Statistics 10 questions.

Test #2A

· Choose 10 of the 14 questions to answer.

· I will grade the first 10 questions you answer. For example: if you answer #1-12, I will grade #1-10 (the first ten). If you answer #2-12, I will grade #2-11 (the first ten).

· Each question is worth 1 point. Zero points will be assigned for questions without answer

1) Why do we use the adjusted r-squared in multiple linear regression?

2) What is the purpose of saying “all-else-equal” or “ceteris paribus” when it comes to multiple linear regression?

3) Why is the F-test important in multiple linear regression more so than in simple linear regression?

4) What is the difference between the AIC and BIC?

5) Why is outlier analysis so important?

6) What is the danger in using stepwise regression?

7) What is the difference between Cook’s Distance and DFFITS?

8) What can you do to remove multicollinearity?

9) What is a VIF?

10) If you have to create dummy variables for the seven continents of the world, how many columns do you create and why?

11) Why do we standardize residuals?

12) Describe a QQ plot and what it can tell us.

13) When you build a confidence interval of coefficients of regression using r, we do not divide by the square root of n. Why?