Q & A With Eli Finkel – The Science Behind Online Dating (Part 2)

Eli Finkel, a social psychologist at Northwestern University, is one of five authors on a new study in Psychological Science in the Public Interest. The study, ‘Online Dating: A Critical Analysis from the Perspective of Psychological Science‘ is co-authored by Paul Eastwick of Texas A & M University, Benjamin Karney of UCLA, Harry Reis of the University of Rochester and Susan Sprecher of Illinois State University.

We invited our Facebook and Twitter followers to submit their questions on love, relationships and online dating to Finkel. Here is the second part of his response. Check out Part 1 of the Q & A here!

When people end up together after online dating, can people argue that it’s all just a coincidence or luck of the draw, or does the initial profile information make a difference?

This is a fascinating question, but, to date, an unanswered one. There’s certainly reason to believe that having a putatively scientific algorithm pick a match for you might make you more attracted to that match (as a result of placebo and expectancy confirmation effects, for example). My guess is that people who have faith in the algorithm will indeed experience greater attraction to an algorithm-selected match than they would have if that same person had been selected at random.

What is unknown to date is the degree to which that artificially elevated initial attraction is likely to yield long-term relationship well-being or disappointment. Research by marriage scholars such as Ben Karney (a PSPI co-author), Jim McNulty, and Lisa Neff suggests that the story is likely to be complex; that is, it’s unlikely to be a main-effects story.

What would you choose to put on a questionnaire if you were writing one?

I would scour the relationships literature to discern those factors that are known to be especially relevant to long-term relationship well-being. In particular, I would structure the dating site in a way that makes it possible for the algorithm not only to incorporate information about two people who are entirely unaware of each other’s existence, but also information about their interpersonal dynamics. One efficient means of doing this would be to build a site that resides more-or-less at the intersection of speed-dating, traditional online dating (like the early iterations of Match.com), and algorithmic-matching online dating (like eHarmony.com).

Such a site could have people browse profiles and set up 3-minute webcam-based “speed-dates” with 15 potential partners, completing a 1-minute questionnaire following each date (total time commitment: 60 minutes, assuming cooperative scheduling logistics). In addition, computer software (and, pending resource availability, perhaps human observers) would code each dyadic interaction for things like facial expressions, word usage, affective tone, etc. Next, all of this dyadic information would be incorporated into an algorithm that would use established relationship science to develop a compatibility index. Finally, the site would invest considerable resources to follow up with its users after they went on dates together, collecting data that would be crucial in refining the algorithm.

Such an online dating would be ambitious, but it’s far from impossible. And it could be both effective and lucrative.

After reading “So Far Away from One’s Partner, Yet so Close to Romantic Alternatives. . .” It is difficult to imagine why an avoidantly attached person would seek to enter into a relationship at all. I was wondering how an avoidantly attached person might approach the world of online dating. Are avoidantly attached persons more likely to try online dating because the initial “meeting” provides the “distance and detachment” they are more comfortable with?

Although I’m not aware of definitive data on this specific topic, it’s worth noting that avoidant people regularly enter romantic relationships. Once they’re in those relationships, however, they are especially likely to avoid deep emotional intimacy. My guess is that avoidant people might be no less likely than secure people to date online because the initial attraction phase isn’t especially threatening or challenging for them.

What socio/psychological effects do you anticipate if popularity of online dating increases for generations?

This is a deep question. As the Internet increasingly embeds itself in the DNA of daily life, online dating, in the broadest sense, is likely to become increasingly prevalent. But online dating seems likely to continue to evolve at a rapid clip. Over time, the process of meeting potential romantic partners through the Internet will increasingly approximate the process of meeting potential romantic partners face-to-face (webcam interactions, avatar-based interactions, and other procedures nobody’s dreamed of yet). My guess is that such computer-mediated communications will never achieve the richness of face-to-face interaction, but getting ever-closer is a great thing.

To the degree that online dating—whether through self-selection or algorithm-selection—increases the degree to which partners in a given relationship tend to be similar to each other, the emergence and increasing pervasiveness of online dating will fail to achieve its potential to introduce people to an increasingly diverse pool of potential partners. Harry Reis, a PSPI co-author, observes (a) that even though online dating enlarges the “pool of eligibles,” people might increasingly date others who are similar to them; and (b) that the consequences of this trend could be significant. Both self-selection and algorithm-selection pull for similarity in significant ways. For example, the various niche sites that are based on religion, occupations, hobbies, etc., are guaranteed to expose you only to people who are like oneself.  Even on less specialized sites, square dancers are likely to indicate interest in other square dancers and not in people who list square dancing as a deal-breaker. Most algorithms also seem to pull for similarity—if nothing else, it’s easy to program some variants of similarity, and most programmers don’t have the expertise in relationship science to look for matches based on more nuanced principles of interpersonal attraction or relationship well-being.

One of the great things about close relationships is the way they can expose you to new experiences. For example, Reis tried skiing because the woman he was dating (who later became his wife) loved skiing. It’s now one of his favorite things to do, though he never would have tried it by himself. Research on self-expansion theory (Aron & Aron, 1997) suggests that people can grow in profound ways from having a partner who is different from themselves, and the increasing prevalence of online dating might be reducing the likelihood and extent of such personal growth.

And the increasing emphasis on similarity that is likely fostered by online dating may have even more profound implications than is initially apparent. Reis discusses biometric genetic models, which “counter-intuitively indicate that assortative mating for a heritable trait will actually increase phenotypic variability, by increasing additive genetic variance, as first demonstrated by Sewall Wright in 1921” (Finkel et al., in press, p. 50 in the uncorrected proofs). “It follows,” he continues, “that if the extent of assortative mating is greater for couples created through matching sites than for traditional couples, then the phenotypic variability of human traits, as well as the genetically determined component of those traits, will increase.” The socio/psychological implications of this increase in phenotypic variability are unknown, but they could be significant.