Eco finance

  1. Use ?evals to get information about the data and variables in it
  2. Simple regression of teaching score on beauty score
  1. Have a scatterplot of the teaching score on beauty score, with a smooth line added. Does there seem to be a linear relationship? A positive or negative linear relationship?
  2. Run the regression. What is the expected increase in teaching score when the beauty score increases by 1?
  3. Get the raw, standardized, and studentized residual of the model.
    Have a scatterplot matrix and correlation matrix for the three residuals.
  4. Using raw residual, get the residual vs. x and histogram of the residuals. Identity if there are any violations of the four assumptions for regression.
  5. What is the R-squared of the model? Interpret it.

3. Multiple regression of teaching score on beauty score and gender

  1. Have a scatterplot of teaching score on beauty score, with different colors for males and females. Add smooth lines separately for male and female
  2. Will you choose a model with interaction or without interaction?
  3. With the model you chose in part b, what is the expected increase in male instructors’ teaching scores when beauty scores increase by 1? What is the expected increase in female instructors’ teaching scores when beauty scores increase by 1?

4. Multiple regression of teaching score on beauty score and age

  1. Will you choose a model with interaction or without interaction?
  2. With the model you chose in part a, what is the expected increase in teaching score when beauty score increases by 1 unit, for all
    faculties with age = 37? With age = 57?
  3. At what age level does beauty score negatively affect teaching score?
  4. Have the regression plane for your model in part a.

5. Stepwise and best subset selection

  1. Use pent=0.05 and prem=0.05 to get the stepwise regression model
  2. Is the ID variable in the model you got in part a? If yes, remove it.
  3. Run the best subset selection with the mode you got in part b.
    According to AIC, which subset model will be the best? According to BIC (Schwarz’s), which subset model will be the best?