who asked more follow-up questions were more likely to receive an offer of a second date. That portion of the study was observational, so it cannot establish that the questions caused the romantic interest. The separate experiments support a causal effect on interpersonal liking, not proof of a universal romantic technique.

An independent audit later found minor reporting errors in the paper. A correction published by the American Psychological Association in 2025 stated that every substantive hypothesis-test result and the paper's conclusions remained unchanged.

The finding is easy to turn into bad advice. Conducting an interrogation is not attractive. A stream of questions can feel like an intake form if the asker never reacts, shares anything or gives the other person room to choose what to reveal.

A genuine follow-up question communicates something more specific: I heard what you said, noticed a detail and would like to understand it better. Relationship researchers describe the broader experience as responsiveness, the sense that another person understands, validates and cares about what you communicate.

You also have to become knowable

Listening is only half of getting to know someone. A classic meta-analysis by Nancy Collins and Lynn Miller found three intertwined relationships: people who disclosed more intimately tended to be liked more, people disclosed more to those they already liked, and disclosing to someone could itself increase liking.

That is a feedback loop, not permission to unload an entire personal history on a stranger. The effects depended on context and the kind of disclosure involved. Intimacy offered too early or inappropriately can repel rather than connect.

A better conversation usually moves in increments. Ask something real. Listen to the answer. Follow what genuinely interests you. Offer something of your own at a comparable depth. Notice whether the other person does the same.

This is how two people generate the information that no pre-date questionnaire contains.

Specific interest is different from wanting everyone

The science of reciprocal liking contains a useful wrinkle. People may appreciate being liked, but attention can lose its meaning when it appears to be distributed indiscriminately.

In a study of 156 undergraduate speed daters, pair-specific romantic desire was positively reciprocated. A participant's unique interest in one partner was associated with that partner's unique interest and sense of chemistry. Participants who expressed high romantic desire for nearly everyone, meanwhile, tended to be desired less.

The study does not establish a universal law, and it does not support manufacturing scarcity or pretending to be aloof. It suggests that interest may be appealing partly when it feels specific: you noticed something about me, rather than you want someone and I happen to be available.

If an interaction feels promising, the practical move is refreshingly ordinary. Make your interest legible. Suggest another conversation or a specific activity. Give the person an easy, pressure-free opportunity to respond.

Then believe the response. Mutuality is not a finishing touch on attraction; it is the point. A refusal, a pattern of evasion or a lack of enthusiasm is not an invitation to deploy more exposure, more questions or a better trick.

The closest thing to an answer

Much of this research studied young, Western men and women considering opposite-sex partners, the same narrow framework embedded in the phrase “the opposite sex.” Its averages should not be treated as laws governing every gender, orientation, culture or age.

Within those limits, the evidence supports possibilities rather than a secret formula.

Spend time in settings where you can meet people you genuinely like. Treat profiles as introductions, not compatibility verdicts. Ask follow-up questions because you care about the answers. Share enough of yourself to become knowable. If your interest is specific, express it clearly and without pressure. Look for enthusiasm coming back.

None of those behaviors can make one chosen person attracted to you. They can create a better interaction and make mutual attraction easier to recognize.

The goal is not to conquer a demographic category. It is to find the rare person whose attention brightens in your company, and whose company does the same to you.