It's a group that took statistics from Wikipedia articles of famous people to then run the headline "rock and rollers twice as likely to divorce." That headline is misleading. They didn't study rock and rollers, they studied celebrities.

This older survey similarly has them playing fast and loose with the statistics to get a flashy headline. In their own data it's clear that online dating has cannibalized dating at bars, and shares a near identical divorce rate - 20% vs 19%. In other words the survey is a nothingburger. All it shows is a shift in dating channels. But that's not the headline they pushed.

If their clickbait strategy isn't clear to you yet, don't take it from me. They claim this is their goal on their own website: "Time and again our research department has injected reality and hard evidence into this debate with eye catching research which the media have broadcast."