Why can’t we deduce the effectiveness of sanctions by only looking at instances where sanctions were imposed?
- A researcher cannot explain the determinants of x by exclusively looking at instances where x is true. This is called selecting on the dependent variable.
- For example, we cannot infer the effect of water on death by only looking at people who died. The vast majority of individuals who have died were consuming water, so such a study would (falsely) lead us to believe that water causes death.
- Similarly, we cannot infer whether sanctions are effective or not by only looking at instances where sanctions were imposed. We need to compare those cases to situations where sanctions were not imposed.
- Because sanctions are often used as a last resort, they appear less effective than they really are.