In fact, eHarmony couples’ divorce rate was under half that of other online dating websites. In turn, couples who met in any online dating venue were slightly likelier to stay together than those who met offline, even if the offline meeting happened at church.
The online dating couples were happiest, too, with eHarmony’s leading—again, slightly.
But does statistical significance imply social relevance? On a 7-point scale, for all measures of love, happiness, trust, chemistry, and compatibility, eHarmony’s married couples scored up to three-tenths of a point higher than couples who met any other way, and folks who met via online dating were slightly happier than those who met otherwise. In a data set of 19,000-plus respondents like this one, even tiny differences will be statistically significant. I asked Cacioppo, did he think these were large differences in satisfaction?
He responded that how you meet your mate is just one of many factors in how happy you’re going to be, so we wouldn’t expect it to have a large effect. But “Whether [the results] are practically important is in the eye of the beholder. If you’re someone who wants to maximize your chances of a happy marriage, then the statistical differences become important whether or not the effect sizes are small.”
So you might also pay attention to the small, single, yet statistically significant finding that eHarmony couples were also the likeliest to say they were “extremely unhappy” with their marriages. No matter how couples met, under 8% of respondents said they were extremely unhappy– but eHarmony topped the list according to their own statistical analyses (see page 7, top).
Mama was right. There’s no pleasing everyone.
Do you wonder how these results can be valid, since eHarmony came out on top in every desirable category and paid for the project? Me, too—even though I met my own sweet husband six years ago on eHarmony. In reply, Warren and Cacioppo underscored that the research was needed, the public funding was non-existent, and so eHarmony decided to step into the breach to further, not bias, the science. Towards that end, Cacioppo obtained a raft of guarantees from Warren beforehand: Results would be published regardless of how they reflected on eHarmony; Cacioppo had complete autonomy; independent statisticians from Harvard planned and executed the analyses; and the entire data set is online even now, so anyone can crunch the numbers and see whether the results hold. A peer-reviewed publication known for integrity,PNAS, published the study, and you can read it here:http://www.pnas.org/content/early/2013/05/31/1222447110.full.pdf+html.
Warren was not surprised that eHarmony, which distinguishes itself through introducing similar people, came out on top. “If you can put people together based on similarity, you have so much better a chance of the marriage succeeding.”
What do I think? It’s just one study covering just seven years. It’s not experimental—meaning, we’ve got no cause-effect. But having read the PNAS paper carefully, I think this is solid research.
And online dating is here to stay…hopefully For Better.
What do you think?