Top 5 factors that make or break a relationship, according to AI

Does it matter more who you love or how you love? A large machine learning study analyzed data from thousands of couples to identify which characteristics are most vital to predicting the success or failure of a relationship. Using artificial intelligence, researchers found that the individual traits of the partners had less to do with making the couple happy than the characteristics of the relationship itself.

To put it another way – the dynamic of the relationship you create, with its shared experiences and in-jokes, is more important than the specific traits of the one you are with.

Psychologist Samantha Joel from Western University in Canada led the research of 11,000 couples and 43 separate self-reported datasets – the first-ever systematic research into people’s love life using AI. The couples came from the U.S., Canada, Israel, the Netherlands, Switzerland and New Zealand.

Joel put their findings in context:

“Relationships-specific variables were about two to three times as predictive as individual differences, which I think would fit many people’s intuitions,” said Joel, adding “but the surprising part is that once you have all the relationship-specific data in hand, the individual differences fade into the background.”

According to Joel, feeling good about your relationship can have profound impacts on your health and well-being, as well as productivity at work. Their team identified five predictors of relationship success:

1. Perceived Partner Commitment – the belief in your partner’s full level of commitment is the most reliable predictor of the health of the relationship.

2. Intimacy / appreciation – feeling close to or being appreciated by your partner.

3. Sexual satisfaction – how satisfied the partners are with the quality of their sex life.

4. Perceived Partner Satisfaction – how happy one thinks their partners feels about the relationship.

5. Conflict – how often the partners fight.

Credit: Western University.

If you notice, love figures on this list, being nearly as important as the last two items, but is not in the top position.

The most important individual differences that affect relationships were found to be satisfaction with life, negativity, depression, attachment anxiety (worrying about the relationship), and attachment avoidance (not getting too attached). Interestingly, all of these accounted for just about 21 percent of variance in the other partner’s relationship satisfaction.

Besides Joel, the study involved Paul Eastwick from University of California, Davis, and 84 other international scholars. “Who I am’ doesn’t really matter once I know ‘who I am when I am with you,” Eastwick noted.

You can read their study, which utilized the machine learning system Random Forests, published in the Proceedings of the National Academy of Sciences.