POL 51 Lecture Notes - Lecture 9: Confounding, Bruce Russett
Document Summary
Wwhen the economy is stronger, media coverage becomes more positive. When the economy is weaker, media coverage is more negative. > we would not expect a larger incvote. If the explanation for why x causes y is z, then if x changes but z does not, we would not expect y to change. Confounding and intervening variables consider an x/y causal hypothesis: z is an intervening variable if x causes z and z causes y. In this case, the x/y correlation dissapears when we control for z. X) in the case of an intervening variable, if the value of x changes, it causees the value of z to change, which causes the value of y to change. Z is sometimes referred to as the causal mechanism. Another example: the democratic peace divide countries into democracies and nondemocracies create dyads, pairs of countries dyads that include two democracies experience less military conflict than dyads that include one or no democracies.