Historical Method and Causal Inference: How We Know What We Think We Know About International Politics

LESSON METADATA

Note on self-containment: Every methodological concept, example, and debate you need is explained in this document. You do not need to consult outside readings. The “Further Reading” section is optional.

SECTION 0 — Why This Lesson Exists

Lessons 1 and 2 gave you the conceptual instruments — puzzles, levels, power, perspectives, law. But those instruments are only as good as the evidence you feed them. The single most common failure in IR analysis is not bad theory; it is bad method — selecting cases that confirm your argument, confusing correlation with causation, reading history backwards from the present, or treating a single case as proof of a general law.

This lesson teaches you how to know what you think you know. It covers four interlocking skills:

1. Historical method: How to read sources critically, how to avoid presentism and teleology, and how to use archives and historiography without being captured by them.
2. Causal inference: How to distinguish causation from correlation, how to use process tracing and comparative method, and how to construct and evaluate counterfactuals.
3. Selection bias: Why the cases you choose to study determine what you can conclude, and how the most common error in IR — selecting on the dependent variable — produces false confidence.
4. Perspective on method: How African, Chinese, Russian, Western, and UN traditions approach historical evidence differently, and why those differences matter for adjudication.

By the end, you will be able to take any causal claim in IR — “NATO caused the war,” “colonialism caused state fragility,” “sanctions caused the JCPOA collapse” — and subject it to disciplined evaluation: what is the evidence, what is the mechanism, what would falsify it, and what would we expect to see if it were false?

SECTION 1 — PRE-CLASS PRIMING

1.1 Retrieval warm-up (from memory)

1. From L1: Name the four levels of analysis and give one cause located at each level for a single event.
2. From L1: Distinguish the first, second, and third faces of power with one observable implication of each.
3. From L2: What did the ICJ hold in Nicaragua about the relationship between Article 2(4) and customary international law?
4. From L2: What is the three-layer reading of a legal document, and apply it in one sentence to Article 4(h) of the AU Constitutive Act.

Check against the previous lessons before proceeding.

1.2 Orientation questions

Hold these in mind:

• What makes a historical explanation causal rather than merely descriptive?
• Can we ever know what would have happened if history had gone differently?
• Why do historians and social scientists often disagree about the same events?
• Whose archives survive, and whose do not? What does that mean for what we can know?

1.3 Vocabulary pre-teach

Provisional definitions; revise by the end:

Process tracing; counterfactual; selection on the dependent variable; endogeneity; presentism; teleology; historiography; primary source; secondary source; source criticism; comparative method; most-similar systems design; most-different systems design; congruence method; archival grain.

SECTION 2 — LESSON CORE

2.1 Learning objectives

By the end of this lesson, you can:

1. Distinguish causal from correlational claims, and identify the mechanism that links cause to effect.
2. Apply process tracing, comparative method, and counterfactual analysis to a single case and explain what each method contributes.
3. Identify selection bias, endogeneity, and presentism in existing IR explanations, and explain how each distorts inference.
4. Read a primary source and a secondary source critically, using external and internal source criticism, and identify what the archive silences.
5. Reconstruct how the African, Chinese, Russian, Western, and UN traditions approach historical method differently, and adjudicate among them on a specific case.

2.2 KEY IDEA 1 — Chronology is not explanation: causal claims require mechanisms and counterfactuals

The claim: A causal claim is not a story about what happened. It is a claim that if the cause had been absent or different, the effect would have been absent or different — and that we can specify the mechanism that links cause to effect.

The problem. Consider this claim: “The Treaty of Versailles caused the Second World War.” It is a chronological sequence — Versailles (1919), then the Second World War (1939). But a sequence is not a cause. To make it a causal claim, you must specify:

• The mechanism: How did Versailles produce the war? Through German resentment and revisionism? Through the punitive reparations that crippled the Weimar economy? Through the territorial settlements that left German minorities outside the Reich? Through the disarmament clauses that made German rearmament a symbol of sovereignty?
• The counterfactual: Would the war have happened without Versailles? If the answer is “probably yes, because the structural conditions of multipolarity and the rise of Nazi ideology would have produced a war anyway,” then Versailles is a contributing condition, not a sufficient cause.
• The scope condition: Under what conditions does a punitive peace produce a revanchist war? The answer must specify what distinguishes Versailles from other punitive settlements that did not produce wars (e.g., the settlement of 1815 after Napoleon, which produced a century of relative peace).

The mechanism requirement. A causal claim without a mechanism is a black box. “Sanctions caused the collapse of the JCPOA” tells you that two things happened; it does not tell you how one produced the other. The mechanism might be:

• Economic pressure: Sanctions crippled the Iranian economy, making the regime unwilling to comply with nuclear restrictions.
• Political signaling: Sanctions signaled that the West would not deliver the economic benefits promised under the deal, undermining the domestic coalition that supported it.
• Reciprocal escalation: Sanctions triggered Iranian counter-escalation (nuclear enrichment, support for proxies), which triggered further sanctions, in a spiral that neither side could control.

Each mechanism implies a different counterfactual and a different observable implication. The economic-pressure mechanism implies that removing sanctions would have stabilized the deal. The signaling mechanism implies that credible commitments would have stabilized it. The reciprocal-escalation mechanism implies that de-escalation on both sides would have been necessary.

The counterfactual requirement. A causal claim is a claim about a comparison — between the world we observe and a world in which the cause is absent. This comparison is always counterfactual (we cannot observe it) and always theory-laden (we need a theory to tell us what the counterfactual world would look like). The discipline of constructing counterfactuals is what distinguishes causal analysis from storytelling.

Scope condition: Counterfactuals are only as good as the theory that generates them. A counterfactual that violates everything we know about the world is not a useful comparison; it is a fantasy. The standard is “minimal rewrite” — change only the cause in question, hold everything else constant, and ask what would plausibly follow.

Observable implication: If a causal claim is well-specified, you should be able to identify (a) the mechanism, (b) the counterfactual, and (c) the observable implications that would differ in the counterfactual world. If you cannot do all three, the claim is not yet a causal claim.

2.3 KEY IDEA 2 — Process tracing, comparative method, and counterfactuals are complementary tools, not rivals

The claim: IR has three main methods for establishing causal claims: process tracing (within-case), comparative method (across-case), and counterfactual analysis (against-case). They are not competitors; they answer different questions, and the strongest analyses use more than one.

Process tracing. Process tracing is the systematic examination of the sequence of events within a single case to identify the causal mechanism that links an independent variable to an outcome. It asks: Did the mechanism that the theory predicts actually operate in this case?

The method was formalized by Alexander George and Andrew Bennett, and is now the dominant qualitative method in IR. It involves:

1. Specifying the mechanism that the theory predicts (e.g., “NATO enlargement → Russian perception of existential threat → decision to invade”).
2. Identifying observable implications at each step (e.g., Russian military planning documents, Putin’s speeches, diplomatic exchanges).
3. Testing the evidence against the mechanism (e.g., did Russian planners actually cite NATO enlargement as a threat? Did Putin’s speeches escalate in response to NATO decisions?).
4. Ruling out alternative mechanisms (e.g., was the invasion driven by domestic politics, economic interests, or Putin’s personal ideology rather than NATO enlargement?).

Process tracing is especially powerful for testing mechanisms — it can show whether the causal process that a theory predicts actually occurred. It is less powerful for estimating causal effects — it cannot tell you how much of the outcome was due to the mechanism in question versus other factors.

The comparative method. The comparative method examines variation across cases to identify causal relationships. It uses John Stuart Mill’s methods of agreement and difference, formalized by Arend Lijphart and others:

• Most-similar systems design: Compare cases that are similar on most variables but differ on the outcome. If they also differ on one independent variable, that variable may be the cause. Example: Compare two African states with similar colonial legacies, similar ethnic compositions, and similar resource endowments, where one experienced a coup and the other did not. If the difference is explained by a single variable (e.g., the presence of a foreign military base), that variable may be causal.
• Most-different systems design: Compare cases that are different on most variables but share the same outcome. If they also share one independent variable, that variable may be the cause. Example: Compare the collapse of the Soviet Union and the collapse of the Ottoman Empire — very different contexts, but both were multi-ethnic empires that collapsed under the pressure of nationalism and great-power competition.

The comparative method is powerful for identifying associations and for ruling out rival explanations (by showing that a variable that is present in one case and absent in another does or does not co-vary with the outcome). It is less powerful for establishing mechanisms — it can show that two variables are associated without showing how the association works.

Counterfactual analysis. Counterfactual analysis constructs an explicit alternative world in which the cause is absent or different, and asks what would plausibly follow. It is the method that makes causation observable by implication rather than by direct observation.

The classic statement is James Fearon’s “Counterfactuals and Hypothesis Testing in Political Science” (1991), which distinguishes causal effects (the difference between the observed outcome and the counterfactual outcome) from causal mechanisms (the process that produces the outcome). Fearon argues that counterfactuals are unavoidable in causal inference, but they must be disciplined by theory and evidence.

The discipline of counterfactual construction involves:

1. Specifying the counterfactual precisely: what is the cause being removed or changed, and what is held constant?
2. Deriving implications from theory: what would we expect to see in the counterfactual world?
3. Testing those implications against evidence: are there historical cases that approximate the counterfactual? Are there observable traces of the counterfactual world in the actual world?

The complementarity. The strongest causal analyses combine all three:

• Process tracing establishes that the mechanism operated in the case.
• Comparative method establishes that the cause and effect co-vary across cases.
• Counterfactual analysis establishes that the effect would not have occurred without the cause.

Worked illustration — the end of the Cold War. The peaceful end of the Cold War (1989–1991) is one of the most studied cases in IR. How would each method approach it?

• Process tracing: Examine the sequence of decisions by Gorbachev, the East European revolutions, and the Soviet withdrawal. Did Gorbachev’s ideas (new thinking, common European home) cause the peaceful transition? Did the material pressures of Soviet decline cause it? Did the transnational diffusion of norms cause it? The method would trace the mechanism through archival evidence, memoirs, and interviews.
• Comparative method: Compare the end of the Cold War with other hegemonic transitions — the British decline, the collapse of the Ottoman Empire, the end of the Napoleonic wars. What distinguishes the peaceful Cold War transition from transitions that produced war? The method would identify the variable that explains the difference.
• Counterfactual analysis: What if Gorbachev had responded to the East European revolutions with force, as Brezhnev had in 1968? Would the Soviet Union have survived? Would the transition have been violent? The method would construct the counterfactual and test it against evidence (e.g., the Soviet leadership’s debates, the military’s reluctance to intervene).

None of these methods alone gives a complete answer. Together, they provide a more robust causal claim.

2.4 KEY IDEA 3 — Selection bias is the most common error in IR, and it is usually invisible to the analyst

The claim: The cases you choose to study determine what you can conclude. If you select cases on the basis of the outcome you are trying to explain, you will systematically overestimate the importance of whatever variable is present in those cases — because you have excluded the cases where the outcome did not occur.

The problem. This is called selection on the dependent variable, and it is the most common methodological error in IR. The classic statement is by Barbara Geddes, who showed that studies of economic development that selected only successful developers would conclude that “the policies of successful developers cause development” — without examining whether unsuccessful developers also adopted those policies.

In IR, the problem takes many forms:

• Studying only wars: If you study only wars, you will find that arms races, alliances, and territorial disputes are present in all of them. But you will not know whether those factors are also present in peaceful cases — and if they are, they do not explain war.
• Studying only successful coups: If you study only coups that succeeded, you will conclude that certain conditions (economic crisis, military factionalism) cause coups. But you will not know whether those conditions are also present in cases where coups were attempted and failed.
• Studying only cases where a theory applies: If you study only cases where realism seems to explain behavior, you will conclude that realism is a powerful theory. But you will not know whether it also explains the cases where it seems not to apply — and if it does not, its scope is narrower than claimed.

The mechanism. Selection bias operates through truncation — the cases you exclude are not random. They are systematically different from the cases you include, in ways that are correlated with both the cause and the effect. The result is biased inference: the relationship you observe in your sample is not the relationship in the population.

How to correct for it. There are three main strategies:

1. Select on the independent variable: Choose cases that vary on the cause you are interested in, and examine whether the outcome varies accordingly. This is what most-similar and most-different systems designs do.
2. Select on the outcome and then examine the counterfactual: If you must select on the outcome (e.g., you are studying wars because wars are what you care about), explicitly examine the cases where the outcome did not occur, and ask whether the cause was present or absent.
3. Use large-N data: If you can measure the cause and the outcome across many cases, you can estimate the relationship without selecting on either. This is what quantitative IR does, and it is one reason why quantitative methods are valuable even for qualitative researchers.

Scope condition: Selection bias is a problem for causal inference, not for description or interpretation. If you are trying to describe how a particular war unfolded, selection bias is irrelevant. If you are trying to explain why wars occur, it is fatal.

Observable implication: If a study is vulnerable to selection bias, you should be able to identify the cases it excludes and show that those cases would change the conclusion. If the excluded cases are similar to the included cases on the cause but different on the outcome, the study’s conclusion is suspect.

2.5 KEY IDEA 4 — The five perspectives approach historical evidence differently, and those differences are themselves evidence

The claim: Historical method is not a neutral set of techniques. Different traditions approach the archive, the source, and the narrative with different assumptions about what counts as evidence, whose voice is authoritative, and what the purpose of historical reconstruction is. These differences are not obstacles to analysis; they are data about how different traditions understand the international.

2.5.1 The African perspective (primary)

Core commitments. African historiography begins from the recognition that the archive is a colonial institution. Michel-Rolph Trouillot’s Silencing the Past (1995) argues that history is produced through a process of silencing — the production of historical narratives always involves the exclusion of certain voices, and the silences are not accidents but constitutive features of the narrative. Ann Laura Stoler’s Along the Archival Grain (2009) shows that colonial archives are not neutral records but epistemic machines that classified, categorized, and governed colonized peoples — and that reading “against the grain” of the archive can recover the agency that the archive was designed to erase.

How this perspective approaches historical method. African historiography emphasizes:

• Oral history and non-archival sources: Written archives are disproportionately colonial; oral traditions, songs, and material culture carry historical knowledge that the archive excludes.
• Reading against the grain: The colonial archive records the colonizer’s categories. Reading against the grain means asking what the colonized were doing that the archive did not record, and what the archive’s silences reveal about power.
• History from below: The experiences of ordinary people, not only elites, are central to understanding how international order was experienced and contested.
• The coloniality of knowledge: The categories of social science — “state,” “sovereignty,” “development” — are themselves products of colonial history. Using them uncritically reproduces colonial ways of knowing.

What this perspective contributes to causal inference. It insists that the evidence you have is shaped by power. The absence of evidence is not evidence of absence; it may be evidence of silencing. It also insists that the categories you use to organize evidence are themselves historical products, not neutral containers.

What it occludes: It can underplay the value of quantitative evidence and formal methods; it can romanticize the “voice” of the colonized; it can make causal inference difficult by emphasizing the uniqueness of each historical experience.

2.5.2 The Chinese perspective

Core commitments. Chinese historical method is shaped by a civilizational and cyclical understanding of history. The “century of humiliation” (1839–1949) is the central historical narrative in Chinese IR: the period of foreign invasion, unequal treaties, and territorial loss that is understood as the foundational trauma of modern China. This narrative is not merely descriptive; it is prescriptive — it defines what China must never allow to happen again, and it frames contemporary sovereignty and non-interference as defenses against a repeat of that history.

How this perspective approaches historical method. Chinese historiography emphasizes:

• Continuity and civilizational time: Chinese history is understood in terms of dynastic cycles and long continuities, not in terms of the sharp ruptures that Western historiography emphasizes.
• The “century of humiliation” as method: The narrative is used to interpret contemporary events — NATO enlargement, sanctions, human-rights criticism are all read through the lens of the humiliation narrative.
• Relational historiography: History is understood as the unfolding of relationships and hierarchies, not as the interaction of discrete sovereign states.
• The state as the subject of history: Chinese historiography tends to treat the state as the primary agent, with less attention to non-state actors and civil society than Western historiography.

What this perspective contributes to causal inference. It insists that historical memory is a causal variable — the way states remember the past shapes how they interpret the present. It also insists that long-run continuities matter more than the short-run events that Western IR tends to study.

What it occludes: It can underplay the diversity of Chinese historical experience; it can treat the state’s narrative as the only narrative; it can make it difficult to distinguish between historical memory as a cause and historical memory as a justification.

2.5.3 The Russian perspective

Core commitments. Russian historical method is shaped by a civilizational and grievance-based narrative. The repeated invasions of Russia from the West (Napoleon 1812, Hitler 1941) are the central historical reference points. The collapse of the Soviet Union is understood as a catastrophe — the “greatest geopolitical catastrophe of the twentieth century,” in Putin’s words — and the post-Cold War settlement is understood as a humiliation imposed by the West.

How this perspective approaches historical method. Russian historiography emphasizes:

• Geopolitical determinism: Geography and the absence of natural defenses are treated as the fundamental drivers of Russian history.
• The “whataboutism” method: Russian official discourse frequently responds to criticism by citing Western historical crimes (colonialism, slavery, the Iraq war) — not as a defense but as a way of delegitimizing Western moral authority.
• Civilizational narrative: Russia is understood as a distinct civilization with its own historical path, not as a backward version of Europe.
• The state as the guardian of history: The Russian state has increasingly regulated historical memory (e.g., laws against “falsifying history,” the rehabilitation of Stalin).

What this perspective contributes to causal inference. It insists that historical grievance and humiliation are causal variables — states that feel historically wronged may behave differently from states that do not. It also insists that the historical narrative itself is a political resource — the way a state narrates its past is a form of power.

What it occludes: It can naturalize aggression by framing it as a response to historical grievance; it can treat the state’s narrative as the only narrative; it can make it difficult to distinguish between grievance as a cause and grievance as a justification.

2.5.4 The Western perspective

Core commitments. Western historical method is shaped by the scientific aspiration — the belief that historical and social phenomena can be studied with the same rigor as natural phenomena. This aspiration produced:

• Positivist historiography: The search for general laws and causal regularities, using comparative method and quantitative data.
• Archival research: The systematic use of primary documents, with a strong emphasis on source criticism and the verification of claims.
• The “linguistic turn”: The recognition that historical narratives are constructed, and that the historian’s own position shapes the narrative.
• The “history from below” tradition: The recovery of the experiences of ordinary people, marginalized groups, and colonized peoples — a tradition that overlaps with African historiography but emerged independently within Western historiography.

What this perspective contributes to causal inference. It provides the methodological toolkit — process tracing, comparative method, counterfactual analysis, source criticism — that the discipline uses. It also provides the self-critique — the recognition that the toolkit is not neutral and that the historian’s position matters.

What it occludes: It can treat Western historical experience as universal; it can underplay the colonial genealogy of its own methods; it can be overly confident in the possibility of objective knowledge.

2.5.5 The UN perspective

Core commitments. The UN approaches historical method through the lens of institutional memory and legal precedent. The UN’s historical work is not primarily academic; it is operational — the documentation of conflicts, the recording of resolutions, and the accumulation of precedent for future action.

How this perspective approaches historical method. The UN’s historical practice emphasizes:

• Documentation: The UN produces vast archives of resolutions, reports, and debates that serve as the institutional memory of the international community.
• Legal precedent: Historical interpretation in the UN is oriented toward establishing what the law requires, not toward explaining why events occurred.
• The “lessons learned” genre: The UN produces retrospective reports on its own failures (e.g., Rwanda, Srebrenica, Sri Lanka) that are intended to improve future practice.
• The normative narrative: The UN’s historical narrative is progressive — the story of the expansion of human rights, decolonization, and international law — even as it acknowledges failures.

What this perspective contributes to causal inference. It provides the institutional record — the documented decisions, resolutions, and debates that are essential evidence for any causal claim about international politics. It also provides the normative benchmark — the standards against which state behavior is judged.

What it occludes: It can be selective in its documentation; it can treat legal precedent as a substitute for causal explanation; it can underplay the power dynamics that shape its own institutional memory.

2.5.6 The comparison matrix

2.6 WORKED EXAMPLE (I do)

The puzzle: The Rwandan genocide of 1994 killed approximately 800,000 people in 100 days. The international community — the UN, the US, France, Belgium — failed to intervene effectively. Why?

The puzzle stated precisely: Why did the international community fail to intervene to stop the Rwandan genocide, despite the clear evidence of mass atrocities and the existence of a UN peacekeeping mission (UNAMIR) on the ground?

Level of analysis: The failure is best explained at the systemic and state levels, with the individual level providing the proximate trigger.

• Systemic level: The post-Cold War moment had produced a normative expectation of humanitarian intervention (the “new world order”), but it had not produced the institutional capacity or political will to act. The Security Council was reluctant to authorize a robust mission; the US had just experienced the “Black Hawk Down” failure in Somalia (1993) and was unwilling to intervene in Africa.
• State level: The US, France, and Belgium had different interests in Rwanda. France had supported the Habyarimana regime and was implicated in the arming of the genocidaires. Belgium was the former colonial power and had a special responsibility. The US was focused on domestic politics and had no strategic interest in Rwanda.
• Individual level: The UN Secretary-General, Boutros Boutros-Ghali, was sympathetic but constrained by the Security Council. The UNAMIR commander, Roméo Dallaire, requested reinforcements but was denied. The US Ambassador to the UN, Madeleine Albright, was constrained by Washington.

Method 1 — Process tracing: Trace the sequence of decisions from April to July 1994:

• April 6: Habyarimana’s plane is shot down; the genocide begins.
• April 7–14: Dallaire requests reinforcements; the Security Council debates.
• April 21: The Security Council votes to reduce UNAMIR’s force from 2,500 to 270.
• May: The genocide escalates; the US resists using the word “genocide” to avoid legal obligations.
• June: France launches Operation Turquoise, a unilateral intervention that is widely criticized for protecting perpetrators.
• July: The RPF (Rwandan Patriotic Front) captures Kigali and ends the genocide.

Process tracing reveals the mechanism: the Security Council’s decision to reduce UNAMIR was not a response to new information; it was a response to the political pressure from the US, which did not want to be drawn into a peace enforcement mission. The mechanism is state-level interest, operating through the systemic institution of the Security Council.

Method 2 — Comparative method: Compare Rwanda with other post-Cold War humanitarian crises:

• Somalia (1992–93): The US intervened, then withdrew after casualties. The failure of Somalia shaped the US response to Rwanda.
• Bosnia (1992–95): The international community intervened belatedly, but eventually with NATO air power. Why Bosnia and not Rwanda? Because Bosnia was in Europe, had a higher strategic profile, and the media coverage was more sustained.
• Kosovo (1999): NATO intervened without UN authorization. Why Kosovo and not Rwanda? Because Kosovo was in Europe and the US had a domestic political interest.

The comparative method reveals that the variable that explains intervention is strategic interest, not the severity of the humanitarian crisis. Rwanda was in Africa, had no strategic resources, and was not a priority for the US or Europe.

Method 3 — Counterfactual analysis: What if the US had been willing to support a robust UNAMIR mission?

• Minimal rewrite: The US changes its position on UNAMIR reinforcements, but all other factors remain constant.
• Theory: The genocide was organized by a small elite (the akazu) and depended on the state apparatus. A robust international force could have disrupted the command structure and protected civilians.
• Observable implications: If the counterfactual is correct, we should see evidence that the genocide was vulnerable to external disruption — e.g., the RPF’s success in ending it, the capacity of UNAMIR to protect civilians when it had the mandate and resources, the rapid collapse of the genocide once the RPF advanced.
• Evidence: The RPF, a relatively small force, was able to defeat the genocidal regime. This suggests that a well-equipped international force could have intervened effectively. Dallaire’s reports suggest that a force of 5,000 could have made a difference.

Perspective reconstruction:

• African perspective: The failure is read through the colonial archive. The genocide was not a “tribal conflict” but a political project organized by a state elite, enabled by the colonial legacy of ethnic classification (the Belgians’ 1933 identity cards) and the impunity of the Habyarimana regime, which was a French ally. The international community’s failure is read as racialized indifference — African lives were deemed expendable because they were African.
• Chinese perspective: The failure is read as a failure of Western interventionism — the West intervened in Bosnia and Kosovo because they were in Europe, but not in Rwanda because it was in Africa. The lesson is that sovereignty must be respected and that external powers cannot be trusted to protect African lives.
• Russian perspective: The failure is read as evidence of Western hypocrisy — the West claims to care about human rights but intervenes only where its interests are at stake. The lesson is that great-power interests determine the fate of small states.
• Western perspective: The failure is read as a failure of political will and institutional capacity. The US was traumatized by Somalia; the UN was constrained by the P5; the media did not cover the genocide adequately. The lesson is that institutions need better mandates and resources and that early warning must be matched by early action.
• UN perspective: The failure is read as a failure of the Security Council to act on the Secretary-General’s warnings. The UN’s own “lessons learned” report (the Carlsson Report) is a primary document of institutional self-criticism.

Adjudication. Which perspective best explains the failure to intervene? The African perspective captures the racialized structure of international concern — the fact that African lives were not considered worth the risk of intervention. The Western perspective captures the proximate political mechanism — the US decision not to support UNAMIR, the Security Council’s reduction of the force, the failure to use the word “genocide.” The Chinese and Russian perspectives capture the structural skepticism of great-power intervention, which is itself a reason why interventions do not happen. The UN perspective captures the institutional failure of the Security Council.

A complete explanation combines the systemic level (the post-Somalia reluctance to intervene) with the state level (US, French, and Belgian interests) and the transnational level (the racialized structure of international concern). The African perspective is essential because it identifies the racial hierarchy that made the failure possible — not as a conspiracy, but as a structured indifference.

2.7 PROBLEM PAIR (We do)

Your task: Analyze the following case using the three methods and the perspective comparison. Write your analysis before reading the model answer.

The case: In July 2023, a military junta overthrew Niger’s elected president, Mohamed Bazoum. France and the US had military bases in Niger and had been key security partners. The junta demanded that French troops leave; they did, by the end of 2023. The junta then deepened ties with Russia (via the Africa Corps) and China. ECOWAS threatened military intervention, but did not intervene. The junta remains in power.

Your tasks:

1. State the puzzle precisely: why did the coup lead to the expulsion of France and the realignment toward Russia and China, rather than to a restoration of the previous alignment?
2. Apply process tracing to the sequence of events from July 2023 to the present. What mechanism does the evidence suggest?
3. Apply the comparative method: compare Niger with Mali (2020 coup) and Burkina Faso (2022 coup). What explains the similar outcomes? What variable distinguishes these cases from coups that did not lead to realignment?
4. Construct a counterfactual: what if France had negotiated with the junta rather than demanding Bazoum’s reinstatement? What would we expect to see if the counterfactual is correct?
5. Reconstruct the African and Russian perspectives on this case. How do they interpret the junta’s actions?
6. Write a 300-word adjudication: which perspective best explains the realignment, and what does it miss?

Model answer (compare after writing your own):

Puzzle: Why did Niger’s coup lead to the expulsion of France and realignment toward Russia and China, rather than to a restoration of the previous alignment or a negotiated settlement?

Process tracing: The sequence is:

• July 26, 2023: The presidential guard, led by General Tchiani, detains Bazoum and declares a coup.
• July 30: ECOWAS issues a seven-day ultimatum demanding Bazoum’s reinstatement and threatens military intervention.
• August: The junta refuses; Mali and Burkina Faso pledge to defend Niger; the AU does not endorse intervention.
• September: France refuses to recognize the junta and insists on Bazoum’s release; the junta demands the withdrawal of French troops.
• October: France announces the withdrawal of its ambassador and troops.
• December: The last French troops leave.
• 2024: The junta deepens ties with Russia (Africa Corps) and China (mining and infrastructure deals).

The mechanism is mutual escalation and the search for external alternatives. France’s refusal to negotiate, combined with ECOWAS’s threat of intervention, made the junta dependent on external security partners. Russia provided security assistance without conditionality; China provided economic engagement without governance conditions. The junta used the external competition to consolidate power and legitimize its rule through anti-French nationalism.

Comparative method: Compare Niger with Mali (2020) and Burkina Faso (2022):

• All three had military coups, all three experienced a breakdown in relations with France, and all three realigned toward Russia.
• The variable that distinguishes them from coups that did not lead to realignment (e.g., the 2021 coup in Guinea, which remained aligned with France) is the presence of a significant French military footprint and the failure of French-led counterterrorism. In Mali and Burkina Faso, French forces were present and had failed to contain the insurgency; in Guinea, there was no French military presence and the junta did not frame its legitimacy around anti-French nationalism.
• The comparative method suggests that the French military presence and the failure of French counterterrorism are the key variables.

Counterfactual: What if France had negotiated with the junta rather than demanding Bazoum’s reinstatement?

• Minimal rewrite: France changes its strategy but all other factors remain constant.
• Theory: The junta’s primary interest was regime survival, not alignment with Russia per se. If France had offered security cooperation without demanding Bazoum’s reinstatement, the junta might have maintained ties with France.
• Observable implications: If the counterfactual is correct, we should see evidence that the junta was open to negotiation with France — e.g., statements from the junta suggesting willingness to cooperate on security, diplomatic back-channels, or comparisons with other cases where negotiation produced a different outcome.
• Evidence: The junta’s early statements emphasized sovereignty and the failure of French counterterrorism, not an ideological commitment to Russia. The junta’s subsequent deals with Russia and China were instrumental, not ideological. This suggests that a different French strategy might have produced a different outcome.

African perspective: The junta’s actions are interpreted as a rational assertion of sovereignty against a former colonial power that had failed to provide security and had imposed conditions. The anti-French sentiment is not irrational; it reflects a long history of colonial and neocolonial relations, and it is shared by significant portions of the population. The junta’s realignment is a strategy of multi-alignment — using external competition to increase Niger’s bargaining power.

Russian perspective: The junta’s actions are interpreted as a sovereign choice by a state that has been abandoned by the West. Russia is a reliable partner that provides security without political conditions. The realignment is a symptom of the collapse of Western hegemony and the emergence of a multipolar order.

Adjudication: The African perspective best explains the legitimacy of the realignment — why the junta’s anti-French posture was popular and why it could sustain itself. The Russian perspective explains the opportunity — why the junta had alternatives to France. The Western perspective explains the proximate cause — the French refusal to negotiate and the ECOWAS threat made the realignment inevitable. The comparative method suggests that the French military presence and the failure of French counterterrorism are the key structural variables. A complete explanation combines the African, Russian, and Western perspectives: the structural failure of French counterterrorism (Western), the availability of Russian alternatives (Russian), and the legitimacy of anti-colonial sovereignty assertion (African).

2.8 INDEPENDENT MEMO (You do)

Produce a one-page analytical memo (500–600 words) on the following case, using the four headings below.

Case: The collapse of the JCPOA and Iran’s withdrawal from the treaty (2018–2025).

Memo headings:

1. Mechanism: What causal process explains the collapse? Specify the level of analysis and the mechanism (economic pressure, political signaling, or reciprocal escalation).
2. Method: Apply process tracing, comparative method, and counterfactual analysis to the case. What does each method contribute?
3. Perspective reconstruction: How would each of the five perspectives explain the collapse and interpret the evidence? (One paragraph or table entry each.)
4. Falsifiers: What evidence would prove your explanation wrong? What would you need to observe to abandon it?

Self-check: Does your memo distinguish mechanism from correlation? Does it specify the counterfactual? Does it identify the evidence that would falsify the claim?

2.9 HINGE QUESTION

A scholar argues: “The Arab Spring happened because of the combination of youth unemployment and social media.” A critic responds: “But youth unemployment and social media were also present in countries that did not experience an Arab Spring.” What methodological problem is the critic identifying?

• A) Endogeneity
• B) Selection on the dependent variable
• C) Presentism
• D) Teleology

Answer: B. The critic is identifying selection on the dependent variable. The scholar has selected cases where the Arab Spring occurred and identified factors present in those cases. But the critic points out that those factors are also present in cases where the Arab Spring did not occur — which means the factors do not explain the outcome. To establish causation, the scholar would need to examine the cases where the factors were present but the outcome did not occur.

If you answered A: Endogeneity refers to a situation where the cause is itself caused by the effect (reverse causation) or where cause and effect are mutually constitutive. The critic’s point is about case selection, not endogeneity. If you answered C: Presentism refers to reading the past through the lens of the present. The critic’s point is about case selection, not presentism. If you answered D: Teleology refers to explaining events as if they were directed toward a predetermined end. The critic’s point is about case selection, not teleology.

SECTION 2A — DATA LAB (30–45 minutes)

Purpose: Ground the lesson’s methodological concepts in empirical work.

Task: Using the summary data below, analyze the relationship between economic inequality and civil war onset.

Table 1: Civil war onset and inequality (illustrative data, simplified)

Analysis questions:

1. Association: Is there an association between inequality and civil war onset in this data? What about colonial legacy and resource dependence?
2. Selection bias: If you selected only the countries with civil war onset (A, C, E), what would you conclude about the causes of civil war? How does including the non-onset cases change your conclusion?
3. Mechanism: What mechanism might link inequality to civil war? What mechanism might link colonial legacy to civil war? How would you test these mechanisms?
4. Counterfactual: Choose one case and construct a counterfactual: what would have happened if the inequality had been lower? What evidence would you need to support the counterfactual?
5. Perspective: How would each of the five perspectives interpret this data? (African: colonial legacy and structural inequality; Chinese: underdevelopment and external interference; Russian: Western hegemony and resource extraction; Western: grievance and opportunity; UN: governance and development.)

Rule: Complete the data lab before the independent memo.

SECTION 3 — DUAL-CODED VISUAL ANCHOR

Task: Draw the following by hand.

Visual 1: The three methods. Draw three overlapping circles labeled “Process tracing,” “Comparative method,” and “Counterfactual analysis.” In the overlap, write “Robust causal claim.” For each circle, write one question it answers and one limitation.

Visual 2: Selection bias. Draw a 2×2 table: rows = “Cause present” and “Cause absent”; columns = “Outcome present” and “Outcome absent.” Shade the cell that selection on the dependent variable excludes. Write a sentence explaining why excluding that cell biases inference.

Visual 3: The five perspectives on historical method. Draw a five-column table with rows: “What counts as evidence,” “Purpose of history,” “Key occlusion.” Fill from memory.

SECTION 4 — CORNELL NOTES

Set up your notes with the following cue column:

SECTION 5 — FEYNMAN COMPRESSION

Step 1: Explain the difference between correlation and causation to a smart non-specialist in 150–200 words, using an example from IR.

Step 2: Mark every point where you hesitated or used a term you could not define.

Step 3: Reread only the gaps. Rewrite.

Step 4: Produce one analogy for process tracing. Then state where the analogy breaks down. (Example: “Process tracing is like a detective reconstructing a crime — but unlike a detective, the IR analyst cannot interview the suspects, and the evidence may have been deliberately destroyed.”)

SECTION 6 — SPACED-REPETITION FLASHCARDS

Produce 12–15 cards. Use the following types:

Tag each card: L3, method, perspective, causation, difficulty.

SECTION 7 — METACOGNITIVE CALIBRATION

7.1 Confidence audit

Rate confidence (0–100%) before checking:

After exercises, rate accuracy. Calculate the gap.

7.2 Three-question debrief

1. What did I misunderstand about the relationship between mechanism and counterfactual?
2. Which method was easiest to apply? Which was hardest? What does that suggest about my default analytical style?
3. What will I do differently in Lesson 4?

7.3 Error log

SECTION 8 — CHEATSHEET (Module 0)

Add this row to your Module 0 cheatsheet:

SECTION 9 — REVIEW SCHEDULE

SECTION 10 — COMPLETION CHECKLIST

• Retrieval warm-up completed from memory
• Five learning objectives read and understood
• Four key ideas studied with mechanisms and scope conditions
• Perspective comparison matrix filled from memory, then checked
• Worked example (Rwanda 1994) studied and self-explained
• Problem pair (Niger 2023) attempted and compared
• Independent memo produced (500–600 words)
• Hinge question answered correctly (or reteaching completed)
• Data lab completed before independent memo
• Three dual-coded visuals drawn
• Cornell notes with cue column, source log, and summary
• Feynman explanation written, gaps identified, rewritten
• 12–15 flashcards created with tags
• Confidence audit and error-log entry
• Cheatsheet entry added
• R2 review scheduled

MODULE 0 SYNTHESIS

You have now completed the orientation and method module. Before proceeding to Module 1 (Epistemological and Ontological Foundations), write a 1,500-word synthesis essay answering the following question:

“What is an explanation in International Relations, and how would I know a good one from a bad one?”

Your essay must:

1. Define the puzzle, the level of analysis, and the concept of power (L1).
2. Distinguish description, explanation, prediction, and prescription (L1).
3. Explain the relationship between legal rules and power politics (L2).
4. Explain the difference between correlation and causation, and identify the three methods of causal inference (L3).
5. Reconstruct the five perspectives on at least two of the three lessons, and adjudicate between them on a specific case.
6. Identify at least three methodological errors that are common in IR, and explain how to avoid them.

This essay is your entry ticket to Module 1. It is assessed on the quality of the argument, the precision of the concepts, and the use of evidence — not on length.

FURTHER READING (Optional — for deepening, not required)

• On counterfactuals: James Fearon, “Counterfactuals and Hypothesis Testing in Political Science,” World Politics (1991); Philip Tetlock and Aaron Belkin, Counterfactual Thought Experiments in World Politics (1996).
• On selection bias: Barbara Geddes, “How the Cases You Choose Affect the Answers You Get,” Political Analysis (1990); Gary King, Robert Keohane, and Sidney Verba, Designing Social Inquiry (1994), ch. 4–6.
• On process tracing: Alexander George and Andrew Bennett, Case Studies and Theory Development in the Social Sciences (2005); Andrew Bennett and Jeffrey Checkel, Process Tracing (2015).
• On comparative method: Arend Lijphart, “Comparative Politics and the Comparative Method,” American Political Science Review (1971); Charles Ragin, The Comparative Method (1987).
• On African historiography: Michel-Rolph Trouillot, Silencing the Past (1995); Ann Laura Stoler, Along the Archival Grain (2009); Toyin Falola, Decolonizing African History (2022).
• On the Rwandan genocide: Romeo Dallaire, Shake Hands with the Devil (2003); Philip Gourevitch, We Wish to Inform You That Tomorrow We Will Be Killed with Our Families (1998); the UN’s Carlsson Report (1999).
• On Chinese and Russian historical narratives: Zheng Wang, Never Forget National Humiliation (2012); Lauri Mälksoo, Russian Approaches to International Law (2015); Maria Todorova, Imagining the Balkans (2009).

End of Lesson 3. End of Module 0.

Say “Next” when you are ready for Lesson 4 (Positivism, Post-Positivism, and the Linguistic Turn), the first lesson of Module 1.