Lecture+1+Review+SLR

# Isanunbiased estimatoroftheparameter if e

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: e actual data points (because your sample is not informative about what is going on far away from the sample data points). If xi = 0 is not sufficiently near the actual data, the intercept does not have much meaning by itself, it only serves to locate the regression line. ˆ • In this case, α = −23.5 implies that for a person with 0 years of schooling, the predicted hourly wage is ‐\$23.5. This does not make much sense. In our data, X is between 12 to 16. This is far away from 0. So the data do not have much information around X=0. Figure: Impact of the range of data on estimates Goodness of fit How do we measure how well the sample regression line fits the data? ˆˆ ˆ ˆ yi = yi + ui (by definition, ui = yi − yi ), so each yi is divided into an explained part ˆ ˆ ( yi ), and an unexplained part ( ui ). We can then define the following: • ∑ • sample variance of Y multiplied by n‐1) n ˆ ˆ ∑ i=1 ( yi − y )2 is the explained sum of squares (SSE) ( =total variation in yi , ˆ ˆ i.e., the sample variance of yi multiplied by n‐1. Recall that y = y ,) n...
View Full Document

## This document was uploaded on 03/11/2014.

Ask a homework question - tutors are online