Experiments

1
Experiments
2
Definition

• Making observations in circumstances so arranged
  or interpreted that we can isolate the factors
  relevant to our particular inquiry

3
Experiments are good for

• Intervening in natural processes
• Isolation of causal mechanism
• Diversity in data

4
Traditional Experimental Methods

• Researcher control over independent variable
• Pretest/posttest
• Treatment/control
• Random Assignment

5
Method Modifications

• More groups
• More treatments

6
Quasi Experiments

• Ex post facto
• Natural Experiments

7
Why Experiments in Political Science?

• Best assessment of Causal Relationships
• Enables complex phenomena to be decomposed
• Accelerate interdisciplinary work
• More likely to produce anomalous facts
• Travels across levels of aggregation

8
Threats to Internal Validity

• Lack of appropriateness
• Treatment misidentification
• Expectancy of Experimenters
• Placebo effect
• Hawthorne effect

• Selection bias.
• Non-Randomization
• Differential attrition
• Endogenous Change
• Testing
• Maturation
• Regression

9
Threats to Internal Validity

• External events
• Contamination
• Compensatory rivalry
• Demoralization

10
Threats to External Validity

• Generalizability of sample
• Results instability