Science of Attraction, Compatibility & Romance | Dr. Paul Eastwick

Eastwick uses the movie 'Say Anything' as a model: John Cusack and Ione Skye go to a party for their first date, float through different groups, talk to others about how the date is going, and observe each other's behavior. The key moment where Skye starts to fall for Cusack is watching him take car keys away from someone who shouldn't drive — she sees his protective, responsible side in action. Eastwick contrasts this with app-style dating, which is essentially a job interview where people display traits. Real long-term compatibility involves the '4,000 daily responsibilities' of life together — setting out coffee, mowing the lawn, navigating stress — that are very hard to forecast from early one-on-one conversations. Watching someone in varied contexts provides much richer data.

Eastwick explains that the marketplace metaphor breaks down in happy relationships because people actually become 'bad barometers of what is good.' A partner who is happy in their relationship will perceive alternative partners as less desirable than any objective metric would indicate. This is a protective mechanism. Simultaneously, recent studies show a parallel effect: sexual fantasy about an alternative can temporarily boost desire for the partner (the 'rebound'). But Eastwick strongly cautions that the problem is never the initial attraction — it's the escalation: repeated hanging out, secretive conversations, emotional intimacy with the alternative. Social media dramatically increases the risk of this escalation because fleeting connections that would have ended in the past now have permanent follow-up vectors.

Eastwick notes that breakups are hard not just because of lost support but because you lose continuity with yourself — the stories and narratives built with the other person. He encourages viewing past relationships (including divorces) as things that 'were real, mattered in that moment' rather than failures. Huberman shares his own rule with his partner: they don't speak negatively about anyone they've previously been with, and they practice 'selective ignorance' — deliberately not probing into each other's pasts. They both express gratitude for the experiences that shaped their current partner. Eastwick endorses this, noting that a lot of cultural feedback treats breakups as failures ('I'm so sorry,' 'What happened?') rather than normal life transitions.

Eastwick's team partnered with a matchmaking service where men in the pool were on average 4 years older than women. When they tracked actual second-date interest after face-to-face meetings, both men and women showed a slight preference for younger partners. Women who explicitly said 'don't set me up with younger guys' would go on those dates and report enjoying them. The 4-year average age gap persists in the couples that form, but the mechanism is not operating at the initial-attraction or first-date level — suggesting it may result from who enters the pool in the first place or later-stage filtering, not from differential in-person desire.

Eastwick explains that 'perceived similarity' is a free-floating sense — when people like someone, they find the similarities and elevate their importance; when they don't, they dismiss mismatches as irrelevant. This motivated reasoning means that objectively measured similarity (e.g., an 83% match on all assessable traits) performs no better than a coin flip at predicting whether two people will click. The only possible exception is demographic/SES similarity, where there may be very small effects (53-47 rather than 50-50). This is why the apps feel useless: they're optimizing on something that doesn't predict attraction when people actually meet. The couples in the happiest relationships are the ones who exhibit these biased similarity perceptions most strongly — and Eastwick says it's 'stupid human tricks, but it kind of works.'