This is from the article you linked to:
The findings are not as straightforward as saying “conservative Protestants are causing trouble for other people’s marriages,” said Charles Stokes, a sociology professor at Samford University, who conducted a separate study on Americans who, on average, got married at a younger age.
In his own research, Stokes found that conservative Protestants who attend church regularly are significantly less likely to have gotten divorced than nonreligious peers.
“The pattern that pops out in this data is that when you look at those who attend church weekly, their divorce rates are the same as other high-attending Christians,” Stokes said. “Nominal Christians are probably getting the community norms but aren’t in a social structure to live the norms out.”
So, once again, the message is that nominal evangelicals have high divorce rates. Practicing evangelicals are less likely to divorce.
To summarize: In areas with high numbers of evangelicals, non-religious people adhere to cultural norms surrounding marriage (such as marrying early) but are not part of church groups that offer them support. This contributes to high levels of divorce in predominantly evangelical regions of the country, even though among regular church-going evangelicals divorce rates are lower.