Episode Summary
Executive Summary: The episode examines whether science can improve online dating by matching compatible partners, using Rose’s frustrating swiping experience as a case study. Research from psychologist Paul Eastwick suggests that personality quizzes and algorithms cannot reliably predict romantic chemistry, even though some traits correlate broadly. The takeaway: dating apps may help with access, but there’s no proven scientific shortcut to finding a soulmate.
Main Topics: The frustration of modern swiping (Priority: 5/5): Rose’s experience illustrates how much time people spend on dating apps for very few dates and even fewer real connections, highlighting the emotional and practical cost of endless swiping. Can algorithms predict compatibility? (Priority: 5/5): The episode tests the core claim of science-based dating apps: that surveys and machine learning can identify who will like whom. The answer from the research is largely no. Paul Eastwick’s speed-dating experiments (Priority: 5/5): Eastwick’s college speed-dating studies collected personality data and partner preferences, then tested whether they could predict mutual attraction; they could not. Limits of broad matching factors (Priority: 4/5): Shared age, education, and political views can increase the odds of attraction, but these broad similarities are too coarse to reliably generate strong matches. The unpredictability of romance (Priority: 4/5): The episode argues that relationship formation resembles an earthquake more than a forecastable weather pattern—something real, but not precisely predictable. Practical takeaways for dating app users (Priority: 3/5): Without trustworthy compatibility algorithms, users still need to navigate profiles, dates, and self-presentation strategically rather than relying on a perfect matching engine.
Key Arguments: Online dating can consume huge amounts of time while producing very few meaningful connections, making its inefficiency a real user pain point. A study of two million first messages found that fewer than 2% led to a phone number exchange, showing how low conversion can be on dating apps. Psychological questionnaires and machine learning models failed to predict which strangers would like each other after meeting. Even people’s stated ideal partner preferences were essentially useless for predicting real-world attraction. Some similarities like age, education, and politics matter, but only at a broad level, so they are insufficient for precise matching. Dating apps claiming scientific soulmate matching should provide evidence, because current research does not support those promises. Romantic connection contains too much unpredictability to be reduced to a stable algorithmic formula.
Data Points: Time spent swiping: About 20 hours per month - Rose estimates she spends half an hour to an hour, four to five days a week on apps. Time spent swiping annually: More than 100 hours a year - Rose calculates the cumulative annual time spent on dating apps. Average dates per month: About 3 dates - Rose describes how many dates she typically gets from her swiping. Meaningful chats from those dates: About 1 out of 3 - Rose says she is lucky if one date out of three leads to easy conversation. First messages leading to phone numbers: Less than 2% - Oxford study of 2 million first messages on dating apps. Participants in Eastwick’s early studies: 350 students - Paul Eastwick’s college-based speed-dating and questionnaire experiments. Date length in experiment: 4 minutes - Participants at Eastwick’s speed-dating event had brief four-minute conversations.
Pivotal Quotes: "We couldn't predict compatibility at all." — Paul Eastwick: Summarizing the results of his questionnaire and machine-learning analysis of speed-dating outcomes. "There's no existing algorithm that is going to predict compatibility." — Paul Eastwick: Direct response to whether any current dating algorithm can find the perfect person. "relationships are more like earthquakes than the weather." — Paul Eastwick: Used to explain why romantic connections are hard to forecast precisely.
Implications: Listeners should treat dating apps as access tools, not soulmate machines. For the industry, the burden is on companies to prove compatibility claims with real outcomes data, because current science does not support reliable perfect-match algorithms.
About Science Vs
There are a lot of fads, blogs and strong opinions, but then there’s SCIENCE. Science Vs is the show from Spotify Studios that finds out what’s fact, what’s not, and what’s somewhere in between. We do the hard work of sifting through all the science so you don't have to and cover everything from 5G and ADHD, to Fluoride and Fasting Diets.