Government Religious Preference 2.0 (GRP 2.0), Composite

Data Archive > International Surveys and Data > Cross-National > Summary

The Government Religious Preference dataset (GRP) measures government-level favoritism toward, and disfavor against, 30 religious denominations. The unit of observation is the state-year. A series of ordered categorical variables index the state’s institutional favoritism in 28 different ways. Those 28 variables are combined to form five composite indices for five broad components of state-religion: official status, religious education, financial support, regulatory burdens, and freedom of practice. The five components’ composites in turn are further combined into a single composite score, the GRP score. All of this is done for each of the 30 religious denominations covered in the dataset. The total number of data points is approximately 42 million, distributed in six data files.

Version 2.0 expands the dataset considerably from the original version. Version 2.0 draws from many more primary and secondary sources. It covers all independent states that are included in the Correlates of War state system. Its temporal coverage ranges from 2015 back to the year of the state’s independence or about 1800, whichever is later. The precise years of coverage for each state is detailed in the Codebook, Appendix A. Each variable-level data point is documented with a Coding Event number to enable users to identify the source used to populate the variable.

This dataset is the first of six files. It contains the Government Religious Preference Composite.

Data File
Cases: 18,240
Variables: 237
Weight Variable:
Non-missing values of composites G1, G2, G3, G4, G5 are weighted equally to form GRP composite.
Data Collection
Date Collected: October 2013 through April 2017
Original Survey (Instrument)
GRP 2.0 Codebook
Funded By
Partially funded by a Jack Shand Research Award of the Society for the Scientific Study of Religion, and multiple Work Study grants from Maryville University of St. Louis and University of Southern California.
Collection Procedures
The following section describes the first of the two phases of the dataset construction process. The first phase is more responsive to the solicitation of “collection procedures” and the second phase more responsive to “sampling procedures,” hence the two phases are presented in reverse order. Researchers methodically consulted the following primary and secondary sources:

Archivio delle Costituzioni Storiche. Accessed October 8, 2016.
Catholic Encyclopedia (1912)
Comparative Constitutions Project Dataset. Accessed March 1, 2016.
Concordat Watch. Accessed December 18, 2017.
Encyclopedia of the Modern Islamic World (1995)
Encyclopaedia Judaica, 2nd ed. (2007)
Fox, Jonathan. A World Survey of Religion and State (2008)
Fox, Jonathan. Political Secularism, Religion, and the State (2015)
Fox, Jonathan. The Unfree Exercise of Religion (2016)
Shaikh, Farzana. Islam and Islamic Groups (1992)
“Israel's Diplomatic Missions Abroad: Status of relations.”
. Accessed May 5, 2017.
U.S. State Department International Religious Freedom Reports (1999-2015)
Jewish Encyclopedia (1906)
New Catholic Encyclopedia (1967)
New Catholic Encyclopedia, 2nd ed. (2003)
O’Brien, David. Constitutional Law and Politics, 9th ed. (2014)
Religion and State Dataset, Round 2 (2012)
Statesman’s Yearbook (1866-2013)
U.S. Federal Research Division Country Study Series (c. 1960-2004)
Die Verfassungen in Deutschland (seit 1806). Accessed October 3, 2016.
World Christian Encyclopedia (1982)
World Christian Encyclopedia, 2nd ed. (2001)
World Constitutions Illustrated (2016). HeinOnline.

For each country covered in the dataset, researchers consulted each source listed above for information on the existence, origination, change, or discontinuation of a law or policy directed toward any (or all) of the following 30 religious denominations:
Christianity (all)
Western Christianity (Catholicism and Extended Protestantism combined)
Extended Protestantism (Protestantism, Anglicanism, Pentecostalism combined)
Orthodox Christianity
Christian Syncretic Religions
Islam (all)
Sunni Islam
Shia Islam
Ibadi Islam
Baha’i Faith
Buddhism (all)
Theravada Buddhism
Mahayana Buddhism
Vajrayana (Tibetan) Buddhism
Chinese Folk Religion
Buddhism Complex (Buddhism, Shintoism, Taoism, Confucianism combined)
Indigenous Ethnoreligion

For further details on the religious denominations and composites thereof, see the GRP2.0 Codebook, pages 6-8.

GRP 2.0 reports governmental laws and policies in 28 issue areas, each of which is assigned to an individual variable as follows:

G10: Establishment of Religion, General
G11: National Establishment of Religion, Detailed
G12: Religionist Law or Legal Systems
G13: Religious Qualifications for Public Officer
G14: Religious Qualifications for Citizenship
G15: Subnational Establishment of Religion
G16: Religious International Relations Policy (in GRP 2.0, provided only for Judaism)
G20: Public Religious Education, General
G21: National Public Religious Education, Detailed
G22: Subnational Public Religious Education
G30: Financial Support of Religion, General
G31: Discriminatory Financial Support of Religion
G32: Discriminatory Religious Tax Exemptions
G33: Levy of Religious Taxes or Tithes
G40: Religious Equality of Regulatory Burdens, General
G41: Equality of Regulatory Burdens of Religionist Persons
G42: Equality of Regulatory Burdens of Religionist-Owned Businesses
G43: Equality of Regulatory Burdens of Religious Institutions
G44: Assertion of Governmental Control over Religious Organizations
G45: Religiously-Based Registration Requirements
G46: Application of Religious Blasphemy Laws
G50: Free Exercise of Religion, General
G51: Religious Practice in Public/Private
G52: Religious Practice by Citizens/Non-Citizens
G53: Criminal Penalties for Religious Practice
G54: Government Tolerance of Social and Extra-Legal Sanctions for Religious Practice
G55: Regulation of Religious Proselytizing
G56: Regulation of Religious Conversion

Upon the discovery of the existence, origination, change, or discontinuation of a law or policy directed toward any religion (or Atheism) covered by one of the 28 variables listed above, the coder creates a “Coding Event” and codes the applicable variable according to the rubric reproduced in the GRP 2.0 Codebook, pages 13-33. Each Coding Event applies to a single country, a single variable, and a single value assigned to that variable. (The same source citation may generate multiple Coding Events).

Each Coding Event is assigned a unique ID number. The Coding Event specifies the country, variable, value assigned to that variable, and the religious denominations to which the Coding Event applies. It further specifies the years of the policy’s origination and discontinuation, if known, and the “specified year,” i.e. the year in which the source documents the policy to be in existence (according to a predetermined rubric reproduced in the GRP 2.0 Codebook, page 35). If the start- and end-years are not stated in the source, the coder assigned the years to which the Coding Event is “carried forward” or “carried back,” according to a predetermined rubric reproduced in the GRP 2.0 Codebook, pages 36-37 & Appendix C).

A random sample of Coding Events was selected for separate coder verification. With each “batch” of approximately 1000 Coding Events, the sample contained a minimum of 5% of all Coding Events for a country, 5% of all Coding Events for a source, and 5% of all Coding Events for a variable. (Thus the actual sample size varied but ranged between 7 and 9 percent of the batch.) For each Coding Event sampled, a separate coder independently verified the source, religions, specified year, variable number, and value assigned to the variable. The intercoder reliability rate exceeded 85 percent for every batch sampled, and exceeded 90 percent for all batches except one.
Sampling Procedures
The first phase of data construction is described below under “Collection Procedures.” The procedure described below is actually the second phase of data construction.

Once the Coding Event list was completed for all countries and sources and all Inter-coder Reliability tests completed, the Coding Events were sorted by country, variable, and source authoritativeness. The latter variable denotes six categories of authoritativeness, as follows:

1. Primary documents.
2. Reports and documents from governments.
3. Sources from well-reputed scholars.
4. Professionally authored encyclopedic sources.
5. Academic and professional datasets.
6. Encyclopedic and other sources of unknown authenticity.

For each country and each variable, country-specific codesheets were populated using the best combination of most specific and most authoritative Coding Events. The precise rules for Coding Event selection are detailed in the GRP 2.0 Codebook, pages 37-38 and Appendix D (Rule P).

To populate variables for individual religious denominations with greater efficiency, codesheet templates were programmed for data entered into “Nonpreferred Religion” columns to populate automatically into the columns for specific religious denominations. Coding Events pertaining to a Preferred religion, or to some other specific religion, were entered into codesheets manually. The construction of component-level variables, including highest and lowest value and composite score, was similarly automated.
Principal Investigators
Davis Brown, Nonresident Fellow, Baylor University Institute for Studies of Religion

Patrick James, Professor, University of Southern California
Related Publications
Brown, Davis. 2016. “The Influence of Religion on Interstate Armed Conflict: Government Religious Preference and First Use of Force, 1946-2002.” Journal for the Scientific Study of Religion 55(4): 800-820. [Based on the original version of the GRP dataset.]

An article of record for GRP 2.0 is currently undergoing the peer-reviewed journal submission process.
Information on indexes or other constructed variables
The 28 variables listed in section (d) above are collapsed into one composite variable for each of five components:

G1: Official Status (G10, G11, G12, G13, G14, G15, G16)
G2: Religious Education (G20, G21, G22)
G3: Financial Support (G30, G31, G32, G33)
G4: Regulatory Burdens (G40, G41, G42, G43, G44, G45, G46)
G5: Free Exercise (G50, G51, G52, G53, G54, G55, G56)

The five components just listed are further collapsed into a single Government Religious Preference composite score:

GRP: Government Religious Preference (G1, G2, G3, G4, G5)

This is accomplished for each of the 30 religious denominations listed in the Collection Procedures. For further detail on construction of composite scores, see the GRP 2.0 Codebook, pages 4-5.
Bookmark and Share