What actually helps a couple, according to trials
Most relationship advice has never been tested. Some of it has — and the tested things are smaller and stranger than the advice industry suggests. Here is what held up in randomised trials and diary studies.
Prepared by NUMERA
The interventions are shorter than you would guess
A trial with 215 couples tested therapist-led Marriage Checkups: assessment and feedback sessions, repeated a year later. Intimacy differences favoured treatment across the two-year follow-up, with a median effect of d = .37. This was a defined clinical intervention, not simply two informal conversations.
In another study, 120 couples were asked to spend seven minutes, three times a year, writing about a recent conflict from the perspective of a neutral third party who wants the best for everyone. Twenty-one minutes across a year. The couples who did it showed none of the decline in relationship quality that the control group showed in the second year.
Both studies tested specific procedures. Their results cannot establish that any structured conversation, journal prompt or commercial reading produces the same benefit.
What affect-labeling research does and does not show
In a laboratory fMRI study, labeling emotions in negative images reduced amygdala responses relative to comparison tasks. This is a finding about that task, not evidence that a written relationship reading calms an argument.
NUMERA can offer words and prompts to consider. Any benefit from using them together remains untested for our product.
Appreciation and an unequal division of labour
Across three pandemic-era samples involving 2,193 people, including 476 couples, unequal or unfair labour was associated with lower satisfaction when people did not feel appreciated; that association was absent when they felt appreciated.
This was not a trial assigning appreciation or workloads. It does not show that thanks can replace a fair division of work. Both recognition and the practical arrangement can belong in the conversation.
What your partner notices is not what your partner does
Phone use during time together correlates with lower satisfaction and more conflict across 52 studies covering 19,698 people. But a 14-day diary study of 100 couples separated the two halves of that: perceived phone-snubbing predicted how the relationship felt, while actual measured phone use predicted nothing — not the next day, not two months later.
The practical reading is that these arguments are about attention and its signals, not screen time. Arguing about minutes will miss it.
The context nobody is accounting for
In the American Psychological Association's 2025 Stress in America survey of 3,199 US adults, 54% said they felt isolated often or some of the time, and 69% said they could have used more emotional support over the past year than they received — up from 65% in 2024.
A lot of what lands inside a relationship is a deficit that was carried in from outside it. That is worth knowing before deciding that the relationship is the problem.
How NUMERA fits this, honestly
We build a reflective reading from two birth dates and your answers. The dates provide symbolic themes, not a validated measurement of compatibility.
The report offers reflection questions and concrete next steps. It has not been tested in these trials, does not reproduce therapist-led care, and cannot claim their outcomes. You can assess whether a prompt is useful without treating it as a prediction.
Questions people ask
What relationship interventions have actually been tested in trials?
The trials discussed here tested therapist-led assessment and feedback with 215 couples, and a specific perspective-taking writing exercise with 120 couples. They found benefits under those protocols; neither tested NUMERA or any generic conversation prompt.
Does dividing household labour more equally improve a relationship?
The appreciation study discussed here was observational, not a test of redistributing work. It found that feeling appreciated moderated the association between unequal labour and satisfaction; it does not show that an unfair arrangement should remain unchanged.
Is phone use really damaging relationships?
What matters is perception. Across 52 studies of 19,698 people, phone-snubbing correlated with conflict, but a 14-day diary study of 100 couples found that perceived phone-snubbing predicted relationship outcomes while actual measured phone use did not.