Couples counselors have been around for decades. But with around 40% of US marriages now ending in divorce, they could probably do with some more modern techniques.
A new AI study led by Western University could provide them with some pointers. The researchers say it’s the first systematic attempt at using machine learning to predict relationship satisfaction.
“Satisfaction with romantic relationships has important implications for health, wellbeing, and work productivity,” said Western Psychology professor Samantha Joel in a statement. “But research on predictors of relationship quality is often limited in scope and scale, and carried out separately in individual laboratories.”
Joel and her colleagues collected information on 11,196 couples from 43 datasets. They then used a random forest classification algorithm to comb through the data for patterns that show what makes a successful relationship.
The AI found that the most reliable predictor of romantic success is a belief that your partner is fully committed to the relationship. The next most important factors were appreciation, sexual satisfaction, perceived-partner satisfaction, and conflict.
Notably, the individual characteristics of each partner — such as income and life satisfaction — were far less influential than relationship-specific factors.
“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. “But the surprising part is that once you have all the relationship-specific data in hand, the individual differences fade into the background.”
In other words, who you are as an individual is far less important than who you are when you’re with your partner.