What the research says about texting and relationships

Two things, mostly. First, that how people communicate matters more than how much — the presence of warmth, responsiveness, and repair predicts more than raw message count. Second, that the reliable signals are patterns over time, not single messages. Almost every finding below is a tendency measured across many interactions or many people, which is exactly why no single text deserves to carry the weight of a conclusion.

What the research does not do is hand anyone a verdict about a specific relationship from a specific thread. The studies describe populations and probabilities. Applied honestly, they help you notice a pattern and read it charitably — never certify what your partner “really” means by one reply.

It is tempting to assume that more contact equals more closeness. The evidence complicates that. In a 2014 study published in Computers in Human Behavior, Luo found that the higher the share of texting out of all communication, the lower people’s reported relationship satisfaction— and that attachment anxiety and avoidance both tracked with more texting, not less. Volume can be a symptom as easily as a sign of health.

Frequency norms are also just norms. Pew Research Center found that 85% of teen daters expect to hear from a partner at least once a day and 11% expect hourly contact, with 92% having spent time texting a partner. Those are descriptions of expectation, not thresholds for a good relationship. A daily text means one thing in one couple and something else entirely in another.

Relationship science has a name for the ingredient that keeps surfacing: perceived partner responsiveness — the sense that someone understands you, values you, and is there for you. It is one of the steadiest predictors of closeness, and it shows up over text too. A 2025 study in PLOS ONE found that messages containing emojis were rated as more responsive than text-only ones, and that perceived responsiveness in turn predicted closeness and satisfaction. The signal was not the emoji itself; it was the felt sense that someone was engaged.

Over text, responsiveness looks concrete: a reply that answers what you actually said, a follow-up on the thing you were nervous about, a question that opens the door rather than closing it. This is one of the patterns ReadBeneath looks for when you upload a conversation — reciprocity and responsiveness read across the whole thread, per speaker, with the messages that support each read cited. It measures the pattern; it does not decide what the pattern is worth to you.

The most durable framework here comes from The Gottman Institute’s long observational studies of couples. Their “Four Horsemen” — criticism, contempt, defensiveness, and stonewalling — are the communication patterns most associated with relationships coming apart, with contempt named the single greatest predictor of divorce. The counterweight they document is balance: successful couples kept roughly five positive interactions for every negative one during conflict, and about twenty to one outside of it.

The word to sit with is pattern. A single sharp message is not contempt; a run of mockery, eye-rolling, and disrespect over time is. That distinction is exactly why ReadBeneath surfaces the four horsemen per speaker and ties each observation to cited messages rather than declaring a relationship healthy or not. The framework describes a tendency worth noticing — it does not certify a verdict about any one exchange.

The most quietly powerful finding in the Gottman work is about the small stuff. Tracking newlyweds over years, they found that couples who were still together had turned toward each other’s everyday “bids” for attention 86% of the time, while couples who divorced had turned toward them only 33% of the time. A bid is tiny — a shared link, a “look at this,” a passing worry mentioned in a text. Answering it is what accumulates into connection.

Alongside turning toward bids sits repair: the ability to circle back after friction, take a slice of ownership, and reach for the other person instead of the last word. In the Gottman research, sturdy couples were not the ones who never argued — they were the ones whose repair attempts landed. Over text, both patterns are visible in the shape of a thread: whether small reaches get met, whether a rough exchange finds its way back.

It is common, measurable, and worth taking seriously without alarm. Pew Research Center reports that 51% of partnered adults say their significant other is at least sometimes distracted by their cellphone when they are trying to talk, including 16% who say it happens often. Researchers call the specific act “phubbing” — phone snubbing — and Roberts and David found that 46.3% of people had been phubbed by a partner, with perceived phubbing linked to more conflict and lower relationship satisfaction.

The wider backdrop is that checking is now the default. The American Psychological Association’s Stress in America work found that 86% of U.S. adults constantly or often check email, texts, and social media, and that these “constant checkers” report higher stress. None of this means a distracted partner is a failing one. It means the pull of the phone is a shared, structural fact — better named gently than treated as a personal indictment.

A few things hold up. Responsiveness beats frequency. Contempt is worth taking seriously; warmth and repair are worth protecting. Small bids, answered, are the raw material of closeness. And distraction is common enough that noticing it says little about either person’s worth. Every one of these is a statement about patterns, which is the quiet thread running through all of it.

That is also the boundary. None of this research can read a single screenshot and tell you what it means. Patterns need a sample, and the honest floor is more history than most people assume — see how many messages it takes to see real patterns. A tool that respects that will show you a pattern, cite the messages under it, offer a fair alternative reading, and stay quiet when the evidence is thin — which is the posture ReadBeneath is built around.

Every figure above is drawn from a named, published source. Links open in a new tab.

• The Gottman Institute

  The Four Horsemen: criticism, contempt, defensiveness & stonewalling

  Names the four corrosive patterns and identifies contempt as the single greatest predictor of divorce.

• The Gottman Institute

  The Magic Ratio: the key to relationship satisfaction

  Stable couples kept about five positive moments for every negative one during conflict, and roughly 20 to 1 outside it.

• The Gottman Institute

  Start paying more attention to bids for connection

  In the newlywed studies, couples who stayed together turned toward each other's bids 86% of the time; those who divorced, only 33%.

• Pew Research Center

  Dating and Relationships in the Digital Age (2020)

  51% of partnered adults say their partner is at least sometimes distracted by their phone mid-conversation; 16% say often. Survey fielded October 2019.

• Pew Research Center

  Teens, Technology and Romantic Relationships (2015)

  85% of teen daters expect to hear from a partner at least once a day, 11% hourly; 92% had spent time texting a partner.

• American Psychological Association

  Constantly checking devices is linked to stress (Stress in America, 2017)

  86% of U.S. adults say they constantly or often check email, texts, and social media; 43% qualify as 'constant checkers.'

• Roberts & David, Computers in Human Behavior (via Baylor University)

  Partner phubbing and relationship satisfaction (2016)

  46.3% of respondents reported being 'phubbed' by a partner; perceived phubbing tracked with more conflict and lower relationship satisfaction.

• Luo, Computers in Human Behavior

  Effects of texting on satisfaction in romantic relationships: the role of attachment (2014)

  A higher share of texting out of all communication was associated with lower relationship satisfaction; attachment anxiety and avoidance both tracked with more texting.

• Huh, PLOS ONE

  Emojis, perceived responsiveness, and satisfaction in text messaging (2025)

  Messages with emojis were rated more responsive than text-only ones, and perceived responsiveness predicted closeness and satisfaction.

• Analyze a WhatsApp conversationUpload a .txt export and get per-speaker patterns — reciprocity, responsiveness, the four horsemen — each tied to cited messages.
• See all conversation analyzersWhatsApp, iMessage, Telegram, and Instagram exports, each read the same honest way: patterns with receipts, never a verdict.
• How many messages reveal a real pattern?The sample-size floors an analysis waits for before naming anything — the boundary this whole roundup keeps pointing at.
• Green flags in textingWhat reciprocity, repair, and attentiveness look like in a thread — the healthy side of the same research, shown in examples.

More on the patterns worth naming lives in the Learn hub.