A worldwide test of the predictive validity of ideal partner preference matching.
Level 3 - non-randomized controlled study
Large-scale cross-sectional observational multi-site study (registered report)
PubMed 39480282 · doi:10.1037/pspp0000524
What was done
A registered report conducted through the Psychological Science Accelerator examined the predictive validity of ideal partner preference matching across 43 countries and 22 languages (N = 10,358). The study tested partnered and single participants using a 35-trait corrected pattern metric (assessing overall trait profile matching against normative profiles) and a level metric (trait-by-trait interaction tests). It also evaluated partnered individuals who recalled ideals held prior to relationship entry and compared stated versus revealed preferences for traits including physical attractiveness and earning potential.
What was found
Significant preference-matching effects were observed across the full sample as well as within single and partnered subgroups. The corrected pattern metric across 35 traits yielded a zero-order effect of β = .19, which remained β = .11 when adjusting for normative preference matching. Specific trait-level interaction tests demonstrated very small associations (average β = .04). Results were similar among partnered participants reporting pre-relationship ideals, with no consistent moderation by individual differences. For physical attractiveness, stated preferences underestimated revealed preferences in both sexes (particularly women). For earning potential, men underestimated and women overestimated their revealed preferences.
Why it matters
This study provides high-powered, cross-cultural evidence clarifying a long-standing debate in human mating research, showing that people evaluate partners more positively when they match their ideals, although the magnitude of this effect is modest at the profile level and very small for isolated traits.
Limits
The study relies on self-reported preference and evaluation ratings rather than objective relationship longevity or behavioral choices. Retrospective recall of pre-relationship ideals in partnered individuals carries risk of recall bias, and observational designs cannot definitively establish causal effects of preference matching.