- A regression analysis of campaign spending and election outcomes in U.S. House races shows that the candidate who spends more money wins the election approximately 90% of the time. A pundit argues this proves that money is the deciding factor in politics. A political scientist would be most likely to challenge this conclusion by pointing out that:
Answer: Donors are strategic and tend to give more money to candidates who are already likely to win, such as popular incumbents.
The correct answer identifies a critical confounding factor and selection effect. The perceived viability or likelihood of winning is a confounder. It influences both how much money a candidate can raise (donors flock to likely winners) and the ultimate election outcome. This makes it appear that money is causing the win, when in fact the likelihood of winning is causing both. This is a more fundamental challenge to the causal claim than the others. (A) identifies a possible omitted variable, but it's less central than candidate viability. (C) relates to the mechanism of spending, not whether spending is causal in the first place. (D) misunderstands statistical analysis; a 90% correlation is very strong, and outliers do not invalidate it, though the inference drawn from it can be flawed.
- A political scientist analyzes historical data and finds a correlation between the expansion of a country's railroad network and its subsequent democratization. They theorize that railroads foster a national identity and connect dissident groups, causing democratic change. Which of the following discoveries would most strongly suggest that a confounding variable is at play?
Answer: The process of industrialization tended to occur at the same time as railroad expansion and also created new social classes that demanded political representation.
When analyzing correlations in political science research, you must always consider whether a third variable might be causing both observed phenomena—this is called a confounding variable. The researcher observes that railroad expansion correlates with democratization and theorizes a causal relationship, but correlation doesn't prove causation.
Answer D identifies the strongest confounding variable: industrialization. If industrialization drives both railroad construction AND creates new social classes demanding political representation, then industrialization—not railroads themselves—could be the true cause of democratization. This would make the railroad-democracy correlation spurious, meaning it exists only because both variables are caused by the same underlying factor.
Answer A suggests timing issues but doesn't eliminate the causal relationship—delayed effects are common in political processes and don't necessarily indicate confounding. Answer B describes the motivations of authoritarian regimes but doesn't provide an alternative explanation for democratization itself; unintended consequences can still create causal relationships. Answer C mentions localized anti-government sentiment from land seizures, but this describes a direct effect of railroad construction rather than suggesting an alternative cause for broader democratization patterns.
The key insight is that industrialization represents a comprehensive economic and social transformation that could plausibly drive both infrastructure development (railroads) and political change (democratization) independently. This makes it a classic confounding variable.
Remember: when evaluating causal claims in political science, always ask whether a broader underlying process could be driving multiple observed changes simultaneously.
- A study of 20 democracies reveals that countries with higher levels of female representation in their national legislatures also tend to have more comprehensive environmental protection laws. An analyst claims that electing more women causes stronger environmental policy. This conclusion is questionable primarily because it may overlook which of the following?
Answer: The possibility that countries with progressive social values are more likely to both elect women and support environmentalism.
This question asks the test-taker to identify a classic confounding variable. Progressive social values (the confounder) could independently cause both higher female representation (the independent variable) and stronger environmental laws (the dependent variable). This would create a spurious, non-causal relationship. (A) discusses policy implementation, which is a separate issue from policy formation and does not challenge the causal link. (C) offers a counter-argument about legislative priorities but doesn't address the methodological reason why the observed correlation might not be causal. (D) suggests reverse causation, which is a plausible alternative, but the presence of a strong underlying societal value system is a more comprehensive and common explanation for this type of macro-level correlation.
- A political scientist finds that when the Supreme Court's public approval rating is high, its rulings are more likely to be successfully implemented by federal agencies. She concludes that high public approval causes greater compliance. An alternative explanation is that a third variable is at play. Which of the following is the most plausible confounding variable?
Answer: The ideological alignment between the median justice on the Court and the prevailing national mood.
A plausible confounder must affect both the independent variable (Court's approval) and the dependent variable (implementation). The ideological alignment between the Court and the national mood (B) does this effectively. When the Court is in sync with the public's general ideology, its approval rating is likely to be high. Simultaneously, federal agencies, being part of a government that is also responsive to the national mood, are more likely to be ideologically aligned with the rulings and implement them successfully. This creates the observed correlation without requiring a direct causal link from approval to compliance. (A) is a plausible factor, but the broader 'national mood' is a more encompassing confounder. (C) and (D) are features of the cases themselves and are less likely to systematically drive both public approval and agency compliance.
- A researcher analyzes data from 50 countries and finds a negative correlation between a country's average tariff rate and its score on a democratic index. The researcher concludes that protectionist trade policies erode democracy. Which of the following describes a plausible alternative explanation involving reverse causation?
Answer: Countries with strong democratic institutions are more likely to have powerful consumer lobbies that advocate for lower tariffs and free trade.
Reverse causation means the presumed dependent variable is actually causing the presumed independent variable. In this case, it would mean that the level of democracy (DV) is causing the tariff rate (IV). Choice (A) provides a clear mechanism for this: established democracies have interest groups (consumer lobbies) that successfully push for lower tariffs. This reverses the causal arrow proposed by the researcher. (B) describes a confounding variable, not reverse causation. (C) describes a mediating variable (economic growth), which supports the researcher's original causal direction. (D) points to a measurement error, which is a different type of methodological critique.
- A researcher finds a strong positive correlation between the number of public libraries per capita in a city and its average citizen satisfaction score. The researcher concludes that building more libraries causes an increase in citizen satisfaction. Which of the following, if true, would most seriously weaken this causal conclusion?
Answer: The study controlled for city population size but not for median household income.
The correct answer identifies a potential confounding variable. Higher median income could independently lead to both more public funding for amenities like libraries and higher levels of citizen satisfaction, creating a spurious correlation. Not controlling for income is a major flaw. (B) mentions measurement error, which is a potential issue in any study but doesn't address the specific causal claim's primary weakness. (C) describes a moderating effect, where the relationship changes under different conditions, but it doesn't invalidate the underlying causal claim, only specifies its scope. (D) describes a mediating variable (literacy) and a causal mechanism, which would actually strengthen the researcher's conclusion, not weaken it.
- Based on the passage, the researcher concludes that the act of displaying a lawn sign causes homeowners to become more politically informed, perhaps to defend their choice to neighbors. Which of the following statements identifies the most likely confounder and the reason it is a confounder?
Answer: The homeowner's level of political engagement is the confounder because highly engaged individuals are more likely to both display a sign and actively seek out political information.
The most direct confounder is a person's pre-existing level of political engagement. A highly engaged person is predisposed to both perform an expressive political act (displaying a sign) and to be politically knowledgeable. The engagement is the common cause of both observed phenomena, making the correlation between the sign and knowledge spurious. (A) and (D) propose plausible confounders (wealth/education), but political engagement (C) is a more proximate and direct cause for both specific behaviors. (B) describes a potential difference between groups but doesn't explain the correlation itself as effectively as the underlying trait of engagement.
- A city government experiments with a new community policing program in one high-crime district. After a year, the crime rate in that district has fallen by 15%. The city council declares the program a success and moves to implement it city-wide. Which of the following potential confounding factors represents a 'history' threat to the council's conclusion?
Answer: During the same year, a national law was passed that provided significant funding for community-led youth programs in high-crime areas nationwide.
A 'history' threat refers to an external event that occurs at the same time as the treatment (the policing program) and could also affect the outcome (the crime rate). The passage of the national law (C) is a perfect example. It's an external event that provides an alternative explanation for the drop in crime in the district, confounding the effect of the community policing program. (A) describes a selection bias, not a history threat. (B) describes an event that would likely increase crime, so it doesn't explain the observed decrease. (D) describes a potential baseline difference, which is a problem for comparison, but not a history threat.
- A political scientist observes that in the United States, states with higher per-capita ice cream consumption have higher rates of voter turnout in presidential elections. Concluding that ice cream consumption boosts civic engagement is flawed. What is the most likely confounding variable responsible for this spurious correlation?
Answer: Median age of the state's population, since older populations might consume less ice cream and also vote at different rates than younger populations.
The most plausible confounder is the median age of the population. Older citizens have significantly higher rates of voter turnout, and they may consume ice cream at different rates (possibly lower) than younger populations. This third variable affects both of the observed variables. (A) and (B) are less direct; while temperature might affect ice cream sales, its link to turnout is tenuous. (C) is a creative but less likely explanation; the link between polarization and ice cream consumption is highly speculative compared to the well-established relationship between age and voting.
- A researcher observes that politicians who use humor more frequently in their public speeches tend to have higher approval ratings. The researcher suspects this correlation is spurious. To test for the influence of a confounding variable, which research design would be most effective?
Answer: Statistically comparing the approval ratings of humorous and non-humorous politicians while controlling for their underlying charisma and public speaking skill.
The most likely confounder is a politician's overall charisma or public speaking ability. A charismatic politician is likely to be skilled at using humor effectively and likely to have high approval ratings for reasons independent of the humor itself. The best way to test this is to statistically control for this third variable (D). By comparing humorous and non-humorous politicians who have similar levels of charisma, the researcher can isolate the specific effect of humor. (A) analyzes the independent variable in more detail but doesn't address the confounder. (B) relies on self-reporting, which can be unreliable. (C) is a better design (a before-and-after study) but is still susceptible to other events causing the change in approval; controlling for the confounder directly is a more robust approach.
- Refer to the scatter plot, which shows the relationship between GDP per capita and a Political Stability Index for a sample of 100 countries. An analyst concludes from the plot that increasing a country's wealth is a direct and effective strategy for ensuring political stability. Which statement represents the most sophisticated methodological critique of this conclusion?
Answer: The relationship appears to be correlational, and the causal direction could be reversed: political stability may be a prerequisite for attracting investment and achieving economic growth.
The most significant challenge to the analyst's causal claim is the problem of reverse causation (also known as endogeneity). While it's plausible that wealth leads to stability, it is equally, if not more, plausible that a stable political environment is necessary to enforce contracts, protect property, and foster an atmosphere conducive to long-term economic development. The scatter plot only shows a correlation; it cannot, by itself, establish the direction of the causal arrow. (B) and (D) are valid observations about the data, but they only qualify the relationship, they don't fundamentally challenge the causal inference in the way that reverse causation does. (C) is a critique of measurement, which is a valid but separate issue from causal direction.
- In a given year, the set of U.S. states that increased funding for public universities also saw a rise in voter participation among 18-24 year-olds. A researcher concludes that increased university funding causes youth political participation to rise. This conclusion is most vulnerable to criticism because:
Answer: the states that increased university funding may have been controlled by a political party that was also more effective at youth voter mobilization.
This question tests your understanding of causation versus correlation and the importance of controlling for confounding variables in research design. When researchers observe two trends occurring together, they must be extremely careful before claiming one causes the other.
The correct answer is D because it identifies the most serious threat to the researcher's causal claim: a confounding variable. If the same political party both increased university funding AND was particularly effective at mobilizing young voters through campaigns, rallies, or other outreach efforts, then the correlation between funding and participation could be entirely explained by this third factor. The university funding itself might have had no causal effect on voter participation at all.
Option A is wrong because identifying specific mechanisms of an effect (financial aid vs. facilities) doesn't challenge whether the effect exists—it just asks for more detail about how it works. Option B is incorrect because comparing college vs. non-college youth would help understand the mechanism but doesn't address whether a causal relationship exists in the first place. Option C suggests an alternative explanation (presidential election excitement), but this would likely affect all states similarly, so it wouldn't explain why only the states with increased university funding saw higher youth participation.
When evaluating causal claims in political science research, always ask: "What else could explain this correlation?" Look for confounding variables—factors that might influence both the supposed cause and the supposed effect. The strongest criticisms of causal arguments typically identify plausible alternative explanations that weren't controlled for in the research design.
- A study shows a positive correlation between the number of journalists a country has per capita and its level of government accountability. A causal conclusion that more journalists lead to more accountability is challenged by the possibility of a confounding variable. Which of the following variables is conceptually distinct from a confounder in this context?
Answer: Increased media reports on corruption, which leads to public outcry, forcing politicians to become more accountable.
This question requires distinguishing a confounding variable from a mediating variable. A confounder is a common cause of both the IV and DV. A mediator is part of the causal pathway between the IV and DV. Options (A), (B), and (D) are all plausible confounders—they could cause both a rise in the number of journalists and an increase in accountability. Option (C), however, describes a mechanism or mediating variable. It suggests the pathway is: more journalists (IV) -> more reports on corruption (mediator) -> more accountability (DV). Because it describes the process through which the causal relationship might work, it strengthens the original causal claim rather than challenging it as spurious.
- A study finds that individuals who report getting most of their political news from social media are more likely to believe in political conspiracies than those who rely on traditional newspapers. If a researcher wishes to test whether a confounding variable is responsible for this correlation, which of the following variables would be most crucial to control for in their statistical model?
Answer: The individual's pre-existing level of trust in government institutions.
This is a case of self-selection bias driven by a confounding variable. A pre-existing low level of trust in institutions (the confounder) could lead an individual to both avoid traditional news sources (like newspapers associated with the 'establishment') in favor of social media, and to be more receptive to conspiracy theories. Controlling for this variable is essential to isolate the true effect of social media. (A), (B), and (D) are plausible demographic or behavioral controls, but none is as directly and theoretically linked to both the choice of media and the belief in conspiracies as pre-existing institutional trust.
- A study using aggregate data from all 50 U.S. states finds that states with a higher percentage of citizens of immigrant origin have lower rates of violent crime. A senator cites this study to argue that immigration causes a reduction in crime. Why is this conclusion, drawn from this specific data, potentially an example of the ecological fallacy?
Answer: The observed state-level correlation may be driven by immigrants moving to economically vibrant states that have declining crime rates for other reasons.
The ecological fallacy occurs when an inference about individuals is drawn from aggregate-level data. The state-level correlation could be misleading. Option (B) points this out perfectly by identifying a confounding variable at the state level (economic vibrancy) that influences both where immigrants choose to settle and the crime rate. It could be that in these prosperous states, native-born citizens are committing fewer crimes, and this effect is stronger than any effect from the immigrant population. The aggregate data masks these individual-level dynamics, making the senator's causal claim about immigrants themselves suspect. (A), (C), and (D) are all valid methodological critiques, but they don't describe the ecological fallacy, which is specifically about the error of moving from aggregate to individual-level inference.
- The table below shows data for four fictional districts. A policy analyst observes that districts with higher spending on public schools per pupil also have higher average voter turnout. The analyst concludes that investing in schools causes an increase in civic participation. Which comparison of districts most effectively highlights a potential confounding variable?
Answer: Comparing District A and District D reveals that both school spending and voter turnout are correlated with the median household income of the district.
This question requires using the table to identify a confounder. By comparing District A and District D, we see a stark difference. District D has much higher school spending and much higher turnout than District A. However, it also has a much higher median household income. This comparison makes it clear that income is a potential confounding variable: higher income could cause both a willingness to pay higher taxes for schools and a higher propensity to vote (as income and voting are strongly correlated). This third variable provides an alternative explanation for the observed correlation. (A) simply restates the initial correlation. (B) shows that the correlation is not perfect but doesn't identify the confounder. (D) discusses the magnitude of the effect, not the validity of the causal claim.
- A study of international relations finds that when two countries increase their bilateral trade volume, the likelihood of them going to war with each other in the following decade decreases. The 'commercial peace' theory argues this is because trade creates economic interdependence, which makes war too costly. Which of the following, if true, would suggest the observed correlation is caused by a confounding factor?
Answer: Pairs of countries that are both established democracies are more likely to both trade heavily with each other and remain at peace.
This question asks to identify a confounder for the 'commercial peace' theory. The confounder must cause both increased trade and a decreased likelihood of war. Joint democratic governance (B) is the classic alternative explanation. The 'democratic peace' theory argues that democracies are unlikely to fight each other. It is also true that democracies are more likely to trade with each other due to shared norms, trust, and legal systems. Therefore, democracy could be causing both increased trade and peace, making the relationship between trade and peace spurious. (A) and (D) discuss aspects of the trade relationship but don't introduce a third variable that causes both trade and peace. (C) provides counterexamples, which may weaken the theory, but it doesn't offer a systematic, confounding explanation for the observed overall correlation.
- A comparative analysis finds that federal states (countries with a formal division of power between central and regional governments) have, on average, higher levels of ethnic conflict than unitary states. A scholar concludes that federalism causes ethnic conflict by institutionalizing and hardening ethnic divisions. Which of the following is a plausible confounding factor that challenges this conclusion?
Answer: Federalism is often adopted in countries that already have deep ethnic divisions, as a political compromise to hold the country together.
This is a problem of selection bias. The choice to adopt a federal system is not random. Countries with significant, pre-existing ethnic cleavages are more likely to choose federalism precisely to manage those tensions. In this case, the pre-existing ethnic diversity is the confounder: it leads to the adoption of federalism (the IV) and is also the underlying cause of ethnic conflict (the DV). This makes it seem like federalism is causing the conflict when it might simply be a response to it. (B) is a similar and related point but (A) more directly states the selection effect. (D) describes a potential mechanism through which federalism might exacerbate conflict, which would support the scholar's claim. (C) is a measurement critique, not a confounding issue.
