A new set of experiments from the University of Cologne’s Social Cognition Center Cologne and the Centre for Social and Economic Behavior reveals a striking asymmetry in how people judge potential partners online. Across four studies, comprising seven experiments with a total of 1,074 participants, researchers find that positive traits are named more frequently and cluster tightly around a shared ideal, while negative traits span a broader semantic range and prove more informative for dating decisions in hypothetical scenarios.
Leading the work are Tabea J. Zorn and Christian Unkelbach. Unkelbach described their framework as part of a broader Evaluative Information Ecology model. “My colleagues and I have a general model of evaluative information (i.e., good and bad behaviors/traits/people, etc.) and what we call its ‘ecology,’” Unkelbach told PsyPost. “In a nutshell: Positive information is more frequent, but negative information is more diverse (or less similar). We call this the structural properties of positivity prevalence and negativity diversity.” He added that the study tests whether these assumptions hold in dating contexts and what the findings might imply for self-presentation online.
Across the four studies, the researchers tested how people generate, evaluate and act upon positive and negative traits in dating scenarios. Every experiment was preregistered, and the team posted their data, materials and analysis code publicly online.
In the first round of experiments, Studies 1a and 1b surveyed two groups of 100 English-speaking adults from the United Kingdom and United States via Prolific, all with prior online dating experience. Study 1c, run in a German laboratory, involved 98 German-speaking adults (after excluding three under 18) with a mix of dating experience. Across these three experiments, between 19 and 31 per cent of raw trait entries were filtered out because they did not describe character or personality traits.
Participants named more desirable qualities than undesirable ones, averaging between 3.61 and 4.26 positive traits per person compared with 2.44 and 2.88 negative traits. In the German sample, neither prior online dating experience nor the order in which traits were presented affected how many traits were generated. Even when the researchers reanalysed all raw entries without filtering out non-personality items, the higher count of positive entries persisted in Studies 1a and 1c; in Study 1b the difference was not statistically significant.
Although the researchers did not set out to test trait variety in these initial experiments, pooling all distinct English trait words from Studies 1a and 1b revealed a broader picture: 181 unique negative traits versus 152 unique positive traits, with four words—“educated,” “flirtatious,” “needy,” and “political”—falling into overlapping categories where some participants listed them as desirable and others as flaws. Negative words included some direct opposites of positives, but also flaws without clear positive counterparts, as well as excesses of desirable traits such as “overconfident.”
Study 2 then mapped the semantic landscape of the traits. Involving 175 English-speaking adults, participants rated seven lists drawn from the pool of 333 traits on three foundational dimensions of meaning: pleasantness, emotional intensity and sense of control. After calibrating responses with a set of 12 reference words, the researchers plotted the traits in a three-dimensional map and calculated each word’s distance from others within the same category. Positive traits clustered more closely to one another, averaging a distance of 1.96, while negative traits were more dispersed, averaging 2.33. Follow-up checks removing the most extreme words did not alter the overall pattern. “Most people agree on what positive traits are in online dating contexts, but there is more variety on what people consider ‘negative,’” Unkelbach explained.
Study 3 examined how a trait’s structural similarity relates to its perceived importance. On average, participants rated avoiding negative traits as modestly more important than seeking positive traits (5.51 vs 5.37 on a 7-point scale). For positive qualities, those closer in meaning to other desirable traits tended to be rated more important, reinforcing classic impression-formation theories that central traits carry extra weight. The same pattern appeared for negative traits, though the link between closeness and importance was stronger for flaws than for positive traits. Across words, “trustworthy” topped the positive dealmakers list (6.73/7), with “abusive” leading negative dealbreakers (6.98), followed by other terms such as “unfaithful,” “dishonest,” and “deceiving.” The researchers note that, among the top items, a word like “fake” appeared in both categories for some participants.
In Studies 4a and 4b, the researchers tested whether structural similarity could predict hypothetical dating app choices beyond stated importance. After excluding words with inconsistent ratings, 308 traits remained (146 positive and 162 negative). In Study 4a, 200 Prolific participants imagined a partner described by a single word and chose “swipe left” or “swipe right.” Study 4b replicated with a 7-point likelihood scale (1 to 7). For positive traits, stated importance strongly predicted swiping decisions in both experiments, with about 95 per cent of partners described by desirable traits accepted in Study 4a, and an average of 5.84 out of 7 in Study 4b. Structural similarity did not significantly predict positive-s Trait swiping in either experiment, suggesting a ceiling effect in the binary task did not mask the effect.
Responses to negative traits varied more widely. In Study 4a, around 15 per cent of participants were willing to swipe right for an undesirable trait, with ratings averaging 2.18 out of 7 in Study 4b. In both experiments, a trait’s importance and its structural similarity each independently predicted swiping decisions: even after accounting for stated importance, flaws that clustered more closely with other negative traits reduced the likelihood of a right swipe. Across all prompts in Study 4a, 52.9 per cent of choices were right swipes, even though negative traits outnumbered positives in the prompts. The researchers describe this as a “progression bias” toward extending a potential match the benefit of the doubt in early dating decisions.
Unkelbach cautioned that the swiping findings should be interpreted with care. “We must acknowledge that these are self-reports of people in hypothetical scenarios,” he said. “The importance for real-life decisions should not be overstated. The model test (i.e., do the structural properties we propose also hold here) is solid, though.” He also noted that the experiments used single-word descriptors and did not include profile photos, factors that strongly influence real dating app decisions.
The team also flagged potential directions for future work, including how combinations of positive and negative traits operate on fuller profiles and whether initial swiping patterns translate into face-to-face dates or apply across non-Western cultures. They observed that most daters tend to share a common cluster of positive qualities, and raised the question of whether acknowledging a minor flaw might ever make a profile seem more genuine. In Studies 4a and 4b, even mild flaws yielded low right-swipe rates when shown alone, suggesting limited value in signalling negative information online—at least in the contexts tested so far.
Asked what surprised the team most, Unkelbach highlighted the selective impact of negative traits. “That we found such a differentiated influence on the negative side – we expected that if one has a negative trait, one is ‘out.’ However, people differentiated,” he said. Across the swiping studies, stated importance remained a stronger predictor of choices than similarity, though similarity added extra predictive power on the negative side. The researchers suggest that negative traits were more informative for distinguishing swiping decisions because consensus on what to seek is higher than how to avoid, when choices are framed around individual traits.
The study, “Valence Asymmetries of Traits in Dating Contexts,” was authored by Tabea J. Zorn and Christian Unkelbach.
