POL SCI 3 Lecture Notes - Lecture 5: Causal Inference, Confounding, Statistical Hypothesis Testing
Document Summary
Hypothesis: as movie ticket prices go up, attendance goes down. X = price of ticket, y = number of times attended. Ce (x=10 to x=15) = y (x=10 to x=15) Change x (actual to cf) change y (actual to cf) = ce. Fundamental problem of causal inference: ce cannot be observed. Education and income, hyp: x=education has ce on y=income. Confounding variables zs that: (a) have an ce on y, (b) either also have a ce on x or are merely correlated with x, (c) are not directly affected by x or y. X and y correlated, z y, z x. Cannot have reverse arrows (x z or y z) according to (c) X and y correlated, z y, z correlated with x. In education and income hypothesis: z as aptitude of achievement (general predisposition) that affects income acquisition and your education.