Dating site matches not so scientific

Though succesful, online data sites can't claim to have truly scientific matching formulas, researchers say

Users flock to online dating sites in ever greater numbers, but despite their marketing claims, services such as Match.com and eHarmony may not be offering potential mates chosen through rigorous scientific methods, a group of psychologists and sociologists have charged.

"Companies have not made their algorithms [for matching potential mates] available to the public, nor even to regulatory authorities. Nobody knows what the algorithms are," said Harry Reis, a professor of psychology at the University of Rochester. "It is certainly possible they have some magic formula no one has looked at that could in fact be effective. However, there is no evidence for that."

Reis and other psychologists and sociologists have published a sweeping review of scientific studies that have been done over the past decade on online dating, which will be published in the February edition of the Psychological Science in the Public Interest journal. Commissioned by the Association for Psychological Science, the study reviews and summarizes more than 400 psychology studies and public interest surveys that have been done on this topic.

The sites themselves point to statistical successes of their services, even if they decline to reveal the proprietary matching algorithms that led to their success. Like other Internet services -- such as Google -- online dating sites jealously guard their algorithms as trade secrets. But such secrecy should prevent them from making claims of scientific accuracy, the researchers charge.

"Online dating is fast becoming one of the major ways in which people meet their partners. It is growing at a very rapid rate," Reis said. "People have always tried to set one another up. So the practices that many sites have are just modern versions of what is going on since time immemorial."

Over the past decade, online dating has become the second most popular way of meeting partners, surpassed only by meeting through friends, the researchers conclude. In the early 1990s, less than one percent of the population met through commercial dating services, including printed personal advertisements. Then, there was still a stigma attached to online dating, the authors note.

Today, that stigma seems to have largely disappeared. By 2005, 37 percent of American single adults had dated someone they connected with online. And by 2009, 22 percent of heterosexual couples and 61 percent of same-sex couples found their partners through the Web.

Online dating provides the convenience -- and fun -- of being able to peruse a list of potential mates, scanning dozens in a few minutes. But this approach has a number of limitations, researchers warn. People become conditioned to a shopping mentality, where they can just pick the desired features from a list.

"The range of choice can be a positive thing, by giving people many more options in a very time efficient way," Reis said. "But it can also encourage a mentality where one goes through a list of partners in the same way one might go through a list of books on Amazon. And often this approach is not helpful."

One aspect of online dating services that the researchers highlighted was how they often implied in their advertisements and promotional material that their ability to match partners is based on scientific algorithms. eHarmony's Web site, for instance, claims to match candidates using "29 dimensions of compatibility," pairing people for such factors as emotional energy, adaptability, romantic passion and other factors.

The researchers didn't single out eHarmony specifically in the study, "but it is certainly the site that makes the strongest claims to have a scientific basis for the algorithms that it uses," Reis said. "I presume that it does real science in developing its algorithm, but it has never made its work available to the scientific community, so nobody knows what's actually in it." In other words, eHarmony markets its service with the patina of scientific legitimacy, but has not gone through standard scientific peer review to be verified as scientifically valid.

In lieu of published research, sites may post research on their own sites, not disclosing study and data collection methods. "I liken this to a drug company putting out a drug, making a claim for it, and then telling no one what is in the drug," Reis said.

The problem with such claims is that people will assume that, through the workings of science, they will be finding the perfect partner, an attitude that can encourage an unrealistic and even destructive viewpoints in regard to relationships. When a relationship doesn't proceed exactly as hoped for, individuals may feel frustrated and insecure, the researchers assert.

Also the approach rests on the notion that a perfect partner can be found by identifying common or complimentary traits, an idea the researchers cast doubt upon.

"It is highly unlikely that what you can learn about two people before they have ever met can account for more than a trivial amount of what determines if a relationship will succeed over a long period of time," Reis said. "Relationship success over a long period depends on how two people interact with one another. It depends on what happens in their lives, the adversities and successes they have together, the way in which their lives mature and grow. These things are simply not knowable before they meet," Reis said.

When contacted, eHarmony declined to specifically respond to the study. "eHarmony's matching system is based on years of empirical and clinical research on married couples," said a company spokeswoman who did not want to be named. She did say that eHarmony's algorithms pair potential partners based on the aspects of personality, values and interest "that are most predictive of relationship satisfaction."

On average, 542 people who met on eHarmony get married each day in the U.S., according to a 2009 study conducted for eHarmony by Harris Interactive. The company has also posted numerous statistics on how and why relationships are successful, using its own data.

Other co-authors of the paper include psychology and sociology researchers from Northwestern University, Texas A&M University, University of California at Los Angeles and Illinois State University.

Joab Jackson covers enterprise software and general technology breaking news for The IDG News Service. Follow Joab on Twitter at @Joab_Jackson. Joab's e-mail address is Joab_Jackson@idg.com

What’s wrong? The new clean desk test
Join the discussion
Be the first to comment on this article. Our Commenting Policies