Lecture 6 - MLR B

Question 1

What information can be gleaned from a Scatter plot regarding an MLR?

Answer 1

Whether a linear relationship is appropriate.

Question 2

What information can be gleaned from a Residual plot regarding an MLR? (2 things)

Answer 2

Whether a linear relationship is appropriate and to verify the error assumptions.

Question 3

If a residual plot shows a pattern of alternating clumps of positive and then negative errors what assumption of MLR has likely been violated?

Answer 3

Independence of error terms (their is some autocorrelation)

Question 4

If a residual plot shows a change in variability w.r.t. x, what assumption of MLR has likely been violated?

Answer 4

Constant variance. The homoscedasticity assumption has been violated.

Question 5

If a residual plot shows error terms significantly more frequently scattered above the zero line, what assumption of MLR has likely been violated?

Answer 5

The normality of errors assumption.

Question 6

What is evaluated by the Durbin-Watson test?

Answer 6

Whether auto-correlation is present.

Question 7

What is the typical range of the Durbin Watson test? What value indicates no auto-correlation?

Answer 7

Range 0 - 4

Values close to 2 show no auto correlation.

Question 8

What is the null hypothesis for the Durbin-Watson test?

Answer 8

That there is no auto-correlation.

Question 9

Is the Durbin-Watson test one or two tailed?

Answer 9

It can be either.

Question 10

What is multi-collinearity? Which of the MLR assumptions does it violate?

Answer 10

Multi-collinearity refers to the situation when two (or more) of the X variables are highly correlated. It violates the independence of variables assumptions.

Question 11

What is the difference between collinearity and autocorrelation?

Answer 11

Collinearity refers to the correlation of X variables while autocorrelation refers to the correlation of error terms.

Question 12

What are the effects of significant multicollinearity?

Answer 12

1) The signs (+ or -) of cooefficients may be the opposite of expectations.
2) The standard errors of the regression coefficients will be overestimated (F, t, R squared and hypothesis testing invalid.

Question 13

How is collinearity assessed?

Answer 13

The correlation coefficient for each pair of variables is computed and assessed.

Question 14

What is the variance inflationary factor?

Answer 14

An objective means of assessing the collinearity of data.

Question 15

What is the range of VIF that indicates multicollinearity?

Answer 15

> 5

Anything less than 5 has insignificant collinearity.

Question 16

How are Qualitative variables accounted for by MLR?

Answer 16

By ‘coding’ dummy variables as 0 for no and 1 for yes.

Question 17

Define an interaction.

Answer 17

An interaction occurs if the effect of an explanatory variable on the dependent variable is, in turn, dependent on the value of another explanatory variable in the economy.

Question 18

How can interaction effects be accounted for in an MLR model?

Answer 18

By including an interaction term which is the product of the two explanatory variables that are believed to have a significant interaction effect.