What does it mean when someone reads your message and doesn't reply?

Seen at 14:12. No reply. It is now 19:40 and you have checked six times. I co-founded Subtext, and being left on read is the single situation people describe to me most, usually with an apology attached for caring about it. If you are the one who read and has not answered, the question from the other chair is why can’t I reply to messages I actually want to answer.

You do not need the apology. The distress is not irrational and the research on that is solid. What is much weaker than the internet suggests is the second half, the part where anyone tells you what the silence means. Nobody has cleanly measured that, and the comforting statistic you have probably seen comes from a study of teenagers responding to automated clinic reminders.

Being ignored hurts, and this part is well established

Kipling Williams’s model holds that being ignored threatens four basic needs: belonging, self-esteem, control, and the sense that your existence registers with other people.

The workhorse experiment is Cyberball, a virtual ball-tossing game where the other players stop throwing to you. A meta-analysis of 120 Cyberball studies covering 11,869 participants found the average effect of being ostracised is large, exceeding d = 1.4, and generalises across the number of players, duration, gender, age, country and outcome measure¹.

What makes it striking is what fails to defuse it. The effect survives being excluded by a computer, by a group you despise, and even when participants are told the exclusion was a technical glitch. Knowing it does not matter does not stop it landing. Which is my read on why a tool like Subtext helps less here than people hope. The distress is not a reasoning error you can argue yourself out of.

There are two limits. Cyberball demonstrates an immediate reflexive drop in mood and need satisfaction under controlled conditions. It does not demonstrate long-term harm, and the staged theory built around it fared less well in that same meta-analysis than the core effect did.

The brain claim needs correcting. You have very likely read that social rejection uses the same neural machinery as physical pain. That comes from a 2003 fMRI study with 13 undergraduates, and it is contested². One line of criticism holds that the regions involved respond to salience generally, so inferring pain from their activation is a reverse-inference error. A later study scanned 60 participants across painful heat and photographs of ex-partners and found pattern classifiers could distinguish physical pain and social rejection at 92 and 80 per cent accuracy, with each performing at chance on the other³. I’d report the overlap. I wouldn’t report the equivalence.

Text specifically

Smith and Williams ran the landmark experiment in 2004⁴. They included forty students in a three-way text conversation, then stopped sending them messages. They reported worse mood and lower belonging, control, self-esteem and meaningful existence, and showed what the authors called virtual bravado, sending more messages to try to re-engage.

The design detail that matters is that, unlike Cyberball, the target could not watch the others being included. Merely being ignored was enough.

Forty undergraduates and 2004-era phones, so treat it as a real causal finding on a small scale.

Newer work tells you how often this happens in ordinary life. Two experience-sampling studies, one with 323 people logging 1,107 ostracism experiences over 14 days and one with 272 people producing 7,943 assessments over 7 days, found ostracism occurs as often as two to three times a week on average⁵. The strongest behavioural response afterwards was avoidance, ahead of approach and ahead of anything antisocial.

And what is missing. No study has experimentally isolated the read-receipt condition, where you know the message was seen, against a matched condition without receipts, and measured need threat with anything like Cyberball’s rigour. What exists on read receipts is surveys and interviews. Given how much commentary this topic generates, that absence is remarkable.

What the read receipt changes

Without receipts, silence is structurally ambiguous. You cannot tell whether the message failed to deliver, is sitting unread in a backlog, or was seen and set aside. The receipt removes the technical ambiguity and converts missing information into confirmed exposure, which moves your processing from a question about the network to a question about the relationship.

That framing is well accepted in the field. It is also a framework, not a tested causal claim, and I’d flag that before you build a theory on it.

What has been measured is thinner than the volume of commentary suggests. Two surveys of 402 senders and 316 recipients found read-but-unanswered messages generate anxiety and negativity in senders, part of why texting gives you anxiety, and that recipients strategically manage the seen signal⁶. A separate survey of 207 people found the seen function pushes users to answer faster than they otherwise would, tied to need to belong and fear of ostracism⁷.

On turning receipts off, there is no empirical study at all. Only advice articles. The reasoning is plausible in both directions, that disabling them restores plausible deniability while introducing a different low-level uncertainty, and neither half has been tested. It is an open question.

One indirect finding does support the recipient side. Some people avoid opening messages at all, precisely to escape the pressure the receipt would create⁸.

The reasons people actually read and don’t reply

The most useful study I’ve found in this whole area is a diary-and-interview study of instant-messaging users in Taiwan, which catalogued why read-receipted messages go unanswered⁸. Eight categories:

• Reading a message to assess urgency during demanding work, then deferring
• The composition cost, where formulating a socially safe reply takes real effort and waits for bandwidth
• Inconvenient context, meaning driving, commuting, in public
• Low informational priority, set aside and then accidentally never answered
• Emotional fatigue, deferring interaction to protect wellbeing
• Memory decay
• Deliberate boundary management
• Technical friction

Category six deserves a paragraph of its own. Opening the message clears the notification badge. The read receipt is generated by the exact act that destroys the reminder. The seen signal and the forgetting have a common cause, which means the thing that makes you feel deliberately ignored is often the mechanism by which you were forgotten.

That typology tells you the space of reasons. It does not tell you their frequencies, and no percentages attach to it. The reasons people go quiet for good are a different list, which I went through in why people ghost.

This is roughly where a tone-reading tool stops being able to help, and I would rather say so. Subtext can tell you what your message conveys and what readings it supports. It cannot tell you why someone else has not answered, because the answer is usually not in the message at all.

The statistic you have seen, and why it does not mean what it says

There is one number in circulation suggesting most unanswered texts are innocent. It comes from structured interviews with participants in a trial of text reminders for contraceptive-clinic attendance, where 78 per cent attributed non-responsiveness to being away from the phone or a conflict such as school or work, 44 per cent cited service interruption from unpaid bills, and 28 per cent each cited phone loss and number changes⁹.

Read the population before you use it. These are urban adolescent girls responding to automated clinical reminders from a health service, in an economically constrained sample where phone-service interruption was a major factor. The messages are not from a person the respondent has feelings about. Using “78 per cent of unanswered texts are innocent” as a general claim about texting would misuse the study.

My bottom line is that nobody has quantified, on a representative sample of ordinary interpersonal messaging, what fraction of unanswered messages reflect disinterest against everything else. That is an evidentiary vacuum. It should not be filled with either a comforting guess or a catastrophising one.

Why the not knowing is the worst part

This has the best evidence in the article.

De Berker and colleagues had 45 volunteers play a task turning over rocks that might conceal snakes delivering a mild shock¹⁰. A 50 per cent chance of shock produced more stress than a 100 per cent certain shock. Salivary cortisol, pupil diameter and skin conductance all tracked uncertainty rather than badness. Certain bad news was less stressful than not knowing.

An EEG study found people high in neuroticism showed stronger neural responses to uncertain feedback than to explicitly negative feedback¹¹. Same direction.

One thing to retire. I wouldn’t use the popular framing that an unanswered message is an open loop your brain cannot close, because of the Zeigarnik effect. A 2025 meta-analysis found no reliable memory advantage for unfinished tasks once Zeigarnik’s own 1927 data are excluded¹². The related tendency to resume interrupted tasks did hold, which is weaker and still serviceable.

If you are someone who reads rejection into ambiguity quickly, that is a measured disposition, not a character flaw, and it is the subject of are they mad at me, or am I overthinking it. Rejection sensitivity is defined as anxiously expecting, readily perceiving and overreacting to rejection, and the experimental evidence shows high-scoring people read intentional rejection into others’ ambiguous behaviour specifically¹³. Ambiguous silence is precisely the cue this trait is built to misread. I’d note that the construct has a fair psychometric challenge attached, in how much it captures beyond general negative emotionality.

What to do

Send the follow-up, once. People systematically underestimate how much others appreciate being reached out to. A series of preregistered experiments found this robustly, and an independent well-powered replication with 742 participants reproduced it at d = 0.25¹⁴. Note the size. Real, replicated, and modest, and the original authors flagged their own focal experiment as underpowered. If the worry is whether your follow-up text sounds desperate, that has its own research.

Do not escalate. No peer-reviewed randomised trial has tested whether double texting increases reply rates in an interpersonal or romantic context. All three research passes I ran behind this article converge on that. What the evidence supports is narrower and still useful. Your catastrophic prediction about how a follow-up will land is probably miscalibrated, and one warm, low-pressure, non-demanding follow-up is more likely to be welcomed than you fear. Repeated emotionally loaded follow-ups have no evidence behind them and a coherent mechanism against them.

Make it easy to answer. A short, specific, low-pressure message asks less than “are you ignoring me”, and the composition cost you are asking someone to pay is one of the documented reasons the reply never came.

Try Subtext in your browserGet the appTry Subtext in your browser

When silence stops being ordinary

Ordinary non-reply, conflict avoidance and needing space are not abuse. The boundary is real and definable, and it turns on pattern, transparency and the conditions under which the silence ends. Not on duration, and certainly not on one unanswered message.

Withdrawing during conflict is often a capacity failure rather than a strategy. In Gottman’s marital work, stonewalling is described as physiologically driven, with the withdrawing partner disengaging to self-soothe. I’d be sceptical of the famous divorce-prediction accuracies attached to that research, which come from models fitted and tested on the same small samples, not from prospective prediction on fresh couples.

Silence used within a pattern of control is a different thing. Three questions separate them:

1. Is the need for space communicated and shaped? Healthy regulation says something like “I am overwhelmed, I need half an hour.” Coercive silence is imposed with no explanation, no timeframe and no exit criteria.
2. What ends it? Withdrawal ends when arousal subsides. Coercive silence ends only when you capitulate, surrender something, or apologise for a perceived offence. This is the sharpest single discriminator.
3. Is there a pattern and a power imbalance? Occasional dysregulation happens in healthy relationships. A repeating cycle of offence, silence, capitulation and reset, producing hypervigilance and changed behaviour in you, is not that.

In England and Wales, controlling or coercive behaviour in an intimate or family relationship has been a criminal offence since the Serious Crime Act 2015 came into force, where the behaviour is repeated or continuous between people who are personally connected and has a substantial adverse effect on day-to-day activities¹⁵. Scotland’s Domestic Abuse (Scotland) Act 2018 takes a similar pattern-based approach.

These criteria come from clinical guidance and are not a validated instrument, and I’d hold them lightly in one direction. A partner who goes quiet during a fight and comes back without asking you to give anything up is most likely regulating an overwhelmed nervous system, not abusing you. The difference is legible, and it is not about how many hours passed. If the second description fits what you are living with, talking to someone outside the situation is worth more than getting the next message right.

Sources

Numbered in the order they appear above. Where a figure could not be verified against the primary source, the text says so.

1. Hartgerink, C. H. J., van Beest, I., Wicherts, J. M., and Williams, K. D. (2015). The ordinal effects of ostracism: A meta-analysis of 120 Cyberball studies. PLOS ONE, 10(5), e0127002. N = 11,869. Funded in part by NSF grant BCS-1339160 and NWO grant 016-125-385.
2. Eisenberger, N. I., Lieberman, M. D., and Williams, K. D. (2003). Does Rejection Hurt? An fMRI Study of Social Exclusion. Science, 302, 290 to 292. Thirteen undergraduates. Funded in part by NIH R21MH66709-01.
3. Woo, C.-W., and colleagues (2014). Separate neural representations for physical pain and social rejection. Nature Communications, 5, 5380. Sixty participants.
4. Smith, A., and Williams, K. D. (2004). R U There? Ostracism by Cell Phone Text Messages. Group Dynamics: Theory, Research, and Practice, 8(4), 291 to 301. Approximately 40 students. Funded by the Australian Research Council. Effect sizes circulating for this study could not be verified.
5. Büttner, C., Ren, D., Stavrova, O., Rudert, S. C., Williams, K. D., and Greifeneder, R. (2025). Ostracism in everyday life. Journal of Personality and Social Psychology, 129(5), 870 to 887.
6. Hoyle, R., Das, S., Kapadia, A., Lee, A. J., and Vaniea, K. (2017). Was my message read? Privacy and Signaling on Facebook Messenger. CHI 2017, 3838 to 3842. Senders n = 402, recipients n = 316.
7. Mai, L. M., Freudenthaler, R., Schneider, F. M., and Vorderer, P. (2015). “I know you’ve seen it!” Computers in Human Behavior, 49, 296 to 302. Online survey, n = 207.
8. Chou, Y.-H., Lin, C., Lin, S.-Y., You, C.-W., and Chang, Y.-J. (2022). Why Did You/I Read but Not Reply? CHI 2022. Sample size unresolved across sources; at least 25. Funded by Taiwan’s National Science and Technology Council.
9. Irons, R., Tomaszewski, K., Muñoz Buchanan, C. R., and Trent, M. (2015). Understanding Adolescent Nonresponsiveness to Text Messages. Journal of Urban Health, 92(3), 502 to 512. Urban adolescent girls responding to automated clinic reminders; interview subsample size not recoverable.
10. de Berker, A. O., and colleagues (2016). Computations of uncertainty mediate acute stress responses in humans. Nature Communications, 7, 10996. Forty-five volunteers. Funded by the UK Medical Research Council.
11. Hirsh, J. B., and Inzlicht, M. (2008). The Devil You Know: Neuroticism Predicts Neural Response to Uncertainty. Psychological Science, 19(10), 962 to 967. Sample size unverified.
12. Ghibellini, R., and Meier, B. (2025). The Zeigarnik effect: a meta-analysis. Humanities and Social Sciences Communications, 12, 962.
13. Downey, G., and Feldman, S. I. (1996). Implications of Rejection Sensitivity for Intimate Relationships. Journal of Personality and Social Psychology, 70, 1327 to 1343. Funded by NIMH R29-MH51113 and the Harry Frank Guggenheim Foundation. Per-study sample sizes not independently verified.
14. Liu, P. J., Rim, S., Min, L., and Min, K. E. (2023). The Surprise of Reaching Out. Journal of Personality and Social Psychology, 124(4), 754 to 771. Independent replication by Transparent Replications, n = 742, d = 0.25.
15. Serious Crime Act 2015, section 76, in force 29 December 2015. Controlling or coercive behaviour in an intimate or family relationship.