Polisci midterm

Polisci midterm flashcards

What are the post-WWII factors that led to the expansion of the comparative politics subfield in the U.S.?

After WWII, U.S. comparative politics grew because of the Cold War, the rise of newly independent states after decolonization, and government funding for area studies and language training (Title VI, Fulbright-Hays). Vanderbilt even established an Institute for Brazilian Studies in 1949 with Carnegie funding Fall 2025. Scholars wanted to understand both allies and adversaries, fueling the expansion of comparative politics.

What is the basic trade-off between the number of cases in your study and the measures you employ in that research?

This is the “many variables, small-N problem”. Large-N studies allow for generalization but sacrifice precision and nuance in measures. Small-N or single case studies allow richer, context-specific measures but cannot be generalized as widely. Researchers must balance 

depth vs. breadth

What are the key issues involved in the “Area Studies/Generalization” debate?

Area studies emphasize deep local expertise, language, and culture, while generalization emphasizes theory building and cross-regional comparison. The tension: “Can’t see the forest for the trees” (area studies) vs. “too broad, not enough depth” (generalization). The discipline has tried to integrate both — using area-specific knowledge to refine broader theories.

Based on the article “The Research Question” in Section 1 of your syllabus, what is the difference between a “descriptive” research question and an “analytical” research question? Provide an example of each.

• Descriptive: “What is happening?” (e.g., “What is the level of voter turnout in Mexico’s 2024 election?”). 

• Analytical/Explanatory: “Why/How is it happening?” (e.g., “Why is turnout higher in Mexico than in the U.S.?”). Slides emphasize IV = explanatory factor, DV = outcome of interest. Descriptive maps phenomena, analytical builds causal arguments.

What are dependent and independent variables? Is it possible for a dependent variable to also be an independent variable (or vice versa)?

A dependent variable (DV) is the “outcome of interest,” while an independent variable (IV) is a factor hypothesized to affect it. Example: DV = regime type; IV = economic inequality. 

Yes, a variable can be both — e.g., democratization is a DV in some studies, but an IV in others (when explaining human development).

On what regions of the world does the Vanderbilt Comparative Politics faculty generally focus?

Vanderbilt comparative faculty emphasize Latin America, with strengths in democratization, state capacity, inequality, and survey research (via LAPOP – Latin American Public Opinion Project) as well as Chinese and Middle Eastern. One relatively weaker side is Western European

As discussed in class, what are the “two faces of power”?

Bachrach & Baratz (1962): 

• Overt power is Decision-making power → ability to push or veto policies. 

• Covert power is Agenda-setting power → ability to control what issues are even discussed. 

The second is subtle but critical: shaping the agenda defines what is politically possible.

What are some ways in which specific system support might affect diffuse system support? How might diffuse support affect specific support?

Specific support = support for what authorities do (policies, performance). 

Diffuse support = support for the underlying system/rules of the game. Consistently poor performance (low specific support) can erode belief in the system itself (diffuse support). Conversely, strong diffuse support (belief in democracy) can sustain specific support even when governments underperform.

 To what type of system support does the phrase “Obedezco pero no cumplo” apply?

This Spanish phrase (“I obey but do not comply”) captures weak diffuse support. Citizens outwardly acknowledge the authority of rules but don’t internalize them, undermining legitimacy.

What is an institution as defined by North? What is a principal challenge in the comparative study of institutions?

North (1991): Institutions are “humanly devised constraints that structure political, economic, and social interaction” — both formal (laws, constitutions) and informal (customs, traditions). Informal rules are difficult to measure. Challenge is establishing a valid basis for comparison, particularly given the subjective nature of informal constraints.

What is an example of a formal institution? What is an example of an informal institution? 

• Formal: U.S. Constitution, independent central bank. 

• Informal: “Usos y Costumbres” in Mexico, unwritten political norms, taboos

What are some key elements of a political regime? How do they help us distinguish between a democracy and an authoritarian regime

A regime defines the rules of the game: 

1. Access to power (inclusive vs exclusive). 

2. Transfer of power (contestation, institutionalized uncertainty – parties can lose). 

3. Application of power (rule of law vs arbitrary). Democracies = inclusiveness + contestation + constitutionalism (Dahl, Przeworski). Authoritarian regimes exclude, transfer unpredictably, and apply power arbitrarily.

What are some examples of a strong/weak state? 

a. Does a military authoritarian regime necessarily have a strong state?

• Strong state: High capacity (effective bureaucracy, policing, taxation) → e.g., South Korea. 

• Weak state: Limited control, corruption, fragility → e.g., South Sudan.

A military authoritarian regime isn’t necessarily “strong” — many authoritarian states are weak in capacity but strong in repression.

What and where is Kurdistan? 

Kurdistan is a stateless nation spanning Iraq, Turkey, Iran, Syria, and Armenia, long seeking statehood. It highlights the mismatch between “nation” and “state” — a nation without sovereignty.

What are some of the common components found in most definitions of democracy?

• Inclusiveness (broad participation). 

• Contestation (genuine competition, parties can lose). 

• Rule of law/constitutionalism (procedures, civil liberties). 

• Other measures (e.g., EIU, Freedom House) add participatory, deliberative, and effective dimensions.

What are some of the challenges faced by scholars attempting to measure democracy? 

Concepts differ (is democracy binary or a continuum?), measurement tools vary (Freedom House, Polity, V-Dem, EIU). Authoritarian regimes often manipulate elections but retain formal institutions (hybrid regimes). “Diminished subtypes” (illiberal democracy, competitive authoritarianism) blur boundaries.

 How does a central bank’s degree of autonomy relate to a country’s regime type? 

• More autonomy = stronger institutional constraint on rulers, typical in democracies. 

• Less autonomy = rulers can use banks for short-term gain, common in authoritarian regimes. It’s an indicator of institutional checks and balances.

• Separating central bank autonomy to regime type autonomy 

• It’s not a defining component of democracy. Central banks might or might not be related but either way it’s not a defining feature 

What are the three dimensions of “development,” and which does Streeten propose we view as the end goal?

• Economic development: growth of national wealth. 

• Political development: conflict resolution, governance for the “greater good.” 

• Human development: expanding choices and potential. 

Streeten argues human development should be the ultimate end goal. It’s good for the other elements of development if human development comes first. Good for the economy. (Ex. Educated workforce)

 Speaking of “development,” do all good things go together? Give an example to support your answer. 

Not always — development involves trade-offs (“cruel choices”). 

Example: Economic growth may fuel inequality or undermine democracy (resource curse, China).

What is the human development index and its three dimensions? What are the measures used to capture those dimensions?

• Health: life expectancy. 

• Education: years of schooling. 

• Living standards: GNI per capita (PPP). Range: 0–1 (higher = more developed).

What does the term “noisy data” mean and why should we be concerned if it is systematic noise vs. random noise? 

“Noisy data” = unreliable measurements (errors, distortions). Data with problems.

Random noise cancels out; systematic noise biases results in one direction, distorting findings.

What type of development data tends to be most vulnerable to systematic noise? Why?

Data on human development (e.g., literacy rates, infant mortality) — governments may manipulate or underreport for legitimacy

Why must we use “per/capita” measures of economic development in comparative politics?

Because total GDP doesn’t reflect average individual well-being. Per capita adjusts for population size, making comparisons possible between countries 

What are the main problems with GDP/capita as a measure of economic and/or human development?

• Doesn’t capture distribution (inequality). 

• Masks regional differences (whole-nation bias). 

• Doesn’t account for sustainability or quality of life. 

• Rough proxy for human development

Given all the problems with GDP/capita, why is it such a commonly used measure of development?

• Widely available across time and space (countries). 

• Relatively reliable compared to other indicators. 

• Correlates fairly well with other development indicators

• Easy to communicate and compare.

• Countries don’t really screw around with their own economic data because of both domestic and international pressures—countries who break these rules typically get punished pretty quickly.

What is the “tunnel effect” (see Hirschman) and how might it relate to the concepts of specific and diffuse support (regime legitimacy)? How might it relate to a society’s cross-cutting vs. overlapping divides?

The tunnel effect = when inequality is tolerated if people believe their own situation will improve soon (like being stuck in traffic while the other lane moves, expecting your turn next). This idea basically describes are rapidly developing country where one group is growing really quickly and one is not or just growing less quickly but essentially this just describes income inequality in getting ahead over time.

• Specific/diffuse support: If citizens expect growth, they may keep supporting the system (diffuse support, overall structural support) despite unequal outcomes (specific support, basically if they support the people in government at that time)). But if expectations collapse, legitimacy erodes. 

• Divides: In cross-cutting contexts (you see people in the other lane that look like you), groups may share optimism, stabilizing democracy. In overlapping divides (everyone in your lane looks the same as you do, everyone in the other lane looks different from you), frustration lines up along identity/economic cleavages, fueling conflict, shorter moment of hope.

Based on class discussions, what are some key trade-offs a principal must consider when delegating power to an agent? Provide some examples of these trade-offs.

Principals face trade-offs in: 

• Discretion vs. rules (e.g., TSA agents given flexibility vs strict procedures). 

• Independence vs. oversight (e.g., lifetime Supreme Court terms vs congressional term limits). 

• Agent incentives vs. long-term interests (e.g., politicians prioritizing reelection over policy). Too much discretion risks agency loss, too much oversight reduces effectiveness. 

• The agent should know more than the principal so they should be able to do what they think is best. Agents have to sometimes make tough choices in the short term to benefit the principal in the long term. Downside of high levels of independence: can go rogue, undermine what the principal really wants them to do, agency loss. High levels of oversight, you get the ability to remove the agent if you need to and remove agency loss. Downside: agent isn't able to do what they potentially need to do, agents will try to please the voters in the short term while potentially not serving the long term interests of the principal

According to Lupia, what is agency loss and how is it minimized?

Agency loss = the gap between what a principal wants and what the agent actually does.

Lupia: It is minimized when 

(1) principal and agent share common interests (selection process determines wether the agent has same interests as the principal)

(2) principal has information to monitor consequences. Example: central bank autonomy reduces agency loss if the bank's goals align with the system's long-term stability.

Why is the information imbalance between principal and agent so important in understanding the nature of the relationship and the potential for agency loss?

Agents often have more specialized information than principals (information asymmetry). This imbalance makes monitoring difficult and increases agency loss. 

Example: security services during the “War on Terror” — agents had more info than legislators, allowing them to pursue actions not fully aligned with oversight

What are the principal-agent trade-offs of a lifetime Supreme Court term vs. a 2-year term for a Representative?

• Lifetime Supreme Court term: high independence, low oversight → risk of unaccountability, but protects against political pressure. 

• 2-year Representative term: frequent accountability, but incentives skew toward short-term reelection goals rather than long-term governance. Both illustrate how delegation design balances independence vs accountability.

What is the difference between an absolute and relative measure of poverty? What does each imply for assessing the poverty levels of a particular country?

• Absolute poverty: Fixed threshold, like the World Bank line ($3.00/day in low-income countries). It measures whether people can meet basic needs. 

• Relative poverty: Defined in comparison to others within the same society (e.g., % below 50% of median income). Implication: A country may have low absolute poverty (basic needs met) but high relative poverty (persistent inequality).

What does the Gini Index measure and what is its range?

The Gini Index measures income inequality by comparing the actual income distribution to perfect equality (the Lorenz curve).

• Range: 0 = perfect equality, 1 = perfect inequality. 

• Example: Brazil’s Gini fell from 63.3 in 1989 to 54.7 in 2009, but inequality remained high.

If you are told that the economy of a country is generally extraction-based developed during colonial rule, what would be some of your hypotheses regarding that country’s characteristics?

Extraction economies usually show: 

• Boom–bust cycles of growth tied to commodity markets. 

• Export-oriented infrastructure (“all roads lead to the port”).

• Authoritarian or oligarchic systems with elites controlling rents. 

• High inequality and weak incentives for education (very poor lower and middle income education systems). 

• Corruption + foreign influence in politics

• Diffuse support is very very low

What is the paradox of plenty according to Karl? According to Hiskey?

• Karl (1999): Oil wealth → centralized authority, rentier politics, corruption, overspending, authoritarian kleptocracy. Cursed by the resource due to (international influence, reenterism - cities can crumble and you can’t complain, corruption) 

• Hiskey: Natural wealth becomes a curse because rulers lose the capacity to say “no.” More wealth leads to more waste, more centralization, and less accountability. Corruption.

What are some of the key elements of the “resource curse”?

• Corruption and patronage networks. 

• Overspending and debt accumulation. 

• Authoritarian consolidation (power centralized, taxation avoided). 

• Weaker institutions and state legitimacy. 

• Volatility from dependence on global commodity prices

What does a "petri dish" experiment have that political scientists want?

Petri dishes allow control over variables — researchers isolate one factor while holding others constant. Political scientists aim for the same: controlling for confounding variables to establish causality

In your view, what is an example of a measure in comparative politics that is fairly close to the concept it is trying to capture? What is an example of a measure that is far away from the concept it is trying to capture?

Close: Life expectancy to measure “health” in human development. GINI index to wealth distribution (there are critiques, still a distance)—still pretty close though.

Far: GDP per capita as a proxy for “quality of life” — economic output ≠ well-being

Why would a researcher decide to conduct a "static single case study"?

It allows rich, detailed analysis of one country at one time, with deep contextual knowledge, intensive and elite interviews, etc. You want to understand why the country works before you understand how it works. Weakness: no variation = no causal inference or generalization. Does not allow for any contextualization. 

What are the factors that might be controlled for in a dynamic case study? What factors would not be controlled?

Controlled: time-invariant variables (culture, geography).

Not controlled: time-varying variables (technology, global conditions, military coup, major accident etc.)

What is the basic logic behind a "most similar" and "most different" systems research design? To what degree do they help a researcher provide support for a theory or hypothesis?

Most Similar Systems (MSSD): Compare similar cases, isolate 1 key difference.

Most Different Systems (MDSD): Compare very different cases that share 1 outcome.

Both help rule out alternative explanations, but don’t fully prove causation

What additional strengths/weaknesses might a subnational most similar/different systems design have?

Strength: More detailed knowledge of subnational units, reduces measurement error.

Weakness: Units may be too similar, limiting meaningful variation.

What are the pros and cons of a large-N, cross-national research strategy?

Pros: Generalizability, statistical control over many variables, confidence estimates.

Cons: Measurement error, loss of detail, “whole-nation bias” (masking variation within countries)

What type of research question might benefit from intensive or elite interview data?

Questions about decision-making processes, elite bargaining, or lived experiences. E.g., how Cuban officials handled the missile crisis. Research questions trying to uncover the nuance. Mostly descriptive.

According to Hochschild, what is the main difference between an intensive interview and an elite interview?

Intensive: Open-ended, long, often with randomly chosen individuals in the target population.

Elite: Interviews with key actors chosen for who they are / their position

What is the key advantage of field and lab experiments for scholars of comparative politics?

They offer control. Causal leverage — random assignment lets researchers isolate treatment effects. The petri dish is gold standard because it is under absolute control.

According to Green, when did field experiment work in political science begin to increase rapidly?

Field experiments grew rapidly in the 1998– early 2000s, especially with randomized control trials (RCTs) in development and get-out-the-vote campaigns 

What are some potential issues that may arise with field experiment-based research?

• Generalizability (results context-specific).

• Attrition bias (participants drop out).

Replication challenges.

What is an example of an ethical issue that might arise in a field experiment?

Deception or lack of consent. Example: withholding social services from a control group in an RCT. Influencing elections.

What is the difference between a cross-sectional and a longitudinal (or panel) survey? What are some strengths and weaknesses of each?

Cross-sectional: One point in time, broad snapshot. → Strength: fully representative sample, fast, cheap. Weakness: no causal dynamics.

Longitudinal/panel: Same respondents over time. → Strength: tracks change-easier to establish a fairly compelling link. Weakness: costly, attrition, if someone drops out from the survey. If there were to be 1500 and 1200 dropped out, then is 300 remaining truly representative of the entire 1500?

Give an example of a natural experiment. What are some potential problems with natural experiments?

Ex: Immigration crackdown → study before/after effects on PTA participation.

Problems: Say before 1/20 and after 1/20,  no true control, other factors may have changed too, researcher lacks randomization.

Ex: Earthquake → beautiful study of natural experiment before the earthquake vs after the earthquake. E.g. a system not responding to the earthquake and diffuse support falling as a result

Field experiment: get out the vote campaign in one village vs. another (created by the researcher)

What is the "law of large numbers" and how does it help survey researchers design their surveys?

The larger the sample, the closer the estimate will be to the true population value. Population size doesn’t matter as long as the sampling procedure is correct. It helps researchers to conduct reasonable amount of surveys instead of millions 

→ Explains why 1,500–2,000 respondents is enough for national surveys

What are the limitations of a twitter poll?

• Sampling bias (users not representative).

• Self-selection bias (respondents not random).

• No control over multiple responses, bots, or question wording

Which of the following problems with survey question do you think is most problematic: Leading questions; Loaded questions; Double-barreled questions?

All distort results, but double-barreled questions are especially problematic because they combine two questions into one, making responses uninterpretable

What are some other issues discussed in class regarding the design and implementation of a survey that might affect the quality of the answers a respondent provides to an interviewer?

Priming / question order effects.

Interviewer effects (gender, race, language).

Context effects (location, presence of others).

Social desirability bias (respondents give “correct” answers)- Regime type of country you are in. It’s very hard to do survey in authoritarian regime, like Nicaragua and Venezuela

What are some of the keys to developing a well-designed research plan?

Recognize trade-offs (cases vs control).

Define scope conditions (time, space).

Be clear about concepts vs measures. Good Data

Choose methods that match your theory of causality (correlation ≠ causation)