{[ promptMessage ]}

Bookmark it

{[ promptMessage ]}

slides_5_conddist

# ∙ intuitively we compute the expected value of y

This preview shows pages 41–48. Sign up to view the full content.

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: ∙ Intuitively, we compute the expected value of Y for each slice of the population determined by the values x 1 , x 2 ,... and then we weight these by the share of the population accounted for by each x r . 41 ∙ To show LIE, suppose Y and X are both discrete. Then, using f X , Y x , y f Y | X y | x f X x , E Y ∑ s 1 y s f Y y s ∑ s 1 ∑ r 1 y s f X , Y x r , y s ∑ s 1 ∑ r 1 y s f Y | X y s | x r f X x r ∑ r 1 ∑ s 1 y s f Y | X y s | x r f X x r ∑ r 1 E Y | X x r f X x r ∑ r 1 x r f X x r E X . 42 ∙ There is a hand-waving step when the order of the summations is switched, but it is valid under the assumption that E | Y | . With a finite number of outcomes there is no issue about switching the order. ∙ A similar argument works in the continuous case, but we need to switch the order of the integration. 43 EXAMPLE : Suppose hourly wage Y is related to education X as E Y | X 5.8 .40 X If the average education level in the population is 12.5, then we can find the average hourly wage: E Y E 5.8 .40 X 5.8 .40 E X 5.8 .40 12.5 10.8, or \$10.80. 44 ( ce4 ) (General Version of the LIE) Let X and Z be random vectors and assume E | Y | , so that E Y | X and E Y | X , Z exist. Call these 1 X E Y | X and 2 X , Z E Y | X , Z . Then E 2 X , Z | X 1 X , which we more commonly write in a simpler form: E E Y | X , Z | X E Y | X 45 EXAMPLE : Suppose E Y | X , Z 1 X 2 Z and E Z | X 1 X . Then E Y | X E 1 X 2 Z | X 1 X 2 E Z | X 1 X 2 1 X 2 1 2 1 X This simple example is fundamental to understanding “omitted variables bias” in regression analysis. 46 ( ce5 ) If Y and X are independent then E Y | X E Y . This is obvious because D Y | X D Y under independence. Somewhat less obvious is that if X , Y is independent of Z , then E Y | X , Z E Y | X ∙ It is not enough to assume X is independent of Z and Y is independent of Z . Generally, to ensure E Y | X , Z E Y | X , we need D X , Y | Z D X , Y , and this does not follow from D X | Z D X and D Y | Z D Y ....
View Full Document

{[ snackBarMessage ]}

### Page41 / 67

∙ Intuitively we compute the expected value of Y for each...

This preview shows document pages 41 - 48. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online