Can An Algorithm Really Predict Love?

eHarmony promises to match singles with potential dates who are "prescreened for deep compatibility with you across 29 dimensions."

But what does that actually mean? How scientific are the algorithms that so many online dating dates claim can predict compatibility? Is a mathematical formula really capable of finding lasting love?

If you ask Eli J. Finkel and Benjamin R. Karney, psychologists and authors of a recent opinion piece on NYTimes.com, the answer is "no."

"It's hard to be certain, since the sites have not disclosed their algorithms," write Finkel and Karney, but "the past 80 years of scientific research about what makes people romantically compatible suggests that such sites are unlikely to do what they claim to do." Dating sites simply fail to collect adequate amounts of important information about their members, they say, and because what data they do gather is based on singles who have never met in person, dating sites are unable to predict how compatible two people will be when they actually do interact face-to-face.

The most telling signs of whether or not a relationship will succeed occur only after a couple has met - like communication patterns, problem-solving tendencies and sexual compatibility - and gotten to know each other. Those factors can't possibly be evaluated by an algorithm.

Dating sites also don't take into account the environment surrounding a potential relationship. Crucial factors like job loss, financial strain, infertility, and illness are completely ignored, despite the large influence they have on long-term compatibility. The information collected by online dating sites focuses instead on personal characteristics, which aren't negligible but only account for a small portion of what makes two people well suited for each other.

There's no doubt that "partners who are more similar to each other in certain ways will experience greater relationship satisfaction and stability relative to partners who are less similar," but online dating algorithms do not address those deep forms of similarity.

"Perhaps as a result," Finkel and Karney theorize, "these sites tend to emphasize similarity on psychological variables like personality (e.g., matching extroverts with extroverts and introverts with introverts) and attitudes (e.g., matching people who prefer Judd Apatow's movies to Woody Allen's with people who feel the same way)," forms of similarity that don't actually predict compatibility in a long-term relationship.

Online dating, the researchers conclude, is not any worse a method of meeting your match, but it also isn't any better than traditional methods. Choose your dates wisely, and don't choose your dating sites based on the promises of a magical algorithm.