Beyond Correlations: Inferences and Theory

Question 1

From Concepts to Variables

Answer 1

• Concepts in political science, such as democracy, power, or ideology, are essential for understanding and studying political phenomena. However, to make them operational in research, scholars often use variables.
• Dependent variables are the outcomes or phenomena that researchers want to explain. For example, when studying political behavior, voter turnout can be a dependent variable.
• Independent variables are factors or variables that researchers believe influence or cause changes in the dependent variable. These are the variables that researchers analyze to see if they have an effect on the outcome.

Question 2

Relationships Between Variables

Answer 2

• Correlation refers to the statistical relationship or covariation between two or more variables. Researchers analyze correlations to identify patterns or associations between variables.
• Correlations can be either direct or inverse (positive or negative). A direct correlation implies that as one variable increases, the other also increases. An inverse correlation means that as one variable increases, the other decreases.
• Some relationships can be nonlinear. For example, the relationship between democratization and civil war, as described by Saideman et al. (2002), is like an “upside-down bowl,” indicating that the relationship is more complex than a simple linear correlation.

Question 3

Spuriousness

Answer 3

• Spuriousness is an important concept in understanding correlations. It refers to situations where correlations between variables are misleading because of hidden or unobserved factors.
• The example used in the presentation involves the correlation between higher temperatures and both increased ice cream sales and drowning deaths. This is a spurious correlation because the actual cause is the summer season, which affects both ice cream sales and swimming, leading to drowning deaths.

Question 4

Real World Example

Answer 4

American Recovery and Reinvestment Act (2009). The distribution of funds from this act was initially attributed to partisanship, with Democratic districts receiving more money.
* Nate Silver challenged this explanation, suggesting that the presence of a state capital was a significant factor influencing the distribution of funds. Districts with state capitals are often controlled by Democrats due to demographic factors.
* This example illustrates the importance of examining correlations carefully and considering alternative explanations, especially in political research.

Question 5

Descriptive and Causal Inference

Answer 5

• Descriptive inference involves using observed data to understand unobserved phenomena. For instance, if data shows that average household spending decreased as pandemic lockdowns increased, descriptive inference can be used to understand why this occurred.
• Causal inference goes beyond simple description. It requires establishing a counterfactual scenario, which is what would have happened if certain conditions had not occurred. Causal inference also requires demonstrating a causal mechanism, which is the process or pathway through which a relationship is realized.
• To establish causal relationships rigorously, researchers must rule out spuriousness and alternative explanations and establish a proper temporal order of events.

Question 6

Intervening Variables

Answer 6

• Intervening variables are the steps or factors that link independent and dependent variables in a causal argument. They help clarify the sequence of events in a causal relationship.
• For example, if the research question is whether increased oil exports lead to authoritarian rule, intervening variables could include taxation (which decreases accountability), spending (due to oil wealth), and repression (fueled by lavish military spending).

Question 7

Causal Mechanisms

Answer 7

• Causal mechanisms are the logical processes or pathways that explain how a relationship between variables is realized. They provide insight into the “how” and “why” behind causal relationships.
• In a study by Thelen (2018) on the introduction of Uber, the political-economic context served as a causal mechanism that channeled conflict in different ways, resulting in varying policy outcomes. This context helped explain why different regions had different responses to Uber.

Question 8

Theories in Political Science

Answer 8

• Theories are comprehensive explanations of why political events and relationships occur the way they do. Theories are fundamental to understanding and making sense of the political world.
• Theories help in hypothesis development, which are educated guesses about some aspect of the political world. Researchers use theories to inform their hypotheses, and these hypotheses guide their research and analysis.
• Scholars often use both deductive (starting with theory and testing hypotheses) and inductive (building theory from data) approaches in their research.