International Religious Freedom Data, Aggregate File (2003-2008)
CitationGrim, B. J., Harris, J. D., Martin, R. R., & Finke, R. (2019, December 19). International Religious Freedom Data, Aggregate File (2003-2008)
SummaryThis file contains aggregate measures from the ARDA's coding of the 2003, 2005, and 2008 U.S. State Department's International Religious Freedom Reports. This coding produced data on 199 countries and territories (see below for list of countries coded), but excluded the United States. It also includes three indexes calculated from these data: the Government Regulation of Religion Index (GRI), the Government Favoritism of Religion Index (GFI), and the Modified Social Regulation of Religion Index (MSRI) [see Grim and Finke (2006) for more information on the GRI and GFI, and see below for more information on the MSRI]. Data in this file represent mean coding responses for three of each variable from all three years of coding unless otherwise noted. Many countries have scores on variables that are expressed as decimals, and which do not correspond with a value label in the variables' descriptions. These decimal values signify that a country's scores on these variables vary over the 2003, 2005 and 2008 Reports.
The ARDA has added three additional variables to the original data set to enhance the users' experience on our site.
Data FileCases: 199
Weight Variable: None
Data Collection2003, 2005 and 2008 aggregate measures compiled in September 2012
Funded ByThe John Templeton Foundation
Collection ProceduresEach year (since 1999), the U.S. State Department releases International Religious Freedom Reports on approximately 199 countries or territories. Based on the text in the 2003, 2005, and 2008 reports, ARDA researchers systematically coded the measures included in this file. Under the direction of Brian Grim, the ARDA's Project Manager for International Data during the coding of the 2003 and 2005 reports, these reports were assigned quantitative measures by using a coding instrument, essentially a survey questionnaire; this technique was utilized again during the coding of the 2008 report completed in February 2010. Although the most immediate goal was to develop measures for religious regulation and favoritism, the questions included measures for specific acts of discrimination, prejudice, persecution, warfare, property rights, forced migration and other acts that might (or might not) be related to the religious life of the country. For all variables, the coders were asked to make substantive observations of the qualitative data and to base their codes on empirical observations of actions or patterns of behavior that were documented in the reports.
Users should be aware of the following limitations: (1) All variables reflect information that was coded from the State Department Reports, and when no problem was reported, then the item was coded as "0," meaning: "According to the Report, the item was not mentioned as a problem." Since the reports tend to simply not report a problem rather than say, "the problem is absent," we are not able to reasonably determine whether the problem was unobserved or absent. This means that the data reflect what was reported. (2) The focus of the reports is on limitations of religious freedom. Thus, we would argue that the most accurate measures are those which address the core issues related to the restriction (or regulation) of religious freedom and religious persecution. For example, government favoritism of religious education could arguably be harmless to religious freedom (helping the poor obtain skills) or harmful (training terrorists based on a religious ideology). Thus, since such issues tend to be reported when there is a problem, they cannot be used to form a full picture of the role of religion in education for a country. (3) The three different years of coding are not three discrete measures, but rather represent trend information that continues to be reported for several years running, which makes sense, for instance, because cases of violence tend to have continuing effects.
Thus, it would not be advisable to treat the data as separate measures from which timelines are developed since it may be possible that later years report newly arising problems in addition to old ones. (4) The aggregate dataset for the three years of coding contains the mean score of each ordinal variable across the three years. We suggest that those using the data for social scientific modeling and analysis use the aggregate data set, which has the benefit of greater variation in the variables and lesser error since random errors from one year will be attenuated in the aggregate data. (5) The data in this file represent mean scores from three years of coding, and, therefore may not fit precisely into category labels due to changes in coding responses over time. Users should use category labels as indicators of a country's average response. To find the modal response for this data, it is recommended that users refer to the three previous files of data.
For a more detailed description of the coding procedures, see Grim and Finke (2006).
Sampling ProceduresPrimary data come from the coding of the 199 countries covered by the U.S. State Department International Religious Freedom Reports.
Principal InvestigatorsThe Association of Religion Data Archives
Brian J. Grim, Project Manager for International Data (2003 and 2005 waves)
Jaime D. Harris, Project Manager for International Data (2008 wave)
Robert R. Martin, Research Associate
Roger Finke, Director
Related PublicationsGrim, Brian J. and Roger Finke. 2006. "International Religion Indexes: Government Regulation, Government Favoritism, and Social Regulation of Religion." Interdisciplinary Journal of Research on Religion 2 (Article 1). https://www.religjournal.com/
List of Countries or Territories Included in this Data FileAfghanistan
Antigua and Barbuda (2003 and 2008 reports only; aggregate measures are means from these two waves)
Bosnia & Herzegovina
Central African Republic
Congo, Democratic Republic of (Congo-Kinshasa)
Congo, Republic of (Congo-Brazzaville)
Israeli Occupied Territories (Palestine)
Korea, Democratic People's Republic of (North Korea)
Korea, Republic of (South Korea)
Kosovo (All three waves; 2003 and 2005 data except for population and area taken from report on Serbia and Montenegro)
Kyrgyz Republic (Kyrgyzstan)
Micronesia, Federated States of
Montenegro (All three waves; 2003 and 2005 data except for population and area taken from report on Serbia and Montenegro)
Northern Cyprus (Turkish Cypriot Admin) (2008 report only)
Papua New Guinea
St. Kitts and Nevis
St. Vincent and the Grenadines
Sao Tome and Principe
Serbia (All three waves; 2003 and 2005 data except for population and area taken from report on Serbia and Montenegro)
Slovak Republic (Slovakia)
Timor-Leste (East Timor)
Trinidad and Tobago
United Arab Emirates
Western Sahara (2003 and 2005 reports only; aggregate measures are means from these two waves)
Note 1: Government Regulation of Religion IndexThe Government Regulation of Religion Index (GRI) is calculated by obtaining the sum of the variables GRI_A_AG through GRI_F_AG and multiplying the sum by 5/3; GRI values have a possible range of 0 to 10.
Note 2: Government Favoritism of Religion IndexThe Government Favoritism of Religion Index (GFI) is calculated by obtaining the sum of the variables GFI_A_AG through GFI_E_AG and multiplying the sum by 2. GFI values have a possible range of 0 to 10.
Note 3: Modified Social Regulation of Religion IndexBeginning in 2008, the Modified Social Regulation of Religion Index (MSRI) is used to measure social regulation of religion in countries. The MSRI is calculated by obtaining the sum of the variables MSRIA_AG through MSRIE_AG and multiplying the sum by 2. Countries may have MSRI values between 0 and 10. The MSRI is a modification of the Social Regulation of Religion Index (SRI) described in Grim and Finke (2006). It replaces (1) a measure of social attitudes toward conversions to other religions (SRI_B_AG in the 2001-2005 Aggregate File) and (2) a measure of the situation regarding religious social movements in a country (SRI_E_AG in the 2001-2005 Aggregate File) with the following measures: (1) a measure of citizens' tolerance of nontraditional faiths (MSRIB_AG) and (2) a measure of citizens' receptivity to proselytizing by nontraditional faiths or faiths other than their own (MSRIE_AG). The other three measures used to construct the SRI are still used in the construction of the MSRI.
Note 4: GVTYPEAG CodingSeveral countries receive a score of nine on the variable GVTYPEAG, which indicates a country's type of state or official religion. This coding indicates that there was a change during the coding of the 2003, 2005 and 2008 Reports in how the original measure GVTYPE (i.e., GVTYPE03, GVTYPE05, and GVTYPE08) was coded. Scores on the GVTYPE measures for each of the countries that received a nine on GVTYPEAG are given below.
Argentina: Secular (0) in 2003 and 2008, Christian (2) in 2005
Belarus: Secular (0) in 2003 and 2005, Christian (2) in 2008
Bulgaria: Secular (0) in 2003, Christian (2) in 2005 and 2008
Cyprus: Secular (0) in 2003 and 2008, Christian (2) in 2008
Dominican Republic: Secular (0) in 2003 and 2005, Christian (2) in 2008
Equatorial Guinea: Secular (0) in 2003 and 2008, Christian (2) in 2008
Lebanon: Islamic (3) in 2003, Secular (0) in 2005 and 2008
Moldova: Atheistic (1) in 2003, Secular (0) in 2005 and 2008
Nepal: Hindu (7) in 2003, Secular (0) in 2005 and 2008
Somalia: Atheistic (1) in 2003, Islamic (3) in 2005 and 2008
Sri Lanka: Buddhist (5) in 2003 and 2008, Secular (0) in 2005
Thailand: Buddhist (5) in 2003 and 2005, Secular (0) in 2008
Tibet: Atheistic (1) in 2003 and 2008, Secular (0) in 2005
Turkmenistan: Christian (2) in 2003, Secular (0) in 2005 and 2008
Tuvalu: Secular (0) in 2003 and 2005, Christian (2) in 2008
Venezuela: Secular (0) in 2003 and 2008, Christian (2) in 2008