“I know what you did last Sunday,” claims the title of a new survey.

You skipped church. And then nearly one in seven of you fibbed about attending.

That’s according to a new survey by the Public Religion Research Institute released Saturday (May 17). The study, to be presented at the national meeting of the American Association for Public Opinion Research, was designed to measure the “social desirability bias in self-reported religious behavior.”

The survey finds that many Christians — and unbelievers, too — will exaggerate about attending worship in live phone interviews. However, when asked in an anonymous online questionnaire, people will answer more realistically.

On the phone, 36 percent of Americans report attending religious services weekly or more, while 30 percent say they seldom or never go.

But online, a smaller share (31 percent) of people surveyed said they attended church at least weekly, while a larger portion (43 percent) admitted they seldom or never go.

People who don’t attend worship — but say they did — may not mean to lie, said PRRI CEO Robert Jones.

People respond to phone surveys as they think “a good Christian” would or should answer, he said. “There’s an aspirational quality here,” he said. “People see themselves as the kind of person who would go.”

Once you remove the social pressure of speaking on the phone, “you see people willing to give answers that are probably closer to reality,” he said. “People feel less pressure to conform.”

Three groups were most likely to inflate attendance:

  • White mainline Protestants: By phone, 29 percent say they don’t go to church. Online, that jumps to 45 percent.
  • Catholics: On the phone, 15 percent. Online, 33 percent.
  • Adults ages 18-29: On the phone, 31 percent. Online, 49 percent.

The PRRI study is an update of studies on inflated church attendance conducted in the 1990s. In those studies research teams surveyed Catholics and Protestants in Ashtabula County, Ohio, and compared self-reported attendance claims with actual head  counts in scores of churches.

The result: “Actual church attendance was about half the rate indicated by national public opinion polls.”

Since there’s no way to do head counts of people not attending, PRRI found a contemporary technological approach — two different survey formats. Both surveys of American adults were conducted in 2013, with 2,002 people interviewed by cell and landline and a demographically comparable group of 2,317 who answered questions online.

People don’t even have to be religious to inflate claims of religiosity, PRRI found.

Those one in five Americans who are “nones” also may feel greater pressure to fib because “they are the farthest outside general social expectations,” said Jones.

On the phone, 73 percent of “nones” say they seldom or never attend, but 91 percent say so when interviewed online.

In the overall study, 19 percent of adults answering online said religion was not important to them; only 13 percent said so on the phone.

However, among the “nones,” the gap on the importance of religion was markedly wider — 49 percent on the phone, compared with 73 percent online.