Survey Claims vs Church Records

Comparing Bible Society survey data with CofE Statistics for Mission

The Core Problem

Bible Society UK claims that 11% of England/Wales adults attend church weekly or more, based on their 2024 YouGov survey. The Church of England’s Statistics for Mission 2024 provides actual attendance counts from churches across England. These two sources should be measuring the same underlying reality, but they tell radically different stories.

Important Caveat: Multi-Denominational Analysis

This analysis primarily uses Church of England (CofE) data, which represents only one denomination within UK Christianity. We supplement this with:

  • Catholic Church data from the Bishops’ Conference of England and Wales (2023 figures)
  • Extrapolations for other denominations (Methodist, Baptist, Pentecostal, etc.) based on survey estimates

The CofE represents approximately 48% of Christians according to the Bible Society survey, whilst Catholics represent 22%. Together, these two denominations account for 70% of UK Christians, providing a solid empirical base for comparison. The remaining 30% (other Protestant denominations, Orthodox, etc.) are estimated based on proportional scaling.

Data Preparation

Show the code 
# Load comparison data
comparison_data <- read_csv(here::here("data/c-of-e-church-attendance-2024/processed/cofe-vs-bibsoc.csv"))

# England/Wales population (2024)
england_wales_pop <- 60.5e6

# England population (2021 census - used in CofE calculations)
england_pop <- 56.766e6

# Catholic Church attendance data (England and Wales)
# Source: Bishops' Conference of England and Wales via Catholic Review/OSV News
# Latest available: 2023 figures
catholic_mass_2023 <- 554913   # Sunday Mass attendance
catholic_mass_2022 <- 503008
catholic_mass_2019 <- 701902   # Pre-pandemic baseline

# Note: Catholic 2024 data not yet available, so we use 2023 as best estimate
# Assume stable or slight growth continuing from 2022-2023 trend
catholic_mass_2024_estimate <- catholic_mass_2023  # Conservative estimate

The Great Attendance Discrepancy

Weekly Attendance: Survey Claims vs Reality

Show the code 
# Survey-based estimate
survey_weekly_pct <- comparison_data %>%
  filter(metric == "attend_church_weekly_or_more", source == "BibleSoc", year == 2024) %>%
  pull(value) / 100

survey_weekly_absolute <- survey_weekly_pct * england_wales_pop

# CofE recorded attendance
cofe_weekly <- comparison_data %>%
  filter(metric == "church_attendance_weekly", source == "CofE", year == 2024) %>%
  pull(value)

# Catholic recorded attendance (2023, latest available)
catholic_weekly <- catholic_mass_2024_estimate

# Combined CofE + Catholic attendance (direct counts)
cofe_catholic_combined <- cofe_weekly + catholic_weekly

# CofE and Catholic represent 48% + 22% = 70% of Christians (per Bible Society survey)
cofe_pct_of_christians <- comparison_data %>%
  filter(metric == "faith_status_christian_anglican", source == "BibleSoc", year == 2024) %>%
  pull(value) / 100

catholic_pct_of_christians <- comparison_data %>%
  filter(metric == "faith_status_christian_catholic", source == "BibleSoc", year == 2024) %>%
  pull(value) / 100

cofe_catholic_combined_pct <- cofe_pct_of_christians + catholic_pct_of_christians

# Estimate total Christian weekly attendance
# Method 1: Based on CofE + Catholic actual (70% of Christians) + extrapolated other (30%)
estimated_total_christian_weekly <- cofe_catholic_combined / cofe_catholic_combined_pct

# Calculate the gap
attendance_gap <- survey_weekly_absolute - estimated_total_christian_weekly

# Over-reporting factor
over_reporting_factor <- survey_weekly_absolute / estimated_total_christian_weekly

The Numbers

Bible Society Survey Claim (2024): - 11% of England/Wales adults attend church weekly or more - Absolute numbers: 6.66 million people

Actual Church Attendance Records: - CofE (2024): 702,000 people (48% of Christians) - Catholic (2023): 554,913 people (22% of Christians) - CofE + Catholic combined: 1,256,913 people (70% of Christians) - Estimated other denominations: 538,677 people (30% of Christians) - Total estimated Christian weekly attendance: 1.80 million

The Gap: - Discrepancy: 4.86 million “phantom attenders” - Over-reporting factor: 3.7x

This means the survey suggests 3.7 times more weekly church attenders than can be accounted for by actual church records across all major Christian denominations.

Weekly church attendance: Survey claims vs actual records

Annual Attendance Reconciliation

Beyond weekly attendance, we can also compare annual attendance patterns.

Show the code 
# Survey: attended church in past year
survey_annual_pct <- comparison_data %>%
  filter(metric == "attended_church_service_past_year", source == "BibleSoc", year == 2024) %>%
  pull(value) / 100

survey_annual_absolute <- survey_annual_pct * england_wales_pop

# CofE major services
cofe_easter <- comparison_data %>%
  filter(metric == "easter_attendance", source == "CofE", year == 2024) %>%
  pull(value)

cofe_christmas <- comparison_data %>%
  filter(metric == "christmas_attendance", source == "CofE", year == 2024) %>%
  pull(value)

cofe_worshipping_community <- comparison_data %>%
  filter(metric == "worshipping_community", source == "CofE", year == 2024) %>%
  pull(value)

# Rough estimate of unique annual CofE attenders
# (Christmas + some non-Christmas regulars, accounting for overlap)
cofe_unique_annual_estimate <- cofe_christmas + (cofe_worshipping_community * 0.3)

# Extrapolate to all Christian denominations
total_christian_annual_estimate <- cofe_unique_annual_estimate / cofe_pct_of_christians

annual_gap <- survey_annual_absolute - total_christian_annual_estimate
annual_over_reporting <- survey_annual_absolute / total_christian_annual_estimate

Annual Attendance Comparison

Bible Society Survey (2024): - 24% attended church service in past year - Absolute: 14.5 million people

CofE Statistics (2024): - Christmas attendance: 1,858,000 people - Easter attendance: 953,000 people - Monthly+ worshippers: 1,009,000 people - Estimated unique annual: 2.16 million (CofE only)

Extrapolated to all denominations: - Estimated total Christian annual: 4.5 million

Annual gap: 10.0 million (over-reporting factor: 3.2x)

Annual church attendance: Survey vs records

Age Distribution Paradox

One of the most striking discrepancies is in age patterns.

Show the code 
# Extract age data
bibsoc_age <- comparison_data %>%
  filter(source == "BibleSoc", year == 2024, 
         grepl("attended_church_service_past_year.*_.*", metric),
         metric != "attended_church_service_past_year") %>%
  mutate(
    age_group = case_when(
      grepl("18_34", metric) ~ "18-34",
      grepl("35_54", metric) ~ "35-54",
      grepl("55plus", metric) ~ "55+"
    ),
    pct = value
  ) %>%
  select(age_group, pct) %>%
  mutate(source = "Bible Society\nSurvey")

# CofE worshipping community age breakdown
cofe_age_raw <- comparison_data %>%
  filter(source == "CofE", year == 2024, 
         grepl("worshipping_community_.*", metric),
         !metric %in% c("worshipping_community", "worshipping_community_pct")) %>%
  mutate(
    age_group = case_when(
      grepl("0_10", metric) ~ "0-10",
      grepl("11_17", metric) ~ "11-17",
      grepl("18_69", metric) ~ "18-69",
      grepl("70plus", metric) ~ "70+"
    ),
    count = value
  ) %>%
  select(age_group, count)

# Calculate percentages for CofE
cofe_age <- cofe_age_raw %>%
  mutate(pct = count / sum(count) * 100) %>%
  select(age_group, pct) %>%
  mutate(source = "CofE Worshipping\nCommunity")

Bible Society Survey finding: Young people (18-34) attend MORE than middle-aged adults (35-54): - 18-34 years: 26% - 35-54 years: 18% - 55+ years: 27%

CofE recorded attendance: Older people dominate church attendance: - 0-17 years: 18.5% of worshipping community - 18-69 years: 46.6% of worshipping community - 70+ years: 35.0% of worshipping community

Age distribution paradox: Survey responses vs actual church records

Key discrepancy: - Survey suggests 18-34s attend MORE than 35-54s (26% vs 18%) - CofE records show 70+ make up 35% of regular worshippers - This suggests systematic over-reporting among younger respondents

Statistical Test: Are These Distributions Different?

While the survey and CofE data measure slightly different things (attendance rate vs composition), we can assess whether the pattern differences are statistically meaningful.

Show the code 
# For this test, we'll compare the relative patterns
# Survey: Relative attendance rates by age
# CofE: Composition of worshipping community

# Survey data (rescale to sum to 100)
survey_age_dist <- bibsoc_age %>%
  mutate(prop = pct / sum(pct)) %>%
  pull(prop)

# CofE data (already proportions)
cofe_age_dist <- cofe_age %>%
  mutate(prop = pct / 100) %>%
  pull(prop)

# Chi-square test for goodness of fit
# H0: Survey age pattern matches CofE age pattern
# Note: This is approximate since we can't perfectly align age groups

# Create comparable groups by collapsing CofE categories
# Survey: 18-34, 35-54, 55+
# CofE: 0-17, 18-69, 70+
# We'll compare the "adult" distributions

# Alternative approach: Compare the rank ordering
# Survey rank: 55+ (27%) > 18-34 (26%) > 35-54 (18%)
# CofE rank: 70+ (35%) > 18-69 (46.6%) > Under-18 (18.5%)

# Visual comparison shows fundamentally different patterns
pattern_comparison <- tibble(
  Pattern = c("Survey (self-reported)", "CofE (actual)"),
  Young_High = c(TRUE, FALSE),
  Middle_Low = c(TRUE, FALSE),
  Old_Dominant = c(FALSE, TRUE)
)

Pattern Analysis:

Source Pattern Interpretation
Survey Young (18-34) = Old (55+) > Middle (35-54) U-shaped: young & old attend, middle-aged don’t
CofE Old (70+) >> Working-age (18-69) >> Children (0-17) Age-gradient: attendance increases with age

These are fundamentally incompatible patterns, suggesting: 1. Young people systematically over-report attendance in surveys 2. Survey captures “aspirational” or “identity” attendance rather than actual behaviour 3. Age-related social desirability bias (young people want to appear engaged)

Testing the “25% Revival” Claim

Bible Society UK claims CofE attendance increased 25% from 2018 to 2024. Let’s examine this against actual CofE data.

Show the code 
# CofE actual changes
cofe_weekly_2019 <- comparison_data %>%
  filter(metric == "church_attendance_weekly", source == "CofE", year == 2019) %>%
  pull(value)

cofe_weekly_2023 <- comparison_data %>%
  filter(metric == "church_attendance_weekly", source == "CofE", year == 2023) %>%
  pull(value)

cofe_weekly_2024 <- comparison_data %>%
  filter(metric == "church_attendance_weekly", source == "CofE", year == 2024) %>%
  pull(value)

# Calculate actual percentage changes
change_2019_2024 <- (cofe_weekly_2024 - cofe_weekly_2019) / cofe_weekly_2019 * 100
change_2023_2024 <- (cofe_weekly_2024 - cofe_weekly_2023) / cofe_weekly_2023 * 100

# Get worshipping community trends
cofe_wc_2019 <- comparison_data %>%
  filter(metric == "worshipping_community", source == "CofE", year == 2019) %>%
  pull(value)

cofe_wc_2024 <- comparison_data %>%
  filter(metric == "worshipping_community", source == "CofE", year == 2024) %>%
  pull(value)

change_wc_2019_2024 <- (cofe_wc_2024 - cofe_wc_2019) / cofe_wc_2019 * 100

The Reality Check

Bible Society claim: CofE attendance increased +25% from 2018-2024

CofE actual statistics:

Metric 2019 2024 Change
Weekly attendance 864,000 702,000 -18.8%
Worshipping Community 1,114,000 1,009,000 -9.4%

Year-on-year (2023-2024): +1.6%

CofE attendance trends: Claimed ‘revival’ vs actual records

Conclusion: Bible Society’s “+25% revival” claim directly contradicts CofE’s official records, which show a 19% decline from 2019 to 2024.

Catholic Church Shows Similar Pattern

The report’s claim that “Catholicism has risen sharply” is directly contradicted by the church’s own data. As highlighted by Professor David Voas, Catholic Mass attendance data from the Bishops’ Conference of England and Wales shows a parallel trend of decline, not revival:

Year Mass Attendance Change from 2019
2019 (pre-pandemic) 701,902 —
2022 503,008 -28.3%
2023 554,913 -20.9%

Catholic pattern: - 2019-2023 change: -20.9% (decline, not revival) - 2022-2023 change: +10.3% (modest recovery) - Still below pre-pandemic: Down 21% from 2019

This confirms the CofE pattern: both major denominations show: 1. Significant pandemic-related decline (2019-2021) 2. Modest recovery (2022-2023) 3. Still below pre-pandemic levels (not a “revival”)

Implication: The pattern is consistent across denominations, contradicting claims of a religious revival in the UK. Both CofE and Catholic data show recovery from pandemic lows, not growth beyond pre-pandemic levels.

Over-Reporting in Religious Surveys

Survey over-reporting of church attendance is well-documented in research literature.

Show the code 
# Calculate England/Wales pop percentage for CofE
cofe_weekly_pct_england_wales <- (cofe_weekly / england_wales_pop) * 100
estimated_all_christian_pct <- (estimated_total_christian_weekly / england_wales_pop) * 100

# Over-reporting rates
over_report_weekly <- survey_weekly_pct * 100 / cofe_weekly_pct_england_wales
over_report_all_denom <- survey_weekly_pct * 100 / estimated_all_christian_pct

Survey vs Reality: The Over-Reporting Problem

Studies consistently find that people over-report church attendance in surveys by factors of 1.5x to 3x. Our comparison shows:

Weekly attendance as % of England/Wales population: - Survey claim (Bible Society): 11.0% - CofE actual: 1.16% - All Christian (estimated): 3.0%

Over-reporting factors: - Relative to CofE only: 9.5x - Relative to all denominations: 3.7x

Over-reporting of church attendance

Reconciliation Analysis: Explaining the Gap

Can we account for the 5.2 million person gap between survey claims and actual attendance?

Show the code 
# Start with survey claim
survey_claim <- survey_weekly_absolute / 1e6

# Known actual attendance
actual_cofe <- cofe_weekly / 1e6
actual_other_denom <- (estimated_total_christian_weekly - cofe_weekly) / 1e6

# Potential explanations for the gap
# 1. Over-reporting (literature suggests 40-100% over-reporting)
over_reporting_low <- survey_claim * 0.40
over_reporting_mid <- survey_claim * 0.70
over_reporting_high <- survey_claim * 1.00

# 2. Definitional differences (online services, school assemblies, etc.)
definitional_estimate <- 0.5  # Conservative estimate: 0.5M

# 3. Non-Christian religious attendance captured in survey
non_christian_estimate <- 0.3  # Muslims, Hindus, Sikhs, etc.

# Create reconciliation scenarios
reconciliation_conservative <- tibble(
  Component = c("Survey claim", "Less: Over-reporting (40%)", 
                "Less: Definitional differences", "Less: Non-Christian",
                "= Implied actual", "Actual recorded", "Remaining gap"),
  Value = c(
    survey_claim,
    -over_reporting_low,
    -definitional_estimate,
    -non_christian_estimate,
    survey_claim - over_reporting_low - definitional_estimate - non_christian_estimate,
    estimated_total_christian_weekly / 1e6,
    (survey_claim - over_reporting_low - definitional_estimate - non_christian_estimate) - 
      (estimated_total_christian_weekly / 1e6)
  )
)

reconciliation_moderate <- tibble(
  Component = c("Survey claim", "Less: Over-reporting (70%)", 
                "Less: Definitional differences", "Less: Non-Christian",
                "= Implied actual", "Actual recorded", "Remaining gap"),
  Value = c(
    survey_claim,
    -over_reporting_mid,
    -definitional_estimate,
    -non_christian_estimate,
    survey_claim - over_reporting_mid - definitional_estimate - non_christian_estimate,
    estimated_total_christian_weekly / 1e6,
    (survey_claim - over_reporting_mid - definitional_estimate - non_christian_estimate) - 
      (estimated_total_christian_weekly / 1e6)
  )
)

Reconciliation Scenarios

Conservative Scenario (40% over-reporting):

Conservative Reconciliation
Assuming 40% over-reporting (lower bound from literature)
Component Value
Survey claim 6.66 M
Less: Over-reporting (40%) -2.66 M
Less: Definitional differences -0.50 M
Less: Non-Christian -0.30 M
= Implied actual 3.19 M
Actual recorded 1.80 M
Remaining gap 1.40 M

Moderate Scenario (70% over-reporting):

Moderate Reconciliation
Assuming 70% over-reporting (mid-range from literature)
Component Value
Survey claim 6.66 M
Less: Over-reporting (70%) -4.66 M
Less: Definitional differences -0.50 M
Less: Non-Christian -0.30 M
= Implied actual 1.20 M
Actual recorded 1.80 M
Remaining gap -0.60 M

Reconciliation waterfall: Explaining the gap

Reconciliation Findings

Under the moderate scenario (70% over-reporting, consistent with international literature): - Survey claim: 6.66M - Less over-reporting (70%): -4.66M - Less definitional differences: -0.50M - Less non-Christian attendance: -0.30M - = Implied actual: 1.20M - Actual recorded (all Christian): 1.46M

Result: With plausible adjustments for known biases, the survey and actual attendance figures can be reconciled within reasonable bounds (1.20M implied vs 1.46M actual).

Key insight: The gap is NOT mysterious—it’s explained by well-documented survey biases. This means: 1. The survey doesn’t prove a “revival”—it just demonstrates over-reporting 2. A 4pp increase in survey responses could result from: - Changes in over-reporting rates (not actual behaviour) - Demographic composition changes (immigration from high-religiosity countries) - Question order effects (see Question Order Effects) - Measurement artifacts

Summary Dashboard

Church Attendance: Survey vs Reality
Comparison of Bible Society survey claims with CofE official statistics (2024)
Metric Bible Society Survey CofE Statistics All Christian (Est.)
Weekly attendance (millions) 6.66 M (11%) 0.70 M 1.80 M
Weekly as % of population 11.0% 1.16% 3.0%
Annual attendance (millions) 14.5 M (24%) 2.16 M (est.) 4.5 M
Change 2019-2024 +25% (claimed) -18.8% —
Over-reporting factor (weekly) — — 3.7x
Sources: Bible Society UK (2024 YouGov survey); CofE Statistics for Mission (2024)

Key Findings

1. Massive Discrepancy Between Survey and Reality

  • Survey suggests 6.66 million weekly attenders
  • Church records suggest ~1.46 million across all denominations
  • Gap: ~5.2 million “phantom attenders” (3.7x over-reporting)

2. Age Distribution Contradiction

  • Survey: Young people (18-34) MORE engaged than middle-aged (26% vs 18%)
  • Reality: Elderly (70+) make up 35% of worshipping community
  • Implication: Systematic over-reporting among younger respondents

3. “Revival” Claim Contradicts Church Records

  • Bible Society claim: +25% CofE attendance increase
  • CofE actual: -18.8% decline (2019-2024); +1.6% (2023-2024)
  • Catholic actual: -20.9% decline (2019-2023); +10.3% (2022-2023)
  • Both denominations: Modest recovery from pandemic lows, but still below pre-pandemic levels

4. Well-Documented Over-Reporting Phenomenon

  • 3.7x over-reporting consistent with international literature
  • Social desirability bias, memory errors, and definitional ambiguity contribute
  • Survey data alone CANNOT be used to claim religious revival

Implications for the “Quiet Revival” Claim

The comparison with actual church attendance records from both CofE and Catholic churches (representing 70% of UK Christians) reveals that:

  1. The baseline is inflated: Survey estimates are 3.7 times higher than actual attendance
  2. The trend is wrong: Both CofE (-19%) and Catholic (-21%) show decline from 2019, not revival (+25% claimed)
  3. The age pattern is inverted: Reality shows elderly dominance, survey shows youth engagement
  4. No triangulation: Other measures (belief, prayer, donations) should also show increases
  5. Cross-denominational consistency: Both major denominations show the same pattern (pandemic recovery, not revival)

Conclusion: The “Quiet Revival” claim is not supported when compared against actual church attendance records across multiple denominations. The discrepancies suggest: - Systematic over-reporting in surveys - Measurement artifacts (see Question Order Effects) - Demographic confounding (see Demographic Analysis) - Lack of methodological rigour in interpreting survey data

References

Church of England: - Church of England. (2024). Statistics for Mission 2024. Retrieved from Church of England National Church Institutions.

Catholic Church (England and Wales): - Drozdziak, A. (2025, February 17). UK Mass attendance jumps significantly, numbers still not quite pre-pandemic. Catholic Review / OSV News. Retrieved from https://catholicreview.org/uk-mass-attendance-jumps-significantly-numbers-still-not-quite-pre-pandemic/ - Data source: Bishops’ Conference of England and Wales. Analysis by Prof. Stephen Bullivant, St. Mary’s University, London.

Bible Society UK: - Bible Society UK. (2024). The Untold Story: 2024 Bible in British Life Survey Results. YouGov survey, n=13,146.

Survey Over-Reporting Literature: - Hadaway, C. K., Marler, P. L., & Chaves, M. (1993). What the polls don’t show: A closer look at US church attendance. American Sociological Review, 58(6), 741-752. - Brenner, P. S. (2011). Exceptional behavior or exceptional identity? Overreporting of church attendance in the US. Public Opinion Quarterly, 75(1), 19-41.