A good listener is often described by what they do not do. They do not interrupt, check a phone or turn every story back towards themselves. But silence alone is ambiguous. It cannot tell a speaker whether the other person is attentive, confused, bored or simply waiting for a turn.

A paper published in Communications Psychology in November 2025 looked instead at the visible work of listening. Across two preregistered studies with a combined 646 participants, researchers coded what people actually did during recorded conversations with unfamiliar partners, then compared those behaviours with several markers of connection.

The clearest pattern appeared in a ten-minute conversation designed to encourage personal disclosure. Participants who asked more relevant follow-up questions and offered more verbal validation tended to have quicker conversational exchanges, partners who reported a stronger shared connection, and interactions that outside observers also judged as more connected.

A second study made the story less tidy. During brief small talk, an overall pattern of attentive verbal and nonverbal behaviour predicted connection more consistently than follow-up questions did on their own.

That difference is the heart of the research. A question is not automatically evidence of care, and a nod is not proof of understanding. Listening becomes socially meaningful when several behaviours fit together well enough for another person to recognise attention, comprehension and goodwill.

The researchers coded behaviour, not reputations

The two studies were secondary analyses of earlier intervention datasets collected at the University of North Carolina at Chapel Hill. Instead of asking participants whether they considered themselves good listeners, Taylor West and colleagues examined recorded conversations frame by frame and turn by turn.

Study 1 drew on 300 undergraduate and graduate students, most of whom were women and whose average age was 20. Participants completed one of two activities, then returned to the laboratory approximately 48 hours later.

One group had watched a short video about the benefits of connecting with strangers and acquaintances, discussed the goal with a virtual character, and made concrete plans to seek moments of connection. The active control group followed a similarly structured programme focused on diaphragmatic breathing.

At the second visit, each participant completed a ten-minute “fast friends” task with an unfamiliar partner. Printed prompts gradually invited more self-disclosure. The partner was always one of ten trained female confederates who did not know which intervention the participant had received.

Five research assistants later divided each recording into 30-second segments. Their coding focused on two behaviours: relevant follow-up questions and verbal validation. Agreement between coders was very high, which matters when categories depend on human judgement.

A follow-up question had to pursue, clarify or better understand something the speaker had said. Verbal validation included paraphrasing, reflecting understanding, offering support or giving a friendly opinion in response to the speaker’s meaning.

A bare “yeah”, “mhmm” or “me too” did not qualify. Neither did quickly pivoting to one’s own similar experience without first acknowledging what the other person had said.

Validation did not mean automatic agreement

The word “validation” can sound as though a good listener must endorse every claim. That is not how it was used here.

A listener could validate by accurately paraphrasing a difficult experience, checking whether they had understood it, or offering a response suited to what had been disclosed. The relevant signal was comprehension and positive intention, not surrendering one’s own judgement.

This distinction also separates responsive listening from social mimicry. Saying “me too” may be warm, but it can also move the conversational centre away from the speaker. Reflecting first shows that the original account was not merely used as a launch point for somebody else’s story.

The same logic applies to follow-up questions. “What happened next?” can communicate interest if it grows from the answer. The identical sentence can feel mechanical if the listener has been working down a list.

Earlier research covered in Scandinavia Standard’s article on why follow-up questions increase liking found that people who asked more questions, particularly follow-ups, were generally better liked. The newer paper extends that idea by coding several forms of listening and comparing them with the rhythm and shared emotional quality of the exchange.

The ten-minute conversation produced the clearest pattern

During the “fast friends” task, more follow-up questions and more verbal validation were each associated with shorter average gaps before both participants and their partners responded. These were differences of fractions of a second, not a contest to speak as fast as possible.

That measure came from prior work showing that faster conversational response times often signal that two people feel they have clicked. The interpretation is not that pauses are bad. In casual conversations between strangers, a smooth turn-taking rhythm can indicate that neither person is struggling to determine whether the other is engaged.

Both coded behaviours were also positively associated with the confederate’s report of “positivity resonance”, a term for a shared experience of warmth, care and emotional synchrony. Independent observers who watched the recordings likewise saw more of those connecting behaviours in interactions where participants asked more follow-ups and validated more.

There was one revealing non-result. Participants who themselves displayed more listening behaviour did not report significantly greater positivity resonance. The benefits appeared more clearly in their partners’ reports and in external observations.

That asymmetry makes intuitive sense. Listening is directed outward. Its first social effect may be that another person feels understood, while the listener’s own sense of togetherness still depends on what comes back.

The randomized intervention adds a limited causal piece. Participants encouraged to seek connection asked 23 per cent more follow-up questions than those in the breathing-control condition. On the raw scale, the groups averaged 2.30 and 1.45 follow-up questions respectively during the ten-minute task.

An exploratory mediation model was consistent with those additional questions partly accounting for the intervention’s improvement in connection. The effect was small, however, and the intervention did not explicitly train question-asking. The model supports a possible pathway; it does not prove that inserting one more question into any conversation will make connection rise.

Small talk did not reward every question

Study 2 analysed 348 working adults aged 20 to 65. After 35 days in one of four behavioural conditions, each participant attended a final laboratory visit where researchers quietly arranged an opportunity for up to five minutes of small talk with the experimenter.

This time coders used the same counts of follow-up questions and verbal validation, plus a global listening score. The broader score considered the exchange as a whole, combining verbal responsiveness with visible cues such as engaged posture, facial expression and eye gaze.

The global score was associated with all of the connection markers examined: faster response times for both people, more observer-coded positivity resonance, and a higher connection rating from the experimenter.

Verbal validation was still associated with the shared emotional quality of the interaction. Follow-up questions alone were not. They predicted a faster response from the partner, but not the broader observer-coded or partner-reported measures of connection.

This is a useful correction to any simple conversational rule. Small-talk questions can express curiosity, but they can also fill silence, satisfy politeness or keep an exchange moving until somebody can leave. Counting a question without seeing the face, posture, timing and response that surround it loses much of its meaning.

The interventions also failed to increase any listening measure in Study 2. A five-minute exchange with an experimenter after 35 days of repeated reminders may simply have been too different from the more personal, ten-minute peer conversation conducted 48 hours after a richer intervention.

Returning to an earlier detail is unusually revealing

When the researchers examined the transcripts, they noticed another pattern among participants coded as strong listeners. These people often remembered a detail from earlier in the conversation and brought it back later to move the exchange forward.

Someone might return to a comment about a professor and ask what the speaker had been thinking at that moment. Another might ask what happened after an earlier confrontation. The follow-up did not merely attach to the final sentence. It showed that an earlier part of the person’s account had remained active in the listener’s mind.

That callback was not coded as its own preregistered variable, and the paper does not provide a separate effect size for it. It is a qualitative observation from the transcripts, not an independently tested third result.

Even with that boundary, the observation helps explain why some follow-ups carry more weight than others. A question that returns to a detail demonstrates attention across time. It is harder to produce convincingly while only waiting for one’s turn to speak.

It also shows why self-disclosure alone is not sufficient. Strangers often enjoy deeper conversation more than they expect. Yet disclosure creates only an opening. Someone still has to receive it, hold it and respond.

Connection is made between people, not by one performer

The authors describe positivity resonance as a collective state. One person may offer careful attention, but an interaction is unlikely to feel fully mutual if the other person never returns it.

This is the basis for the paper’s emphasis on reciprocity. The listener’s behaviour was linked more consistently to the partner’s experience than to the listener’s own. The researchers suggest that togetherness may be most pronounced when both people display high-quality listening to each other.

The last part should be read carefully. The studies did not randomly assign both members of a pair to listen well or poorly, then compare every possible combination. Nor were the conversations between two equally naive participants. In Study 1 the unfamiliar partner was a trained confederate; in Study 2 it was the experimenter. Both were blind to intervention condition, but both had repeated experience with the protocol.

Reciprocity is therefore a theoretically grounded interpretation of the dyadic pattern, not a separate causal experiment proving that a precise amount of mutual listening produces a precise increase in connection.

Still, it captures an important feature of conversation. Listening is not a solo display performed at somebody. Each person’s response changes what the other person can comfortably say next. Acknowledgement can invite detail, detail can invite curiosity, and curiosity can invite acknowledgement in return.

The practical lesson is a pattern, not a script

The simplest practical sequence is ordinary: notice a detail, check that it was understood, ask where it leads, and leave room for the other person to do the same. None of that requires an impressive story or a memorised list of clever questions.

It is just as important not to turn listening into performance. Too many questions can feel like an interview. A paraphrase delivered after every sentence can sound clinical. Quick replies are not universally better, particularly during conflict, grief or a subject that deserves reflection.

Nonverbal conventions also vary. Eye contact, facial expressiveness and response timing can differ across cultures, disabilities, personalities and levels of familiarity. A person’s comfort with those cues should not be mistaken for the depth of their care.

The evidence itself has clear limits. Both samples came through one US university setting and were predominantly women. The conversations were brief, recorded and involved partners who had practised their role. The analyses connected behaviour with momentary measures immediately after one exchange, not with friendships that lasted.

The researchers have made their data and analysis code available through the Open Science Framework, and they explicitly call for replication in natural settings with partners who are not trained in a laboratory protocol.

The careful conclusion is narrower than a formula for instant intimacy. People felt and appeared more connected when attention became observable in the shape of the conversation. Listening mattered not as quiet politeness, but as a response another person could recognise, build upon and, at its best, return.