She not only shows that there is a correlation between broadband adoption and rising marriage rates, but uses "sharp temporal and geographic variation in the pattern of consumer broadband adoption" to shed light on the causal question. Broadband became available in different places at different times, in other words, so you can see that marriage rates spike first where broadband becomes available. The magnitude of the effect is estimated as leading to a roughly 13-30 percent increase in marriage rates among the benchmark population of young people.
I have a short thing about online dating in Wired's 20th Anniversary Issue—I take a personal interest in the subject because that's how I met my wife. So speaking from personal experience, I would say that this is an underrated benefit of recent technological innovation. It has become somewhat fashionable to dismiss the web and digital communication as not all that significant in economic terms. But if you dial back to 1993, you'd find that the U.S. was already a land of material abundance by any realistic standard. Innovations that have helped us build and maintain richer connections with other people are in fact extremely valuable relative to that baseline. We see more marriages, more people forming communities of interest, people keeping in closer touch with old friends, more engagement in politics and civil affairs, and all kinds of other good stuff. I hope more people will pursue Bellou's line of research here.
Correction, May 2, 2013: This post originally misspelled Brenda Cronin's last name.