What the research says

This field has a habit of repeating striking numbers that don't survive scrutiny. We cite sources, and we say when a finding is contested.

Our standard for citing things

If we give a number, we link to where it came from. If a number is contested in the literature, we say so — as with the widely repeated claim about predicting divorce with over 90% accuracy, which Heyman & Slep (2001) challenged on the grounds that the models were fitted and evaluated on the same sample. The Gottman Institute's own research FAQ frames the finding more carefully than most secondary coverage does.

We describe published concepts and cite their sources. Nearvo is not affiliated with, endorsed by, or licensed by any research institute or author whose published work is discussed on this site.

Everything in this section

• Attachment styles, and where the framework stops — Attachment theory is well studied and routinely over-applied. What the research supports, why the quiz version fails, and how to use the idea without
• Does a daily question actually do anything? — What the research supports, what it does not, and the honest version of the claim a couples app can make. Includes the study everyone cites and what i
• What this category claims, and what it can evidence — A line-by-line audit of the claims made by relationship apps, including ours, against the research they cite. Most of the citations do not support the
• What perpetual conflict means — A substantial share of couple disagreement is rooted in stable differences and will not be resolved. Why treating a perpetual problem as solvable prod
• Why most relationship advice does not generalise — The advice industry runs on rules derived from unrepresentative samples and presented as universal. What the better research says about that, and how
• Why relationship quizzes do not change anything — The most common complaint about couples apps is that a quiz produces a result and no next action. Why a category label is not an instruction, and what