Relationship Surveys May All Be Measuring The Same Feeling

Why Relationship Research May Be Predicting Itself

Ask someone how much they trust their partner, and their answer will probably sound a lot like how they’d rate their communication, their commitment, and their overall happiness. That’s no coincidence, according to a new study. It raises the possibility that a large chunk of relationship research has been asking dozens of different questions and often getting back the same underlying answer.

A team of researchers found that when people fill out surveys about their romantic relationships, their responses to a huge range of supposedly separate topics, from trust and intimacy to passion, appreciation, conflict, and more, are overwhelmingly driven by a single underlying feeling: how good or bad they think their relationship is, period. The researchers call this feeling “Q,” a general factor representing overall relationship sentiment. It swallowed up the vast majority of meaningful variation across 34 commonly used relationship measures.

Relationship science has spent years developing hundreds of specialized tools to measure specific parts of how couples function, assuming that trust is different from commitment, which is different from intimacy, which is different from satisfaction. If people are largely telling researchers the same thing in different words, then a large portion of the field’s measurement practices, and some conclusions that depend on treating these measures as distinct, may deserve another look.

A Single Feeling Explains Most of the Overlap

The study, published in PLOS ONE, unfolded across two investigations involving a combined 3,439 participants recruited from a U.S. census-matched national panel. All participants were adults between 18 and 75 currently in romantic relationships.

In the first study, 2,000 participants completed 206 items drawn from 29 widely used relationship satisfaction scales, including the Dyadic Adjustment Scale, the Couples Satisfaction Index, and the Investment Model Scale. The original pool of 512 items was trimmed for duplicates, preserving items spanning 27 supposedly separate relationship topics.

The researchers used a statistical technique that reveals whether survey questions are measuring several distinct things or mostly one big thing. A single feeling dominated the results. In plain terms, about three-quarters of the meaningful overlap in people’s answers about trust, appreciation, communication, and commitment could be chalked up to that one general feeling about whether the relationship was good or bad.

A Second Study Finds the Same Result Again

The second study expanded the scope to 1,439 new participants, 34 relationship measures, and topics including gratitude, forgiveness, and sexual satisfaction. The same pattern showed up again, just as strong. Sexual satisfaction was the one area that stood somewhat apart from the general feeling in both studies. Trust, commitment, intimacy, love, and forgiveness were largely absorbed by it.

Trust and Satisfaction May Be Predicting Each Other

The researchers frame their results through a concept called “sentiment override,” the idea that when people evaluate specific aspects of a relationship, their answers appear to be heavily shaped by an overall evaluation of the relationship rather than by weighing each dimension independently. This phenomenon was described in relationship research decades ago, but the current study provides new empirical evidence for its scope in self-report surveys.

Consider a common study design: a researcher measures trust and satisfaction separately, then reports that trust “predicts” satisfaction. If both measures are mostly tapping the same underlying sentiment, that finding could be partly circular. As the paper’s title provocatively asks: is the field “predicting relationship quality with itself?”

The researchers aren’t arguing that trust, commitment, and intimacy don’t exist as real, distinct aspects of relationships. They’re arguing that standard self-report surveys, the workhorse of relationship science, may not be capable of capturing those distinctions. A recent review cited in the paper found self-report data appeared in 96% of 771 relationship studies published between 2014 and 2018.

The paper draws a parallel to personality psychology, where a similar problem was eventually addressed through consolidation efforts like the Big Five model. Relationship science, the authors suggest, may need its own version: a systematic effort to sort which ideas are genuinely distinct from which are just different labels for the same feeling.

Researchers May Need New Tools Beyond Self-Report

None of this means relationship science should be tossed out. But the field may need to rethink its measurement toolkit, leaning more on observational methods, behavioral tasks, or daily diary studies that don’t rely so completely on subjective self-evaluation.

The findings may also matter for couples therapy and relationship education research, where self-report scales are often used to track outcomes. If several scales capture much of the same general sentiment, they may provide less distinct information than their labels suggest.

People know whether things feel good or bad in a relationship, and that feeling shapes nearly everything they report about it. Decades of carefully labeled measures may, in many cases, have been asking different versions of essentially the same question.

---