Module 1: Multiple Regression Analysis

Question 1

When do we have statistical significance?

Answer 1

When we reject our null hypothesis and have low p-values and high F-statistics

Question 2

T/F: Failure to reject the null is the same as proving x does not influence y.

Answer 2

F: We cannot draw that conclusion from lack of statistical significance. Only that we cannot reject the null in this experiment/regression

Question 3

Lack of staitstical significance is not:

Answer 3

1)Evidence that x should be removed fro the regression
2) Proof of no relationship between x and y

Question 4

Statistical significance is not:

Answer 4

1) Evidence that x belongs in the regression
2) Proof of a causal relationship between x and y
3) Evidence that the magnitude of the relationship is meaningful

Question 5

Given the following regression, interpret tenure.

wage=-0.815+0.537educ+0.164tenure -1.790female + 0.132nonwhite

where wage is in $ and tenure is in years

Answer 5

Each year longer with one’s current employer is associated with $0.164 greater hourly wage holding ceteris paribus.

Something along these lines is fine.

Question 6

What are null and alternative hypothesis?

Answer 6

Null is the one we want to reject H0: Bj = 0
Alternative is the one we are trying to prove Ha: Bj !=0

Question 7

Interpret female for the following regression

wage=-0.815 0.537educ+0.164tenure-1.79female-0.132nonwhite

where wage is in $ and if female = 1

Answer 7

We estimate that women earn $1.79 less per hour compared to men holding all else equal.