Ethnolinguistic Fractionalization (ELF) Indices, 1961 and 1985

Philip G. Roeder
Department of Political Science
University of California, San Diego
La Jolla, CA  92093-0521
858 / 534-6000
Fax: 858  / 534-7130


I computed these indices from population estimates in the following sources:

Bromlei, Iu. V., ed. 1988.  Narody mira: istoriko-etnograficheskii spravochnik. Moscow: Sovetskaia Entsiklopediia.

Bruk, S[olomon] I. 1986.  Naselenie mira: etnodemograficheskii spravochnik, 2d rev. ed. Moscow: Izdatelístvo "Nauka".

Bruk, S. I. and V. S.  Apenchenko, eds.  1964.  Atlas narodov mira.  Moscow: Glavnoe upravlenie geodezii i kartografii gosudarstvennogo geologicheskogo komiteta SSSR and Institut etnografii im. H. H. Miklukho-Maklaia, Akademiia nauk SSSR.

In addition, I filled in missing cases with data from:

Europa World Yearbook (for the Czechoslovakian and Yugoslavian successor states and for Bulgaria)

USSR.  Gosudarstvennyi komitet SSSR po statistike. 1992. Itogi vsesoiuznoi perepisi naseleniia 1989 goda. Minneapolis: East View Publications.  (for the Soviet successor states)

Alternative Estimates

Column 1 lists country codes. Most correspond to the Singer-Small list, except in the cases of successor states and newly unified states, where I invent a few codes in order to distinguish the cases. (For example, the ELF for Russia is different from the Soviet Union and so these cases do not have the same code.)

Column 2 lists country names.

Columns 3 to 5 reprint  the indices published in Charles Lewis Taylor and Michael C. Hudson.  1972.  World Handbook of Political and Social Indicators, 2d ed.  New Haven: Yale University Press. Pp. 271-274.

In my own data for 1961 and 1985 I calculate the ELF in three different ways.  Each index uses the Taylor and Hudson formula, but the three indices differ from one another in what they consider an ethnic group.  You must make a decision that will be determined in large part by your research question and your own definitions of ethnicity:

Columns 6 and 9 report an ELF Index (for 1961 and 1985, respectively) that uses none of the groupings reported in the sources when data on sub-groups are available.  (For example, it treats separate Native American groups as separate ethnic groups rather than combining these in a catch-all "Indigenous Peoples".  Similarly,  it treats  Hutus and Tutsis as separate ethnic groups rather than grouping these as  Banyarwanda in Rwanda or Barundi in Burundi). In addition, in settler societies of the Western Hemisphere, this index treats racial distinctions within ethnolinguistic groups (Afro-Americans versus White Americans or Afro-Colombians versus Euro-Colombians) as separate ethnic groups.

Columns 7 and 10 report the ELF Index (for 1961 and 1985, respectively) if racial distinctions within linguistic groups do not constitute separate ethnic groups. (In this index, for example, White and Black Americans constitute a single ethnic group. Similarly, Afro-Colombians, Euro-Colombians, Mestizo-Colombians, and Mullato-Colombians are simply Colombians.) Note that Taylor and Hudson are inconsistent in this; they make this distinction in the United States but not in other Western Hemisphere societies. Note: Columns 7 and 10 show only the ELF Indices that differ from Columns 6 and 9, respectively. (That is, to use this column you must fill in the blank spaces from the previous column.)

Columns 8 and 11 report an ELF Index (for 1961 and 1985, respectively) that groups racial and cultural sub-groups within linguistic groups in ethnic groups defined by language. (In this index, for example, all Arabs in Kuwait constitute a single ethnic group rather than separate Kuwaiti, Syrian, Palestinian, and other ethnic groups. The groupings are indicated in column 12. Note: Columns 7 and 10 show only the ELF Indices that differ from the previous columns. (That is, to use this column you must fill in the blank spaces from the previous columns.)

Column 12 (Annotations) lists the groups that affect the calculation of the ELF Index in columns 8 and 11. A name without parentheses (for example, Arab) indicates the group that is made up of sub-groups. A list of names linked by pluses and placed in parentheses indicates the constituents of a group.


If you use these data please cite this web-site.  I would recommend the following citation:
Philip G. Roeder.  2001.   "Ethnolinguistic Fractionalization (ELF) Indices, 1961 and 1985." February 16.  <http//\~proeder\elf.htm>.  (Date you consulted web-site).
Thank you most sincerely for that consideration.

Please address questions, suggestions, and corrections to me at:

The data are available as an Excel spreadsheet. You may copy these to your computer and incorporate them into your own data sets for analysis.

Posted: February 16, 2001. (Most Recently Up-dated: September 18, 2001)

Link to Excel file:[ELF.xls]