Deck 17: Multiple Regression

Question 1

Suppose you have four predictor variables and you want to use them to statistical predict an outcome variable. Which of the following techniques would be best to generate a predication equation?

A) Correlation
B) Multiple Regression
C) Exploratory Factor Analysis
D) Discriminant Function Analysis

Answer 1

B

Question 2

If you are interested in the original metrics of your variables, you should use ______ regression coefficients.

A) unstandardized
B) standardized
C) centered
D) uncentered

Answer 2

A

Question 3

If we want to see the percentage of variance in our outcome variable accounted for by our predictor variables, we need to examine which of the following statistics? 

A) rₓᵧ
B) B
C) R²
D) R

Answer 3

C

Question 4

When three predictors are highly correlated with each other, which of the following is likely?

A) interpretation of individual regression coefficients is difficult
B) the overall R² of the prediction equation may be inflated
C) more predictor variables are needed to achieve significance
D) All of the Above

Question 5

A regression equation that was computed on one sample should be applied to a new sample to determine whether the regression coefficients from the first sample capitalize on chance. What factor is associated with less capitalization on chance?

A) Large number of predictor variables
B) Large sample size
C) Samples that are older in age
D) Skewed predictor variables.

Question 6

When building a multiple regression equation to increase prediction, you should choose variables that ______.

A) have high correlations with outcome variables
B) have weak correlations with other predictor variables
C) can be reliably measured
D) all of the above

Question 7

The adjusted R² controls for which of the following factors to correct the obtained R² ______________.

A) reliability of predictor variables
B) reliability of predictors and outcome variable
C) sample size and number of predictors
D) multicollinearity among predictors

Question 8

If your boss wants to figure out how much variance your regression equation should account for in a new sample, which statistic should you report to her?

A) R²
B) Cross-validated R²
C) Adjusted R²
D) Standard Error of Prediction

Question 9

Which of the following statistics provides the most direct estimate of the average amount of error in regression equation predictions?

A) Standard Error of the Estimate
B) Standard Error of Measurement
C) Standard Error of Dependency
D) Standard Error of Develoment

Question 10

Which of the following are key challenges in combining predictors in multiple regression as discussed by Shultz, Whitney, and Zickar (2021)? ________.

A) Choosing predictors that do not overlap with each other
B) Choosing predictors that will maximize prediction
C) Making sure to avoid capitalizing on chance
D) All of the Above

Question 11

What are strategies for avoiding capitalizing on chance in multiple regression?

Question 12

What are problems that multicollinearity cause with multiple regression in terms of prediction.

Question 13

Why might you use a standardized regression equation versus an unstandardized regression equation?

Question 14

What are key statistics that should be used to evaluate the effectiveness of a regression equation for prediction?

Question 15

Suppose you are designing a selection battery that could be used to predict success for a particular job, food service preparation employee. You have several suggested predictor variables and your boss suggested that you use multiple regression to develop an equation used for weighting and hiring. You need to collect data on existing employees and so, knowing what you've read in Shultz, Whitney, and Zickar (2021) what are some steps and principles that will help guide data collection.