Deck 9: Regression

Question 1

If Forbes includes years of education and age to predict CEO salary, what would multicollinearity represent?

A) High correlation between CEO age and years of education
B) A linear relationship of both predictor variables and CEO salary
C) Both predictor variables correlating with variables not included in the regression model
D) Independence between years of education and age

Answer 1

High correlation between CEO age and years of education

Question 2

Which of the following is not a reason for multicollinearity being a problem in regression?

A) It limits the size of R.
B) It makes it difficult to assess the importance of individual predictors.
C) It leads to unstable regression coefficients.
D) It creates heteroscedasticity in the data.

Answer 2

It creates heteroscedasticity in the data.

Question 3

Which of the following statements about the t-statistic in regression is not true?

A) The t-statistic tests whether the regression coefficient, b, is equal to 0.
B) The t-statistic provides some idea of how well a predictor predicts the outcome variable.
C) The t-statistic can be used to see whether a predictor variable makes a statistically significant contribution to the regression model.
D) The t-statistic is equal to the regression coefficient divided by its standard deviation.

Answer 3

The t-statistic is equal to the regression coefficient divided by its standard deviation.

Question 4

The distances of cases from the model mean are called?

A) Leverage values
B) Hat values
C) Standard distances.
D) Mahalanobis distances

Question 5

Which regression assumption does the Durbin-Watson statistic assess?

A) Linearity
B) Independence of errors
C) Homoscedasticity
D) Multicollinearity

Question 6

The following graph shows evidence of?

A) Heteroscedasticity
B) Non-linearity
C) Heteroscedasticity and non-linearity
D) Regression assumptions that have been satisfied.

Question 7

A business analyst wanted to predict the success of a government intervention in failing firms, predicted by a categorical variable, 'duration of financial difficulty prior to the intervention'. The predictor variable had four categories: less than 6 months, 6-12 months, 1-2 years, more than 2 years. They needed to code these variables into dummy variables for the regression using less than 6 months as their control category. Which of the following represents the correct coding scheme?

A)
B)
C)
D)

Question 8

Forbes has modelled CEO salary by age for an exclusive feature in the magazine. Which of the following statements is not true about outliers in the model?

A) Outliers are values where CEO salary against age is very different from others in the data set
B) Outliers bias the mean.
C) Outliers bias regression parameters (if included in the model).
D) Outliers only represent influential CEOs in the data set.

Question 9

Which of these statements is not true?

A) If the average variance inflation factor is greater than 1, the regression model might be biased.
B) Tolerance values above 0.2 may indicate multicollinearity in the data.
C) Multicollinearity in the data is shown by a VIF (variance inflation factor) greater than 10.
D) The tolerance value is calculated by 1 / VIF (variance inflation factor)

Question 10

Which of the following statements about the F-ratio is true?

A) The F-ratio is the ratio of variance explained by the model to the error in the model.
B) The F-ratio is the ratio of variance explained by the model to the total variance in the outcome variable.
C) The F-ratio is the ratio of error variance to total variance.
D) The F-ratio is the proportion of variance explained by the regression model.

Question 11

Using the model in equation (8.16), how many records would be sold if £29,000 was spent on brand advertising, the track was played 19 times on radio and the band were rated 7 on the attractiveness scale?

A) 2,461,660 records
B) 2435 records
C) 2488 records
D) 2, 435, 050 records

Question 12

Which of these problems arise as collinearity increases?

A) Untrustworthy beta coefficients.
B) The size of R is constrained
C) It is more difficult to estimate the individual importance of a predictor
D) All of the above.

Question 13

An online retailer uses the number of website clicks to predict sales revenue for an online hair-replacement (wig) company. The R² statistic represents?

A) The percentage of variance in the predictor (clicks) variable accounted for by the outcome (sales revenue) variable.
B) The proportion of variance in the outcome variable (sales revenue) accounted for by the predictor variable (clicks).
C) The proportion of variance in the predictor variable (clicks) accounted for by unknown variables.
D) The percentage of variance in the outcome variable (sales revenue) accounted for by unknown variables.

Question 14

The difficulty with using one regression equation to predict values in a different set of data is called?

A) Shrinkage
B) Contraction
C) Reduction
D) Washing

Question 15

Management Research magazine has recently published an article on discrete categories. Which of the following strategies for modelling discrete data do you think they suggest?

A) Stupid variables
B) Dummy variables
C) Imitation variables
D) Faking variables