Texting anxiety is real, and the research explains almost all of it

A while back someone I like replied to a long message of mine with “ok”. Nothing else. I read it maybe fifteen times over the next two hours, looking for the tone. Was it a cold ok or a busy ok. I drafted three replies and sent none of them. Then they messaged again an hour later about something completely unrelated, perfectly warm, and it turned out the ok had been typed in a car park.

I’ve spent a lot of time since then reading the actual research on this, partly out of self-interest and partly because most of what’s written about texting anxiety online is nonsense with a percentage sign attached.

Full disclosure before we go any further: I co-founded Subtext, an app that reads the emotional subtext of a message before you send it. So I have an obvious commercial interest in texting anxiety being a real thing. That’s precisely why I went to the primary sources rather than trusting the listicles, and why I’ve flagged every number I couldn’t verify.

No diagnostic manual contains the words “texting anxiety”, and that explains a lot

There’s no clinical diagnosis called texting anxiety. No validated texting anxiety scale either, which surprised me more. What exists instead is a cluster of well-studied constructs that overlap with it, and if you want honest answers you have to go and read those.

The oldest is communication apprehension, defined by James McCroskey in 1977¹ as “an individual’s level of fear or anxiety associated with either real or anticipated communication with another person or persons”. The word “anticipated” is doing a lot of work in that sentence. Most texting anxiety happens before anything has been sent.

Then there’s telephone anxiety, which has a longer research history than the text version. A 2023 national survey of 520 US adults by Leanna Kim and Sang-Hwa Oh² found that heavier use of digital communication technologies correlated with higher telephone anxiety, and that the effect was stronger among people speaking a non-native language.

The closest thing to a purpose-built instrument is the Digital Stress Scale, developed by Jeffrey Hall and colleagues in 2021³ with 735 adolescents and young adults. It measures five separate things: availability stress, approval anxiety, fear of missing out, connection overload, and online vigilance. Internal consistency across the subscales ran from .85 to .93, and a shortened ten-item version was later validated across German, Italian and Japanese samples⁴. If you recognise yourself in this article, availability stress and approval anxiety are probably the two doing the damage.

I find the absence of a proper texting anxiety scale genuinely odd, given how many people describe the experience. It’s a gap in the research, not a verdict on whether the thing is real.

The empty reply box is harder than a phone call, and researchers know why

The intuitive explanation for texting anxiety is that it’s social anxiety with a screen in front of it. The research says something more interesting.

Text is asynchronous. You can take as long as you want. That’s supposed to be the relief, and for a lot of people it’s the opposite, because unlimited time to compose means unlimited time to second-guess. A 2023 study in Media Psychology⁵ found that people higher in social anxiety preferred texting to face-to-face conversation, and that the preference got stronger as the conversation got more threatening, moving from small talk to difficult conversations to breakups. What mediated the preference was the perceived importance of editability, meaning the ability to redraft before anyone sees it.

A 2025 study in Communication Reports⁶ found the same mechanism under a different name. Socially anxious people scored higher on what the authors called perceived controllability in texting, and that controllability predicted both a preference for texting and higher anxiety about phone calls. Younger adults in their twenties and thirties reported stronger texting preference and more call anxiety than adults in their forties and fifties.

None of this is new. Reid and Reid found it back in 2007 with a sample of 158⁷: lonely people preferred voice calls and rated texting as less intimate, while socially anxious people preferred texting and rated it as a better medium for expressive contact. The medium that removes the anxiety of speaking hands you a different anxiety about writing, and the people who need one relief tend to inherit the other problem.

Your brain fills in the missing tone, and it guesses badly

This is the mechanism that made the most sense to me personally.

Text strips out vocal tone, facial expression, timing, and posture. Your social cognition still needs that information, so it manufactures it. What it manufactures depends on what you were already worried about.

Kingsbury and Coplan tested this directly⁸ in a study titled, wonderfully, “RU mad @ me?”. They built a measure of interpretation bias for ambiguous text messages with 215 undergraduates and found that negative readings of ambiguous messages tracked with social anxiety symptoms and with established face-to-face interpretation-bias measures. A second study with 353 participants found that ambiguous messages from female senders were interpreted more negatively, an effect especially pronounced among male readers.

So the “ok” that ruined my afternoon contained no information about mood at all. I supplied that part myself.

There’s some neuroscience worth knowing here too, though I’d hold it loosely. A pre-registered EEG hyperscanning study published in Scientific Reports in 2024⁹ measured brain-to-brain synchrony in 65 mother and adolescent pairs, comparing face-to-face conversation with texting from separate rooms. Texting did produce genuine neural synchrony, which was the first demonstration of its kind. Face-to-face produced more, with eight fronto-temporal links significantly stronger. I’ve seen this study cited online as proof that texting doesn’t connect people. That’s not what it found. It found that texting connects people somewhat less.

Waiting for a reply is a completely different problem from writing one

Most people fold these together, and they have separate research literatures.

The best-studied version of waiting stress comes from workplace psychology. Larissa Barber and Alecia Santuzzi named and validated it in 2015 as telepressure¹⁰, “the combination of a strong urge to be responsive to people at work through message-based ICTs with a preoccupation with quick response times”. People scoring high on it reported more burnout, worse sleep quality, and more health-related absences, and the effect held after controlling for job involvement and general workload. Follow-up work with samples of 254 and 409 workers¹¹ showed the damage runs through failed psychological detachment, which is a technical way of saying you never fully leave the conversation.

The same tension exists in friendships. Jeffrey Hall and Nancy Baym’s 2012 study of mobile maintenance expectations¹² found a genuine paradox: expecting constant contact increases dependence, which raises closeness and satisfaction, while simultaneously increasing overdependence and entrapment, which lower it. The thing that makes you close is the thing that makes you feel trapped.

Interface design pours petrol on this. A CHI 2018 analysis of stress on mobile messengers¹³ found read receipts, typing indicators and contact syncing running through nearly every stress theme users reported. A separate 2018 study measured how long it takes for sadness, anxiety, anger and guilt to appear¹⁴ while waiting for a reply, split by whether the message showed as read or unread. The read receipt removes your plausible deniability and theirs at the same time.

And the checking itself carries a cost. APA’s 2017 Stress in America survey¹⁵ found 43% of Americans qualified as constant checkers of email, texts and social media, and they reported higher average stress than everyone else, 5.3 out of 10 against 4.4. Among constant checkers who also checked work email on days off, it rose to 6.0. In the same survey, 18% named technology use as a significant source of stress in their lives.

Karla Murdock’s 2013 study of 83 first-year university students¹⁶, who averaged around 114 texts a day, found more texting associated with more sleep problems regardless of baseline stress, and interpersonal stress predicting burnout only among the heavier texters. It’s a small correlational study, so I’d treat it as suggestive rather than settled.

Some people get hit much harder than others

If you’re autistic, this literature will read less like news and more like transcription. Philippa Howard and Felicity Sedgewick surveyed 245 autistic adults¹⁷ for a paper titled “Anything but the phone!”, and found written communication preferred across most contexts, valued specifically for thinking time, predictability and lower sensory load. Participants described phone calls causing “debilitating anxiety”, and one described other people’s insistence on phone contact as “disabling to the point of outright discrimination”. The sample’s mean GAD-7 score was 11.56, which sits in the moderate anxiety range, so this is a group already carrying a lot.

Adults with ADHD describe a different shape of the same problem, usually through rejection sensitivity. One qualitative study captured a participant awake until four in the morning¹⁸, checking their phone and asking why the other person hadn’t replied. Rejection sensitive dysphoria isn’t a formal diagnosis and the evidence base is mostly small and qualitative, so I’d call it well-observed rather than well-measured.

Attachment style shows up too, though the evidence is thinner than the confident blog posts suggest. Robert Weisskirch found anxious attachment associated with more texts¹⁹ sent to and received from partners, but with only 31 participants. Shanhong Luo’s larger study of 395 people²⁰ found that the share of your communication conducted by text correlated negatively with relationship satisfaction, while raw text volume barely mattered. The proportion seems to matter more than the count.

One useful complication comes from a three-wave longitudinal study of 523 adolescents by Van Zalk²¹. Excessive chatting predicted compulsive internet use mainly among teenagers low in social anxiety. For socially anxious teenagers, chatting looked compensatory rather than compulsive. The same behaviour means different things in different people, which is one reason blanket screen-time advice keeps failing.

Most texting anxiety statistics you’ll see are invented

This section is why the article took three times longer than planned.

Search for texting anxiety statistics and you’ll find confident numbers. 31% of people experience it daily. 47% of Gen Z say it’s keeping them single. The average person spends 40 minutes drafting an important text. A 73% draft abandonment rate. I tried to trace all of them. Nearly every one leads back to an app company’s own blog post or a syndicated press release with no sample size, no fieldwork dates, and no disclosed methodology.

The 31% figure at least has a real ancestor. It comes from a 2018 Viber survey run by Pollfish with 2,400 adults across the UK and US, and the original wording was that 31% found texting a daily source of stress. Somewhere in the retelling, stress became anxiety. I couldn’t find a live copy of the original report, so I’d treat even that one as directional.

There’s also a 42% telephobia figure attached to Gen Z in a lot of places. The nearest real study is a 2024 survey of 300 medical students in India²², where 42% showed some degree of phone ring phobia, broken down as 33.0% mild, 7.67% moderate and 1.33% severe. Medical students at one Indian college is not Gen Z globally, and “mild” is carrying most of that 42%.

For a sense of how unstable prevalence figures get around here, look at nomophobia, the fear of being without your phone, which has been studied far more than texting anxiety. A systematic review in PLOS ONE²³ found estimates of people suffering from it ranging from 6% to 73% depending on where you draw the cutoff. A later meta-analysis of 52 studies covering 47,399 people across 20 countries²⁴ landed at roughly 20% mild, 50% moderate and 20% severe. One construct, two very different pictures, depending on the threshold.

If you want defensible numbers on how people feel about messaging versus calling, the honest ones come from named pollsters with published methodology. Pew Research’s 2015 study of teens and friendships found 58% of teens with smartphone access preferred texting their closest friend, and a 2024 Uswitch survey of 2,000 UK adults found 70% of 18 to 34 year olds preferring text and social media over calls, with 56% assuming an unexpected call means bad news. Neither measures anxiety. I’d still rather quote a real number about preference than an invented one about distress.

The scales researchers reach for when they can’t measure texting anxiety directly

Worth knowing these exist, if only so you can spot when someone quotes one incorrectly.

The Nomophobia Questionnaire from Caglar Yildirim and Ana-Paula Correia²⁵ runs to 20 items scoring 20 to 140, with standard cutoffs at 21 to 59 mild, 60 to 99 moderate, and 100 upwards severe. The Self-Perception of Text-Message Dependency Scale from Tasuku Igarashi and colleagues²⁶ is 15 items covering emotional reaction, excessive use and relationship maintenance. Bianchi and Phillips’s Mobile Phone Problem Use Scale²⁷ is 27 items with an original alpha of .91. And the Fear of Missing Out scale from Andrew Przybylski and colleagues²⁸, validated on a nationally representative sample of 2,079 UK adults, keeps turning up as a correlate of almost everything in this space.

Notice what’s missing from that list. Every one of them measures dependence, phobia or missing out. None measures the specific experience of staring at a half-written reply.

What actually helps, according to people who tested it

The strongest single piece of evidence here is unglamorous. Kostadin Kushlev and Elizabeth Dunn ran a two-week within-subjects experiment with 124 adults²⁹. In one week people checked email as often as they liked, which at baseline was about 15.5 times a day. In the other they were asked to check three times daily with notifications off. The limited week produced significantly lower daily stress, with an effect size of .45. Kushlev’s own summary is the line I keep coming back to:

People find it difficult to resist the temptation of checking email, and yet resisting this temptation reduces their stress.

There’s an honest caveat inside the paper, which is that people in the limited condition actually managed about 4.7 checks a day rather than three. Trying and partly failing was still enough.

Second, there’s now a proper randomised trial of text-message-delivered CBT. Mason and colleagues randomised 102 young adults aged 18 to 25³⁰ with GAD-7 scores of 10 or above, recruited across 33 US states, to eight weeks of automated CBT by text or a waitlist. Around a quarter of the treatment group reached minimal-symptom functioning against 5.5% of controls, and the effect ran through increased behavioural activation and reduced perseverative thinking. It targets generalised anxiety rather than texting anxiety specifically. The mechanism it moves, perseverative thinking, is exactly the one that has you rewriting a message nine times.

Third, and this matters if you manage people: policy alone doesn’t fix response pressure. Barber, Santuzzi and Hu surveyed two worker samples³¹ and found that having a formal disconnection policy showed no association with telepressure at all. What predicted it was implicit norms about after-hours availability. Eurofound reached the same conclusion across Belgium, France, Italy and Spain³², finding that a right-to-disconnect policy on its own “is insufficient to bring about cultural change in the workplace”. Your team watches what you do at 11pm, not what the handbook says.

Beyond that, the practical moves follow fairly directly from the mechanisms above. Turn off read receipts, since they mostly manufacture pressure in both directions. Batch your checking instead of living inside the thread. When a message reads as cold, write down one alternative explanation before you reply, which is standard cognitive restructuring aimed at the exact bias Kingsbury and Coplan measured. And agree explicit norms with the people who matter about what a slow reply does and doesn’t mean, because the Hall and Baym finding suggests the expectation causes more harm than the delay ever does.

That last cluster is roughly the problem we built Subtext around, so weigh the recommendation accordingly.

The advice I think is wrong

Two things get repeated constantly that I don’t accept.

The first is that texting anxiety is a symptom of phone addiction and the answer is a digital detox. The Van Zalk finding cuts against that neatly: for socially anxious teenagers, messaging was compensatory rather than compulsive. Removing the medium that lets an anxious person communicate at all, then calling it treatment, looks like a bad trade to me.

The second is the instruction to “just call instead”. For a large group of people, including many autistic adults, calling is the harder thing rather than the easier one, and Howard and Sedgewick’s participants say so in fairly blunt language. For most people the difficulty is a fear of being read badly, and that follows you from one channel to the next.

I’m also suspicious of how fast this whole area got commercialised, including by companies like mine. The research base is thinner than the marketing implies. There’s no validated texting anxiety scale and no randomised trial of anything targeting texting anxiety specifically. Anyone quoting you a precise percentage is quoting a press release.

When it’s bigger than texting

One thing I want to leave you with. A meta-analysis of 41 studies covering 41,871 children and young people³³ found problematic smartphone use at a median prevalence of 23.3%, associated with substantially raised odds of anxiety and depression. Those are correlations, and they don’t tell you which came first. If you’re finding that message dread is affecting your sleep or whether you leave the house, that’s worth talking to a GP or a therapist about rather than solving with a settings change. Cognitive behavioural therapy has a strong evidence base for anxiety generally, and the specific thoughts driving texting dread respond well to exactly that kind of work.

Being bad at texting is an extremely ordinary thing to be. Almost everyone in the studies above is, one way or another.

---

If you think I’ve read any of this wrong, or you know of research I’ve missed, tell me. I’d rather be corrected than quoted.

Samet Durgun is the co-founder of Subtext, an app that catches the emotional tone of your messages and rewrites them in your own voice. He’s based in Berlin.

---

Sources

Every link above goes to the primary source where one exists. Where I could only find secondary reporting, or couldn’t verify a figure at all, I’ve said so in the text rather than dressing it up.

1. McCroskey, J. C. (1977). Oral communication apprehension: A summary of recent theory and research. Human Communication Research, 4(1), 78-96. jamescmccroskey.com
2. Kim, L. T., & Oh, S.-H. (2023). Predicting telephone anxiety. Communication Research Reports, 40(3). tandfonline.com
3. Hall, J. A., et al. (2021). Development and initial evaluation of a multidimensional digital stress scale. researchgate.net
4. Funke, C., et al. (2025). The digital stress scale: cross-cultural application, validation, and development of a short scale. Review of Managerial Science. link.springer.com
5. To Text or Talk in Person? Social Anxiety, Media Affordances, and Preferences for Texting Over Face-to-Face Communication in Dating Relationships (2023). Media Psychology, 27(3). doi.org
6. Avoidance and Anxiety About Phone Calls in Young Adults: The Role of Social Anxiety and Texting Controllability (2025). Communication Reports, 39(2). tandfonline.com
7. Reid, D. J., & Reid, F. J. M. (2007). Text or Talk? Social Anxiety, Loneliness, and Divergent Preferences for Cell Phone Use. CyberPsychology & Behavior, 10(3), 424-435. pubmed.ncbi.nlm.nih.gov
8. Kingsbury, M., & Coplan, R. J. (2016). RU mad @ me? Social anxiety and interpretation of ambiguous text messages. Computers in Human Behavior, 54, 368-379. sciencedirect.com
9. Generation WhatsApp: inter-brain synchrony during face-to-face and texting communication (2024). Scientific Reports. nature.com
10. Barber, L. K., & Santuzzi, A. M. (2015). Please respond ASAP: workplace telepressure and employee recovery. Journal of Occupational Health Psychology, 20(2), 172-189. pubmed.ncbi.nlm.nih.gov
11. Barber, L. K., Conlin, A. L., & Santuzzi, A. M. (2019). Workplace telepressure and work-life balance outcomes. Stress and Health, 35(3), 350-362. pubmed.ncbi.nlm.nih.gov
12. Hall, J. A., & Baym, N. K. (2012). Calling and texting (too much): Mobile maintenance expectations, (over)dependence, entrapment, and friendship satisfaction. New Media & Society, 14(2), 316-331. journals.sagepub.com
13. Shin, et al. (2018). Understanding Stress on Mobile Instant Messengers based on Proxemics. CHI 2018. interruptions.net
14. Time Before Negative Emotions Occur While Waiting for a Reply (2018). ACHI 2018. personales.upv.es
15. American Psychological Association (2017). Stress in America: Coping with Change. apa.org
16. Murdock, K. K. (2013). Texting While Stressed: Implications for Students’ Burnout, Sleep, and Well-Being. Psychology of Popular Media Culture, 2(4), 207-221. murdocklab.academic.wlu.edu
17. Howard, P. L., & Sedgewick, F. (2021). ‘Anything but the phone!’: Communication mode preferences in the autism community. Autism, 25(8), 2265-2278. journals.sagepub.com
18. The lived experience of rejection sensitivity in ADHD: a qualitative exploration. ncbi.nlm.nih.gov
19. Weisskirch, R. S. (2012). Women’s Adult Romantic Attachment Style and Communication by Cell Phone with Romantic Partners. Psychological Reports, 111(1), 281-288. pubmed.ncbi.nlm.nih.gov
20. Luo, S. (2014). Effects of texting on satisfaction in romantic relationships: The role of attachment. Computers in Human Behavior, 33, 145-152. sciencedirect.com
21. Van Zalk, N. (2016). Social anxiety moderates the links between excessive chatting and compulsive Internet use. Cyberpsychology, 10(3). cyberpsychology.eu
22. Bairwa, Y. P., Udayaraj, A., & Manna, S. (2024). The fear of smartphone notifications and calls among medical students. Journal of Family Medicine and Primary Care, 13(5), 1850-1855. pmc.ncbi.nlm.nih.gov
23. León-Mejía, A. C., et al. (2021). A systematic review on nomophobia prevalence. PLOS ONE, 16(5). journals.plos.org
24. Jahrami, H., et al. (2023). The Prevalence of Mild, Moderate, and Severe Nomophobia Symptoms. Behavioral Sciences, 13(1), 35. ncbi.nlm.nih.gov
25. Yildirim, C., & Correia, A.-P. (2015). Exploring the dimensions of nomophobia. Computers in Human Behavior, 49, 130-137. doi.org
26. Igarashi, T., Motoyoshi, T., Takai, J., & Yoshida, T. (2008). No mobile, no life: Self-perception and text-message dependency among Japanese high school students. Computers in Human Behavior, 24(5), 2311-2324. tasukuigarashi.com
27. Bianchi, A., & Phillips, J. G. (2005). Psychological predictors of problem mobile phone use. CyberPsychology & Behavior, 8(1), 39-51. pubmed.ncbi.nlm.nih.gov
28. Przybylski, A. K., Murayama, K., DeHaan, C. R., & Gladwell, V. (2013). Motivational, emotional, and behavioral correlates of fear of missing out. Computers in Human Behavior, 29(4), 1841-1848. selfdeterminationtheory.org
29. Kushlev, K., & Dunn, E. W. (2015). Checking email less frequently reduces stress. Computers in Human Behavior, 43, 220-228. interruptions.net
30. Mason, M. J., et al. (2025). Treating young adult generalized anxiety disorder with text-message delivered cognitive behavioral therapy. sciencedirect.com
31. Barber, L. K., Santuzzi, A. M., & Hu, X. (2023). Tackling the Problem of Workplace Telepressure: Are Disconnection Policies Helpful? Group & Organization Management. journals.sagepub.com
32. Eurofound (2023). Right to disconnect: Implementation and impact at company level. eurofound.europa.eu
33. Sohn, S. Y., et al. (2019). Prevalence of problematic smartphone usage and associated mental health outcomes amongst children and young people. BMC Psychiatry. ncbi.nlm.nih.gov

Three sources I deliberately left out. A Cambridge journal article titled “Text Anxiety in Adolescents” is about test anxiety, meaning exams, and the word “Text” in the title is a typo that has since propagated into a lot of AI-written articles on this subject. A 2025 SciTePress paper on “text syndrome” contains apparently fabricated references, and “text syndrome” appears nowhere in the peer-reviewed literature. The widely circulated Gen Z statistics about texting anxiety and dating trace back to an app company’s own press release with no methodology attached.