The science we build on

We did not invent these ideas.
We just made them daily.

LVRS FRVR is built on three decades of peer-reviewed relationship research. The most rigorous modern study on what predicts a strong relationship was published in 2020. It identified five factors. We track all five. Every day.

The 5 predictors of relationship quality

Joel and her co-authors analyzed 43 longitudinal datasets covering 11,196 couples. These five factors explained the most variance in relationship quality. Source listed below.

01

Perceived partner commitment

Whether you feel your partner is in this with you for the long haul.

What we built for it

Pulse score

Your daily Pulse tracks the quiet signals of commitment, not just the loud ones.

02

Appreciation

How often each of you feels seen and valued by the other.

What we built for it

Tiny Reaches and shared sparks

Small daily moments compound. We make them visible so they do not slip past.

03

Sexual satisfaction

How connected you feel to each other physically and emotionally.

What we built for it

Daily ritual prompts

Gentle, private check-ins that surface drift before it becomes distance.

04

Perceived partner satisfaction

Whether you sense your partner is happy in the relationship.

What we built for it

Combined Pulse and asymmetry alerts

When one partner pulls away, we flag it early and together. No surprises.

05

Conflict patterns

How quickly the two of you come back after a hard moment.

What we built for it

Repair Velocity

We measure how fast you reconnect, not whether you fight. Speed matters.

Source: Joel et al. (2020). Machine learning uncovers the most robust self-report predictors of relationship quality across 43 longitudinal couples studies. Proceedings of the National Academy of Sciences, 117(32), 19061-19071.

The studies behind the product

Three decades of converging evidence. The science is not new. The daily practice is.

1994

Gottman, J. M. What Predicts Divorce?

Predicted divorce with 93.6% accuracy from observed communication patterns. The foundation paper.

Hillsdale, NJ: Erlbaum.

2019

Yöntem et al. Divorce Predictor Scale with Machine Learning

Replicated and extended the Gottman framework with modern ML algorithms. Reached 97.6% accuracy across 170 couples.

International Journal of Innovative Studies in Sciences and Engineering Technology.

2020

Joel et al. The 11,196 Couples Study

Used machine learning across 43 longitudinal datasets and 11,196 couples to identify the most robust predictors of relationship quality. Published in PNAS, the most prestigious general-science journal in the United States.

Proceedings of the National Academy of Sciences, 117(32), 19061-19071.

2007

Amato and Hohmann-Marriott. High vs Low-Distress Divorce

Found that the majority of divorces happen in low-distress marriages. Couples who described things as fine often drift quietly until separation. Drift hides in plain sight.

Journal of Marriage and Family, 69(3).

1974

Minuchin, S. Families and Family Therapy

Identified the boundary types that govern healthy family systems: clear, diffuse, and rigid. Our WARD assessment maps onto these directly. Watchtower reads as clear boundaries, Absorber as diffuse, Rampart as rigid. Drifter is the gap Minuchin's frame implies but does not name. The boundary you cannot see is the one that erodes the bond.

Harvard University Press.

1989

Bolger, DeLongis, Kessler and Wethington. The Contagion of Stress Across Multiple Roles

Showed that stress from outside the relationship, especially work, crosses into the couple through one partner and lands on both. The empirical grounding for our Protection corner. Outside pressure is not just an individual problem. It shapes the dyad. WARD measures how each partner responds when it arrives.

Journal of Marriage and the Family, 51(1), 175-183.

1999

Boss, P. Ambiguous Loss: Learning to Live with Unresolved Grief

Introduced boundary-ambiguity theory: when a partner cannot say what is in or out of the relationship, outside forces erode the bond silently. This is the third leg of our WARD assessment, and the empirical grounding for the Drifter style specifically. Drift is not coldness. It is the absence of a boundary anyone can see.

Harvard University Press.

1993

Ickes, W. Empathic Accuracy

Introduced the empathic-accuracy paradigm: measuring how accurately one partner infers another's thoughts and feelings in real time. The foundational construct behind our READ assessment. Reader and Empath both score high here. The split is how they get there. Reader checks. Empath senses.

Journal of Personality, 61(4), 587-610.

2003

Simpson, Orina and Ickes. When Accuracy Hurts, and When It Helps

Showed that empathic accuracy is not uniformly good. Its effect on the relationship depends on context. The empirical reason our READ results elevate Reader (asked) over Empath (sensed) when the read matters most, even though intuition is often right.

Journal of Personality and Social Psychology, 85(5), 881-893.

1997

Swann and Gill. Confidence, Accuracy and Consensus in Person Perception

Found that people are more confident than accurate about intimate partners, and that the confidence gap widens with relationship length. The grounding for the Assumer style: a trusted model of your partner that has quietly fallen out of date.

Journal of Personality and Social Psychology, 73(4), 747-757.

What this is not

Not therapy

We are not licensed counselors. We are an instrumentation layer. If you are in crisis, please find a human.

Not a guarantee

Statistical patterns describe groups, not your specific relationship. The research is a map, not a forecast.

Not a verdict

A low Pulse is not a verdict. It means there is something worth talking about, with care.

What the research cannot tell us

Joel and her team explained roughly 45% of the variance in relationship quality from these five factors. That leaves 55% unexplained. Some of that is noise. Some is the part of human connection that no model has yet captured.

Self-report has biases. People answer how they wish they felt, not always how they actually feel. We design our daily ritual to be short, gentle, and repeated. Repeated honest signals beat single perfect ones.

No study can predict your specific relationship. The science is a starting point. The daily practice is yours to build.

Important transparency

  1. LVRS FRVR is a relationship wellness app. It is not therapy, not medical advice, and not a substitute for licensed mental-health care. If you or your partner are in crisis, please contact a licensed professional or, in the United States, the 988 Suicide and Crisis Lifeline.
  2. All third-party studies cited on this page are independent peer-reviewed work. LVRS FRVR is not affiliated with the Gottman Institute, Dr. John Gottman, the authors of any cited paper, the Journal of Marriage and Family, or the Proceedings of the National Academy of Sciences. We are a consumer of their published findings, nothing more.
  3. Research findings describe groups of couples on average. They cannot predict outcomes for any individual relationship. A daily Pulse score reflects patterns found in the published literature. It is a signal, not a verdict.
  4. LVRS FRVR has not yet conducted or published its own peer-reviewed research. When we do, through Love Pulse Labs, we will pre-register every study, follow IRB or equivalent review, and disclose methodology and conflicts of interest in line with academic norms.
  5. Statistics shown elsewhere on lvrsfrvr.com that include a number or percentage are sourced from the studies listed above. If you ever see a claim without a source, please email us so we can correct it.

See where you stand. In 5 minutes.

The Drift Check is free. No signup needed. The science behind it is on this page.

Take the Free Drift Check