Take the quiz test your understanding of the key concepts covered in the chapter. Try testing yourself before you read the chapter to see where your strengths and weaknesses are, then test yourself again once you’ve read the chapter to see how well you’ve understood.

1. Which one of these statements is not a Gauss-Markov assumption?

That the error term has a conditional mean of zero

Absence of influential observations.

That the error term has constant variance

That the errors are uncorrelated

Answer:

b. Absence of influential observations

2. Why should we not include irrelevant variables in our regression analysis?

Your R-squared will become too high

Because of data limitations

It is bad academic fashion not to base your variables on sound theory

We increase the risk of producing false significant results

Answer:

d. We increase the risk of producing false significant results

3. How can we deal with the breach of the assumption of linearity?

Include a squared term

Include an interaction term

Use robust regression

Use the margins command

Answer:

a. Include a squared term

4. What is the best way to find the exact top or bottom point of a squared effect?

Through derivation using values from the two coefficients

Excluding the squared term and predicting

Including the squared term and predicting

Graphing the results and comparing the top/bottom point with the value on the X-axis

Answer:

a. Through derivation using values from the two coefficients

5. Name another way of modeling non-linearity

Using the linktest command

Using an interaction term

Using dummy variables

Using a bivariate regression model

Answer:

c. Using dummy variables

6. Which statistics can help us detect multicollinarity?

Variance inflation factor (VIF)

F-statistic

Durbin-Watson

Tolerance values (1-VIF)

Answer:

a. Variance inflation factor (VIF)

d. Tolerance values (1-VIF)

7. What does heteroskedasticity mean?

The variance in the residuals are the same regardless of their predicted values.

There is variance in the residuals

That we are unable to produce residuals

The variance in the residuals differ depending on their predicted values

Answer:

d. The variance in the residuals differ depending on their predicted values

8. What are the two ways we can check for heteroskedasticity?

We can examine a plot of predicted values vs the residuals

We can run the Hausman test

We can run the hettest command

We can compare the F-test of two models

Answer

a. We can examine a plot of predicted values vs the residuals

c. We can run the hettest command

9. Which one is not a measure of influential (or potentially influential) units?