LinearRegression2

# Or in other words what is var b and var b 0 1 to

This preview shows page 1. Sign up to view the full content.

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

Unformatted text preview: er words, what is Var (b ) and Var (b ) ? 0 1 To answer this question, we need an additional assumption:2 Homoskedasticity: Assume Var (u | x ) = s 2 [SLR.5] VER. 9/25/2012. © P. KOLM 12 What Does Homoskedasticity Mean? VER. 9/25/2012. © P. KOLM 13 What Does Heteroskedasticity Mean? VER. 9/25/2012. © P. KOLM 14 Sampling Variance of the OLS Estimators (2/2) It can be shown that: ˆ Var (b1 ) = s2 n å (x i =1 2 ˆ Var (b0 ) = sn i -1 - x )2 n åx i =1 n 2 i å (x i - x )2 s2 = SSTx 2 sn = -1 n åx i =1 2 i SSTx i =1 VER. 9/25/2012. © P. KOLM 15 n ˆ Proof: Recall: b1 = b1 + å (x i =1 i - x ) ui SSTx Therefore, n æ ö ÷ çb + 1 ˆ ) = Var ç Var (b1 å (x i - x )ui ÷ ÷ ç 1 SST ÷ ç è ø x i =1 æn ö 1 ç (x - x )u ÷ ÷ = Var çå i i÷ ç i =1 ÷ SSTx2 è ø = = = = n 1 Var ((x i - x )ui ) 2å SSTx i =1 n 1 å (x i - x )2Var (ui ) SSTx2 i =1 s2 n å (x i - x )2 SSTx2 i =1 s2 s2 SSTx = 2 SSTx SSTx ˆ Var (b0 ) follows in a similar way (For you: Do the calculation!) VER. 9/25/2012. © P. KOLM 16 s2 ˆ Var (b1 ) = SSTx 2 ˆ Var (b0 ) = sn -1 n åx i =1 2 i SSTx Remarks3 Let’s make sure the following make sense: The larger the error variance, s2 , the larger the variance of the intercept and slope estimates The larger the variability in the x i , the smaller the variance of the slope estimate A larger sample size typically decreases the variance of the intercept and slope estimates...
View Full Document

## This document was uploaded on 02/17/2014 for the course COURANT G63.2751.0 at NYU.

Ask a homework question - tutors are online