Codebook - U

advertisement
Codebook: Regions, Regional Power, and
Status Attribution
October 2009
Table of Contents
Variables: ........................................................................................................................................ 2
Works Cited: ............................................................................................................................... 6
Appendices: Coding Processes ....................................................................................................... 7
Appendix 1: Process by which missing SIPRI data was imputed .............................................. 7
Appendix 2: Coding the Cultural Indicators ............................................................................. 16
Appendix 3: ISO Country Codes .............................................................................................. 18
Variables:







cowabb1: 3 letter country abbreviation, state 1
cowabb2: 3 letter country abbreviation, state 2
ccode1: COW Country code, state 1
ccode2: COW Country code, state 2
iso1: ISO 3 letter country code, state1
iso2: ISO 3 letter country code, state2
year: Year of observation
From COW’s CINC data:
 irst: Iron and steel production (thousands of tons)
 milex: Military Expenditures (thousands of current year US Dollars)
 milper: Military Personnel (thousands)
 energy: Energy consumption (thousands of coal-ton equivalents)
 tpop: Total Population (thousands)
 upop: Urban population (population living in cities with population greater than 100,000)
 cinc: Composite Index of National Capability (CINC) score
From the SIPRI data:
 milexp: Military expenditure (in millions of 2005 USD)
o missing data was imputed; process is listed in Appendix 1
From the WDI: (*need to check these explanations)
 gdp: GDP (how is this measured?)
 captl: market capitalization of domestic firms
 pop: total population
 resid: residual of regression population on GDP
From EUgene: (*need to check these explanations)
 contig: Direct contig (1=Lnd, 2=1-12 miles watr, 3=13-24, 4=25-150, 5=151-400, 6=?)
 distance: Distance, from the capitol of country 1 to the capitol of country 2, in milse
From the IDEA Events Data:
 statevisits: Number of state visits from state 1 to state 2 per year
 cc: sum of the absolute value on the Goldstein scale of events between country 1 and
country 2 per year
 cc_nosv: Identical to cc, but excludes all state visits from state 1 to state 2
 cc_count: A count of all events, conflictual, cooperative, and neutral, from state 1 to state
2 per year
 cc_countnosv: Identical to cc_count, but excludes all state visits from state 1 to state 2






coop: Sum of all cooperative events between state 1 and state 2 per year, using weighted
values of events from the Goldstein scale
coop_nosv: Identical to coop, but excludes all state visits from state 1 to state 2
conflict: Sum of all conflictual events between state 1 and state 2 per year, using
weighted values of events from the Goldstein scale
coop_count: Count of all cooperative events between state 1 and state 2 per year
coop_countnosv: Identical to coop_count, but excludes all state visits from state 1 to state
2
conf_count: Count of all conflictual events between state 1 and state 2 per year
From Goertz and Powers (2009):
 rei: 1 signifies the dyad shares membership to a common REI, 0 not; this data is coded
from the list provided by Goertz and Powers (2009)
From The Military Balance (MB) (range 2002-2005):
 milpers: number of active military personnel per country; countries for which ranges are
given (e.g., 11-15000) are treated as missing data
 lowbound: if a range is given for military personnel, the lower bound of the range is listed
 highbound: if a range is given for military personnel, the upper bound of the range is
listed
 milpersavg: lists the number of active military personnel per country, using the average if
a range is given
 milperslow: lists the number of active military personnel per country, using the lower
bound if a range is given (a more conservative measure of active personnel)
 milpershigh: lists the number of active military personnel per country, using the upper
bound if a range is given (a more liberal measure of active personnel)
 milpersavg_th: identical to milpersavg, but divided by 1000
 milperslow_th: identical to milperslow, but divided by 1000
 milpershigh_th: identical to milperhigh, but divided by 1000
Military Personnel Variables:
 milper_avg: combines the CINC milper variable (ranges from 1988-2001) and the MB’s,
milpersavg_th variable (ranges from 2002-2005), to obtain a count of military personnel
in thousands from 1988 to 2005
 milper_h: identical to milper_avg, but uses the upper bound of the MB range
 milper_l: identical to milper_avg, but uses the lower bound of the MB range
Barbieri Trade Data (Non-directed dyad bilateral trade data)

flow1: imports of state 1 from state 2 (rename to imports)

flow2: imports of state 2 from state 1 (rename exports)
From CIA world fact book?












langmaj: 1 if the majority language family of state 1 and state 2 is the same, 0 otherwise
langmin: 1 if the minority language family of state 1 and state 2 is the same, 0 otherwise
langminmaj: 1 if the minority language family of state 1 matches the majority language
family of state 2, 0 otherwise
langmajmin: 1 if the majority language family of state 1 matches the minority language
family of state 2, 0 otherwise
scriptmaj: 1 if the majority writing script of state 1 and state 2 is the same, 0 otherwise
scriptmin: 1 if the minority writing script of state 1 and state 2 is the same, 0 otherwise
scriptminmaj: 1 if the minority writing scrip of state 1 matches the majority script of
state 2, 0 otherwise
scriptmajmin: 1 if the majority writing scrip of state 1 matches the minority script of state
2, 0 otherwise
religionmaj: 1 if the majority religion of state 1 and state 2 is the same, 0 otherwise
religionmin: 1 if the minority religion of state 1 and state 2 is the same, 0 otherwise
religionmajmin: 1 if the majority religion of state 1 matches the minority religion of state
2, 0 otherwise
religionminmaj: 1 if the majority religion of state 1 matches the minority religion of state
2, 0 otherwise
Our Opportunity and Willingness Measures:









milexp_total: sum of all military expenditure in the system per year
milpers_total: sum of all military personnel in the system per year
milexp_cinc1: the military expenditure of country 1 (milexp) divided by the sum of all
military expenditure in the system that year (milexp_total)
milexp_cinc2: the military expenditure of country 2 (milexp) divided by the sum of all
military expenditure in the system that year (milexp_total)
milper_cinc1: the military personnel of country 1 (milper_avg) divided by the sum of all
military expenditure in the system that year (milpers_total)
milper_cinc2: the military personnel of country 1 (milper_avg) divided by the sum of all
military expenditure in the system that year (milpers_total)
military_cinc1: the average value of the military expenditure “cinc” score and the
military personnel “cinc” score for country 1
military_cinc2: the average value of the military expenditure “cinc” score and the
military personnel “cinc” score for country 2
captl_total: the total amount of capital in the system per year













captl_cinc1: the amount of capital of country 1 (captl1) divided by the total amount of
capital in the system
captl_cinc2: the amount of capital of country 1 (captl1) divided by the total amount of
capital in the system
lofp_military: loss of power gradient for military strength
(military_cinc1^log10[(distance/500) + 10 – e ] )
lofp_economic: loss of power gradient for military strength
(captl_cinc1^log10[(distance/500) + 10 – e ] )
milpower_comp: a comparison of country 1’s military power at home versus at country
2’s capital (lofp_military / military_cinc1)
econpower_comp: a comparison of country 1’s economic power at home versus at
country 2’s capital (lofp_econmic / captl_cinc1)
mil_opp: coded 1 if country 1 has at least 50 % of its military power at country 2’s capital
as compared to its military power at home (mil_opp = 1 if milpower_comp > .5; 0 if
milpower_comp ≤ .5)
econ_opp: coded 1 if country 1 has at least 50 % of its economic power at country 2’s
capital as compared to its economic power at home (mil_opp = 1 if milpower_comp > .5;
0 if milpower_comp ≤ .5)
mill_will: coded 1 if country 1 directs at least 1 event towards country 2
econ_will: coded 1 if country 1 and country 2 trade at least 1 USD worth
mil_oxw: military opportunity times military willingness; 1 indicates countries are able
and willing to interact diplomatically/militarily)
econ_oxw: economic opportunity times military willingness; 1 indicates countries are
able and willing to interact economically)
max_oxw: the maximum value of mil_oxw and econ_oxw
Works Cited:
Barbieri, Katherine, Omar Keshk, and Brian Pollins. 2008. Correlates of War Project Trade
Data Set Codebook, Version 2.0. Online: http://correlatesofwar.org.
Goertz, Gary, and Kathy L. Powers. 2009. “The economic–institutional construction of regions:
conceptualization and operationalization.”
COW
SIPRI
Military Balance
IDEA Events Data
EUgene Data
WDI indicators
ATOP data?
Appendices: Coding Processes
Appendix 1: Process by which missing SIPRI data was imputed
Added data is color coded in excel and details for each added entry are listed below by country in
the order which they appear in the revised SIPRI dataset.
For countries with red numbers, data was calculated for the missing state (state A) using
expenditures of the neighboring state with the largest military expenditures (state B) for all years
missing in state A. The ratio of State B to State A's military expenditures was calculated using
data from the correlates of war. That ratio was then multiplied by the SIPRI data for State B.
The resulting figure was entered for the missing value of State A for each year available.
For countries with green 0's, COW data was not available for the missing state. These missing
points typically correspond with periods of domestic turmoil and conflict, or the missing entry
falls after 2001.
For countries with blue numbers, data was provided in a previous version of SIPRI, but
the most recent version of SIPRI has removed that entry for review.
Libya
88-92; 94-96 extrapolated from Libya-Algeria COW ratio
93 data coded 0, missing in COW
Angola
88-91; 93-97 extrapolated from Angola-Zambia COW ratio
92 data coded 0, missing in COW
98 extrapolated from Angola-Namibia COW ratio
Benin
91-92; 94-98 extrapolated from Benin-Nigeria COW ratio
93 data coded 0, missing in COW
Cape Verde
89-90 data coded 0, missing in COW
91-92 extrapolated from Cape Verde-Senegal COW ratio
Central African Republic
88-90; 97-01 extrapolated from Central African Republic-Cameroon COW ratio
Chad
88-92 extrapolated from Chad-Nigeria COW ratio
Congo
88-90; 92; 94-00 extrapolated from Congo-Cameroon COW ratio
91; 93 coded 0, missing in COW
Congo, Dem. Rep.
88-92; 93-95; 01 extrapolated from Cong, Dem. Rep.-Sudan COW ratio
92; 02 coded 0, missing in COW
Cote d'Ivoire
95 extrapolated from Cote d'Ivoire-Ghana COW ratio
98-01 extrapolated from Cote d'Ivoire-Guinea COW ratio
02 coded 0, COW ends in 01
Equatorial Guinea
88-93 coded 0, missing from COW
96-01 extrapolated from Equatorial Guinea-Cameroon COW ratio
02-05 coded 0, COW ends in 01
Eritrea
88-92 coded "." Eritrea not a state
04-05 coded 0, COW ends in 01
Gabon
88-90; 94-99 extrapolated from Gabon-Cameroon COW ratio
91-93 coded 0, missing in COW
Guinea
88-90; 95-96 extrapolated from Guinea-Senegal COW ratio
05 coded 0, COW ends in 01
Guinea-Bissau
88; 91-93 coded 0, missing in COW
90; 99 extrapolated from Guinea-Bissau-Senegal COW ratio
04 coded 0, COW ends in 01
Liberia
88-91;93-97; 00 extrapolated from Liberia-Sierra Leone COW ratio
92 coded 0, missing in COW
01-02 coded from 06 SIPRI data
03 coded 0, COW ends in 01
Mali
91-92 extrapolated from Mali-Algeria COW ratio
Namibia
88-89 coded "." not a state
90 coded from 06 SIPRI data
Niger
88-93 extrapolated from Niger-Algeria COW ratio
Sierra Leone
98-99 extrapolated from Sierra Leone-Guinnea COW ratio
Somalia
88-90; 95-01 extrapolated Somalia-Ethiopia COW ratio
91-94 coded 0, missing in COW
02-05 coded 0, COW ends in 01
Sudan
89 extrapolated from Sudan-Egypt COW ratio
Tanzania
88 extrapolated from Tanzania-Kenya COW ratio
Togo
96 coded from 06 SIPRI data
97-01 coded from Togo-Ghana COW ratio
02 coded 0, COW ends in 01
Zambia
98; 00 coded from 06 SIPRI data
01 extrapolated from Zambia-Angola COW ratio
02-03 coded 0, COW ends in 01
Cuba
88-91;94-01 extrapolated from Cuba-USA COW ratio
92-93 coded 0, missing from COW
02-05 coded 0, COW ends in 01
Haiti
88-92; 95-01 extrapolated from Haiti-Dominican Republic Ratio
93-94 coded 0, missing from COW
02-05 coded 0, COW ends in 01
Jamaica
88-89 extrapolated from Jamaica-Dominican Republic
Trinidad and Tobago
88;90;93 coded 0, mising in COW
89 extrapolated from Trinidad and Tobago-Guyana COW ratio
91-92; 96-01 extrapolated from Trinidad and Tobago-Venezuala COW ratio
94-95 coded from 2006 SIPRI data
02-05 coded 0, COW ends in 01
Belize
88; 98-99 extrapolated from Belize-Mexico COW ratio
Costa Rica
88-99 extrapolated from Costa Rica-Panama COW ratio
00-01 extrapolated from Costa Rica-Nicaragua COW ratio
02-05 coded 0, COW ends in 01
Honduras
88-99 extrapolated from Honduras-Guatemala COW ratio
Nicaragua
88;90 extrapolated from Nicaragua-Guatemala COW ratio
89 coded 0, missing in COW
Panama
00-01 extrapolated from Panama-Colombia COW ratio
02-05 coded 0, COW ends in 01
Bolivia
88 extrapolated from Bolivia-Chile COW ratio
Guyana
97-01 extrapolated from Guyana-Brazil COW ratio
02-05 coded 0, COW ends in 01
Paraguay
88 coded from the 06 SIPRI data
Peru
88 extrapolated from Peru-Chile COW ratio
Venezuala
88-90 extrapolated from Venezuala-Brazil COW ratio
Kazakhstan
88-90 coded "." state does not exist
91 coded 0, missing in COW data
92 extrapolated from Kazakhstan-Russia COW ratio
Kyrgyzstan
88-90 coded "." state does not exist
91 coded 0, missing in COW data
Tajikistan
88-90 coded "." state does not exist
91 coded 0, missing in COW data
05 coded 0, COW ends in 01
Turkmenistan
88-90 coded "." state does not exist
91 coded 0, missing in COW data
92-93; 00-01 extrapolated from Turkmenistan-Iran COW ratio
02-05 coded 0, COW ends in 01
Uzbekistan
88-90 coded "." state does not exist
91-93 coded 0, missing in COW data
98l 00 extrapolated from Uzbekistan-Kazakhstan COW ratio
02; 04-05 coded 0, COW ends in 01
North Korea
88-94; 98-04 coded from the 06 SIPRI data
95-97 coded "0" COW ratio unreliable relative to SIPRI data
Laos
88;90-91 coded 0, missing in COW data
89 extrapolated froms Laos-China COW ratio
Mongolia
88-89 extrapolated from Mongolia-Russia COW ratio
Myanmar
88-01 extrapolated from Myanmar-China COW ratio
02-05 coded 0, COW ends in 01
Vietnam
88-94 coded from 06 SIPRI data
95-01 extrapolated from Vietnam-China COW ratio
02 coded 0, COW ends in 01
Afghanistan
88-89; 91-94 coded 0, missing in COW
90; 95-01 extrapolated from Afghanistan-China COW ratio
02 coded 0, COW ends in 01
Tongo
88 coded "." state does not exist
Albania
88-89; 91 extrapolated from Albania-Greece COW ratio
Armenia
88-90 coded "." state does not exist
91 extrapolated from Armenia-Turkey COW ratio
94 coded 0, missing in COW
Azerbaijan
88-90 coded "." state does not exist
91 extrapolated from Azerbaijan-Iran COW ratio
Belarus
88-90 coded "." state does not exist
91 coded 0, missing in COW
Bosnia-Herzegovina
88-91 coded "." state does not exist
92-01 extrapolated from Bosnia-Herzegovina-Croatia Ratio
Croatia
88-91 coded "." state does not exist
Czech Republic
88-92 coded "." state does not exist
Czechoslovakia
88-92 extrapolated from Czechoslovakia-Germany
93-05 coded "." state does not exist
EStonia
88-90 coded "." state does not exist
91 coded 0, missing in COW
Georgia
88-90 coded "." state does not exist
91-95 coded 0, COW data unreliable
Germany, DR
88-89 extrapolated from Germany, DR-Germany FR COW ratio
90 coded 0, COW data missing
Latvia
88-90 coded "." state does not exist
91 coded 0, COW data missing
92 extrapolated from Latvia-Russia COW ratio
Lithuania
88-90 coded "." state does not exist
91 coded 0, COW data missing
92 extrapolated from Lithuania-Russia COW ratio
Macedonia
88-92 coded "." state does not exist
93-95 coded from 06 SIPRI data
Moldova
88-90 coded "."
91-92 coded 0, COW data unreliable
Montenegro
88-05 coded "." state does not exist
Russia
91 coded from 06 SIPRI data
Serbia/Yugoslavia
88-91 coded as 0, COW data unreliable
92-95 coded using 06 sIPRI data
Slovak Republic
88-92 coded "." state does not exist
Slovenia
88-81 coded "." state does not exist
Ukraine
88-90 coded "." state does not exist
91-92 coded 0, COW data unreliable
Iran
88 coded from 06 SIPRI data
Iraq
88-01 extrapolated from Iran-Iraq COW ratio
02-04 coded 0, COW data missing
Lebanon
89 extrapolated from Lebanon-Israel COW ratio
Qatar
88; 91; 93-01 extrapolated from Qatar-Saudi Arabia COW ratio
89-00; 92 coded 0, COW data missing
02-05 coded 0, COW data ends in 01
Yemen
88-89 coded "." state does not exist
Yemen, Arab Rep
88-90 extrapolated from Yemen, Arab Rep-Saudi Arabia COW ratio
91-05 coded "." state does not texist
Yemen, PR
88-89 extrapolated from Yemen, PR-Saudi Arabia COW ratio
90-05 coded "," state does not exist
Yugoslavia (former) entry merged with Serbia for consistency with COW country codes
Appendix 2: Coding the Cultural Indicators
Majority language family:
 Refers to single biggest language in the country – not the language family. For example,
the population of the Cote d’Ivorie speak Niger-Congo languages. However, there are
more than 250 different languages in the CDI, spoken by a varying number of people.
Instead, French is the language spoken by the most Ivorians, and the CDI is coded as a
“Majority Romance” country.
 Once the majority language has been identified, it is assigned to one of a number of broad
categories. If it does not fit into any of them, it is called an “Isolate”. The most prominent
isolates are Celtic and Hebrew.
 A country can only have one majority language. If it has several official languages, the
most widespread one is coded as the majority language; the others become minority
languages.
Minority language families:
 Includes any language beyond the majority language spoken by at least 10% of the
population.
 These are coded in the same way the majority languages. If there are no languages
beyond the official one spoken by at least 10% of the population, this category is coded
“0”.
 A country can have any number of minority languages.
Majority script:
 Refers to the alphabet used to write the majority language.
 Once the majority script has been identified, it is assigned to one of a number of broad
categories. If it does not fit into any of them, it is called an “Isolate”. The most prominent
isolates are Greek and Hebrew.
 A country can only have one majority script. If several alphabets are used, the most
widely used is coded as the majority script.
 Generally, all African and Latin American native languages were converted to Latin
characters under colonial administration. Thus, though pre-colonial alphabets exist for
many of these languages, they currently use Latin characters and are coded as “Latin
script”.
Minority scripts:
 Includes any alphabet(s) beyond the one used by the majority language.
 These are coded in the same way the minority languages. A minimum of 10% of the
population must use the script, otherwise it is coded as “0”
 A country can have any number of minority scripts.
Majority religion:
 This category refers to the country’s religion with the most followers.



Once the majority religion has been identified, it is assigned to one of a number of broad
categories. If it does not fit into any of them, it is called an “Isolate/Local/Tribal”. The
most prominent isolates are voodoo in the Caribbean and animism and ancestor worship
in Africa and SE Asia.
A country can only have one majority religion. If a country has several official religions,
the largest of these is coded as the “Majority religion”.
If a country has an official religion, but this religion is smaller than another religion, the
largest religion is still coded as the “Majority religion”.
Minority religions:
 Includes any religion beyond the majority religion practiced by at least 10% of the
population.
 These are coded in the same way the majority religion. If there are no religions beyond
the official one practiced by at least 10% of the population, this category is coded “0”.
 A country can have any number of minority religions.
Special:
 All categories are binary. States are or are not members of each category – this is
represented, respectively, with a “1” or a “0”.
 Countries are not coded as both “Majority” and “Minority” if their languages belong to
the same broad language families.
 The same applies to scripts.
 The one exception to this rule is Eritrea. Most Eritreans speak Tigrinya written in the
Ge’ez alphabet, while a minority speak and write Arabic. Both languages belong to the
Afro-Asiatic category, but the languages and the alphabets are mutually unintelligible.
Eritrea is thus coded as both majority and minority Afro-Asiatic, with a majority Ge’ez
and a minority Perso-Arabic scripts.
Appendix 3: ISO Country Codes
Countries without ISO codes—pre-2008
 East Germany
 West Germany
 Czechoslovakia
 South Yemen (Democratic Republic of Yemen)
 North Yemen (Yemen Arab Republic)
 Yugoslavia/Serbia and Montenegro/Serbia
Download