The controversy surrounding the now-withdrawn Quiet Revival report should serve as a reminder that eye-catching statistics require careful scrutiny rather than immediate acceptance, two of Britain’s leading social researchers have said.
Speaking at the Religion Media Centre Festival 2026 in Westminster, Professor Sir John Curtice, senior research fellow at the National Centre for Social Research, and Professor Bobby Duffy, director of the Policy Institute at King’s College London, reflected on the lessons from the controversy.
The report, published by Bible Society in April 2025, attracted widespread attention after suggesting church attendance each month in England and Wales had increased significantly between 2018 and 2024, with participation among Generation Z supposedly rising four-fold.
It was later retracted after polling company YouGov discovered problems with its methodology that rendered the data unreliable.
During a panel chaired by journalist Tim Wyatt, the researchers said it had highlighted the importance of interpreting religious statistics within the context of long-term evidence rather than relying on a single survey.
Professor Duffy said he was sceptical of the findings from the outset because they conflicted with decades of established research.
“We could see nothing in our data … that looked like that,” he said, referring to the long-running World Values Survey.
While acknowledging that sudden changes are always possible, he said: “Shifts in religious belief and religious practice is quite generational because it's quite imbued in people at a young age. It tends to shift in large proportion due to generational replacement. Some people flow in and out of religion, but the big moves are generational. So, you don’t tend to see those quick turnarounds in people's behaviours or beliefs.”
Professor Curtice said the appeal of the original story reflected journalism’s natural attraction to unexpected developments.
“The exceptional gets noticed,” he observed, whereas the long-term decline in Christian observance is often regarded as less newsworthy.
Drawing on data from the long-running British Social Attitudes survey, Professor Curtice said the evidence instead points to church attendance recovering somewhat after Covid-19 but remaining below pre-pandemic levels.
“The pandemic has had a deleterious impact on both Christian and non-Christian observance,” he said.
According to the survey, around 5% of adults now attend a Christian service weekly, compared with 8% prior to Covid-19.
Monthly attendance has also fallen from around 12% before the pandemic to about 9% today.
Among those who identify as Christian, the proportion attending at least monthly now stands at around 14%, after falling to 11% during the pandemic from approximately 16% to 17% previously.
He added that reduced attendance has not been confined to Christianity. Other faith communities have also seen declines in religious observance since the pandemic, with attendance among Muslim and other non-Christian groups falling from close to 40% to around 25%.
Although Christianity continues to experience a long-term process of secularisation, Professor Curtice noted that Britain’s changing population means overall patterns of religious practice are becoming more complex.
Younger age groups now include larger numbers of people from religious minority communities, particularly Muslims, whose levels of observance remain comparatively higher.
As a result, he said: “The overall link between religious observance and age is now flatter than it once was.”
Professor Curtice also said churches may appear younger today partly because older worshippers were less likely to return after breaking long-established attendance habits during the pandemic, rather than because large numbers of young people had begun attending.
Christian identification, meanwhile, has stabilised at around 40%, he said.
Younger people who describe themselves as Christian are more likely to practise their faith than previous generations, suggesting that “it’s becoming less of a social symbol and more indeed a religious practice”.
Professor Duffy said belief in God has steadily declined since the early 1980s, falling from around three-quarters of the population to roughly half today.
Belief in heaven has followed a similar trajectory.
By contrast, belief in life after death has remained remarkably consistent over several decades, with around 4 in 10 people continuing to hold that belief despite broader changes in religious affiliation.
Asked whether there is evidence of a rise in spirituality among people who identify with no religion, Professor Curtice pointed to previous academic research that found little support for the idea of a large movement towards being “spiritual without being religious”.
The researchers also discussed why the ‘Quiet Revival’ figures proved so misleading.
Professor Duffy said surveys measuring relatively uncommon behaviours can be especially vulnerable to small errors, which become “magnified” in the final results.
He also warned against what he described as “motivated reasoning” - the temptation to accept surprising findings because they fit a compelling narrative.
“If something surprises you that much, it’s probably wrong,” he said, recalling advice from his former polling mentor, Bobby Worcester.
Professor Curtice explained that online surveys face an ongoing challenge in verifying that respondents are genuinely who they claim to be.
While companies such as YouGov invest heavily in fraud detection, he said people attempting to complete surveys for financial reward may falsely claim to belong to unusual demographic groups, distorting the results.
“Being a 21-year-old white man who attends church once a week is a pretty good way of making yourself a rare being,” he said.
He added that the original study had attempted to establish a trend using only two surveys, warning that “you cannot make a trend out of two time points.”
Even so, both speakers cautioned against dismissing online polling altogether.
Professor Duffy said online panels remain an important tool for understanding public opinion quickly, particularly when high-quality probability surveys can take years to complete.
The latest World Values Survey, for example, is being carried out in 2026 but is not expected to produce results until early 2027.
Instead, he argued, the priority should be on strengthening safeguards against increasingly sophisticated fraud, including the growing use of artificial intelligence to generate fake survey responses.
He said AI had altered the financial incentives behind survey fraud because automated agents could generate large numbers of responses at very little cost.
“The sector is taking that incredibly seriously,” he said, though he acknowledged that AI continues to make the problem more difficult.
Professor Curtice said all surveys - whether using traditional face-to-face interviews or modern online surveys - should be approached with “a degree of scepticism” as every research method has limitations.
Declining response rates and the rising cost of in-person interviews mean even probability-based studies have increasingly moved online, making the way participants are recruited especially important in ensuring reliable results.
