Country codes - Revision history

Nils Duepont: /* Misc */

2025-03-31T11:20:28Z

‎Misc

Nils Duepont: /* Contributors and revisions */

2025-03-31T09:47:06Z

‎Contributors and revisions

Nils Duepont: /* Misc */

2025-03-31T09:46:31Z

‎Misc

Nils Duepont: /* A note on "counterfactual and related unit" coding */

2025-03-31T09:27:59Z

‎A note on "counterfactual and related unit" coding

Nils Duepont: /* A note on "counterfactual and related unit" coding */

2025-03-31T09:25:04Z

‎A note on "counterfactual and related unit" coding

Show changes

Nils Duepont: /* Exemplifying a related unit coding */

2025-03-31T09:18:41Z

‎Exemplifying a related unit coding

Nils Duepont: /* A note on “counterfactual and related unit” coding */

2025-03-31T09:15:18Z

‎A note on “counterfactual and related unit” coding

Nils Duepont: /* Amendments and limitations */

2025-03-31T09:10:21Z

‎Amendments and limitations

Nils Duepont: /* A note on “counterfactual and related unit” coding */

2025-03-31T09:10:09Z

‎A note on “counterfactual and related unit” coding

Nils Duepont: /* Amendments and limitations */

2025-03-31T09:05:10Z

‎Amendments and limitations

@@ Line 4,151: / Line 4,151: @@
 * related_list corrected for South Korea on March 10, 2021
 * 35 International Organizations added as entities on May 27, 2024
-* Section on "related unit coding" added as well as new entities, namely: Soviet Union, Yugoslavia, Greenland, French Guyana and New Caledonia and Dependencies on March 31, 2025
+* Section on "related unit coding" added as well as new entities, namely: Soviet Union, Yugoslavia, Greenland, French Guiana and New Caledonia and Dependencies on March 31, 2025
 == Sources ==

@@ Line 4,135: / Line 4,135: @@
 </table></shtml>
-== Contributors and revisions ==
+== Misc ==
 Contributors: Nils Düpont, Fabian Besche, Bastian Becker, and Heiner Fechner

@@ Line 4,135: / Line 4,135: @@
 </table></shtml>
-== Misc ==
+== Contributors and revisions ==
 Contributors: Nils Düpont, Fabian Besche, Bastian Becker, and Heiner Fechner

@@ Line 48: / Line 48: @@
 * restrictions to the scope of certain COW codes, while
 * adding new entities to the country scheme that COW does not cover.
 ==== Exemplifying a related unit coding ====
-In the case of Yugoslavia, for example, data for Bosnia and Herzegovina, Croatia, Macedonia, Montenegro, Serbia, or Slovenia may be coded in a counterfactual sense looking back thereby creating a "consistent time-series" for todays’ entity. This implies
+In the case of Yugoslavia data for Bosnia and Herzegovina, Croatia, Macedonia, Montenegro, Serbia, or Slovenia may be coded in a counterfactual sense looking back thereby creating a "consistent time-series" for todays’ entity. This implies
 * to fill each country's time-series with information that applied to entire Yugoslavia ("related unit coding");
 * to restrict the COW code of Serbia (345) to today’s Serbia thereby diverging from COW by "excluding" Yugoslavia;
@@ Line 181: / Line 185: @@
 | consistent time-series looking back from today’s entity
 |}
 ==== Idiosyncratic codings ====

@@ Line 62: / Line 62: @@
 Use case #1 applies the "counterfactual coding" resulting in a consistent time-series for each focal entity. For use case #2, information from the federal level that diverges from information for the republics is still accurately captured in that today’s countries bear their proper data even for times being part of Yugoslavia while the latter information for the higher-level unit does not get lost.
-[INSERT TABLES]
+{| class="wikitable" border="1" style="display: inline-table; margin: 20px; width: 40%"
 ==== Idiosyncratic codings ====

@@ Line 33: / Line 33: @@
 There are many other country coding schemes out there like [https://en.wikipedia.org/wiki/ISO_3166-1 ISO 3166-1] or the scheme applied by the "Varieties of Democracy" Project ([https://www.v-dem.net/en/ V-Dem]) - each with a different scope regarding time and entities. Given the analytical scope of the [http://www.socialpolicydynamics.de/en/ CRC 1342], it was decided to opt for the COW scheme in the first place, and modify and ammend it where necessary. For the most parts, the WeSIS numerical codes are in line with COW. Since COW has become a standard on its own, V-Dem includes COW codes as well making it easy to merge WeSIS, COW and/or V-Dem data. Likewise, packages like [https://github.com/vincentarelbundock/countrycode countrycode] for R or [https://ideas.repec.org/c/boc/bocode/s453301.html kountry] for Stata provide valuable tools to convert from and to COW codes if needed.
-== A note on “counterfactual and related unit” coding ==
+== A note on "counterfactual and related unit" coding ==
-Each country coding scheme reflects the analytical lens of its (scientific) project at a given time point. COW, for example, treats as many entities as separate as possible to allow for adequately describing and analyzing inter- and intra-state wars – the founding principle of COW. V-Dem, in turn, emphasizes the continuity with a contemporary nation-state being interested in the state of democracy and how it changes. Being published in 1974, the ISO 3166 scheme was introduced to reflect the need for harmonized and machine-readable standards, foremost for official statistical purposes, but with a focus on the then existing countries and their subdivisions. Given that countries, borders and names change, there is no “one-size-fits-all scheme”; for the social sciences, the question rather is which scheme “does the least harm” when assigning values to entities. In this sense, the coding scheme used for WeSIS sits in-between COW and V-Dem (while still providing ISO 3166 codes when downloading data).
+Each country coding scheme reflects the analytical lens of its (scientific) project at a given time point. COW, for example, treats as many entities as separate as possible to allow for adequately describing and analyzing inter- and intra-state wars – the founding principle of COW. V-Dem, in turn, emphasizes the continuity with a contemporary nation-state being interested in the state of democracy and how it changes. Being published in 1974, the ISO 3166 scheme was introduced to reflect the need for harmonized and machine-readable standards, foremost for official statistical purposes, but with a focus on the then existing countries and their subdivisions. Given that countries, borders and names change, there is no "one-size-fits-all scheme"; for the social sciences, the question rather is which scheme "does the least harm" when assigning values to entities. In this sense, the coding scheme used for WeSIS sits in-between COW and V-Dem (while still providing ISO 3166 codes when downloading data).
-Analyzing the introduction of social policies and the interplay between national determinants and international and transnational interdependencies, for example colonial legacies, the COW scheme best reflected the CRC’s need at that time. Once introduced, there has been an analytical shift though to describe and analyze the dynamics of the focal policy within a country over time (i.e., more like V-Dem’s focus). To maintain consistency, comparability, and compatibility, WeSIS still relies on the COW coding scheme, but certain modifications and deviations were needed to adequately describe and map the dynamics of social policies.
+Analyzing the introduction of social policies and the interplay between national determinants and international and transnational interdependencies, for example colonial legacies, the COW scheme best reflected the CRC's need at that time. Once introduced, there has been an analytical shift though to describe and analyze the dynamics of the focal policy within a country over time (i.e., more like V-Dem's focus). To maintain consistency, comparability, and compatibility, WeSIS still relies on the COW coding scheme, but certain modifications and deviations were needed to adequately describe and map the ''dynamics'' of social policies.
-While any deviation is noted in the above table in the column remarks, certain coding issues arise when assigning values to the “proper” entity at a given point in time: Georgia, for example, did not “suddenly” introduce a health-care system upon independence in 1991 but rather “inherited” the system from the time being part of the Soviet Union. For this reason – and unless noted otherwise on an indicator page – in WeSIS a “counterfactual and related unit coding” is applied.
+While any deviation is noted in the above table in the column remarks, certain coding issues arise when assigning values to the "proper" entity at a given point in time: Georgia, for example, did not "suddenly" introduce a health-care system upon independence in 1991 but rather "inherited" the system from the time being part of the Soviet Union. For this reason – and unless noted otherwise on an indicator page – in WeSIS a '''''counterfactual and related unit coding''''' is applied.
-A “counterfactual coding” assigns the values of related entities to the “lower” units in cases like the Soviet Union, Yugoslavia, Austria-Hungary and the like for times where today’s (!) focal entity was not (yet) independent. The related units thereby “inherit” the values of the related entity.
+A "counterfactual coding" assigns the values of related entities to the "lower" units in cases like the Soviet Union, Yugoslavia, Austria-Hungary and the like for times where today’s (!) focal entity was not (yet) independent. The related units thereby "inherit" the values of the related entity.
-One issue with this kind of coding, still, is that it may be misleading for some entities if differing information is available for both the related unit and the “lower” units. A case in point is Yugoslavia's 1974 constitutional reform which significantly decentralized social policy, delegating extensive authority to the constituent republics. While the federal government only set minimum standards, the republics were granted the authority to expand both the scope and coverage of benefits within their territories. Consequently, data on Yugoslavia may indicate a reduction in benefits, as it reflects only the minimum federal standards.
+One issue with this kind of coding, still, is that it may be misleading for some entities if differing information is available for both the related unit and the "lower" units. A case in point is Yugoslavia's 1974 constitutional reform which significantly decentralized social policy, delegating extensive authority to the constituent republics. While the federal government only set minimum standards, the republics were granted the authority to expand both the scope and coverage of benefits within their territories. Consequently, data on Yugoslavia may indicate a reduction in benefits, as it reflects only the minimum federal standards.
 For this reason, in WeSIS we opted for
-* combining the “counterfactual coding” with
+* combining the "counterfactual coding" with
 * restrictions to the scope of certain COW codes, while
 * adding new entities to the country scheme that COW does not cover.
@@ Line 51: / Line 51: @@
 ==== Exemplifying a related unit coding ====
-In the case of Yugoslavia, for example, data for Bosnia and Herzegovina, Croatia, Macedonia, Montenegro, Serbia, or Slovenia may be coded in a counterfactual sense looking back thereby creating a “consistent time-series” for todays’ entity. This implies
+In the case of Yugoslavia, for example, data for Bosnia and Herzegovina, Croatia, Macedonia, Montenegro, Serbia, or Slovenia may be coded in a counterfactual sense looking back thereby creating a "consistent time-series" for todays’ entity. This implies
-* to fill each country’s time-series with information that applied to entire Yugoslavia (“related unit coding”);
+* to fill each country's time-series with information that applied to entire Yugoslavia ("related unit coding");
-* to restrict the COW code of Serbia (345) to today’s Serbia thereby diverging from COW by “excluding” Yugoslavia;
+* to restrict the COW code of Serbia (345) to today’s Serbia thereby diverging from COW by "excluding" Yugoslavia;
-* to add a new code for Yugoslavia, strictly limited to the time from 1918 to 1992.
+* to add a new code for Yugoslavia (3451), strictly limited to the time from 1918 to 1992.
 Below are two use cases taking the Yugoslavian example. For the sake of ease assume two indicators:
-* one (e.g., a macro-economic indicator) for which information is only available for entire Yugoslavia (aka the “related entity”);
+* one (e.g., a macro-economic indicator) for which information is only available for entire Yugoslavia (aka the '''''related entity''''');
-* and a second capturing the scope of benefits as in the example given above, i.e., information for the “related units” is available and either no information for the higher-level unit is available or is available but diverges.
+* and a second capturing the scope of benefits as in the example given above, i.e., information for the '''''related units''''' is available and either no information for the higher-level unit is available or is available but diverges.
-Use case #1 applies the “counterfactual coding” resulting in a consistent time-series for each focal entity. For use case #2, information from the federal level that diverges from information for the republics is still accurately captured in that today’s countries bear their proper data even for times being part of Yugoslavia while the latter information for the higher-level unit does not get lost.
+Use case #1 applies the "counterfactual coding" resulting in a consistent time-series for each focal entity. For use case #2, information from the federal level that diverges from information for the republics is still accurately captured in that today’s countries bear their proper data even for times being part of Yugoslavia while the latter information for the higher-level unit does not get lost.
 [INSERT TABLES]
@@ Line 69: / Line 69: @@
 * '''Germany''': COW provides three codes to capture Germany (255), the Federal Republic of Germany (260) and the German Democratic Republic (265). Given that COW resumes the 255-code after 1990 raises the question how to capture/display a continuity. For this reason, it was agreed to rely on the latest codes (i.e., 255 and 265), and code the data as follows:
-** Duplicate any 255-data before 1954/1955 for the GDR thereby creating a “265-timeline” – which is similar to use case #1 (“counterfactual coding”). 265 is then restricted temporally to end in 1990.
+** Duplicate any 255-data before 1954/1955 for the GDR thereby creating a "265-timeline" – which is similar to use case #1 ("counterfactual coding"). 265 is then restricted temporally to end in 1990.
 ** Assign any data of the Federal Republic of Germany (260) to the latest code, i.e., 255; this way, the code of 260 becomes obsolete and future data collections shall only use 255 for data between 1945 and 1990.
-* '''Vietnam''': Vietnam (816) and Republic of Vietnam (817): Prior to independence, use case #1 applies in that the time-series for Vietnam (816) and the Republic of Vietnam (817) bear the values from the “dominant” colonial power; upon unification the time-series of 817 simply ends (like the time-series of the GDR) whereas 816 continues.
+* '''Vietnam''': Vietnam (816) and Republic of Vietnam (817): Prior to independence, use case #1 applies in that the time-series for Vietnam (816) and the Republic of Vietnam (817) bear the values from the "dominant" colonial power; upon unification the time-series of 817 simply ends (like the time-series of the GDR) whereas 816 continues.
-* '''Yemen''': COW provides three codes to capture Yemen (679), the Yemen Arab Republic (678) and Yemen People’s Republic (680; “South Yemen”). In line with V-Dem and many studies in piece research, it was agreed to still continue the code 678 as the “main” code for Yemen (instead of using the latest “unification code”, thereby making 679 obsolete). Analogous to the case of the GDR, it was agreed to duplicate any 678-data before 1967 for “South Yemen” thereby creating a “680-timeline” which temporally ends in 1990 (and which is similar to use case #1, i.e., a “counterfactual coding”).
+* '''Yemen''': COW provides three codes to capture Yemen (679), the Yemen Arab Republic (678) and Yemen People’s Republic (680; "South Yemen"). In line with V-Dem and many studies in piece research, it was agreed to still continue the code 678 as the "main" code for Yemen (instead of using the latest "unification code", thereby making 679 obsolete). Analogous to the case of the GDR, it was agreed to duplicate any 678-data before 1967 for "South Yemen" thereby creating a "680-timeline" which temporally ends in 1990 (and which is similar to use case #1, i.e., a "counterfactual coding").
 == Relation types, or: how to read the table ==

← Older revision		Revision as of 09:10, 31 March 2025
Line 32:		Line 32:

	There are many other country coding schemes out there like [https://en.wikipedia.org/wiki/ISO_3166-1 ISO 3166-1] or the scheme applied by the "Varieties of Democracy" Project ([https://www.v-dem.net/en/ V-Dem]) - each with a different scope regarding time and entities. Given the analytical scope of the [http://www.socialpolicydynamics.de/en/ CRC 1342], it was decided to opt for the COW scheme in the first place, and modify and ammend it where necessary. For the most parts, the WeSIS numerical codes are in line with COW. Since COW has become a standard on its own, V-Dem includes COW codes as well making it easy to merge WeSIS, COW and/or V-Dem data. Likewise, packages like [https://github.com/vincentarelbundock/countrycode countrycode] for R or [https://ideas.repec.org/c/boc/bocode/s453301.html kountry] for Stata provide valuable tools to convert from and to COW codes if needed.		There are many other country coding schemes out there like [https://en.wikipedia.org/wiki/ISO_3166-1 ISO 3166-1] or the scheme applied by the "Varieties of Democracy" Project ([https://www.v-dem.net/en/ V-Dem]) - each with a different scope regarding time and entities. Given the analytical scope of the [http://www.socialpolicydynamics.de/en/ CRC 1342], it was decided to opt for the COW scheme in the first place, and modify and ammend it where necessary. For the most parts, the WeSIS numerical codes are in line with COW. Since COW has become a standard on its own, V-Dem includes COW codes as well making it easy to merge WeSIS, COW and/or V-Dem data. Likewise, packages like [https://github.com/vincentarelbundock/countrycode countrycode] for R or [https://ideas.repec.org/c/boc/bocode/s453301.html kountry] for Stata provide valuable tools to convert from and to COW codes if needed.
−

	== A note on “counterfactual and related unit” coding ==		== A note on “counterfactual and related unit” coding ==

← Older revision		Revision as of 09:10, 31 March 2025
Line 35:		Line 35:

	== A note on “counterfactual and related unit” coding ==		== A note on “counterfactual and related unit” coding ==
		+
		+	Each country coding scheme reflects the analytical lens of its (scientific) project at a given time point. COW, for example, treats as many entities as separate as possible to allow for adequately describing and analyzing inter- and intra-state wars – the founding principle of COW. V-Dem, in turn, emphasizes the continuity with a contemporary nation-state being interested in the state of democracy and how it changes. Being published in 1974, the ISO 3166 scheme was introduced to reflect the need for harmonized and machine-readable standards, foremost for official statistical purposes, but with a focus on the then existing countries and their subdivisions. Given that countries, borders and names change, there is no “one-size-fits-all scheme”; for the social sciences, the question rather is which scheme “does the least harm” when assigning values to entities. In this sense, the coding scheme used for WeSIS sits in-between COW and V-Dem (while still providing ISO 3166 codes when downloading data).
		+
		+	Analyzing the introduction of social policies and the interplay between national determinants and international and transnational interdependencies, for example colonial legacies, the COW scheme best reflected the CRC’s need at that time. Once introduced, there has been an analytical shift though to describe and analyze the dynamics of the focal policy within a country over time (i.e., more like V-Dem’s focus). To maintain consistency, comparability, and compatibility, WeSIS still relies on the COW coding scheme, but certain modifications and deviations were needed to adequately describe and map the dynamics of social policies.
		+
		+	While any deviation is noted in the above table in the column remarks, certain coding issues arise when assigning values to the “proper” entity at a given point in time: Georgia, for example, did not “suddenly” introduce a health-care system upon independence in 1991 but rather “inherited” the system from the time being part of the Soviet Union. For this reason – and unless noted otherwise on an indicator page – in WeSIS a “counterfactual and related unit coding” is applied.
		+
		+	A “counterfactual coding” assigns the values of related entities to the “lower” units in cases like the Soviet Union, Yugoslavia, Austria-Hungary and the like for times where today’s (!) focal entity was not (yet) independent. The related units thereby “inherit” the values of the related entity.
		+	One issue with this kind of coding, still, is that it may be misleading for some entities if differing information is available for both the related unit and the “lower” units. A case in point is Yugoslavia's 1974 constitutional reform which significantly decentralized social policy, delegating extensive authority to the constituent republics. While the federal government only set minimum standards, the republics were granted the authority to expand both the scope and coverage of benefits within their territories. Consequently, data on Yugoslavia may indicate a reduction in benefits, as it reflects only the minimum federal standards.
		+
		+	For this reason, in WeSIS we opted for
		+	* combining the “counterfactual coding” with
		+	* restrictions to the scope of certain COW codes, while
		+	* adding new entities to the country scheme that COW does not cover.
		+
		+	==== Exemplifying a related unit coding ====
		+
		+	In the case of Yugoslavia, for example, data for Bosnia and Herzegovina, Croatia, Macedonia, Montenegro, Serbia, or Slovenia may be coded in a counterfactual sense looking back thereby creating a “consistent time-series” for todays’ entity. This implies
		+	* to fill each country’s time-series with information that applied to entire Yugoslavia (“related unit coding”);
		+	* to restrict the COW code of Serbia (345) to today’s Serbia thereby diverging from COW by “excluding” Yugoslavia;
		+	* to add a new code for Yugoslavia, strictly limited to the time from 1918 to 1992.
		+
		+	Below are two use cases taking the Yugoslavian example. For the sake of ease assume two indicators:
		+	* one (e.g., a macro-economic indicator) for which information is only available for entire Yugoslavia (aka the “related entity”);
		+	* and a second capturing the scope of benefits as in the example given above, i.e., information for the “related units” is available and either no information for the higher-level unit is available or is available but diverges.
		+
		+	Use case #1 applies the “counterfactual coding” resulting in a consistent time-series for each focal entity. For use case #2, information from the federal level that diverges from information for the republics is still accurately captured in that today’s countries bear their proper data even for times being part of Yugoslavia while the latter information for the higher-level unit does not get lost.
		+
		+	[INSERT TABLES]
		+
		+	==== Idiosyncratic codings ====
		+
		+	Few cases still exist in the COW coding scheme where use cases #1 and #2 do not work immediately; for these, the following codes and coding rules were agreed upon:
		+
		+	* '''Germany''': COW provides three codes to capture Germany (255), the Federal Republic of Germany (260) and the German Democratic Republic (265). Given that COW resumes the 255-code after 1990 raises the question how to capture/display a continuity. For this reason, it was agreed to rely on the latest codes (i.e., 255 and 265), and code the data as follows:
		+	** Duplicate any 255-data before 1954/1955 for the GDR thereby creating a “265-timeline” – which is similar to use case #1 (“counterfactual coding”). 265 is then restricted temporally to end in 1990.
		+	** Assign any data of the Federal Republic of Germany (260) to the latest code, i.e., 255; this way, the code of 260 becomes obsolete and future data collections shall only use 255 for data between 1945 and 1990.
		+
		+	* '''Vietnam''': Vietnam (816) and Republic of Vietnam (817): Prior to independence, use case #1 applies in that the time-series for Vietnam (816) and the Republic of Vietnam (817) bear the values from the “dominant” colonial power; upon unification the time-series of 817 simply ends (like the time-series of the GDR) whereas 816 continues.
		+
		+	* '''Yemen''': COW provides three codes to capture Yemen (679), the Yemen Arab Republic (678) and Yemen People’s Republic (680; “South Yemen”). In line with V-Dem and many studies in piece research, it was agreed to still continue the code 678 as the “main” code for Yemen (instead of using the latest “unification code”, thereby making 679 obsolete). Analogous to the case of the GDR, it was agreed to duplicate any 678-data before 1967 for “South Yemen” thereby creating a “680-timeline” which temporally ends in 1990 (and which is similar to use case #1, i.e., a “counterfactual coding”).

	== Relation types, or: how to read the table ==		== Relation types, or: how to read the table ==

← Older revision		Revision as of 09:05, 31 March 2025
Line 32:		Line 32:

	There are many other country coding schemes out there like [https://en.wikipedia.org/wiki/ISO_3166-1 ISO 3166-1] or the scheme applied by the "Varieties of Democracy" Project ([https://www.v-dem.net/en/ V-Dem]) - each with a different scope regarding time and entities. Given the analytical scope of the [http://www.socialpolicydynamics.de/en/ CRC 1342], it was decided to opt for the COW scheme in the first place, and modify and ammend it where necessary. For the most parts, the WeSIS numerical codes are in line with COW. Since COW has become a standard on its own, V-Dem includes COW codes as well making it easy to merge WeSIS, COW and/or V-Dem data. Likewise, packages like [https://github.com/vincentarelbundock/countrycode countrycode] for R or [https://ideas.repec.org/c/boc/bocode/s453301.html kountry] for Stata provide valuable tools to convert from and to COW codes if needed.		There are many other country coding schemes out there like [https://en.wikipedia.org/wiki/ISO_3166-1 ISO 3166-1] or the scheme applied by the "Varieties of Democracy" Project ([https://www.v-dem.net/en/ V-Dem]) - each with a different scope regarding time and entities. Given the analytical scope of the [http://www.socialpolicydynamics.de/en/ CRC 1342], it was decided to opt for the COW scheme in the first place, and modify and ammend it where necessary. For the most parts, the WeSIS numerical codes are in line with COW. Since COW has become a standard on its own, V-Dem includes COW codes as well making it easy to merge WeSIS, COW and/or V-Dem data. Likewise, packages like [https://github.com/vincentarelbundock/countrycode countrycode] for R or [https://ideas.repec.org/c/boc/bocode/s453301.html kountry] for Stata provide valuable tools to convert from and to COW codes if needed.
		+
		+
		+	== A note on “counterfactual and related unit” coding ==

	== Relation types, or: how to read the table ==		== Relation types, or: how to read the table ==