Social media and communication
How everyday communication changed
Traditional calls and texts are in measured decline while app messaging and calling carry several times the volume, but the effects on comprehension are barely measured.
The channel shift is documented precisely, because operators and platforms report the traffic. UK voice minutes on fixed and mobile networks are declining slowly, text messages are declining fast, and messaging apps carry several times the volume of the networks they displaced. What that shift does to people is documented far less well. The experimental literature amounts to a handful of studies, mostly small, showing that written messages are harder to read correctly than their senders assume, that voice builds connection more than people expect, and that in-person groups generate more ideas than video calls without choosing between them any better. None of that supports the confident claims in circulation about misunderstanding rates or generational phone anxiety. For a workplace-oriented example of technology measurement in practice, see chronemics.
Traditional voice is shrinking and SMS is collapsing
Ofcom collects traffic returns from UK operators, so these are counted minutes rather than recalled ones.
Total UK fixed and mobile voice call volumes fell 6% to 176 billion minutes in 2024, with landline calls down 23% and mobile calls down 4%.
Direction: Decrease. Strength of evidence: Strong.
Caveat The figures exclude voice minutes carried inside apps, so they measure the decline of one channel rather than a decline in talking.
Text messaging is going faster, and part of the drop is an accounting artefact rather than behaviour.
UK mobile messages fell 24.3% year on year to 5.24 billion in Q3 2025, while mobile voice minutes fell only 3.9% over the same period.
Direction: Decrease. Strength of evidence: Strong.
Caveat Part of the SMS decline reflects reclassification of business messaging and RCS silently replacing SMS on Android handsets.
The traffic moved rather than disappeared
Treating the network figures as evidence that people communicate less is a mistake the same report corrects. Platform-reported volumes run several times higher than the regulated networks.
Online communication services carried 820 billion outgoing messages, up 6%, and 415 billion voice and video call minutes, up 4% — roughly 2.4 times total traditional voice minutes.
Direction: Increase. Strength of evidence: Strong.
Caveat Supply of this data is voluntary and coverage across services is incomplete; message counts exclude reactions and likes.
Total time online is also measured rather than assumed. Ofcom's Online Nation research put UK adults at 4 hours 30 minutes a day online in 2025, with children at about two hours a day at ages 8–9 rising to about four hours by 13–14 (Online Nation, 2025). Almost all television, music and reading now sits inside that total, so it is not a proxy for time spent talking to people, and not a proxy for social media use either.
Written messages are less legible than senders believe
The most-cited experiment on text-based misunderstanding is small and dated, and its actual finding is narrower than its reputation.
Senders expected recipients to identify sarcasm in their emails 75% of the time, but accuracy was about 50% — chance level — while the same messages delivered by voice hit the predicted 75%.
Direction: Decrease. Strength of evidence: Mixed.
Caveat A lab task with small undergraduate samples in which participants were explicitly barred from using emoticons, a constraint that no longer reflects how people write.
Voice does more than people expect, and people still avoid it
A later set of studies asked people to reconnect with an old friend and randomly assigned the medium. The pattern was a forecasting error: participants predicted awkwardness that did not materialise.
Reconnecting by phone produced a stronger bond than by email, 5.27 against 4.62 with d = 0.40 and no more actual awkwardness, yet 67% of participants said they would prefer email.
Direction: Increase. Strength of evidence: Mixed.
Caveat Effect sizes are modest, follow-up completion in the field study was 51.5%, and the design confounds voice with synchrony so it cannot isolate which does the work.
Going remote changed who talks to whom
The largest natural experiment on workplace communication came from a single firm sending tens of thousands of people home at once. The finding was structural rather than about volume: networks tightened around existing contacts and shifted out of scheduled time.
Collaboration networks became more siloed, with time spent on bridging ties down 0.41 of a February baseline unit, while instant messages rose 0.50, emails 0.08 and scheduled meeting hours fell 0.16.
Direction: Decrease. Strength of evidence: Mixed.
Caveat One technology firm over a three-month window overlapping the onset of the pandemic, with no productivity outcome measured and several authors employed by Microsoft with pending related patents.
None of that measured output. A randomised trial at Trip.com found that two days a week at home cut attrition from 7.2% to 4.8% with no detectable difference in performance grades or promotion (Bloom, Han & Liang, Nature, 2024), which is covered further in the material on what is actually known about remote and hybrid arrangements and on work that never switches off.
Video calls cost ideas but not judgement
The one clean medium comparison on creative work found a deficit in generating options and no deficit in picking among them.
In-person pairs generated roughly 15–20% more creative ideas than videoconferencing pairs, though videoconferencing groups were no worse at selecting which idea to pursue.
Direction: Decrease. Strength of evidence: Mixed.
Caveat Both studies used short, artificial brainstorming tasks between newly paired strangers, so the finding concerns idea generation in first meetings rather than remote work in general.
The short version
- UK voice minutes across fixed and mobile networks fell 6% to 176 billion in 2024, a slow decline rather than a collapse (Ofcom, 2025).
- Text messaging is falling roughly six times faster than voice, down 24.3% year on year in Q3 2025, partly for accounting reasons (Ofcom, 2026).
- App traffic more than compensates, at 820 billion messages and 415 billion call minutes in a year (Ofcom, 2025).
- Senders overestimate how clearly their written tone reads, with sarcasm recognised at chance level in the only well-known test of it (Kruger et al., 2005).
- The effect of any of this on everyday understanding is unmeasured: the relevant experiments are lab tasks with strangers, and none tracks real conversations over time.