Young adults turn crushes into love, study suggests

The image of young adults living in a hookup culture with emotionally meaningless relationships might be a common theme in movies and daytime talk shows. But it does not seem to be the norm in real college life, suggests a new study from University of California, Davis, researchers.

The study, published May 28, is the first of its kind to look at early relationship development—the time period in which people experience rising and falling romantic interest for partners who could, but often do not, become committed partners; in other words, "crushes." Previous studies have examined committed relationships on one extreme, and first impressions on the opposite extreme.

To obtain their data, researchers surveyed 208 heterosexual college students at a Midwestern university about their dating habits, their various likes and dislikes, and attraction to potential partners over a seven-month period. Participants described an average of five crushes during this stretch of time and reported about 15% of them turning into dating relationships at some point. They collected a total of over 7,000 reports on these potential partners.

What predicted whether or not these attractions ultimately fizzled?

"What took us by surprise is that many of the important factors were the same things you would have seen in a committed relationship," noted Paul Eastwick, UC Davis professor in the Department of Psychology and lead author of the study. "This supposed hookup melee actually looks a lot like people taking relationships for a test run."

The authors used machine-learning approaches to identify the strongest predictors of romantic interest in each crush. Machine learning, or the use of algorithms and statistical models to analyze and draw inferences from patterns in data, is especially useful at identifying predictors that are likely to be robust and replicable, the authors said.

Over the course of the study, some of the best predictors of sustained interest in a partner turned out to be markers of attachment, such as seeking out someone's presence as much as possible, feeling distressed when separated from them, and wanting to tell them about successes. These features are traditionally considered markers of pair-bonded relationships.

"When feelings of attachment and emotional connection start to kick in, young adults seem to take it as a sign that this is a crush worth pursuing," said study co-author Samantha Joel, an assistant professor of psychology at Western University. "Sexual and emotional attraction seem to go hand-in-hand, even before a committed relationship materializes."

Other factors that were known to be critical in initial impression contexts had no effect at all in the current study. Specifically, physical attractiveness—the most commonly studied variable in the whole initial attraction literature—was surprisingly weak.

Participants also uploaded photographs of their crushes, and the researchers used a team of coders, who didn't know the subjects of the photographs or anything about them, to rate how physically attractive the crushes were on a 1-10 scale. This variable turned out to be completely irrelevant to whether participants were romantically interested in the crushes.

"If we had been looking at a bar, or speed-dating—a setting where you have to compete to be noticed—these coder ratings of physical attractiveness should have been exceptionally good at predicting which partners were highly desired and which ones were not," Eastwick explained. "But that isn't what the data revealed at all."

According to Eastwick, these findings imply that early relationship development is a mating context in which people search for evidence of compatibility. "It isn't about fighting to get the 'most valuable' partner you can," he said. "It's about trying to find someone who inspires both a sexual and emotional connection. That's how young people initiate relationships."

The study, "Predicting Romantic Interest During Early Relationship Development: A Preregistered Investigation Using Machine Learning," was published in The European Journal of Personality. Co-authors were Joel; Kathleen Carswell, assistant professor, Durham University; Daniel C. Molden and Eli J. Finkel, professors, Northwestern University; and Shelley A. Blozis, professor of psychology, UC Davis. This study was designed by Molden and Finkel.