Why does small talk feel harder than it used to?

From 10 May through 31 December 2020, people aged 15 and older in the United States averaged 0.54 hours a day in the Bureau of Labor Statistics category “socializing and communicating,” down from 0.66 hours in the same months of 2019. That is about seven minutes off a habit of roughly forty, and the share of people who did any face-to-face socializing on a given day fell from 35.2 percent to 27.1 percent [1].

Small talk is where a thinner habit shows up later: a hello, one question, a few extra minutes beside a group. The layers below ask what the ear is timing, what the studies actually found, and what a small repetition can practise.

A reply starts before the other sentence ends

Tanya Stivers, N. J. Enfield, Stephen Levinson and colleagues timed the gap between a question and the next turn in ten languages, including English, Japanese and Danish [2]. The most common gaps fell between 0 and 200 milliseconds, and the average was 208 milliseconds. Japanese turns in the sample averaged about 7 milliseconds and Danish turns about 469, so one language’s ordinary beat can sound hesitant in another. Answers, and nods made with the voice, came faster than voice alone [2].

Listeners treat the speed as connection. Emma Templeton, Luke Chang, Thalia Wheatley and colleagues had friends and strangers talk, then rate how connected they felt [3]. Faster replies lined up with stronger connection, driven more by the partner’s speed than by one’s own, and slowed recordings sounded less connected to new listeners [3]. Under about 250 milliseconds a reply is too quick to choose, so it has to be underway before the other person finishes [3]. A one-second search is several times that hole.

Calls made the hole easy to misread. “Telephone calls, mail, and e-mail” rose from 0.15 hours a day to 0.22 as face-to-face socializing fell [1]. Katrin Schoenenberg, Alexander Raake and Judith Koeppe added delay to phone calls [4]. Up to 1.2 seconds, technical-quality ratings held. At high delays, partners were rated less attentive, less extraverted and less conscientious [4]. A line’s pause was filed as a personal trait, the same story a slow sentence invites about oneself. Stivers timed questions, Templeton rated felt connection, and Schoenenberg added the delay on purpose [2][3][4].

The lost minutes did not land evenly

The national drop hides the ages [1]. From 15 to 19, the average fell from 0.79 hours a day to 0.60, about eleven minutes. From 20 to 24 it rose from 0.56 to 0.66, about six minutes, and from 25 to 34 it rose from 0.63 to 0.66. People who kept meeting may have stayed longer, while seminars, new shifts and crowded birthdays stayed scarce. Collection stopped mid-March to mid-May 2020, so the window opens on 10 May [1].

The feeling, asked later, did not simply follow those minutes. YouGov, for the Prince’s Trust NatWest Youth Index, questioned 2,106 people aged 16 to 25 in the UK from 22 December 2021 to 12 January 2022 [5]. Forty percent said they felt anxious about socialising because of the pandemic, 36 percent said making friends had got harder, and 45 percent said their social life had got worse [5]. That is a weighted online report of a feeling, not a skills exam. Minutes and confidence can part company.

Repeated approaches moved the forecast

If ease were a fixed trait, a few days of it would leave self-ratings where they were. Gillian Sandstrom, Erica Boothby and Gus Cooney watched a narrower shift [6]. Students on a US campus and a UK campus did weekday errands, such as finding someone wearing interesting shoes, or drinking a coffee. One group talked. The other only watched. Of 454 who began, 286 remained (at least four days, an honesty check, average age about 20, mostly course credit): a university points-game sample, not a portrait of 2020 [6].

Talkers expected 0.89 rejections before someone would speak, then 0.23 by Friday and 0.32 a week later. Watchers, asked at the end, expected 0.99 and then 1.11 [6]. A 1-to-7 score covered starting, continuing and ending, flipped so higher meant more able. Talkers moved from 4.34 to 4.77 by Friday and 4.81 a week later. Watchers stayed near 4.2 to 4.3 (Friday gap d = 0.43). Expected awkwardness among talkers fell from 3.78 to 3.34, a small edge over watchers on the usual cutoff (d = 0.35, p = 0.05). Expected enjoyment rose from the talkers’ own Monday and did not clearly beat the watchers. Watching was not the same practice as speaking first [6].

Three repetitions for a cold skill

Three repetitions put the first line back in reach.

Repetition	Smallest form	What it is practising
Hello first	Greet someone before a topic is ready	Loading the opening until it can run on its own
One question	Ask one thing, then listen to the answer	Handing the next beat to the other person
Five more minutes	Stay past the first invented exit, once	The middle, where one specific detail can appear

Hello first is not a drill for a 200-millisecond gap. That speed sits below deliberate control [3]. One greeting is the outing. One question follows the result that connection tracked the partner’s reply [3]: ask whether the rain will keep up, then hear the answer.

Five more minutes has a limit beside it. Mastroianni, Gilbert, Cooney and Wilson studied 932 conversations [7]. In the lab, both people got the ending they both wanted 1.59 percent of the time, and the chat ended when even one of them wanted to 29.37 percent of the time. Both had wanted to stop sooner in 46.83 percent of pairs and to continue in 9.52 percent. Across the studies the miss was about half the length. Recalled everyday chats averaged 13.97 minutes (median 10): people wished for 1.91 minutes more on average, while the absolute miss was 6.60 [7]. Staying five minutes is for someone about to flee. Leaving when you are done fits the same evidence [6][7].

A week of ordinary approaches moved the score people were giving themselves: practised timing, with the next line slow to load [6].

On the walk back, wind pushes at a coat and rain ticks against a hood. A voice on a phone nearby rises and falls through a story meant for someone else, and for a moment the only task is listening.

Where these numbers stop

A US time diary missing spring 2020, an online UK survey of feelings, question timing in ten languages, campus and lab chats, delayed phone calls, and one week of student approaches: each describes a group, not one evening [1][2][3][4][5][6][7]. A pause can be thoughtfulness, a language’s tempo, or someone who wants to leave. If joining people has felt heavy for a long time, say so to someone you trust. This page describes research. It is not personal advice, and it does not claim that a sound or a product changes the difficulty.

Sources

1. U.S. Bureau of Labor Statistics (22 July 2021). American Time Use Survey, May–December 2019 and 2020, Tables 1 and 9. https://www.bls.gov/news.release/archives/atus_07222021.htm
2. Stivers, T., Enfield, N. J., et al. (2009). Universals and cultural variation in turn-taking in conversation. PNAS 106(26), 10587–10592. https://doi.org/10.1073/pnas.0903616106
3. Templeton, E. M., Chang, L. J., Reynolds, E. A., Cone LeBeaumont, M. D., and Wheatley, T. (2022). Fast response times signal social connection in conversation. PNAS 119(4), e2116915119. https://doi.org/10.1073/pnas.2116915119
4. Schoenenberg, K., Raake, A., and Koeppe, J. (2014). Why are you so slow? Misattribution of transmission delay to attributes of the conversation partner at the far-end. International Journal of Human-Computer Studies 72(5), 477–487. https://doi.org/10.1016/j.ijhcs.2014.02.004
5. The Prince’s Trust and YouGov (2022). The Prince’s Trust NatWest Youth Index 2022. Online survey of 2,106 people aged 16–25 in the UK, 22 December 2021–12 January 2022. https://assets.ctfassets.net/qq0roodynp09/IDAYdbIqAgNf8OAThpO9z/12dbc21ae9e2195595978a35db6ca220/Prince-s_Trust_Youth_Index_2022_DSN5322_FINAL.pdf
6. Sandstrom, G. M., Boothby, E. J., and Cooney, G. (2022). Talking to strangers: A week-long intervention reduces psychological barriers to social connection. Journal of Experimental Social Psychology 102, 104356. https://doi.org/10.1016/j.jesp.2022.104356
7. Mastroianni, A. M., Gilbert, D. T., Cooney, G., and Wilson, T. D. (2021). Do conversations end when people want them to? PNAS 118(10), e2011809118. https://doi.org/10.1073/pnas.2011809118