Assessing the model

Slides

• “Goodness of fit”
• “Criticizing the model and checking assumptions”

Control questions:

• What is the interpretation of \(R^2=0.80\)?
• If the correlation between \(X\) and \(Y\) is \(0.5\), what is \(R^2\) in the regression \(Y=\beta_0+\beta_1 X+\epsilon\)?
• What are the assumptions of a linear regression?
• How do we assess whether the linearity assumption is fulfilled?
• How do we assess whether the residuals are homoskedastic (have constant variance)?
• What is QQ-plot and what do we use it for?
• Do we need normality of the errors when estimating coefficients?
• Do we need normality when making prediction intervals?