Political Science 101 POLS: Political Methodology - Causal Inference Counterfactuals RCTs Instrumental Variables DifferenceinDifferences Regression Discontinuity

1. What This Is & Why It Matters

Causal Inference is the process of establishing cause-and-effect relationships between variables in social science research. This concept is central to political science because it allows researchers to identify the underlying factors driving policy outcomes, electoral behavior, and institutional performance. Without understanding causal inference, you can't explain why some democracies collapse, why certain policies fail, or why some leaders succeed.

Consider the example of the 2016 US presidential election. Researchers used causal inference techniques to investigate the impact of social media on voter behavior. By employing instrumental variables analysis, they found that exposure to fake news on Facebook significantly increased the likelihood of voting for Donald Trump. This study highlights the importance of causal inference in understanding the complex relationships between variables in politics.

2. Essential Thinkers, Concepts & Models

• Counterfactuals: A hypothetical scenario that explores what would have happened if a particular event or decision had occurred differently. Why it's still relevant: Counterfactuals are used in historical analysis to evaluate the consequences of alternative outcomes.
• Randomized Controlled Trials (RCTs): An experimental design that randomly assigns participants to treatment or control groups to measure the effect of a variable. Why it's still relevant: RCTs are widely used in policy evaluation to assess the impact of interventions.
• Instrumental Variables (IVs): A variable that affects the outcome variable but is not directly related to the treatment variable. Why it's still relevant: IVs are used to address endogeneity and selection bias in causal inference.
• Difference-in-Differences (DiD): A method that compares the change in outcomes between treatment and control groups over time. Why it's still relevant: DiD is used to evaluate the impact of policies or interventions that are implemented at different times.
• Regression Discontinuity (RD): A method that exploits the discontinuity in a regression line at a specific threshold to estimate the effect of a variable. Why it's still relevant: RD is used to evaluate the impact of policies or interventions that are implemented at a specific threshold.
• Neyman-Rubin Causal Model: A framework that defines causality in terms of counterfactuals and potential outcomes. Why it's still relevant: The Neyman-Rubin model provides a foundation for causal inference in social science research.
• Rubin's Causal Model: An extension of the Neyman-Rubin model that defines causality in terms of treatment and control groups. Why it's still relevant: Rubin's model provides a framework for evaluating the impact of interventions.

3. Step-by-Step 'Political Analysis'

1. Formulate a research question: Identify a policy or phenomenon that you want to investigate and formulate a clear research question.
2. Choose a research design: Select a research design that is appropriate for your research question, such as an RCT, IV, DiD, or RD.
3. Collect data: Gather data that is relevant to your research question and design.
4. Estimate the effect: Use statistical methods to estimate the effect of the variable on the outcome variable.
5. Interpret the results: Interpret the results in the context of your research question and design.

4. Common Student Mistakes

• Misconception: Assuming that correlation implies causation.
• The right view: Correlation does not imply causation. You need to use causal inference techniques to establish a cause-and-effect relationship.
• Misconception: Thinking that RCTs are the only way to establish causality.
• The right view: While RCTs are a powerful tool, they are not the only way to establish causality. Other methods, such as IVs and DiD, can also be used.
• Misconception: Believing that causal inference is only relevant for experimental designs.
• The right view: Causal inference is relevant for all research designs, including observational studies and surveys.

5. Exam/Essay Tips

• Multiple-choice questions: Be careful to distinguish between different causal inference techniques, such as RCTs and IVs.
• Free-response questions: Use specific examples to illustrate the application of causal inference techniques.
• Trap distinctions: Be aware of the differences between different research designs, such as RCTs and DiD.
• Integrating into an argumentative essay: Use causal inference techniques to support your argument and provide evidence for your claims.

6. Quick Practice Scenario

A researcher wants to evaluate the impact of a new policy on voter turnout. They use a DiD design and find that the policy increased voter turnout by 10%. However, they also find that the policy had a larger impact on voters who were already likely to vote. What is the main limitation of this study?

Answer: The main limitation of this study is that it does not account for selection bias, which may have affected the results.

7. Last-Minute Cram Sheet

• Counterfactuals: A hypothetical scenario that explores what would have happened if a particular event or decision had occurred differently.
• Randomized Controlled Trials (RCTs): An experimental design that randomly assigns participants to treatment or control groups to measure the effect of a variable.
• Instrumental Variables (IVs): A variable that affects the outcome variable but is not directly related to the treatment variable.
• Difference-in-Differences (DiD): A method that compares the change in outcomes between treatment and control groups over time.
• Regression Discontinuity (RD): A method that exploits the discontinuity in a regression line at a specific threshold to estimate the effect of a variable.
• Neyman-Rubin Causal Model: A framework that defines causality in terms of counterfactuals and potential outcomes.
• Rubin's Causal Model: An extension of the Neyman-Rubin model that defines causality in terms of treatment and control groups.
• 'Separate but equal' was overturned by Brown v. Board – Plessy v. Ferguson was the earlier, racist ruling.
• Correlation does not imply causation.
• RCTs are not the only way to establish causality.

8. Further Study Resources

• Textbooks: American Government: Stories of a Nation by Brinkley, Government by the People by Magleby, et al.
• Khan Academy units: Causal Inference, Experimental Design, and Regression Analysis.
• YouTube channels: Crash Course Government, 3Blue1Brown, and CGP Grey.