ㅉW3: Simple + Multiple regression

Question 1

Regression assumption

Answer 1

1. Independence of observation
2. Normal distribution (histogram, P-P plot, shapiro wilk
3. Homoscedasticity (constant variance)
4. Linearity
5. No colinearity (ONLY multiple regression)

Question 2

Assumptions for homoscedesticity

Answer 2

1. Assumption check via bivariate scatter plot, and residual-vs-fitted plot
   
   • rvfplot: fanning indicates violated assumption

Question 3

Assumption for linearity

Answer 3

1. scatter plot (is there any clear non-linear trend)
2. rvfplot, yline(0) - wants no trend, no picture, no pattern

Question 4

Normal probability plot (P-P plot) assumption

Answer 4

• Most common non-normal esiduals are snaking around the line
• Stata syntax: pnorm residvarname