Project 4: Exploratory Data Analysis
An Analysis of International Crisis Behavior
Matthew R. Versaggi, MS, MBA - Artificial Intelligence Engineer
June 23, 2016
- 1 Introduction
- 2 Explorations of the Dataset Structure
- 3 Validate Datset with external Sources
- 4 Exploration of Selected Crisis Attributes
- 4.1 Attributes List used in this exploration:
- 4.2 Definitions of Attributes:
- 4.3 Exploration of Crisis Attributes by ERA (WW2, Cold War, Super Power):
- 4.3.1 Crisis Participants (Actors)
- 4.3.2 Crisis Belligerents (Trigents)
- 4.3.3 Trigger of the Crisis
- 4.3.4 Issue of the crisis
- 4.3.5 Crisis Management I: Major Response to Crisis Trigger
- 4.3.6 Crisis Management II: Principal Coping Technique
- 4.3.7 Crisis Outcome (Win, Lose, Draw)
- 4.3.8 Power Status of Actor
- 4.3.9 Power Status of Threat
- 4.3.10 Violence experiences by Crisis Actor
- 4.3.11 Political Regime of Crises Actor.
- 4.3.12 Gravest threat during crisis.
- 4.3.13 Nuclear capability of Crisis Actor
- 4.3.14 Alliance Patterns
- 5 Explorations of Actors (Participants)
- 6 Exlporations of USA crises during the Cold War ERA
- 7 Final Plots
- 8 Reflections
1 Introduction
This is an structured exploration of the The International Crisis Behavior (ICB) Project dataset, which can be found here: http://sites.duke.edu/icbdata/data-collections/. The aim of the ICB Project is to shed light on a pervasive phenomenon of world politics, military conflict, and sovereign state crises.
Key Goals: (1) Explore key crisis attributes, interactions and relationships by ERA: (WW2, Cold War, Super Power), (2) explore the behaviors of Major (those with the highest frequencies) crisis “Actors” by ERA, (3) explore the USA crises during the Cold War ERA, and finally (4) explore some interesting multidimensional crisis characteristics of the dataset in unique ways.
The prime objective is to guide the reader through the exploration of the four key goals stated above, in a methodical, pragmatic fashion, in order to provide an appreciation for the richness (and gravity) of this dataset.
A secondary objective is to provide a structured framework in which the data could speak for itself and tell its own story. Every effort will be made to adhear to the spirit of this objective.
2 Explorations of the Dataset Structure
This dataset consists of 470 unique crises, 95 attributes (variables), 1,036 crisis actors (records), including 35 protracted conflicts, all from the period of 1918 to 2013.
- It is a rich dataset consisting almost entirely of categorical (no ordering) and ordinal (categorical but with ordering) data.
I have augmented the dataset with a half dozen transformed variables to better facilitate the conveying of the information in some of the various plots used in the explorations.
There are only a dozen attributes of the 95 available which will be the focus of this exploration. These are listed in the section “Exploration of Selected Crisis Attributes”.
Below is the structure of the dataset:
## 'data.frame': 1036 obs. of 95 variables:
## $ icb2 : Factor w/ 2 levels "ICB2","UCB2": 1 1 1 1 1 1 1 1 1 1 ...
## $ crisno : int 1 2 2 3 4 4 4 4 5 5 ...
## $ cracno : int 1 2 3 4 5 6 7 8 9 10 ...
## $ cracid : int 365 93 94 365 365 366 368 367 315 290 ...
## $ actor : Factor w/ 143 levels "AFG","ALB","ALG",..: 109 91 29 109 109 42 77 73 33 103 ...
## $ systrgyr: int 1918 1918 1918 1918 1918 1918 1918 1918 1919 1919 ...
## $ systrgmo: int 5 5 5 6 11 11 11 11 1 1 ...
## $ systrgda: int NA 25 25 23 18 18 18 18 15 15 ...
## $ crisname: Factor w/ 469 levels "AALAND ISLANDS",..: 377 113 113 378 27 27 27 27 431 431 ...
## $ triggr : int 9 7 4 7 6 9 9 9 2 7 ...
## $ yrtrig : int 1918 1918 1919 1918 1918 1918 1918 1918 1919 1919 ...
## $ motrig : int 5 5 1 6 11 11 12 12 1 1 ...
## $ datrig : int NA 25 25 23 18 22 NA NA 15 23 ...
## $ trigent : int 996 94 996 997 366 365 365 365 290 315 ...
## $ trigloc : int 2 1 2 1 1 1 1 1 1 1 ...
## $ southv : int 220 94 93 200 366 365 365 365 290 315 ...
## $ southpow: int 3 1 1 3 1 3 3 3 2 1 ...
## $ sizedu : int 1 NA NA 1 1 NA NA NA NA NA ...
## $ strcdu : int 1 1 NA 1 1 1 1 1 1 1 ...
## $ comlev : int 7 1 1 7 8 8 8 8 3 3 ...
## $ majres : int 8 3 6 8 8 9 9 9 8 8 ...
## $ yerres : int 1918 1918 1919 1918 1918 1918 1918 1918 1919 1919 ...
## $ monres : int 5 5 1 7 11 11 12 12 1 1 ...
## $ dayres : int 28 30 28 1 22 22 NA NA 23 23 ...
## $ trgresra: int 14 6 4 9 5 1 NA NA 9 1 ...
## $ crismg : int 8 4 4 8 8 8 8 8 8 8 ...
## $ cenvio : int 4 1 1 4 4 4 4 4 4 4 ...
## $ sevvio : int 3 1 1 3 3 3 3 3 3 3 ...
## $ usinv : int 7 7 7 3 3 3 3 3 3 3 ...
## $ usfavr : int 3 1 3 3 3 1 1 1 3 3 ...
## $ suinv : int 9 1 1 9 9 8 8 8 1 1 ...
## $ sufavr : int 8 5 5 8 8 3 3 3 5 5 ...
## $ gbinv : int 7 1 1 8 3 6 3 3 3 3 ...
## $ gbfavr : int NA 5 5 NA NA NA NA NA 3 3 ...
## $ frinv : int 2 1 1 8 3 1 3 3 3 3 ...
## $ frfavr : int NA 5 5 NA NA NA NA NA 3 3 ...
## $ itinv : int 1 1 1 1 1 1 1 3 3 3 ...
## $ itfavr : int NA 5 5 NA NA NA NA NA 3 3 ...
## $ grinv : int 1 1 1 1 8 1 1 8 1 1 ...
## $ grfavr : int NA 5 5 NA NA NA NA NA NA NA ...
## $ jpinv : int 7 1 1 1 1 1 1 1 3 3 ...
## $ jpfavr : int NA 5 5 NA NA NA NA NA 3 3 ...
## $ globorg : int 1 1 1 1 1 1 1 1 1 1 ...
## $ globact : int 1 1 1 1 1 1 1 1 1 1 ...
## $ globfavr: int 1 1 1 1 1 1 1 1 1 1 ...
## $ regorg : int 0 1 1 0 0 0 0 0 0 0 ...
## $ regact : int 0 1 1 0 0 0 0 0 0 0 ...
## $ rofavr : int 0 5 5 0 0 0 0 0 0 0 ...
## $ outcom : int 1 1 4 1 4 1 1 1 2 2 ...
## $ outfor : int 6 4 4 6 9 8 8 8 9 9 ...
## $ outevl : int 2 2 3 2 3 2 2 2 4 4 ...
## $ outesr : int 1 2 2 1 2 2 2 2 2 2 ...
## $ yrterm : int 1920 1918 1919 1919 1920 1920 1920 1920 1920 1920 ...
## $ moterm : int 4 12 9 9 8 2 7 8 7 7 ...
## $ daterm : int 1 15 3 27 11 2 12 11 28 28 ...
## $ trgterra: int 686 205 222 462 632 438 574 603 560 552 ...
## $ resterra: int 673 199 218 453 627 438 574 603 551 552 ...
## $ actloc : int 30 42 42 30 30 34 34 34 31 31 ...
## $ geog : int 30 42 42 30 34 34 34 34 31 31 ...
## $ cractloc: int 1 1 1 1 1 2 2 2 1 1 ...
## $ noactr : int 7 5 6 5 8 8 8 8 3 3 ...
## $ stainsys: int 47 47 49 47 47 47 47 47 49 49 ...
## $ period : int 1 1 1 1 1 1 1 1 1 1 ...
## $ syslev : int 2 1 1 2 2 1 1 1 1 1 ...
## $ pc : int 2 2 2 2 1 1 1 1 1 1 ...
## $ pcid : int 27 6 6 27 0 0 0 0 0 0 ...
## $ viol : int 3 1 2 3 3 3 3 3 3 3 ...
## $ iwc : int 1 1 1 1 1 1 1 1 1 1 ...
## $ powdis : int NA 1 -1 NA 12 -12 -12 -12 -1 1 ...
## $ gpinv : int 7 4 4 7 7 7 7 7 3 3 ...
## $ powinv : int 1 1 1 1 1 1 1 1 1 1 ...
## $ age : int 1 3 3 1 1 3 3 3 3 3 ...
## $ territ : int 3 1 1 3 3 1 1 1 1 2 ...
## $ regime : int 2 2 2 2 2 2 2 2 1 4 ...
## $ durreg : int 1 2 1 1 1 1 1 1 1 1 ...
## $ allycap : int 4 2 1 4 4 2 2 2 1 1 ...
## $ globmemb: int 4 4 4 4 4 4 4 4 4 4 ...
## $ nuclear : int 1 1 1 1 1 1 1 1 1 1 ...
## $ powsta : int 3 1 1 3 3 1 1 1 1 2 ...
## $ issue : int 1 1 1 1 2 2 2 2 1 1 ...
## $ chissu : int 4 6 6 4 6 6 6 6 6 6 ...
## $ gravty : int 2 1 2 2 3 6 6 6 3 3 ...
## $ pethin : int 0 0 0 0 0 0 0 0 0 0 ...
## $ col : int 1 NA NA 1 1 4 4 4 4 4 ...
## $ unemp : int NA NA NA NA NA 4 4 4 4 4 ...
## $ inflat : int 1 1 1 1 1 4 4 4 4 4 ...
## $ foodpr : int 1 NA NA 1 1 4 4 4 4 4 ...
## $ labstr : int 1 NA NA 1 1 4 4 4 4 4 ...
## $ short : int 1 NA NA 1 1 4 4 4 4 4 ...
## $ econdt : int 1 NA NA 1 1 4 4 4 4 4 ...
## $ regrep : int NA 1 1 NA NA 4 4 4 4 4 ...
## $ socunr : int 1 NA NA 1 1 4 4 4 4 4 ...
## $ massvl : int 1 1 1 1 1 4 4 4 4 4 ...
## $ gvinst : int 1 2 NA 1 1 4 4 4 4 4 ...
## $ sourdt : int 3 2 2 3 3 3 3 3 3 3 ...## icb2 crisno cracno cracid actor systrgyr systrgmo systrgda
## 1 ICB2 1 1 365 RUS 1918 5 NA
## 2 ICB2 2 2 93 NIC 1918 5 25
## 3 ICB2 2 3 94 COS 1918 5 25
## 4 ICB2 3 4 365 RUS 1918 6 23
## 5 ICB2 4 5 365 RUS 1918 11 18
## 6 ICB2 4 6 366 EST 1918 11 18
## crisname triggr yrtrig motrig datrig trigent trigloc southv
## 1 RUSSIAN CIVIL WAR I 9 1918 5 NA 996 2 220
## 2 COSTA RICAN COUP 7 1918 5 25 94 1 94
## 3 COSTA RICAN COUP 4 1919 1 25 996 2 93
## 4 RUSSIAN CIVIL WAR II 7 1918 6 23 997 1 200
## 5 BALTIC INDEPENDENCE 6 1918 11 18 366 1 366
## 6 BALTIC INDEPENDENCE 9 1918 11 22 365 1 365
## southpow sizedu strcdu comlev majres yerres monres dayres trgresra
## 1 3 1 1 7 8 1918 5 28 14
## 2 1 NA 1 1 3 1918 5 30 6
## 3 1 NA NA 1 6 1919 1 28 4
## 4 3 1 1 7 8 1918 7 1 9
## 5 1 1 1 8 8 1918 11 22 5
## 6 3 NA 1 8 9 1918 11 22 1
## crismg cenvio sevvio usinv usfavr suinv sufavr gbinv gbfavr frinv frfavr
## 1 8 4 3 7 3 9 8 7 NA 2 NA
## 2 4 1 1 7 1 1 5 1 5 1 5
## 3 4 1 1 7 3 1 5 1 5 1 5
## 4 8 4 3 3 3 9 8 8 NA 8 NA
## 5 8 4 3 3 3 9 8 3 NA 3 NA
## 6 8 4 3 3 1 8 3 6 NA 1 NA
## itinv itfavr grinv grfavr jpinv jpfavr globorg globact globfavr regorg
## 1 1 NA 1 NA 7 NA 1 1 1 0
## 2 1 5 1 5 1 5 1 1 1 1
## 3 1 5 1 5 1 5 1 1 1 1
## 4 1 NA 1 NA 1 NA 1 1 1 0
## 5 1 NA 8 NA 1 NA 1 1 1 0
## 6 1 NA 1 NA 1 NA 1 1 1 0
## regact rofavr outcom outfor outevl outesr yrterm moterm daterm trgterra
## 1 0 0 1 6 2 1 1920 4 1 686
## 2 1 5 1 4 2 2 1918 12 15 205
## 3 1 5 4 4 3 2 1919 9 3 222
## 4 0 0 1 6 2 1 1919 9 27 462
## 5 0 0 4 9 3 2 1920 8 11 632
## 6 0 0 1 8 2 2 1920 2 2 438
## resterra actloc geog cractloc noactr stainsys period syslev pc pcid viol
## 1 673 30 30 1 7 47 1 2 2 27 3
## 2 199 42 42 1 5 47 1 1 2 6 1
## 3 218 42 42 1 6 49 1 1 2 6 2
## 4 453 30 30 1 5 47 1 2 2 27 3
## 5 627 30 34 1 8 47 1 2 1 0 3
## 6 438 34 34 2 8 47 1 1 1 0 3
## iwc powdis gpinv powinv age territ regime durreg allycap globmemb
## 1 1 NA 7 1 1 3 2 1 4 4
## 2 1 1 4 1 3 1 2 2 2 4
## 3 1 -1 4 1 3 1 2 1 1 4
## 4 1 NA 7 1 1 3 2 1 4 4
## 5 1 12 7 1 1 3 2 1 4 4
## 6 1 -12 7 1 3 1 2 1 2 4
## nuclear powsta issue chissu gravty pethin col unemp inflat foodpr labstr
## 1 1 3 1 4 2 0 1 NA 1 1 1
## 2 1 1 1 6 1 0 NA NA 1 NA NA
## 3 1 1 1 6 2 0 NA NA 1 NA NA
## 4 1 3 1 4 2 0 1 NA 1 1 1
## 5 1 3 2 6 3 0 1 NA 1 1 1
## 6 1 1 2 6 6 0 4 4 4 4 4
## short econdt regrep socunr massvl gvinst sourdt
## 1 1 1 NA 1 1 1 3
## 2 NA NA 1 NA 1 2 2
## 3 NA NA 1 NA 1 NA 2
## 4 1 1 NA 1 1 1 3
## 5 1 1 NA 1 1 1 3
## 6 4 4 4 4 4 4 33 Validate Datset with external Sources
Rationale:
Part of the process of any exploratory data analysis is to perform a preliminary “gut check” of the data with an external source just to verify that it “makes sense”, if even only in a cursory fashion. This section is that vetting process.
NOTE: The colored dots on the bar chart correspond to the heightened key international crisis levels during the time period of 1918 - 2013. They also correspond to the key international crises below (left to right on the chart is top to bottom of the list). This is a reasonable “gut check” that the data is indeed not bad data. The table below was constructed via commonly found historical records of the time period.
- Russian Civil War
- World War 2 starts
- Cold War Starts
- The 1947 Arab-Israeli war
- Berlin Crisis
- The 1973 Arab-Israeli War
- Soviet-Afghan War
- 1987 Sino-Indian skirmish
- Gulf War
- Kosovo War
- Iraq War
- Syrian Civil War
4 Exploration of Selected Crisis Attributes
- There are only a dozen or so key attributes (of the 96 variables) explored over three distinct ERAs of the time period 1918 - 2013. Those are illustrated in the following graph, listed below, and subsequently defined in the next section.
4.1 Attributes List used in this exploration:
- “Crisis Trigger”
- “Crisis Issue”
- “Major Response”
- “Coping Technique”
- “Actor Power Status
- “Threat Power Status”
- “Crisis Violence”
- “Political Regime”
- “Alliance Patterns
- “Nuclear Capability”
- “Gravest Threat”
- “Crisis Outcome”
4.2 Definitions of Attributes:
- NOTE: It is prudent to remind the reader to peruse this section carefully as the precise definitions of these attributes and their values takes on special meanings in some cases, and will aide in understanding some of the explanatory graphs later on.
“Crisis Trigger”
- The trigger or precipitating cause of a foreign policy crisis refers to the specific act, event or situational change which leads decision-makers to perceive a threat. A trigger may be initiated by: an adversary state; a non-state actor; or a group of states (military alliance).
Verbal act - protest, threat, accusation, demand, etc.
Political act - subversion, alliance formation by adversaries, diplomatic sanctions, severance of diplomatic relations, violation of treaty
Economic act - embargo, dumping, nationalization of property, withholding of economic aid.
External change - intelligence report, change in specific weapon, weapon system, offensive capability, change in global system or regional subsystem, challenge to legitimacy by international organization.
Other non-violent act.
Internal verbal or physical challenge to regime or elite - incitement by media, proclamation of new regime, fall of government, coup d’etat, sabotage act, terrorism, assassination, riot, demonstration, strike, arrest, martial law, execution, mutiny, revolt.
Non-violent military act - show of force, war game or maneuvers, mobilization, movement of forces, change of force posture to offensive.
Indirect violent act.
Violent act - border clash, border crossing by limited force, invasion of air space, sinking of ship, sea-air incident, bombing of large target, large-scale military attack, war.
“Crisis Issue”
- This variable identifies the most important initial issue area of the crisis as perceived by the crisis actor.
Military-security - territory, borders, access to the sea, free navigation, irredentism, change in the military balance, military incidents, war.
Political-diplomatic - sovereignty, hegemony, international status, change in international system, colonialism, cold war.
Economic-developmental - treatment of property, raw material resources, oil, economic recession, economic pressure, currency problems.
Cultural-status - ideology, threat to non-material values, internal problems, state of regime, population problems.
Other.
“Major Response”
- Once the decision makers of a state perceive the act/event/change which triggers its crisis, they will decide on an appropriate response. The major response is that specific action which captures the major thrust of its behavior.
No response-inaction.
Verbal act - protest, threat, accusation, demand, etc.
Political act - subversion, alliance formation by adversaries, diplomatic sanctions, severance of diplomatic relations, violation of treaty threat to commit economic or military act..
Economic act - embargo, dumping, nationalization of property, withholding of economic aid.
Other non-violent act.
Non-violent military act - declaration of war, show of force, war games or maneuvers, mobilization, movement of forces, with demand of forces, change of force posture to offensive, military aid.
Multiple including non-violent military act.
Violent military act - border clash, border crossing by limited force, invasion of air space, sinking of ship, sea-air incident, bombing of large target, large-scale military attack, war.
Multiple including violent military act.
“Coping Technique”
- States employ a variety of techniques to cope with foreign policy crises. This variable identifies the primary crisis management technique used by a crisis actor, as distinct from a specific act, which was the focus of the major response variable.
Negotiation - formal, informal, bilateral, multilateral, international, diplomatic exchange.
Adjudication or arbitration.
Mediation - by global or regional organization, ally, or alliance personnel.
Multiple not including violence.
Non-military pressure - e.g., withholding of promised economic aid.
Non-violent military - physical acts (maneuvers, repositioning of forces); verbal acts (oral and written statements by authorized leaders threatening to use violence).
Multiple including violence.
Violence.
“Actor Power Status”
- The power status of the actor (the one who has to respond to the crisis).
Small power.
Middle power.
Great power.
Superpower.
“Threat Power Status”
- This variable assesses the power status of the state or other entity identified as the source of the threat.
Small power.
Middle power.
Great power.
Superpower.
“Crisis Violence”
- This variable identifies the extent of violence experienced by a crisis actor, regardless of its use or non-use as a crisis management technique.
No violence.
Minor clashes .
Serious clashes.
Full-scale war .
“Political Regime”
- This variable distinguishes between authoritarian and democratic regimes, as well as between civil and military regimes at the time of the crisis. Criteria for identifying democratic regimes are: competitive elections; pluralist representation in the legislature; several autonomous centers of authority in the political system; competitive parties; and a free press.
Democratic regime
Civil authoritarian regime
Military-direct rule
Military-indirect rule
Military dual authority
“Alliance Patterns”
- This variable identifies the type of alliance pattern which existed for the crisis actor at the time of its crisis.
Non-aligned or neutral.
Informal alliance with superpower or great power.
Formal alliance with superpower or great power.
Alliance leader - superpower or great power.
“Nuclear Capability”
- Did the crisis actor possess nuclear weapons or access to them at the time the crisis began?
No (foreseeable) nuclear capability - the actor did not possess a nuclear capability with any operational military significance when the crisis began; moreover, the international consensus at the time was that it could not develop or acquire such capability within five years.
Foreseeable nuclear capability - the actor could develop or acquire operational nuclear military capability within five years of the beginning of the crisis.
Possession of nuclear capability - the actor had nuclear military capability (weapons) and delivery means but no second-strike capability.
Developed nuclear capability, with second strike capability superpower or great power with ability to absorb a first strike and retaliate.
“Gravest Threat”
- This variable identifies the object of gravest threat at any time during the crisis, as perceived by the principal decision makers of the crisis actor.
Economic threat.
Limited military threat.
Political threat - threat of overthrow of regime, change of institutions, replacement of elite, intervention in domestic politics, subversion .
Territorial threat - threat of integration, annexation of part of a state’s territory, separatism.
Threat to influence in the international system or regional subsystem - threat of declining power in the global system and/or regional subsystem, diplomatic isolation, cessation of patron aid .
Threat of grave damage - threat of large casualties in war, mass bombings.
Threat to existence - threat to survival of population, of genocide, threat to existence of entity, of total annexation, colonial rule, occupation.
Other.
“Crisis Outcome”
- This variable deals with the content of crisis termination. Did a crisis actor yield? Did it triumph? Was a compromise reached? Was there a blurred outcome regarding goal achievement? The outcome is indicated from the perspective of a specific actor. The values which fall under the categories listed below should be thought of in terms of achievement/nonachievement of basic goals(s) by a crisis actor in the context of a specific crisis.
Victory - achievement of basic goal(s); the crisis actor defeated a threatening adversary by counter-threats .
Compromise - partial achievement of basic goal(s).
Stalemate - no effect on basic goal(s); no clear outcome to the crisis; no change in the situation.
Defeat - non-achievement of basic goal(s); the crisis actor yielded or surrendered when an adversary threatened basic values.
Other.
4.3 Exploration of Crisis Attributes by ERA (WW2, Cold War, Super Power):
This section explores key crisis attributes from the entire dataset segmented by the three ERA’s. Each attribute is examined individually to highlight it’s nuances, gravity, patterns, and trends.
- NOTE: It si worth articulating that this section is about general trends, not specific interactions (discussed later on), so most of the analysis is quite general in nature, will cover all of the participants, over all of the years in question.
4.3.1 Crisis Participants (Actors)
- Major Actors (Countries with the highest crisis counts) Worldwide
| Country | Country | Country | Country |
|---|---|---|---|
| China | Egypt | France | Germany |
| Greece | India | Iraq | Israel |
| Japan | Libya | Pakistan | Russia |
| Syria | Turkey | UK | USA |
| Yugoslavia |
ANALYSIS:
CAVEAT: These next two graphs obscure a good bit of the data on the X axis, that is OK because the key takeaways are simply observing the simple macro trends that are discussed below.
The key observation from this graph is that the “frequency” of crises and actors increases significantly during the cold war period, and then sharply subsides during the superpower period. There is significant instances of crises during WW2 (which is expected) but the increase during the Cold war period is notable, as in the steep decline in the following era.
There is an outlier phenomenon* (a legitimate one) in the cold war era which will continue to manifest itself through out this exploration and should be noted. That outlier is the influence of “Non-State Actors”** (eg: terrorist groups) and “Multi-State Actors” (alliances of nations) in the crisis activities. Since they are not specific sovereign states, they are coded as such, and their presence tends to overshadow the activities of the other single state actors.
NOTE: For highlighting purposes, observe the rise and fall of crisis frequency between the red and blue lines in the graphs.
4.3.2 Crisis Belligerents (Trigents)
- Major Trigger Actors (Belligerents with the highest crisis counts) Worldwide
| Country | Country | Country | Country |
|---|---|---|---|
| China | France | Germany | India |
| Iran | Iraq | Israel | Italy |
| Japan | Libya | North Korea | Russia |
| South Sudan | Turkey | UK | USA |
| Yugoslavia |
ANALYSIS:
- The same overall trends exist for the Crisis Belligerents (causal actors) as the actors (indicating a correlation) except that they are slightly muted. Even so, it is still worth noting.
4.3.3 Trigger of the Crisis
ANALYSIS:
Violent acts and political acts continue to dominate in roughly the same frequency proportions in each key era, except that their (as well as all of the others) frequency during the cold war is greatly pronounced.
Notice that in the superpower era, overall trigger activity reduces significantly (perhaps due to advanced military and nuclear arms proliferation).
4.3.4 Issue of the crisis
ANALYSIS:
- The dominant crisis issue across the eras is military-security followed by political-diplomatic. The same trend of a sharp increase during the cold war era is evident along with the subsequent sharp decline in the super power era.
4.3.5 Crisis Management I: Major Response to Crisis Trigger
ANALYSIS:
When it comes to the major responses of a crisis, the overall proportions and trends are the same between eras (sharp increases during the cold war with notable declines during the superpower era), but within the cold war era, the increases of political, non-violent military, and violent military are “significantly”" increased.
A quick look back into the various US-Russian tension levels, the introduction of nuclear weapons, and the struggles for international dominance between two emerging superpowers might lend understanding to the movements of those specific parameters.
One nuance that should be noted is the emergence of “Economic” response to crises as a viable alternative along with “Verbal” options in the superpower period. I wonder if the emergence of intertwined economies and advanced military weapons could account for the appearance of these?
4.3.6 Crisis Management II: Principal Coping Technique
ANALYSIS:
Not surprisingly, the combination of violence and negotiation (in it’s many forms) dominate principle coping techniques during all eras (as well as all of written history). Again, the frequencies of the instances are significantly higher in the cord war era, which is expected considering the history of the time period.
It should be noted that non-violent techniques also increased during the cold war indicating that in the mix of events, there were some cooler heads prevailing. The would make good sense, given the advancement and proliferation of weapons of mass distruction.
4.3.7 Crisis Outcome (Win, Lose, Draw)
ANALYSIS:
In the WW2 era, outcomes of victory or defeat dominate all other categories with compromise emerging as a distant third option.
In the Cold war era, this mix changes significantly with victories dominating the frequency chart. It is interesting to note that Stalemates rises significantly in this era along with compromises and defeats. I’d argue that this is also a function of the power struggles of that period between the US-USSR, the advancement of WMD’s, and jockying for geopolitical / economic dominance.
In the superpower era, the levels return to nearly of the WW2 era, but with a unique phenomenon of compromise surpassing all of the others for the first time. Again, it could be argued that because of the gravity of nuclear weapons, compromise is a much preferred alternative to the gravity of the other outcomes.
4.3.8 Power Status of Actor
ANALYSIS:
- The distribution of Actor power status makes sense given the eras in consideration, and that changes accordingly over time. In the WW2 era, there were a lot of small and great powers with a few middle powers emerging. After WW2, in the cold war era, that mix changed significantly both in gross numbers of small and middle powers and the increase of great powers, plus the emergence of super powers for the first time in history. Those same proportions continue in the superpower era but the frequency is greatly reduced, which is notable given the backdrop of history in that period.
4.3.9 Power Status of Threat
ANALYSIS:
The power status of threats during the WW2 era is what one would expect, dominated by great powers with an introduction of superpowers (at the end of the war).
The cold war era is different, the power status of the threats are dominated by small and middle powers with the great and super powers also rising in frequency, but less than the small and middle powers.
By the time the superpower era takes hold the overall frequency is reduced substantially, but of that, small and medium threat powers still dominate in frequency. I wonder if the nuclear stalemate among the maturing of superpowers has made the relevancy of conventional weapons and localized conflicts significant again?
4.3.10 Violence experiences by Crisis Actor
ANALYSIS:
- One trend that is notable is that in the WW2 and cold war eras, non-violence levels were higher than all others significantly. That does “not” continue in the superpower era, where there is equal frequencies of non-violence as other forms of violence. This is not a particularly encouraging macro trend to observe, even though the overall levels are lower in the superpower era than the cold war, one would hope to see non-violence levels higher than the others.
4.3.11 Political Regime of Crises Actor.
ANALYSIS:
- I find this particular analysis interesting in that democracies along with civil authoritarian regimes are most often embroiled with interstate crises as opposed to some form of a military government. That trend goes counter to the conventional narrative about democracies and civil authoritarian regimes, although historically it has definitely been the case that democracies and civil authorities are more prone to get involved in conflicts and then stay involved until the very end.
4.3.12 Gravest threat during crisis.
ANALYSIS:
The WW2 era distribution of gravest threats is predictible with the highest frequency of threats being to a states actual existence, along with corresponding territorial and grave damage threats.
The cold war era saw a dramatic rise in political, territorial, and influence threats, and a continued rise in grave damage. This (again) is understandable given the introduction of nuclear WMD’s and the capabilities of intercontinental delivery systems of those weapons move the crisis focus to other less grave areas.
Continuing the trend into the Superpower era, the overall levels drop significantly and we see a rise in limited military threats slightly overshadowing all of the others, which again seems typical of this time period.
4.3.13 Nuclear capability of Crisis Actor
ANALYSIS:
There are a few interesting aspects to exploring the nuclear capabilities across the three eras in question. The first being that in the cold war era there were a decent number of “foreseeable” actors who “might” develop nuclear weapons sometime in the future sometime. By the time the superpower era dawns, that figure is drastically reduced, which is understandable given the gravity of consequences given the enhanced distructive power of nuclear weapons.
The next (also understandable for the same reasons) is that the number of actors in “posession” of nuclear weapons has also drastically reduced.
Also note in the WW2 era, it was observed that some actors “could” develop nuclear weapons. Hindsight in this case is very interesting.
4.3.14 Alliance Patterns
ANALYSIS:
The latent dominant trend across all eras is for a high frequency of non-alligned (or neutral) parties. The alliance leaders have shrunk (to just a few) proportionally, but have kept those who are formally alligned with them (as historical treaties like NATO) would suggest.
One of the more interesting developments is that the informal (or loose) alliances have significantly reduced in the superpower era. I wonder if that is a function of the necessity of the informal alliances has dropped off, or their usefulness has.
5 Explorations of Actors (Participants)
This section explores the set of world actors who have the highest crisis frequences (as shown by the size of the pink circles on the map below.) by examining a similar set of attributes as the previous section.
5.1 Major Crisis Actors World Wide
5.2 Major Actors (those with the highest crisis counts) Worldwide
- “China”
- “Egypt”
- “France”
- “Germany”
- “Greece”
- “India”
- “Iraq
- “Israel
- “Japan”
- “Libya”
- “Pakistan
- “Russia
- “Syria”"
- “Turkey”
- “United Kingdom”
- “USA”
- “Serbia”
5.3 Exploration of Major Actors World Wide by ERA (WW2, Cold War, Super Power)
5.3.1 Actors and Trigger Acts of Crisis
ANALYSIS:
The trigger acts in the WW2 era are as expected, high frequency in the case of the UK, Germany, France, Japan, and Russia, while low for the US (a late entrant into the conflict).
The cold war in contrast, sees the US frequency skyrocketing dominated by poltiical, and violent (direct and indirect) acts. Russia, Israel and Egypt also have nearly identical levels of trigger acts with a mix of acts.
The superpower era sees all of those levels recede drastically across the board with the notables being Israel (which experiences exclusively violent trigger acts), and the USA (which still leads in trigger act frequencies with a relatively equal representation of trigger acts across the board).
5.3.2 Crisis Issue
ANALYSIS:
Not surprisingly Military-Security along with Political-Diplomatic issues dominate across the various eras and actors.
In the cold war era Political / Diplomatic issues actually begin to surpass military / security issues.
Interesting to note that in the superpower era, Turkey, Syria, Russia, Pakistan, Libya, Japan, Israel, India and Greece all have trigger issues completely comprised of military-security trigger issues - a fairly violent mix of concerns.
5.3.3 Power Status: Actor
ANALYSIS:
During the WW2 era there were 6 great powers and no super powers, with the UK, Russia, Japan, Germany, and France comprising the strongest great powers and the USA the weakest of the six.
In the wake of that conflict into the cold war, the US and Russia emerge as the lone pair of superpowers with the US far ahead in the power factor than any of the others.
Fast forward to the superpower era and we still see the US as the lone military superpower with Russia greatly weakened but still considered a superpower.
Again, the same patterns of heightened frequency during the cold war with a steep decline across the board into the superpower era is observed.
5.3.4 Power Status: Threat
ANALYSIS:
- Pretty much everyone of the major crisis actors threats were dominated by great powers during the WW2 era, whereas the threats in the cold war now involved superpowers and far more small and medium powers. In the super power era the threat posed were exclusively small and medium powers unless the actor was an adversary of the USA, where it was by default, a superpower.
5.3.5 Gravest Threat
ANALYSIS:
The WW2 era witnessed the gravest threats of territory, influence, grave damage, and a states very existence in various mixes for all of the key players.
Moving to the cold war era, influence of the international system dominates the gravest of threats across all of the major players.
Predictibly, in the superpower era, there is more limited military issues as the gravest threats in the mix than were previously observed. Again, this is indicative of the period in question.
5.3.6 Major Response
ANALYSIS:
- Interesting trends consistent in both the WW2 and Cold War eras is the presence of “Verbal” and “non-violent military” responses to the trigger acts of the belligerents. In the cold war, the incidences of non-violent responses of all sorts grew in frequency across the board whereas in the superpower era there was more instances of inaction and non-violent responses which suggest a decent measure of calmer heads when faced with interstate challenges.
5.3.7 Principal Technique
ANALYSIS:
WW2 witnessed violence as the principle coping technique across the spectrum of key players with negotiation seeing some frequency.
In the cold war however, there were a lot more diplomatic pressures used as a coping technique, but were offset by the amount of violence also incorporated into the era.
The superpower era saw a lot less “”direct" violence, and more “multiple” including violence techniques which would make sense given the hsitorical amount of intel and covert ops used by the superpowers during that era.
5.3.8 Violence Levels
ANALYSIS:
- The WW2 era witness it’s fair share of full scale war, but it also saw a lot of non-violence as well by the major players of that time period. The Cold War encompassed more frequency of non-violence, minor clashes, with just a few serious clashes and some bouts of full scale war. Interestingly (or predictably), the superpower era saw a dramatic reduction in counts across the board with all players, with only a few instances of serious clashes and full scale war.
5.3.9 Political Regime
ANALYSIS:
The single most obvious trend among the major players is for democracies and civil authorities to be embroiled in interstate crises more than any other political regime in existance, far more than any type of military shared government. This might be because the majority of the major entities were either democratic or civil authortarian with only a few military type governments available, but that trend spans three eras as well.
As an acid test I ran the plot again on the “entire” dataset (not shown), and that general trend held up fairly well. It has been suggested by many historians that democracies and civil authorities are not good international neighbors precisely because they tend to be frequently involved in interstate crises (and protracted conflicts), and tend to see them through until the very end.
5.3.10 Nuclear Capability
ANALYSIS:
The general trends concerning nuclear capabilities between eras continues with none existing during WW2, then in the Cold War the US, Russia, UK, Pakistan, Israel, India, France and China all come into posession of such WMDs. The US, UK, France, Russia, and China have all developed their own nuclear programs with Pakistan, Israel and India posessing nuclear weapons but not having developed them. The real concern at this point in history is the number of states who “might” be able to develop them in the future.
In the superpower era, we see the former developed and posession states continue their status, but with those who “might” develop them no longer having the capability to produce them. This is probably a function of the international non-proliferation treaties being enforced.
5.3.11 Alliance Patterns
ANALYSIS:
- There are two major trends observable with alliance patterns, the first is the blooming of the informal alliances of the cold war era among the major players (which completely disappears in the superpower era). The second is the consolidation of alliance leaders from six to two when perusing the period spanning from WW2 to the superpower era with the USA and Russia emerging as the remaining alliance leaders. There are a good number of non-alligned states though, which is notable.
5.3.12 Crisis Outcome
ANALYSIS:
Victory and defeats are the dominant crisis outcome during WW2, that much is expected. In the cold war era, victories still dominate but compromise and stalemate appear more frequently as outcomes while defeats are still a significant occurance. It could be argued that the presence of nuclear WMD have brought back the significance of conventional militaries (and localized conflicts) which could contribute to this trend.
In raw counts, during the cold war, the USA, Russia, and Israel see significant victories, but that trend evaporates almost entirely in the superpower era giving way to compromise and stalemates as the more frequently observed outcome. The overarching trend of a significant overall frequency reduction persists in this era.
6 Exlporations of USA crises during the Cold War ERA
- The USA is a unique case in the Cold War ERA and warrants a closer inspection of it’s behavior relative to the trigger actors (belligerents) of that period, particularly considering the backdrop of various technological advances, economic progress, globalization, and other key influential trends of that era.
6.1 USA Cold War Crisis Map
ANALYSIS:
- For the purposes of this particular exploration, it is beneficial to peruse the list of USA adversaries, the corresponding crisis, and its outcome below:
6.1.1 USA Cold War Crisis List: (Adversary, Crisis, Outcome)
## Cambodia
## crisname Crisis_Result
## 1 MAYAGUEZ Victory
##
## China
## crisname Crisis_Result
## 1 KOREAN WAR II Compromise
## 2 TAIWAN STRAIT I stalemate
## 3 TAIWAN STRAIT II Victory
##
## Cuba
## crisname Crisis_Result
## 1 BAY OF PIGS Defeat
##
## Vietnam
## crisname Crisis_Result
## 1 GULF OF TONKIN stalemate
## 2 VIETNAM SPRING OFF. Victory
## 3 INVASION OF CAMBODIA stalemate
## 4 VIETNAM PORTS MINING Compromise
## 5 CHRISTMAS BOMBING Compromise
##
## France
## crisname Crisis_Result
## 1 DIEN BIEN PHU Defeat
##
## Grenada
## crisname Crisis_Result
## 1 INVASION OF GRENADA Victory
##
## Guatemala
## crisname Crisis_Result
## 1 GUATEMALA Victory
##
## Iran
## crisname Crisis_Result
## 1 US HOSTAGES IN IRAN Compromise
## 2 IRAN NUCLEAR II Compromise
##
## Iraq
## crisname Crisis_Result
## 1 IRAQ/LEB. UPHEAVAL Victory
## 2 IRAQ DEPLOY./KUWAIT Victory
## 3 DESERT STRIKE Victory
## 4 UNSCOM I Victory
## 5 UNSCOM II Compromise
## 6 IRAQ REGIME CHANGE Victory
##
## Jordan
## crisname Crisis_Result
## 1 BLACK SEPTEMBER Victory
##
## Libya
## crisname Crisis_Result
## 1 LIBYAN JETS Victory
## 2 LIBYAN CIVILWAR Victory
##
## MULTI_STATE_ACTOR
## crisname Crisis_Result
## 1 KOREAN WAR III Victory
## 2 OCTOBER-YOM KIPPUR WAR Victory
## 3 GULF WAR Victory
##
## NON_STATE_ACTOR
## crisname Crisis_Result
## 1 CHINA CIVIL WAR Defeat
## 2 PATHET LAO OFFENSIVE Victory
## 3 VIET CONG ATTACK Victory
## 4 NAM THA Victory
## 5 CONGO II Victory
## 6 PLEIKU stalemate
## 7 DOMINICAN INTERVENTN. Victory
## 8 TET OFFENSIVE Defeat
## 9 WAR IN ANGOLA Defeat
## 10 SHABA II Victory
## 11 GULF OF SYRTE II Victory
## 12 HAITI MIL. REGIME Victory
## 13 US EMBASSY BOMBINGS Compromise
## 14 AFGHANISTAN/US Victory
##
## Panama
## crisname Crisis_Result
## 1 PANAMA FLAG Compromise
## 2 INVASION OF PANAMA Victory
##
## Korea
## crisname Crisis_Result
## 1 KOREAN WAR I Victory
## 2 PUEBLO Defeat
## 3 EC-121 SPY PLANE stalemate
## 4 POPLAR TREE Compromise
## 5 N. KOREA NUCLEAR I Compromise
## 6 N. KOREA NUCLEAR II Compromise
## 7 N. KOREA NUCLEAR III Compromise
## 8 N. KOREA NUCLEAR IV-SATELLITE LAUNCH stalemate
## 9 N. KOREA NUCLEAR V stalemate
##
## Russia
## crisname Crisis_Result
## 1 AZERBAIJAN Victory
## 2 TURKISH STRAITS Victory
## 3 BERLIN BLOCKADE Victory
## 4 SUEZ NATN.-WAR Victory
## 5 BERLIN DEADLINE stalemate
## 6 BERLIN WALL Compromise
## 7 CUBAN MISSILES Victory
## 8 SIX DAY WAR Victory
## 9 CIENFUEGOS SUB. BASE Victory
## 10 AFGHANISTAN INVASION stalemate
## 11 NICARAGUA MIG-21S stalemate
##
## Syria
## crisname Crisis_Result
## 1 SYRIA/TURKEY CONFRNT. Victory
## 2 SYRIA CHEMICAL WEAPONS Compromise
##
## United Kingdom
## crisname Crisis_Result
## 1 TRUMAN DOCTRINE Victory
## 6.2 Exploration: USA vs “Others” - Cold War ERA
NOTE: This exploration will frame the USA vs “Other” major crisis players during the Cold War Era (which should not be confused with their “adversaries”).
NOTE: It is important to highlight the significant difference in the “Counts” of the USA compared to “others”, it is roughly double, which is quite significant. The USA was significantly more involved in the Cold war than the other major players of that era.
6.2.1 Crisis Triggering Act
ANALYSIS:
- Roughly half of the acts triggering crisis for the USA during the cold war were non-violent in nature, the other half incorporated some form of violence. This overall trend also holds for the other actors of that period as well regardless of their crisis frequencies.
6.2.2 Crisis Issue
ANALYSIS:
- During the cold war, the dominant crisis issues for the USA were security and diplomatic in nature, with the emergence of economic as a small but viable crisis issue. Diplomatic crisis were by far the largest in frequency and is mirrored by the key USA adversary - Russia. Of the other major players in this era, those in the middle east displayed domonant crisis issues of “Military-Security” (with the notable exception of China), and lesser diplomatic issues. This is reflected in the history of the middle east during that time period. The last two notable item were (1) the emergency of “cultural” as a legitimate crisis issue in the case of Yugoslovia , Turkey, Israel, Iraq, Egypt and China, and (2) the pronounced “economic” crisis frequencies of the UK, Syria, Greece and Egypt of that period.
6.2.3 Actor (USA) Power Status
ANALYSIS:
- The power status of the USA is as expected for that period, as is its adversary Russia. Most of the other players are considered great powers with just a few small powers represented. There are however, a decent representation of middle powers in the mix most notable Pakistan.
6.2.4 Threat Power Status
ANALYSIS:
- Roughly half of the threats counts involved with the USA were small to medium powers and the other half were great or super powers. The only other players to witness such big/small proportions in the strength of their threats were the UK, Israel, India, Greece, France and Egypt. The rest had mixes containing more great and super powers as threats than small and medium powers.
6.2.5 Gravest Threat
ANALYSIS:
- The gravest threat to the US by far was Territorial in nature with Influencing the International System accounting for a small percent of the overall counts. The “other” players all had instances of Economic, Military, Political, as well as territorial and influence representations as gravest threats with just Israel and Pakistan fearing for their very existence at some point in that era.
It is worth pointing out that the dominant representations in the gravest threat categories of the “others” were “international influence” and “territorial”.
6.2.6 Major Response
ANALYSIS:
The majority of the primary responses of the USA were of non-violent nature and is roughly the same for Russia except that Russia’s “mix” of responses is quite different in that Russia did not use economic or “other” non-violent responses.
The other major players of this era also employed non-violence about roughly half the time as well.
The key observations for the ‘others’ category were that (1) very little economic (and other) responses were employed, (2) a great deal of political and non-violent military responses were leveraged as major responses, and (3) there were a lot more variations of violent responses employed across the board particularly with reguard to the middle east players, France and China.
6.2.7 Principle Coping Technique
ANALYSIS:
- Again, the non-violent techniques are the dominant choices of both the US and Russia during the cold war era, which, given the historical context, is understandable. On the “others” side, the only coutnries to have similar mixes of violence to non-violence techniques were Yoguslovia, the UK, Turkey and Greece. The rest employed more violence than non-violence as a principle technique to their respective crises.
6.2.8 Violence
ANALYSIS:
About half time the USA witnessed no violence, and of the haf in which it did see some, about half of that were minor clashes, whereas the other half of that was serious clashes or full scale war. Russia on the other hand enjoyed a lot more non-violence, an equal number of minor clashes, some serious clashe, and no full scale war at all.
In the case of the “others”, everyone except Yugoslavia, and Russia has experienced a lot of full scale war during that period, along with significant amounts of serious clashes.
6.2.9 Political Regime
ANALYSIS:
The USA being a Democracy and Russia being a Civil Authority is well established, as is the latent trend for both democracies and civil authorities to be enbroiled in a large number of interstate crises historically.
Interesting to note is the clustering of military political regimes in the middle east: (Syria, Pakistan, Libya, Iraq, and Egypt).
6.2.10 Nuclear Capabilities
ANALYSIS:
It is clear that both the USA and Russia had both “posession of” and a “developed” nuclear weapons program. The only others to do so were: China, France and the UK. Israel had posession of nuclear weapons but did not have a developed nuclear weapons program.
It was widely believed that Libya, Israel, Iraq, and Egypt “could have develop” nuclear weapons sometime in the future, but they currently did not.
One note is that both Pakistan and India did make entrances onto the nuclear scene independently later on and are in legitimate posession of a nuclear weapons program.
6.2.11 Alliance Patterns
ANALYSIS:
- The two givens of this era are that both the USA and Russia are alliance leaders with France emerging on the scene later on. Everyone else has formal (or informal) alliances with the leader and only a few instances of neutrality are observed with the notable exception of India.
6.2.12 Crisis Outcome
ANALYSIS:
More than half of the crisis outcomes for the USA ended in a “Victory”, and the remaining half was roughly equally divided between compromise, stalmalmate, and defeat. A similar observation can be made of the Russian experience in terms of proportions of outcomes.
There was a consistent presence of compromise and stalemate during this period across almost all the players, which is notable.
Other notable exceptions are Israel, who has never known defeat, only victory, compromise and stalemate during this period, and Iraq which only experienced stalemate and defeat.
7 Final Plots
This section is dedicated to exploring interesting multidimensional interactions of selected attributes, and other interesting phenomenon of the dataset using Polar, Mosaic and Alluvial plots.
The topics explored are (1) the “Power Difference” between the Actor and Adversary using Polar Coordinate Plotting, (2) the relationship between “Nuclear Capability, Actor Power Status and Crisis outcome” using Mosaic Plotting, and (3) the “Many Pathways to Victory (and defeat)”, using Alluvial diagrams.
For the convenience of the reader, I’ll provide a brief refresher on the background and interpretation of the data and/or plot types when it is useful.
7.0.1 Exploration of Power Difference: USA vs Adversaries
- Refresher: A power score was determined for each crisis actor and its principal adversary on the basis of the total of six separate scores measuring: (1) size of population, (2) GNP, (3) territorial size, (4) alliance capability, (5) military expenditure, and (6) nuclear capability, at the onset of the crisis. The power of a crisis actor and the power available to it from tight alliance partners was then compared to that of its principal adversary to create a final power discrepancy score.
ANALYSIS:
Note: Both the direct and inverse relationships of the power differentials are examined to highlight the nuances.
USA vs Adversaries: The most striking observation of the USA vs its adversaries in the cold war era is that the USA nearly always enjoyed a huge power advantage as evidenced by the small dotted circle line which is the Power Parity threshold line. The rest of the radius of the circle is the positive power difference.
Adversaries vs USA: Russia obviously has achieved some measure of a positive power differential, but aside from them, only China and North Korea actually approach power “significance” with the USA. Others like Libya, Vietnam, and Panama actually have a “large negative power differential” relative to the USA, whereas the remaining group has only achieved a power differencial of “one”.
7.0.2 Exploration of Nuclear Capability and Power Status on Crisis Outcome
Refresher: Mosaic plots visualize multivariate categorical data very well, and represent variable counts as rectangular areas on the plot. Each categorical variables goes to one edge of the square, which is subdivided by its labels into rectangles. The size of the rectangles is proportional to frequency. Mosaic plots represent the data as is, and does not make any attempt to “generalize to the full population”. To make inferences about the population, we need to provide measures of statistical significance. Inspired by the chi-square test, we can define Pearson residuals which measure the departure of each cell from “independence”. The units are in “standard deviations”“, so a residual greater than 2 or less than -2 represents a departure significant at the 95% level. When testing for statistical significance: Grey means that data are consistent with (you cannot reject the hypothesis of) variable independence, Blue means that a positive association exists (more observations present than expected under the null model), and Red means a negative association esists (fewer observations than would be expected under the null model).
NOTE: This exploration comprises the entire ICB dataset, there is no subsetting of the data.
NOTE: Unfortunately R-Markdown shrinks this graph, which makes it difficult to fully appreciate the nuances, but there is a lot of complex variable relationship information there which could still be brought to light.
NOTE: On the bottom right hand side of this plot, the illegible text should reflect the 4 outcomes: (Defeat, Stalemante, Compromise, and Victory) for each of the four labels of Nuclear Capability (Developed, Has, Maybe, and None).
ANALYSIS:
First: If an actor is small and does not have nuclear capabilities then it is not clear at all that it could “ever” eperience stalemate, compromise, or victory in a interstate crisis.
Second: If an actor has developed nuclear weapons its probably a superpower, and will generally experience victory in interstate crisis situations, but it could also experience stalemate, compromise or defeat.
Third: If an actor has posession of nuclear weapons then its most likely a great or super power and can expect to experience victory and stalmate as a “great power”, and victory if it is a “superpower”.
Forth: If an actor might develop nuclear weapons in the future, but does not have them currently, it has better chances for vistory, but can definitely experience defeat as well.
7.0.3 Exploration of the “Many Pathways to Victory (and Defeat)”
- NOTE: These Alluvial diagrams explores how (1) Alliance Patterns, (2) Actor Power Status, and (3) Nuclear Capabilities affect the (4) Outcome (“Result”) of the crisis in question. The “pathways” to victory are marked in red to make the identification of the possibilities more clear.
Alliance Patterns are comprised of (None, Leader, Informal, and Formal) and are the first set of blocks on the left hand side of the diagrams with the subheading Alliances
Actor Power Status is comprised of (Small, Middle, Great and Super), and are the second set of blocks in the diagram with the subheading of Power_Status.
Nuclear Capabilities are comprised of (None, Maybe, Has, and Developed), and are the third set of blocks in the diagram with a subheading of Nuclear.
Outcome is comprised of (Victory, Stalemate, Compromise, Defeat and NA), and is the forth set of blocks in the diagram with a subheading of Result
7.0.4 Pathways to “Victory”
7.0.5 Pathways to “Defeat”
Refresher: Alluvial diagrams were originally developed to visualize structural change in large complex networks, but they can be used to visualize any type of change in group composition between states and include statistical information to reveal significant change. Alluvial diagrams highlight important structural changes that can be further emphasized by color, and make identification of major transitions easy. We will use them to highlight the many pathways to victory (and defeat) during interstate crisis analysis.
NOTE: This exploration comprises the entire ICB dataset, there is no subsetting of the data.
ANALYSIS:
What is truly fascinating about the things this type of diagram highlights is those aspects of sovereign state crises that go in stark against “conventional” wisdom. For example: it is quite possible to be a state that with no alliances, small power, with no nuclear capabilities, and still achieve victories in conflicts. Conversely, it is also quite possible to be an alliance leader, a superpower, with a developed nuclear weapons program, and still suffer defeat.
The other interesting notable from these sets of diagrams are the widths (frequencies) and directions of the pathways to the result as experienced through each variable in the analysis. The magnitudes of each variables influence on the outcome is can be clearly seen. For example, defeat can quite literally come from anywhere.
Walk through - On the first plot, as one follows the red lines, the following inferences can be made as to the many pathways to victory of a crisis:
- No alliances, Small Power, No Nuclear Capabilities.
- No alliances, Medium Power, No Nuclear Capabilities.
- No alliances, Great Power, No Nuclear Capabilities.
- Informal alliances, Small Power, No Nuclear Capabilities.
- Informal alliances, Medium Power, No Nuclear Capabilities.
- Informal alliances, Great Power, (maybe, has, or developed) nuclear capabilities.
- Formal alliances, Small Power, No Nuclear Capabilities.
- Formal alliances, Medium Power, No Nuclear Capabilities.
- Formal alliances, Great Power, (maybe, has, or developed) nuclear capabilities.
- Alliance Leader, Great Power, (maybe, has, or developed) nuclear capabilities.
- Alliance Leader, Super Power, (maybe, has, or developed) nuclear capabilities.
8 Reflections
FUll Disclosure - I currently work in the defense industry, and also have a passion for world history, particularly the behavior of sovereign states during conflicts, so to have the chance to analyze a dataset such as this was a real joy.
I encountered various “novelties” in this dataset immediately, the first being that nearly all of the data is either categeorical or ordinal and not the continuous data that we have seen so much of during the trainning phase of our studies. That presented some interesting challenges in terms of making the right choices for plotting vehicles since categorical variables restrict much of those choices. Second, nearly all of this data was coded numerically, and that necessitated some transformation of the data into a form which could more readily be used by the Mosaic and Alluvial plots. Third, constructing plots of this sort that captured the complex relationships and yet maintained enough elegance to make conveying the key points possible was also challenging. Forth, adhering to the spirit of “letting the data speak for itself” was difficult as it required the imposing of an overall structure that would guide the reader without getting in the way of the story. Lastly, the dataset contains 96 variables, and for the sake of brevity I only used a dozen or so very general ones in this exploration. Many of the variables not chosen contain rich information about very narrowly defined topics within the dataset which would be an entire study in their own right.
I did engage in a great deal of “developmental detours” during this project, particularly to comprehend what EDA really was across the board in data science, to understand the proper “care and handling” of categorical and ordinal variables, the intricacies of ggplot2, mosaic, alluvial diagrams, network maps, and spatial mappings in the R world, the behavioral patterns and best practices of the ggplot2 system, and a host of other adjacent fields to EDA and R which appeared useful. The price of these developmental detours was time expended to stop moving forward with the project, explore a complementary field, return to the project, incorporate that individual additions and then repeat. In truth, that made this project take up the most time of any project thus far in the program for me.
To enrich the analysis moving forward, I’d begin to make more data transformation to begin capturing more of the nuances of the data in route to uncovering the gravity of the decisions that the leaders of those sovereign states mades. I’d also begin to include more of the 96 variables (instead of just 12) in the explorations to add richness to the mix or explore more narrowly defined topics.
The success I found were in the quality of the data, which is a direct reflection of the great care and immense effort it took the ICB researchers to compile, vet, and organize that data. It made constructing a structure such that the data could readily “tell it’s own story” a reachable goal, which isn’t often the case.
I also found success in being able to procure useful R libraries such as the Alluvial library, the SQLDF, and others. They made some complex tasks much easier and extracting / visualizing interesting combinations of variables possible.
In the end I’m quite satisified with the results of this project, which for me are first and foremost the procurement of the EDA skills and implementation expierience garnered from the process, I felt like I procured a fine return on investment from this endeavor.