A
Primary Evidence
Direct RMBI scores derived from qualifying datasets or studies analyzed within the source framework.
RESEARCH & EVIDENCE
Follow the Result Back to the Evidence.
Explore the datasets, published studies, evidence grades, proxies, and methodological rules underlying the Religious Member Benefit Index.
NATIONAL DATASETSPUBLISHED RESEARCHTRACEABLE EVIDENCE
Evidence Standards
Every Score Shows Its Evidence Strength
RMBI keeps evidence strength visible. Direct evidence, published comparative evidence, documented proxies, and missing evidence are treated differently.
B
Published Comparative Evidence
Published comparative research directly supporting a group or category value, transformed to the RMBI scale when needed.
C
Credible Proxy
Broad-family, constituent-denomination, or benchmark evidence used only when direct evidence is unavailable and the crosswalk is documented.
N/A
No Qualifying Evidence
If no defensible observation exists, RMBI records N/A. Missing evidence is not converted to zero or filled with a neutral value.
Documented Evidence Sources
The Research Evidence Behind the RMBI
The RMBI evidence base includes 13 research study/data families: 6 core scoring studies and 7 supporting research families. The 2020 U.S. Religion Census is shown separately as an administrative dataset used for Institutional Reach and population/rollup checks. The current public comparison focuses on Christian communities and the religiously unaffiliated.
13Research families
6Core scoring studies
7Supporting research families
+1Administrative dataset
Evidence-grade note: A/B or B/C means the grade can differ by the specific variable, transformation, or way that source is used. The source itself is not being assigned two grades simultaneously.
NATIONAL SURVEY
A
CORE SCORING STUDY
Pew Religious Landscape Study 2014
Nationally representative religious-landscape survey used directly in RMBI scoring across multiple dimensions.
NATIONAL SURVEY
A
CORE SCORING STUDY
Pew Religious Landscape Study 2023–24
Updated nationally representative religious-landscape evidence used directly in RMBI scoring and comparison.
YOUTH STUDY
A
CORE SCORING STUDY
National Study of Youth & Religion — Wave 1
National youth study used directly for verified youth-outcome and related category evidence within the final scoring framework.
NATIONAL SURVEY
A
CORE SCORING STUDY
Cooperative Election Study 2024
Large national survey used directly for Religious Engagement measures, including religion importance, attendance, and prayer.
HOUSEHOLD PANEL
A/B
CORE SCORING STUDY
Philanthropy Panel Study 2001
Household giving study used directly for religious giving, total charitable giving, and giving relative to household income.
LONGITUDINAL DATA
B
CORE SCORING / ADJUSTMENT SOURCE
2023 PSID Charitable-Giving Analysis
Panel Study of Income Dynamics research used in the documented charitable-giving treatment for the religiously unaffiliated.
CONGREGATIONAL STUDY
B
SUPPORTING RESEARCH
National Congregations Study
Congregation-level evidence used to triangulate the availability and structure of member resources across religious communities.
CONGREGATIONAL STUDY
B
SUPPORTING RESEARCH
Faith Communities Today 2020
National congregational research used as supporting evidence when evaluating member-access resources and institutional offerings.
PUBLISHED RESEARCH
B
SUPPORTING RESEARCH
Pew Congregational Services Research
Published Pew research on congregational services and support used to corroborate the Member Resources framework.
LDS-SPECIFIC STUDY
B
SUPPORTING RESEARCH
UPenn / Cnaan — Called to Serve
Large Latter-day Saint study used to corroborate patterns in charitable giving, volunteering, and service.
YOUTH RESEARCH
B
SUPPORTING RESEARCH
SHARP Religion-Specific Youth Research
Religion-specific youth research used as corroborating evidence for youth outcomes. Multiple SHARP papers are treated as one study family.
NATIONAL RESEARCH
B
SUPPORTING RESEARCH
Pew Research Center — Religious Nones in America
Supporting evidence used to evaluate volunteering and civic-life outcomes among religiously unaffiliated Americans.
GSS-BASED RESEARCH
B
SUPPORTING RESEARCH
Taggart & Jensen — Religious Disaffiliation Research
GSS-based published research used as corroborating evidence for charitable behavior, volunteering, and religious disaffiliation.
ADMINISTRATIVE DATA
A
SEPARATE ADMINISTRATIVE DATASET
2020 U.S. Religion Census
Census-style adherent and congregation data used for Institutional Reach and for population/rollup checks. It is not counted among the 13 research study/data families.
Where the Evidence Feeds the Index
Seven Research Questions. One Comparative Framework.
The Research page documents the source evidence; Methodology explains how those sources are standardized and combined.
Transparent by Design
From a Score Back to Its Evidence
RMBI is designed so a reader can move backward from a result and understand how that result was constructed.
1
Score
Start with the overall or dimension-level result.
2
Dimension / Study
Identify the dimension and the qualifying study or source composite contributing to the result.
3
Variable & Raw Evidence
Trace standardized variables back to the observed percentage, dollar, rate, census measure, or documented coding input.
4
Evidence Strength & Limits
Interpret the result alongside evidence grade, missing-data treatment, and the methodological limits of comparative research.
Stronger Evidence Should Carry More Confidence
An A-grade result and a documented proxy should not be interpreted as equally strong evidence. RMBI keeps evidence strength visible so users can separate direct comparisons from crosswalks or aggregated proxy evidence. Unsupported missing observations are reported as N/A rather than imputed.
Likewise, comparative outcomes do not by themselves establish that religion caused the observed difference. Demographics, selection, geography, socioeconomic conditions, and other factors can contribute to measured differences between communities.
See What the Evidence Produces.
Move from the source framework into the final RMBI rankings and dimension-level comparisons.
EXPLORE THE RESULTS