creg: Composition of Religious and Ethnic Groups (CREG)...
In peacesciencer: Tools and Data for Quantitative Peace Science Research
creg R Documentation
Composition of Religious and Ethnic Groups (CREG) Fractionalization/Polarization Estimates
Description
This is a data set with state-year estimates for ethnic and religious fractionalization/polarization,
by way of the Composition of Religious and Ethnic Groups (CREG) project at the
University of Illinois. I-L-L.
Usage
creg
Format
A data frame with 11523 observations on the following 9 variables.
ccodea Correlates of War state code
gwcodea Gleditsch-Ward state code
creg_ccodea numeric code for the state, mostly patterned off Correlates of War codes but with important differences. See details section for more.
yearthe year
ethfracan estimate of the ethnic fractionalization index. See details for more.
ethpolan estimate of the ethnic polarization index. See details for more.
relfracan estimate of the religious fractionalization index. See details for more.
relpolan estimate of the religious polarization index. See details for more.
Details
The data-raw directory on the project's Github contains more information about how these data were
created. Pay careful attention to how I assigned CoW/G-W codes. The underlying data are version 1.02.
The state codes provided by the CREG project are mostly Correlates of War codes, but with some differences.
Summarizing these differences: the state code for Serbia from 1992 to 2013 is actually the Gleditsch-Ward code (340).
Russia after the dissolution of the Soviet Union (1991-onward) is 393 and not 365. The Soviet Union has the 365 code.
Yugoslavia has the 345 code. The code for Yemen (678) is effectively the Gleditsch-Ward code because it spans the entire
post-World War II temporal domain. Likewise, the code for post-unification Germany is the Gleditsch-Ward code (260) as well.
The codebook actually says it's 265 (which would be East Germany's code), but this is assuredly a typo based on the data.
The codebook cautions there are insufficient data for ethnic group estimates for Cameroon, France, India,
Kosovo, Montenegro, Mozambique, and Papua New Guinea. The French case is particularly disappointing but the
missing data there are a function of both France's constitution and modelling issues for CREG (per the
codebook). There are insufficient data to make religious group estimates for China, North Korea, and the
short-lived Republic of Vietnam.
The fractionalization estimates are the familiar Herfindahl-Hirschman concentration index. The polarization
formula comes by way of Montalvo and Reynal-Querol (2000), though this book does not appear to be published
beyond its placement online. I recommend Montalvo and Reynal-Querol (2005) instead.
You can cite Alesina (2003) for the fractionalization measure if you'd like.
In the most literal sense of "1", the group proportions may not sum to exactly 1 because of rounding in the
data. There were only two problem cases in these data worth mentioning. First, in both data sets, there would
be the occasional duplicates of group names by state-year (for example: Afghanistan in 1951 in the ethnic group
data and the United States in 1948 in the religious group data). In those cases, the script I make available
in the data-raw directory just select distinct values and that effectively fixes the problem of duplicates,
where they do appear. Finally, Costa Rica had a curious problem for most years in the religious group data. All
Costa Rica years have group data for Protestants, Roman Catholics, and "others." Up until 1964 or so, the "others"
are zero. Afterward, there is some small proportion of "others". However, the sum of Protestants, Roman Catholics, and
"others" exceeds 1 (pretty clearly) and the difference between the sum and 1 is entirely the "others." So, I drop
the "others" for all years. I don't think that's terribly problematic, but it's worth saying that's what I did.
References
Alesina, Alberto, Arnaud Devleeschauwer, William Easterly, Sergio Kurlat and Romain Wacziarg. 2003.
"Fractionalization". Journal of Economic Growth 8: 155-194.
Montalvo, Jose G. and Marta Reynal-Querol. 2005. "Ethnic Polarization, Potential Conflict, and Civil Wars"
American Economic Review 95(3): 796–816.
Nardulli, Peter F., Cara J. Wong, Ajay Singh, Buddy Petyon, and Joseph Bajjalieh. 2012.
The Composition of Religious and Ethnic Groups (CREG) Project. Cline Center for Democracy.
peacesciencer documentation built on March 31, 2023, 8:37 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
| creg | R Documentation |
Composition of Religious and Ethnic Groups (CREG) Fractionalization/Polarization Estimates
Description
This is a data set with state-year estimates for ethnic and religious fractionalization/polarization, by way of the Composition of Religious and Ethnic Groups (CREG) project at the University of Illinois. I-L-L.
Usage
creg
Format
A data frame with 11523 observations on the following 9 variables.
ccodea Correlates of War state code
gwcodea Gleditsch-Ward state code
creg_ccodea numeric code for the state, mostly patterned off Correlates of War codes but with important differences. See details section for more.
yearthe year
ethfracan estimate of the ethnic fractionalization index. See details for more.
ethpolan estimate of the ethnic polarization index. See details for more.
relfracan estimate of the religious fractionalization index. See details for more.
relpolan estimate of the religious polarization index. See details for more.
Details
The data-raw directory on the project's Github contains more information about how these data were
created. Pay careful attention to how I assigned CoW/G-W codes. The underlying data are version 1.02.
The state codes provided by the CREG project are mostly Correlates of War codes, but with some differences. Summarizing these differences: the state code for Serbia from 1992 to 2013 is actually the Gleditsch-Ward code (340). Russia after the dissolution of the Soviet Union (1991-onward) is 393 and not 365. The Soviet Union has the 365 code. Yugoslavia has the 345 code. The code for Yemen (678) is effectively the Gleditsch-Ward code because it spans the entire post-World War II temporal domain. Likewise, the code for post-unification Germany is the Gleditsch-Ward code (260) as well. The codebook actually says it's 265 (which would be East Germany's code), but this is assuredly a typo based on the data.
The codebook cautions there are insufficient data for ethnic group estimates for Cameroon, France, India, Kosovo, Montenegro, Mozambique, and Papua New Guinea. The French case is particularly disappointing but the missing data there are a function of both France's constitution and modelling issues for CREG (per the codebook). There are insufficient data to make religious group estimates for China, North Korea, and the short-lived Republic of Vietnam.
The fractionalization estimates are the familiar Herfindahl-Hirschman concentration index. The polarization formula comes by way of Montalvo and Reynal-Querol (2000), though this book does not appear to be published beyond its placement online. I recommend Montalvo and Reynal-Querol (2005) instead. You can cite Alesina (2003) for the fractionalization measure if you'd like.
In the most literal sense of "1", the group proportions may not sum to exactly 1 because of rounding in the
data. There were only two problem cases in these data worth mentioning. First, in both data sets, there would
be the occasional duplicates of group names by state-year (for example: Afghanistan in 1951 in the ethnic group
data and the United States in 1948 in the religious group data). In those cases, the script I make available
in the data-raw directory just select distinct values and that effectively fixes the problem of duplicates,
where they do appear. Finally, Costa Rica had a curious problem for most years in the religious group data. All
Costa Rica years have group data for Protestants, Roman Catholics, and "others." Up until 1964 or so, the "others"
are zero. Afterward, there is some small proportion of "others". However, the sum of Protestants, Roman Catholics, and
"others" exceeds 1 (pretty clearly) and the difference between the sum and 1 is entirely the "others." So, I drop
the "others" for all years. I don't think that's terribly problematic, but it's worth saying that's what I did.
References
Alesina, Alberto, Arnaud Devleeschauwer, William Easterly, Sergio Kurlat and Romain Wacziarg. 2003. "Fractionalization". Journal of Economic Growth 8: 155-194.
Montalvo, Jose G. and Marta Reynal-Querol. 2005. "Ethnic Polarization, Potential Conflict, and Civil Wars" American Economic Review 95(3): 796–816.
Nardulli, Peter F., Cara J. Wong, Ajay Singh, Buddy Petyon, and Joseph Bajjalieh. 2012. The Composition of Religious and Ethnic Groups (CREG) Project. Cline Center for Democracy.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.